CRITICAL METALS HANDBOOK Critical Metals Handbook

Edited by

Gus Gunn British Geological Survey Keyworth Nottingham UK

Published in collaboration with the British Geological Survey

This work is a co-publication between the American Geophysical Union and Wiley This edition first published 2014 © 2014 by John Wiley & Sons, Ltd This work is a co-publication between the American Geophysical Union and Wiley

Registered Office John Wiley & Sons, Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Offices 9600 Garsington Road, Oxford, OX4 2DQ, UK The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK 111 River Street, Hoboken, NJ 07030-5774, USA

For details of our global editorial offices, for customer services and for information about how to apply for permission to reuse the copyright material in this book please see our website at www.wiley.com/wiley-blackwell.

The right of the author to be identified as the author of this work has been asserted in accordance with the UK Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

Designations used by companies to distinguish their products are often claimed as trademarks. All brand names and product names used in this book are trade names, service marks, trademarks or registered trademarks of their respective owners. The publisher is not associated with any product or vendor mentioned in this book.

Limit of Liability/Disclaimer of Warranty: While the publisher and author(s) have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. It is sold on the understanding that the publisher is not engaged in rendering professional services and neither the publisher nor the author shall be liable for damages arising herefrom. If professional advice or other expert assistance is required, the services of a competent professional should be sought.

Library of Congress Cataloging-in-Publication Data Critical metals handbook/edited by Gus Gunn. pages cm Includes bibliographical references and index. ISBN 978-0-470-67171-9 (cloth) 1. Metals–Handbooks, manuals, etc. I. Gunn, Gus, 1951- TA459.C75 2014 669–dc23 2013022393 A catalogue record for this book is available from the British Library.

Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic books.

Cover image: The Spor Mountain open-pit beryllium mine in Utah operated by Materion Brush Natural Resources Inc. (Courtesy of Materion Corp.) Cover design by Steve Thompson

Set in 9/11.5pt Trump Mediaeval by SPi Publisher Services, Pondicherry, India

1 2014 Contents

List of Contributors, xi Policy issues, 38 Acknowledgements, xiii Notes, 39 References, 39 1 Metal resources, use and criticality, 1 T.E. Graedel, Gus Gunn and Luis Tercero 3 Recycling of (critical) metals, 41 Espinoza Christian Hagelüken The geology and technology of metals, 1 Rationale and benefits, 41 Key concepts, 1 The urban mine, 41 Definitions and terminology, 3 Recycling benefits, 43 Will we run out of minerals?, 5 Status and challenges of recycling critical Geological assessment, 6 metals, 45 Considerations of supply and demand, 6 The metals life cycle, 45 Recycling and reuse of metals, 9 Waste and resource legislation, 47 The concept of criticality, 10 The recycling value chain, 47 Assessments of criticality, 11 Recycling challenges, 48 Improving criticality assessment, 14 The seven conditions for effective Implications of criticality for corporate and recycling, 50 governmental policy, 16 Recycling technologies, 51 Outlining this book, 16 Collection and pre-processing, 52 Acknowledgements, 17 Metallurgical recovery, 54 Note, 18 Status of recycling of the EU critical References, 18 metals, 57 The significance of life-cycle 2 The mining industry and the supply of structures, 58 critical minerals, 20 Case study 1: Industrial PGM David Humphreys applications, 59 Suppliers of minerals – miners and Case study 2: Automotive PGM explorers, 21 applications, 60 Industry dynamics, 23 Case study 3: Electronic PGM Constraints on mineral supply response, 27 applications, 60 Natural constraints, 27 Global flows of old products, 60 Economic constraints, 29 Differences in recycling rates and Institutional constraints, 31 pathways for improvement, 61 Critical minerals and the role of China, 34 Conclusion and the way forward, 62 vi Contents

Innovation needs, 62 Uses of beryllium, 100 Resource security as a societal driver for Alloys containing less than 2% beryllium, recycling, 64 especially copper–beryllium, 101 Mining and recycling as complementary Pure beryllium metal and alloys containing systems, 64 over 60% beryllium, 102 Conclusions, 66 Beryllia (BeO) ceramics, 103 Notes, 66 World production, 103 References, 67 World trade, 105 World resources, 106 4 Antimony, 70 Mineralogy of beryllium, 106 Ulrich Schwarz-Schampera Beryllium deposits, 107 Introduction, 70 Pegmatite deposits, 107 Definitions and characteristics, 70 Hydrothermal deposits, 110 Abundance in the Earth, 71 Mining and processing of beryllium, 110 Mineralogy, 71 Beryl ores, 110 Major deposit classes, 72 Bertrandite ores, 110 Gold–antimony (epithermal) deposits, 74 Processing of beryl and bertrandite to Greenstone-hosted quartz-carbonate vein beryllium hydroxide, 111 and carbonate replacement deposits, 77 Production of metal and alloys from Reduced magmatic gold systems, 78 beryllium hydroxide, 113 Extraction methods and processing, 78 Production of beryllium oxide from Mining, 78 beryllium hydroxide, 113 Ore processing, beneficiation and Recycling, 115 conversion to metal, 79 Substitution, 116 Specifications, 82 Environmental aspects, 116 Uses, 82 Prices, 118 Antimony trioxide, 84 Outlook, 118 Sodium antimonate, 84 Note, 119 Other non-metallurgical uses, 85 References, 119 Antimony metal, 85 Recycling, 85 6 Cobalt, 122 Substitution, 86 Stephen Roberts and Gus Gunn Resources and reserves, 86 Introduction, 122 Production, 87 Physical and chemical properties, 122 Projects under development, 90 Distribution and abundance in the World trade, 91 Earth, 122 Prices, 92 Mineralogy, 122 Environmental aspects, 94 Deposit types, 123 Outlook, 95 Hydrothermal deposits, 123 References, 96 Magmatic deposits, 129 Laterites, 130 5 Beryllium, 99 Manganese nodules and cobalt-rich David L. Trueman and Phillip Sabey ferromanganese crusts on the Introduction, 99 seafloor, 132 Properties of beryllium, 99 Extraction, processing and Distribution and abundance in the Earth’s refining, 134 crust, 100 Cobalt from nickel sulfide ores, 134 Contents vii

Cobalt from nickel laterite ores, 134 Outlook, 170 Cobalt from copper–cobalt ores in DRC Acknowledgements, 171 and Zambia, 135 References, 172 Other sources of cobalt, 136 World production and trade, 138 Resources and reserves, 139 8 Germanium, 177 Uses, 140 Frank Melcher and Peter Buchholz Recycling, 142 Introduction, 177 Substitution, 142 Physical and chemical properties, 177 Environmental issues, 143 Distribution and abundance in the Prices, 144 Earth, 177 Outlook, 144 Mineralogy, 178 Acknowledgements, 146 Deposit types, 179 Notes, 146 Accumulation of germanium in sulfide References, 146 deposits, 181 Enrichment of germanium in lignite and 7 Gallium, 150 coal, 185 Thomas Butcher and Extraction methods, processing and Teresa Brown beneficiation, 186 Introduction, 150 Extraction, 186 Physical and chemical Processing, 186 properties, 150 Specifications, 188

Mineralogy and distribution, 150 Germanium tetrachloride, GeCl4, 188

Sources of gallium, 151 Germanium dioxide, GeO2, 188 Bauxite, 151 First reduction metal, 188 Sphalerite (ZnS), 151 Production of zone-refined metal Other geological settings, 152 (‘intrinsic’ metal), 188 Recovery methods and refining, 152 Single crystals, 188 Primary recovery, 152 Uses, 189 Secondary recovery, 153 Recycling, re-use and resource Refining and purification, 155 efficiency, 189 Gallium in GaAs semiconductors, 155 Substitution, 191 Specifications and uses, 157 Environmental aspects of the Gallium metal, 157 life cycle of germanium and its Gallium antimonide, 157 products, 192 Gallium arsenide, 157 Resources and reserves, 192 Gallium chemicals, 159 Production, 194 Gallium nitride, 160 Future supplies, 196 Gallium phosphide, 162 World trade, 197 Photovoltaics, 162 Prices, 197 Substitution, 163 Outlook, 198 Environmental aspects, 163 Supply challenges, 198 World resources and production, 164 Demand drivers, 199 Production in 2010, 164 Supply and demand scenario, 200 Future supplies, 166 Acknowledgments, 200 World trade, 167 Notes, 200 Prices, 167 References, 200 viii Contents

9 Indium, 204 Extraction methods and processing, 236 Ulrich Schwarz-Schampera Specification and uses, 238 Introduction, 204 Recycling, 240 Physical and chemical properties, 204 Substitution, 240 Abundance in the Earth’s crust, 205 Environmental factors, 241 Mineralogy, 205 World resources and production, 241 Major deposit classes, 206 Reserves and resources, 241 Base-metal sulfide deposits, 209 Production, 244 Polymetallic vein-type deposits, 209 Current producers, 245 Base-metal-rich tin–tungsten and skarn Production costs, 248 deposits, 210 Future supplies, 249 Base-metal-rich epithermal deposits, 210 Pegmatite-based projects, 249 Extraction methods and processing, 210 Continental brines, 250 Mining, 210 Geothermal brine, 251 Processing, beneficiation and conversion Oilfield brine, 251 to metal, 212 Hectorite, 252 Indium production from copper ores, 213 Jadarite, 253 Indium production from tin ores, 214 World trade, 253 Indium recovery from secondary Prices, 254 sources, 214 Outlook, 255 Specifications and uses, 214 Acknowledgements, 258 Indium–tin oxide (ITO), 215 Notes, 258 Alloys and solders, 215 References, 258 Semiconductors, 216 Others, 216 11 Magnesium, 261 Resources and reserves, 217 Neale R. Neelameggham and Bob Brown Production, 218 Introduction, 261 Production from residues and scrap, 220 Physical and chemical properties, 261 Projects under development, 221 Distribution and abundance in the Abandoned production, 221 Earth, 262 World trade, 222 Mineralogy, 262 Prices, 223 Deposit types, 263 Recycling and substitution, 224 Extraction methods, processing and Environmental aspects, 225 beneficiation, 263 Outlook, 226 Nineteenth-century magnesium References, 227 production processes, 266 Commercial magnesium production 10 Lithium, 230 processes of the twentieth century, 266 Keith Evans Specifications and uses, 267 Introduction, 230 Recycling, re-use and resource Properties and abundance in the Earth, 230 efficiency, 269 Mineralogy and deposit types, 230 Substitution, 271 Pegmatites, 232 Environmental aspects, 272 Continental brines, 232 Non-greenhouse-gas regulations – Geothermal brines, 234 electrolytic magnesium production, 272 Oilfield brines, 234 Non-greenhouse-gas regulations – thermal Hectorite, 234 magnesium, 273 Jadarite, 235 Greenhouse-gas emission studies, 273 Contents ix

World resources and production, 275 Other hydrothermal veins, 324 Future supplies, 277 Iron oxide–apatite deposits, including World trade, 277 iron-oxide–copper–gold (IOCG) Prices, 277 deposits, 324 Outlook, 279 Placer deposits (mineral sands), 324 References, 281 Ion adsorption deposits, 324 Seafloor deposits, 325 12 Platinum-group metals, 284 By-products, co-products and waste Gus Gunn products, 325 Introduction, 284 Extraction methods, processing and Properties and abundance in the Earth, 284 beneficiation, 325 Mineralogy, 285 Mining, 325 Major deposit classes, 285 Beneficiation, 325 PGM-dominant deposits, 286 Extraction and separation of Nickel–copper-dominant deposits, 292 the REE, 327 Other deposit types, 293 Specifications and uses, 328 Extraction and processing, 294 Recycling, re-use and resource Extraction methods, 294 efficiency, 328 Processing, 294 Substitution, 330 Specifications and uses, 297 Environmental aspects, 330 Uses of platinum, palladium and World resources and production, 331 rhodium, 297 Future supplies, 332 Uses of ruthenium, iridium and World trade, 333 osmium, 300 Prices, 334 Recycling, re-use and resource Outlook, 336 efficiency, 300 Note, 337 Substitution, 301 References, 337 Environmental issues, 301 World resources and production, 302 14 Rhenium, 340 Resources and reserves, 302 Tom A. Millensifer, Dave Sinclair, Production, 302 Ian Jonasson and Anthony Lipmann World trade, 304 Introduction, 340 Prices, 306 Physical and chemical properties, 340 Outlook, 306 Distribution and abundance, 341 Acknowledgements, 309 Mineralogy, 341 Note, 309 Deposit types, 342 References, 310 Porphyry deposits, 342 Vein deposits, 345 13 Rare earth elements, 312 Sediment-hosted copper deposits, 345 Frances Wall Uranium deposits, 346 Introduction, 312 Magmatic nickel–copper–platinum- Physical and chemical properties, 312 group element (PGE) deposits, 346 Distribution and abundance in the Earth’s World resources and production, 346 crust, 313 Future supplies, 348 Mineralogy, 315 Extraction methods, processing and Deposit types, 317 beneficiation, 350 Carbonatite-related REE deposits, 319 Specifications and uses, 352 Alkaline igneous rocks, 323 Recycling and re-use, 354 x Contents

Catalysts, 354 Mineralogy, 386 Superalloys, 355 Deposit types, 386 Substitution, 355 Vein/stockwork deposits, 387 Environmental issues, 356 Skarn deposits, 389 World trade, 356 Disseminated or greisen deposits, 390 Prices, 357 Porphyry deposits, 390 Outlook, 358 Breccia deposits, 391 References, 359 Stratabound deposits, 391 Pegmatite deposits, 392 15 Tantalum and niobium, 361 Pipe deposits, 392 Robert Linnen, David L. Trueman and Hot-spring deposits, 392 Richard Burt Placer deposits, 392 Introduction, 361 Brine/evaporite deposits, 392 Physical and chemical properties, 361 Extraction methods, processing and Distribution and abundance in the beneficiation, 392 Earth, 361 Extraction, 392 Mineralogy, 362 Processing, 393 Deposit types, 363 Specifications and uses, 395 Carbonatite deposits, 363 Specifications, 395 Alkaline to peralkaline granites and Uses, 396 syenites, 367 Recycling, re-use and resource Peraluminous pegmatites, 368 efficiency, 398 Peraluminous granites, 370 Old scrap, 398 Extraction methods and processing, 371 New scrap, 398 Specifications and uses, 374 Unrecovered scrap, 399 Recycling, re-use and resource Recycling methods, 399 efficiency, 375 Substitution, 399 Substitution, 375 Environmental aspects of the life Environmental aspects of niobium and cycle of the metal and its tantalum, 376 products, 399 Geopolitical aspects, 376 World resources and production, 400 World resources and production, 377 Resources and reserves, 400 Future supplies, 379 Production, 401 Prices, 380 Future supplies, 402 Outlook, 381 World trade, 404 Note, 382 Prices, 406 References, 382 Outlook, 406 Acknowledgements, 409 16 Tungsten, 385 References, 409 Teresa Brown and Peter Pitfield Introduction, 385 Appendices, 414 Physical and chemical properties, 385 Glossary of technical terms, 419 Distribution and abundance in the Earth’s Index, 431 crust, 385 Contributors

Bob Brown Keith Evans Publisher Independent Consultant Magnesium Monthly Review San Diego Prattville California Alabama USA USA T.E. Graedel Teresa Brown Center for Industrial Ecology British Geological Survey Yale University Keyworth New Haven Nottingham Connecticut UK USA

Gus Gunn Richard Burt British Geological Survey GraviTa Inc. Keyworth Elora Nottingham Ontario UK Canada Christian Hagelüken Thomas Butcher Director EU Government Affairs Independent Consultant Umicore AG & Co. KG New York Hanau-Wolfgang USA Germany

Peter Buchholz David Humphreys Mineral Resources Agency (DERA) at the Federal Independent Consultant Institute for Geosciences and Natural Resources London (BGR) UK Dienstbereich Berlin Wilhelmstraße 25-30 Ian Jonasson 13593 Berlin-Spandau Formerly research scientist at Geological Survey Germany of Canada xii Contributors

Ottawa Stephen Roberts Ontario School of Ocean and Earth Science Canada National Oceanography Centre University of Southampton Robert Linnen Southampton Robert W. Hodder Chair in Economic UK Geology Department of Earth Sciences Phillip Sabey University of Western Ontario Manager London Technology and Quality Ontario Materion Natural Resources Canada Delta Utah Anthony Lipmann USA Managing Director Lipmann Walton & Co Ltd Ulrich Schwarz-Schampera Walton on Thames Federal Institute for Geosciences and Natural Surrey Resources (BGR) UK Stilleweg Hannover Frank Melcher Germany Federal Institute for Geosciences and Natural Resources (BGR) Dave Sinclair Stilleweg Formerly research scientist at Geological Survey Hannover of Canada Germany Ottawa Ontario Canada Tom A. Millensifer Executive Vice President and Technical Director Luis Tercero Espinoza of Powmet, Inc. Fraunhofer Institute for Systems and Innovation Rockford Research ISI Illinois Karlsruhe USA Germany

Neale R. Neelameggham David L. Trueman ‘Guru’ Consulting Geologist Ind LLC Richmond 9859 Dream Circle British Columbia South Jordan Canada Utah USA Frances Wall Head of Camborne School of Mines and Peter Pitfield Associate Professor of Applied Mineralogy British Geological Survey Camborne School of Mines Keyworth University of Exeter Nottingham Penryn UK UK Acknowledgements

I would like to thank the authors and reviewers gratitude to the Natural Environment Research of each chapter who worked hard to deliver high- Council (NERC) UK for provision of funding quality content suitable for the intended non- through a knowledge exchange grant that allowed specialist readership of this book. I am particularly me to work on this project. Finally, I would like grateful to colleagues at the British Geological to thank my wife, Barbara, for her patience, Survey for their expert contributions: Teresa understanding and support throughout the prepa- Brown for many contributions to editing, map ration of this book. preparation, provision of statistical data, and compilation of appendices and the glossary; Gus Gunn Debbie Rayner for preparing all the diagrams and tables; Ellie Evans for formatting text and refer- British Geological Survey ences; and Chris Wardle for assisting with the Keyworth, Nottingham, UK cover design. I would also like to express my April 2013 1. Metal resources, use and criticality

T.E. GRAEDEL1 , GUS GUNN2 AND LUIS TERCERO ESPINOZA3

1 Center for Industrial Ecology, Yale University, New Haven, Connecticut, USA 2 British Geological Survey, Keyworth, Nottingham, UK 3 Fraunhofer Institute for Systems and Innovation Research ISI, Karlsruhe, Germany

The geology and technology of metals purification, a mineral, often in combination with certain other minerals, is incorporated into a com- Key concepts ponent which is used in a product. It is the need or In a book such as this, which is intended for a desire for the products that generates a demand for broad audience, it is important to discuss some minerals, rather than demand for the mineral itself. key concepts and terminology relating to min- As a result, there is always the possibility of finding erals and metals which, although widely used, are an alternative material to provide the required func- seldom defined. In some cases the meaning may tionality. The only exceptions to this possibility are be obvious, while in others they are anything but nitrogen, phosphate and potash, which are essential obvious. To avoid confusion and misuse, and to life itself and cannot be substituted. to minimise the risks of misunderstanding, we The term ‘mineral’ is used to describe any nat- define in the first part of this chapter certain urally occurring, but non-living, material found fundamental terms that will provide a foundation in, or on, the Earth’s crust for which a use can be for the chapters which follow. found.1 Four principal groups of minerals may be Minerals are essential for economic develop- distinguished according to their main uses: ment, for the functioning of society, and for main- 1. Construction minerals – these comprise bulk taining our quality of life. Everything we have or minerals such as sand and gravel, crushed rock use is ultimately derived from the Earth, produced and clay, which are used for making concrete and either by agricultural activities or by the extrac- bricks to provide foundations and strength in tion of minerals from the crust. Unlike crops, buildings, roads and other infrastructure. They which are grown for the essential purpose of main- are produced in large quantities at low cost from taining life by providing the nutrients we need to extensive deposits that are widely distributed at survive, mankind does not generally need the min- shallow depths in the Earth’s crust. erals themselves. Rather, minerals are extracted 2. Industrial minerals – these are non-metallic for the particular physical and chemical properties minerals that, by virtue of specific chemical or their constituents possess and which are utilised physical properties, are used for particular appli- for specific purposes in a huge range of goods and cations in a wide range of industrial and consumer products. Following some form of processing and products. There are numerous industrial minerals

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 2 t.e. graedel, gus gunn and luis tercero espinoza but the most widely used include salt, gypsum, that are common in the crust, such as magnesium, fluorspar, and kaolin. They tend to occur in large aluminium and titanium, occur in forms that quantities but only at relatively few locations. need a high input of energy to separate them from They generally require specialist processing in their ores, thus making them relatively expensive. their production and consequently they are It is also important to note that the localised relatively expensive. concentrations of metals that can be exploited 3. Energy minerals – these are minerals such as economically result from unusual geological oil, gas and coal that are used to generate energy processes. Consequently, the distribution of that is captured when they are burned. They are economic deposits in the Earth’s crust is highly used in the production of electricity, in fuels for dispersed, with some regions richly endowed in transportation and heating, and also in the metals and others largely devoid of them. manufacture of plastics. Coal is relatively easy to Furthermore, our knowledge of the processes that find and cheap to extract; in contrast, oil and gas lead to the concentration of particular metals are generally difficult to find and extract and, in the Earth’s crust varies widely. For metals that therefore, command high prices. are used in large quantities, such as copper and 4. Metals – metals are distinguished by distinc- zinc, we have a reasonably good idea of where and tive chemical and physical properties, such as how to locate new deposits. However, for many high electrical and thermal conductivity, mallea- of the scarcer metals, especially those that have bility, ductility and the ability to form alloys. been brought into wide use relatively recently, They are exploited for a multitude of purposes and information on their occurrence, concentration some, such as iron, aluminium and copper, are and processing is generally very limited. used in huge quantities. Other metals with fewer It is a complex and expensive process to prove or more specialised applications, such as platinum, economic viability once an unusual enrichment of indium and cobalt, are used in much smaller a potentially useful mineral or assemblage of quantities, ranging from tens to hundreds or thou- minerals, commonly referred to as a ‘mineral sands of tonnes per year. Economic deposits of occurrence’, is discovered. This involves determi- metals are rare and difficult to locate. The metal- nation of the quantity of mineral present and the bearing ores are expensive to mine and to process, assessment of the optimum methods for mining and consequently metals command a high price. and processing the ore. Apart from geological Another term in common usage is ‘mineral processes that determine the physical availability commodity’ which is used to refer to any mineral of a metal there are a host of other factors that raw material that can currently be extracted from influence access to the resources in the ground – the Earth for a profit. cheap labour or cheap power may confer a The abundance of individual metals in the competitive advantage to a particular country or Earth’s crust varies greatly (Figure 1.1) and influ- region while, on the other hand, government regu- ences the costs involved in locating, mining lation, fiscal and administrative requirements, or and preparing the metals for use. Some of the social and cultural constraints may restrict or major industrial metals, like iron, aluminium prevent access to potentially valuable deposits. and calcium, have crustal abundances similar to The timescale from discovery of a mineral the main rock-forming elements, such as oxygen, occurrence to mine production is generally a long silicon and calcium, and are several orders of one. It commonly takes more than ten years to magnitude more abundant than many of the evaluate the mineral resource in the ground, to widely used base metals such as copper, lead raise the funds to build a mine, to acquire the and zinc. Many others, such as the precious necessary regulatory approvals and to secure the metals gold and platinum, are considerably rarer. trust and cooperation of the local communities. However, crustal abundance is only one factor Once these are in place, and provided that favour- that influences production costs. Some metals able economic conditions prevail, the mine and Metal resources, use and criticality 3

109

Rock-forming elements

O Si 106 Ai H Na Ca K Fe Mg atoms of Si Ti 6 C F P Mn 103 S Sr Li Zr Ci Zn B N Cu Ce V Rb Cr Nd Be Ga Nb Pb Sc Co Ni Y Sn La SmGdDyEr As Pr Yb 0 Ge Cs 10 Br Mo W Sb Ti Cd Eu Ho I Tb Lu Ag Tm Se In Hg B

–3 Ru Te Au Abundance, atoms of elements per 10 10 Pd Re Pt Major industrial metals in red Rh Os Precious metals in purple Rare earth elements in blue Rarest ‘metals’ Ir

10–6 0 102030405060708090 Atomic number, Z

Figure 1.1 The abundance of the chemical elements in the Earth’s upper continental crust as a function of atomic number. Many of the elements may be classified into partially overlapping categories. (Modified from USGS, 2002.)

supporting infrastructure can be built and min- the size and duration of these investments it is eral extraction can commence. essential that all parties – the mining company, investors, local communities, governments and regulators – ‘speak the same language’ and fully Definitions and terminology understand their obligations and expectations The costs involved in bringing a new mine into throughout the life of the mine, from construction production today commonly amount to hun- to operation, closure and site rehabilitation. dreds of millions of dollars or, in the case of Without effective communication, based on a large new mine on a greenfield site, more than clear unambiguous terminology, such under- a billion dollars. A metal mine typically operates standing can never be attained and problems may for a minimum period of a decade although, well arise at some stage. depending on economic and other circumstances, The first steps in determining the economic it may continue for more than 100 years. Given viability of a mineral deposit are the exploration 4 t.e. graedel, gus gunn and luis tercero espinoza and resource assessment stages which involve paring reports on exploration results, mineral drilling and detailed sampling to determine the resources and ore reserves. quantity of material present and its quality – or, in the case of a metallic mineral deposit, its grade, Resources and reserves which is the percentage of metal that the rock contains. The consistent and correct use of termi- The key elements of the reporting codes are the nology is essential for the reporting and assessment terms ‘resources’ and ‘reserves’, which are of exploration results and to underpin sound frequently confused and/or used incorrectly. decision making. Without this, discrimination They are, in fact, fundamental to the distinction between genuinely economic deposits and those between a mineral deposit that is currently of marginal or unproven economic significance is economic (reserves) and another which may impossible. become economic in the future (resources). The assessment is, therefore, based on a A mineral ‘resource’ is a natural concentration system of resource classification the main of minerals or a body of rock that is, or may objective of which is to establish the quantities of become, of potential economic interest as a basis minerals likely to be available in the future. for the extraction of a commodity. A resource has Many governments now require that resources physical and/or chemical properties that makes it and reserves are reported according to interna- suitable for specific uses and is present in tionally accepted codes in countries where the sufficient quantities to be of intrinsic economic company’s stock is listed. Adherence to such interest. To provide more information about the reporting standards ensures full and transparent level of assurance, resources are divided into dif- disclosure of all material facts and is intended to ferent categories which, in the JORC code, are provide all parties with reliable information on referred to as measured, indicated and inferred which to base investment decisions. Such codes resources, reflecting decreasing level of geolog- include the Joint Ore Reserves Committee ical knowledge and hence decreasing confidence (JORC) code in Australia and the Canadian in their existence. Institute of Mining, Metallurgy and Petroleum It is important to note that identified resources (CIM) reporting standard which is referred to as do not represent all the mineral resources present in National Instrument (NI) 43-101. Following an the Earth, a quantity that is sometimes referred to era of industry self-regulation, these codes were as the ‘resource base.’ In addition to identified developed in response to scandals in Australia resources, there are resources that are undiscovered and Canada where many people were misled by or unidentified (Figure 1.2). Undiscovered resources speculation and rumour leading to unfounded may be divided into hypothetical and speculative spectacular rises in share prices and, soon after, categories. Hypothetical resources are those which rapid falls. In the short term these led to huge may reasonably be expected to occur in deposits financial losses and, in the longer term and more similar to those known in a particular area under significantly, to a prolonged loss of investor similar geological conditions. Speculative resources confidence in the mining industry. Accordingly are those which may be present either in known these, and other codes, were developed to set deposit types in areas with favourable geological minimum standards of reporting of exploration settings but where no discoveries have yet been results, mineral resources and ore reserves. They made or in new types of deposit whose economic provide a mandatory system of classification of potential has not yet been recognised. tonnage and grade estimates according to geolog- A mineral ‘reserve’ is that part of a mineral ical confidence and technical/economic consider- resource that has been fully geologically evaluated ations. They require public reports to be prepared and is commercially and legally mineable. by appropriately qualified persons and provide Mineral reserves are divided in order of increasing guidance on the criteria to be used when pre- confidence into probable and proved categories. Metal resources, use and criticality 5

Undiscovered resources (hypothetical and speculative) Identified resources Reserves

Figure 1.2 Schematic representation of the relative size of the quantities represented by the terms resources and reserves. Reserves generally represent only a tiny fraction of resources.

The ultimate fate of a mineral reserve is either to USGS abandoned use of the reserve base category be physically worked out or to be made non-via- in 2010 (USGS, 2010). ble, either temporarily or permanently, by a change in circumstances (most often economic, regulatory or social). So-called ‘modifying factors’ Will we run out of minerals? (economic, mining, metallurgical, marketing, social, environmental, legal and governmental) We are using minerals and metals in greater quan- contribute to the viability of a mineral deposit tities than ever before. Since 1900 the mine pro- and determine whether or not it will be exploited. duction of many metals has grown by one, two, Figure 1.2 is a simple graphical depiction of the or even three orders of magnitude (Graedel and relative sizes of the quantities represented by the Erdmann, 2012). For some metals, especially terms undiscovered and identified resources and those used in high-tech applications, the rate of reserves. If this figure were drawn to scale the use has increased particularly strongly in recent circle representing the reserves would be very decades, with more than 80 per cent of the total small relative to the resources because reserves are global cumulative production of platinum-group only a tiny fraction of the resources of any metals (PGM), indium, gallium and rare earth mineral. elements (REE) having taken place since 1980 The term ‘reserve base’ was also formerly used (Hagelüken et al., 2012). We are also using a when discussing mineral resources and mineral greater variety of metals than ever before. For availability. This term, introduced by the United example, turbine blade alloys and coatings make States Geological Survey (USGS) and the United use of more than a dozen metals and high-level States Bureau of Mines (USBM) in 1980, was used technological products, such as those used in as an estimate of the size of the mineral reserve medicine, incorporate more than 70 metals. In and those parts of the resources that had reason- the quest for improved performance, microchips able potential for becoming economic within now use about 60 metals, whereas in the 1980s planning horizons beyond those that assume and 1990s only about 20 were commonly incor- proven technology and current economics. porated into these devices. However, the reserve base estimates were gener- The main reasons for these changes are ally based on expert opinion rather than on data increased global population and the spread of and were not readily defensible, especially at prosperity across the world. New technologies, times of rapid growth in mineral demand and such as those needed for modern communication consequent massive increases in exploration and computing and to produce clean energy, also expenditure, as happened during much of the first require considerable quantities of numerous decade of the 21st century. Consequently, the metals. In the light of these trends it has become 6 t.e. graedel, gus gunn and luis tercero espinoza important to ask if we can continue to provide in South America (Cunningham et al., 2008). the minerals required to meet this demand, and This study concluded that there may be a huge also to question whether our resources will amount of copper to be discovered to a depth of ultimately be exhausted. one kilometre below the Earth’s surface in the Andes, equivalent to 1.3 times as much as has already been found in porphyry copper deposits in Geological assessment this region. Estimates derived in this way are In general, our knowledge of the geology and very useful, not only to mining companies but industrial uses of those metals used in greatest also to planners, economists, governments and amounts, such as iron, aluminium and copper, is regulators. The approach also has real practical extensive. There is a reasonably good idea of the value because it assesses the availability of geological processes responsible for the formation resources of a type that are well known and can of economic deposits of these metals, and conse- be mined and processed economically with quently how to identify the best places to look for current technology. However, this method is additional resources. Experience over many dependent on the availability of high-quality geo- decades and centuries has taught geologists and logical data and on a sound understanding of the mining engineers how to find, extract and process target mineral deposit class. Unfortunately, such these metals to provide the goods and services we geological information is not generally available need. As a result it has been possible to find new and knowledge of many mineral deposit classes deposits to replace those that are worked out, and that may contribute to global metal production is economic development has not been constrained poor. Consequently, this approach is not likely to by metal scarcity. yield reliable estimates of global metal avail- However, reliable estimates of the total ability in the near future; rather, its application amount of any metal that may be available in the will be restricted to a particular deposit type Earth’s crust are not in place. Various authors within specific areas. Of course, rather than hav- have calculated the maximum quantities present ing accurate estimates of what might ultimately based on estimates of mean elemental crustal be available to us, what really matters is how can concentrations and have concluded that the we be sure that we have enough metal to meet amounts potentially available are huge (e.g. our needs and that we will not run out in the Cathles, 2010). Although these estimates provide future as demand grows. upper limits to availability, they have little real practical value because they take no account of Considerations of supply and demand the costs, economic, environmental or social, that would be involved in extracting metals from Much of the recent debate has focused on the ade- these sources. Some researchers have adopted a quacy of mineral deposits to meet future demand different, ‘bottom up’ approach based on probabi- rather than on the political and economic barriers. listic estimates of the crustal endowment of Several authors have concluded that mineral scar- particular metals in specific deposit types. city and, ultimately, depletion are unavoidable Perhaps the best known and largest study of (Ragnarsdottir, 2008; Cohen, 2007). Some have made this type is the United States Geological Survey’s alarmist forecasts that suggest that for some min- Global Mineral Resource Assessment Project, erals and metals depletion may occur over relatively which is being undertaken to assess the world’s short timescales of a few decades or even years. undiscovered non-fuel mineral resources. One However, these predictions are based on ‘static life- of the first studies completed was a quanti- times’ derived from existing known resources or tative mineral resource assessment of copper, reserves divided by current or projected future molybdenum, gold and silver in undiscovered demand (Cohen, 2007; Gilbert, 2009; Sverdrup et al., porphyry deposits of the Andean mountain belt 2009). These forecasts fail to recognise that resources Metal resources, use and criticality 7

(a) Copper (b) Nickel

18 1.8 2011: 16.2 Mt 2010: 1.54 Mt 16 1.6 14 1.4 12 1.2 10 1 8 0.8 6 1982: 8 Mt 0.6 1987: 0.82 Mt 4 0.4 2 0.2 Mine production (Mt) Mine production (Mt) 0 0 1980 1985 1990 1995 2000 2005 2010 1985 1990 1995 2000 2005 2010

50 70 1987: 62 years 1982: 44 years 2011: 43 years 60 2010: 50 years 40 50 30 40

20 30 20 10 10 Static lifetime (years) Static lifetime (years) 0 0 1980 1985 1990 1995 2000 2005 2010 1985 1990 1995 2000 2005 2010

Figure 1.3 Despite escalating global production of metals, reserves have continually been replenished. These graphs show that static lifetimes (number of years’ supply remaining equals reserves divided by annual production), in this case of (a) copper and (b) nickel, are extended ahead of production (Mt, million tonnes, metal content). (Mine production data from BGS World Mineral Statistics Database; reserve data from USGS Mineral Commodity Summaries, 2012 and earlier editions.) and reserves are neither well known nor fixed. global copper reserves in the early 1930s were Reserves are economic entities that depend on reported to be about 100 million tonnes, thought scientific knowledge of minerals and on the price of at the time to be sufficient for about 80 years. the target metal or mineral. As our scientific under- However, in 2010 the USGS reported copper standing has improved, reserves have continually reserves of 540 million tonnes (USGS, 2010) and been replenished through new discoveries, by in 2011 the estimate was again revised upwards improved mining and processing technology, and by to 630 million tonnes, an increase of more than improved access to deposits. Furthermore, market 16 per cent in a single year (USGS, 2011). Similar mechanisms help to overcome supply shortages for trends can be seen in the global reserve levels for major metals – if prices rise, then reserves will extend some minor metals. For example, tungsten to include lower-grade ore; if prices fall then they reserves grew by more than 50 per cent between will contract to include higher-grade material. High 2000 and 2011, while reserves of REE grew by 25 prices will also stimulate increased substitution, per cent between 2008 and 2011. It is clear, there- recycling and resource efficiency and thus will con- fore, that reserve estimates are unreliable indica- tribute to improved security of supply. tors of the long-term availability of metals as Crowson (2011) has discussed changes in their definition depends on current science, tech- reserve levels of some major industrial metals nology and economics (Figure 1.3). since 1930. He showed that, despite escalating A type of scarcity referred to as ‘technical scar- production, reserve levels have actually grown city’ or ‘structural scarcity’ presents a particular over time and outpaced production. For example, challenge and may be difficult and expensive to 8 t.e. graedel, gus gunn and luis tercero espinoza

Table 1.1 By-product metals derived from the production of selected major industrial metals (top row, bold). Those metals shown in italics may also be produced from their own ores. (PGM, platinum-group metals; REE, rare earth elements.)

Copper Zinc Tin Nickel Platinum Aluminium Iron Lead

Cobalt Indium Niobium Cobalt Palladium Gallium REE Antimony Molybdenum Germanium Tantalum PGM Rhodium Niobium Bismuth PGM Cadmium Indium Scandium Ruthenium Vanadium Thallium Rhenium Osmium Tellurium Iridium Selenium Arsenic

alleviate. Technical scarcity applies chiefly to a In some situations certain elements which are range of rare metals used mostly in high-tech normally mined as by-products may also be applications. Many of these are not mined on their mined in their own right if their concentrations own; rather they are by-products of the mining of and mode of occurrence allow it. For example, the ores of the more common and widely used cobalt is generally a by-product of copper mining, metals, such as aluminium, copper, lead and zinc but, exceptionally, it can be mined on its own. (Table 1.1). These by-product or companion metals Similarly, the PGM are commonly by-products of are present as trace constituents in the ores of the nickel mining but most production is from host metals and, under favourable economic condi- PGM-only mines in South Africa. tions, they may be extracted from these ores, or In some instances groups of metals have to be from concentrates and slags derived from them. For produced together as coupled elements because example, indium and germanium are chiefly they are chemically very similar and cannot be by-products from zinc production, while tellurium easily separated from the minerals in which they is mainly a by-product of copper mining. However, occur. The best examples of coupled elements are the low concentration of the companion metal in the platinum-group metals (PGM: rhodium, ruthe- the host ores means that there is little economic nium, palladium, osmium, iridium and platinum) incentive to increase production at times of short- and the rare earth elements (REE comprising 15 age. For example, only about 25–30 per cent of the lanthanides, scandium, and yttrium). In these cases 1000 tonnes of indium that is potentially available there is no major carrier metal, but normally one or globally each year from mining indium-rich zinc two of the group determines production levels and ores is actually recovered. The rest ends up in the economic viability of the extractive operations. wastes because it is not economic to install the In the case of the PGM, platinum is commonly the additional indium extraction capacity at zinc refin- main driver for production, with palladium, iridium eries or because the efficiency of the indium and ruthenium derived as by-products. recovery is poor (Mikolajczak and Harrower, 2012). The petroleum industry’s debate about ‘peak It is therefore difficult to predict the capacity of oil’ has been extended to the non-fuel minerals the supply chain to meet increased demand for the industry. The peak concept was developed from by-product. If the high level of by-product demand the work of oil geologist Hubbert in the 1950s is expected to be sustained, for example because of who predicted, on the basis of the existence of a a particular well-established technological require- well-known ‘ultimately recoverable reserve’, that ment such as indium in flat-panel displays and oil production in the USA would peak about 1970 portable electronic devices, then a good economic and then enter a terminal decline (Hubbert, 1956). case for increased indium production can be made. Others extended this approach to predict that Metal resources, use and criticality 9 global oil production would peak in 2000. These potentially serious implications for policy making predictions proved largely correct, although and investment decisions. global oil production peaked a few years later than forecast. Hubbert’s model is based on sym- metrical (bell-shaped) curves, with the produc- Recycling and reuse of metals tion peak occurring when approximately half of the extractable resource has been extracted. More Modern technology is largely designed around recently various authors have advocated ‘peak the use of virgin materials extracted from geolog- metals’ as a tool for understanding future trends ical sources. It is increasingly apparent, however, in the production of metals (Bardi and Pagani, that materials that have been incorporated into 2007; Giurco et al., 2010). Bardi and Pagani (2007) products no longer in use (secondary materials, examined global production data for 57 minerals scrap) can provide a valuable supplement to vir- and concluded that 11 of these had clearly peaked gin stocks. This reuse will generally require that and several others were approaching peak the secondary materials are comparable in quality production. to those generated from the virgin stocks. The application of the peak concept to metals Primary metals are produced through a production has been criticised by various authors sequence of actions following their discovery and who have questioned both the validity of the evaluation: mining the ore, milling it (crushing the assumptions underlying the model when applied rock and separating the metal-containing minerals to metals and also the failure to address the real from the waste material), smelting (to transform causes of variations in production and consumption the metal oxides and sulfides into impure metal), in the mineral markets (Crowson, 2011; Ericsson and refining (to purify the smelted material). None and Söderholm, 2012). Records from the last 200 of these processes is perfect, so metal is lost at each years show that the prices of major metals are stage. The sequence for secondary metals has some cyclical, with intermittent peaks and troughs of the same characteristics. It begins with collec- closely linked to economic cycles. Declining pro- tion of the discards, separation of the metals in the duction is generally driven by falling demand rather discards, sorting of the separated metals, and than by declining resources or lack of resource dis- smelting or similar metallurgical processes to covery. At times of increasing scarcity the price of transform the results of the previous processes minerals will increase, which, in turn, will tend to into metals pure enough for reuse. As with primary stimulate increased substitution and recycling processes, metal is lost at each stage. and encourage investments in new capacity and In a world of increasing resource use, secondary more exploration. High prices may also lead to supplies of metals will, however, be insufficient more focus on improving current exploration and to meet overall demand. Even if all the metals production technologies. Historically, technolog- incorporated into products were collected and recy- ical innovation has often succeeded in developing cled with 100 per cent efficiency at the end of their new lower-cost methods for finding and extracting useful life, there would inevitably be a shortfall in mineral commodities. supply which would have to be filled through It is concluded, therefore, that the peak con- production from primary resources (Figure 1.4). cept is not valid for modelling mineral resource Nonetheless, secondary supplies provide a depletion and cannot provide a reliable guide to resource supplement that generally requires less future metal production trends. Furthermore, energy than primary metals (often much less), estimates of reserves and resources, and the static and has generally lower environmental impacts. lifetime of mineral raw materials calculated from Through recycling activities, most metals have them, should not be used in the assessment of the potential for reuse over and over again, but future mineral availability as they are highly only if product designers enable recycling by judi- likely to give rise to erroneous conclusions with cious choice of metal combinations and assembly 10 t.e. graedel, gus gunn and luis tercero espinoza Demand Figure 1.4 When demand for a D 2 commodity increases over time recycling alone cannot meet the higher demand. At Minimum gap to be filled the beginning of the lifetime of a product, by primary resources today T1, demand is at a level D1. At the end of D its lifetime, T , demand has risen to D 1 2 2 but the amount potentially available

from recycling will be D1. The gap in T T 1 Product 2 Time supply (D2–D1) can only be met from lifetime primary resources. practices, if governments and individuals opti- This book deals with certain metals that have mise product collection at end of life, and if recy- become increasingly important in recent years for a cling technology is able to produce secondary variety of purposes and for which demand is rapidly material whose quality is sufficiently high to increasing. For example, as technology has pro- enable reuse without downgrading. Certain ele- gressed so new markets for metals, which were pre- ments in specific applications are used in a highly viously little used, have arisen or, in some instances, dispersed state and cannot be recovered. For greatly expanded in response to society’s needs. Of example, potassium, phosphate and nitrogen in particular importance are so-called ‘green’ technol- fertilisers are dissipated in use, as are metals like ogies, especially as the major world economies zinc and magnesium, which are also used for attempt to shift from carbon-based energy systems. agricultural purposes. Other unrecoverable losses What is meant by the ‘criticality’ of metals? of metals include titanium in paint pigments, and Dictionary definitions (e.g. “the quality, state, or platinum and ruthenium used in very thin layers degree of being of highest importance”) suggest in hard-disc drives. A wide range of other metals that the term relates to ‘essential’ or nearly so. In is also lost due to wear and corrosion in use. the first few years of the 21st century the label was Recycling of metals and minerals and the applied to metals, and particularly to the possi- challenges associated with improving its uptake bility that some metals might become scarce and efficiency are discussed in more detail in enough to cease being routinely available to tech- Chapter 3 of this book. nology. This is more than an idle concern: there have been a number of instances in the past few decades when war, technological change or geopo- The concept of criticality litical decisions have resulted in temporary short- ages. We ask a more fundamental question here, Without minerals we would not enjoy the lifestyle however: might some metals be particularly sus- that we enjoy in the West and to which many ceptible to long-term scarcity regardless of the others aspire. Without the continued development reason or reasons? If we entertain this possibility, in the twentieth century of technology for min- could we forecast this situation far enough in eral exploration, processing and manufacturing advance to mitigate some of its most challenging we would not benefit from cheap and reliable implications? Or, to simplify, can we determine a products ranging from aeroplanes and cars, to metal’s criticality and turn that knowledge to use? computers, mobile phones and a panoply of other The first complexity to point out is that criti- portable personal electronic products that are cur- cality is a matter of degree, not of state. Figure 1.5 rently proliferating, such as tablet computers. makes this point graphically: criticality is not the Metal resources, use and criticality 11

(a) (b) ty cali riti c No Yes ng si 40 60 a re c In

Not critical Critical

20 80

0 100

Figure 1.5 Criticality is not simply a designation of ‘critical’ or ‘not critical’ as indicated in (a); rather it is a matter of degree, as indicated in (b) where an arbitrary ‘criticality level’ (here 70) is defined. position of a switch, such that a metal is either does not utilise gold, nor the jeweller’s copper critical or non-critical (Figure 1.5a), but rather a (i.e. for those users, either gold or copper cannot position on a dial where any position above a be deemed a critical metal). In sum, the degree of certain level could arbitrarily be designated as the criticality of a metal is related to the physical dividing line between critical or not. The next and chemical properties of the metal itself, to a complexity concerns the metric itself: what is the number of factors influencing supply and demand, dial measuring? As we will see, methodologies for and to the questioners themselves. determining degrees of criticality can be very com- plex and are generally multi-dimensional, so the Assessments of criticality arrow in Figure 1.5b points to a location in two-dimensional or three-dimensional space. This As mentioned earlier in this chapter, concerns reflects the fact that scarcity may be a consequence about the possible scarcity of natural resources are of geological factors, economic factors, technology a recurring theme in history. The main focus has evolution, potential for substitutes, environmental been on the potential impacts of supply disrup- impacts, and many more. This complexity has tions to the economy, especially where it is spawned a variety of analytical approaches and, dependent on imported materials. In the minerals unfortunately for those wishing to employ the industry finding rapid solutions is particularly information from those studies, a variety of results. challenging because of the high costs and long lead It is also important to point out that criticality times required to make new mineral supplies is not a property whose determination is iden- available. Buijs and Sievers (2012) noted that the tical to all potential users. For a company whose criticality studies conducted in the USA and EU in business is making electrical cables, copper is the 1970s and 1980s adopted basically similar essential. For a maker of fine jewellery, gold is approaches to those used today to identify critical essential. However, the cable-maker’s business raw materials. Nevertheless, the critical minerals 12 t.e. graedel, gus gunn and luis tercero espinoza

to assess it was that of a committee of the US

High National Research Council (2008). The committee proposed that criticality was a two-parameter vari- 2 1 able, one parameter being supply risk and the other the impact of supply disruption. Figure 1.6 shows the concept, in which an element falling in the area 1 quadrant was deemed more critical than those in other areas of the diagram. Further, each of those parameters in turn was regarded as some sort of aggregation of a number of contributory metrics: the committee suggested geological availability, political factors, technological capacity and other Impact of supply restriction 4 3 factors for Supply Risk, and substitutability, impor- tance of applications and other factors for Impact of Supply Restriction. The committee did not select Low specific components nor delineate the method- Low High ology in detail, but did make rough criticality Risk of supply restriction approximations for 11 metals and groups of metals. Those showed the most critical to be rhodium, the Figure 1.6 The criticality matrix originated by the U.S. least copper, and the others at various locations in National Research Council (2008), as revised by between. The committee emphasised that the eval- Graedel and Allenby (2010). Metals falling in sector 1 uations were largely to demonstrate the concept, are more critical than those falling in sectors 2 or 3, not in any way to be definitive. and much more critical than those falling in sector 4. A second important evaluation was initiated by the European Commission (EC) in 2009, with identified in those earlier assessments differ from a report published in the following year (European those now classified as critical, thus highlighting Commission, 2010). The EC working group that such studies provide only a ‘snapshot’ of a retained the two-axis concept, with supply risk dynamic system and have little predictive value. being one of the parameters, but defined the sec- However, Buijs and Sievers also observe that the ond axis on the basis of the potential economic analysis conducted in the earlier studies and the impact of supply disruption on European industry. solutions proposed at that time are similar to those Supply risk was further defined as an aggregate of of today. Then, as now, it was concluded that, three parameters: the political stability of the although geological scarcity was highly unlikely, producing countries, the potential to substitute the main supply risks were companion/host rela- the metal being evaluated, and the extent to tionships, import dependence, the concentration which the metals are recycled. The evaluation of production in a small number of politically also included environmental risks as a separate unstable countries, and increased resource nation- concern, and the classification ‘critical’ was alism in various forms as the governments in pro- assigned to a raw material if a certain threshold ducing countries seek to derive greater benefits for both economic importance and at least one of from the exploitation of indigenous resources. the complementary metrics was exceeded. In The measures proposed to alleviate future supply practice, the metals ranked of most environ- shortages include stockpiling of raw materials, mental concern were already designated as criti- establishment of long-term supply contracts and cal based on other factors. exploitation of indigenous resources. The EC working group evaluated forty one The first recent attempt to define metal criti- metals and minerals. The result is shown cality and suggest metrics that might be employed in Figure 1.7. Arbitrarily drawing lines of Metal resources, use and criticality 13

5.0 REE 4.5

4.0 PGM 3.5

3.0 Ge Nb Sb Mg 2.5 Ga

Supply risk 2.0 In W Bt Fl 1.5 Be Gr Co Ta 1.0 Re Mt Cr Li Ls V Bo Te 0.5 Dt Pe Mo Zn Mn Gy Fp Bn Sl Fe Al Ni Tc Cy Ag Ti Cu Bx 0.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Economic importance

Non-metals Metals described in this book Other metals

Ag, silver; Al, aluminium; Be, beryllium; Bt, barytes; Bx, bauxite; Bn, bentonite; Bo, borate; Co, cobalt; Cr, chromium; Cu, copper; Cy, clays; Dt, diatomite; Fe, iron; Fp, feldspar; Fl, fluorspar; Ga, gallium; Ge, Germanium; Gr, graphite; Gy, gypsum; In, indium; Li, lithium; Ls, limestone; Mg, magnesium; Mn, manganese; Mo, molybdenum; Mt, magnesite; Nb, niobium; Ni, nickel; Pe, perlite; PGM, platinum-group metals; Re, rhenium; REE, rare earth elements; Sb, antimony; Sl, silica; Ta, tantalum; Tc, talc; Te, tellurium; Ti, titanium; V, vanadium; W, tungsten; Zn, zinc.

Figure 1.7 The criticality matrix of the European Commission (2010). The horizontal axis reflects the economic impact of supply restriction on a broad group of European industries; supply risk constitutes the vertical axis. The 14 raw materials falling within the top-right cluster are regarded as critical to the European Union. (Modified from European Commission, 2010.)

demarcation, the working group designated ten Buijs and Sievers (2012). They found that the metals as critical: antimony, beryllium, cobalt, great differences in methodology, the sets of gallium, germanium, indium, magnesium, nio- metals reviewed, and selection criteria render it bium, tantalum and tungsten, as well as two less than convincing at present to single out groups of metals, the rare earth elements and the some metals for special attention while neglect- platinum-group metals. ing others, as distinctions between critical and There have been other efforts to designate non-critical metals are too complex to be easily metals as critical, including those of Morley and resolved. It is clear that, although this topic is Etherley (2008), the U.S. Department of Energy generating a high level of interest from govern- (2010 and 2011) and the Joint Research Council ments and corporations throughout the world, of the EC (JRC, 2011). These, together with the methodology is immature and the results the National Research Council and European are not necessarily helpful to all parties whose Commission studies and others, have been ultimate aim is to secure future supplies of min- reviewed by Erdmann and Graedel (2011) and erals (Buijs et al., 2012). 14 t.e. graedel, gus gunn and luis tercero espinoza

The availability of suitable high-quality data used in high-tech application for which little data is a serious issue that can impact on the results are available. Supply risk is estimated for both of the criticality assessment. For example, in the medium term (5–10 years, with corporations the EU study (EC, 2010) the diagram (Figure 1.7) and governments in mind) and for the longer suggested that the highest level of concern term (a few decades, of interest to planners and should be for the rare earth and platinum-group the academic community concerned with sus- elements. These groupings turn out not to be tainable resource management). Environmental particularly helpful so far as criticality is Implications address both issues of toxicity and concerned, in view of the fact that some ele- of energy use (and thus climate impact), and is of ments in each group (e.g. platinum, neo- particular interest to designers, governments dymium) are widely used and have a possible and non-government agencies. Vulnerability claim to criticality, while others in each group to Supply Restriction (VSR) varies according to (e.g. osmium, holmium) are rarely employed organisational level: a particular metal may and clearly not critical. This situation arose be crucial to the products or operations of one because some data used in the analysis was company but of little or no importance to another. available only for the element groups and not An example of the results of this approach is for individual PGM and REE. Similarly, for shown in Figure 1.8. some minor metals trade data is not available in sufficient detail to allow accurate definition Improving criticality assessment of global import and export patterns. Given the inherent complexities and the While it is clear that no single criticality data shortcomings it is inevitable that such assessment is universally applicable, shortlists criticality assessments will not deliver results of critical raw materials have an important of universal application, and also that they may role to play in warning decision makers in fail to identify potential problems. They may government and industry about current issues of suggest that certain materials are at risk when, concern and possible impacts on security of in fact, market forces may be able to solve the supply in the short term. Development of a problems in the short or medium term. They longer-term capacity to explore potential supply may also produce false negatives whereby issues is the ultimate goal of such assessments, supplies of some materials are incorrectly iden- but there are many intricacies to address before tified as secure. However, as these limitations this can be achieved. Key requirements include have come to be appreciated and while interest the necessity to analyse individual metals and in criticality remains at a high level, so there underlying issues in more detail, to acquire have been continual refinements of the meth- better data, and to analyse trends and patterns of odology, adapting it for particular purposes, dif- future demand. ferent organisational levels (corporate, national One of the challenges of providing perspec- and global), and over different timescales. tive on the long-term supply and demand More recently, Graedel and co-workers at Yale of metals is that their uses evolve in ways University have proposed a comprehensive and not always predictable. Nonetheless, various flexible methodology for the determination of studies have attempted to consider technology metal criticality by enhancing the US National scenarios considering how wind power, photo- Research Council approach (Graedel et al., 2012; voltaic solar power, automotive fuel cells, and NRC, 2008). This method involves three dimen- other technologies could develop in the next sions: Supply Risk, Environmental Implications few decades (e.g. European Commission, 2003; and Vulnerability to Supply Restriction. It uses a IEA, 2008; Shell, 2008). In a typical study, Kleijn combination of data and expert judgement, the and van der Voet (2010) evaluated the resource latter especially important for speciality metals requirements needed to meet several technology Metal resources, use and criticality 15

Copper

Arsenic

Selenium

Silver

Tellurium

Gold

Mean 60 80 100

100 0 80

60 Vulnerability to supply restriction 04 40

20 Enviromental implications

02 0 0 20 40 60 80 100 Supply risk

Figure 1.8 The criticality of the geological copper group of metals as determined by the Yale University methodology. (After Nassar et al., 2012.)

projections. They found that substantial deploy- Angerer et al., 2009) and most have focused on ment of wind turbines, photovoltaic solar cells, material requirements for the clean energy sector hybrid vehicles, enhanced transmission grids, (e.g. U.S. Department of Energy, 2010 and 2011; among others, have a strong potential to be JRC, 2011). In general, the inclusion of projections restricted because of the large quantities of in criticality assessment will be a step forward metal that would be required. Their study indi- because it will reduce reliance on the future cates that future technology planning will need validity of indicators compiled from historic and to have at its centre an assessment of the current data. However, projections inevitably impacts on metal demand, especially for the represent a present view of future market states scarce metals that are acquired as by-products. and, though useful for orientation, cannot be Very few studies have attempted to predict relied upon to provide accurate assessments of demand for a broad spectrum of technologies (e.g. future demand. 16 t.e. graedel, gus gunn and luis tercero espinoza

Ensuring supplies of critical materials to cor- Implications of criticality for corporate porations, countries or regions inevitably involves and governmental policy international trade, because no country or region Modern technology makes extensive use of the possesses the full palette of materials – one area metals designated as critical by the various assess- may have good platinum-group metal deposits ments discussed above. In virtually all cases, but few or no rare earth deposits, while another these uses result in improved product performance: may be rich in copper deposits but lacking those faster computers, sharper images on the display of nickel. Because metal use is diverse, the screen, wider ranges of operating temperatures, etc. world’s countries and continents are linked by Sometimes no suitable substitute for a critical their mutual need for the full spectrum of mate- metal in a particular use is known, as with rhodium rials, and this situation requires continued inter- (employed in automobile catalytic converters to national collaboration. oxidise harmful nitrogen oxide gases, NOx), or neo- Recycling efficiency remains a major challenge dymium (a component of high-strength magnets for most metals. In principle, metals are endlessly used in hybrid vehicles to facilitate electric motor reusable. In practice, they are typically reused performance). In other circumstances a substitute only once or twice (Eckelman et al., 2011). Social might be available, but its use would downgrade a commitment and policy initiatives can play product’s utility, as would be the case for hafnium major roles in improving this picture. in computer chips or samarium in missiles. Thus, Thus, designation of metals or metal groups as the potential or actual scarcity of one of these mate- critical carries with it policy implications for cor- rials has dramatic implications for the industrial porations and governments. The responses need using sectors, or for countries or regions containing to be focused, forward-looking and pursued with those sectors. dedication if the consequences of critical metal There exist a number of possible responses to supply constraints are to be minimised or avoided. the realisation that a particular material is or may be critical. For corporations (e.g. Duclos et al., 2010): ● vigorously investigate possible substitute materials; Outlining this book ● improve material utilisation in manufacturing; ● redesign products to eliminate or reduce critical It is not possible in a single book to cover the material use; entire range of potentially critical metals, nor ● investigate the potential for recycled materials to unambiguously select those that might be to replace or supplement virgin material supplies; of most concern. As a practical and reasonable ● consider entering into long-term contracts or choice, however, we address those deemed critical creating stockpiles to ensure supplies for future by the European Union working group (2010): anti- manufacturing activities. mony, beryllium, cobalt, gallium, germanium, For governments: indium, magnesium, niobium, the platinum-group ● support geological research to locate new metals, the rare earth elements, tantalum and mineral deposits and to better evaluate known tungsten. Lithium is included as well, on account deposits; of its increasing importance in battery technology ● support research into improved technologies and current concerns over its long-term avail- for recycling; ability. A chapter on rhenium has also been added. ● consider voluntary programmes or legislation Following this first chapter, two chapters address to improve rates of collection and appropriate topics generic to all the metals. The first treats the processing of discarded products containing mining industry, explaining its nature and how it recyclable materials. responds to changing demand. The second is on Metal resources, use and criticality 17

Group Group New IUPAC New IUPAC 1 2 34 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Old IUPAC Old IUPAC IA IIA IIIAIVA VA VIA VIIA VIII VIII VIII IB IIB IIIB IVB VB VIB VIIB 0

1 2 1 H He Period

3 4 5 1 6 1 7 8 9 10 2 Li Be B C N O F Ne

11 12 13 14 15 16 17 18 3 Na Mg Al Si P S Cl Ar

19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 4 K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr

37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 5 Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe

55 56 57-71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 6 Cs Ba La-Lu Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn

87 88 89-103 104 105 106107 108 109 11 0 111 11 2 11311 4 115 116 117 118 7 Fr Ra Ac-Lr Rf Db SgBh Hs Mt Ds Rg Cn Uut Fl Uup Lv Uus Uuo

Lanthanide 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 6 La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu

Actinide 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 7 Ac Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr

Rare Earth Elements (REE) Platinum-Group Metals (PGM) Others

Figure 1.9 The Periodic Table of the chemical elements highlighting the metals described in this book. The new IUPAC (International Union of Pure and Applied Chemistry) naming system for the groups is used in this book. The old IUPAC system is shown for comparison. recycling, and provides the basis for an under- as well as a perspective on its supply, demand and standing of recycling prospects and limitations. prospects. These metals and the metal groups Each of the individual metals or metal groups covered are shown in the Periodic Table (Figure 1.9). listed above is then given its own chapter, which provides a summary of appropriate information, including physical and chemical properties, Acknowledgements geology, production, trade, recycling and future outlook. While not exhaustive, this information Gus Gunn publishes with the permission of constitutes a basic understanding of the element or the Executive Director of the British Geological element group’s criticality aspects and challenges, Survey. 18 t.e. graedel, gus gunn and luis tercero espinoza

Note European Commission (2003) Hydrogen Energy and Fuel Cells: A Vision of our Future. EUR 20719 EN. 1. In mineralogy and petrology a different definition is European Commission (EC) (2010) Critical Raw used and a mineral is defined as an inorganic sub- Materials for the EU. Report of the ad-hoc working stance with a definite chemical composition and a group on defining critical raw materials. http:// characteristic crystal structure. ec.europa.eu/enterprise/policies/raw-materials/ critical/index_en.htm. Gilbert, N. (2009) The disappearing nutrient. Nature, References 461, 716–718. Giurco, D., Prior, T., Mudd, G., Mason, L and Behrisch, Angerer, G., Erdmann, L., Marscheider-Weidemann, F., J. (2010) Peak minerals in Australia: A review of Lüllmann, A., Scharp, M., Handke, V. and Marwede changing impacts and benefits. Prepared for CSIRO M. (2009) Rohstoffe für Zukunftstechnologien. Minerals Down Under Flagship, by the Institute of Einfluss des branchenspezifischen Rohstoffbedarfs in Sustainable Futures (University of Technology, rohstoffintensiven Zukunftstechnologien auf die Sydney) and Department of Civil Engineering zukünftige Rohstoffnachfrage. Stuttgart: Fraunhofer (Monash University), March 2012. Institut für System- und Innovationsforschung ISI; Graedel, T.E. and Allenby, B.R. (2010) Industrial Ecology Fraunhofer-IRB-Verlag, pp. 404. and Sustainable Engineering, Prentice Hall, Upper Bardi, U. and Pagani, M. (2007) ‘Peak Minerals’ posted Saddle River, NJ, USA. by Vernon C. in The Oil Drum: Europe. http://www. Graedel, T.E. and Erdmann, L. (2012) Will metal scar- theoildrum.com/node/3086. city impede routine industrial use. MRS Bulletin, 37, Buijs, B. and Sievers, H. (2012) Critical thinking about criti- 325–331. cal minerals: assessing risks related to resource security. Graedel, T.E., Barr, R., Chandler, C. et al. (2012) Briefing paper within the Polinares FP7 project. Available Methodology of metal criticality. Environmental from: http://www.clingendael.nl/ciep/publications/2011/ Science and Technology 46 1063–1070. Buijs, B., Sievers, H. and Tercero Espinoza, L.A. (2012) Hagelüken, C., Drielsmann, R. and Ven den Broeck, K. Limits to the critical raw materials approach. Waste (2012) Availability of metals and materials. In: and Resource Management, 165, WR4, 201–208. Precious Materials Handbook (Hanau-Wolfgang, Cathles, L. M. (2010) A path forward. Society for Economic Germany: Umicore AG Co. KG.) Geologists Newsletter, No. 83, October 2010. Hubbert, M.K. (1956) Nuclear energy and the fossil Cohen, D. (2007) Earth’s natural wealth: an audit. New fuels. Presentation at the Spring Meeting of the Scientist, 23 May, 34–41. Southern District, American Petroleum Institute, Crowson, P.C.F. (2011) Mineral reserves and future min- San Antonio, Texas, March 1956. erals availability. Mineral Economics 24, 1–6. IEA (2008) World Energy Outlook 2007. China and India Cunningham, C.G., Zappettini, E.O., Waldo Vivallo S. Insights. OECD/IEA 2007, Paris, France. et al. (2008) Quantitative mineral resource assessment Kleijn, R. and van der Voet, E. (2010) Resource con- of copper, molybdenum, silver and gold in undiscov- straints in a hydrogen economy based on renewable ered porphyry copper deposits in the Andes mountains energy sources: An exploration. Renewable and of South America. USGS Open-File Report 2008–1253. Sustainable Energy Reviews 14, 2784–2795. Duclos, S. J., Otto, J.P. and Konitzer, D.G. (2010) Design Joint Research Council (2011) Critical metals in stra- in an era of constrained resources. Mechanical tegic energy technologies. (Luxembourg: Publications Engineering 132/9, 36–40. Office of the European Union.) Eckelman, M., Reck, B.K. and Graedel, T.E. (2011) Mikolajczak, C. and Harrower, M. (2012) Indium Exploring the global journey of nickel with Markov Sources and Applications. Minor Metals Conference, models. Journal of Industrial Ecology 16/3, 334–342. February 2012. Erdmann, L. and Graedel, T.E. (2011) The criticality of Morley, N. and Etherley, D. (2008) Material Security: non-fuel minerals: A review of major approaches and Ensuring Resource Availability to the UK Economy. analyses. Environmental Science and Technology 45, Oakedene Hollins: C-Tech Innovation Ltd: Chester, 7620–7630. UK. Ericsson, M and Söderholm, P. (2012) Mineral depletion Nassar, N.T., Barr, R., Browning, M. et al., (2012) Criticality and peak production. Polinares, Working Paper No. 7, of the geological copper group. Environmental Science September 2010. and Technology 46, 1071–1076. Metal resources, use and criticality 19

National Research Council (2008) Minerals, Critical U.S. Department of Energy (2010) Critical Materials Minerals, and the U.S. Economy. Washington, DC: Strategy. Washington, DC. National Academy Press. U.S. Department of Energy (2011) Critical Materials Ragnarsdóttir, K.V. (2008) Rare metals getting rarer. Strategy. Washington, DC. Nature Geoscience. 1, 720–721. USGS. (2002) Rare Earth Elements – Critical Resources Shell (2008) Shell Energy Scenarios to 2050. Shell for High Technology. U.S. Geological Survey Fact International BV. Sheet 087-02. Sverdrup, H., Koca, D. and Robert, K.H. (2009) Towards USGS. (2010) Mineral Commodity Summaries 2010. a world of limits: The issue of human resource follies. USGS. (2011) Mineral Commodity Summaries: Goldschmidt Conference Abstracts 2009. Copper. 2. The mining industry and the supply of critical minerals

DAVID HUMPHREYS

Independent Consultant, London, UK

Mineral products are bought for their utility, this similar duality. The economic function of mining utility being reflected in the price which companies is to respond to the requirements of consumers are prepared to pay for them. Properly the market, as expressed through mineral prices. functioning markets should ensure that an appro- For the most part, the industry does this quite priate supply of such products is available to effectively. The industry has always had a strong meet consumer demand. A shortage of the enterprise culture and rising mineral prices can sought-after mineral serves to push prices up and usually be relied upon to prompt mining and stimulate companies to invest in new production exploration companies to develop mines and capacity. A surfeit of supply leads to a fall in price search for new mineral deposits. and a curtailment of output. As with mineral consumers, producers operate The issue of a mineral’s ‘criticality’ enters into in a national setting. National authorities are the equation because the global economy is com- responsible for establishing the legal, fiscal and posed not just of companies and consumers but environmental parameters within which mining also of nations, and nations have strategic interests. companies work. However, like consuming nations, Within the broader, strategic, context, mineral producing nations have strategic objectives. In this products are viewed not only as having utility to context, mining may be perceived as a vehicle consumers but also in terms of the contribution for the promotion of broader objectives such as they make to national projects, such as raising the economic development, the reduction of poverty or living standards of the nation’s citizens, maintain- the assertion of national self-determination. In a ing a capability to produce certain important direct parallel with the process of securitisation in industrial goods, or ensuring that the nation has consuming countries, the assertion of these stra- the ability to defend itself militarily. In making the tegic priorities results in the politicisation of the transition from being simply ‘useful’ to being ‘criti- mineral products and conditions the ability of the cal’, minerals and their supply become not just mining industry to respond to market signals and matters for the market but also matters of national thus to supply the minerals that consumers require. security. The process of transition is thus often This chapter is divided into five sections. The referred to as ‘securitisation’. first looks at the mining industry and its major The role played by the mining industry in corporate components, the miners and explorers. meeting the demand for minerals is subject to a The second discusses how the mining industry

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. The mining industry and the supply of critical minerals 21

Table 2.1 World’s largest mining companies by market capitalisation, mid-March 2013. (Data from author’s estimates based on web sources.)

Rank Company Country Market Cap $bn

1 BHP Billiton Australia 190 2 Rio Tinto UK 92 3 Vale Brazil 90 4 Xstrata Switzerland 51 5 Anglo American UK 39 6 Freeport McMoRan USA 34 7 Grupo Mexico Mexico 32 8 Norilsk Nickel Russia 32 9 Barrick Gold Canada 29 10 Goldcorp Canada 26 11 Newmont Mining USA 20 12 Newcrest Mining Australia 18 13 Teck Resources Canada 17 14 Antofagasta UK 16 15 Fresnillo UK 16 16 AngloGold Ashanti South Africa 13 17 Fortescue Metals Group Australia 13 18 Yamana Gold Canada 11 19 Impala Platinum South Africa 9 20 Kinross Gold Canada 9

responds to the demand for minerals and to US$2100 billion (Citi, 2011a). London lies right at changes in the level of demand. The third exam- the heart of this industry, and is host to the head- ines the factors which inhibit the mining indus- quarters of several of the world’s largest mining try’s responses to changes in demand. The fourth companies. As of March 2013, there were thir- looks at some of the specific issues posed for teen mining and metals companies in the FTSE miners by the minerals currently deemed ‘critical’ 100 having a combined market capitalisation of and at the role of China in mineral markets. The US$340 billion, 12.7 per cent of the total value of fifth considers some of the things that govern- the FTSE100 (FTSE, 2013). Seven years earlier, the ments of consuming countries can do to promote share was six per cent. the supply responsiveness of the mining industry. The structure of the global mining industry today is the product of a long and complex his- tory. The largest and most publicly visible com- Suppliers of minerals – miners and explorers panies are the so-called ‘global diversified miners’, or mining ‘majors’. These are, by any standards, The mining industry exists to meet the mineral large companies, operating across many geogra- requirements of consumers and, in so doing, phies and minerals. Following a period of consol- make profits for shareholders. Although not on idation during the first decade of the century, this the scale of the oil and gas industries, the mining group currently comprises BHP Billiton, Vale, Rio industry is, nevertheless, a very large industry. Tinto, Anglo American and Xstrata.2 The market The enterprise value1 of the global mining capitalisation of the world’s largest mining industry in 2010 is estimated to have been around companies is shown in Table 2.1. The country 22 david humphrey indicated is the country of the company’s pri- economies and correspondingly large capital out- mary stock market listing. The table, it should lays, like iron ore. They can, however, operate in be noted, excludes aluminium companies, this markets where demand is small or where ore because most of the value of aluminium, like deposits can be worked on a relatively small scale, steel, is created through metallurgical processing like precious metals or semi-precious stones. At the rather than through mining. extreme end of this part of the industry are the arti- At the next level down in terms of scale, com- sanal miners. These are very small, maybe even panies tend to be more focused with respect to part-time, operators, recovering minerals that can either commodity or country. Freeport McMoRan, be easily mined near surface (such as alluvial gold, Grupo Mexico and Antofagasta, for example, are tin, tantalum and diamonds) using very little focused on copper, while Barrick Gold, Goldcorp capital. Such production activity is commonly and AngloGold Ashanti are, as their names suggest, lightly regulated or indeed wholly unregulated, focused on the production of gold. Companies with miners operating under very basic, and often which produce a variety of products, but which unsafe and environmentally unsound, conditions. operate predominantly in one country, include Artisanal mines do, nonetheless make a significant, several from the former Soviet Bloc, most notably if not always terribly reliable, contribution to the Norilsk Nickel, but also Kazakhmys and ENRC supply of several critical minerals. (Eurasian Natural Resources Corp.) which fall just The other key players in the mineral supply outside the top twenty companies listed. equation are exploration companies. This is Most of the world’s largest miners, and all of the entrepreneurial end of the business – the those in Table 2.1, are public companies, quoted equivalent of technology start-ups – the end on stock markets (from which their market where small companies go out to find mineral capitalisations are derived). There are, in addition, deposits in the hope either of being able to mine a few mining companies comparable in the scale them themselves or else (and more often) sell of their mineral output to those listed in the table them on at a good profit to a larger company for which are either wholly or predominantly owned development. Since exploration can create enor- by the state. These include the world’s largest mous value for shareholders, turning what might copper producer, Codelco, which is owned by the otherwise be a fairly worthless piece of land into state of Chile, and a handful of Chinese companies a profitable business opportunity, exploration such as China Shenhua, Yanzhou Coal, China companies have a strong pioneering quality. The Minmetals Corporation (Minmetals), Chinalco, highest rewards typically go to those with inno- Metallurgical Corporation of China, (MCC) China vative ideas about ore genesis (an example might Nonferrous Metal Mining Corp. (CNMC) and the be those which uncovered significant diamond Jinchuan Group. Although production from state- resources in Canada) or which are prepared to go owned enterprises is significant and growing, the looking in remote and difficult places. By the extent of state ownership in mining is still very same token, exploration is also an extremely much less than is the case with oil and gas. high-risk activity, and much exploration ends in Beyond the larger and mid-sized mining com- failure and in investors losing their money. panies, there are huge numbers of smaller miners, Accordingly, exploration companies have their ranging from quoted companies with two or three own particular economics and their own specialist mines to small family enterprises. Some produce investors. Banks, which might well be interested for international markets and some just for local in helping a mining company with proven mineral markets. The nature of the mineral product and the reserves to finance the construction of a mine, are form of its occurrence play an important part in not generally interested in financing exploration. determining what products such producers focus Exploration companies therefore tend to have to on. Small miners do not generally try to compete rely on equity (i.e. stock market) financing for in mineral markets where producers need scale their activities or on the support of large private The mining industry and the supply of critical minerals 23

70

60

50

40

30

20 Percent of total expenditure

10

0 Figure 2.1 Worldwide exploration by 1997 1999 2001 2003 2005 2007 2009 2011 company type: per cent shares, 1997– Majors Intermediates Juniors Govt/others 2012. (Data from MEG, 2012.)

investors. Some stock markets specialise in the 50 per cent of total spend in 2006 and 2007. A provision of this sort of financing, notably the high proportion of exploration spending by Toronto stock exchange (TSX) the Australian juniors is accounted for by gold, the small scale of stock exchange (ASX) and the alternative many gold deposits combined with the easy sale- investment market (AIM) of the London stock ability of the product making this metal the exchange (LSE). Because of the nature of its activ- target of choice for many juniors. A final point to ities and of its financing, this is much the most note is that MEG data is focused on private- sector responsive part of the mining industry and the exploration and accordingly does not take full part that is quickest to adjust to changes in market account of exploration by state companies and perceptions. other state organisations. In light of the fast Metals Economics Group (MEG) has, for many growth of state-funded exploration in countries years, compiled data on global exploration such as Russia, India and, above all, China, in spending. For 2012, it estimated that expenditure recent years, Raw Materials Group of Sweden was at a record level of US$21.5 billion (MEG considers that MEG’s data understate the total 2012). Figure 2.1 shows the distribution of explo- exploration spend (Ericsson, 2011a). ration expenditure in recent years split between that undertaken by mining majors, by intermedi- ates, by juniors and by government or other orga- Industry dynamics nisations. Two points are apparent from this figure. First, spending by the juniors was much The larger mining companies do not generally more responsive to rising prices during the course give much thought to a mineral’s perceived criti- of the metal price boom in 2004–2007 and more cality when evaluating an investment. Their role responsive also to the falling off of prices in is to produce minerals for which there is a proven 2008–2009. Secondly, despite the small size of market and to make a profit by so doing. the companies in this sector, the juniors It is certainly the case that part of the collectively account for a very large proportion assessment of whether something can be mined of total exploration, this share rising to over profitably resides in a miner’s judgement about 24 david humphrey the strength of demand for the mineral in question in time to relieve any shortage, it is simply that and the price that consumers will be prepared to funding is more readily available at such times. pay for it. However, for the most part these The identification of rare earth elements and cannot be very accurately determined. Mineral lithium as critical minerals in recent years has demand and mineral prices are functions of the helped generate huge interest in exploration for economic cycle, the forecasting of which is a very these minerals. There are believed to be some inexact science. Moreover, proving up resources three hundred rare earth deposits under evalua- and bringing them into production is a process tion (Chegwidden and Kingsnorth, 2011) and over that can take several years and a lot can change in one hundred lithium projects (Mining Journal, the condition of markets during that time. Thus, 2011). This gold rush mentality – wherein high while a miner must have some general level of levels of exploration feed expectations about the confidence that a market will exist for the product demand prospects for a mineral, and vice versa – to be produced and that prices will be sufficient is an age-old feature of the mining industry. to generate a positive return on capital, detailed Only a very few of the many thousands of projections of demand growth are not normally mineral prospects that are explored ever actually the primary factor behind a decision to invest. make it through to production. And when it Mining companies cannot realistically lay claim comes to the determination of whether a mineral to any particular comparative advantage in the deposit is to be developed, then judgements about art of economic forecasting and will generally, the outlook for demand may well take second and rightly, be sceptical about the claims which place to judgements about the economics of appear in the popular press from time to time production. After all, if too many companies are about the glittering prospects of this or that pursuing the same growth segment for a given exotic-sounding mineral. mineral, then there is always the risk that the The situation with junior miners and explora- market will at some point tip over into serious tion companies is a little different. As already oversupply, at which point the relative competi- noted, these companies are generally dependent tiveness of producers becomes rather important. on equity markets for their financing. Their Many large mining companies, it might be noted, survival thus depends on their ability to spark talk about their strategic objective as being to and to sustain interest amongst investors. secure and operate low-cost, long-life, mines Accordingly, they tend to be rather more sensitive without reference to any particular mineral or its to market perceptions about the desirability of demand outlook. different minerals than are large mining com- In order for a prospect to be developed, a mining panies and will often creatively talk up the pros- company will generally want to be sure that the pects for the products which they are hoping to resource is of a scale, quality and consistency to find and to mine. support production long enough to permit the This being the case, exploration companies recovery of the initial capital investment. It will and junior miners are that much more likely than need to be sure that the conditions of the rock are larger, well-established, mining companies to be such as to permit safe and efficient mining. It will responsive to the notion of a mineral’s criticality. need to be sure that power and water are available A project becomes easier to promote if the prod- to the project and that transport exists to get the uct it is expected to recover is viewed as having product to market. In essence, what this will all an exciting growth prospect, or is used in new ultimately boil down to is that the company will and exotic applications; especially when this is want to be confident, or as confident as it is pos- reflected in strongly rising prices. It may not be sible in business to be, that it will be able to pro- that the mineral in question is suffering from duce at costs which will make it profitable over insufficient investment, or even that there is a the long term. This will, of course, depend in part realistic prospect of getting a mine into operation on its assessment of the long-run price of the The mining industry and the supply of critical minerals 25

product to be produced. However, because of the deemed critical, and whose demand is expected uncertainties attaching to forecasting long-run to grow rapidly in coming years, there is no commodity prices, the mining company will also reason in principle why the industry should not be seeking the comfort of knowing that its costs be able to respond to the challenge and to match are competitive relative to those of others in the increases in demand. industry. Aside from encouraging the development of Such comparative cost assessments play an known high-quality deposits, higher mineral absolutely vital part in mining company decision prices have the effect of converting what were making. Having low production costs is not once marginally economic resources into mine- in itself sufficient to justify investment in a able reserves, and may similarly convert waste project. The object of the exercise, after all, is to dumps from earlier workings into sources of make money, and while low operating costs are recoverable product. Sustained interest in these clearly better than high operating costs, if low minerals will also likely stimulate an interest in production costs can only be achieved through investing in, and improving the technologies very high levels of spending on capital, it may used for, the recovery of these minerals. It is still be that a project does not merit development. interesting to note that Molycorp Minerals’s However, comparative assessments of operating Mountain Pass rare earths mine in California, costs remain important to miners for several which has re-opened in response to higher reasons. They permit companies to benchmark rare earth element prices, has been wholly themselves against others in the industry and to reconfigured since it was closed in 1995 follow- determine how efficiently they are producing. ing a thorough review of all aspects of its technical They provide information about where prices and environmental performance. The new mine might trend longer term, in as far as prices in a will have process recoveries of 95 per cent as competitive market tend towards the cost of the against 60–65 per cent at the old mine, will use marginal producer; that is, the highest cost 30 per cent less reagents and only four per cent of producer required in production to meet prevail- the fresh water. Accordingly, it will be able to ing demand. And, by the same token, they will produce rare earth elements at much lower costs provide information about which companies than it could when it operated previously. are likely to be making positive cash returns Mining is, however, a capital-intensive industry throughout the cycle, something in which those with long lead times from discovery to production financing a mine’s development, be they share- and its responses are necessarily lagged. For new holders or banks, will be particularly interested. mine production to flow, mining companies have For the most part, the mining industry has to be convinced that they have a viable project and been successful in responding to the changing then secure funding for it. Combined with the requirements of the market, and indeed it is need to assess the environmental impacts of a organised and incentivised so to do. Rising prices mine, to acquire permits to mine and, frequently, provide a signal of actual or impending shortage forge agreements with local communities, this and companies accordingly respond by increasing process can take several years. And then there is output from existing operations and by insti- the not insignificant matter of building the mine gating searches for new resources. Rising prices itself. This will require ground preparation, the also have a stimulus effect on financial markets construction of plant, the acquisition of specialised and facilitate the raising of debt and equity fund- equipment and the creation of facilities for mine ing by miners and explorers. Reflecting these waste. Not uncommonly, it will also require the factors, global mining investment in 2011 soared building of railways, ports and power stations. to US$175 billion, the highest level ever recorded Accordingly, while it may be the case that (UBS, 2011) and investment in 2012 is believed eventually miners will catch up with imbalances to have been higher still. For minerals currently in supply and demand for minerals, there can 26 david humphrey

500 18 450 400 16 350 300 14 250 200 12 US cents/Ib 150 Million tonnes 100 10 50 0 8 1985 1989 1993 1997 2001 2005 2009 2012 Mine production Price

Figure 2.2 Copper prices and mine production, 1985–2012. (Price data are annual averages calculated from London Metal Exchange monthly averages; production data compiled from World Metal Statistics Yearbook, published annually by World Bureau of Metal Statistics.) nonetheless be periods of shortage while they are sometimes referred to as demand rationing or bringing on new supply, these periods potentially demand destruction. lasting several years. Such long lags are one Demand rationing works in several ways. of the defining characteristics of the mining First and foremost, high prices encourage con- industry and the basis of its sticky supply sumers to use less of a product or to use it more responses. The point can be illustrated at a high efficiently. Thus, high copper prices have level with the example of copper. Figure 2.2 resulted in the thin-walling of copper pipes and shows a plot of world mine production against in gauge reduction. High prices of nickel have the world copper price. A close examination of encouraged a shift towards the production of this figure reveals that there was an eight-year stainless steels which use less, or indeed, no, gap between the uptick in prices which took nickel. High prices also encourage consumers place in 1987 and the acceleration of mine pro- who can do so to switch to using cheaper duction growth in 1995. A similar delayed supply alternative materials. Thus, high platinum prices response is evident in the cyclical downswing. have led jewellers to substitute white gold and The decline in prices which occurred following palladium for platinum, while high copper prices the price peak of 1995 did not result in a visible have encouraged the substitution of plastic reduction in global production until 2002. In the plumbing pipe for brass pipe. Finally, high prices most recent cyclical upswing, the price increases result in a reduced call on mined materials by which began in 2004 finally resulted in an encouraging recycling (see Chapter 3). A key acceleration in mine production in 2012. Like element in the economics of recycling is played the proverbial oil tanker, mineral production can by the cost of collection and separation. Although take a long time to turn around. the effects tend to be highly mineral specific, as During these lengthy periods of supply mineral prices rise so generally they provide an adjustment, prices are required to take the strain incentive for the collection of old scrap and an of forcing supply and demand into alignment increase in the rate of recycling. by rising to levels which choke off the portion These effects of higher prices are part of the of demand which cannot be satisfied. This is normal mechanism of adjustment in mineral The mining industry and the supply of critical minerals 27

markets. Indeed, some of the changes brought reserves of a mineral by annual production to get about by price-induced changes in demand have a ‘static life index’; that is, an estimate of the beneficial long-run effects in terms of increasing expected life of the remaining reserves expressed the efficiency of materials use and promoting in years. Such estimates are, unfortunately, com- advances in technology. Although it is not so monly subject to misrepresentation in that casual hard to point to examples where mineral mar- users of these data have a tendency to overlook kets have suffered sustained shortfalls in supply – the fact that reserves is a dynamic concept. For the shortages of cobalt arising from civil war in many of the most commercially important Zaire in the late 1970s, for example – it is quite minerals, including copper, nickel, lead, tin and difficult to think of peacetime examples where zinc, reserves life tends to fall into a range of shortages of minerals have had serious adverse twenty to fifty years (USGS, 2012). However, it long-run effects on an economy. Normally, should be noted that these reserves lives have not markets adjust in the short term through price- much changed in many years (Crowson, 2011). induced demand rationing and in the longer run As production has risen, so reserves have gone through increased supply. up. The commercial incentives simply do not exist for companies to go out and prove up reserves which will be required more than fifty Constraints on mineral supply response years out. Moreover, it should also not be forgotten that most mineral materials are not While it may generally be the case that properly destroyed by use and that in addition to reserves functioning markets will provide solutions to in the ground there are substantial amounts of mineral shortages, there is a variety of natural, above-ground materials available for re-use and economic and institutional factors which in prac- recycling. tice can inhibit the responses of miners and For the minor metals which are the focus of explorers to shortages and thereby prolong the concern in the debate on critical minerals, the period of supply adjustment. These are important picture on reserves life is more varied. Some, in assessing the likely future availability of criti- such as the rare earths, lithium and tantalum, cal minerals. The next three sections look at have reported reserves which are very large these factors in turn. indeed relative to current levels of production, stretching out in the case of the first two to sev- Natural constraints eral hundred years. For others, because they are recovered as by-products, there are no meaningful One of the defining characteristics of the min- estimates of reserves so a calculation of reserves erals industry is that mineral resources are life cannot be made. depleted through exploitation. (For a fuller If the physical availability of minerals is not discussion on resource definitions and related a major constraint on the supply response of the issues, see Chapter 1.) This gives rise to a common minerals industry, the geographic concentration perception that physical availability may become of mineral resources can be. Such a concentration a constraint on future mineral supplies. In point exists in the case of the platinum group metals of fact, physical availability of minerals in the (PGMs).3 (A more detailed discussion on the ground is scarcely, if ever, a constraint on mineral PGMs is to be found in Chapter 12.) According to supply. While some minerals are inevitably easier the US Geological Survey, some 95 per cent of to find and develop than others, most minerals the world’s reserves of PGMs are located in South mined commercially are available in quantities Africa, in the Bushveld Complex (USGS, 2012). which are adequate for very many years to come. Such concentration of reserves does not in itself A simple way of looking at the adequacy of represent a constraint on supply, but it does mineral reserves is to divide through the reported make supply more susceptible to constraint. 28 david humphrey

Thus, while South Africa’s reserves are sufficient be an issue in supply vulnerability as the case of to supply world markets with PGMs for many the PGMs also illustrated. Examples of minerals years to come, several factors have restricted where corporate concentration of production access to the reserves. is unusually high are niobium (where CBMM is A large part of the better reserves are tied up the dominant producer) and tantalum (Talison by one of the existing major producers and are Minerals). therefore not available to industry newcomers. Another natural constraint on the ability of Those that are available often present serious the mining industry to respond to shortages is the challenges with respect to mining conditions, fact that some metals are produced predomi- permitting and access to smelting and refining nantly as by-products of other, economically facilities. Meanwhile, all mining companies more important, metals. (For more on the sub- operating in South Africa, including the majors, ject, see Chapter 1.) This is not to say that they face tough challenges associated with ownership cannot be recovered in their own right, only that requirements (including the strictures of black the cost of doing so will be very much higher, per- economic empowerment (BEE)), health and safety haps prohibitively so. Thus, cobalt is produced standards, power availability, rising labour costs largely as a by-product of copper and nickel. and a strong Rand. It is largely for these reasons Cadmium, indium and germanium are produced that, despite the existence of strong demand for as by-products of zinc production. Germanium is platinum over the past decade, driven by demand also recovered from coal fly ash. Gallium is pro- for autocatalysts, investment in the industry has duced as a by-product of bauxite. Molybdenum, been constrained and production has grown rhenium, selenium and tellurium are all pro- scarcely at all (see Figure 12.5), forcing up prices duced as by-products of copper. and, as noted above, resulting in a substantial The problem for these metals is that their displacement of jewellery demand for platinum. supply is largely dependent on the production Other cases where a high concentration of economics of the metal of which they are a reserves represents a potential constraint on by-product. If production of the principal metal supply are phosphates in Morocco and cobalt ceases to be economically viable then the in DR Congo. In the case of Morocco, which by-product will cease to be recovered too. It is accounts for three-quarters of the world’s for this reason that the supply of by-products reserves of phosphate rock (USGS, 2012), access tends to be insensitive to changes in the level to the industry is restricted by the fact that the of demand and why the prices of mineral recov- industry is wholly state-owned. With respect to ered as by-products can fluctuate wildly. The cobalt, the DR Congo accounts for around half price of molybdenum (Figure 2.3) shows a fairly the world’s reserves of cobalt and around half classic by-product profile, which is to say, long its production (USGS, 2012). The challenging periods of low and stable prices (reflecting investment climate in DR Congo, and the reluc- times when markets are well supplied) inter- tance of large international companies to operate spersed by explosive price peaks (reflecting there, has long made this a matter of concern for times when supply is unable to respond consuming nations. The invasion of the Katanga to increased demand). Rhenium, another province in DR Congo in the 1970s (or the Shaba by-product of copper which is discussed in province and Zaire as they were respectively Chapter 14, shows a similar price pattern to known at the time), and the disruptions to molybdenum. Quite a few of the minerals supply that this caused, were a key event in trig- deemed critical in the US and the EU fall into gering the ‘resource war’ concerns of the 1980s. this by-product category and accordingly In addition to the concentration of reserves by display the rather erratic supply and price country, corporate concentration can sometimes behaviour of such metals. The mining industry and the supply of critical minerals 29

45

40

35

30

25

20

15 US$/lb, dealer oxide 10

5

0 1973 1977 1981 1985 1989 1993 1997 2001 2005 2009 2013

Figure 2.3 Molybdenum price, 1973–2012. (Price data are quarterly averages from Metal Bulletin.)

1.00 88 0.95 87 0.90 0.85 86

0.80 85 0.75 84 0.70

Copper grade % 0.65 83

0.60 Copper recovery rate % 82 0.55 0.50 81

1980 19821984 198619881990 19921994 19961998 200020022004 200620082010 Figure 2.4 Copper mine grades and Copper grade Copper recovery rate recoveries, 1980–2010. (After Citi, 2011b.)

Economic constraints To some degree, the upward pressure on A second set of constraints on mineral industry costs which results from these trends can development – referred to here as economic con- be offset – or even more than offset – by cost- straints – are a product of the fact that the quality reducing improvements in technology, and histor- of mineral resources has a tendency to deteriorate ically this has been the general experience of the over time. Ores become lower in grade or more industry. However, there is no law which says difficult to treat, while ore deposits are found at that this has to be the case and, for a number of greater depth or in more difficult locations. As an mineral commodities, it would appear that the illustration of this, Figure 2.4 shows the recent declining quality of reserves, combined with other declining trend in copper ore grades and in recov- factors such as higher energy prices, water avail- eries from those ores. ability and tougher environmental requirements, 30 david humphrey

7000

6000

5000

4000

3000

2000 US$/tonne (2011US$)

1000 Figure 2.5 Long-run marginal cost of 0 copper production. (Adapted from 2004 2007 2009 2011 Robinson, 2011.) are pushing up net production costs, notwith- remote from infrastructure, and as the environ- standing continuing technological progress. mental and political challenges of mining mount, Sticking with copper, Figure 2.5 shows an so the cost of building mines has escalated too. analysis by the consultants CRU of what their Figure 2.6 shows estimates of the capital costs database is telling them has happened to the of some large greenfield copper mines currently long-run marginal cost of producing copper. in development or undergoing evaluation. The (These are notionally the operating costs of the capital costs of these mines typically fall in the last (i.e. highest cost) producer required in pro- range US$10,000–20,000 per tonne of annual mine duction to balance the market. For practical capacity. Capital costs historically have generally reasons, they are more usually derived by been below US$7500 per tonne of capacity, with mechanically taking a reading off the industry US$5000 per tonne for a long time being used by cost curve at fixed point, for example, the 90th or the industry as a rough rule of thumb. 95th percentile.) These costs leapt from around In principle, higher costs of production US$2400 per tonne in 2004 to some US$5800 per should eventually result in higher prices, which tonne in 2011; a real terms increase of 13 per cent should in turn contribute towards bringing a year over the period. Considering that industry forward the necessary investment to balance operating costs declined for the twenty-five years the market. However, there are lags in the prior to this the increase is extraordinary. Partly, system. The long-run prices used by companies of course, the effects are cyclical, but it seems in the evaluation of their projects have been probable that, underlying these cyclical influ- rising, but companies, and the banks providing ences, a structural shift is taking place. Moreover, them with finance, have to be absolutely con- this experience is not exclusive to copper. Similar vinced that prices are going to stay substan- evidence of deteriorating quality of ore resources tially higher on a sustainable basis before and rising production costs can be adduced for risking a commitment to large long-life pro- nickel, PGMs and gold. At the same time, it jects. It also might be noted that exchange rates should be noted, evidence of declining ore quality can be as important as mineral prices in is less evident in other cases, for example, in iron determining a mining project’s viability. The ore, coal and bauxite. emergence of an increasingly multi-polar global What applies to operating costs applies also to economy, and the associated decline in the role capital costs. As mines become deeper and more of the US dollar, is likely to bring with it The mining industry and the supply of critical minerals 31

25000

20000

15000

10000 US$ per tonne of mine capacity 5000

0

Andina I Pebble Aynak Solobo Oyu Tolgoi TampakanToromocho Quellaveco Las BambasKOV restart Olympic Dam Tenke Fungarume Historical (top end)

Figure 2.6 Capital cost of green-field copper mines. (Data from various industry sources.) increased currency instability, adding a further leisure use, and concerns about impacts of min- layer of complexity and risk to mine project ing on the environment and on local commu- evaluation. In short, while companies may nities, make many developed countries highly be investing heavily in new capacity, they are ambivalent about the industry. On one level, this having to overcome higher economic barriers, simply reflects the widespread perception that and assume greater exposure to risk, to do so. mining is a dirty and unsightly business and the fact that developed countries have large numbers Institutional constraints of articulate people with the leisure to fight min- eral projects. Such opposition to mining activ- A third set of constraints on mineral supply ities is often underpinned by the view that involve institutional factors. For the purposes of developed countries can make their living in the section, these are taken to include the laws other, cleaner, ways and import the mineral raw and taxes to which mining companies are subject materials they want from elsewhere. wherever they operate, through to intermittent While many mining companies will want to geopolitical interventions in the industry. persist with mining in the more developed coun- Although a widespread perception exists in tries because of the political stability and legal the more economically developed countries that protections which such countries typically offer, issues of institutional risk is largely a matter for mineral projects in developed countries often emerging and developing nations, this is far from confront extremely demanding and lengthy the case where the mining industry is concerned. permitting procedures and very tight restrictions Pressures on land for housing, agriculture and on emissions, noise, visibility, effluents and 32 david humphrey transportation. Moreover, these restrictions are economic and social development, and a symbol getting tougher with time. Although the provi- of a country’s sovereign right of self-determina- sions in themselves may be entirely reasonable, tion. As the Washington consensus gives ground their cumulative effect can sometimes render a to the Beijing consensus at the level of the global project marginal and encourage miners to go economy, so the emphasis on the role of the where they feel more appreciated and the wealth nation and of the state is becoming more and employment they create are more highly prominent within the confines of the resources valued. It does, however, leave developed coun- sector. The forces for economic liberalism, as rep- tries more heavily dependent on imports than resented, for example, by the attempt to complete they might otherwise be and in a morally weak the World Trade Organisation’s Doha Round, are position to demand that others supply them with in retreat and in international institutions like products they are demonstrably reluctant to pro- the World Trade Organisation the sovereign duce themselves. rights of nations over natural resources are A growing impediment to the ability of miners increasingly being asserted in opposition to the and explorers to respond to changes in mineral principles of economic efficiency which underpin demand is resource nationalism in mineral-rich and legitimise the free trade system. Less and countries. As was the case with the commodity less, it seems, will mineral-rich countries accept boom of the 1970s, the commodity boom which the idea that other countries, or multilateral started in 2004 helped stoke up a debate in institutions, have the authority to determine mineral-producing countries over whether host how they develop their resources and how much nations were receiving a sufficient share of the of their mine production they must make avail- proceeds from mining’s growing success. Countries able on international markets. throughout the world have taken the opportu- In the course of the recent minerals boom, a nity to increase taxes and royalties on the number of countries have come to the view industry. These are factors which mining com- that their national interests are best served by panies have to take into account when assessing insisting that the state has a stake in mining the likely returns to shareholders from an operations on their home soil, or else that investment and, at the margin, can be an impor- mine developments are undertaken by domestic tant factor influencing the decision whether or private companies. Bolivia, for example, not to proceed with an investment. The threat embarked on a programme of nationalisation for of a wide-ranging Resource Super Profits Tax in the mining industry in 2005. Mongolia has Australia in the first half in 2010 resulted in insisted on a major holding for the state in the many mining companies pointedly cancelling large Oyu Tolgoi copper mine. Zimbabwe passed and deferring projects. (A revised Minerals an Act in 2008 to promote the 51 per cent ‘indi- Resource Rent Tax was subsequently intro- genisation’ of mining companies operating within duced in July 2012.) its borders. The Government of Guinea stripped However, the concerns of resource nation- Rio Tinto of some of its permits to mine iron ore alism are not confined to the distribution of in 2008, on the grounds that the company was income. They stretch also into the ownership and not advancing the projects quickly enough, and control of the industry. As noted in the introduc- now requires a substantial direct holding in all tion to this chapter, this politicisation of min- large new mining projects undertaken within the erals in mineral-producing countries is a direct country. In DR Congo, the government in 2010 parallel to the securitisation of minerals in con- expropriated two mines belonging to TSX-quoted suming nations. Minerals viewed within this First Quantum Minerals. broader political context become not just the Nor should it be supposed that such interven- basis of wealth-generating economic activity but tionism is confined to developing countries. a potential component in a project of national The Australian government blocked the purchase The mining industry and the supply of critical minerals 33

100

80

60

40 Policy potential index

20

0

US* Chile Brazil Peru India PNG Russia Bolivia China FinlandSweden Guinea Zambia Canada* Australia* Argentina Colombia Mongolia DR Congo Venezuela Indonesia Philippines Kazakhstan South Africa * Average for states listed

Figure 2.7 Policy attractiveness of large mining countries. (Data from Fraser Institute, 2011.) of the Prominent Hill copper mine by China Congo, Guinea and Mongolia in the hope and Minmetals in 2009, while the Canadian government expectation that they can manage the geopolit- blocked BHP Billiton’s proposed take-over of the ical risks involved and not become victims Potash Corporation of Saskatchewan in 2010 on several years down the road of the ‘obsolescing the grounds that it was not in the national interest. bargain’ (the situation in which the investment State interventionism inevitably adds another has been made and the rules are changed). Time layer of uncertainty to investment decision mak- will tell whether this confidence is justified. The ing by mining companies. Political risk experience of the oil sector, it must be said, which assessment is difficult and unreliable and there is now wholly dominated by state firms, provides are only so many things that companies can do to a somewhat discouraging example. mitigate risk. Many available strategies for risk An objective assessment of the nature and mitigation, such as bringing in partners or buying scale of geopolitical risk across the industry political risk cover, result in reduced control over poses obvious problems and, in the last resort, projects and/or increased costs. Despite this, on it is the geopolitics of the particular country in the basis that mining companies have to go where which a miner is thinking of investing which the minerals are found (and presumably also matter. However, attempts are routinely made because mineral prices remain high), companies to try and provide some comparative context are continuing to commit to invest in what might for the assessment of this issue. Figure 2.7 be regarded as ‘difficult’ countries such as DR shows the results of an investor perceptions 34 david humphrey survey carried out annually by the Fraser at the time was that the USSR was broadly self- Institute of Canada. Amongst other things, this sufficient in minerals. survey seeks to capture the mining industry’s A second wave of concern over the supply of perception of the relative attractiveness of critical minerals followed (not coincidentally) the mining policy in a variety of mineral-rich coun- commodities boom of the 1970s. Although the tries, taking account of political risk alongside Cold War was still on-going at that time, and a range of other factors such as taxation, the the USSR was still perceived as a threat to the administration, interpretation and enforcement West, the primary focus of concern over mineral of mining laws, and environmental regulation. supplies at that time had shifted to South Africa. While such analyses have their limitations, The policy of apartheid in South Africa had alien- Figure 2.7 serves to make the point that percep- ated many western states and there was a wide- tions of policy attractiveness vary significantly spread view that the USSR was seeking to across different countries, with some regimes capitalise on the situation through its support for (those on the left-hand side of the chart) viewed the African National Congress (ANC) and the as essentially supportive and others as effec- socialist regimes of neighbouring Angola and tively no-go zones. It also suggests that the Mozambique. The government of South Africa perceived attractiveness of many mineral-rich was, at the same time, using the threat of disrup- countries in the developing world is considered tion to mineral supplies to the region to bolster quite low. its position in western capitals. The specific min- erals whose supply was deemed under threat by the US, Western Europe and Japan at this time Critical minerals and the role of China included the PGMs, manganese, chromium and vanadium, for all of which South Africa was the The specific minerals which are the object of western world’s leading supplier (House of Lords, concern for mineral-consuming countries vary 1982 and Maull, 1986). Because of Zaire’s4 through time, as also do the countries viewed as proximity to South Africa and its dependence on unreliable sources of supply. This in turn has South Africa’s transport routes, supplies of cobalt implications for the producers whose role it is were also considered vulnerable to unfolding to seek to ensure that adequate supplies of the events in South Africa (this quite apart from minerals are forthcoming. issues related to Zaire’s own political instability). The concerns which arose in the immediate The concern over mineral supplies at this post-WWII era over mineral supply were focused time was less to do with the threat to the mili- largely on the military requirements of having to tary capabilities of mineral consuming coun- fight a sustained conventional (i.e. non-nuclear) tries and more to do with the threat of economic war. Efforts to address the threat in the US – disruption. Manganese, chromium, vanadium which included the creation of a large materials and cobalt were used in the production of high- stockpile – were therefore focused on a lot of performance steels, such as stainless steels and relatively basic industrial raw materials, particu- high-strength low-alloy (HLSA) steels, as well larly those which were not abundant in the US, as superalloys, which were in turn used for such as bauxite, manganese, zinc, lead, nickel, the manufacture of important high-technology chromium and tin (Anderson and Anderson, products such as petrochemical plant, oil pipe- 1998). With respect to the threat of nuclear con- lines and jet and gas turbines. The fear was that frontation, there were parallel concerns relating curtailment of supplies of these metals from to the availability of uranium. The USSR, the South Africa would cause serious dislocation in source of the presumed threat to the USA in the strategically important industrial sectors, from context of the Cold War, would have had similar energy production to aerospace. These concerns concerns, although the perception in the West gave rise in a number of countries, including The mining industry and the supply of critical minerals 35

Japan, the Republic of Korea and the UK, to stra- is a major producer and supplier of many tegic stockpiling of the threatened metals and high-technology minerals, and that western to the adoption of schemes to incentivise min- consumers became heavily dependent on eral exploration. supplies from China during years when China In the most recent manifestation of concern was offering these minerals at substantially over the supply of critical minerals, the focus of lower prices than were available from suppliers concern has once more shifted. With respect elsewhere. Of the fourteen minerals judged to the specific minerals which are deemed most critical by the European Commission – under threat, the focus has shifted to a range of antimony, beryllium, fluorspar, gallium, specialised, low-volume metals used in the pro- germanium, graphite, indium, magnesium, rare duction of technologically advanced consumer earth elements, tungsten, niobium, PGMs, electronics, green energy products and defence cobalt and tantalum – no less than ten (the first applications. Many of these are discussed in ten minerals listed) are sourced by the EU sub- detail elsewhere in this book. stantially from China (European Commission, The sophisticated nature of the products in 2010). The concern is that China’s own growing which minerals now designated as critical are domestic use of these minerals is reducing used and the growing complexity of linkages the supplies being made available for export, between different sectors of modern economies creating increased competition for supplies (as well as the blurring of the distinction bet- amongst other users of these minerals and ween commercial and military products), putting upward pressure on prices. makes the threat which their non-supply would Not surprisingly, given that it is the world’s pose rather harder to evaluate than was the case largest and fastest growing market for minerals, in earlier era (Anderson and Anderson, 1998). China shares many of the concerns of the US While there is a clear sense in consuming coun- and the EU about minerals availability. In fact, tries that these metals are important for certain because of the need for a good supply of raw cutting-edge applications, it is evident from materials to support the rapid industrialisation studies published in the US and EU that the and urbanisation of the country, and because authors of these studies have struggled with the legitimacy of China’s leadership depends the matter of how to assess the relative impor- in no small part on its ability to sustain high tance of different end-uses of the minerals growth rates, China takes the matter of min- designated as critical and the likely economic eral supplies very seriously indeed (FT.com, impact of their non-availability (NRC, 2008 2011 and Ericsson, 2011b). For those minerals and European Commission, 2010). While the which it can source internally, the Chinese approaches adopted in the US and EU reports government has generally encouraged local mine differ – with the US study resting more on development. Recently, however, this objective expert judgment and the EU study adopting a has awkwardly become conflicted with another more quantitative approach – both generate policy objective, namely the need to regulate some rather counter-intuitive results. Thus, for the mining industry more tightly so as to example, the US study determines that the improve its environmental performance and con- economic impact of restrictions on the supply serve resources, and led the Chinese government of rhodium would be greater than those on to seek to restrict the export of certain minerals copper, while in the EU study the economic considered important to national economic importance of tellurium and rhenium is rated development. as higher than that of copper and the PGMs. There is, however, a long list of minerals The geographic focus of concern has also which China cannot source wholly from domestic shifted since the 1980s. It has shifted towards sources. As of 2010, China had to import 100 per China. This follows from the facts that China cent of its PGMs, 85 per cent of its copper and 36 david humphrey nickel, and 70 per cent of its iron ore. For these Total value in 2009: US$386 billion minerals, China has had to turn to international 9% markets and its purchases of these and many other minerals have been a major factor driving 2% 2% global markets and mineral investment in recent years. In addition, since 2004, and the promulga- 3% 34% tion of its ‘go out’ policy, the Chinese Government 4% has been actively encouraging its companies to invest in mining overseas as a means to secure 4% supplies for its domestic metallurgical opera- tions. Chinese companies have been particularly active in the pursuit of iron-ore investment 5% opportunities overseas, notably in Australia, but they have also invested in other mineral projects such as those for copper, nickel and coal. In 5% addition to China’s direct investment in foreign mining projects, Citi analysts have identified 217 M&A (mergers & acquisition) deals involving Chinese companies in recent years, totalling 14% 18% almost US$50 billion (Citi, 2011a). The Metal- lurgical Miners’ Association of China (MMAC) Iron ore Copper Gold has said it would like to see 40 per cent of imports Zinc Nickel Manganese of iron ore coming from Chinese-invested mines Molybdenum Diamonds Silver by 2015 (China Economic Net, 2011a). Also, as a means to secure adequate supplies of mineral, the Platinum Other Chinese Government operates a strategic stock- pile, the State Reserves Bureau (SRB), to hold and Figure 2.8 Metals value at the mine in 2009 (includes manage supplies of metals it deems critical to its diamonds and uranium). (Data from Raw Materials industrial development such as aluminium, Group, Stockholm, 2010.) copper, nickel and zinc. The particular range of the minerals designated as critical in the US and the EU has comparison, the value of mined cobalt in that important implications for the nature of the year was around US$2 billion, the value of response from the world’s miners. For the most rare earths was around US$1 billion, the value part, these are not minerals of any great interest of antimony was somewhat less than US$1 to the major miners. They are simply too small billion, while the markets for gallium, germa- in terms of their market size. Minerals do not, nium, indium and tantalum combined amounted it should be noted, attract public interest in to less than US$1 billion. The value of rare direct proportion to the scale of their markets. earths production, it might be noted, was Raw Materials Group of Stockholm has calcu- around one per cent the value of iron ore pro- lated that global mine output of metals plus duction, revealing a striking difference between diamonds and uranium was worth around the importance accorded these minerals by US$386 billion at the mine in 2009. Its analysis policy makers and the importance accorded is shown in Figure 2.8. (Note that it excludes them by the industry. coal, which would add very substantially to The large mining companies, having revenues the total.) Two thirds of the total was accounted measured in tens of billions of dollars a year, nat- for by iron ore, copper and gold. By way of urally like to focus their financial resources and The mining industry and the supply of critical minerals 37 management time on commodities which can while in Mongolia, Monros is expanding its oper- make a material contribution to their businesses. ations. In South Africa, Sephaku is expanding its As a result of this, the development of projects operations by developing a new mine at Nokeng producing many of the minerals deemed critical and ENRC is planning a mine at Doornhoek. is often left to smaller companies which, while In Canada, Canada Fluorspar is re-activating the they may be enterprising, often lack the experi- St Lawrence fluorspar mines in Newfoundland. ence, political clout and financial muscle of the In the US, the Klondike II fluorspar mine in big companies, making the route from discovery Livingston County, Kentucky, has been per- to production lengthier and more uncertain. mitted for re-opening. In Vietnam, Dragon Capital These smaller companies also face the challenge Vietnam Resource Investments is building the that the markets for many minor metals lack Nui Phao tungsten-fluorspar project in North transparency. The absence of exchange pricing Vietnam while in Thailand, SC Mining Co is and forward markets for these metals inevitably developing the Doi Ngom deposit in the north of makes some potential financial backers nervous the country. about investment. All of which serves to make the point that That said, smaller miners have been extremely resources of many of the minerals currently active in pursuing projects targeted on these criti- sourced from China are, at a price, available cal minerals and on others facing declining elsewhere in the world. It is just that China’s supplies from China. Thus, for example, in the preparedness in the past to supply these com- case of the rare earth elements, Molycorp modities at low prices made it uneconomic for Minerals has re-opened the Mountain Pass mine many producers elsewhere to do so. Consumers, in California, Lynas Corporation has re-opened who themselves operate in competitive mar- the Mount Weld mine in Western Australia, kets, were opportunistically led towards buying Great Western Minerals Group (GWMG) is re- cheap Chinese minerals, in doing so creating a opening the Steenkampskraal mine in South degree of dependency on China that was, in Africa’s Western Cape, Toyota is planning to retrospect, perhaps unwise. As supplies from open a mine at Dong Pao in Vietnam with Sojitz China have diminished and prices have and the Vietnamese Government, while there is a increased, so miners in these other countries raft of other projects in Canada, Australia and have been granted the opportunity to start, or elsewhere undergoing exploration and evalua- re-start, production. The same applies in the tion. In the case of tungsten, North American case of the critical materials sourced signifi- Tungsten Corp. has re-opened its Cantung Mine cantly from DR Congo, namely cobalt and tan- in the Northwest Territories of Canada and is talum. Substantial resources of these metals evaluating the Mactung deposit in Yukon. There exist outside DR Congo (for tantalum in Brazil are also advanced plans in Australia to develop and Australia, for example). It is just that in the the King Island Scheelite mine on Tasmania and past there was insufficient economic incentive the Molyhil project in Northern Territory. Woulfe for producers in these other regions to grow Mining Corp. is hoping to re-open the Shangdong their output. A diminution of supply out of DR tungsten mine in South Korea. In the UK, Wolf Congo, or heightened concern over political Minerals has conducted a feasibility study on the risk in the country, would provide the required Hemerdon tungsten deposit in Devon and has incentive. raised funds to re-open the mine there. For minerals such as gallium, germanium and Much the same goes for fluorspar, another indium which are recovered as by-products and mineral on the EU’s list of critical minerals. which are sourced from China as a result of In Mexico, there are expansions planned at China’s rapid development as the world’s largest Mexichem Fluor SA de CV (the world’s largest processor of metals, there is very little the min- producer) and at Fluorita de Mexico SA de CV, ing industry can do to relieve supply shortages. 38 david humphrey

Generally, it will not be economic for miners to and result from the lagged supply responses pursue production of these minerals in their own which are an unavoidable feature of the mining right, and the issue of a supply response rests industry. High mineral prices are part of the rather with metals processors outside China and mechanism for transition; they force supply and on the question of whether it is profitable for demand into alignment, in the short run by chok- them to add recovery circuits to existing plants to ing off demand and in the longer run by stimu- produce the relevant metals. In the longer term, lating new supply by encouraging increased it may depend on an ability to find new resources exploration and technical innovation. of these metals or to work different types of ores Policy makers, whose time horizons, being containing them. politically determined, are generally shorter than With respect to large-scale, more basic mineral are those of the mining industry, need to be products, China represents a rather different aware of the underlying reasons why the industry challenge for western states. Certainly it is the suffers from lagged responses and why therefore case that China’s growing demand for these prod- the adjustment to imbalances in supply and ucts has resulted in a tightening of global supplies demand takes time. This is normal, if frustrating. and increased prices, much as it has for more spe- They also need to be aware of the characteristics cialist metals. However, China’s attempts to source of the individual mineral products under threat an increasing amount of minerals from their own and understand better where and how they are overseas mines – which may serve to ease the produced, especially those which come as pressure on global supplies in all regions – is an by-products of other minerals, or flow from the additional competitive pressure on western min- processing of imported ores. Europe, for example, ing companies. Just as western consumers are is a major producer of cobalt but the cobalt it pro- becoming more conscious of competition from duces all comes as a by-product from the processing Chinese manufacturers and purchases of min- of imported copper and nickel ores. Without these erals so western miners are feeling similar pres- metal processing activities in Europe, there would sures in their business. As of the moment, this be no cobalt produced. threat is relatively modest (Ericsson, 2011b) but For miners and explorers to perform their it is one that is likely to increase with time. functions effectively, markets must be allowed to (Humphreys, 2011). operate. Price signals must be reliable and com- panies must be allowed – or, better, encouraged – to respond to these signals. The following comment, Policy issues which was made at a time when governments were busy disengaging themselves from involve- In the main, mineral markets work and deliver ment in the minerals sector, still seems relevant to an appropriate level of supply to mineral users. today’s challenges. Mining is a highly adaptable and enterprising industry and miners are constantly on the “Fair and efficient markets are man-made con- lookout for opportunities to make money by structs. Their effective functioning depends not identifying gaps in the market and filling them. on the absence of policy but on a particular type Although concerns over the availability of min- of policy. They do not create or maintain them- eral supplies in consuming nations are under- selves. Those who decide to trust to the markets for their minerals therefore need to accompany standable when markets are tight, there are few this decision with a commitment to ensure that examples one can point to in history where the mineral products are able to flow without undue non-availability of mineral supplies has resulted hindrance from tariffs, subsidies or spurious envi- in serious economic trauma. The shortages ronmental conditions, and that investment in the which have given rise to concerns over critical industry can similarly flow to where it can most minerals in recent years are largely transitional productively be employed.” (Humphreys, 1995.) The mining industry and the supply of critical minerals 39

With respect to the last point relating to the free minerals which cannot be supplied locally, flow of investment, here there do appear to be there is a need to fight for open and competitive some significant challenges facing the industry. mineral markets both within multilateral While there are few physical impediments to forums and through bilateral agreements. It investment in new mineral supply, there are may be (and this would represent a departure some significant economic and institutional con- from past and present practices) that govern- straints and these appear to be getting more ments should stand up more prominently for severe with time. In particular, there are the their companies where their legal rights are growing pressures from resource nationalism and being flouted by host countries or where they from the growing involvement of the state in the are subject to unfair competition by state- mining sector in many countries. These are owned enterprises. They could also support and matters about which the mining industry can do encourage institutions which provide finance relatively little but which should be of interest to for exploration and mining, including providing policy makers. guarantees for companies making investments While current and recent concerns over criti- in higher-risk countries, and provide more cal minerals naturally lead governments of con- support for educational establishments which suming countries to want to do something to are training up the next generation of mining prevent a recurrence of supply shortages, it carries industry personnel. The best protections against the risk of fighting the last war. As has been sustained mineral shortages are efficiently working shown, problems of minerals supply do not markets and free-flowing investment. Although always come back in the same form or apply to the point has a tendency to get lost in the world the same minerals. Policies such as stockpiling of of political cut and thrust, the US, the EU and critical minerals have a superficial appeal but China are all major mineral consuming and they are cumbersome and costly and have not in importing regions and share a common interest the past proven very effective; in addition, the in healthily supplied global markets. buying of minerals for a strategic stockpile always risks aggravating supply problems by pushing up prices and distorting markets. Notes There are, nonetheless, things that policy makers can usefully do to assist with the adjust- 1. Enterprise value (EV) is calculated as market capi- ment to supply shortages and to support future talisation plus debt, minority interests and preferred industry supply responses. However, these tend shares minus total cash and cash equivalents. to be more long term and structural in nature. 2. In late 2012, Glencore and Xstrata agreed terms for the Governments of consuming countries can, for merger of the two companies, to take place in 2013. 3. These comprise platinum, palladium, rhodium, example, help consumers adjust to mineral ruthenium, iridium and osmium. shortages and accompanying high prices by 4. Since 1997, Zaire has become the Democratic encouraging R&D in materials technologies Republic of the Congo. and facilitating recycling. With regard to supporting future industry supply responses, governments of mineral- References consuming regions can promote local mine pro- duction, where this could be viable. They could Anderson, E.W. and Anderson, L.D. (1998) Strategic also do more to promote the development of Minerals: Resource Geopolitics and Global Geo- new technologies for mining and mineral Economics. John Wiley and Sons, Chichester, UK. processing. Neither of these are things for which Chegwidden, J. and Kingsnorth, D. (2011) Rare earths – an the authorities in the US and the EU have evaluation of current and future supply. Presentation to shown much enthusiasm in recent years. For the Institute for the Analysis of Global Security, 2011. 40 david humphrey

China Economic Net. (2011) http://en.ce.cn/Industries/ Humphreys, D. (1995) Whatever happened to security EnergyandMining/201104/19/t20110419_22371368. of supply? Minerals policy in the post-Cold War shtml world. Resources Policy 21 (2), 91–97. Citi (2011a) The Changing Face of Global Mining Humphreys, D. (2011) Emerging Miners and their Markets. Citibank Global Markets. Growing Competitiveness. Mineral Economics 24 Citi (2011b) Good Assets or Good Management. (1), 7–14. Citibank Global Markets. Maull, H.W. (1986) South Africa’s Minerals: The Crowson, P.C.F. (2011). Mineral reserves and future Achilles Heel of Western Economic Security? minerals availability. Mineral Economics 24, (1), 1–6. International Affairs, RIIA 62 (4). Ericsson, M. (2011a) Private communication. MEG. (2012) Corporate Exploration Strategies. SNL Ericsson, M. (2011b) Mineral Supply from Africa: China’s Metals Economics Group, Halifax, Nova Scotia. Investment Inroads into the African Mineral Resource Mining Journal (2011) Answering the call. Lithium Sector. The Journal of The Southern African Institute Focus, Mining Journal. of Mining and Metallurgy 111 (July). NRC. (2008) Minerals, Critical Minerals, and the US European Commission (2010) Critical Raw Materials Economy. National Research Council, The National for the EU. Report of the Ad-hoc Working Group on Academies Press, Washington D.C. defining critical raw materials. Robinson, P. (2011) Copper – Returning to Normality? Fraser Institute (2011) Survey of Mining Companies: Presentation to CRU Copper Conference, Santiago, 2010–2011, Chile. FT.com (2011) Minmetals warns of a gulf in resources. UBS. (2011) Mining and Steel Primer. UBS Investment FTSE. (2013) Available at http://www.ftse.com/ Research. House of Lords (1982) Strategic Minerals, 20th report of U.S. Geological Survey, USGS. (2012) Mineral Commodity the House of Lords Select Committee on the European Summaries. US Geological Survey. Communities, Session 1981–82, HMSO. 3. Recycling of (critical) metals

CHRISTIAN HAGELÜKEN

Director EU Government Affairs, Umicore AG & Co. KG, Hanau-Wolfgang, Germany

Rationale and benefits The urban mine The world faces major societal challenges Unlike energy raw materials, metals are, in prin- including climate change, energy supply and ciple, not consumed but can be kept in an ‘eternal’ availability of critical resources. Metals are life cycle. After initial extraction from mineral essential to the economy as a whole and to the deposits in the ‘geosphere’ metals are used to products and technologies that will allow construct infrastructure and to manufacture transition towards a society that is able to products. The use of metals, particularly of tech- address these challenges. Technology metals,1 nology metals in industrial and consumer prod- which include the critical metals, play a key ucts, has grown rapidly in recent decades. For role in this context not only for the European example, more than 80 per cent of the global Union but for all industrialised regions in the mine production of the platinum-group metals world. Moving towards a competitive and sus- (PGM), the rare earth elements, indium and tainable economy requires access to these mate- gallium since 1900 took place in the last three rials in adequate quantities and at competitive decades (Hagelüken and Meskers, 2010), and costs. It is increasingly recognised that, among many of these metals are still bound in the ‘tech- others, one important use area for technology nosphere’ or ‘anthroposphere’. As a result, prod- metals is energy generation (Achzet et al., 2011; ucts such as cars, electronics, batteries and Graedel, 2011; and Resnick Institute, 2011). As industrial catalysts have evolved into a potential especially Europe is highly dependent on ‘renewable’ metal resource for the future, often imports of metals and metal concentrates, and referred to as an ‘urban’ or ‘above ground mine’, primary resources from European deposits are that must not be wasted. Moreover, compared to not available in sufficient amounts for most natural ores, metal concentrations in many prod- metals, more recycling is a cornerstone in any ucts are relatively high. For example, a typical strategy for securing long-term supply and primary gold mine will yield around five grams of becoming more independent of geopolitical con- gold per tonne (g/t) of ore. In electronic scrap, this straints, as described in Chapter 2. figure rises to 200–250 g/t for computer circuit

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 42 christian hagelüken

Table 3.1 Average content of precious metals, copper and cobalt in mobile phones and computers, and resulting metals demand from global sales in 2010, compared with world mine production.

a) Mobile phones b) PCs and laptop computers

1600 million units/year, each 350 million units/year, of which ca with a lithium-ion battery 180 million have a lithium-ion battery a + b = Urban mine

Global mine Share a + b of Unit metal Total metal Unit metal Total metal production global mine Metal content content content content (2010) production

Silver 250 mg 400 t 1000 mg 350 t 22,900 t 3% Gold 24 mg 38 t 220 mg 77 t 2650 t 4% Palladium 9 mg 14 t 80 mg 28 t 225 t 19% Copper 9 g 14,000 t 500 g 175,000 t 18 Mt <1% Cobalt 3.8 g 6100 t 65 g 11,700 t 88,000 t 20% t, tonnes; Mt, million tonnes; g, grams; mg, milligrams boards and as high as 300–350 g/t for mobile over 40 different chemical elements including phone handsets, grades which are very uncommon base metals such as copper (Cu), nickel (Ni) and in primary gold deposits. An autocatalyst con- tin (Sn), speciality metals including cobalt (Co), tains some 2000 g/t of PGMs in the ceramic brick, indium (In), and antimony (Sb), and the precious compared to average PGM concentrations of less metals silver (Ag), gold (Au) and palladium (Pd). than 10 g/t in most mines which produce PGMs. Metals, mostly copper, make up about one quarter Taking the high environmental impact of primary of the weight of such a telephone. production of precious metals due to low ore con- In one tonne of mobile phones (equivalent to centrations, difficult mining conditions and other about 13,000 units without batteries) this adds up factors into account, scrap recycling becomes even to an average of 3.5 kg Ag, 340 g of Au, 130 g of more beneficial from an environment protection Pd, and as much as 130 kg of Cu. The value of viewpoint. Furthermore, if state-of-the-art tech- these metals can exceed €10,000 per tonne, with nologies are used, the environmental benefits of more than 80 per cent of the total attributed to the the recycling of scrap metals are further enhanced precious metals. For a single unit, however, the (Hagelüken and Corti, 2010). precious-metal content is only of the order of mil- The challenge, however, is that most consumer ligrams: 250 mg Ag, 24 mg Au, 9 mg Pd on average, products are widely distributed and often difficult while Cu is not more than nine grams and the to trace around the planet and that, although rechargeable battery contains about four grams of metal concentrations are high, the absolute pre- Co (Hagelüken and Meskers, 2008). Thus, the net cious metal content in a single device is often metal value of a single mobile phone is only about very low. Accordingly, in order to economically one euro even at the high metal prices prevailing exploit these urban mines, it is first necessary to in 2011, and so does not per se provide a real gather sufficient quantities of these dispersed economic incentive for recycling. It is the sheer products to create a true above-ground deposit number of mobile phones in use that makes this that will fulfil the second basic criterion for an metal resource significant: about 1.6 billion units economically viable ‘ore body’ apart from its were sold worldwide in 2010 alone (Gartner, concentration, that is sufficient volume. This is 2011), and, by the end of 2010, total global sales illustrated in Table 3.1 using the example of mobile had reached around 10 billion units, roughly phones and computers. Mobile phones contain equivalent to 1.5 telephones for every human Recycling of (critical) metals 43 being currently living on the planet. They contain significance of life-cycle structures’) cannot be a total of 2400 tonnes of Ag, 230 tonnes of Au, and overcome, the potential of this urban mine will be 90 tonnes of Pd. The gold and silver content of the significantly undermined. combined 2010 sales volumes of mobile phones Based on an in-depth analysis of PGM ows and computers are equivalent to approximately from automotive catalysts (GFMS et al., 2005) four per cent of the global mine production of sil- and ongoing industrial experience in the autocat- ver and gold and 20 per cent of palladium and alyst recycling market, I estimate that the cumu- cobalt (Table 3.1). But how much of this ‘urban lated wastage from car catalysts already add up mine’ will eventually be recycled and what is to some 900 tonnes of PGM globally, of which required to do this effectively? Europe accounts for about 200 tonnes. PGM losses The use of metals in cars provides another occur at all stages of the life cycle, including dissi- example of metals potentially available from the pative driving losses ‘through the exhaust pipe’ ‘urban mine’. In cars, the catalyst in the exhaust (if road conditions are bad or the car is not well system has accounted for well over 50 per cent maintained), losses from cars that were not recy- of the mine production of PGMs (in particular cled, losses from catalysts that were not removed platinum, palladium and rhodium) for many years from a car before shredding, losses during (inappro- and the recent proliferation of electronic compo- priate) handling and mechanical pre-treatment of nents in road vehicles (such that a car is effec- car catalysts, and losses during metallurgical tively a ‘computer on wheels’) will have an catalyst processing. increasing impact on the demand for gold, silver, and a number of critical metals. Given the antici- Recycling benefits pated boom in the use of electric vehicles, even more technology metals will be required. These There is a broad consensus today that recycling will include more PGM for fuel-cell stacks, cobalt, offers significant benefits, from an environmental lithium and rare earths for batteries and electric perspective as well as by increasing the supply motors, and bismuth and semiconductor metals, security of metals. It is generally acknowledged such as antimony, tellurium, germanium, for that recycling: thermo-electric applications in which waste heat ● reduces the environmental burden that would is used to generate electric power. Figure 3.1 otherwise occur (by preventing emissions from shows annual demand figures for PGM in auto- discarded products and landfills into soil, water motive catalysts, both for the whole world and and air, and by reducing the use of land needed to separately for Europe. Cumulative gross demand accommodate waste materials); (5430 tonnes PGM globally) and recycling vol- ● mitigates the environmental impact of mining umes (730 tonnes) up to 2010 are also shown. In (by reducing the amount of mining required and Europe, the large increase in demand for PGM in thus decreasing energy demand, carbon dioxide this application started in the mid-1990s. As a emissions, land and water use, and impacts on result, in view of the relatively long lifetimes of the biosphere, e.g. in rain forests, Arctic regions, cars (10–15 years), this means that some 1100 ocean oors); tonnes of PGM are still in use in European cars ● extends the lifetime of and preserves primary and up to 3000 tonnes globally. Most car catalysts geological resources (and thus contributes to that are recycled today are from the late 1990s, so buying time to develop improved mining in the future the potential for recycling PGM from and processing techniques that might facilitate cars on the road today is huge. However, there are the extraction of metals from low-grade or deep major deficiencies in the recycling of automotive ore bodies); catalysts, mainly related to the failure to collect ● reduces geopolitical dependencies arising vehicles at the end of their lives. If these structural where critical metal resources are concen- problems of this ‘open cycle’ (see Section ‘The trated in a few mining countries and/or are in 44 christian hagelüken

(a) 350

Cumulative demand (tonnes),1980–2010 300 Total Recycling Rh 420 64 Pd 2200 285 250 Pt 1900 381 Rh+Pd+Pt 5430 730 200

150

100 Global PGM demand (tonnes/year)

50

0 1980 1984 1988 1992 1996 2000 2004 2008

Pt - Platinium Pd - Palladium Rh - Rhodium

(b) 120

Cumulative demand (tonnes),1980–2010 100 Total Recycling Rh 97 20 Pd 600 62 80 Pt 710 70 Rh+Pd+Pt 1410 152 60

40

20 European PGM demand (tonnes/year)

0 1980 1984 1988 1992 1996 2000 2004 2008

Pt - Platinium Pd - Palladium Rh - Rhodium

Figure 3.1 (a) Gross global demand for platinum, palladium and rhodium for automotive catalysts (annual and cumulative) and cumulative recycling volumes. (b) Gross European demand for platinum, palladium and rhodium for automotive catalysts (annual and cumulative) and cumulative recycling volumes. (Data from Johnson Matthey, 2011.) Recycling of (critical) metals 45 the hands of a small number of companies (recy- ● Extract metals in an efficient manner from cling therefore effectively creates a significant the ores. local resource); ● Use them as efficiently as possible in the ● contributes to the supply security of minor and manufacturing of products. critical metals by partially decoupling their produc- ● Avoid dissipation during use and at their end- tion from the primary production of the associated of-life (EoL). major carrier metals (the strong link between the ● Minimise losses of metals into residues during mine production of carrier metals and by-product all phases of the life cycle. Such residues may minor metals limits the primary supply of impor- include tailings and slags from primary produc- tant technology metals – see Chapters 1 and 2 for tion, production scrap from manufacturing, pro- further discussion); ducts discarded during their useful lifetime, ● supports ethical sourcing of raw materials non-recycled streams, and slags, effluents and (through improved transparency of the supply final waste at the end of life. chain, ensuring that no ‘conict metals’ are Where such residues cannot be avoided they used that originate from war regions in Africa or need to be recycled with high efficiencies. In this elsewhere); way metals can be re-used at a high rate in suc- ● dampens metal price uctuations by improving cessive life cycles. This implies that both scrap the demand–supply balance and limiting specula- generated during production and derived from tion by broadening the supply base; end-of-life products are new sources of raw mate- ● creates significant employment potential, rials. It is therefore desirable to reduce unrecy- including high-technology jobs and infrastructure. cled residues derived from all stages of the life cycle. This demands ‘intelligent’ product design (durable products, appropriate material combina- tions, ease of disassembly, etc.) and the use of Status and challenges of recycling appropriate and innovative processes throughout critical metals the value chain as described later in this chapter. Historic wastes should also be regarded as addi- The metals life cycle tional future resources in addition to primary Fully effective recycling could, in principle, lead geological resources. Moreover, substitution in to an infinite cycle of metal use. The quality of product manufacturing may also help mitigate recycled metals matches that of primary metals raw materials supply constraints, but it could and they are traded accordingly at identical lead to new challenges for product recyclability prices. Hence, once metals are recycled there is or for sourcing the substitutes. The encourage- no need to incentivise their use. In practice, how- ment of increased re-use and extended lifetimes ever, some quantity of metal is inevitably lost of products and materials is also beneficial, from product life cycles because they cease to be providing it ultimately leads to recycling at the accessible for recovery for various reasons. Some end of the extended lifetime. In this sense re-use materials, like plastics, often face a ‘downcycling’ is not an alternative option to recycling, rather it issue because the recycled material is of lower delays the time until recycling should take place quality than the virgin material and hence can (Eurometaux, 2011). only be used in lower-value applications. Primary metal production is not usually The key challenge, therefore, is to minimise considered when recycling is discussed. This metal losses. As Figure 3.2 shows, the metals life misses the fact that improved treatment of tail- cycle starts with exploration and mining which ings, slag, or other side streams from mining, bring materials from the geosphere into the tech- smelting and refining can contribute signifi- nosphere. Thereafter, the basic concept for sus- cantly to resource efficiency and to the supply tainable metals use is simple in principle: of technology metals. Together, improved 46 christian hagelüken

Dissipation Residues

Residues Use

Product New End-of-life manufacture scrap

M et from g als, yclin alloys industrial Rec & c omp materials ounds Residues

Raw materials production Historic wastes Natural from (tailings, landfills) concentrates, Residues resources ores

Figure 3.2 Sustainable use of metals along product life cycles – efficient recycling of residues generated at all stages can minimise use of primary raw materials. (Adapted from Meskers, 2008.) efficiency in primary production and reworking finally from entire out-of spec or obsolete LCD of historic primary wastes, such as mine tail- monitors or photovoltaic modules becomes ings and smelter slags, comprise a large and increasingly difficult. In the early stages of the accessible potential additional source for most manufacturing process there is significant recy- by-product metals, such as indium, germa- cling potential for many technology metals. In nium, molybdenum, rhenium and gallium. the case of indium, germanium and ruthenium, However, for most precious metals, such as the these have been realised to a growing extent in PGM, these inefficiencies in the primary supply recent years, again driven by rising prices. Metal chain have already been largely overcome due losses occurring during the use of a product are to their high value. not generally recoverable due to their mostly dis- In manufacturing, recycling production scrap sipative nature (Hagelüken and Meskers, 2010). usually becomes more challenging when moving Recycling EoL products will be key to achiev- along the production process. For example, while ing sustainable use of metals. This has been recycling indium from sputter target manufac- recognised by governmental bodies such as the turing and spent ITO (indium–tin oxide) targets European Commission (EU-COM, 2012), the is relatively simple, reclaiming the metal from USA (US NRC, 2008; US DOE, 2010) and Japan sputtering chamber scrapings, from broken or seeking to curb waste generation and to use waste out-of-spec liquid crystal display (LCD) glass or as a resource. Recycling of (critical) metals 47

product design as well as innovative business Waste and resource legislation models, well-organised and efficient recycling sys- EU legislation like the Directive on End-of-life tems, and high-performance technical processes Vehicles (ELV) of September 2000 (EU-ELV, over the entire value chain. 2000) and the Directive on Waste Electrical and Electronic Equipment (WEEE) of January 2003 The recycling value chain (EU-WEEE, 2003) demonstrates the approach in Europe. However, this legislation was devel- As shown in Figure 3.3 recycling needs to be con- oped in the 1990s and its main motivation was sidered as a system comprising a number of key environmental protection through the avoid- parts. It goes beyond process technology and ance of hazardous emissions and the recovery requires a complete chain that begins with col- of mass materials, such as steel and plastics. lecting, sorting, and dismantling/pre-processing The recovery of critical metals was hardly con- in order to separate components containing sidered at that time. Hence, these directives valuable metals or to upgrade relevant fractions define only mass-based recycling targets (e.g. 80 prior to final metallurgical processing. per cent of the weight of a car needs to be recy- Each of the main steps involves different stake- cled) and the recovery of critical and technology holders and companies. The process complexity, metals is not specifically addressed.2 The effect the skills and infrastructure requirements, and the of this has, in fact, been counter-productive investment volumes rise significantly along the with the recycling technologies applied to process chain: recover the mass materials actually leading ● Collection: this can be organised by national or in some case to losses of technology/critical municipal take back schemes, original equip- metals.3 ment manufacturer (OEM) systems, commercial The current calculation method for the EU or charity organisations and involves logistical recycling rates has two additional deficiencies: companies. Collection generally takes place on a (1) it considers only the EoL products which are local or regional level and commonly involves recycled within a country, i.e. the major exports small and medium enterprises (SMEs). of old cars or electronics from some countries, ● Dismantling and pre-processing: this uses such as Germany, are not reected in the manual and/or mechanical processes, often in reported recycling rate; and (2) it ignores the combination. It takes place on a local, regional real recycling efficiency of the final (metallur- and inter-regional level, again involving many gical) recycling process, since for this step only SMEs, but also some large waste management input data are considered (regardless of which companies. output of a specific metal is achieved). In other ● Metallurgical metals recovery: comprises a words, the difference in efficiency attained in a combination of pyrometallurgical (smelting) state-of-the-art metallurgical process and a and hydrometallurgical (chemical) processes, primitive ‘backstreet’ operation are not taken followed by a final step of metal refining. The into account. Although for steel, copper and key players are usually large metallurgical precious metals losses are small if state-of-the-art companies, such as Aurubis, Boliden, Dowa processes are used, they can be significant for and Umicore, many of which are also active in many technology metals. Hence, the high processing primary materials. The plant feed is recycling rates reported in the EU for WEEE sourced globally and technological sophistica- or ELV create a false impression that the recy- tion and economies of scale are crucial for cling systems are uniformly advanced and success. effective across the board. To obtain a true The resulting overall recycling efficiency is circular economy for technology metals, new the product of the efficiency of each stage – approaches are needed, which demand better recycling is only as good as its weakest link. 48 christian hagelüken

Efficiency at each stage: 30% X 60% X 95% = 17%

End-of-life Dismantling & Smelting & Recycled Collection products pre-processing refining metals

Reuse Separated components Final waste & fractions

Figure 3.3 The main stages in a recycling chain for consumer products. The total recycling efficiency results from the product of the single stage efficiencies – the example shown is for average gold recovery from electronic scrap.

For example, the overall efficiency of gold and products such as vehicles and computers pose palladium recycling from WEEE in Europe is particular technical challenges (Reuter et al., estimated to be below 20 per cent (Hagelüken 2005; Van Schaik and Reuter, 2012). and Corti, 2010). This reects the combined Recycling technologies for precious metals efficiency of collection (30 per cent), pre-pro- and base metals, such as copper or lead, have cessing (60 per cent) and metallurgical recovery been developed over centuries and are mostly (over 95 per cent) (Figure 3.3). capable of achieving high recovery yields today While recycling of mass metals works when it comes to the final metallurgical step. relatively well and further improvements may Recycling of speciality metals is often connected be possible, there are significant constraints and to base and precious metal metallurgy and still deficits for many technology metals, including offers significant potential for optimisation in some identified as critical. The main deficits many cases. Metals that are dissipated during usually occur at the collection stage (either their path along the recycling chain, for example because the products are not collected at all or into dust fractions, or that are diverted into are exported for sub-standard treatment after an inappropriate final process by unintended collection) and at the pre-processing stage co-separation are lost. Examples are precious (Meskers et al., 2009). In contrast, for many, but metals which end up in steel mills or aluminium not all, metals the metallurgical smelting and smelters. refining stages are already very efficient. In order For speciality metals, the situation is usually to improve metal recycling overall it is necessary worse because it is far more difficult to recycle to develop specialised, systematic approaches technology metals from complex products than along the chain for specific products and metal to recycle mass materials from waste streams combinations. such as bottle glass or steel scrap. While the latter deals with ‘mono-substance’ materials largely without hazardous ingredients, the Recycling challenges former has to cope with ‘poly-substance’ com- Metal combinations in products usually differ positions, including hazardous materials, that from those in primary deposits, which results in are present in complex components within new technological challenges for their efficient complex products. In the one case, the focus is recovery. In many cases, very low concentrations on volume and cost optimisation, while envi- or ‘inappropriate’ combinations of technology ronmentally sound value recovery from mate- metals in certain products will set economic and rials present in low concentrations is of key technical limits to their reclamation. Complex importance in the other case. It is evident that Recycling of (critical) metals 49

GROUP 1 18 1 2 1 H He PERIOD 2 13 14 15 16 17 3 4 5 1 6 1 7 8 9 10 2 Li Be B C N O F Ne

11 12 13 14 15 16 17 18 3 Na Mg Al Si P S Cl Ar 364 5 7891011 1 2 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 4 K Ca Sc Ti V Cr Mn Fe Co Ni Cu Zn Ga Ge As Se Br Kr

37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 5 Rb Sr Y Zr Nb Mo Tc Ru Rh Pd Ag Cd In Sn Sb Te I Xe

55 56 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 * 6 Cs Ba Hf Ta W Re Os Ir Pt Au Hg Tl Pb Bi Po At Rn

87 88 104 105 106107 108 109 11 0 111 11 2 11 4 ** 7 Fr Ra Rf Db SgBh Hs Mt Uun Uuu Uub Uuq

* LANTHANIDE > 50% 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 > 25–50% 6 La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu > 10–25% ** ACTINIDE 1–10% 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 7 Ac Th Pa U Np Pu Am Cm Bk Cf Es Fm Md No Lr < 1%

Figure 3.4 Global end-of-life recycling rates for metals in metallic applications. (Modified after UNEP, 2011.) recycling of technology and critical metals in jewellery, coinage and investment ingot uses, requires a fundamentally different recycling as well as the highly efficient (>90 per cent) chain setup and that a ‘mass-materials approach’ industrial life cycles of PGMs, e.g. in (petro) will inevitably lead to significant losses of trace chemical process catalysts. PGM recycling rates materials. from automotive catalysts are only 50–60 per These challenges are underlined in the report cent, while for gold, silver and PGM in electronic “Recycling Rates of Metals” published by the applications they fall below 15 per cent. As will UNEP Resource Panel (UNEP, 2011). It shows be shown later in this chapter recycling tech- that, while for most base and precious metals nology is not the main reason for these losses. end-of-life recycling rates of over 50 per cent are The recycling potential or recyclability of a achieved on global average, recycling rates for product or material depends on various technical, most speciality metals are extremely low, less economic, structural and organisational factors: than one per cent (Figure 3.4). ● The intrinsic metal value of the base material. Even for the valuable precious metals a divi- This depends on its absolute metal content and sion of recycling rates by main application fields the metal price, and determines the economic reveals significant deficits in most consumer attractiveness of recycling. It sets a benchmark for applications (Figure 3.5). the recycling technology and the overall recycling The high average precious metals recycling costs. Materials containing precious metals often rates are attributed to the very low level of losses make recycling quite attractive. 50 christian hagelüken

EOL recycling Sector-specific EOL recycling rates Jewellery, Importance of end use sector for each metal rates coins listed (% of total gross Vehicles Electronics Industrial Dental Others metal demand)* applications 1) 2) 3) 4) 5) > 50% Ru 5–15 0–5 40–50 0–5 > 25–50% Rh 50–60 45–50 5–10 80–90 30–50 40–50 > 10–25% Pd 60–70 50–55 5–10 80–90 15–20 15–20 90–100 < 10% Ag 30–50 0–5 10–15 40–60 40–60 90–100 nil Os no relevant end use

Ir 20–30 0 0 40–50 5–10 * Including metal demand for closed systems (e.g. Pt 60–70 50–55 0–5 80–90 120 10–20 90–100 process catalysts, glass and 15–20 0–5 10–15 70–90 15–20 0–5 90–100 other industrial applications) Au

1) Total without jewellery, coins (no 3) Including process catalysts/ 4) Including decorative, medical, typical end-of-life managements for electrochemical, glass, mirror (Ag), sensors, crucibles, photographic (Ag) these products). batteries (Ag). In some cases, the photovoltaics (Ag). 2) Autocatalysts, spark plugs, available EOL metal is reduced 5) Including medals and silverware. conductive Ag-pastes, excluding car due to prior in-use dissipation (e.g. electronics. homogeneous Pt-catalysts).

Figure 3.5 Global end-of-life recycling rates for precious metals by important application fields (after UNEP, 2011). The shading in the boxes indicates the importance of that sector as a proportion of total gross demand e.g. for ruthenium (Ru), >50% is used in electronics and between 25–50% in industrial applications. The actual recycling rates achieved are indicated by the numbers in the boxes. (Ag, silver; Au, gold; Ir, iridium; Os, osmium; Pd, palladium; Pt, platinum; Rh, rhodium; Ru, ruthenium.)

● The material composition: this goes beyond The seven conditions for the chemical composition to include physical effective recycling characteristics such as shape, size and the type of connection between materials or components. It For effective recycling of a product, material or inuences the choice of the technical recovery metal, seven basic conditions must be met: methods used for sorting, pre-treatment and 1. Technical recyclability of the material or metal refining. The product or material composition combination. For example, all precious metals, usually has a significant impact on recovery costs most base metals and a number of speciality metals as well as on the technically achievable recovery can be recovered from a printed circuit board if yields. state-of-the-art processes are used, but some speci- ● The application field of a product and how it is ality metals such as gallium, germanium or rare used: the former refers to the area of use in earth elements contained in this mix are not recov- consumer or business markets and the latter to erable (see Section ‘Metallurgical recovery’). new or re-use products, user behaviour, risk of dis- 2. Accessibility of the relevant components. For sipation, product mobility, country of usage, etc. example, an underoor automotive catalyst, a These factors together determine the probability lead-acid car battery, or a personal computer (PC) of particular products and metals entering into an motherboard is easily accessible for dismantling, appropriate recycling channel at their end-of-life. whereas a circuit board used in car electronics (e.g. (see section ‘The significance of life cycle struc- in the engine management system) usually is not. tures’ below) (Hagelüken and Meskers, 2010). As long as such components are dismantled before Recycling of (critical) metals 51 the car is put through the shredder, most metals channelled into a shredder process without prior they contain are technically recyclable (condition 1; removal of the precious metal-containing circuit see Section ‘Collection and pre-processing’). boards. The same applies to the PGM-containing 3. Economic viability, whether intrinsically catalyst in a car or fuel cell. (triggered by a positive net value) or externally 7. Sufficient capacity along the entire chain to (by authorities, retailers, etc. to compensate a make comprehensive recycling happen. Once negative net value) created. Dismantled automo- conditions 1–6 are met, the only requirement is tive catalysts or PC motherboards have a positive to ensure that there is sufficient capacity to pro- net value, therefore recycling is economically cess the volume of material available for recy- viable by itself. In contrast, a dismantled ultra- cling. Metal refiners are willing to invest in thin PGM-coated PC hard disk or a LCD screen building up such capacities provided there is coated with indium–tin oxide (ITO) usually has a sufficient security of feed later on. Conditions 4 negative net value due to the cost of processing it. to 6 are thus crucial to trigger timely investments Recovering the critical metals from such prod- in the growing market for metals recycling. ucts would currently not be economically viable Particular end-of-life products and materials unless paid for externally or subsidised. However, will satisfy certain conditions in this sequence; it should be noted that such external creation of the further they get, the easier it will be to find economic viability is not unusual in recycling, appropriate measures to make use of this recy- e.g. in the case of treatment of household waste cling potential. Figure 3.10, later in this chapter, (paid by consumer fees) or in deposit fund sys- shows a schematic representation of a life cycle tems for empty beer bottles or aluminium cans for a typical critical-metals-containing product, (see also condition 4). highlighting the various points at which losses of 4. Collection mechanisms to ensure the product metal can occur. Each of these provides an oppor- is available for recycling. If collection mecha- tunity for making an improvement to the process nisms are not in place, items such as old PCs or or for adding incentives that encourage recycling mobile phones may end up being stored in house- to take place. holds or discarded into the waste bin for landfill or municipal incineration. The metals they con- tain would effectively be lost to the recycling Recycling technologies chain. 5. Entry into the recycling chain and remaining Before the actual recovery of the technology therein up to the final step. Items such as PC metals can take place, the recycling material motherboards, mobile phones or cars containing needs to be conditioned in most cases as shown catalysts are often sent (either legally or ille- in Figures 3.3 and 3.6. Examples of such condi- gally) to countries without the proper infrastruc- tioning include dismantling and/or mechanical ture for recycling at their end of life. The same pre-processing (e.g. by shredding and sorting) of applies for components or fractions from EoL electronic scrap, de-canning car catalysts (extract- products which escape at a later stage from a ing the catalyst monolith from the steel case) and state-of-the-art recycling chain. This usually burning oil refining catalysts contaminated with results in technology metals being lost to the carbon. Whatever pre-processing is employed, it recycling chain. should be conducted in a way that its output frac- 6. Optimal technical and organisational setup of tions provide an optimal fit to the subsequent this recycling chain. Comprehensive recycling metallurgical recovery processes and that losses chains exist within Europe and in other industri- of valuable substances during pre-processing are alised regions, though it is important that certain minimised. items, such as PCs and mobile phones, are not Different metals require different final path- mixed with other low-grade electronic waste and ways for metallurgical processing. Three main 52 christian hagelüken

Waste electrical & electronic equipment (WEEE)

WEEE- WEEE not collection collected ‘universal’ integrated smelter-refineries for Cu, precious metal & some Pre-processing special metals

Rare Earth Indium Precious Disposal of Co-Li recycling Fe- AI- Cu- Plastics- from hazardous recycling from metal LCD recovery recovery recovery recycling materials from rechargeable recovery magnets batteries screens

Dedicated specialised processes for certain special metal fractions Precious & Slags & special metals other residues

Figure 3.6 Dismantling and pre-processing are crucial to channel substances into the most appropriate metallurgical recovery facilities for final processing. This is illustrated here for waste electrical and electronic equipment. (Al, aluminium; Co, cobalt; Fe, iron; Li, lithium.) routes and a large number of dedicated processes tions that contain the target metals do not reach are currently used (Figure 3.6): the appropriate final process. Hence, the organi- ● Ferrous metals recovery takes place in steel sation of the recycling chain and the management plants. of the interfaces between the main steps are cru- ● Aluminium recovery is conducted in alu- cial for overall recycling success (Hagelüken, minium refineries and re-melters. 2006a). Furthermore any system should be ● Copper, lead, nickel and precious metals are designed for a specific type of material (Van recovered in modern integrated smelter-refineries. Schaik and Reuter, 2010). A number of speciality metals, such as tellurium, selenium and antimony, also fit in this pathway Collection and pre-processing and can be recovered in sophisticated ‘universal processes’ as by-products. Effective collection systems are a prerequisite for However, a number of other speciality metals reclaiming metals and the infrastructure used for are lost in such universal owsheets and require this purpose must be suited to local conditions. dedicated processes, as shown in Figure 3.6. Some Collected EoL products are usually sorted into of the critical metals from the EU list (EC, 2010), several categories, which, in the EU, are pre- such as PGM, antimony and some indium, can be scribed by legislation at the country level. In recovered efficiently via the integrated smelter- order to simplify logistics, attempts are often refinery pathway. The other critical metals require made to reduce the number of categories. dedicated processes, although for all of them met- However, too much reduction will result in a allurgical solutions do exist (for specific challenges heterogeneous mixture of material types, which see Section ‘Metallurgical recovery’). tends to reduce the effectiveness of the sub- It is clear that even the best metallurgical sequent pre-processing and recovery processes. recovery process is useless if the material frac- A balance must be struck between too many and Recycling of (critical) metals 53 too few categories to maximise the overall only 12 per cent, 26 per cent, and 26 per cent, recovery rates of technology metals. A separate respectively (Chancerel et al., 2009; Chancerel, collection category for small ICT equipment, 2010). Automotive catalysts, batteries, high-grade such as mobile phones, digital cameras and USB circuit boards, mobile phones and MP3 players memory sticks, is sensible because they contain thus need to be separated prior to mechanical significant concentrations of technology metals pre-processing to prevent irrecoverable losses. that would be largely lost if mixed with other These components and devices can be fed into small low-grade electric household appliances a smelter-refinery process directly and most such as razors, tooth brushes and toasters. metals recovered with an efficiency of greater Pre-processing is required for most complex than 90 per cent. products in order to prepare for subsequent For larger and for low-grade electronic scrap, effective recovery of the contained metals. It such as small domestic appliances and most must be able to cope with a material feed that audio-visual equipment, direct feeding to a changes over time and includes many different smelter is usually not applicable and some degree models and types of equipment (Reuter et al., of mechanical pre-processing is required. Instead 2005). While this works quite well for base of intensely shredding the material, a coarse size metals, it is much more difficult to achieve for reduction, followed by manual or automated technology metals. Losses during pre-processing removal of circuit board fractions, can be a viable for the various recycling routes and the impact alternative. Trained workers are often able to of different material combinations (product remove certain complex target components more design) need to be quantified to evaluate the selectively than automated sorting. Wherever efficiency of the processes and the potential for trained manual labour is available and afford- improvement, mainly in the technical interface able, it can be a viable alternative to dismantle, between (mechanical) pre-processing and sort and remove critical fractions, such as cir- metallurgy. The complexity of high technology cuit boards or batteries, by hand if this is products leads to incomplete extraction of mate- combined subsequently with state-of-the-art rials at EoL, as these are strongly interlinked. industrial metal recovery processes (UNEP-StEP, For example, precious metals contained in cir- 2009; Wang et al., 2012). Often, better results can cuit boards are associated with other metals in be achieved by a ‘negative selection’: rather than contacts, connectors, solders, etc.; with ceramics removing clearly identified items or components in multi-layer capacitors, integrated circuits of a specific type (e.g. positive selection of circuit (ICs), hybrid ceramics, etc.; with plastics in cir- boards) it can be more effective to remove every- cuit board tracks, interboard layers, ICs, etc. thing which is clearly free of these items. Small-size material connections, coatings and Compared to a direct positive selection here the alloys cannot be separated sufficiently by shred- output stream with target material is still partly ding. Hence, incomplete liberation and diluted with remaining non-target material, subsequent incorrect sorting result in losses of hence the concentration of target material technology metals to side streams, including after sorting is lower. However, it avoids the dust, from which they cannot be recovered by critical slip of value into the wrong fraction. metallurgical treatment (Hagelüken, 2006a; With the reduced amount of the remaining Meskers et al., 2009; Van Schaik and Reuter, less heterogeneous material then a second, 2010 and 2012). more sophisticated (and costly) sorting step An industrial test with mixed IT electronic becomes affordable. A common misconception scrap treated in a modern shredder without prior in mechanical pre-processing is that the aim dismantling of circuit boards revealed that the should be to generate highly concentrated (‘pure’) percentages of silver, gold and palladium ending output streams, since this can usually only be up in fractions from which they could be recov- achieved at the cost of significantly reduced ered (circuit board and copper fractions) were recovery rates. To mitigate this ‘concentration 54 christian hagelüken dilemma’ it is necessary to find the right sorting material mixes are best recovered by initial pyro- intensity, which enables sufficient separation metallurgy (‘smelting’) followed by combined and concentration without causing counter-pro- hydro (‘leaching’) and pyrometallurgical processes. ductive effects on metal yields. For specific Chemically noble metals (such as copper, PGM, material mixes and separation technologies this selenium and tellurium) can be concentrated into can be expressed by the so-called grade-recov- a metallic phase, for example by alloying with ery curve, in the same way that separation copper, lead or nickel. Volatile elements (such as efficiency is measured in the beneficiation of mercury, cadmium, zinc and rhenium) tend to go primary metallic ores (Hagelüken, 2006a; Van to the offgas stream, while metals that oxidise Schaik and Reuter, 2010 and 2012). easily are transferred to the slag phase (REE, For improving the recycling efficiency of rare lithium, aluminium, silicon, magnesium, tan- earth elements (REE), high-strength magnets, talum and germanium). While most metals are such as those used in electric motors and hard extracted via the metallic phase, some can be disk drives, are a potentially important source recovered from offgas or slags if concentrations as they account for a significant share of total are high enough and/or a ‘property hook’ (thermo- REE demand and the concentrations of REE dynamic properties, vapour pressure, density, within them are relatively high. However, magnetic properties, etc.) is available to enable these magnets are usually very brittle, so a the process to be economic. Traces of metal oxides shredder process would break them into many embedded in large volumes of slag volumes com- tiny still highly magnetic pieces which would prising, for example, mostly ceramics or silica, are stick to any iron surface and would not be not economic to extract. accessible for REE recovery. Consequently, Consequently, high-technology metallurgical these magnets must be removed prior to any processes are required to recover technology metals shredding process. from complex components efficiently. For example, Overall, the optimisation of product design, Umicore’s integrated smelter refinery in Antwerp combined with the use of appropriate sorting recovers in its universal owsheet fourteen differ- methods and depth, and routing of the various ent precious and speciality metals together with fractions produced, can lead to a substantial the base metals copper, lead and nickel, which are increase in yields, especially for technology used as metallurgical collectors (Figure 3.7). metals. New developments in products continu- For precious metals from circuit boards or cat- ally lead to new recycling challenges. An example alysts, yields of close to 100 per cent are achieved is the modern car. Such ‘computers on wheels’ in spite of their low concentrations, while, at the include numerous electronic components, con- same time, tin, lead, copper, bismuth, antimony, taining significant amounts of technology metals, indium, selenium and others are reclaimed which are scattered throughout the vehicle. This (Hagelüken, 2006b). Other integrated smelter prevents current shredder technology from recov- processes are operated by Aurubis, Boliden, Dowa ering them, while cost-effective manual disman- and Xstrata (see website information under refer- tling is not currently feasible. ences). Furthermore, in dedicated processes Umicore recovers cobalt, nickel and copper from rechargeable batteries (Dewulf et al., 2010) Metallurgical recovery (Figure 3.8), indium and gallium from photovol- Although in principle, every metal can be recov- taic sputter chamber scrapings (Meskers et al., ered on its own, the mix of substances in a 2010), germanium from wafer production scrap, material stream creates challenges for metallur- and indium from indium–tin oxide (ITO) sputter- gical recovery. The more complex this mix the ing targets. more challenging is the metallurgy and the Research is underway to extend the range of greater are the inevitable metal losses. Complex feed materials and recover additional special Recycling of (critical) metals 55

SULFURIC ACID PLANT

Matte Process Gas BLAST FURNACE

sulfur dioxide

SMELTER Lead Aggregate Slag bullion for concrete Cu bullion LEACHING & ELECTRO- WINNING LEAD Figure 3.7 Metal combinations that fit Speiss REFINERY Ni thermodynamically can be recovered in a sophisticated integrated smelter-refinery process where copper, lead and nickel act Precious metals residues as collectors for precious and some speciality metals. The example shown here is the Hoboken universal process of PRECIOUS METAL SPECIAL METALS Umicore. The large red arrows indicate CONCENTRATION REFINERY where recycling materials can be fed into the process (depending on concentration and properties). The main feed stream goes into the smelter (upper left arrow).

(Ag, silver; As, arsenic; Au, gold; Bi, PRECIOUS METAL bismuth; In, indium; Ir, iridium; Ni, REFINERY nickel; Pd, palladium; Pt, platinum; Rh, rhodium; Ru, ruthenium; Sb, antimony; Sulfuric Cu Ag, Au, Pt, In, Se, Te Pb, Sn, Se, selenium; Sn, tin; Te, tellurium.) acid Pd, Rh, Ru, Ir Sb, Bi, As metals, such as REE and lithium (Umicore-Rhodia, difficult or even thermodynamically impos- 2011). As described in the metal-specific chapters sible. Hence product design, mechanical pre- of this book, further dedicated processes exist for a processing and metallurgical processing are number of other materials, like tantalum recovery linked in a highly interdependent system. The from capacitor scrap or tungsten recovery from laws of thermodynamics determine which drilling tools. metals might ultimately be recovered under For metals that already follow other metal optimal metallurgical conditions (Habashi, streams or can be separated from offgas or efflu- 1997; Nakajima et al., 2011). It is important ents, economic recovery might be achieved that product design avoids – if product func- through adjustments to the ow sheet and/or tionality permits – incompatible combinations by developing appropriate after-treatment and ensures that materials which cannot be steps. In contrast, recovering metals economi- recovered in a common final treatment process cally that oxidise easily and are dispersed as a can be separated during pre-processing (‘design low-grade slag constituent can be extremely for disassembly’). Sophisticated modelling of 56 christian hagelüken

End of life batteries Further production scraps refining of Cu, Ni, Co, Zn, Fe, Mn

Ultra-high Solvent temperature extraction smelting

AI/Ca/Li/REE Co/Cu/Ni/Fe CoCl2 NiSO4 Fe, Mn, Zn, slag alloy Cu removal Figure 3.8 Certain metal combinations & recovery need dedicated processes. The Umicore Cement R&D Oxidation process for lithium-ion and nickel-metal industry for Li & and firing hydride batteries is shown here. (Al, REE with LiCo3 aluminium; Ca, calcium; Co, cobalt; Cu, recovery copper; Fe, iron; Li, lithium; Mn, manganese; Ni, nickel; REE, rare earth LiCoO2 Ni(OH)2 elements; Zn, zinc. CoCl2, cobalt chloride; pure new battery LiCoO2, lithium cobalt oxide; Ni(OH)2, precursors nickel hydroxide; NiSO4, nickel sulfate.) this interlinked system has been undertaken by very serious effects on health and environment, Reuter and Van Schaik (Van Schaik and Reuter, but their efficiency is also very low. An investiga- 2010 and 2012). tion in Bangalore, India revealed that only 25 per The combination of metals with toxic and/or cent of the gold contained in circuit boards was organic substances with halogens in many EoL recovered, compared to over 95 per cent in products requires specialised equipment and integrated smelters (Keller, 2006; Rochat et al., processes. Considerable investment is required 2007). A UNEP report provides a comprehensive for offgas and effluent management to ensure overview on the situation in developing coun- environmentally sound operations, preventing tries (UNEP-StEP, 2009). heavy metal and dioxin emissions. In practice, at Recovering metals from combinations that the present time many recovery plants are not do not occur in nature presents a particular adequately equipped. In particular in the emerg- challenge. Most metallurgical recovery processes ing Asian economies, electronic scrap is often were developed over centuries around the combi- ‘industrially’ treated in non-compliant smelters nations of metal families and gangue minerals, as or leached with strong acids in hydrometallur- shown in the Web of Metals (Reuter et al., 2005). gical plants with untested or ineffective effluent Some processes have been adjusted to secondary management, and with a primary focus on recov- materials but the same laws of chemistry and ering gold and copper. Most electronic scrap in thermodynamics apply. Most primary concen- these countries is handled in the informal sector trates fit ‘automatically’ into one of the in thousands of backyard recycling facilities. This established metallurgical recovery routes. This includes open-air incineration to remove plastics, is not the case for EoL products that contain ‘cooking’ of circuit boards over a torch for de- man-made combinations, which are commonly soldering, cyanide leaching and mercury amal- very different from those found in nature. Once gamation (Puckett et al., 2002 and 2005; Kuper precious metals and speciality metals enter a and Hojsik, 2008; Sepúlveda et al., 2010; Chi steel plant or aluminium smelter it is almost et al., 2011). Not only do these practices have impossible to recover them. Recycling of (critical) metals 57

In most cases, however, metallurgical tech- indium or antimony would not be economically nology itself is not the barrier preventing good viable. In the future, platinum-containing fuel recycling rates. Appropriate processes for new and cell stacks may become an important source for difficult materials (e.g. mobile phones, lithium-ion PGM recycling. If end-of-life products are prop- batteries, new process catalysts, diesel particulate erly collected the main reason for metal losses filters and fuel cells) will be developed as long as in this group is inappropriate pre-processing. there is an economic incentive to do so. The (b) Critical metals, without associated ‘paying’ complex interplay between metallurgy and other precious metals, that can be technically recycled technical and non-technical impact factors on using (new) dedicated processes metal recycling rates is described in depth in the As for the first group, technical recyclability is report ‘Metal Recycling – Opportunities, Limits, a basic requirement. This is affected by material Infrastructure’ (UNEP, 2013) which was pre- composition and complexity as well by the pared by a group of international recycling concentration of these metals and can be techni- and metallurgy experts in the context of the cally more challenging than for precious metals. International Resource Panel. Further differences occur here in the (current) economic viability. Recycling is most economi- cally attractive for relative high-grade concen- Status of recycling of the EU trates or components which contain critical critical metals metals. These include: indium from spent ITO With respect to technical and economical recy- targets; indium and gallium from high-grade pho- clability the critical metals defined by the EU tovoltaic (or other) production scrap; germanium working group (EC, 2010) can be divided into from Ge wafers; beryllium from certain high three main groups: grade production scraps; tantalum from capacitor production scrap; tungsten, cobalt, tantalum and (a) Critical metals combined with precious niobium from machine tools, superalloys and metals that can be recycled economically using alloyed steel (if used in steel alloys, recovery of existing metallurgical processes. these metals generally takes place by alloy recy- The PGMs are the best performers in this cling, without separating and purifying the group because their high price drives them into individual element); and cobalt from certain recycling. Technically, metallurgical yields of rechargeable batteries or process catalysts. well over 95 per cent can be achieved for platinum It is generally not currently economic to and palladium. For rhodium, ruthenium and recycle critical metals where their concentra- iridium metallurgical yields are also high but tions are low or for which new large-scale somewhat lower than for platinum and palla- processes are still under development. These dium. The most important PGM-containing include: REE from lamp phosphors, batteries, products for recycling are car and process cata- magnets;4 indium from LCD screens or photovol- lysts, equipment used in the glass industry, jew- taic panels; cobalt from rechargeable batteries ellery, circuit boards and mobile phones. In the with relative low cobalt content or high disman- latter two uses PGM (mainly palladium) are tling/pre-treatment requirements. Whether new used in combination with other critical metals processes will eventually become economically and with gold and silver. Some sophisticated viable depends on many factors, including future integrated smelter-refinery processes can co- metal prices, available quantities and associated recover antimony and indium from such com- economies of scale, R&D efforts and technical plex metal combinations (Hagelüken, 2006b). progress to improve metal yields and to reduce This co-recovery is triggered by the presence of process costs. Effective pre-processing of the criti- the PGM and/or gold as ‘paying metals’, without cal metal-containing components and materials which the recovery of low concentrations of is an essential prerequisite to ensure they are 58 christian hagelüken directed to the most appropriate metallurgical possible without sacrificing other valuable and processes. For example, without removal of the critical metals although tantalum can be recov- REE-containing magnets from electric motors ered efficiently from many alloys or capacitor recycling of the REE will be impossible. Similarly, production scrap. Research on REE recovery batteries must be removed from electrical devices should focus on magnets, lamp phosphors and and ITO coatings from glass surfaces to create an rechargeable batteries which account for a indium concentrate that is amenable to existing major part of current total demand (Oakdene indium recovery processes (Böni and Widmer, Hollins, 2010; Schüler et al., 2011). Research on 2011). the recovery of REE from car catalysts is unlikely (c) Critical metals as part of complex material to be worthwhile because of the likelihood of mixes with thermodynamic constraints associated negative impacts on the efficiency of Electronic circuit boards are a good example of PGM recycling from these sources. this group. Although antimony, and to a lesser extent indium, thermodynamically fit with the recovery of copper and precious metals, other crit- The significance of life-cycle structures ical metals present in circuit boards do not. Tantalum, gallium, germanium and the REE are The largest challenges for recycling are to over- likely to oxidise in the existing recycling processes come the low levels of collection of consumer and thus would end up highly diluted in the goods and their inefficient handling within the smelter slag. Recovery from this slag would not be recycling chain. The life-cycle structure for economic. Hydrometallurgical processes for the consumer goods differs fundamentally from that treatment of circuit boards do not offer a better of industrial products. This was investigated in a option, since the range and yields of recovered comprehensive research project on the life cycle metals would become worse and treatment of of PGMs (GFMS et al., 2005). Good examples of strong leaching effluents is challenging and costly. efficient industrial cycles (business-to-business, Furthermore, sole hydrometallurgical treatment B2B) are PGM-containing process catalysts. For of complex products will lead to numerous other the PGM contained therein, ownership usually output streams and residues that need to be dealt remains with the industrial user, the product with, ranging from precipitates and unclean resins location is well known, and handling throughout to vapours or sludges. Shifting the focus from the the life cycle is conducted in a professional, precious-metals recovery to tantalum, REE or transparent way. This is known as a ‘closed-loop’ gallium does not make economic sense either recycling system. In contrast, ownership of since it would lead to high losses of the valuable consumer items (business-to-consumer, B2C) precious metals. There are a number of other tends to shift frequently, goods such as mobile examples where incompatible material combina- phones and cars are moved around the globe, tions will inevitably lead to the loss of one critical manufacturers lose track of their devices and the metal or another (e.g. PGM and REE in automo- ow of products becomes almost impossible to tive catalysts). Hence, it is almost inevitable that trace. This forms an ‘open loop’ in which recy- in complex mixes technical conicts of interest cling cannot be guaranteed. Even after an item will exist. Consequently, 100 per cent recycling of reaches the recycling chain, the first steps in critical metals will never be possible and priority particular are not always handled by reputable choices need to be made (see also UNEP, 2013). agencies (Hagelüken et al., 2009). The structural factors for PGM recycling may Such prioritisation has also to be made bet- also be extended to other metals in industrial and ween the target materials for recovery of critical consumer products. Figure 3.9 illustrates the metals. For reasons given above, e.g. tantalum difference between closed and open loop systems. recovery from EoL circuit boards will not be The structure of the latter is considerably more Recycling of (critical) metals 59

(a) Oil refining catalyst Reactor in oil refinery Spent catalyst Product manufacturing Product use in process Product recycling Metal Metal Pre- Metal Charge Use Discharge precursor product (regeneration) processing refining

Metals

(b) Integrated circuits, Printed circuit board, Mobile phone, PC, capacitors, etc battery, display laptop computer

Component Assembly End product User 1 Final user

Alloys, metal compounds Exchange of Return point Precursor Production scrap, component collection rejects, overstock Metals

Metals Dismantling and pre-processing Sorting recovery (possibly multistep)

Metal goes to wrong No removal of Losses/system outflows for product output fraction (e.g. plastics) metal component

Losses/system outflows (component/metals)

Manufacturing Use Recycling logistics Physical metals recovery, refining

Figure 3.9 (a) Closed-loop recycling systems for industrial applications, illustrated by process catalysts. (b) Open-loop recycling systems for consumer goods, illustrated by electronic items. complex, hence the probability of effective recy- petrochemical processing, and in PGM equipment cling of technology metals in these systems is used in the glass industry, are the current bench- much less. The following case studies illustrate mark for PGM recycling, with end-of-life recy- the current recycling rates for PGM materials cling rates comfortably exceeding 80 per cent. from industrial, automotive and consumer Recycling in this case is solely market driven and electronic applications (Hagelüken, 2012). Those is an integral part of the product life cycle. Each of of the seven conditions for effective recycling, as the conditions 1–7 for effective recycling are met, described above, which are not met are also indi- strongly supported by the high value of the PGM cated. All numbers refer to global averages (see materials involved. This limits the need for pri- also Figure 3.4). mary-metal supply for these industries to cov- ering the small life-cycle losses and to keeping up with market growth and new applications. As Case study 1: Industrial PGM applications most industrial users own the PGM contained in Industrial applications, such as catalysts used in their products and subcontract to specialists the the production of fine chemicals and in PGM recovery from spent catalysts and the 60 christian hagelüken manufacture of fresh catalysts, they are also less when end-of-life electronics are exported out of vulnerable to metal price volatility. The metals Europe, but failure at condition 4, in which items are largely turned around on the basis of the are stored by consumers or disposed of through weights involved and market prices impact only municipal waste collection, is also significant. on the small gaps that need to be closed. Conditions 2 and 3, the accessibility and Furthermore, the closed-loop system offers full economic viability of components for recycling, transparency on metal sources, thus avoiding the can also be an issue in some cases, as can use of ‘conict metals’ in these applications. condition 6, inappropriate handling within the recycling chain. The EU WEEE Directive helps to Case study 2: Automotive PGM applications stimulate recycling of electrical and electronic products but its enforcement is often weak. End-of-life recycling rates for PGM in automotive Collection rates in many countries are low and applications have a global average in the range illegal or dubious exports out of Europe are a 50–60 per cent. Automotive PGM recycling in serious problem, especially for IT and telecom- Europe is partly impacted by legislation such as munication equipment, which are of high rele- the EU End of Life Vehicles (ELV) Directive, but vance for critical metals (Sander and Schilling, the dominant driver is economic. As for all PGM 2010; Prakash et al., 2010; Basel Convention, applications, technical recyclability is not a 2011). More transparency and better monitoring problem. However, in automotive applications, of end-of-life chains would improve the rates of the PGM-containing catalyst is only a sub-sys- recycling of these products. Most importantly, a tem of a larger product, i.e. the car, which is shift of focus of the current legislation away from driven by its own market mechanisms at the end mass and towards a more pragmatic approach to of its life. European car catalyst recycling mainly the introduction of treatment standards and a fails at condition 5. Many old cars are exported to mandatory process certification system along the countries outside Europe which lack an appro- recycling chain would help to increase the recy- priate recycling chain, and it is only due to the cling rates of PGM and other technology metals. excellent recyclability and the intrinsic economic value of automotive catalyst recycling that the PGM losses here are not even higher. Better Global flows of old products enforcement of international waste shipment Trading of old equipment and donations to rules to limit the export of genuine scrap cars charity have led to a steady but hardly trans- could enhance recycling rates within Europe. parent ow of EU devices to eastern Europe, Africa and Asia (Buchert et al., 2007; Sander and Schilling, 2010). Whether a product is ready for Case study 3: Electronic PGM applications recycling or will be re-used largely depends on Recycling rates for PGM in electronic applica- where it happens to be – what is regarded as waste tions are currently only about 5–10 per cent. The in Europe or elsewhere in the developed world main driver for recycling in Europe is legislation may well be re-usable in Africa. Dealers take such as the EU Waste Electrical and Electronic advantage of this by ‘exporting for re-use’ and a Equipment (WEEE) Directive. The largest use of fair amount of these exports evade the Basel PGM in electronics is palladium used in circuit Convention waste-export provisions. The Basel boards, and market mechanisms at the end of life Convention restricts the exports of hazardous of products such as PCs, TVs, mobile phones and wastes, which includes most electronics scrap, car electronics play an important role in the rates and hence is only allowed to be exported to of collection and recycling of the circuit boards OECD countries under a notification procedure. they contain. As in the case of automotive cata- However, a ‘product’ (i.e. not a waste) does not lysts, a major challenge occurs at condition 5 fall under the Basel regime, so declaring old Recycling of (critical) metals 61 equipment as a re-usable product is a way of facilities) to state-of-the-art facilities in industri- by-passing the regulation. alised countries. Beyond a significant reduction In fact, used products collected in good faith of hazardous emissions and an increase in the for recycling or re-use are known to vanish in a amount of recycled metals, this approach also dubious way only to resurface in primitive land- generates better net revenues in the developing/ fills or substandard backyard ‘recycling’ facilities transition countries. This “Best of 2 Worlds” con- in developing countries (Sander and Schilling, cept is elaborated in (Rochat et al., 2008; Wang 2010). Insufficient cooperation within the life et al., 2012). cycle and recycling chain, together with poor tracking of product and material streams along Differences in recycling rates the entire chain, can explain why inefficient open and pathways for improvement cycles continue to exist. Mobile phones provide a striking example The main reason for the differing recycling rates (Nokia, 2008): although the technology is in place in the case studies is less a question of market to recycle them economically and with high and legislative drivers, and more connected to the efficiency, the actual global recycling volume step a product reaches in the sequence of condi- (excluding re-use) is less than 2000 tonnes per tions listed above. It is important to note that in annum. This is less than 5 per cent of the global all applications technical recyclability is not an total of 80,000 tonnes per annum which is poten- issue: for example, yields of well over 95 per cent tially available for recycling. Most of the old can be achieved if the PGM-containing material telephones that are collected are exported to reaches a state-of-the-art precious metals refinery. developing countries for re-use (partly justified, Hence, neither PGM product manufacturers nor partly ‘fake re-use’), where efficient recycling PGM refiners can act alone to improve the does not take place at their final end of life. In situation; it is the life-cycle system as a whole many cases, collected phones are exported in and the interactions of the stakeholders within bulk, without prior testing, mainly to Hong Kong. that system which will make the difference. There specialised companies select re-usable and What has been shown here for PGMs is basi- repairable phones from the rest. While the re- cally true for most other critical metals. However, usable phones are traded within Asia and to PGMs benefit from a very good metallurgical Africa, the non-reusable ones mostly go to local recyclability even from low-grade and complex backyard recyclers. In order to overcome this products as well from a high economic attractive- practice mandatory testing of such equipment ness due to their value. As described earlier in before export is required to ensure that all non- this chapter, for metals contained in complex reusable phones are identified as scrap. These mixes with thermodynamical constraints even should remain in Europe and be directed into under an optimal life cycle system setup, a 95 per legitimate recycling channels. cent yield will not be possible. This is elaborated If mobile phones are truly exported for re-use in detail in a UNEP report on opportunities and they can help improve the communication infra- limits of metal recycling (UNEP, 2013). However, structure in developing countries at affordable relatively high metallurgical yields will be pos- prices. On the other hand, the current lack of any sible for these metals once it is assured that the effective global recycling infrastructure inevitably relevant fractions in which they are contained leads to the significant loss of valuable raw mate- are directed into state-of-the-art metallurgical rials. It is therefore advisable to set up global processes. structures which promote the shipment of criti- The only way to develop a true circular cal fractions derived from EoL products from economy for technology metals hence is to shift developing and transition countries (when they the open-loop systems of consumer products are lacking in appropriate metallurgical recovery towards closed loops similar to those that prevail 62 christian hagelüken in industrial applications (Hislop and Hill, 2011). recycling and access to secondary raw materials New business models, closer stakeholder cooper- (Eurometaux, 2010). ation and more transparency on material ows are a prerequisite. Deposit funding systems (e.g. for mobile phones or batteries) can incentivise Conclusion and the way forward consumers to hand back their old devices into an appropriate recycling chain. Leasing models or As illustrated in Figure 3.11 consideration needs other innovative ways of product service systems to be given both to the actual materials that are are another option, especially for products with a used and to how and why they are used in order to longer lifetime. If a manufacturer or retailer achieve high recycling rates of critical metals (see retains ownership of his products and sells also UNEP, 2013). Figure 3.10 shows that losses instead the access to its function he keeps track occur at every stage of the product life-cycle, and of the products throughout their entire time of the factors which inuence the magnitude of use and eventually can take it back. As in the these losses differ widely and must be addressed case example of PGM process catalysts he could individually through appropriate measures rang- then re-use certain parts internally and subcon- ing from product design and technical innovation in tract the remaining to recycling specialists, while recycling, to new business models, training and still maintaining the ownership of the contained education, and socio-political framework condi- metals. By doing so he could partly secure his tions. Substantially improving the recycling rates supply of critical metals for the next product of metals in order to boost secondary supply cycle and largely rule out unethical sourcing requires innovation throughout the life cycle. (conict metals). A manufacturer would also However, the required quantum leap can only be directly benefit from innovative design for achieved by a holistic system approach and ade- recycling and disassembly in his products. He quate policy support which takes the interdepen- further could build up a strong form of contin- dencies of life-cycle steps, impact factors and uous interaction with his customers, using their measures into account. feedback for product improvement and service innovation which should have a positive impact Innovation needs on his economic performance. Hence, innovative business models have the potential to create win– In order to minimise the material losses, win situations for improved resource security improved and systematic knowledge development and business profit. and transfer is needed. In general, education on The above discussion has focused on the PGM, sustainability, legislative support and science- which, due to their high value and perfect technical based innovation are needed. Education on sus- recyclability, occupy a unique position among the tainability will generate engaged professionals critical metals. As shown in the previous section, that are aware of the issues related to recycling for most other critical metals technical and and resource scarcity and that have the appro- economic challenges also need to be overcome. priate technical background to tackle these However, the PGM example underlines that cre- issues. Legislative support is needed to create ating closed-loop systems in combination with a incentives as recycling of some critical metals carefully constructed infrastructure is the key has macroeconomic benefits without being cur- to achieving a circular economy. Without this rently economically viable for every value-chain approach, research efforts to improve recyclability element. Support to transfer innovative labora- and economic viability will largely be wasted. As a tory-scale and pilot processes into industrially contribution to the Raw Materials Initiative of the successful operations is required to reduce major European Commission, Eurometaux and the Öko- investment risks. Legislation, knowledge Institute have proposed 10 measures to improve development and cooperation will accelerate the Disposal Dissipation Manufacturing Use & reuse End-of-life P,SL, P,T P,SL P,SL,LC LC Obsolete/ Waste hibernating dump

Raw materials Conditioning production Collection Excavation Pre-treatment Primary | Secondary P,T P,T,SL P,SL,LC P,T

Impact factors for losses: Processing P Material properties: physical P,T composition; value T Technology: selectivity; efficiency/ Tailings Mining yield; environmental impact P,T SL Societal & legislative: awareness, Waste rock incentives; take-back infrastructure

Ores LC Lifecycle type, closed or open

Losses/sink Process Stocks Main flows Minor flows

Figure 3.10 Life cycles of metals and products, and impact factors for losses at various stages. (After McLean et al., 2010.)

6 S oci 5 P o-e rodu co ct no 4 Ec des m onom ign ic ic v a iab tt 3 Colle ilit ra ction y c /l t og i 2 Di ist v smantlin ics e g/Pr n e-p e ro s

c s

e

s

1 Metallurgy s i n g

Product perspective

Material perspective

Figure 3.11 Recycling success factors – the material and product perspectives must be combined to ensure that the appropriate materials are used in the best way in products that meet the requirements of the consumer. The material perspective focuses on (technical) material properties and related impacts on recyclability, while the product perspective covers mainly socio-economic aspects such as consumer awareness, business models and design aspects. 64 christian hagelüken realisation of breakthrough technologies in dem- products in a way they can be optimally traced onstration projects. Science-based innovation is and recycled. also needed as it is only by thorough under- standing of fundamental processes that sound Resource security as a societal solutions to technical problems can be devel- driver for recycling oped. In order to maximise the impact of fundamental knowledge, alignment between Currently, recycling is driven either by value, when competitive and pre-competitive research is the value of recovered substances is significantly essential. As the challenges faced are huge, the higher than the cost of separating them from a fundamental knowledge build-up needs to be waste stream, or by societal concerns related to the tackled globally in order to combine the world’s environment, human health and human safety, or top research centres. This also requires support to volume aspects of waste streams. A new societal through appropriate legislation and regulation in driver that is gaining importance is the sustainable order to preserve and enhance the existing supply of critical resources. Legislation best knowledge base (Eurometaux, 2011). addresses societal concerns. For example, the The recognition by many governments that WEEE Directive was developed because of the waste should be considered as a resource and not impacts on environment, health and safety and the to consider disposal as an option is a major advance volume of the devices and fractions for which no accomplished in recent years. As a result, waste is value driver was present. With regard to the impor- recycled to recover the valuable materials when tance of technology metals, resource aspects need technology is available, or it is stockpiled while to be an essential part of legislation in the future as waiting for suitable technology to be developed. well. The use of sub-standard recycling technol- For example, instead of sending indium-containing ogies, as well as the large amounts of electronic LCD screens to municipal waste incinerators it waste not entering the recycling chain or being could be an option to store them in controlled lost along the way, are in urgent need of major intermediate depots in order to allow time for the improvement. development of appropriate recycling technologies In the same way that environmental and and to build up an adequate volume of material health dimensions alone are not powerful enough suitable for recycling which later could form the drivers to promote recycling, market forces alone base load for a new process. will not be sufficient to improve the recycling of In order to improve this process good docu- critical metals. Favourable economic conditions mentation of waste stream compositions is cannot be relied upon, and the metals from EoL needed in order to incorporate the material in the devices will inevitably be lost in the meantime. recycling ow sheet more effectively and to A policy framework for the sustainable supply of improve separation in temporary stockpiles critical metals from secondary sources is there- (avoiding the formation of diluted complex waste fore needed to address this societal concern. This streams). Furthermore, better and smarter collec- should also include economic incentives to tion is generally needed to ensure better separa- trigger recycling of important waste streams tion in the early stages and to increase the fraction which contain critical metals. of materials that enter the recycling stream directly. This goes hand-in-hand with the preven- Mining and recycling as tion of dubious and illegal export of waste complementary systems streams. Two major challenges ahead are to treat the complex waste streams available in landfills As has been shown, some losses of metals are and to further improve collection and separation inevitable, even in an idealistic static system by more recycling-oriented design and tagging where demand remains at a constant level. Recycling of (critical) metals 65

(a) Static system (b) Dynamic system

Year 3 2 Year Year 1 Inevitable loss Year 0

n n o io ti t c c u u U U d d s s o o r e r e P P

Inevitable loss

Inevitable loss

Recycling Recycling (0) (1) (2) (3) Inevitable loss Inevitable loss Mining Mining

160 160 140 140 120 120 100 100 80 80 60 60 40 40

Market demand % 20 20 Market demand % 0 0 012345678910 012345678910 Time Time Recycling Mining

Figure 3.12 Mining and recycling are complementary systems to cover inevitable losses and market growth.

These losses would occur even with 100 per cent In reality the system is dynamic, i.e. prod- collection and using best available technology ucts have a lifetime that needs to be bridged throughout the chain (Figure 3.12). In addition, before end-of-life recycling can take place and economic constraints, which weigh raw material while market demand for raw materials may be value against recycling costs, will increase losses growing. In this case even immediate recycling, further, usually significantly beyond the if it were possible, would not be sufficient to technical minimum. Such losses will therefore meet demand (Figure 3.12b). Hence, although occur again in every new product/raw materials recycling is a cornerstone in securing supply cycle and eventually the initial raw materials of critical metals it cannot solve it alone. We stock will be used up and will need to be need to use and combine the complete ‘sus- complemented by supply from primary mined tainability toolbox’ in an optimal way, ores (Figure 3.12a). including mining, recycling, substitution and 66 christian hagelüken resource savings (both through higher effi- for consumer products, there will need to be a ciencies in material use and more sufficiency in gradual shift towards more industrial style life styles). It is not helpful to play one of these practices, which means that new business ‘tools’ off against the others, rather the ‘low- models will need to be introduced to provide hanging fruits’ in each part of the life cycle strong incentives for returning products at should be identified and appropriate measures their end of life. This may include deposit fees prioritised accordingly. on new products or innovative product service systems. For emerging technologies (such as Conclusions fuel cells, electric vehicle batteries and photo- voltaics), setting up ‘closed-loop structures’ The potential for recycling critical metal- containing from the beginning will be essential, and man- waste can be thought of as an ‘urban mine’, and can ufacturers that put successful models in place complement the primary source of metal supply will be able to secure their supply of critical from mining. However, in order to fully utilise this metals for the future. In the EU (and Japan) the potential source of secondary supply, the following principal understanding about the need is more changes will need to take place: advanced than in other regions of the world Attitudes need to change from ‘waste management’ (EU-COM, 2008; EU-COM, 2012). Hence EU to ‘resource management’ to ensure the collec- and national legislation should act as frontrun- tion and appropriate treatment of end-of-life ners here, by encouraging manufacturers and products and to encourage the enforcement of consumers in that direction and simulta- legislation. This is particularly important where neously creating supportive regulatory and economic or environmental drivers are cur- administrative conditions for the implementa- rently absent. tion of new models. Targets need to be adapted accordingly, with emphasis on the quality and efficiency of recy- cling processes and the recovery of precious and Notes other critical metals, rather than on the overall mass of materials such as plastics or steel. 1. The term ‘technology metals’ is used in this chapter Recycling practice needs to reect the new require- for the broader group of speciality and precious metals which are – due to their specific physical- ments. In place of the traditional structures of chemical properties – essential for the functionality a scrap business, high-tech recycling can sit of high-tech products. alongside clean-tech manufacturing and renew- 2. At the end of 2011 the revision of the WEEE direc- able energy generation in terms of company tive was agreed between the EU parliament, the structures, appearance and stakeholder cooper- council and the EU commission. It contains a ation, with increased emphasis on transpar- number of improvements (e.g. own-collection cate- ency and business ethics. Reality in many cases gory for small WEEE, increased collection targets, is still far from this and the recycling industry increased reporting requirements of input and as a whole has to undergo fundamental changes. output data of all actors involved in the recycling This will probably leave many traditional chain, stricter provisions to tackle illegal exports, players behind unless they adapt to the new development of harmonised standards for treatment operations) (EU Parliament News 2012; EU requirements. Legislative Observatory 2012) The manufacturers’ vision needs to change as 3. For example, if one aims to achieve the required 65 well. Rather than a burden imposed by legisla- per cent recycling rate for mobile phones, it will, tion, recycling can be seen as an opportunity with a narrow definition of recycling, be necessary to for manufacturers to sustainably increase recover the plastic fraction. In a mechanical process access to the raw materials needed for their this only can be achieved by intense shredding and future production. To close the recycling loop sorting of the phones which will lead to high losses Recycling of (critical) metals 67

of gold and other technology metals, which occur in rials, Enterprise and Industry Directorate General. traces only (Chancerel et al., 2009). (http://ec.europa.eu/enterprise/policies/raw-materials/ 4. A comprehensive overview on the use and recycling documents/index_en.htm) challenges for REE is given in Schüler et al. (2011). EU-ELV (2000) European Parliament and Council Directive 2000/53/EC of the European Parliament and the council of 18 September 2000 on end of life vehi- References cles, Official. Journal of the European Communities, 21-10-2000, Brussels Belgium. Achzet, B., Reller, A., Zepf, V., Rennie, C., Ashfield, M. EU-WEEE (2003) European Parliament and Council and Simmons, J. (2011) Materials Critical to the Directive 2002/96/EC of the European parliament Energy Industrie. An introduction. University of and of the council of 27 January 2003 on waste Augsburg. electrical and electronic equipment (WEEE). Official Aurubis: www.aurubis.com/en/our-business/raw-materials/ Journal OJ L 37 of 13.02.2003, Brussels Belgium. recycling/technology/ EU-COM (2008) 699 Communication from the Basel Convention (2011) Where are WEEE in Africa? Commission to the European Parliament and the Findings from the Basel Convention e-Waste Africa Council. The Raw Materials Initiative – Meeting our Programme. Secretariat of the Basel Convention, Critical Needs for Growth and Jobs in Europe, Brussels Geneva/Switzerland (Dec. 2011) (2008). Boliden: www.boliden.se EU-COM (2012) 82: Communication from the Böni, H. and Widmer, R. (2011) Disposal of Flat Panel commission to the European Parliament, the Council, Display Monitors in Switzerland, EMPA, March the European Economic and Social Committee and 2011. the Committee of the Regions. Making Raw Materials Buchert, M., Hermann, A., Jenseit, W., Stahl, H., B., Available for Europe’s Future Well Being - Proposal for Osyguß, B. and Hagelüken, C. (2007) Optimisation of a European Innovation Partnership on Raw Materials, precious Metals Recycling: Analysis of exports of used Brussels 2012. vehicles and electrical and electronic devices at EU Parliament News (2012) Better management of e-waste Hamburg port, Dessau: Federal Environmental Agency www.europarl.europa.eu/news/en/headlines/content/ of Germany. 20120106FCS34949/8/html/Better-management-of- Chancerel, P. (2010) Substance ow analysis of the recycling e-waste. of small waste electrical and electronic equipment – EU Legislative Observatory (2012) European Parliament An assessment of the recovery of gold and palladium. 2008/0241 (COD) – 19/01/2012 Text adopted by Doctoral thesis, Technische Universitaet Berlin, Parliament, 2nd reading, 19.1.2012. Germany http://opus.kobv.de/tuberlin/frontdoor.php? Eurometaux (2010) Eurometaux’s Proposals for the Raw source_opus=2590. Materials Initiative, Darmstadt / Brussels, 11th June Chancerel, P., Meskers, C., Hagelüken, C. and Rotter, S. 2010 (http://www.oeko.de/oekodoc/1069/2010-115-en. (2009) Assessment of Precious Metal Flows During pdf and appendix). Preprocessing of Waste Electrical and Electronic Eurometaux (2011) Roadmap for securing a sustainable Equipment. Journal of Industrial Ecology 13 (5), 791–810. supply basis for Europe (draft), Brussels, Nov. 2011. Chi, X., Streicher-Porte, M., Wang, M.Y.L. and Reuter, Gartner (2011) Gartner Says Worldwide Mobile Device M.A. (2011) Informal electronic waste recycling: A Sales to End Users Reached 1.6 Billion Units in 2010; sector review with special focus on China. Waste Smartphone Sales Grew 72 Percent in 2010 Gartner Management 31, 731–742. Newsroom, February 9, 2011. http://www.gartner. Dewulf, J., Van der Vorst, G., Denturck, K., Van com/it/page.jsp?id=1543014 (July 15, 2011.) Langenhove, H., Ghyoot, W., Tytgat, J. and Vandeputte, GFMS, Umicore and Öko Institute (2005) Materials K. (2010) Recycling rechargeable lithium ion batteries: Flow of Platinum Group Metals; GFMS: London, critical analysis of natural resource savings. Resources 2005. Conservation and Recycling 54, 229–234. Graedel, T.E. (2011) On the Future Availability of the DOWA: www.dowa.co.jp/en/jigyo/metalmine_summary. Energy Metals. Annual Review of Materials Research html 41, 323–335. EC (2010) Critical raw materials for the EU: Report of the Habashi, F. (1997) Handbook of Extractive Metallurgy. Ad-hoc Working Group on defining critical raw mate- Wiley-VHC, Weinheim. 68 christian hagelüken

Hagelüken, C. (2006a) Improving metal returns and eco- Meskers, C.E.M., Hagelüken, C., Salhofer, S. and efficiency in electronics recycling – a holistic approach Spitzbard, M. (2009) Impact of pre-processing routes for interface optimisation between pre-processing on precious metals recovery from PCs. Proceedings of and integrated metals smelting and refining. In: the European Metallurgical Conference, 527–540. Proceedings of the IEEE International Symposium Meskers, C.E.M., Vandenbroeck, K., Vliegen, J., De on Electronics and the Environment, 8–11 May, Ruijter, I., Dalle, T. and Rigby, P. (2010) Recycling San Francisco, 218–223. technologies to close the loop for PV materials. In: Hagelüken, C. (2006b) Recycling of Electronic Scrap at Proceedings 25th European Photovoltaic Solar Energy Umicore’s Integrated Metals Smelter and Refinery. Conference and Exhibition, 5th World Conference on World of Metallurgy – Erzmetall 59 (3), 152–161. Photovoltaic Energy Conversion, Valencia, 3683–3687. Hagelüken C., Buchert, M. and Ryan, P. (2009) Materials Nakajima, K. Takeda, O. Miki, T. Matsubae, K. and ow of platinum group metals in Germany. Nagasaka, T. (2011) Thermodynamic analysis for the International Journal of Sustainable Manufacturing controllability of elements in recycling process of 1 (3), 330–346. metals. Environmental Science and Technology 11, Hagelüken, C. and Meskers, C.E.M. (2008) Mining our 4929–4936. computers – opportunities and challenges to recover Nokia (2008) Global consumer survey reveals that scarce and valuable metals from end-of-life electronic majority of old mobile phones are lying in drawers at devices. In: Reichl, H., Nissen, N., Müller, J. and home and not being recycled. Nokia Press Room. July Deubzer, O. (eds.) Electronics Goes Green 2008+. 8, 2008. http://press.nokia.com/2008/07/08/global- Fraunhofer IRB Verlag, Berlin, 623–628. consumer-survey-reveals-that-majority-of-old- Hagelüken, C. and Meskers, C.E.M. (2010) Complex mobile-phones-are-lying-in-drawers-at-home-and- lifecycles of precious and special metals. In: Graedel, not-being-recycled/ (18th September 2011). T. and Van der Voet, E. (eds.) Linkages of Sustainability; Oakdene Hollins (2010) Lanthanide Resources and MIT Press: Boston. Alternatives. A report for Dept. for Transport and Hagelüken, C. and Corti, C.W. (2010) Recycling of gold Dept. for Business, Innovation and Skills, Aylesbury, from electronics: Cost-effective use through ‘Design UK, March 2010. for Recycling’, Gold Bulletin 43 (3), 209–220. Prakash, S., Manhart, A., Amoyaw-Osei, Y. and Hagelüken, C. (2012) Recycling the Platinum Group Agyekum, O. (2010) Socio-economic assessment and Metals: A European Perspective. Platinum Metals feasibility study on sustainable e-waste management Review 56 (1), 29–35. in Ghana, Öko-Institut Freiburg. Hislop, H. and Hill, J. (2011) Reinventing the wheel: A Puckett, J., Byster, L., Westervelt, S., Gutierrez, R., circular economy for resource security. Green Davis, S., Hussain, A. and Dutta, M. (2002) Exporting Alliance, London 2011. Harm - The high-tech trashing of Asia; Basel Action Johnson Matthey (2011) PGM Market review archive. Network and Silicon Valley Toxics Coalition: Seattle, www.platinum.matthey.com/publications/pgm- February 25. market-reviews/archive/platinum-2011/ Puckett, J., Westervelt, S., Gutierrez, R. and Takamia, Keller, M. (2006) Assessment of gold recovery processes Y. (2005) The digital dump:-exporting re-use and in Bangalore, India and evaluation of an alternative abuse to Africa; Basel Action Network: Seattle, recycling path for printed wiring boards. M.Sc. thesis, October 24, 2002. ETH Zürich, Zürich, Switzerland. Resnick Institute Report (2011) Critical Materials for Kuper, J. and Hojsik, M. (2008) Poisoning the poor. Sustainable Energy Applications. Californian Electronic waste in Ghana; Greenpeace International: Institute of Technology, Pasadena, Sept. 2011. Amsterdam. Reuter, M.A., Boin, U.M.J., van Schaik, A., Verhoef, E., McLean, H., Duchin, F., Hagelüken, C. et al. (2010) Heiskanen, K., Yang, Y. and Georgalli, G. (2005) The Mineral Resources: Stocks, ows, and prospects. In: Metrics of Material and Metal Ecology. Elsevier, Graedel, T., and van der Voet, E. (eds.) Linkages of Amsterdam. Sustainability. Strüngmann Forum Report, Volume. 4. Rochat, D., Hagelüken, C., Keller, M. and Widmer, R. Cambridge, MA: MIT Press. (2007) Optimal recycling for printed wiring boards. Meskers, C.E.M. (2008) Coated magnesium designed In: Hilty, L.M., Edelmann X., and Ruf, A. (eds.) India. for sustainability? Ph.D. thesis, Delft University of R’07, World Congress – Recovery of materials and Technology, Delft, the Netherlands, 2008. energy for resource efficiency. Recycling of (critical) metals 69

Rochat, D., Rodrigues, W. and Gantenbein, A. (2008) Metal Flows to the International Resource Panel. India: Including the existing informal sector in a Graedel, T.E., Alwood, J., Birat, J.-P., Reck, B.K., clean e-waste channel. Proceedings of the Waste Sibley, S.F., Sonnemann, G., Buchert, M. and Management Conference (WasteCon2008), Durban, Hagelüken, C. www.unep.org South Africa. UNEP (2013) Metal Recycling – Opportunities, Limits, Sander, K. and Schilling, S. (2010) Transboundary ship- Infrastructure. In: Reuter, M. et al. A report of the ment of waste electrical/electronic equipment/ Working Group on the Global Metal Flows to the electronic scrap – optimisation of material ows and International Resource Panel. control, Federal Environment Agency Report No. US DOE (2010) Department of Energy. Critical (UBA-FB) 1331, Dessau. Germany. Materials Strategy. Washington, Dec 2010. Schüler, D., Buchert, M., Liu, R., Dittrich, S. and Merz, C. US NRC (2008) National Research Council: Minerals, (2011) Study on Rare Earths and their Recycling. Final Critical Minerals, and the U.S. Economy. National Report for The Greens/EFA Group in the European Academic Press, Washington. (www.nap.edu/ Parliament. Öko-institut e.V. Darmstad, Germany. catalog/12034.html) Sepulveda, A., Schluep, M., Renaud, F., Streicher, M., Van Schaik, A., and Reuter, M.A. (2010) Dynamic mod- Kuehr, R., Hagelüken, C. and Gerecke, A.C. (2010) A elling of e-waste recycling system performance based review of the environmental fate and effects of haz- on product design. Minerals Engineering 23, 192–210. ardous substances released from electrical and electronic Van Schaik, A. and Reuter, M.A. (2012) Shredding, sort- equipment during recycling: examples from China and ing and recovery of metals from WEEE: linking design India. Environmental Impact Assessment Review 30 to resource efficiency. In: Goodship, V. and Stevels, A. (1), 28–41. (eds.) Waste electrical and electronic equipment Umicore/Rhodia (2011) Press release. CP-2011-18-R (WEEE) handbook, Woodhead Publishing Ltd, Umicore and Rhodia develop unique rare earth Cambridge, UK, 163–211. Recycling process for rechargeable batteries. 16th Wang, F., Huisman, J., Meskers, C.E.M., Schluep, M., June 2011 www.rhodia.com Stevels, A. and Hagelüken, C. (2012) The Best-of-2- UNEP StEP Report (2009) Recycling – From E-waste to Worlds philosophy: Developing local dismantling and Resources. Berlin, Germany. (www.unep.org and global infrastructure network for sustainable e-waste www.step-initiative.org). treatment in emerging economies. Waste Management UNEP (2011) Recycling Rates of Metals – A status WM 05/2012; 32 (11), 2134–46. report, A report of the Working Group on the Global Xstrata: http://www.xstrata.com/ 4. Antimony

ULRICH SCHWARZ-SCHAMPERA

Federal Institute for Geosciences and Natural Resources (BGR), Stilleweg, Hannover, Germany

Introduction rhombohedral crystalline structure in the hexag- onal system and forms elongated crystals with Antimony metal, also referred to as regulus for a perfect cleavage along (010). The melting point impure forms of the element, and the natural sul- of antimony metal is very low (630.6 °C). Its fide of antimony, stibnite, have been known since common oxidation states are Sb5+, Sb3+, Sb0, and 4000 B.C. The metal was used as a coating to harden Sb3-, although other oxidation states have been copper between 2500 B.C. and 2200 B.C. The sulfide recognised. Unlike typical metals, it is not mal- was used as eyebrow paint in pre-dynastic Egypt leable, but hard and brittle and can be crushed and early biblical times. Antimony is rarely found to a powder. Together with the other semi-metals in nature as a native metal because of its strong (silicon, germanium, arsenic and tellurium), it is affinity for sulfur and metals like copper, lead and positioned in Period 5, Group 15 of the Periodic silver. In fact, it is believed that the name ‘anti- Table along the boundary between the metals mony’ is derived from the Greek words ‘anti’ and and the non-metals. Selected key characteristics ‘monos’, which together mean ‘not alone’. and physicochemical properties of antimony are Today, antimony has a range of industrial uses in listed in Table 4.1. batteries, chemicals, ceramics and glass but by far The geochemical properties of antimony are the most important is in fire retardants. The eco- such that it tends to occur in nature with base nomically most important ore mineral and principal metal Groups 11 (Cu, Ag), 14 (Sn, Pb), and with source for the production of antimony is stibnite. Groups 15 (Bi) and 16 (Se, Te) of the Periodic Table. In ore deposits it is closely associated with the precious metals, gold and silver, and in minerals, Definitions and characteristics antimony usually occurs with sulfur in a trivalent state (Sb3+). Natural antimony consists of a mix- Antimony (chemical symbol Sb) is a lustrous ture of two stable isotopes that have the atomic silvery white, brittle, crystalline semi-metal (or weights 121 (57.25 per cent by weight of total) and metalloid) that exhibits poor conductivity of 123 (42.75 per cent); in addition, 29 radioactive iso- electricity and heat. Antimony has a face-centred topes with mass numbers 104 to 136 are known.

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Antimony 71

Table 4.1 Selected properties of antimony. Table 4.2 Antimony minerals with antimony concen- trations greater than 70% Sb and additional antimony Property Value Units minerals (in bold) which typically occur in antimony- bearing ore deposits. Symbol Sb Atomic number 51 Mineral Formula % Sb Atomic weight 121.75 Density at 25°C 6692 kg/m3 Antimony Sb 100.00

Hardness (Mohs scale) 3.0 Senarmontite Sb2O3 88.39

Melting point 631 °C Valentinite Sb2O3 83.53

Boiling point 1587 °C Nisbite NiSb2 80.58

Specific heat capacity at 25°C 0.21 J/(g °C) Onoratoite Sb8O11Cl2 79.78 3+ 5+ Electronegativity (Pauling scale) 2.05 Cervantite Sb Sb O4 79.19 3+ 5+ Electrical potential 0.21 V Stibiconite Sb Sb 2O6(OH) 76.37

Electrical resistivity at 25°C 0.40 µΩ m Sarabauite CaSb10O10S6 75.62

Thermal conductivity 25 W/(m °C) Kermesite Sb2S2O 75.24

Latent heat of fusion 163 kJ/kg Coquandite Sb6O8(SO4) (H2O) 75.11

Latent heat of vaporisation 1602 kJ/kg Stibnite Sb2S3 71.68 Breithauptite NiSb 67.47 Stibarsen SbAs 61.91 Abundance in the Earth Gudmundite FeSbS 58.07 Ullmannite NiSbS 57.29 Berthierite FeSb S 56.94 The abundance of antimony in the Earth’s crust is 2 4 Aurostibite AuSb 55.28 estimated to be about 0.2 ppm for the continental 2 Chalcostibite CuSbS2 48.81 crust, slightly less than arsenic and thallium. Jamesonite Pb FeSb S 35.39 Concentrations in oceanic basalts vary from 0.02 4 6 14 Tetrahedrite (Cu,Fe)12Sb4S13 29.64 to 0.8 ppm. Antimony is a moderately siderophile Famatinite Cu3SbS4 27.63 element which behaves like moderately incom- Dyscrasite Ag3Sb 27.34 patible and lithophile elements (such as the light Boulangerite Pb5Sb4S11 26.44 rare earth elements) during magmatic processes. Bournonite PbCuSbS3 24.91 Pyrargyrite Ag SbS 22.48 The volatile behaviour of antimony during sub- 3 3 Freibergite (Ag,Cu,Fe) (Sb,As) S 18.93 duction processes and crust formation is similar to 12 4 13 Stephanite Ag SbS 15.42 that of lead. The meteoritic abundance of anti- 5 4 Vinciennite Cu Fe Sn(As,Sb)S 3.83 mony is 0.142 ppm for chondrite (CI), while sea- 10 4 16 water contains 0.15 ppb. Geochemically, antimony is related to mercury and arsenic. In ore-forming processes, antimony is highly volatile and is chal- Antimony substitutes for bismuth, lead, arsenic cophile in character. An enrichment of about and sulfur in a variety of ore minerals and tends 150,000 times its crustal content is required to to concentrate in sulfide ores along with copper, reach potentially economic concentrations of lead and silver. Antimony rarely occurs in native about three weight per cent, or 30 kg of antimony form and the most important antimony minerals per tonne of antimony ore. in economic deposits are the sulfide stibnite, var- ious complex sulfosalts (berthierite, boulangerite, stephanite, jamesonite, bournonite, tetrahedrite, Mineralogy freibergite, gudmundite, ullmannite), antimonides (dyscrasite, breithauptite), and oxides (kermesite There is a wide variety of antimony minerals in and valentinite/senarmontite) (Table 4.2). natural systems. More than 264 different antimony- Stibnite is the principal source of mined bearing mineral phases have been defined so far. antimony and, although it is widely distributed, 72 ulrich schwarz-schampera ore-grade concentrations are not common. In deposits; and (ii) complex polymetallic deposits major antimony deposits like the Chinese with variable contents of elements of the ‘epith- Xikuangshan deposit in central Hunan, and the ermal suite’ including gold, silver, tellurium, Antimony Line, Murchison belt, South Africa, selenium, mercury, arsenic, antimony and thal- stibnite is the dominant ore mineral (Figure 4.1). lium, and, locally, base metals (copper, lead, bis- Complex sulfosalts occur as trace minerals in a muth, zinc). wide variety of ore deposits. The mineralogy of Antimony deposits are commonly associated antimony-bearing deposits is often characterised with active continental plate margins and oro- by complex intergrowths containing significant genic belts with steep geothermal gradients due concentrations of metals like copper, lead, sil- to enhanced magmatic activity. Collision tec- ver, bismuth, selenium, arsenic and mercury. tonics, mobilisation from the subducting slab, The close mineralogical association with silver magmatic recycling, and metallogenetic processes and the affinity for gold deposits makes anti- within the continental crust or continental frag- mony a useful indicator for the precious metal ments are most important. The majority of potential of some ore-deposit types. Where stib- antimony-bearing ore deposits is associated with nite has been exposed at the surface, it is com- the subduction-related western Pacific plate monly weathered to an oxide phase such as boundaries, especially in east and south-east bindheimite, kermesite, stibiconite or valenti- Asia. Wu (1993) defined three so-called ‘antimony nite/senarmontite. belts’ in China. A second antimony province can be delineated at the Nazca-South American plate boundary in Bolivia and Peru, which includes Major deposit classes more than 500 antimony and antimony–gold deposits, and which continues along the western Antimony occurs in several different types of plate margins of North America. Other provinces ore deposits of all ages, from the presently form- are related to different metallogenic epochs in ing fumarole precipitates of active volcanoes, central Europe covering the Hercynian and for example in New Zealand, to deposits in the Alpine belts. Archaean volcanic strata of greenstone belts, for The highest antimony concentrations com- example in Canada and South Africa (Table 4.3; monly occur in low-temperature magmatic- Figure 4.2). However, ore-grade antimony con- hydrothermal systems in the epithermal centrations are not common and economically environment. Antimony is typically enriched in exploitable deposits of stibnite are generally the distal portions of these systems at shallow small and discontinuous. depths and close to the surface. It is closely asso- Most antimony deposits are of hydrothermal ciated with hydrothermal silica and carbon origin (Obolensky et al., 2007). Three main anti- dioxide, and the common wall-rock is carbonate, mony deposit types can be distinguished, based either as sedimentary limestone or as a prominent on fluid generation and metal source: (i) low- hydrothermal alteration mineral phase. temperature hydrothermal (epithermal) origin in Antimony-bearing deposits are generally asso- shallow crustal environments associated with ciated with calc-alkaline to peralkaline, porphy- magmatic fluids; (ii) metamorphogenic hydro- ritic felsic to intermediate volcanic and intrusive thermal origin in consolidated crustal environ- wall rocks, often in a volcanic cauldron setting. ments derived from crustal fluids, triggered by, The original host rocks are variably affected by and with contributions from, magmatic heat and hydrothermal alteration processes such as silici- expelled fluids; and (iii) reduced intrusion-related fication, carbonatisation, sericitisation, chloriti- gold systems. Antimony deposits may also be sation and greisenisation. Volcanic host rocks distinguished by their metal and mineralogical are typically submarine, subaerial, and pyro- composition: (i) simple stibnite (plus gold) clastic deposits of rhyodacite, dacite and andesite. Antimony 73

(a) (b)

cn

stb

0.1 mm

(c) (d)

0.1 mm

(e)

1cm

Figure 4.1 (a) Stibnite ore in association with carbonate-quartz alteration (white), Beta mine, Antimony line, South Africa; (b) Photomicrograph of stibnite (stb)- cinnabar (cn) intergrowth in quartz-carbonate matrix, Monarch mine, South Africa; (c) Same as B, crossed nicols; (d) Needles of stibnite, Antimony line, South Africa; (e) Sample of stibnite-gold ore, associated with coarse-grained quartz-carbonate alteration (white). Native gold is associated with the contact between stibnite ore and quartz-carbonate. 74 ulrich schwarz-schampera

Table 4.3 Size and grade of antimony deposits (grades and tonnages are very variable between deposits and figures given are indicative only). (Source: BGR database.)

Deposit size Typical grade Estimated antimony metal Per cent

Deposit type range (tonnes) (Sb2S3%) content of known deposits (tonnes) of total

Gold-antimony 104–10 6 0.1–3.5 580,000 20 (epithermal) vein-type Carbonate 106–10 8 1.5–25 2,500,000 60 replacement Reduced-magmatic 106–10 8 0.1–1.5 320,000 10 Polymetallic base 104–10 6 0.1–0.5 175,000 8 metal vein Hot springs 104–10 6 0.1–0.2 2500 2 3,577,500 100

Subvolcanic to hypabyssal intrusive host rocks metamorphic rocks. The deposits may be found comprise granite–granodiorite–quartz–monzo- in association with hot springs and frequently nite differentiates. Intrusion-related antimony occur at the centres of young volcanism. deposits (skarns, replacement ores, vein-type) Three main sub-types of epithermal deposits that form within the region of hydrothermal are distinguished on the basis of alteration and influence surrounding the magmatic source are ore mineral assemblages: the quartz-(kaolinite)- related to plutons of diverse types with charac- alunite or high-sulfidation sub-type; the teristics of I-, S-, and A-type granitoids (Hart, intermediate-sulfidation sub-type; and the adu- 2007). laria-sericite or low-sulfidation sub-type The key characteristics and examples of the (Hedenquist et al., 2000). Each sub-type is char- major types of antimony deposits are given in acterised by hydrothermal fluid of a particular Table 4.4. oxidation state and acidity. The close association of antimony deposits with carbonate sequences and intense quartz-carbonate alteration indicate Gold–antimony (epithermal) deposits moderate acidity for the mineralising fluids. Antimony belongs to a group of metals and met- Temperatures of formation of epithermal alloids, usually referred to as the ‘epithermal’ deposits range from about 100 °C for hot spring suite of elements, which includes gold, silver, or steam-heated deposits to about 350–400 °C tellurium, selenium, mercury, arsenic, antimony for deeper vein and replacement deposits. Ore and thallium (Taylor, 2007). The geochemical deposition is related to pronounced changes in association of these elements is a common and the physical, thermal and chemical properties generic characteristic of the epithermal deposit of the hydrothermal solutions which may occur group. over short distances. The principal commod- Epithermal deposits typically form within ities in epithermal deposits are precious metals 1.5 km of the Earth`s surface. They originate at (gold, silver), although co-enrichment in ele- active subduction zones and are closely associ- ments of the epithermal suite, as well as locally ated with centres of magmatism, typically linked copper, lead, bismuth and zinc, gives rise to to the shallow crustal emplacement of magmatic polymetallic deposits and provides potential for porphyry copper systems. They comprise veins by-products, especially in deposits with high and disseminations in volcanic, sedimentary and silver grades. Figure 4.2 The global distribution of antimony mines, deposits and major occurrences. (Polymetallic base-metal vein deposits are not shown separately.) Some of the symbols on the map represent a single important deposit or resource, while others represent a cluster of deposits in one area or region. Antimony is also known to occur in Algeria, Bosnia and Herzegovina, Brazil, Burma (Myanmar), Ecuador, Greece, Honduras, Japan, Kazakhstan, New Zealand and Pakistan. 76 ulrich schwarz-schampera

Table 4.4 Key characteristics and examples of antimony deposit types (grades and tonnages are very variable between deposits and figures given are indicative only). (Source: BGR database.)

Deposit type Brief description Features Examples

Gold-antimony Medium to large, low-grade Deposits lack significant copper, Hemlo (Canada), Yanacocha (epithermal) vein stockwork-type quartz-stibnite ± lead, zinc and nickel sulfide and (Peru), El Indio (Chile), tetrahedrite veinlets and sulfosalt minerals; Sb-As-Hg-Au- Goldfield (USA). disseminations in shale, Ag-Te assemblage of the limestone, quartzite, volcanic low-sulfidation subtype in rocks, granite in greenstone epithermal environments; veins of belts of a potential quartz cored by massive stibnite; subduction zone and island transitions to mesozonal arc setting. (orogenic) deposits. Greenstone-hosted Numerous small to large size, Deposits lack any significant base or Xikuangshan (China), quartz-carbonate vein high-grade vein-stockwork precious metal assemblages; Kadamdzhay and carbonate deposits of almost pure lenticular bodies of quartz and (Russia), Antimony line replacement stibnite in sedimentary, stibnite within limestone and (South Africa), Olympiada metasedimentary (limestone) intense quartz-carbonate (Russia). and highly altered volcanic alteration, at contacts with sequences. Syn- to post- overlying shale, near high-angle collisional tectonic settings. faults; silicification extends tens of metres into the host metasedimentary or volcanic units. Reduced-magmatic Regional arrays of sheeted Deposits occur in a mineral system Tintina gold province (USA, auriferous quartz-carbonate with an outward zonation of an Canada), Timbarra, Kidston, veins around source plutons; Au-Bi-Te-W-As-Sb-Ag-Pb-Zn (Australia), Niuxinshan small to intermediate size assemblage; skarn-like and (China). proximal As-Sb-Au veinlets. replacement bodies and veins; Weak post-collisional associated with volatile-rich quartz extension behind a thickened monzonite melts. continental margin. Polymetallic base metal Small to medium polymetallic Polymetallic base metal and Ag-rich Cobalt district (Canada), vein deposits; structurally ores; densely intergrown ore Bolivian antimony belt, controlled in post-collisional minerals (telescoping); Sb hosted Bau district (Malaysia). vein breccia zones of clastic by stibnite, tetrahedrite and a metasedimentary or variety of simple and complex magmatic-dominated sulfosalts; quartz-carbonate terranes; mineralisation by gangue. basinal brines. Hot spring exhalative Siliceous precipitates deposited Low temperature fluids forming Kudryavyi, Kuril islands by hydrothermal fluids, hot sulfide and sulfosalt segregations (Russia), springs, fumaroles; volcanic in altered intermediate to felsic Merapi (Indonesia), activity. wall rocks; Sb co-enrichment with Taupo volcanic zone As and Hg, and locally Te, Se, (New Zealand). Au, Ag.

Ag, silver; As, arsenic; Au, gold; Bi, bismuth; Hg, mercury; Pb, lead; Sb, antimony; Se, selenium Te, tellurium; W, tungsten; Zn, zinc . Antimony 77

Epithermal-type antimony deposits include Greenstone-hosted quartz-carbonate stratabound mantos (e.g. Wadley, Sierra de vein and carbonate replacement deposits Catorce antimony district, Mexico), vein deposits controlled by fault zones (e.g. Bolivian antimony Gold-quartz-carbonate vein deposits occur in belt), fault-controlled veins and limestone replace- deformed greenstone belts. Common lithological ments (e.g. Bau district, Sarawak, Malaysia) and associations include tholeiitic basalts and ultra- subordinate Carlin-type gold (arsenic, mercury, mafic komatiite flows intruded by intermediate antimony) deposits (e.g. Getchell, Turqoise Ridge, to felsic porphyry intrusions, accompanied by Twin Creeks, Nevada; Zarshuran, Iran; Alšar, swarms of albitites in some areas (Dubé and Macedonia). Hot spring deposits may have signi- Gosselin, 2007). They are distributed along major ficant antimony enrichments (e.g. New Zealand, compressional to transtensional crustal-scale California, Nevada). Some deposits may be tran- fault zones in deformed greenstone terranes, and sitional to those related to Cordilleran granitoids. mark the convergent margins between major lith- These include mesothermal lead–zinc–silver ological boundaries, such as volcano-plutonic (antimony) veins (e.g. Coeur d’Alene district, and sedimentary domains. Large greenstone- Idaho; Bawdwin, North Shan State, Myanmar), hosted quartz-carbonate vein deposits are com- hydrothermal antimony deposits in, and associ- monly spatially associated with fluvio-alluvial ated with, granitoids (e.g. Southern Bolivia, conglomerate distributed along major crustal Sierra de Catorce, Mexico), and deposits in fault fault zones. This association suggests an empirical zones with scheelite-gold in silicate rocks (e.g. time and space relationship between large-scale Yellow, Idaho). deposits and regional unconformities. Antimony deposits in Bolivia are typically Although these deposits occur in Proterozoic epigenetic vein-type deposits crosscutting early and Palaeozoic terranes they are more abundant in Paleozoic fine-grained clastic rocks. The formation Archaean terranes where their gold and antimony age is considered to be younger than early contents are greatest. They are epigenetic deposits Cretaceous (Dill et al., 1995). They are dominated characterised by networks of gold-bearing, lami- by stibnite, accompanied by pyrite, arsenopyrite nated quartz-carbonate veins (Dubé and Gosselin, and minor jamesonite, berthierite and base metal 2007). They are hosted by greenschist- to amphib- sulfides. Mexico hosts a number of epithermal- olite-facies metamorphic rocks of dominantly style mono- and polymetallic antimony deposits mafic composition and formed at intermediate (Simmons et al., 2005). The vast majority occurs as depth (5–10 km). The mineralisation is syn- to vein-type and replacements in bedded carbonate late-deformation and typically post- to syn-peak rocks. The regional association with sub-volcanic metamorphism. The deposits are typically associ- rocks and co-enrichment with other metals and ated with iron-magnesium-carbonate alteration. metalloids of the epithermal suite are indicative of The gold is largely confined to the quartz- a magmatic-hydrothermal origin. Stibnite pre- carbonate vein network but may also be present dominates in stratabound mantos replacing in significant amounts within iron-rich sulfidised limestone beds; gold may be a potential by-product. wall-rock selvages or within silicified and arseno- Hot-spring deposits enriched in the epither- pyrite- and stibnite-rich replacement zones. mal element suite comprising silica sinters and Examples of synorogenic hydrothermal siliceous muds are found at several locations, e.g. antimony–gold deposits occur in Archaean green- in the Coromandel and Taupo volcanic zones, stone terrains in Southern Africa (e.g. the Antimony New Zealand and at Monte Amiata, Italy. Hot- Line, Murchison, South Africa; and the Kadoma spring deposits also form gold-stibnite ores in and Kwekwe goldfields, Zimbabwe; Vearncombe veins and disseminations in sinters e.g. at Sulfur et al., 1992; Buchholz et al., 2007). In the Murchison Bank and McLaughlin, California and Round and the Barberton greenstone belts of South Africa Mountain, Nevada. the synorogenic mineralisation is sporadically 78 ulrich schwarz-schampera developed in deformed thick quartz-carbonate and have characteristics of I-, S-, and A-type gran- alteration zones within highly deformed and modi- itoids. The deposits typically form within the fied komatiitic basalts. The antimony mineralisa- region of hydrothermal activity around the causa- tion consists of stibnite and minor berthierite. Gold tive pluton. The systems include a wide variety of is an important by-product of antimony mining in intrusion-related mineral deposit styles including the so-called Antimony Line of the Murchison skarns, replacements, veins and disseminations greenstone belt. and they originate from a depth of five to seven km The largest antimony deposits in China are (Hart, 2007). The mineral systems show a pro- epigenetic carbonate-hosted stibnite deposits nounced zonation with proximal gold-tungsten- located in central Hunan, China (Wu, 1993). The arsenic, intermediate arsenic–antimony–gold replacement orebodies occur within Middle and and distal silver–lead–zinc metal associations. Upper Devonian limestones which were subject Antimony-rich deposits typically occur as fault- to hydrothermal alteration in late Jurassic to filled veins. Stibnite predominates, but more com- early Cretaceous (Penga et al., 2003). The deposits plex sulfosalts, in association with bismuth, lead are several metres thick, usually stratiform, but and tellurium, are usually present. locally irregular in shape. Stibnite, the only ore Type-examples of reduced intrusion-related gold mineral, is associated with intense silicification systems are known in the Fairbanks area of central and carbonatisation. Alaska (e.g. Fort Knox deposit) and central Yukon Synorogenic antimony deposits may grade into (e.g. Tombstone gold belt; Hart, 2007). The deposits more complex antimony–gold, antimony–arsenic– of the wider Tombstone gold belt across central gold and antimony–tungsten deposits and their Yukon and Alaska are hosted by, and formed from, combinations (e.g. Xikuangshan, Hunan, China; reduced mid-Cretaceous plutons. Other examples Wadley, Sierra de Catorce, Mexico; Kadamzhai, are known from southern New Brunswick, the Kyrgyzstan; Turhal, Turkey; Hillgrove, New South Bolivian polymetallic belt and the Yanshanian oro- Wales, Australia). Mesothermal gold, transitional gen of the North China craton (Thompson et al., to synorogenic gold deposits, have moderate signif- 1999). Archaean examples of similar type have been icance for the antimony production (e.g. Olimpiada, suggested to occur in the southern Superior Russia). Hydrothermal deposits in intracratonic Province of Ontario and Quebec (Robert, 2001). orogens may also be associated with granites and Reduced intrusion-related gold–antimony systems include multi-metal zoned tin, copper, lead, zinc, share a number of characteristics with orogenic silver (antimony) skarn-replacement-vein systems gold deposits as well as with epithermal-porphyry (e.g. Dachang ore field, Guangxi, Guangdong). ore-forming systems. Critical features include the Synorogenic mercury–antimony deposits are association with reduced intrusions, the lack of known from the Khaidarkan, South Ferghana copper and the presence of tungsten. mercury–antimony belt in Kyrgyzstan and share characteristics with continental epithermal-type deposits (Obolensky et al., 2007). Extraction methods and processing

Reduced magmatic gold systems Mining Reduced magmatic or intrusion-related gold– Of the 18 countries that have produced primary bismuth–tellurium–tungsten–antimony deposits antimony in the past decade, the most impor- are related to the brittle carapace at the top of tant have been, in order of decreasing produc- small metaluminous, moderately reduced and tion, China, Bolivia, South Africa, Russia, fractionated post-collisional quartz monzonite Tajikistan, Turkey and Australia. Antimony plutons. The magmas have a reduced primary mining in those countries is dominated by one oxidation state that form ilmenite-series plutons or a few stibnite-rich deposits, with possible Antimony 79 by-products including gold, silver, tungsten and from the fine fraction using dense media and the mercury. Underground extraction is used in one heavy product is further prepared for flotation. of the world’s major antimony-producing mines, A pure stibnite concentrate is produced following the Consolidated Murchison Mine in South cleaning of the concentrate produced by flota- Africa, which uses variations of shrinkage stop- tion. Overall, 33 per cent of the ore is treated by ing. On the surface, the waste is removed from hand sorting, seven per cent by heavy media sep- the high-grade ore, and the latter is trucked to a aration and 60 per cent by flotation. Typical ben- mill for the separation and concentration of gold eficiation routes, based upon the Xikuangshan and antimony. Where antimony is a minor South ore dressing plant and the Sunshine Mining constituent of a metalliferous ore the mining and Refining Co. (closed in 2001) antimony pro- method, either open pit or underground, is cess are shown in Figure 4.3. designed to optimise the recovery of the principal At least six principal methods have been used metals, such as gold, lead or silver. In some to extract antimony from its ores. The chosen countries (e.g. Bolivia, Peru, Zimbabwe) small technique depends on the oxidation state (anti- mines work irregular and scattered antimony- mony sulfide, oxide or complex, mixed oxide– bearing orebodies that cannot be readily sulfide ore) and the ore grade. Stibnite ores can exploited by large-scale mining methods. These generally be processed more efficiently and at mines are entered by a shallow shaft or short lower cost than oxide ores. adit, developed by drifting in the vein, and Roasting the ore to yield a volatile trioxide or stoped by simple overhand methods between the stable non-volatile trioxide is the only pyro- raises driven on the footwall of the ore. metallurgical procedure suitable for low-grade ores (5 to 25 per cent antimony content). The control of volatilisation conditions produces a high-grade oxide that can be sold directly to con- Ore processing, beneficiation and sumers. Intermediate grade ores, containing 25 conversion to metal to 45 per cent antimony, are smelted at high Techniques for the processing of antimony- temperatures in blast furnaces to produce crude bearing ores encompass a wide range of tech- antimony metal. Oxides, sulfides or mixed ores, niques from traditional hand sorting, which is residues, mattes, slags and briquetted fines or dependent on plentiful and cheap labour, to tech- flue dusts can be used as blast-furnace charges. nologically advanced, capital-intensive mineral The method employs a high smelting column processing. The mineral processing stages gener- and comparatively low air pressure, and slag and ally include conventional crushing and grinding, metal are separated in a forehearth. Considerable followed by combined gravity concentration and quantities of slag are formed, which is desired flotation. because it tends to reduce volatilisation losses. A typical stibnite deposit in China, such as High-grade ores that contain 45 to 60 per cent Xikuangshan, contains stibnite, together with antimony are melted in a crucible or reverberatory pyrite and associated gangue minerals such as furnace with a reducing atmosphere to prevent quartz, calcite, barite, kaolinite and gypsum (Wu, oxidation and loss by volatilisation. On cooling, 1993; Yang et al., 2006). The ore, which grades the solidified product is referred to as ‘crude anti- about 2.7% Sb, is treated by hand sorting, crush- mony’, ‘crudum’, or ‘needle antimony’. Alterna- ing, heavy medium separation and flotation. tively, high-grade sulfide ores can be treated by a Hand sorting includes the recovery of antimony precipitation process in which scrap iron is used lump concentrates before and between the crush- to reduce sulfur and yield impure antimony ing stages. Crushing takes place in a two-stage metal. closed circuit consisting of a jaw and cone crusher Oxide ores that contain about 30 per cent anti- (Anderson, 2000). Gangue material is separated mony are reduced in blast furnaces to crude 80 ulrich schwarz-schampera

Ore

Primary crushing

Screening Heavy media Tailings Hand sorting Tailings separation

Crushing and grinding

Flotation Tailings

Cleaning Tailings

Antimony concentrate

Depending on Pyrometallurgical ore type and grade Hydrometallurgical processing processing

Volatilisation Smelting Heating and Leaching Leaching in a iron with with Tailings blast precipitation alkaline acid furnace under sulfide chloride reducing Slag conditions

Precipitation Iron Electrowinning Antimony sulfide chloride

Antimony Antimony Antimony Electrowinning Hydrolysis trioxide metal metal and precipitation

Antimony Antimony Ammonia metal metal treatment

Antimony oxide

Figure 4.3 Generalised beneficiation flow diagrams, based on various ore dressing and beneficiation processes including the Xikuangshan South ore dressing plant, the Sunshine antimony process and the U.S. Antimony Corp. process. (Modified from Anderson (2000) and U.S. Antimony Corporation (2011).) metal, while richer oxide ores, with about 50 per Mixed sulfide and oxide ores are usually cent antimony, are reduced and refined to com- smelted in blast furnaces, although they are mercial-grade metal in reverberatory furnaces sometimes processed by selective leaching fol- using coke and a suitable flux, such as soda ash. lowed by electrolysis of the leachate to recover Antimony 81 the metal (Butterman and Carlin, 2004). A typical house, a Cottrell precipitator or any combination process uses an alkali hydroxide or sulfide as the of the above. Antimony sulfide may be ineffi- solvent. The filtered leach solution, which con- ciently separated from the gangue of sulfide ore. tains sodium thioantimonate, is electrolysed in a An efficient method uses a reverberatory furnace diaphragm cell using an iron or lead anode and an and continuous liquation under reducing condi- iron or mild-steel cathode. The cathode metal tions. The oxide volatilisation process to recover obtained ranges from 93 to 99 per cent antimony additional antimony is used to treat residues con- metal. taining 12–30 per cent antimony. The liquated The metal produced by pyrometallurgical product is sold for applications requiring anti- processes does not meet the quality requirement mony sulfide or is converted into metallic anti- for commercial products and must be further mony. Antimony oxides are reduced to metal refined. The major impurities are usually arsenic, with charcoal in reverberatory furnaces at about copper, iron, lead and sulfur. The iron and copper 1200 °C. An alkaline flux consisting of soda, concentrations may be lowered by treating the potash and sodium sulfate is commonly used to metal with stibnite or a mixture of sodium sul- minimise volatilisation and dissolve residual sul- fate and charcoal to form an iron-bearing matte, fides and gangue. The loss of antimony from the which is skimmed from the surface of the molten charge by volatilisation is usually high (12–20 per metal. The metal is then treated with an oxidising cent) and lowered by the use of Cottrell precipita- flux that consists of caustic soda or sodium car- tors or baghouses. Rich sulfide ore or liquated bonate and sodium nitrate to remove the arsenic antimony sulfide (crude antimony) is reduced to and sulfur. Lead cannot be readily removed from metal by iron precipitation. This process includes antimony, but material high in lead may be used the heating of molten antimony sulfide in cruci- in the production of antimony-bearing lead-base bles with fine iron scrap. A light fusible matte alloys. with iron sulfide is formed to facilitate separation Antimony metal is also recovered from slags of the metal. A second fusion with liquated and residues produced during the processing of antimony sulfide follows for purification. other metals, such as lead and gold. In the cyani- Intermediate grades of antimony ores, liquation dation of gold ores stibnite is oxidised to various residues, mattes, slags and flue dusts are pro- species which inhibit gold dissolution by con- cessed in blast furnaces at 1300 to 1400 °C in a suming oxygen and cyanide. process similar to that used for the recovery of For primary production, the antimony content lead. It is the favoured method of smelting of of the ore has traditionally determined the pyro- these ores to minimise volatilisation losses. metallurgical recovery. The lowest grades of anti- Hydrometallurgical methods are employed for mony sulfide ores containing 5–25 per cent simple as well as complex antimony ores. A antimony are volatilised to antimony trioxide, two-stage process of leaching and subsequent 25–40 per cent antimony ores are smelted in a electrodeposition is generally involved. Alkaline blast furnace, and 45–60 per cent antimony ores sulfide and the acidic chloride are used as solvent are treated by liquation or iron precipitation systems. The former predominates and is (Anderson, 2000). The removal of antimony as employed in the CIS, China and the USA the volatilized trioxide is the only pyrometallur- (Anderson, 2000). The lixiviant is a mixture of gical method suitable for low grade ores. The sodium sulfide and sodium hydroxide which form combustion of the sulfide components of the ore a sodium thioantimonite (Na3SbS3) when applied supplies some of the energy and fuel require- to stibnite. The dissolution of elemental sulfur in ments are minor. The volatilisation process sodium hydroxide is also used as a lixiviant for includes sulfur combustion at about 1000 °C with alkaline sulfide leaching. The electrodeposition of coke or charcoal and the recovery of volatile anti- the antimony from the alkaline sulfide solution to mony trioxide in flues, condensing pipes, a bag- cathode metal is carried out via electrowinning in 82 ulrich schwarz-schampera diaphragm or non-diaphragm cells. The antimony purity antimony metal, used in thermoelectric metal product may attain a grade of 99.5%. The devices and semiconductors, is produced in the acidic chloride hydrometallurgy uses hydro- form of ingots weighing 0.5 to 3.0 kg and grading chloric acid in conjunction with iron chloride 99.99% Sb (referred to as 4 N) to 99.99999% (7 N).

(FeCl3) to produce antimony chloride (SbCl3). The A number of antimony compounds are also iron chloride acts both as an oxidiser and as a traded on the international markets. These chloridising agent. The dissolved antimony chlo- include antimony trioxide (Sb2O3; 83.5% Sb), ride can be electrowon from solution in diaphragm antimony oxychloride (SbOCl; 70.3% Sb), anti- cells to produce cathode antimony metal. mony pentoxide (Sb2O5, 75.3% Sb), antimony

Alternatively, the antimony chloride solution can trichloride (SbCl3; 53.4% Sb), antimony trisul- be treated by hydrolysis precipitation of the anti- fide (stibnite Sb2S3; 71.1% Sb), sodium antimo- mony from solution as a solid oxychloride (SbOCl, nite (NaSbO3, 63.2% Sb) and stibine (SbH3;

Sb4O5Cl2). The precipitated solid is treated with 97.6% Sb). ammonia to produce antimony oxide. Antimony trioxide, commonly referred to as ATO, is the most widely used antimony compound.

Crude antimony trioxide grades below 98% Sb2O3, while commercial grades contain 99.2 to 99.9%

Specifications Sb2O3. Several commercial specifications are avail- able, each characterised by specific tinting strengths The main antimony products in international and/or the content of particular impurities such as trade are stibnite and subordinate stibnite- arsenic, iron and lead (Amspec, 2011). berthierite and tetrahedrite ores and concentrates, antimony metal, antimony trioxide and antimo- nial lead. Antimony ore concentrates contain 5 to Uses 60% Sb. Chemical grade ores are sufficiently pure to be used directly in the production of antimony Various unique properties of antimony determine trioxide, antimony chloride or other compounds. its use in a diverse range of products and applica- The total impurity level must not exceed 0.25%, tions. These properties include: with arsenic and lead concentrations less than ● low melting point enhancing workability at 0.1% (Anderson, 2000; Roskill, 2007). In the USA low temperatures; two grades of antimony metal are specified ● stability in air at room temperature and in (Butterman and Carlin, 2004): grade A has a water up to 250 °C; minimum content of 99.80% Sb, with maximum ● resistance to most cold acids; values of arsenic (0.05%), sulfur (0.10%) lead ● dissolution in some hot acids and in aqua regia; (0.15%) and other elements (0.05% each); and ● incompatibility with strong oxidising agents, grade B metal must have a minimum content of chlorine, fluorine; 99.5% Sb, with maximum values of 0.1% arsenic, ● reaction with materials that do not react with sulfur and other elements, and 0.2% lead. hydrochloric and nitric acids separately; Antimony metal is commonly traded in ingots ● metastability at rapid cooling providing exo- and slabs weighing between 20 and 50 pounds thermic reaction; and also in the form of granules, cast cake, ● high density (6692 kg/m3), but weak bonding powder, shot and single crystals. The Minor leading to low hardness and brittleness; Metals Trade Association (MMTA) grade I specifi- ● two allotropic forms of antimony: stable cation for antimony metal used for the produc- metallic and amorphous grey; tion of antimony trioxide is minimum 99.65% ● low electrical and thermal conductivity; Sb, and maximum 0.15% As, 0.005% Se, 0.01% ● expansion on freezing, like silicon, bismuth, Bi, 0.02% Pb and 0.02% Fe (MMTA, 2012). High- gallium and germanium. Antimony 83

Table 4.5 Estimated global consumption of antimony by end-use in 2000 and 2011. (Data from Roskill, 2012.)

Consolidated 2000 2011 annual growth rate

Non-metallurgical Tonnes Tonnes % of non- uses antimony % of total antimony % of total metallurgical use % Flame retardants 70,000 47.4 108,250 52.4 83.8 4.0 Plastic catalysts 6000 4.1 12,100 5.9 9.4 6.6 Heat stabiliser 1400 0.9 2700 1.3 2.1 6.1 Glass 16,000 10.8 1650 0.8 1.3 −18.6 Ceramics 1700 1.2 2550 1.2 2.0 3.8 Other 1500 1.0 1900 0.9 1.5 2.2 Sub-total 96,600 65.4 129,150 62.5 100 2.7 Tonnes Tonnes Metallurgical uses antimony % of total antimony % of total % of metallurgical use Lead-acid batteries 40,000 27.1 53,600 25.9 69.2 2.7 Lead alloys 11,000 7.5 23,850 11.5 30.8 2.7 Sub-total 51,000 34.6 77,450 37.5 100 3.9 Total 147,600 100 206,600 100 2.7

Antimony is consumed in the following forms: Antimony trisulfide was used in artillery projec- ● antimony trioxide, used mainly in flame tiles, in bomb fuses and for the generation of retardants and in PET (polyethylene tere- white clouds on detonation. Antimony alloyed phthalate); with lead was used as long as the nineteenth ● sodium antimonite, used mainly in cathode ray century as bearing metal (Babbitt metal) and, tube glass; alloyed with tin, to produce Britannia metal used ● primary metal, used mainly in lead-acid batteries; in items such as eating utensils, teapots and can- ● antimonial lead, mainly recycled from and dlesticks. Various antimony salts have long been re-used in lead-acid batteries. used for medical and veterinary purposes. Today, The early technical use for antimony was antimony compounds are still used for the related to the development of cast metal printing treatment of two parasitic diseases, schistosomi- types, mirrors, bell metal and pigments. The asis and leishmaniasis. The rapid growth in the main antimony-producing countries in the 18th use of plastics after 1950 led to today’s dominant century were France, Germany and Italy. By the market for antimony-based flame retardants in a early 19th century the major uses of antimony wide range of products. Pigments (glass and were in pharmacology, agriculture, artillery, dye- ceramic industries), lead-acid batteries and metal ing, pigments and paints for colouring glass, alloys together account for about 30 per cent of ceramics, cloth and paper, printing type, bearing current antimony use. and anti-friction alloy metal, vulcanising rubber, The most important end uses of antimony in the manufacture of safety matches and in thermo- 2000 and 2011 are summarised in Table 4.5. In electric couples. Mine production increased that period total antimony consumption grew by sharply during World War I as shrapnel was about 40 per cent from 147,600 tonnes to 206,600 hardened with 10 to 13 weight per cent antimony. tonnes. In fire retardants, the largest application, In World War II antimony was used in the lead- antimony use grew by 54 per cent in the same acid batteries of military vehicles and in period, while use in catalysts for PET production flame-retardant compounds for heavy textiles. more than doubled. 84 ulrich schwarz-schampera

Table 4.6 Estimated world consumption of antimony by main product in 2001, 2006 and 2011. (Data from Roskill, 2012.)

2001 2006 2011 2011

Tonnes Tonnes Tonnes antimony % antimony % antimony % % of primary

Primary Product Antimony trioxide1 72,600 53.3 103,950 57.9 124,950 60.5 74.1 Sodium antimonate 15,600 11.5 12,150 6.8 4250 2.1 2.5 Antimony metal 16,000 11.7 23,000 12.8 39,450 19.1 23.4 Sub-total 104,200 76.5 139,100 77.5 168,600 81.7 100.0 Secondary Product Antimonial lead 2 32,000 23.5 40,500 22.5 38,000 18.3 Total 136,200 100 179,600 100 206,600 100

1includes other antimony compounds such as tetroxide, pentoxide, oxychloride 2antimony content

Antimony trioxide estimated to have increased by a compound annual growth rate of four per cent, or 38,250 tonnes, Antimony trioxide, ATO, is the most important between 2000 and 2011. antimony compound produced. Global consump- The other major markets for antimony in non- tion of antimony in ATO is estimated at almost metallic uses are ATO as a heat stabiliser in PVC 125,000 tonnes (equivalent to 150,000 tonnes ATO), and as a catalyst in the production of polyethylene which accounted for 74 per cent of 2011 primary terephthalate (PET), which is used in the manu- antimony consumption (Table 4.6; Roskill, 2012). It facture of synthetic textiles (polyester) and plastic is prepared by the volatilisation of antimony metal containers, such as drinking bottles, and polyester in an oxidising furnace. Of itself ATO has no film mostly used for packaging. ATO is also used flame-retardant property, but when it is combined in the glass and ceramics sector as a degassing with halogenated (brominated, chlorinated) flame- agent and as an opacifier in porcelain enamels and retardant compounds it provides the most effective pottery glazes. and widely used flame-retardant system for plas- tics. The halogenated antimony compounds act as Sodium antimonate dehydrating agents and inhibit ignition and pyrol- ysis in the solid, liquid and gas phases. They also Sodium antimonate (NaSb(OH)6) is mainly used as promote the formation of char on the substrate, a fining and degassing agent in the production of which acts as a barrier and reduces oxygen avail- high-quality clear glass. The antimonate decom- ability and volatile-gas formation. poses in the molten glass generating large bubbles Flame retardants are by far the largest market which rise to the surface scavenging much slower for antimony today. The majority of flame retar- moving fine bubbles leading to the purification and dants are used in plastics, with smaller amounts in homogenisation of the glass batch. Sodium anti- rubber, textiles, paints, sealants and adhesives. The monate is also a decolourant for glass as it removes compounds account for 52 per cent of total anti- traces of iron which can give rise to a greenish tint. mony consumption and over 84 per cent of non- It also has antisolarant properties which protect metallurgical antimony consumption (Table 4.6). against colouring caused by sunlight or fluorescent The world demand for ATO in this market is lights during the lifetime of the glass. Antimony 85

Some sodium antimonate is also used in fire- Recycling retardant and smoke-suppressant compounds and formulations and as feed for metal and trioxide From most products, such as fire retardants, anti- production. World consumption of sodium anti- mony compounds cannot be recycled because monate has steadily fallen from 15,600 tonnes in they are dissipated in use although recycling of 2001, when it accounted for 11.5 per cent of total PET containers makes an indirect contribution. consumption, to 4250 tonnes in 2011, 2.1 per cent However, antimony can be recovered from most of total consumption (Table 4.6; Roskill, 2012). applications where it is used as an additive in lead alloys. Secondary antimony is a significant source Other non-metallurgical uses of supply in many countries and accounts for about 20 per cent of total antimony used and approxi- Minor non-metallurgical markets include lubri- mately 40,000 tonnes per year (Roskill, 2007). cants, ammunition primers, textiles, pharmaceu- Most secondary antimony metal is obtained ticals, pesticides, fluorescent lamps, fireworks from recycled lead-acid batteries which contain and matches, zinc electrowinning and the refining between 0.6 and 1.5% Sb. The recovery of anti- of sour crude oil. Very high purity antimony (5 N mony used in alloys for small-arms ammunition, to 7 N) is increasingly being used in the semicon- semiconductors, bearings and solders cannot be ductor industry as a dopant for ultra-high conduc- recycled. The availability of secondary antimony tivity n-type silicon wafers which are utilised in depends almost entirely on the extent of secondary diodes, infrared detectors and Hall-effect devices. lead recovery and therefore on the market condi- tions for lead and lead battery scrap. It accounts Antimony metal for the entire US-sourced production and makes The largest market for metallic antimony is in an important contribution in the EU, Canada, lead alloys adding hardness and smoothness of Japan, South Korea and Taiwan (Roskill, 2007). finish. The higher the proportion of antimony in Despite stable lead recycling rates, the amount of the alloy, the harder and more brittle it will be. In secondary antimony recovery continuously lead-acid batteries, the addition of antimony decreased in the USA from 10,500 tonnes in 1995 improves the charging characteristics and reduces to 3450 tonnes in 2011. The decline mainly generation of unwanted hydrogen during charging. reflects reductions in battery weight, lower anti- The property of some antimony alloys to expand mony content and growth in antimony-free bat- on cooling is the basis of their use in some applica- teries. In industrialising countries, the recovery of tions such as making typefaces for printing. secondary antimony has strongly increased due to Other metallurgical applications of antimony the higher demand for lead-acid batteries and metal are in solders, fusible alloys, type metals, rising recycling rates. Secondary antimony lead weights, rolled and extruded antimony alloys, recovery in highly industrialised countries, like pewter, Britannia metal, shot and ammunition. It is the USA, Canada and the original EU-15, amounts also used to harden low-tin alloys, to reduce friction to 5.6 kg to 8.7 kg per tonne of secondary refined and wear in bearing alloys in machinery and to lead production (Roskill, 2007). Southern improve the durability of lead sheathing for cables. European older generation battery recycling in Antimony metal is estimated to account for Spain and Italy amounts to 15 kg per tonne, and about 23 per cent of primary antimony consump- Japan, South Korea and Taiwan to 10 kg per tonne. tion (Table 4.6). Primary consumption of antimony The use of antimony in lead-acid batteries may be metal increased by 72 per cent between 2006 and much reduced in the future as alternative vehicle 2011, from 23,000 tonnes to almost 40,000 tonnes technologies are increasingly adopted worldwide. (Roskill, 2012). Excluding its use in lead-acid bat- Blast furnaces, reverberatory furnaces and teries, alloys account for about 30 per cent of metal- rotary furnaces may be used for the recovery of lurgical consumption of antimony (Roskill, 2012). antimony from secondary materials. Blast 86 ulrich schwarz-schampera

furnaces are used for a continuous and regular 4% 2% supply of coarse charge material (i.e. battery 2% scrap) and produce a metal output of uniform 4% composition. Reverberatory furnaces are suitable for finely divided feed material and produce an 36% antimony-rich slag grading 5 to 9% Sb. Antimony 11% oxide in the slag is subsequently reduced in a blast furnace to produce antimony metal. Rotary furnaces produce either a single grade of alloy or both, a high-antimony and a low-antimony alloy.

12% Substitution

There is a range of fire-retardant materials on the market. They include alumina trihydrate, magnesium hydroxide, calcium carbonate, zinc 13% borate, zinc stannate, zinc hydroxyl stannate, 15% melamine, and phosphorus-based, nitrogen- based, and phosphorus–nitrogen-based combina- China Thailand Russia tions. However, their performance is generally Bolivia Kyrgyzstan Turkey inferior to ATO-based fire retardants and accord- South Africa Tajikistan Other countries ingly the scope for substitution where high performance is required is restricted. Several types of plastic catalysts and plastic Figure 4.4 The distribution of world antimony reserves stabilisers, which contain barium, cadmium, in 2011. (Data from: USGS, 2012; Roskill, 2011; Village calcium, germanium, lead, tin, titanium, and zinc Main Reef Ltd, 2012.) in various combinations, may compete with anti- mony but their use generally leads to increased producer, has the largest reserves (1,225,000 production costs. Compounds of cadmium, chro- tonnes, 36 per cent of total). Other producing mium, tin, titanium, zinc and zirconium can sub- countries with major antimony reserves include stitute for antimony chemicals in paint, pigments Thailand (525,000 tonnes), Russia (455,000 and enamels. tonnes), Bolivia (420,000 tonnes) and Kyrgystan A lead–calcium–tin alloy could be a possible (385,000 tonnes). These five countries together alternative to lead-acid batteries, while lead-free account for approximately 87 per cent of world solders based on tin–silver–copper alloys can antimony reserves. Significant reserves also exist substitute for antimony in solder. A wide range in Turkey, South Africa and Tajikistan. In the of materials, mostly metals, may substitute for 1990s reserves were also reported in Guatemala, antimony alloys in friction bearings, ammunition Canada, Peru, Vietnam, Laos and Mexico. A and cable sheathing (Butterman and Carlin, 2004). minor contribution to antimony supply origi- nates as a by-product from gold and base metal refining, for example, from the Mississippi Resources and reserves Valley-type lead deposits in the eastern United States, from base-metal–silver operations in New Global antimony reserves compiled for this book Brunswick, Canada, in La Oroya, Peru and in have been estimated at about 3.4 million tonnes Coahuila and Nuevo Leon, Mexico and from gold in 2011 (Figure 4.4). China, the largest antimony deposits in Australia. Antimony 87

200 180 160 140 120 100 80 60 40 20

Figure 4.5 World antimony mine Thousand tonnes (metal content) production between 1992 and 2010. (Data 0 from British Geological Survey World

Mineral Statistics database.) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

It is significant to note that the USGS estimate Since the nineteenth century antimony has of global antimony reserves in 2011 was 1.8 mil- been produced in more than 40 countries. In 1900 lion tonnes, considerably less than the total pre- global antimony production was only about 7710 sented here (USGS, 2012). This discrepancy tonnes. However, in response to greatly increased highlights the difficulty of obtaining reliable and demand for ammunition in the First World War, consistent global reserve estimates for any min- production rose sharply to 81,600 tonnes in 1916. eral commodity, as discussed in Chapter 1 of This level was only reached again in 1994 when this book. Another factor which contributes the annual production exceeded 100,000 tonnes to this uncertainty is that estimates of antimony for the first time. Since then, the worldwide resources are generally based on pure stibnite production levels have been strongly affected by deposits and do not take account of antimony as a the global markets and have fluctuated bet- by-product of gold mining. Greenstone-hosted ween 103,000 tonnes in 1995 to 181,000 tonnes, quartz-carbonate vein deposits are major sources an all-time high, in 2007 (BGS, 2012). The overall of gold but often include significant quantities doubling of global production since the early of stibnite, berthierite and tetrahedrite, which are 1980s is a consequence of growing demand for not identified as potential antimony sources. The antimony in flame retardants and associated same is true for small-scale gold deposits widely increases in production capacity, especially in present in Europe, Africa, Asia and the USA. China. Production levels fell back sharply in 2008 Resource and reserve assessments for antimony and 2009 in response to the global economic should therefore be treated with caution as they recession but have begun to recover since then likely underestimate the geological availability. (Figure 4.5). China is by far the largest producer of primary antimony with production in 2010 of about Production 130,000 tonnes, equivalent to about 88 per cent of the world total (Figure 4.6). Bolivia is the next Large mines with antimony-bearing ore are in largest producer at about 5000 tonnes or three per production on all continents. Currently, the most cent of the total, followed by Tajikistan, Russia important deposits are located in China, where and South Africa, each contributing a further the majority is produced at Xikuangshan in three per cent. Hunan Province, and in Bolivia, Russia, South The refinery capacities for antimony are esti- Africa, Tajikistan, Canada and Australia. mated to be in excess of 163,360 tonnes per annum 88 ulrich schwarz-schampera

3% (Table 4.7; Roskill, 2011). China has the largest 2% 2% 2% share with 120,000 tonnes (nearly 74 per cent of 3% total), followed by Russia, South Africa, Myanmar, Canada, Tajikistan and Bolivia, each contributing approximately four per cent of the total. Global antimony production exceeds the refinery capacity by about nine per cent, which is likely due to a contribution to supply from stock inventories. Antimony derived from recycling is also significant, while ‘unofficial’ production is thought to have accounted for about 16 per cent of total antimony supply in 2010 (Roskill, 2011). Total antimony supply in 2010 was 196,484 tonnes, a 10 per cent increase on 2009. The total global capacity for antimony production is esti- mated to be about 30 per cent higher than the current production level. The Chinese government is pursuing a programme of closing antimony mines and smelters 88% in order to improve the control of environmental and safety issues. In 2010 they also imposed an China Bolivia Tajikistan antimony production quota of 100,000 tonnes, Russia South Africa Other countries comprising 62,520 tonnes for primary production and 30,480 tonnes from recycling. This compares with a quota of 90,180 tonnes in 2009. China used Figure 4.6 World antimony mine production in 2010, by country. (Data from British Geological Survey, 2012.) to operate about 50 industrial antimony mines and

Table 4.7 Global production capacity of refined antimony in 2010. (Source: Roskill, 2011.)

Company Country Capacity (tonnes/year antimony)

Hsikwangshan Twinkling Star China 55,000 Liu Zhou China Tin Group Co Ltd China 30,000 Hunan Chenzhou Mining Co. Ltd China 20,000 Shenyang Huacheng Antimony China 15,000 GeoProMining Russia 6500 Village Main Reef (former Metorex) South Africa 6000 Various (Shwe Zin Htut, Thu Ya Kan Chun) Burma (Myanmar) 6000 Beaver Brook Canada 6000 Comsup (former Anzob GOK) Tajikistan 5500 Various (Emusa, COMISAL, Bernal Hermanos, CBPA) Bolivia 5460 Mandalay Resources Australia 2750 Cengiz & Özdemir Antimuan Madenleri Turkey 2400 Kazzinc Kazakhstan 1000 Various (Hong Xin, Siam) Thailand 600 Kadamdzhai Kyrgyzstan 500 SRS Laos 500 US Antimony Mexico 150 Total 163,360 Antimony 89 many small production sites, with more than 300 South Africa is among the largest antimony pro- antimony smelters of varying sizes and a large ducers in the world. Since 1932, the Consolidated number of antimony mining and manufacturing Murchison antimony deposits (Alpha/Gravelotte, companies. Now only six companies account Beta, Athens, and Monarch) have produced more for more than 90 per cent of China’s reported than 550,000 tonnes of antimony in concentrate production of antimony. Reported production of and more than 25 tonnes of gold. In March 2011, antimony in China fell in 2010 and is unlikely to Consolidated Murchison was acquired by Village increase in the future, despite the fact that the Main Reef Limited. At present, activities are country is facing a serious shortage of antimony. focusing on ramping up production as more mining The local government in Lengshuijiang, Hunan areas become available and existing shafts are Province, which accounts for about 60 per cent of deepened. Consolidated Murchison has a total world antimony supply, closed almost all its antimony resource base of 200,000 tonnes of which mines and smelters in 2011 (USGS, 2012). No more than 10 per cent is in the reserve category significant antimony deposits have been developed (Village Main Reef, 2012). The 2011 mineral for about ten years and the remaining economic resource and reserves for the underground opera- reserves are being rapidly depleted, with only five tions are 3480 million tonnes of antimony ore at years of mining life remaining in Lengshuijiang 1.87% Sb, which is equivalent to 65,130 tonnes of (USGS, 2012). antimony. Inferred reserves comprise 6051 million The leading global producers of antimony ores tonnes of antimony ore at 2.34% Sb resulting in an and concentrates are listed in Table 4.8. additional 141,718 tonnes of antimony. Gold is the

Table 4.8 Leading global producers of antimony ores and concentrates in 2010. (Source: Roskill, 2011.)

Company Country Estimated production (tonnes antimony)

Hsikwangshan Shanxing Antimony Co. Ltd. China 35,000 Liu Zhou China Tin Group Co Ltd China 30,000 Chenzhou Mining Co. Ltd China 15,000 GeoProMining Russia 6500 Shwe Zin Htut, Thu Ya Kan Chun Burma (Myanmar) 5897 Beaver Brook Canada 5669 Comsup (former Anzob GOK) Tajikistan 5370 Yunnan Muli Antimony Industry Co. Ltd China 4500 Comite Boliviano de Productores de Antimonio Bolivia 3800 Banxi antimony mine (Taojiang Jiutong Group) China 3500 Village Main Reef (former Metorex) South Africa 2257 Nandan Cheshan antimony mine China 2000 Eti Holdings Turkey 2000 Mandalay Resources Australia 1106 Emusa and COMISAL Bolivia 1000 Kazzinc Kazakhstan 840 Hong Xin, Siam Thailand 600 SRS Laos 493 Kadamdzhai Kyrgyzstan 480 Sontrarem Morocco 280 Empresa Minera Bernal Hermanos Bolivia 180 Doe Run Peru Peru 120 Total 126,592 90 ulrich schwarz-schampera traditional by-product at Consolidated Murchison Projects under development and significant gold reserves have been delineated. The estimated life of the mine is 11 years. The The growth in antimony consumption, together potential for reworking of the Consolidated with the concentration of production and falling Murchison Tailings Dump is also under investiga- resource levels in China, have led to an increase tion. Resources have been estimated at 6.18 mil- in exploration activities for antimony-bearing lion tonnes, with an average grade of 0.75% Sb and deposits, mainly with antimony as a potential 0.46 g/t Au, containing a total of 46,461 tonnes of by-product. Most current projects are focusing on antimony and 2.84 tonnes of gold. epithermal and orogenic hydrothermal targets. The Beaver Brook Mine in central Newfound- Mandalay Resources Corp. operates two silver– land, Canada, is North America’s only producing gold–antimony mines at Cerro Bayo in Patagonia. antimony mine. It was re-opened in 2008 after it It recently announced plans to expand its opera- had been closed for 10 years and is now reported to tions to four mines with future annual produc- be the largest antimony mine outside China and tion of 1870 kg of gold and 1600 tonnes of antimony South Africa (Roskill, 2011). Resources are esti- (Mandalay Resources Corporation, 2012; Roskill, mated at about 100,000 tonnes of antimony and 2007). The company is working epithermal vein- the current production rate is 235 tonnes of type mineralisation at Dagny, Fabiola, Delia NW concentrate per week. The stibnite concentrate is and Yasna. Mandalay acquired the Cerro Bayo trucked to Halifax for shipping to China. In 2009, mining and concentrator complex from Coeur Hunan Nonferrous Metals Corporation (HNC), the d’Alene Mines Corp. in August 2010. largest antimony company in the world, acquired In 2012 U.S. Antimony Corporation com- 100 per cent equity of Beaver Brook Antimony pleted the construction of a flotation mill in the Mine Inc. At the end of 2009, China Minmetals State of Guanajuato, Mexico, for processing the Corporation, China’s biggest metal trader, acquired ores from its new mine at the Los Jaurez 51% equity of Hunan Holdings Group (HNG), antimony–silver deposit in the Soyatal mining the state-owned parent of HNC. The state- district and from other properties (U.S. Antimony owned Assets and Supervision Administration Corporation, 2012). U.S. Antimony Corporation, Commission of Hunan Province reserved the through its wholly owned subsidiary USAMSA, remaining 49 per cent of HNG. Small quantities of also owns and operates a smelting facility at antimony concentrates and sodium antimonate Estacion Madero in the Municipio of Parras de la were also produced as by-products from base-metal Fuente, Coahuila, Mexico. Currently, crude anti- operations in Canada, for example, at Sullivan in mony oxide and antimony metal are produced at British Columbia and Belledune in New Brunswick this smelter for further processing at the compa- (Roskill, 2011). ny’s smelter in Montana, USA. Concentrates Bolivia has a long history of antimony produc- and hand-sorted rock from Newfoundland, Peru, tion from small, high-grade vein-type deposits Honduras, Mexico, and other areas are processed which are typically associated with gold-, lead-, at the Mexican facility. zinc- and tin-bearing ores. Three distinctive belts There are several potential sources of anti- of orogenic gold and antimony deposits containing mony under investigation in Canada. These more than 500 known deposits and occurrences include the Lake George Mine in New Brunswick, are recognised along the length of the Eastern owned by Apocan Inc. (an Amspec Chemical Cordillera of the Andes. The gold ores, particu- Corp. subsidiary), which is currently on care-and- larly those of the Caracota–Carma–Candelaria maintenance (USGS, 2011). Exploration for anti- belt, may contain as much as 10 to 20% Sb and mony is also underway at other properties in were originally mined for antimony. These British Columbia, Yukon and Newfoundland. deposits have been exploited on a small scale from Companies working on antimony as a by-product pre-colonial days up to the present. include Equity Silver Mines Ltd. which is Antimony 91

assessing silver–gold–copper ores at a formerly (AMG), which is Europe’s largest ATO manufac- important silver-producing site in central British turer, operates an antimony mine and adjacent Columbia. A flotation plant could produce 1700 smelter through its subsidiary Suda Maden AS. tonnes antimony annually. Another is the Lead Tri-Star Resources plc has been investigating Smelting and Refining Factory in Trail set up by antimony resources at the historic small-scale Cominco Ltd. It produced lead-antimony alloy as workings at Göynuk in the Kutahya Province of a by-product of silver-lead ores (USGS, 2011). north-west Turkey. The company has also been In Australia antimony resources are under granted permission to construct a small-scale investigation at a number of properties, some of processing facility at Göynuk to treat stockpiled which were former gold–antimony producers. material (Tri-Star Resources plc, 2012a). These include the Hillgrove antimony–gold mine In Italy, Adroit Resources Inc. is investigating in New South Wales, which closed in 2009 and for the possibility of reactivating antimony mining which the present owner, Straits Resources Ltd, is in the Manciano area of Tuscany. This area has currently seeking a buyer. The Costerfield gold– several small and medium-sized deposits some of antimony mine in Victoria, owned by Mandalay which were worked in the past (Adroit Resources, Resources Corp., is an underground mining opera- 2012). The company is aiming to develop a new tion which produced 1577 tonnes of antimony in mine producing 10,000 tonnes per annum anti- 2011 and is planning to exceed this in 2012 mony, together with gold and silver (Clarke, (Mandalay Resources Corporation, 2012). The 2012). company has identified new reserves in the In south-east Asia both Myanmar and Laos Augusta deposit which is currently being mined produce antimony ore which is exported to and is also carrying out an economic analysis of China. Tri-Star Resources is collaborating with high-grade antimony ores in the Cuffley lode RDP Singapore Ltd to undertake a technical which was discovered in 2011. Anchor Resources, assessment of the latter’s antimony projects owned by China Shandong Jinshunda Group Co which include exploration rights around two pro- Ltd., is exploring for antimony–gold mineralisa- ducing deposits (Tri-Star Resources plc, 2012b). tion at the Bielsdown project in north-eastern New South Wales. Anchor has published new resource estimates for the Wild Cattle Creek anti- World trade mony mine, located in the Bielsdown licence area, which was worked intermittently on a small scale Antimony is traded in the form of ores and con- from the late 1800 s up to the 1970s (Anchor centrates, semi-refined unwrought antimony Resources Ltd, 2012). Northwest Resources Ltd is metal, antimony trioxide and antimonial lead. planning to develop a gold–antimony mine at its The international trade is traditionally character- Blue Spec Shear project in Western Australia. A ised by the distinction between primary anti- preliminary technical and economic assessment mony production of ores, concentrates and was completed in 2012 which concluded that the unrefined metal and trioxide in industrialising project has the potential to be a low-cost operation countries and the refining and consumption of producing 1900 tonnes of antimony and 68,500 antimony in industrialised countries. ounces of gold per annum over a mine life of five Over the last decade this pattern has changed years (Northwest Resources Limited, 2012). due to the strong growth in the manufacturing There are a number of active antimony-related sectors in the industrialising countries. In ventures in Turkey. Production has historically particular, China has pursued a value-added man- been dominated by Ozdemir Antimony Mining ufacturing policy and imposed export restrictions Joint Stock Co from mines in the Turhal district on crude antimony products. It has become the of northern Turkey (Clarke, 2012). The Dutch major producer and exporter of refined antimony corporation Advanced Metallurgical Group NV metal, compounds and intermediate products 92 ulrich schwarz-schampera

(Roskill, 2007). The combination of mine and abundant minor metals such as molybdenum and smelter closures, production controls, export tungsten. This may be attributed to the fact that quotas, environmental restrictions and dwindling mining costs for antimony are low because it resources in China and the support provided by occurs in highly concentrated, almost pure stib- the government to Chinese companies operating nite deposits, locally with gold as a by-product. overseas have, together, contributed to high There has been significant price volatility in levels of interest in antimony mining in Bolivia, the antimony market over the past 40 years. In Mexico, Australia and Canada and has led to 1970 the rapidly rising production of plastics and increased trade in antimony concentrates. The the establishment of laws that regulated flam- main exporters of concentrate in 2010 were mability of textiles and other materials, together Myanmar, Russia, Canada, Tajikistan and with the concentration of mine production in a Australia (Figure 4.7a); the overwhelming few countries, caused a considerable price spike majority of shipments from these countries went (Figure 4.8). In 1974 sharply increased demand, to China, which accounted for about 90 per cent especially for antimony trioxide, and supply dis- of the total (Roskill, 2007 and 2011). Recorded ruptions from China resulted in another major exports of antimony metal do not entirely reflect price peak. Since the mid-1980s antimony price the trade volume as most of the exports from instability may be largely attributed to the China are not included in the official export data. concentration of production in China and to The main trade flows, once unofficial shipments Chinese government policy. In early 1995 the are taken into account, are from China to imposition of a 20 per cent export tax on anti- Belgium, France, South Korea, Japan and the USA mony ores and concentrates contributed to a (Figure 4.7b; Roskill, 2011). Until 2005, South steep price rise, which reached US$6000 per Africa was the main exporter of antimony oxide, tonne. Other control measures, including the but China now accounts for about 55 per cent of suspension of the issuance of new mining export shipments. Other leading exporters of licences, the imposition of export quotas and a antimony oxide are the companies converting crackdown on illegal mining and smuggling, antimony metal to oxide in Belgium and France. failed to rectify an oversupply situation and low prices persisted from 2001 to 2004. However, prices rose sharply between 2004 and 2008 as Prices demand strengthened, stockpiles were depleted and Chinese control measures took effect. From Antimony is not traded on international metal early 2009, the antimony prices declined consid- exchanges and prices are agreed between producer erably because of the global economic crisis, or trader and consumer, depending on the quality with European prices at the end of the year and form of the product sold. Antimony metal falling to US$6050 per tonne for antimony produced by smelters, generally with purity bet- trioxide. ween 99.5% and 99.6%, is traded in US dollars In 2010 mine closures in China restricted per pound. Further refining is carried out in global supplies and led to the doubling of anti- smelters or by secondary processors and manu- mony metal prices to more than US$12,000 per facturers to produce antimony trioxide and anti- tonne at the end of the year. The upward trend monial lead. The estimated value of primary continued in the first part of 2011 with prices antimony mine production in 2011, based on the climbing to more than US$16,000 per tonne by annual average New York dealer price, was about April. Thereafter, the price declined gradually to US$ 1794 million. end 2011 at about US$13,000 per tonne. During Despite its geochemical scarcity, given its the first three quarters of 2012 antimony metal important industrial uses the commodity price has traded in the range US$12,000–14,000 per of antimony is low when compared to more tonne. Antimony 93

(a) Export 35

30

25

20

15

Thousand tonnes Thousand 10

5

0

Russia Mexico Bolivia Turkey CanadaTajikistan VietnamAustralia Lao, PDR South AfricaKazakhstan United States Other Countries Burma (Myanmar) China (inc Hong Kong)

(b) Import 35

30

25

20

15

Thousand tonnes Thousand 10

5

0

Japan India Spain Italy Brazil FranceBelgium Thailand Austria Kyrgyzstan United States Netherlands Korea (Rep. of) Other Countries China (inc Hong Kong) Ores and concentrates Metal

Figure 4.7 (a) Main antimony exporting countries, 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013) (b) Main antimony importing countries, 2009 (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) 94 ulrich schwarz-schampera

16,000

14,000

12,000

10,000

8000

6000

4000

2000

Annual average antimony price (US$/tonne) 0 1900 1904 1908 1912 1916 1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2000 2004 2008 2012

Figure 4.8 Annual average antimony prices between 1900 and 2011. (Data from BGR database.)

Environmental aspects considered to be a highly reactive element in the oceans, with an average concentration of about Antimony is widely dispersed in the environment 200 ng/l. Ultimately, antimony in the environ- as a result of natural processes such as volcanic ment will end up either in soil or sediment, where eruptions and erosion of antimony-bearing rocks its mobility is controlled mainly by pH and by the and minerals. Anthropogenic emissions of anti- presence of hydrous oxides of iron, manganese mony from burning fossil fuels, mining, smelting, and aluminium to which the antimony is refining and waste incineration are also significant, adsorbed. Sorption of antimony by organic matter but there is little modern data available to assess may also be an important factor in some condi- their magnitude (Belzile et al., 2011). Antimony is tions. Elevated concentrations of antimony in found in very low concentrations in soils, waters soils and sediments are generally related either to and air. It occurs mainly as antimony(III) and anthropogenic sources or are associated with high antimony(V) in environmental, biological and arsenic concentrations in sulfide ores. Information geochemical samples (Filella et al., 2002). Most on the processes controlling the solubility and studies report the dominance of antimony(V) speciation of antimony in water, sediment and under oxic conditions although the presence of soil was reviewed in a risk assessment report significant proportions of antimony(III) and low for ATO prepared for the European Commission concentrations of methylated antimony species (EU RAR, 2008). are also documented (Nishimura and Umetsu, Antimony has no known biological function 2000). Typical concentrations of total dissolved and is not known to be used by any organism. antimony are less than 1.0 μg/l in unpolluted However, given its currently high level of use, waters although the range is quite large, reflecting there is concern about potential pollution and the wide range of conditions in natural systems risks to human health. Long-term, low-level oral and proximity to pollution sources (Filella et al., and dermal exposure from food, air and drinking 2002). Much lower values have been reported in water has negligible health effects (e.g. EU RAR, unpolluted groundwaters in southern Ontario, 2008). However, respiratory irritation, derma- Canada (Shotyk et al., 2006). Antimony is not titis, pneumoconiosis and gastrointestinal Antimony 95

symptoms have been reported in workers in the adverse health effects (International Antimony antimony processing industry, but none of these Association, 2012). symptoms have been directly linked with The US National Institute of Occupational antimony exposure (Sundar and Chakravarty, Safety and Health (NIOSH) has set a recommended 2010). exposure limit for antimony of 0.5 mg/m3 averaged Antimony trioxide (ATO) is classified in the over a 40-hour working week (NIOSH, 2010). In EU as ‘suspected of causing cancer via inhalation’ addition, various countries and the World Health according to Regulation (EC) 1272/2008. The EU Organisation (WHO) have established safe limits for RAR (2008) concluded that there was no human the antimony content of drinking water at which no health risk to consumers (through the use of adverse health effects are likely to occur – for products such as electrical and electronic equip- example, 5 μg/l in the European Union (European ment, PET bottles and flame-retarded items such Commission, 1998), 6 μg/l in the USA (US as textiles and carpets) or to those exposed to Environmental Protection Agency, 2012) and 20 μg/l ATO through environmental sources. It also by WHO (WHO, 2003). stated that there was no need at that time for Given the increasing range of uses for anti- further information or testing and no need for mony, the various routes through which it can risk-reduction measures beyond those being get into the environment and the potential effects applied already. These conclusions have since of a number of antimony species, it is clear that been validated by international experts and are more research and monitoring is needed on its considered to remain valid today (e.g. OECD, biogeochemical cycling and fate in the environ- 2008; Environment Canada, 2010). ment. Some studies have detected antimony in Those most at risk from exposure to antimony bottled waters and commercially available fruit are the workers in industries that process juices although the antimony levels were all antimony ore and metal, and those that make below the WHO guideline for drinking water and antimony chemicals or are involved in the manu- the source of the antimony is unknown (Shotyk facture of products in which they are incorpo- et al., 2006; Hansen et al., 2010). rated. The EU RAR assessed a wide range of possible health effects related to ATO exposure based on human exposure data and animal labo- Outlook ratory studies. Three areas of possible concern were identified: skin irritation for workers under The use of antimony in flame retardants is conditions of perspiration; toxicity to the lung as expected to remain its principal market in the a result of repeated long-term exposure through future, although its application in the production inhalation; and the development of tumours in of polyethylene terphthalate (PET) for plastic bot- the lungs of female rats as a result of inhalation tles and synthetic textiles and for the vulcanisa- exposure. It was concluded that additional data is tion of rubber is likely to increase. only needed for repeated long-term inhalation The demand for use in flame retardants has exposure to better understand the risks to human increased in recent years and is likely to continue health. Additional research is ongoing to investi- to grow as fire regulations become more stringent gate the mechanism and inhalation effects of ATO and widely imposed worldwide, particularly in on lungs. In general, the utilisation of effective Asia, eastern Europe and Latin America. The trend working practices and equipment, combined with of use of antimony in lead-acid batteries is difficult strict adherence to guidelines, ensures that to assess. Alloys of antimony with lead and anti- occupational exposure and environmental dis- monial lead have shown a decrease in use as they charges are minimised and meet the official are increasingly substituted on environmental requirements. Long-term monitoring of workers grounds in some applications. However, the market in an ATO production plant has not identified any for antimony in lead-acid batteries may expand 96 ulrich schwarz-schampera

significantly as the automotive sector, especially References in Asia, continues to grow. The production rate of electric and hybrid vehicles with lead-acid bat- Adroit Resources Inc. (2012) Valle Lupara. http:// teries is also expected to increase. The demand for adroitresources.ca/valle-lupara/ antimony in other metallic uses, such as bearings, Amspec (2011) Flame retardants: antimony trioxide. tends to be stable. Health and environmental con- http://www.amspec.net/products/flame-retardants/ cerns may limit the use of antimony in some appli- antimony-trioxide/ cations but the important role of antimony in Anchor Resources Ltd. (2012) Bielsdown Project, NSW. catalysts for PET production is expected to con- http://www.anchorresources.com/index.php/projects/ nsw/bielsdown-nsw/ tinue to be important. On balance it is anticipated Anderson, C.G. (2000) A survey of primary antimony that worldwide demand for antimony will con- production. In: Young, C. (ed.) Minor Elements 2000: tinue on an upward trend at least in the short term. Processing and Environmental Aspects of As, Sb, Se, Antimony production in China is unlikely to Te, and Bi. SME, Littleton, Colorado, 261–275. increase in the near future and could even decrease Belzile, N., Chen, Y. and Filella, M. (2011) Human as a consequence of mine and smelter closures. Exposure to Antimony: I. Sources and Intake. Critical There is evidence of increased activity in several Reviews in Environmental Science and Technology countries, such as Thailand and Myanmar, but 41, 1309–1373. assessing future production trends is difficult British Geological Survey (2012) World mineral production because of lack of information. No Asian country 2006–10. (Keyworth, Nottingham: British Geological Survey). other than China produces more than a few thou- Buchholz, P., Oberthür, T., Lüders, V and Wilkinson, J. sand tonnes a year and there may be limited scope (2007) Multistage Au-As-Sb Mineralization and to expand output. Production in South Africa is Crustal-Scale Fluid Evolution in the Kwekwe District, likely to increase with the new ownership at Midlands Greenstone Belt, Zimbabwe: A Combined Consolidated Murchison, while production in Geochemical, Mineralogical, Stable Isotope, and Fluid Bolivia will likely fluctuate in line with demand Inclusion Study. Economic Geology 102, 347–378. and will probably increase with the planned Butterman, W.C. and Carlin Jr. J.F. (2004) Antimony. restart of the government-owned Vinto smelter. U.S. Geological Survey Open-File Report 03-019. Australian production could increase by as much Clarke, G. (2012) Antimony (trioxide!) on the watch as 5500 tonnes per annum if the Hillgrove opera- list. Industrial Minerals, September 2012. tion emerges from care and maintenance and if Dill, H.G., Weiser, T., Bernhardt I.R., and Kilibarda, C.R. (1995) The composite gold-antimony vein deposit at the new investors in Anchor Resources, China Kharma (Bolivia). Economic Geology 90, 51–66. Shandong Jinshunda Group, bring the Bielsdown Dubé, B. and Gosselin, P. (2007) Greenstone-hosted deposit on stream (Roskill, 2011). The new US quartz-carbonate vein deposits. In Goodfellow, W.D. Antimony Corporation operation in Mexico (ed.) Mineral deposits of Canada: A synthesis of major should be fully operational by 2013, but will only deposit types, district metallogeny, the evolution add a maximum of 1600 tonnes per annum. of geological priovinces, and exploration methods. Production from Russia could be augmented by Geological Association of Canada, Mineral Deposits 2800 tonnes per year if RusAnt succeed in devel- Division, Special Publication No.5, 49–73. oping the Iliskoye deposit (Roskill, 2011). In total, European Commission (1998) Council Directive 98/83/ new projects could add up to 11,200 tonnes per EC of 3 November 1998 on the quality of water intended for human consumption. http://eur-lex. year antimony to world mine capacity within the europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:319 next four to five years (Roskill, 2011). 98L0083:EN:NOT Antimony prices are expected to remain at Environment Canada (2010) Draft screening assessment high levels as a result of reduced supply from for the challenge. Antimony oxide. Environment Canada China and increasing demand, especially as legis- Health Canada, March 2010. http://www.ec.gc.ca/ lation for fire proofing becomes prevalent in substances/ese/eng/challenge/batch9/batch9_1309- industrialising countries. 64-4_en.pdf. Antimony 97

EU RAR (2008) European Union risk assessment report http://webnet.oecd.org/hpv/UI/handler.axd?id= [draft]. Diantimony trioxide CAS No.: 1309-64-4. 13e93c97-6605-4eac-961f-8af23cc6ad32. Luxembourg: Office for Official Publications of the Penga, J.-T., Hua, R.Z. and Burnardb, P.G. (2003) European Communities. http://ecb.jrc.ec.europa.eu/ Samarium–neodymium isotope systematics of hydro- DOCUMENTS/Existing-Chemicals/RISK_ASSESS thermal calcites from the Xikuangshan antimony MENT/REPORT/datreport415.pdf. deposit (Hunan, China): the potential of calcite as a geo- Filella, M., Belzileb, N., Chen, Y.-W. (2002) Antimony in chronometer. Chemical Geology 200 (1–2), 129–136. the environment: a review focused on natural waters: Robert, F. (2001) Syenite-associated disseminated gold I. Occurrence. Earth-Science Reviews 57, 125–176. deposits in the Abitibi greenstone belt, Canada. Hansen, C., Tsirigotaki, A., Bak, S.A., Pergantis, S.A., Mineralium Deposita 36, 503–516. Stürup, S., Gammelgaard, B. and Hansen, H.R. (2010) Roskill (2007) The Economics of Antimony. Roskill Elevated antimony concentrations in commercial juices. Information Services Ltd., 10th edition, London, Journal of Environmental Monitoring 12, 822–824. pp. 231. Hart, C.J.R. (2007) Reduced intrusion-related gold sys- Roskill (2011) ANCOA Ltd. Study of the antimony tems. In: Goodfellow W.D. (ed.) Mineral deposits of market. 17 October 2011, London, pp. 16. Canada: A synthesis of major deposit types, district Roskill (2012) Antimony: Global Industry Markets and metallogeny, the evolution of geological priovinces, Outlook, 11th Edition 2012. Roskill Information and exploration methods. Geological Association of Services Ltd.. Canada, Mineral Deposits Division, Special Publication Shotyk, W., Krachler, M. and Chen, B. (2006) No.5, 95–112. Contamination of Canadian and European bottled Hedenquist, J.W., Arribas, A.R., and Gonzalez-Urien waters with antimony from PET containers. Journal (2000) Exploration for epithermal gold deposits, Chapter of Environmental Monitoring 8, 288–292. 7. In: Hagemann, S.G. and Brown, P.E., (eds.) Gold in Simmons, S.F., White N.C. and John, D.A. (2005) 2000. Reviews in Economic Geology 13, 245–277. Geological characteristics of epithermal precious and International Antimony Association (2012) http:// base metal deposits. In: Hedenquist, J.W., Thompson, www.antimony.be/en/detail_44.aspx. J.F.H., Goldfarb, R.J. and Richards, J.P. (eds.) Economic Mandalay Resources Corporation (2012). Annual Report Geology 100th Anniversary Volume, 485–522. 2011. http://www.mandalayresources.com/index. Sundar, S. and Chakravarty, J. (2010) Antimony cfm?pagepath=Investors/Annual_Reportsandid=20047 Toxicity. International Journal of Environmental MMTA (2012) Minor Metals in the Periodic Table: Sb - Research and Public Health 7, 4267–4277. Antimony http://www.mmta.co.uk/metals/Sb/1 Taylor, B.E. (2007). Epithermal gold deposits. In: NIOSH (2010) NIOSH Pocket Guide to Chemical Goodfellow W.D. (ed.) Mineral deposits of Canada: A Hazards: Antimony. http://www.cdc.gov/niosh/npg/ synthesis of major deposit types, district metallog- npgd0036.html eny, the evolution of geological priovinces, and explo- Nishimura, T. and Umetsu, Y. (2000) Chemistry on ration methods. Geological Association of Canada, elimination of arsenic, antimony, and selenium from Mineral Deposits Division, Special Publication No.5, aqueous solution with iron(III) species. In: Young, C. 113–139. (ed.) Minor Elements 2000: Processing and Thompson, J.F.H., Sillitoe, R.H., Baker, T., Lang, J.R. Environmental Aspects of As, Sb, Se, Te, and Bi. SME, and Mortensen, J.K. (1999) Intrusion-related gold Littleton, Colorado, 105–112. deposits associated with tungsten-tin provinces. Northwest Resources Limited (2012) Positive scoping Mineralium Deposita 34, 323–334. study for the Blue Spec Shear Gold-Antimony Project. Tri-Star Resources plc (2012a) Operational Update, http://www.nw-resources.com.au/images/north- June 20th. http://www.tri-starresources.com//news/ west---iepho.pdf detail/77 Obolensky, A.A., Gushchina, L.V., Borisenko, A.S., Tri-Star Resources plc (2012b) Antimony technical col- Borovikov, A.A., Pavlova, G.G. (2007) Antimony in laboration in Myanmar. http://www.tri-starresources. hydrothermal processes: solubility, conditions of com//news/detail/82 transfer, and metal-bearing capacity of solutions. UN Comtrade (2013) United Nations Commodity Trade Russian Geology and Geophysics 48 (12), 992–1001. Statistics Database, Department of Economic and OECD (2008) SIDS Initial Assessment Profile for Social Affairs/ Statistics Division, http://comtrade. diantimony trioxide. SIAM 27, 14-16 October 2008. un.org/db/ 98 ulrich schwarz-schampera

U.S. Antimony Corporation (2011) Processing of anti- Murchison Schist Belt, Rooiwater Complex and 1348 mony ores: Metallurgy. http://www.usantimony. surrounding granitoids, Memoir of the Geological com/metallurgy.htm Survey of South Africa, 81, pp. 139. U.S. Antimony Corporation (2012) U.S. Antimony Village Main Reef (2012) Annual report for 2011. www. reports start up of Mexican flotation mill. http:// villagemainreef.co.za/ir/f/Village-Main-Reef-annual- www.usantimony.com/USAC_news.htm report-2011.pdf US Environmental Protection Agency (2012) Basic WHO (2003) Antimony in drinking water. Background information about antimony in drinking water. http:// document for development of WHO Guidelines for water.epa.gov/drink/contaminants/basicinformation/ Drinking-water Quality. Available from http://www. antimony.cfm who.int/water_sanitation_health/dwq/chemicals/ USGS (2011) Minerals Yearbook 2010, Antimony. antimony.pdf http://minerals.usgs.gov/minerals/pubs/commodity/ Wu, J. (1993) Antimony vein deposits of China. Ore antimony/myb1-2010-antim.pdf. Geology Reviews 8, 213–232 USGS (2012) Mineral commodity summaries 2012, Yang, D.-s, Shimizu, M., Shimazaki, Li, X.-h. and Xie, Antimony. http://minerals.usgs.gov/minerals/pubs/ Q.-I. (2006) Sulfur Isotope Geochemistry of the commodity/antimony/mcs-2012-antim.pdf Supergiant Xikuangshan Sb Deposit, Central Hunan, Vearncombe, J. R., Barton, J. M., Cheshire, P. E., De China: Constraints on Sources of Ore Constituents. Beer, J. H., Stettler, E. H., and Brandl, G. (1992) 1347 Resource Geology 56 (4), 385–396. Geology, geophysics and mineralization of the 5. Beryllium

DAVID L. TRUEMAN1 AND PHILLIP SABEY2

1 Consulting Geologist, Richmond, British Columbia, Canada 2 Manager, Technology and Quality, Materion Natural Resources, Delta, Utah, USA

Introduction them and to manufacture beryllium products. These are the United States, Kazakhstan, China Beryllium, chemical symbol Be, is the fourth and India. At present, Materion Corporation1 element in the Periodic Table. It is a constituent ( formerly Brush Wellman Inc.) in the United element in various gemstones, including emerald, States is the only fully integrated beryllium chrysoberyl and aquamarine, which appear to producer in the world, involved in the mining, have first been traded in Egypt about 100 BC ore processing, manufacture, sale and recycling of (Harell, 2004). beryllium-bearing products. It is estimated that Beryllium was discovered in 1797 by the in 2010, Materion produced over 70 per cent of French chemist Vauquelin, who suspected the the world’s mined beryllium ore (Merchant existence of a common element in the minerals Research and Consulting, 2012). The remaining beryl and emerald. It was named glucina by production comes from stockpiled ores in Vauquelin in 1798 after its sweet-tasting salt. Kazakhstan and China, together with minor con- Klaproth (1801) proposed the alternative name tributions from beryl sourced from South beryllia to avoid potential confusion with the America, Africa and India. sweet salt of yttria. However, the name glucin- Several countries, such as Japan and France, ium continued in use, largely in France, until have the capability of processing intermediate the name beryllium was formally adopted by the compounds of beryllium, such as beryllium International Union of Pure and Applied hydroxide, into final products. Chemistry (IUPAC) in 1949. In 1828 elemental beryllium was isolated inde- pendently by two chemists, Wohler in Germany Properties of beryllium and Bussy in France, through the reduction of the salt, beryllium chloride. Its commercial use began Beryllium is a silver-grey metal noted for its light in 1926 and, by 2006, the most recent year for weight with a density of 1846 kg/m3, comparable which figures are available, grew to become a to that of magnesium, and its remarkable stiff- $700 million dollar a year business (Sabey, 2006). ness. Beryllium has the unique property for a Four countries currently have the resources metal of being virtually transparent to X-rays. and capability to mine beryllium ores, to process The fact that sound travels through beryllium

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 100 david l. trueman and phillip sabey

Table 5.1 Selected properties of beryllium. Distribution and abundance in the Earth’s crust Property Value Units

Symbol Be Beryllium has an average crustal abundance of Atomic number 4 about 2.1 ppm (Rudnick and Gao, 2004). Atomic weight 9.01 Systematic studies of the beryllium contents of Density at 25°C 1846 kg/m3 over 500 samples of various rock types by Beus Melting point 1287 °C (1966) gave mean beryllium values of less than Boiling point 2475 °C 2 ppm in most sedimentary rocks and in mafic Hardness (Mohs scale) 5.5 and ultramafic igneous rocks. Felsic igneous Specific heat capacity at 25°C 1.82 J/(g °C) rocks returned higher values with 4 ppm in Latent heat of fusion 1350 J/g Coefficient of linear thermal 11.5 × 10−6 /°C biotite granites and 9 ppm in two-mica granites. expansion In various unusual rock types from the Khibiny Thermal conductivity 210 W/(m °C) and Lovozero alkaline complexes in the Kola Electrical resistivity at 25°C 0.38 µΩ m Peninsula, north-west Russia, mean values of Young’s modulus 287 GPa 4 ppm Be were reported in mela-nepheline sye- Mass magnetic susceptibility 1.26 × 10−8 m3/kg nites and 20 ppm Be in clinopyroxene-nepheline Brinell hardness 600 MN/m2 rocks. faster than any other metal is also important in Uses of beryllium certain applications. When added to other metals, particularly Many of the physical properties of beryllium, copper, as an alloying element, beryllium pro- such as its rigidity, low weight, heat-absorbing vides controllable strengthening mechanisms, capability and dimensional stability, are exploited along with many other valuable attributes such in its applications as a metal, yet the greatest use as good electrical and thermal conductivity and is in its alloys with other metals and in ceramics. very low friction against most bearing surfaces. Figure 5.1 summarises the main end uses of beryl- It also has non-magnetic and non-sparking lium, divided into four main sectors. It should be properties. stressed that there is considerable overlap among Beryllium is the first element of Group 2, the these categories and hence an accurate breakdown alkaline earth elements, in the Periodic Table. In of their relative importance is difficult to obtain. natural systems its oxidation state is Be2+ and, For example, a copper–beryllium alloy may with a high charge / ionic radius ratio, it tends to find use in components of aerospace, military, form strong covalent bonds. The beryllium electronic, automotive or other hardware. nucleus also contains an extra neutron which Similarly, electronic uses of copper–beryllium can be easily dislodged by gamma radiation. include connector terminals that may be used in This property is useful in nuclear applications applications as diverse as aircraft guidance sys- and in the detection, evaluation and mining of tems, mobile device electromagnetic radiation beryllium deposits. shielding, automotive airbag impact sensors and Beryllium has 12 known isotopes of which medical diagnostic hardware. only one is stable, 9Be. Two beryllium isotopes, The three primary forms of beryllium used 7Be and 10Be, are cosmogenic in nature and 10 Be commercially are: can be used for age dating of very young and ● alloys containing small amounts of beryllium, recent geological events (Dunai, 2010). especially copper–beryllium; Physical properties germane to the end uses of ● pure beryllium metal; beryllium are listed in Table 5.1. ● beryllia (BeO) ceramics. Beryllium 101

500 450 400 350 300 250 200 150 100 50 0 Beryllium contained (metric tonnes) 2011 2005 2006 2007 2008 2009 2010 1999 2000 2001 2002 2003 2004

Figure 5.1 Consumption of beryllium by Aerospace Electrical Electronic end-use sector, 1999–2011. (Data from BeST, 2012a.) Other (mechanical, oil & gas drilling/exploration)

Alloys containing less than 2% beryllium, utilised for the manufacture of components used especially copper–beryllium in a myriad of applications, for example: The largest end use of beryllium, comprising 75 ● connector terminals for high-reliability per cent of the total world production, is in the electrical and electronic connections between form of alloys of copper containing less than two circuit boards, wires and components in elec- per cent beryllium. Smaller quantities of beryl- tronics, telecommunications and appliance lium are also used to make alloys of aluminum equipment, automobile systems such as air- and nickel. The beryllium content provides the bag crash sensor and deployment mechanisms, alloys with valuable physical property enhance- anti-lock brake systems, dynamic suspensions, ment as a result of the beryllium reacting with etc.; supplementary alloying elements, such as nickel ● relays for controlling industrial, domestic and and cobalt, to form local concentrations of inter- automobile electrical equipment; metallic compounds known as berylides, which ● electromagnetic radiation shielding spring locate themselves within the crystalline struc- strips used to prevent ‘leakage’ from electronic ture of the metal matrix as precipitates. There, wireless devices that can interfere with other they produce local distortions or other irregular- equipment or become life threatening; ities in the crystal structure of the matrix that ● diaphragms for pressure sensing in aircraft impede the smooth ow of dislocations caused altimeters, medical stethoscopes and sphygmo- by stresses applied to the alloy and thus offer manometers, aneroid barometers and automobile enhanced strength, stiffness and hardness to the engine timing sensors; alloy, while retaining relatively good ductility, ● long service-life springs such as for fire sprinkler machinability, electrical and thermal conduc- water-control valves. tivity. The highly predictable and precise attain- Copper–beryllium alloys in thick plate, rod ment of a desired combination of those properties and tube form are used for the manufacture of: offers many options for the design of critical ● non-magnetic equipment components used in components. oil and gas exploration and production equip- The vast majority of copper–beryllium alloys ment such as directional drilling systems; coal are sold in thin strip between 0.1–0.4 mm and rod and minerals mining equipment; mine detection form in diameters from 0.5–7.5 mm, and they are and minesweeping; 102 david l. trueman and phillip sabey

● undersea cable signal amplification ‘repeater’ freedom of design. Its formability, machinability housings; and joinability allow for relative ease of manufac- ● low-friction, high-strength aircraft landing ture of complex structures. The isotropy and gear, control rod and wing aileron / ap-bearing thermal properties of beryllium serve to minimise bushings; distortions in sophisticated dimensional applica- ● non-sparking, high-strength tools used in anaes- tions, as exemplified by the ability to machine thetic gas controls; petrol refinery tools, chemical complex curved surfaces, such as the faces of the and explosives manufacture; 6.3-metre diameter astronomical telescope mirrors ● plastic injection and blow-moulding moulds in the Hubble and James Webb space-based obser- where the high thermal conductivity, strength vatories, which also rely upon its ability to be and ease of machining properties are exploited. highly polished and accept coatings for enhanced reectivity at operating wavelengths. Another unique property of the metal is that it Pure beryllium metal and alloys containing is highly transparent to X-rays. In thin foil form, over 60% beryllium beryllium is used as the window material for X-ray The second largest consumption of beryllium, tubes and detectors to permit the X-rays through accounting for 20 per cent of the volume of while maintaining vacuum. It is especially useful beryllium produced, is in the form of beryllium in security devices and high-resolution imaging metal of purity >99.5 per cent and alloys con- technology, such as medical applications like taining over 60 per cent beryllium. The principal mammography to detect breast cancer. properties of beryllium metal of commercial Beryllium is a very efficient moderator of interest are its low density, high strength, high neutrons, slowing and reecting them, a property rigidity, structural stability at high tempera- that finds application in materials test reactors tures, thermal conductivity, high transmission and in fundamental particle-physics research. of sound and its transparency to X-rays. It is the material of choice for the wall-lining Beryllium is notable among metals in terms of material relied upon to control the high-tempera- its specific rigidity; i.e. the ratio of modulus to ture gas plasma of fusion processes, such as that density, which is approximately fifty per cent greater than that of steel, while its density (1846 kg/m3) is about 30 per cent less than that of aluminium. Beryllium metal is typically used to produce components utilised in high-technology equip- ment. In situations where weight is a critical factor, such as for structures to be launched into space, it is imperative that such structures are rigid and not subject to distortions or resonant vibrations which might reduce the accuracy of their instrumentation. Beryllium metal is the optimal material for these purposes, principally because of its high specific rigidity. It also has attractive thermal properties that reduce thermal distortions, both at high temperatures experi- Figure 5.2 Six of the 18 mirrors of the James Webb enced during launch and descent, and also at the space telescope undergoing cryogenic testing at the extremely low temperatures of space. Marshall Space Flight Center, Huntsville, Alabama, Beryllium is an isotropic material, having uni- USA. (Photo: K. Hutchison, courtesy of Ball Aerospace form properties in all directions, which increases and Technologies Corp.) Beryllium 103 used in the current JET (Joint European Taurus) mining of beryl, associated with gemstone min- reactor and the larger-scale ITER (International ing, continues globally and provides a small Thermonuclear Experimental Reactor) fusion proportion of the feedstock for USA operations, reactor project, now under construction in France. but a significant share for Chinese operations. The physical properties of beryllium can be Since 1962 a dedicated mine operated by modified by the addition of up to 62 per cent by Materion Corporation in the Spor Mountain weight of aluminium, thereby producing an alloy region of Utah, USA, has provided an estimated with enhanced ductility which may be easily 84 per cent of global production which is machined by conventional metal-cutting tech- predominantly derived by the refining of the niques. This alloy finds many applications bertrandite ores extracted from that site, aug- for lightweight high-strength components of mented with a small volume of imported beryl aerospace and electronics systems such as struc- and recycled materials. tural components of aerospace and munitions In 2010, three countries, the USA, Kazakhstan guidance systems. and China, produced most of the commercial beryllium and derived products. Production of beryllium in China (about 16 per cent of Beryllia (BeO) ceramics the world total) is from both local ores and Beryllium oxide, or beryllia (BeO), ceramic imported beryl, while Kazakhstan derives its accounts for the remaining five per cent of beryl- beryllium from historic inventories of Russian lium consumption. ore concentrates. A breakdown of global pro- Beryllia ceramics have an exceptional combi- duction by region for the main forms in which nation of properties, the most important of which is beryllium is used commercially is given in the combination of high electrical insulation, cou- Table 5.2. The estimates presented include pled with a hardness only slightly lower than that beryllium from all sources including both newly of diamond, and thermal conductivity an order of mined and stockpiled ore, government inven- magnitude greater than that of alumina. Its low tories and scrap. dielectric constant and low loss index permit its use Walsh and Vidal (2009) indicate that China has as an electronics circuit substrate with extremely an annual capacity of about 1500 tonnes of beryl good performance at high frequencies. production, about half of which is located in The coefficient of thermal expansion of beryl- Xinjiang Province and the remainder in Guangdong lia is intermediate between that of silicon or Province, with a fraction of this currently being gallium arsenide and that of typical metals, mak- worked. Processing of the ore from the Koktokay ing it compatible with many adjacent structural mine in Guangdong Province was carried out at components. the Fuyun facility, and was originally exported to Beryllia is widely used as an electrical insu- the former USSR. Fuyun is reported to have an lator, and in heat sinks for radio-frequency and annual capacity of 100 tonnes/year of beryllium radar equipment, automotive electrical systems oxide and 1000 tonnes/year of copper–beryllium and for laser bores and microwave waveguides. In alloys. This area also hosts about 70 per cent of all of those applications, the combination of heat China’s beryllium resources which are in the extraction and electrical insulation is essential. form of the mineral beryl. In addition, China imports an estimated 800 tonnes of beryl annually, which, at an average World production BeO content of 11 per cent, would provide a total annual production of approximately 35 tonnes of Until 1962, beryllium mine production, mostly beryllium at normal refinery efficiencies. in the form of beryl from artisanal operations, The British Geological Survey (BGS) compiles had come from at least 14 countries. Artisanal annual statistics for beryllium production from 104 david l. trueman and phillip sabey

Table 5.2 Estimated global production of beryllium from all sources by region in 2011 (Source: Beryllium Science and Technology Association, 2012a.)

USA, Japan and China (b) Others (b) Total Kazakhstan (a) kilograms kilograms kilograms kilograms

Be contained in metal 43,000 5000 1500 49,500 >60% Be content Be contained in alloys 283,000 59,000 5000 347,000 <60% Be content Be contained in 2500 1000 0 3500 BeO ceramics Total Be contained 328,500 65,000 6500 400,000 in all products produced in 2011

(a) Total aggregated by an independent auditor from data provided by BeST members. (b) Estimates.

9

8

7

6

5

4

3 Thousand tonnes Thousand 2 Figure 5.3 Global annual mine production of beryl for the period 1 1992–2010 (production from bertrandite 0 ore is calculated as equivalent to beryl containing 11% beryllium oxide). (Data

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 from British Geological Survey, 2012.) currently operating mines (Figure 5.3). The data its present very small size. However, it is also are shown as tonnes of beryl equivalent, based on important to note that between 2000 and 2010 a BeO content for beryl of 11 per cent. Materion Corporation temporarily stopped pro- The rapid decline in global production after duction of beryllium metal to allow construction 1998 is thought to be due in part to a slowing of of new facilities in Ohio for the refining of high- the ‘technology revolution’ (Cunningham, 2004a), purity beryllium metal. During that period, and ongoing miniaturisation of electronic devices Materion replaced production of metallic beryl- (Figure 5.3). The drive for miniaturisation in elec- lium from newly mined ore with beryllium metal tronics has significantly reduced the amount of from the stockpile held by the US Defense copper alloys used, and hence that of beryllium, Logistics Agency. Other beryllium users also in this sector. This is illustrated graphically by took advantage of government surplus sales at the transformation of the mobile phone from that time, so the world requirement for newly near brick-size proportions of the early 1990s to mined beryl and beryllium ores was depressed Beryllium 105 temporarily. Materion’s new plant, which com- The large import of beryllium products into menced production in 2011, was built with US France and the USA reects the intake of copper– government financial assistance and is currently beryllium intermediate shapes by the NGK producing high-grade pure beryllium. Another Metals Corporation copper–beryllium alloy finish- contribution to the recent general trend of ing faci lities in Couëron, near Nantes, France and increasing production may be due to increases in Sweetwater, Tennessee in the USA. These are demand for copper–beryllium alloy which may sourced from the parent company’s alloy produc- reect the growing use of electronics, and conse- tion facility in Chita, Japan which imports refined quently the number of copper–beryllium alloy beryllium oxide from Materion USA for reduction connectors, in automotive applications (Jaskula, to copper–beryllium alloys (Jaskula, 2012b). 2012a).

300

World trade 250

Global trade in beryllium in 2009 is summarised 200 in Figures 5.4 and 5.5. 150 Chinese beryl and beryllium-containing Tonnes products have been intermittently available in 100 the global market place since at least the 1960s. The Ulba Metallurgical Plant JSC company in 50 Kazakhstan has developed a growing beryllium- 0 product export business, based upon a former

Soviet facility at Ust Kamenogorsk. The Ulba USA Brazil Other Russia Japan operations rely on imported beryl concentrates, Norway imported at that time from mines in Russia. countries However, the inventory is finite and, when Figure 5.4 Global exports of beryllium by country depleted, Kazakhstan will need to develop an in 2009. The data are shown in terms of contained alternative source of beryl or beryllium-bearing beryllium metal in ores, concentrates and scrap. material. (Data from UN Comtrade, 2013.)

300

250

200

150 Tonnes 100

50

0 Figure 5.5 Global imports of beryllium USA Spain India Other by country in 2009. The data are shown France China Norway IcelandSweden Belgium Singapore in terms of contained beryllium metal countries in ores, concentrates and scrap. (Data from UN Comtrade, 2013.) (Inc Hong Kong) 106 david l. trueman and phillip sabey

Table 5.3 Published beryllium resources (as beryllium oxide, BeO). (After Sabey, 2006.)

Deposit(s) Location Resource, tons BeO Grade, % BeO

Various North Carolina, USA 122,800 0.05 Spor Mountain Utah, USA 72,315* 0.71 McCullough Butte Nevada, USA 47,000 0.027 Various Brazil 42,000 0.04 Strange Lake Canada 42,000 0.08 Aqshatau Kazakhstan 16,000 0.03−0.07 Thor Lake Canada 13,300 0.76 Various, Black Hills South Dakota, USA 13,300 na Sierra Blanca Texas, USA 11,300 >2.0 Lost River Alaska, USA >10,000 0.3−1.75 Yermakovskoye Russia >10,000 1.3 Seal Lake Canada 6800 0.35−0.40 Gold Hill Utah, USA >5000 0.5 Tanco Canada 1800 0.20 Boomer Colorado, USA <1000 2.0−11.2 Hellroaring Creek Canada <1000 0.10 Iron Mountain New Mexico, USA <1000 0.2−0.7 Mount Wheeler Nevada, USA <1000 0.75

*Remaining reserves 2004

World resources other minerals”. Selected beryllium-bearing minerals are listed in Table 5.4 according to their Beryllium (BeO) resources in selected countries BeO content. The most important minerals from are shown in Table 5.3. The estimates for Spor an economic standpoint are bertrandite and Mountain, Utah, USA are proven reserves that meet beryl. Other minerals which have been produced international reporting standards such as NI 43-101 intermittently or on a small scale in the past in Canada or the JORC code in Australia, but many include phenacite. Potential future production of the other estimates are dated and would not may come from the restart of former phenacite qualify under present standards. Significant mining operations in the Kalesay deposit in resources may also exist in China, France, var- Khazakstan. ious African and South American countries, and Most of the other minerals listed in Table 5.4 Afghanistan, but these have not been quantified. typically occur in minor or trace amounts in nature and are most commonly found in sub- alkaline granitic pegmatite suites. However, Mineralogy of beryllium some, such as bertrandite and phenacite, are found in both sub-alkaline and alkaline systems. Beryllium can behave as a lithophile, chalcophile The relationship between a large number of and siderophile element. As a result, it forms a beryllium-bearing minerals and the main beryl- variety of silicates, sulfates, carbonates, several lium-bearing deposit types is provided by Barton hydrated mineral species and a single known and Young (2002). The high-grade mineral behoite, oxide. According to Grew (2002) and Walsh and which is naturally occurring beryllium hydroxide, Vidal (2009), beryllium occurs as an “essential represents a potential work-environment concern ingredient in approximately 45 minerals” and as in some deposits, for example in the Sierra Blanca an “occasional constituent in approximately 50 in Texas. Beryllium 107

Table 5.4 Selected beryllium minerals and their Pegmatite deposits compositions. Those of current economic importance, bertrandite and beryl, are in bold. Pegmatites are defined as very coarse-grained, crystalline rocks commonly of granitic composi- Mineral Formula BeO content (wt %) tion. It is the large size of the individual crystals, Bromellite BeO 98.02 which may attain metre or larger scale in pegma-

Behoite Be(OH)2 58.13 tites, that allows the processing of relatively

Hambergite Be2(BO3)(OH) 53.5 coarse fragments, and/or hand sorting (‘cobbing’)

Phenacite Be2SiO4 45.5 of those minerals of economic interest. Bertrandite Be Si O (OH) 39.6–42.6 4 2 7 2 Granitic pegmatites may be divided into two Moraesite Be (PO )(OH)4H O 25–28 2 4 2 classes: those of the Lithium-Caesium-Tantalum Hurlbutite CaBe (PO )21.3 2 4 (LCT) association and those of the Niobium- Berylonite NaBe(PO ) 19–20 4 Yttrium-Fluorine (NYF) association (Cˇ erný, 1991a). Chrysoberyl BeAl2O4 19.8 Euclase BeAI(SiO )(OH) 17.0–21.8 For a detailed description of these classes, the 4 reader is referred to Chapter 15. The LCT pegma- Herderite CaBe(PO4)(OH,F) 15–16 tites generally occur in pegmatite ‘swarms’, Barylite BaBe2(Si2O7)16

Beryl Be3Al2Si6O18 11.0–14.3 whereas those of the NYF kindred tend to occur

Helvite Mn8(BeSiO4)6S2 11–14.2 within restricted areas, commonly proximal to

Danylite Fe8(BeSiO4)S2 12.7–14.7 their parental rocks. The LCT pegmatites are the Genthelvite Zn (BeSiO ) S 12.6 8 4 6 2 more important source of beryl. Eudidymite Na(BeSi O )(OH) 10.5–11.2 3 7 Pegmatites have long been exploited as sources Gadolinite (Y,Ca) Fe(BeSiO ) (O,OH) 5.5–12.9 2 4 2 2 of mica, feldspar, quartz, gemstock (such as chryso- beryl, aquamarine, morganite, hiddenite, kunzite and polychrome tourmalines) and the elements tin, tantalum, niobium, lithium, rubidium, caesium Beryllium deposits and beryllium (London, 2008). In the LCT system micas, feldspars and quartz are ubiquitous but the Commercial production of beryllium has been rare elements in pegmatites display a zonal pattern derived from two principal types of deposit: of element enrichment and mineral variation granitic pegmatite deposits and hydrothermal- resulting from increased fractionation with dis- metasomatic deposits. Historically, beryllium tance of the pegmatite from its parental source has been extracted from pegmatitic deposits of (Figure 5.7). No scale is given in Figure 5.7 as the all ages and distributed widely around the globe. field or swarm of pegmatites can be highly variable However, since 1969 most of the world’s in extent, ranging from tens of metres to several beryllium production has been derived from kilometres (Cˇ erný, 1991a). Beryllium, usually in one large low-grade hydrothermal deposit of the form of beryl, makes an early appearance in the Tertiary age located at Spor Mountain, Utah, zoned sequence close to the source granite intru- USA. The locations of the most important sion and continues to occur into the most highly beryllium deposits are shown in Figure 5.6, fractionated and distal members of the pegmatite divided into those hosted in pegmatites and field. However, with increasing fractionation the those of other types. It is important to note that elements sodium and caesium, for example, sub- pegmatites commonly occur in large numbers stitute for beryllium in the beryl lattice and lower (a ‘swarm’) within a particular area (pegmatite the quality of the beryl as a feedstock. ‘field’). Within any one field mineral extraction The Harney Peak Granite of the Black Hills, may be focused on a few large deposits or, more South Dakota, is an excellent example of such a commonly, small-scale mining is carried out at pegmatite field and was a primary source of beryl many sites. for the US industry until 1962 (Shearer et al., 1992). Figure 5.6 The global distribution of significant deposits of beryllium. Some of the symbols on the map represent a single important deposit or resource, while others represent a cluster of deposits in one area or region. Beryllium 109

Caesium, lithium, Distance and tantalum, beryllium fractionation increasing

Lithium, beryllium, tantalum > niobium

Tantalum < niobium, beryllium

Beryllium

Barren

Granite

Figure 5.7 Rare-element zoning in pegmatite fields. (After Trueman and Černý, 1982.)

The Arquehanna River pegmatites in the Aracuai metallurgical processes could change this in the area in Minas Gerais Province, Brazil, which have future. A novel and unique otation process was hosted intermittent beryl, tin, tantalum, lithium developed by Bulatovic (1988) that recovered and gemstock production, display a convergence both phenacite and bertrandite from the Thor and overlap of three such suites of pegmatites. Lake deposits in Canada. However, these deposits Individual pegmatite deposits vary in size from remain undeveloped. metres to more than a kilometre in length and Pegmatite mineral deposits are difficult to hundreds of metres in width. The largest, so-called evaluate for their contained beryllium resources. ‘giant’, pegmatites may display internal zoning The minerals of particular interest occur in a in a similar manner to the field or swarm in wide range of sizes, from microscopic to metre which they are located (Cˇ erný, 1991b). Examples scale, and are commonly widely dispersed, even include the Tanco pegmatite in Manitoba, Canada within internally zoned pegmatites. Beus (1962) (Cˇ erný, 2005 and Linnen et al., this volume), the provides comprehensive procedures for evalu- Greenbushes pegmatite in Western Australia ating pegmatitic beryllium deposits. These (Fetherston, 2004) and the Bikita pegmatite in include stripping, trenching, drilling and under- Zimbabwe (Cˇ erný et al., 2003). ground development. Channel sampling may No commercial beryllium production is need to be conducted at intervals as closely known from NYF pegmatites but increased spaced as two metres. This rigorous sampling beryllium demand and the availability of suitable process allowed Beus to categorise beryllium 110 david l. trueman and phillip sabey resources on the contemporary protocols of the precipitated the uorite and led to the deposition USSR, but would probably also be compliant of bertrandite. with resource-reporting standards used today. Mining and processing of beryllium Hydrothermal deposits Beryl ores This broad category includes deposits derived from hydrothermal processes in various geological set- The small size of most granitic pegmatite deposits tings, including replacement, skarn, greisen and makes them uneconomic to mine by mechanical vein deposits, several of which are described by methods, and so beryl extraction from pegmatites Barton and Young (2002). is carried out by a manual process of simulta- Hydrothermal deposits account for the world’s neous extraction, cleaning and concentration largest known resources of beryllium and include which in practice means that this process, known the Lost River Tin deposit in Alaska (Sainsbury, as ‘hand cobbing’, is confined to those parts of the 1963), the Spor Mountain beryllium deposits world where labour costs are low. The beryl ore in Utah (Davis, 1979), the Sierra Blanca deposit concentrates are usually produced in conjunction in Texas (Price et al., 1990) and Thor Lake in with the recovery of gemstones or other valuable Canada’s Northwest Territories (Trueman et al., minerals and by-products, such as columbo-tan- 1988). Resource estimates are available for a talite, feldspars, micas, and lithium minerals. number of these deposits, some of which have Attempts to use photometric sorting of beryl to BeO grades in the per cent range (Table 5.2). These achieve the minimum 11.0% BeO required by contrast markedly with grade estimates in beryl processors have been unsuccessful due to the pegmatite deposits which commonly contain masking caused by gangue minerals adhering to tens or hundreds of ppm beryllium. the beryl surface. The Lost River, Spor Mountain and Sierra Beryl can be concentrated by various novel o- Blanca deposits are related to felsic magmatism of tation processes (Vidal et al., 2009). However, latitic composition. The extrusive host rocks none of these are economically attractive for the (termed ‘ongonites’ by Kovalenko et al., 1971) and extraction of beryllium on a commercial scale their intrusive equivalents are enriched in uorine, because of the relatively high cost of hydrouoric beryllium and uranium, while Lost River is also and oleic acids, the principal reagents. The beryl well known for its tin deposits (Sainsbury, 1963). otation concentrates formerly produced and The high uorine content depresses the freezing offered commercially from China were found to point of these magmas and allows protracted contain an undesirable mix of beryllium minerals enrichment of incompatible elements, such as that proved difficult to process because of excess beryllium, into the late hydrothermal phase. When remaining oleic acid. the hydrothermal uid encountered carbonate rocks uorite was formed, while beryllium was Bertrandite ores deposited in the host volcanic rocks as bertrandite and phenacite and, in the Sierra Blanca deposit also Materion has been mining bertrandite ores at as the rare mineral behoite (Sabey, 2006). Spor Mountain, Utah since 1969 (Figure 5.8), The Spor Mountain deposits formed from a low- from eight open pits located along a complex temperature beryllium–uranium-uorine-enriched fault-segmented ore body. An earlier attempt by hydrothermal uid, circulating through Oligocene Anaconda Minerals to conduct underground min- base surge rhyolite deposits (Lindsey et al., 1973, ing of the bertrandite ore was abandoned due to and Burt et al., 2006). The rhyolite deposits lie poor ground conditions. unconformably on Palaeozoic dolomites which are Prior to mining, rigorous delineation of the thought to have been the source of calcium that bertrandite orebodies is undertaken and close Beryllium 111

Processing of beryl and bertrandite to beryllium hydroxide There are essentially three processes for treating beryllium minerals to produce beryllium chemi- cals which are the starting points for beryllium- bearing commercial products: ● The Materion Corporation uses the Kjellgren– Sawyer sulfate extraction method to treat beryl ores in a mixed owsheet, as described below. ● Walsh et al. (2009) reported that beryllium processing in the CIS and in China probably uti- lises the Kjellgren–Sawyer process as well. In the author’s (PS) experience, however, the CIS and China have used the alkaline ux, or Degussa Figure 5.8 Materion Brush open-pit mining at Spor fusing process, in which beryl ores are mixed with Mountain, Utah. (Courtesy of Materion Corp.) quicklime, heated to 1500 °C and fritted. The frit or glass is then taken into solution in sulfuric control of grade, or beryllium content, con- acid and beryllium hydroxide precipitated with tinues through stockpiling of ore before delivery ammonia. to the processing plant. Most assaying for grade ● The Copaux–Kawecki process, consists of control is carried out with beryllometers, which heating beryl with sodium silicouoride to pro- are hand-held, portable instruments that use a duce sodium beryllium uoride which is in turn radioactive gamma radiation source to displace aqueous leached to produce a beryllium solution an excess loosely attached neutron in the beryl- (Walsh et al., 2009). lium nucleus. The neutron ux events are Different processes are required to treat the counted by a scintillometer that provides a beryl and bertrandite which enter the Materion quantitative analysis of the beryllium content plant as hand-cobbed beryl concentrates and at that location. direct shipping ore, respectively (Figure 5.9). The shallow, westerly dipping orebodies at ● Beryl concentrates, after crushing, are heated Spor Mountain are first drilled on 30-m centres to 1700 °C and quenched rapidly in water to form and assayed at vertical intervals of 0.6 m. a frit or glass. The frit is heat treated at 1000 °C, Overburden, hanging wall rhyolites and alter- ground to finer than 200 mesh (75 μm), and then ation clays are removed and a one-metre layer of leached with a concentrated sulfuric acid solu- clay is left on top of the ore. The overburden and tion at 250 to 300 °C. This process extracts the waste are stockpiled for later restoration of the beryllium, forming a beryllium sulfate solution. mined area. The orebody is next drilled and ● Bertrandite ore is crushed and wet-milled to assayed on 7.5-m centres. The remaining clays yield a fine slurry, which is then leached with are removed and the orebody is mined by loaders a sulfuric acid solution at about 95 °C to extract and scrapers guided by in-pit assaying utilising the beryllium, forming a beryllium sulfate beryllometers. solution. Mined ore is gathered on a stockpile which is, After separating the solids from both streams in turn, drilled and assayed for grade control. of beryllium sulfate solutions, the two solutions There is no additional concentration process, and are combined. Solvent extraction is used to these ‘direct shipping’ ores, at a grade of about remove additional elements that were extracted 0.265% Be are carried by truck 80 km to the with the beryllium. During solvent extraction, Materion process plant, north of Delta, Utah. the beryllium sulfate solution contacts an organic 112 david l. trueman and phillip sabey

Beryl Bertrandite

Crushing and Crushing Wash water from thickening wet grinding

Melting Wet screening

Sulfuric acid Fritting Acid leaching steam

Sludge Heat treatment Thickening Beryl leach discard solution Raffinate discard Grinding Solvent extraction Sulfuric Converted organic acid Sulfation Stripped Strippingorganic Acid conversion

Sludge Dissolving Iron hydrolysis Steam discard Sludge Thickening discard Steam A-hydrolysis

Barren Filtration Beryllium filtrate carbonate Deionized water Repulping

Steam B-hydrolysis Figure 5.9 Beryllium Barren hydroxide production from Product Beryllium Filtration filtrate bertrandite and beryl. (Source: drumming hydroxide Materion Corp.)

solution of di-2-ethylhexylphosphoric acid in a leach solution. Heating the strip solution to kerosene-type organic solvent. Beryllium is selec- about 70 °C separates the iron and aluminum as tively dissolved in the organic solution. The slow hydroxide or carbonate precipitates. Two hydro- rate of extraction at room temperature is acceler- lysis steps at 95 °C and 165 °C, respectively, ated by warming the extractant and leach solu- remove the ammonia and carbon dioxide and tion. The loaded organic phase is treated with an result in the formation of beryllium hydroxide, aqueous ammonium carbonate solution and which is filtered and then drummed for shipment beryllium is stripped from the organic phase, to Materion’s plant in Elmore, Ohio for further forming tetraammonium beryllium tricarbonate processing to finished products containing beryl-

((NH4)4Be(CO3)3). lium. The efficiency of this processing route is In addition to beryllium, the solvent extrac- high, with approximately 80 per cent of the beryl- tion step dissolves iron and small amounts of lium content of both beryl and bertrandite recov- aluminum and uoride present in the sulfate ered (Walsh, 1979). Beryllium 113

quently melted with additional copper and some Production of metal and alloys critical alloying elements, such as nickel and from beryllium hydroxide cobalt, to produce a range of alloy compositions Beryllium hydroxide is the starting point for pro- containing between 0.3–2.0% Be. By varying the duction of beryllium metal and a number of composition, a controlled range of physical and alloys and ceramics that are produced widely mechanical properties can be obtained. Generally, across the United States of America and also at two alloy families are used: those containing 1.8– other facilities in Europe and Asia. The majority 2.0% Be, called ‘gold alloys’ due to their colour, of the beryllium hydroxide produced at Delta, which are used where strength is paramount; and Utah is processed at Materion’s operations in those containing 0.10–0.25% Be called ’red Elmore, Ohio. alloys’, which are used where electrical and Beryllium metal is produced using the thermal conductivity properties are critical. Schwenzfeir process (Figure 5.10). The hydroxide Before use, alloys are heated to fully dissolve is reacted with ammonium uoride to form all of the elemental additions, and to homogenise ammonium uoroberyllate ((NH4)2BeF4) and then and refine the size, shape and location of the heated to produce amorphous beryllium uoride beryllide precipitates, as well as soften the copper

(BeF2). The uoride is then reduced with magnesium matrix. This is termed the solution-annealed at temperatures between 900 °C and 1300 °C to pro- state. The intermediate shapes are subsequently duce beryllium ‘pebbles’ and beryllium uoride/ hot worked by such traditional processes as hot magnesium uoride slags. These are water leached rolling, extrusion or forging to a semi-finished and the remaining beryllium pebbles are melted size and shape. Further cold forming is carried under vacuum to remove entrained gas or trapped out by conventional metal working processes slags and then cast into ingots (Figure 5.11). Vacuum such as rolling to thinner gauge strip products melting is also a start point for recycling of beryl- ready for stamping to size/shape, or by cold forg- lium scrap (Stonehouse, 1985). ing or drawing to rod, wire or tube. The ingots are subsequently machined into Solution-annealed material can be hardened chips or cuttings that are ground and screened to by two complementary methods, work hardening produce a specific particle-size distribution of and heat treatment, which alone, or in combination, rounded particles. These are next compacted by provide a powerful control mechanism to obtain cold, hot or isostatic pressing into ‘blocks’, that a desired combination of properties. The heat- can be made to have specific combinations of prop- treatment process can be carried out by the erties designed for a variety of end uses (Figure 5.11). copper–beryllium manufacturer to deliver a prod- Many beryllium alloys are produced commer- uct termed ‘mill hardened’ that is ready to be cially, the most important of which are those of used by manufacturers to form the desired shape copper, nickel and aluminum. These alloys, without further processing. In other cases, it can which vary in their beryllium content from be more convenient for additional forming opera- 0.15% Be up to 62% Be, are produced in many tions, such as machining and stamping, to be car- forms including casting alloys, plate, rod, bar, ried out by a component manufacturer starting tube, strip and wire products. from an alloy delivered in the more ductile and Commercial copper–beryllium alloys are man- softer solution-annealed form. ufactured by first making a master alloy contain- ing 3.5–10% Be. The most widely used process Production of beryllium oxide utilises an electric arc furnace to produce a stoi- from beryllium hydroxide chiometric composition of 3.3–3.5% Be by adding calcined beryllium hydroxide to pure copper, Beryllium oxide is produced by dissolving beryl- together with a source of carbon, such as graphite, lium hydroxide in sulfuric acid to produce as a reducing agent. The master alloy is subse- hydrated beryllium sulfate. The sulfate is 114 david l. trueman and phillip sabey

Lime Beryllium hydroxide

Batch Recycled material makeup Ammonium beryllium fluoride Slurry (ABF) solution storage

Rotary vacuum Lime sludge filter

Sulfide Ammonium sulfide treater Ammonium bifluoride

Filter Sulfide sludge

Solution storage

Hydrofluoric acid Salt Evaporator

Ammonium bifluoride Centrifuge makeup

Ammonium fluoride Dryer holding tank ABF salt NH4F Fluoride Scrubbers furnaces Beryllium fluoride Reduction furnaces Magnesium

Water Cooling Crushing leaching

Beryllium pebbles

Figure 5.10 Production of beryllium metal (Source: Materion Corp.) Beryllium 115

Beryllium in space or because of their sensitive military pebbles Solids Scrap Chips nature, do not return at all. When these compo- separation nents do finally return at end-of-life, they can be easily recycled. The aerospace industry is the larg- est source of pure beryllium new scrap, as many of Charge Degreasing the end uses only have short production runs. makeup Because of its hardness, and the relative difficulty to cut it at high rotational surface speeds with con- Float Vacuum ventional tooling, beryllium is considered a diffi- melting sink Discard cult metal to machine extensively. Accordingly, it Vacuum- is machined to final form from near net shapes, cast ingot Sizing made by hot isostatically pressing or casting, resulting in significant amounts of chips (i.e. new Skin scrap). A good example is the machining of the ingot Hand back reinforcing web structure of the James Webb inspection telescope mirrors which resulted in 92 per cent of the original rough beryllium shapes being returned. Chipping Magnetic When beryllium metal is recycled, the Beryllium lathe separation Science and Technology Association (2012b) esti- mates a 70 per cent energy saving over the cost of producing newly won beryllium from ore. Sizing Compacting Overall, ‘new scrap’ generated in the course of inspection manufacturing beryllium-bearing products makes an important contribution to supply and may Grind, mill account for as much as 10 per cent of apparent classify consumption (Cunningham, 2004b). In Europe, a 45 per cent net recovery of new scrap copper– Beryllium beryllium alloys from component manufacturers powder back to alloy producers has been reported (European Commission, 2010). Figure 5.11 Final purification, grinding and sizing of When it can be isolated and recovered, copper– beryllium powders. This is also a starting point for beryllium or nickel–beryllium alloy end-of-life beryllium recycling. (Source: Materion Corp.) scrap is directly recycled to produce new alloys. A significant premium is paid by copper–beryllium precipitated, concentrated and calcined at 1430 °C manufacturers to stimulate the return of such to produce the oxide (Figure 5.12). The beryllium scrap in order to take advantage of the energy oxide, or beryllia, is supplied to the electronic conservation and sustainability advantages com- and electrical sectors as fired ceramic compo- pared to extraction from ore. nents for use in many different applications. It is generally not economic to recover beryl- lium metal from copper–beryllium alloys, such as those used in electronic components, because the Recycling beryllium content in each device is small and the beryllium content in the alloy is very low (<1.25%). The pure beryllium metal components used in Furthermore, the components are generally small technological applications have extremely long and difficult to separate. The United Nations lifetimes, and, therefore, return to the recycle Environment Programme (UNEP, 2011) notes that stream very slowly. Some, because of applications most components of electronic and electrical 116 david l. trueman and phillip sabey

Beryllium Sulfuric hydroxide acid

Pressure Storage Dissolver filter tank Condensate Condensate Pressure Evaporator tank filter

Mother Crystallisers liquor

Centrifuge

Beryllium sulfate tetrahydrate salt

Water vapour Calciners Screen SO2 and SO3 to scrubbers Figure 5.12 Flow sheet for production of Beryllium beryllia or beryllium oxide. (Source: oxide powder Materion Corp.) devices are recycled for their copper value alone sections and hence higher weight. Aluminium and end up in a copper smelter where the beryl- nitride or boron nitride may be substituted for lium is captured in the slag and effectively lost. As beryllium oxide in some applications (Jaskula, a result, the recycled beryllium content of old 2012a), but usually provide inferior thermal scrap is low and the end-of-life recycling rate very management. low, at less than one per cent (UNEP, 2011).

Environmental aspects Substitution A disease of the lungs, called chronic beryllium On account of its high cost beryllium is com- disease (CBD) or berylliosis, has been known in monly used only where its specific properties are Europe and the USA since the 1930s and 1940s. crucial and consequently substitution is difficult Most reports were derived from the nuclear in many applications (Sabey, 2006). However, for weapons industry and from the manufacture of some non-critical purposes certain composite uorescent lamps containing beryllium-bearing materials, high-strength grades of aluminium, phosphors. However, the health effects are now pyrolytic graphite, silicon carbide, steel, or known to depend on numerous factors including titanium for example, may be substituted in the nature of the beryllium-bearing material place of beryllium metal or beryllium composites (metal, alloy, ceramic, salt, etc.), the solubility (Jaskula, 2012a). Many different copper alloys, and the level of exposure. According to the such as phosphor bronzes, may be substituted for Beryllium Science & Technology Association, copper–beryllium alloys but these invariably “Beryllium metal, copper–beryllium alloys result in some loss of performance, and energy (CuBe), aluminium–beryllium alloys (AlBe) and loss due to the requirement to use thicker nickel–beryllium alloys (NiBe), in solid form and Beryllium 117 as contained in finished products, present no spe- There has been considerable debate as to cial health risks.” Similarly, it has been shown whether beryllium should be regarded as carcino- that there is relatively little risk of CBD in the genic to humans (Fulton and Goldberg, 2009). In mining and processing of beryllium ores (Deubner the USA the National Toxicology Program (2002) et al., 2011). lists beryllium and certain beryllium compounds In order to contract CBD, an individual must as substances reasonably anticipated to be carcin- be exposed to airborne beryllium in the form of a ogens, while the International Agency for dust, mist or fume and become sensitised to Research on Cancer (2012) classified beryllium beryllium. This sensitisation is an immunolog- and beryllium compounds as carcinogenic to ical lymphocyte proliferation response in some humans. The United States Environmental individuals (Donovan et al., 2007). Only those Protection Agency classifies inhaled beryllium as persons who are genetically susceptible can a probable human carcinogen (USEPA, 1998). become sensitised to beryllium (Wang et al., 1999 However, the latest genotoxicity test results, and 2001). Before the late 1980s, workers were together with expert reviews which highlighted diagnosed with CBD only when they exhibited inherent aws in the epidemiological assess- clinical (observable) symptoms of CBD and ments used in previous studies, have indicated changes in their chest X-ray or lung function test. that beryllium metal is unlikely to be carcino- During the late 1980s and early 1990s, the cri- genic to humans (Hollins et al., 2009; Strupp, teria by which CBD was diagnosed changed, and 2011a and 2011b; Rothman and Mosquin, 2011; workers began to be diagnosed with CBD without and Boffetta et al., 2012). clinical symptoms or measurable impairment. Beryllium is typically present at concentra- This diagnosis became possible as a result of the tions of 0.5 to 2 ppm in soils and rocks throughout application of new technology in medical testing the world. Consequently it is commonly found at and evaluation. Workers diagnosed with CBD in low levels in natural products like coal, wood, the absence of X-ray or lung function changes or vegetables, foodstuffs and gemstones (ATSDR, symptoms of disease are referred to as having 2002). Beryllium intake from the air and dust can sub-clinical CBD, meaning that they have no be increased by 2–3 orders of magnitude in the clinical symptoms or measurable impairment. vicinity of a point source, such as a coal-fired Workers with subclinical CBD may never develop power plant. Beryllium is found in tobacco and clinical CBD or may develop clinical CBD over therefore tobacco smoke is a potential source of time (Borak et al., 2006). exposure to beryllium in the general population. For situations where risk of inhalation In plants and vegetables beryllium is found at low exposure existed, an airborne standard was devel- (ppb) levels that pose no risk to human health. oped in the USA in the late 1940s that was gener- Beryllium has been measured (fresh weight) in ally adopted internationally and is still used rice at 72 μg/kg, lettuce at 16 μg/kg, kidney beans today. This standard is two micrograms beryl- at 2200 μg/kg, peas at 109 μg/kg and potatoes at lium per cubic metre of air (2 μg/m3), averaged 0.59 μg/kg (US Department of Health & Human over an eight-hour work period. Companies pro- Services, 2002). ducing beryllium have more recently imple- It is estimated that within the United States mented a preventative model that incorporates about 45 per cent of airborne beryllium is due to an eight-hour exposure guideline of 0.2 μg/m3. anthropogenic releases (US Department of Health & Scientific research and actual workplace experi- Human Services, 2002). Natural sources, such as ence in beryllium production facilities in the windblown dust and volcanic activity, account USA indicate that the use of engineering and for the remainder of the beryllium released to the work practice controls have been effective in pre- atmosphere. Electric utilities account for about venting any health effects including sub-clinical 80 per cent of the anthropogenic emissions, with CBD and clinical CBD. industry and metal mining contributing the rest. 118 david l. trueman and phillip sabey

250

200

150

100

US$ per pound Figure 5.13 Annual average value of copper–beryllium alloy, per pound of 50 contained beryllium, 1996–2011. (Compiled from USGS annual Mineral 0 Commodity Summaries for beryllium 1996 1998 2000 2002 2004 2006 2008 2010 from 1996–2012; and Cunningham, 1999.)

Beryllium released during the erosion of rocks Long-term contracts for the supply of beryl and soils or derived originally from airborne concentrates normally specify BeO content and emissions generally forms insoluble compounds sieve specifications. A premium is often paid and complexes: the mobility of these is further for BeO contents exceeding 11% as an incen- limited by adsorption onto organic matter, iron– tive for suppliers to remove excess gangue manganese oxyhydroxides and clay minerals. minerals, usually quartz and feldspar, which Therefore, under most environmental conditions consume excess acid during mineral-processing beryllium concentrations in groundwater and stages. surface water are very low. The USGS calculates the value of a pound of In air, beryllium compounds are present mostly beryllium contained in copper–beryllium as fine dust particles, which ultimately settle over alloys, the most widely used beryllium product the land and water surfaces of the globe. Fish do (Figure 5.13). The significant increase in prices not accumulate beryllium from water into their over the period 2005–11 was driven by several bodies to any great extent. Concentrations of key mining and refining cost factors. These beryllium in drinking water range from 0.010 to include increases in the price of refining pro- 1.22 μg/l with an average of 0.19 μg/l (Kolanz, cess chemicals, especially sulfuric acid and 2001). According to the World Health Organisation ammonia, the increase in the price of diesel (2009), “Beryllium is rarely, if ever, found in fuel used in mining, depletion of resources that drinking-water at concentrations of concern. had low overburden removal requirements and Therefore, it is not considered necessary to set a the lower beryllium contents in the newly formal guideline value. A health-based value for worked deposits. beryllium in drinking-water would be 12 μg/l.”

Outlook Prices Since the early 1990s, much of the world’s beryl- As a result of its relatively small market, beryl is lium raw material supply has been derived from a not traded on international commodity exchanges. single source in North America. China has devel- Instead, prices are negotiated between buyers and oped substantial production capacity for the sellers and the few posted prices for beryl contain- production of beryllium hydroxide and copper– ing an average of 11% BeO have remained beryllium hydroxide from imported beryl and unchanged for at least 35 years at US $1600 per domestic ores. A new, high-purity beryllium metric tonne. metal production facility was brought on stream Beryllium 119 in the USA in 2011 to replace the plant closed by upon the real need for the properties of the mate- Materion in 2000. In addition to providing new rials that cannot be met by alternatives without beryllium metal, this plant will also serve as an loss of function. The major commercial producers effective beryllium metal recycling facility. The of beryllium have seen steady growth, inuenced mines at Yermakovskoye, in the Russian State of by global economic cycles, but trending to a 3–5 Buryatia, which were exploited during the Soviet per cent compounded annual growth rate. era, are reported to be undergoing renovation with Although one US company is the current plans to re-open in 2017 (MBC Resources, 2011). market leader in the beryllium business, the Copper–beryllium alloys are widely used to resources of beryllium (Table 5.3), while not closely provide an unmatched physical property set to defined, are large and should be considered more electronic and electrical components. This has led than adequate to supply the small annual tonnages to an increasing demand for beryllium in both new needed in the foreseeable future. It is anticipated consumer applications and in green technologies, that further installed production capabilities will such as wind power generation and hybrid or pure emerge, notably in Russia, Kazakhstan and China, electric vehicles. The relevant physical property set to respond to future market demands. Consumers for application in power generation is the unique of beryllium metal, alloys and ceramic would combination of high strength, high conductivity welcome alternative sources both for security of and high resistance to loss of yield strength at ele- supply and for competitive market pricing. vated temperatures. In hybrid and pure electric vehicles, those properties provide a high current carrying capacity in applications such as battery- Note cell connectors and electric motor components. The same properties promote the use of the alloy in 1. Materion Corporation includes: Materion Brush connector terminals used for solar power systems, Beryllium & Composites, Materion Brush Performance where temperatures in roof-top or desert locations Alloys, Materion Ceramics, Materion Natural Resources. can exceed 75 °C for extended periods of time. Ongoing research has demonstrated that by mixing uranium oxide with beryllium oxide, References greater fuel rod utilisation efficiencies may be achieved in nuclear power generation as a result Agency for Toxic Substances and Disease Registry (2002) of the improved thermal conductivity of the fuel Toxicological Profile for Beryllium. ATSDR, Atlanta. rods (McDeavitt et al., 2011). Nuclear power gen- Barton, M.D. and Young, S. (2002) Non-pegmatitic eration can be expected to grow and beryllium Deposits of Beryllium: Mineralogy, Geology, Phase will continue to play a conventional role as clad- Equilibria and Origin. In: Grew, E.S. (ed.) Beryllium: ding and moderators in these systems. Mineralogy, Petrology and Geochemistry, Reviews in As a result of concerns over possible health Mineralogy and Geochemistry, v.50. issues related to the use of beryllium and the Beryllium Science and Technology Association (2012a) requirement for dust control in its handling, new Production statistics. http://beryllium.eu/about- applications for beryllium might have been beryllium-and-beryllium-alloys/facts-and-figures/ production-statistics/ expected to be taken up only slowly. In fact, the Beryllium Science and Technology Association (2012b) number of new uses for beryllium has increased, Recycling of beryllium. http://beryllium.eu/about- in parallel with a virtually equivalent reduction beryllium-and-beryllium-alloys/facts-and-figures/ in the size of components due to miniaturisation, recycling-of-beryllium/ leading to a relatively at demand for existing Beus, A.A. (1962) Beryllium: Evaluation of Deposits applications, but a steady increase in the number During Prospecting and Exploratory Work. W.H. of new applications. It can be assumed that any Freeman and Co., London. substitution for beryllium-containing materials Boffetta, P., Fryzek, J. and Mandel, J. (2012) Occupational has already occurred, and future demand is based exposure to beryllium and cancer risk: a review of the 120 david l. trueman and phillip sabey

epidemiologic evidence. Critical Reviews in beryllium blood lymphocyte proliferation test based Toxicology 42 (2), 107–118. on a long-term occupational surveillance program. Borak, J., Woof, S.H. and Fields, C.A. (2006) Use of International Archive of Occupational Environmental Beryllium Lymphocyte Proliferation Testing for Health 81, 165–178. Screening of Asymptomatic Individuals: An Evidence- Dunai, T.J. (2010) Cosmogenic Nuclides, Principles, Based Assessment. Journal of Occupational and Concepts and Applications in the Earth Surface Environmental Medicine 48 (9), 937–947. Sciences, Cambridge University Press, Cambridge. British Geological Survey (2012) World mineral produc- European Commission (2010) Critical Raw Materials for tion 2006–10. British Geological Survey Keyworth, the EU. Annex V to the Report of the Ad-hoc Working Nottingham. Group on defining critical raw materials. http://ec. Bulatovic, S. (1988) United States Patent 4,735,710. europa.eu/enterprise/policies/raw-materials/files/ Burt, D.M., Wohletz, K.H. and Sheridan, M.F. (2006) docs/report_en.pdf “Water-laid Tuff” of the Utah Desert and Similar Fetherston, J.M. (2004) Tantalum resources in Western Surge Deposit Misinterpretations: A Possible Lesson Australia. Geological Survey of Western Australia, for Mars. Lunar & Planetary Science Conference, Mineral Resources Bulletin 22. XXXVII, No. 2295. Fulton, G. and Goldberg, A. (2009) Medical Aspects of Cˇ erný, P. (1991a) Rare-element Granitic Pegmatites. the Toxicity of Beryllium and Beryllium Alloys. In: Part II: Regional to Global Environments and Walsh, K.A. (ed.) Beryllium Chemistry and Processing, Petrogenesis, Geoscience Canada, 18 (2). ASM International, Materials Park, Ohio. Cˇ erný, P. (1991b): Rare-element Granitic Pegmatites. Grew, E.S. (2002) Mineralogy, Petrology and Part I: Anatomy and Internal Evolution of Pegmatite Geochemistry of Beryllium: An introduction and list Deposits, Geoscience Canada, 18 (2). of beryllium minerals. In: Grew, E.S. (ed.) Beryllium: Cˇ erný, P. (2005) The Tanco rare-element pegmatite Mineralogy, Petrology and Geochemistry, Reviews in deposit, Manitoba: Regional context, internal Mineralogy and Geochemistry. Volume 50. anatomy, and global comparisons. In: Linnen, R.L. Harrel, J.A. (2004) Archaeological Geology of the World’s and Samson, I.M. (eds.) Rare element geochemistry First Emerald Mine, Geoscience Canada, 31 (2). and mineral deposits. Geological Association of Hollins, D. M., McKinley, M.A., Williams, C. et al. (2009) Canada Short Course Notes, 17, 127–158. Beryllium and lung cancer: A weight of evidence eval- Cˇ erný, P., Anderson, A.J., Tomascak, P.B. and Chapman, uation of the toxicological and epidemiological litera- R. (2003): Geochemical and morphological features of ture. Critical Reviews in Toxicology 39 (S1), 1–32. beryl from the Bikita granitic pegmatite, Zimbabwe. International Agency for Research on Cancer (2012). The Canadian Mineralogist 41, 1003–1011. IARC Monographs on the Evaluation of Carcinogenic Cunningham, L.D. (1999) Metals prices in the United Risks to Humans. Volume 100C, A Review of Human States through 1998: Beryllium. United States Carcinogens: Arsenic, Metals, Fibres, and Dusts. Geological Survey. Jaskula, B.W. (2012a) Beryllium, USGS Mineral Cunningham, L.D. (2004a) Beryllium, United States Commodity Summaries. Geological Survey Mineral Commodity Summaries. Jaskula, B.W. (2012b) USGS 2010 Minerals Yearbook: Cunningham, L D. (2004b) Beryllium recycling in the Beryllium. United States in 2000. United States Geological Klaproth, M.H. (1801) Analytical Essays Towards Survey Circular 1196-P. Promoting The Chemical Knowledge Of Mineral Davis, L.J. (1979) Beryllium and Uranium Mineralization Substances, G. Woodfall, London. in the Spor Mountain Area, Juab County, Utah, In: Kolanz, M.E. (2001) Introduction to beryllium: uses, Basin and Range Symposium and Great Basin Field regulatory history and disease. Applied Occupational Conference, Rocky Mountain Association of Geologists Environmental Hygiene, 16, 559–567. and Utah Geological Association. Kovalenko, V.I., Kuz’min, M.I., Antipin, V.S. and Petrov, Deubner, D.C., Sabey, P., Huang, W. et al. (2011) L.L. (1971) Topaz-bearing quartz keratophyre Solubility and Chemistry of Materials Encountered (ongonite), a new variety of sub-volcanic igneous vein by Beryllium Mine and Extraction Workers: Relation rock. Acad. Sci. USSR Doklady Earth Sci. Sec 199. to Risk. Journal of Occupational and Environmental Lindsey, D.A., Ganow, H. and Mountjoy, W. (1973) Medicine 09/2011. Hydrothermal Alteration Associated with Beryllium Donovan E.P., Kolanz, M.E., Galbraith, D.A., Chapman, Deposits at Spor Mountain, Utah, USGS Professional P.S. and Paustenbach, D.J. (2007) Performance of the Paper 818-A. Beryllium 121

Linnen, R.L., Trueman, D.L. and Burt, R. O. (2013) Strupp, C. (2011b) Beryllium Metal II. A Review of the Niobium and tantalum. In: Gunn, Gus. (ed.) Critical Available Toxicity Data. The Annals of Occupational Metals Handbook (Wiley-Blackwell). Hygiene 55 (1), 43–56. London, D. (2008) Pegmatites. The Canadian Mineralogist, Trueman, D.L. and Černý, P. (1982) Exploration for Special Publication 10. rare-element granitic pegmatites. In: Černý, P. (ed.) MBC Resources (2011) (Division of Metropol Group, Granitic Pegmatites in Science and Industry: A short Moscow). News release 29 August 2011. Available from: course. Volume 8. Mineralogical Association of http://www.steelguru.com/russian_news/Berillium_ Canada. Hydrometallurgical_Plant_to_be_built_in_Buryatia/ Trueman, D.L., Pedersen, J.C., de St. Jorre, L. and Smith, 222584.html D.G.W. (1988) The Thor Lake rare-metal deposits, McDeavitt, S.M., Ragusa, J., Revankar, S.T., Solomon, Northwest Territories. In: Recent Advances in the A.A. and Malone, J. (2011) A High-conductivity Oxide Geology of Granite-Related Mineral Deposits, Fuel Concept, Nuclear Energy International, Globe Canadian Institute of Mining and Metallurgy Special Trade Media. http://www.neimagazine.com/story.asp Vol. 39. ?sectioncode=76andstoryCode=2060405 UN Comtrade (2013) United Nations Commodity Trade Merchant Research and Consulting Limited (2012) Statistics Database, Department of Economic and Beryllium Market Review. http://marketpublishers. Social Affairs/ Statistics Division, http://comtrade. com/report/industry/metallurgy/beryllium_market_ un.org/db/ review.html UNEP (2011) Recycling rates of metals – a status report. National Toxicology Program (2002) 10th Annual Report of the Working Group on Global Metal Flows Report on Carcinogens – December 2002. to the International Resource Panel. Price, J.G., Rubin, J.N., Henry, C.D., Pinckston, T.L., U.S. Department of Health and Human Services (2002) Tweedy, S.W. and Koppenaal, D.W. (1990) Rare-metal Agency for Toxic Substances and Disease Registry, enriched peraluminous rhyolites in a continental arc, Toxicological Profile for Beryllium. Sierra Blanca area, Trans-Pecos Texas; Chemical U.S. Environmental Protection Agency (1998). modification by vapor-phase crystallization. GSA Toxicological Review of Beryllium and Compounds. Special Paper 246. Vidal, E.E., Foley, J. and Olson, D.L. (2009) Mineral Rothman, K.J. and Mosquin, P.L. (2011) Confounding after Processing. In: Walsh, K.A. (ed.) Beryllium Chemistry risk-set sampling in the beryllium study of Sanderson and Processing, ASM International, Materials Park, et al. Annals of Epidemiology 21 (10), 773–779. Ohio. Rudnick, R.L. and Gao, S. (2004) Composition of the Walsh, K.A. (1979) Extraction. In: Floyd D.R. and Lowe Continental Crust. In: Treatise on Geochemistry, J.N. (eds.) Beryllium Science and Technology, Plenum Volume 3 The Crust (Elsevier Pergamon), 1–64. Press, New York. Sabey, P. (2006) Beryllium Minerals. In: Kogel, J.E., Walsh, K.A. and Vidal, E.E. (2009) Sources of Trivedi, N.C., Barker, J.M. and Krukowski, S.T. (eds.) Beryllium. In: Walsh, K.A. (ed.) Beryllium Chemistry Industrial Minerals and Rocks, Society for Mining, and Processing, ASM International, Materials Park, Metallurgy and Exploration. Ohio. Sainsbury, C.L. (1963) Beryllium Deposits of the Western Walsh, K.A., Vidal, E.E. and Mishra, B. (2009) Extractive Seward Peninsula, Alaska, USGS Circular 479. Metallurgy. In: Walsh, K.A. (ed.) Beryllium Chemistry Shearer, C.K., Papike, J.J. and Jolliff, B.L. (1992) and Processing, ASM International, Materials Park, Petrogenetic links among granites and pegmatites Ohio. in the Harney Peak rare-element granite-pegmatite Wang Z., White, P.S., Petrovi, M. et al. (1999) system, Black Hills, South Dakota. Canadian Differential susceptibilities to chronic beryllium Mineralogist 30, 785–809. disease contributed by different Gly69 HLA DPB1 Stonehouse, A.J. (1985) Physics and chemistry of beryl- and -DPA1 alleles. The Journal of Immunology 163, lium. (February, 2012). http://materion.com/~/media/ 1647–1653. Files/PDFs/Beryllium/BerylliumMaterials-MB/ Wang, Z., Farris, G.M., Newman, L.S.et al. (2001) MB-003PhysicsandChemistryofBeryllium.pdf Beryllium sensitivity is linked to HLA-DP genotype. Strupp, C. (2011a) Beryllium Metal I. Experimental Toxicology 165, 27–38. Results on Acute Oral Toxicity, Local Skin and Eye World Health Organisation (2009) Beryllium in Drinking- Effects, and Genotoxicity. The Annals of Occupational water Background document for development of WHO Hygiene 55 (1), 30–42. Guidelines for Drinking-water Quality. 6. Cobalt

STEPHEN ROBERTS1 AND GUS GUNN2

1 School of Ocean and Earth Science, National Oceanography Centre, University of Southampton, UK 2 British Geological Survey, Keyworth, Nottingham, UK

Introduction Distribution and abundance in the Earth

Cobalt has been utilised by society since the Estimates of the crustal abundance of cobalt vary Bronze Age, mainly to impart a rich blue colour to between 15–30 ppm, not dissimilar to the other glass, glazes and ceramics. However, it was only first-period transition metals such as scandium, isolated as a pure metal in 1735 by Swedish chem- copper, zinc and nickel. In particular, cobalt is ist Georg Brandt and demand for cobalt remained most abundant in ultramafic rocks with an subdued until the turn of the 20th century and the average concentration of about 110 ppm. Recent development of cobalt–chromium alloys. In estimates of the upper continental crustal abun- particular, the demand for cobalt increased con- dance of cobalt suggest a concentration of 15 siderably after the Second World War, driven by ±1 ppm (Hu and Gao, 2008). The concentration of the use of high-purity cobalt in jet engines and gas cobalt in sea water is very low (generally less than turbines. Cobalt demand has further accelerated 10 ppt) which in part reflects its short residence in the past 30 years, reflecting the increased use of time1 of 340 years. In contrast, the residence cobalt as an essential constituent of materials times of other base metals are much longer, used in high-technology industries including (nickel, 6000 years; copper, 5000 years; and zinc, rechargeable batteries, superalloys and catalysts. 50,000 years) and their concentrations in sea water are much greater (Broecker and Peng, 1982).

Physical and chemical properties Mineralogy Cobalt (chemical symbol, Co) is a d-block transition metal, silver in colour, with an atomic number of Pure cobalt is not found in nature, but, as a result 27, appearing in the first long period of the Periodic of its chalcophile and siderophile properties, it Table between iron and nickel. Cobalt has two preferentially bonds with iron, nickel, copper and main oxidation states (2+ and 3+) and one naturally sulfur rather than with oxygen into a number of occurring isotope (59Co). Cobalt shows siderophile sulfide and sulfarsenide phases. In particular, it and chalcophile tendencies, has a high melting forms cobalt sulfides and arsenides, such as point of 1493 °C and is ferromagnetic (Table 6.1). cobaltite (Co,Fe)AsS, carrollite (CuCo2S4),

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Cobalt 123

Table 6.1 Selected properties of cobalt. Table 6.2 The most important cobalt-bearing minerals.

Property Value Units Mineral name Chemical formula Cobalt content (wt %)

Symbol Co Cobaltpentlandite Co9S8 67.40 Atomic number 27 Heterogenite-2H Co3+O(OH) 64.10 Atomic weight 58.93 Heterogenite-3R Co3+O(OH) 64.10 3 2+ 3+ Density at 25°C 8800 kg/m Linnaeite Co Co 2S4 57.95

Melting point 1493 °C Sphaerocobaltite CoCO3 49.55

Boiling point 3100 °C Cattierite CoS2 47.89 Electrical resistivity at 25°C 5.81 µΩ m Cobaltite CoAsS 35.52

Specific heat capacity at 25°C 0.42 J/(g °C) Erythrite Co3(AsO4)2.8(H2O) 29.53

Latent heat of vaporisation 6490 J/g Carrollite Cu(Co,Ni)2S4 28.56 Latent heat of fusion 263 J/g Glaucodot (Co,Fe)AsS 26.76

Hardness (Mohs scale) 5.0 Safforite (Co,Fe)As2 21.25 Thermal conductivity 100 W/(m °C) Willyamite (Co,Ni)SbS 20.78

Skutterudite (Co,Ni)As3-x 17.95

Kolwezite (Cu,Co)2(CO3)(OH)2 17.84

Siegenite (Ni,Co)3S4 14.51

content is rare, with the majority of cobalt production achieved through recovery as a by-product of copper and nickel mining in three principal geological settings: hydrothermal, mag- matic and lateritic (Figure 6.2). A fourth, significant but as yet unexploited, cobalt resource resides in iron–manganese nodules and crusts developed on the ocean floor which may contain substantial quantities of base metals including cobalt. In the following sections the geology Figure 6.1 Skutterudite, a cobalt–nickel arsenide, from which accounts for the majority of the world’s Bou Azer mine, Morocco. Maximum dimension of cobalt production is described, with a particular specimen is 7 cm. (BGS © NERC.) focus on the major cobalt-producing regions of the Democratic Republic of Congo (DRC) and linnaeite (Co,Ni)3S4 and skutterudite (Co,Fe,Ni) Zambia, where cobalt is a by-product of copper

As2 (Figure 6.1), which are commonly associated mining, and in Canada, Russia and Australia with the iron sulfides, pyrite, arsenopyrite and where cobalt is produced mainly as a by-product pyrrhotite. However, it also occurs as a carbonate of nickel mining. phase, sphaerocobaltite (CoCO3), and a hydroxide, heterogenite (CoO(OH)) (Table 6.2). Hydrothermal deposits These are cobalt ore deposits derived from hydro- Deposit types thermal fluids which have interacted with a variety of mafic and/or ultramafic basement Despite its low crustal abundance cobalt is rocks or are substantially derived from within concentrated by various geological processes to sedimentary basins. These include the deposits at concentrations suitable for mining. However, Bou Azer in Morocco which are the only cur- mining of metallic ores chiefly for their cobalt rently working mines that produce cobalt as a Figure 6.2 Major cobalt-producing mines and districts. The majority of new production is anticipated to be from laterite deposits. Cobalt 125 primary product. The major producing deposits ratios. Deposition of ore minerals occurred in of the DRC and Zambia are also classified as response to increasing pH, by mixing between hydrothermal in origin. magmatic brines and meteoric water. The pre- dominance of cobalt over nickel arsenide min- Bou Azer, Morocco erals in the Bou Azer mineralisation may be Located 320 km to the east of Agadir, more than attributable to the different solubilities of nickel 60 orebodies comprise the cobalt–nickel–arsenic– and cobalt in the hydrothermal system (Ahmed gold–silver mines of Bou Azer and adjacent areas. et al., 2009). These mines produced about 1800 tonnes of cobalt in 2011 (Cobalt Development Institute, 2012a), with major by-products including nickel, Idaho Cobalt Belt gold and arsenic. These deposits are spatially The Idaho Cobalt Belt (ICB) in north-western associated with serpentinised ultramafic rocks United States is a north-west-trending zone of of a Neoproterozoic ophiolite complex, which cobalt occurrences, approximately 64 km long comprises a mantle sequence of serpentinised and 10 km wide, centred on the former Blackbird peridotites, ultrabasic and basic cumulates, mine. Past production from the Blackbird mine stocks of quartz diorite, basic lavas, and a mixed complex amounts to approximately 1.7 million volcanic and sedimentary sequence (Leblanc and tonnes of high-sulfide cobalt–copper ore, with Kroener, 1981). the district as a whole potentially hosting 50 The Bou Azer cobalt mineralisation is domi- million tonnes of ore (Bookstrom et al., 2007). nated by arsenides, sulfarsenides and sulfides in a The ICB hosts cobalt–copper deposits in quartz-carbonate gangue. Cobalt-bearing arsenide sequences of siliciclastic metasedimentary rocks minerals include skutterudite (CoAs3), safflorite of Mesopro terozoic age deposited within intrac-

(CoAs2), loellingite (FeAs2), nickeline (NiAs), ontinental extensional basins. Styles of miner- rammelsbergite (NiAs2), and sulfur-rich nickel– alisation vary from stratabound, which is the cobalt diarsenide, with accessory copper sulfides, most important, to locally discordant sulfide molybdenite and gold. Cobalt-rich ores predomi- lenses, and discordant tourmaline breccias (Nash nate in almost all the deposits, which contrasts and Hahn, 1989; Bending and Scales, 2001; with the high Ni/Co ratio of the underlying ultra- Bookstrom et al., 2007). Principal ore minerals mafic rocks. are cobaltite and chalcopyrite which occur in a The cobalt mineralisation is located within 5 gangue comprising mostly quartz and biotite to 20 m thick and 50 to 600 m long veins, lenses (Nash and Hahn, 1989). Other ore minerals and stockworks and is structurally controlled include pyrite, pyrrhotite, arsenopyrite, glau- within shear zones. The shear zones are concen- codot ((Co0.5Fe0.5)AsS), safflorite (Co,Fe,Ni)As2), trated along irregular contacts between serpenti- bismuthinite, native bismuth, bismuth tellu- nites and Precambrian volcanic rocks or quartz rides and gold. diorites (Leblanc and Billaud, 1982; Leblanc and The deposits show a variety of deformational Fischer, 1990). The timing of the Bou Azer miner- features and are often localised along faults, alisation remains controversial. However, the shear zones and fold axes. However, the origin of present consensus is that the Bou Azer arsenide these deposits remains controversial with the mineralisation postdates the obduction of the stratabound cobalt–copper–gold sulfides consid- Neoproterozoic ophiolite fragments (Leblanc and ered to have formed either as synsedimentary Billaud, 1982; Oberthur et al., 2009). deposits prior to deformation and regional meta- The cobalt (nickel–iron–arsenic) mineralisa- morphism or during syntectonic metamorphism. tion at Bou Azer developed through the leaching It has also been suggested that they may be a var- of serpentinites by magmatic fluids under moder- iant of iron-oxide–copper–gold (IOCG) deposits ately reducing conditions and at high fluid/rock (Slack, 2006). 126 stephen roberts and gs gnn

N Democratic Republic of Congo Musonoi/T17/Kananga Tenke-Fungurume

Mukondo Masamba/KOV/ Kamoto (KTO) Tilwezembe Kolwezi Kambove 11˚S Mutoshi 11˚S Likasi Zambia Luiswishi Etoile Ruashi Lubumbashi

0 100km Solwezi Nchanga Chingola Mufulira Cobalt-rich copper deposits Chambishi Nkana Other copper deposits Zambia Kitwe Towns Kundelungu and younger Ndola Roan Group Pre-Katangan Basement 26˚E 28˚E

Figure 6.3 Simplified geology of the Central African Copperbelt in the Democratic Republic of Congo (DRC) and Zambia showing the location of selected cobalt–copper and copper mines and deposits. (Adapted from Trans Continent Exploration and Mining Company, 2011.)

rift (Kampunzu et al., 2000; Cailteaux et al., 2005). Central African Copperbelt Closure of the Katangan basin during the Lufilian The Neoproterozoic Katangan Copperbelt of Orogeny resulted in north-verging folds, thrusts Central Africa, located on both sides of the border and nappes of the ‘Lufilian Arc’. The Zambian between north-western Zambia and the Katanga Basin appears to have had a protracted history Province of southern Democratic Republic of with the onset of sedimentation at about 877 Ma2, Congo (DRC), hosts one of the world’s greatest eclogite formation at about 600 Ma and final clo- concentrations of copper and cobalt (Figure 6.3). sure and uplift of the basin at about 530 Ma (John It currently produces about two thirds of the et al., 2003). world’s cobalt with a production of 70,000 tonnes In Zambia the copper–cobalt deposits are in 2010 (BGS, 2012). hosted in para-autochthonous siliciclastic rocks The copper–cobalt ores are hosted by siliclastic close to basement terrains, whereas in the DRC and carbonate sedimentary rocks and volcanic and the deposits and their host rocks define thrust plutonic mafic rocks of the Katangansupracrustal sheets and nappes formed during the Lufilian sedimentary succession, emplaced in a continental Orogeny and the dominant lithological units are Cobalt 127 dolomites and dolomitic shales (Cailteaux et al., reactivated during basin inversion (McGowan 2005). The majority of the copper–cobalt miner- et al., 2003, 2006) or the development of mineral- alisation was deposited prior to the Lufilian ized shear zones within the basement (Bernau compressional tectonics both in the DRC and et al., 2013). There is also evidence that struc- Zambia, evidenced by the folding and thrusting tures developed during the Lufilian orogeny of orebodies. resulted in the modification and redistribution The copper and cobalt ores comprise mainly of high-grade ore horizons (Brems et al., 2009; disseminated sulfides, forming stratiform ore- Muchez et al., 2010). bodies hosted in fine-grained siliciclastic or The post-orogenic history is an important dolomitic sedimentary rocks. The copper is pre- factor in the development of economic cobalt dominantly hosted in chalcopyrite, bornite, chal- mineralisation in central Africa. During various cocite and malachite. The cobalt occurs within episodes of weathering, uplift and erosion the sul- cobaltite, carrollite, cattierite, cobalt pentlandite fide ore deposits were partly oxidised, commonly and siegenite, and as solid solution in pyrite (up down to a depth of about 100 m. This process was to 20 per cent cobalt) (Annels and Simmonds, particularly common in the DRC and resulted in 1984). The primary sulfides are commonly over- local cobalt enrichment in the upper part of the printed by secondary supergene ore minerals, oxidised zone, often referred to as a ‘cobalt cap’, with heterogenite the most abundant oxidised due to the downward leaching of copper and cobalt mineral. The weathering process is cobalt by meteoric fluids. These oxide ores can economically significant because it strongly constitute a major part of the cobalt resource in a concentrates cobalt in the near-surface oxidised particular deposit; for example, they account for ore. Hydrothermal minerals associated with the about half of the cobalt resource at Tenke mineralisation include potassium feldspar, Fungurume. phlogopite, sericite, muscovite, albite, carbonate, quartz, and rutile. These assemblages are indica- Other cobalt deposits in sedimentary basins tive of calcium–magnesium, potassic, and sodic alteration (Selley et al,. 2005). Textural evidence The Boléo district, located in the Santa Rosalía indicates that alteration events occurred at mul- basin of Baja California, Mexico, consists of tiple stages during the basin history, and can multiple, laterally extensive stratiform units of vary between and within deposits. These rela- laminated claystone and claystone breccia that tionships result from a combination of: (1) the contain finely disseminated copper–cobalt–zinc widespread and protracted nature of fluid flow; sulfides and oxidised sulfides. The Boléo clastic (2) variability in the composition of host strata; sequence consists of a series of upward coarsening (3) variability in conditions at the sites of ore fan–delta cycles which include laterally extensive formation; and (4) the effects of subsequent zones of conformable sulfide mineralisation with regional metamorphism (Selley et al., 2005). pervasive supergene oxide overprint in the basal In Zambia, although the copper–cobalt deposits claystone of each fan–delta cycle. Mineralisation are predominantly located in the Lower Roan extends for more than 90 km2 with delineated strata, significant examples are also found in the reserves of 85 million tonnes grading 1.33% Cu, underlying basement (e.g. at Lumwana (Bernau 0.08% Co, and 0.55% Zn, (Baja Mining Corp, et al., 2013) and in the Upper Roan (e.g. at 2010). An exhalative- infiltration model is pro- Kansanshi, Broughton et al., 2002). Therefore, the posed to explain the mineralisation with upflow long-lived Zambian Basin appears to preserve a and discharge onto the basin floor of saline metal- protracted history of ore deposition. A role for tec- liferous brines along basin growth faults. tonics in ore formation is recognised with copper– Mineralisation of the basal claystone is due to the cobalt ores related to early sub-basin extensional downward infiltration of brine and replacement faults (Annels, 1989; Selley et al., 2005), faults of diagenetic pyrite (Conly et al., 2006). 128 stephen roberts and gs gnn

SW NE Deforming basin margin Migration of Cu/Co-rich fluids into Lower Roan rocks

Katangan Supergroup

Basement Supergroup

Nchanga (detail in B Leaching of Cu and Co Cu and C below)

Co Generation Dissolution of Cu of hot saline evaporites Cu fluids deep (source of Cu within basin Mg,Ca,fO ,SO 2-) 2 4 Co Co (a) SW NE SW NE Fluid flow promoted within host Hydrocarbon arenites by overlying shale seals Upper Orebody

Fractured arenite Refractory ore and arkose units

Hydrocarbon

Lower Orebody Inverted structures

Basement = fluid flow mineralisation 2- (b) SO4 & Cu (c)

Upper Roan Feldspathic Arenite Lower Banded Shale Basement Shale Mixed Arenites and Shales Lower Arkose Orebody

Figure 6.4 Schematic diagrams illustrating the development of the copper-cobalt deposit at Nchanga, Zambia. (a) Shows basinal brines migrating up reactivated faults at the margins of the Zambian Basin; (b) Shows migration of brines into gas-filled hydrocarbon traps, with overlying shales acting as seals; (c) Shows final distribution of 2– orebodies at Nchanga. (Cu, copper; Co, cobalt; Mg, magnesium; Ca, calcium; SO4 , sulfate; fO2, oxygen fugacity.) (After McGowan et al., 2006.) Cobalt 129

Iron-oxide–copper–gold deposits (IOCG) and komatiitic basalts which host more than twenty nickel sulfide deposits (Lesher, 1989). IOCG deposits represent an enigmatic style of These lenses of massive sulfide are composed mineralisation recently defined as magmatic- of chalcopyrite, pentlandite and pyrrhotite, typi- hydrothermal deposits that contain economic cally located toward the base of komatiitic lava copper and gold grades. They are structurally con- flows within pronounced linear depressions with trolled, contain significant volumes of breccia the sulfides containing between 1–4% Ni and and are commonly associated with pre-sulfide 0.01–0.33% Co. sodic or sodic-calcic alteration. IOCG deposits have a clear temporal, but not a close spatial, relationship to major magmatic intrusions Sudbury Igneous Complex (Groves et al., 2010). The most striking examples The magmatic nickel–copper–platinum-group are the Olympic Dam deposit of South Australia, element (PGE) deposits of the Sudbury Igneous Candelaria in Chile and Salobo in Brazil. Cobalt Complex (SIC) in Ontario, Canada were discov- in the form of carrollite and cobalt-rich pyrite ered in 1883 and by 2002 had produced in the (containing up to 3 wt% Co) occurs within the region of 9.69 million tonnes of nickel, 9.59 mil- stratabound bornite–chalcopyrite–pyrite miner- lion tonnes of copper and 69,600 tonnes of cobalt alisation at Olympic Dam with an estimated (Mudd, 2010). The SIC remains the world’s cobalt grade for the deposit of 0.02% (Williams, largest nickel producer and also produced about 1999). Significant cobalt reserves have also been 2209 tonnes of cobalt in 2011 (Xstrata, 2012); reported in the NICO IOCG deposit located largely as a minor component of pyrrhotite and 160 km north-west of Yellowknife, in Canada. pentlandite which contain between 0.4–1.3% Co. Fortune Minerals Ltd aims to produce 1800 The Sudbury Igneous Complex is a layered tonnes of cobalt per annum from the NICO intrusion emplaced more than 2600 Ma ago at the deposit (Fortune Minerals Limited, 2012). centre of the Sudbury structure, which comprises a series of rocks widely regarded to represent an Magmatic deposits eroded and tectonised remnant of an originally Concentrations of nickel and copper with recov- 200- to 250-km wide meteorite impact basin erable by-product cobalt, typically between 0.04 (Dietz, 1964; Therriault et al., 2002; Lightfoot and 0.08% Co, are present in mafic and ultra- and Doherty, 2001). Three major lithologies are mafic magmas, characterised by low total silica generally recognised within the Sudbury Igneous and alkalis and high magnesium contents. Complex: traditionally termed norite, quartz Sulfides such as chalcopyrite, pentlandite and gabbro and granophyre. At the base of the Sudbury pyrrhotite precipitate from these magmas either Igneous Complex, the sublayer contains discon- from lava flows or within vast magma chambers, tinuous kilometre-sized bodies characterised by particularly during early Earth history. In all the abundant xenoliths and copper–nickel sulfide magmatic deposits the formation of an immis- mineralisation. The sulfide mineralisation is cible sulfide phase is an essential part of the ore- mainly chalcopyrite and pyrrhotite–pentlandite, forming process. Copper, nickel and cobalt are with more massive sulfides towards the base, efficiently scavenged into this sulfide phase on becoming more disseminated upwards. Some of account of their chalcophile and siderophile the xenoliths are identifiable as locally derived properties. country rocks, others constitute a suite of ultra- mafic and mafic rocks of unknown source, which may be genetically linked to the Sudbury Kambalda Complex. There is a clear genetic relationship The Kambalda Dome, located 700 km east of between the Sudbury sulfide orebodies and the Perth, preserves a 3-km thick pile of tholeiitic location of the contact sublayer and offset dikes, 130 stephen roberts and gs gnn which also contain steeply plunging sulfide min- formation of the ores, with the Palaeoproterozoic eralisation (Lightfoot et al., 1997a, 1997b). Tasiuyak gneiss the probable source of certain key elements (Ryan et al., 1995; Li and Naldrett, 2000). Voisey’s Bay The Voisey’s Bay mine is located approximately Norilsk 35 km south-west of Nain in northern Labrador, Canada. In 2012 cobalt production from Voisey’s Igneous intrusions of Triassic age in the Norilsk Bay was 1221 tonnes, derived from ores grading region of northern Russia contain one of the about 0.12% Co (Vale, 2012). The Voisey’s Bay largest known resources of nickel, copper and intrusion is a member of the Nain Plutonic platinum-group element (PGE)-enriched sulfide Suite and comprises gabbroic-troctolitic rocks mineralisation with significant cobalt production emplaced at 1.334 Ga within the boundary zone as a by-product. The sub-volcanic intrusions feed between the Churchill and Nain structural the lower Triassic members of a sequence of provinces (Amelin et al., 1999). It is widely Permo-Triassic flood basalts. The sulfide ores regarded as a prime example of sulfide mineralisa- occur in various forms: (1) disseminated within tion associated with a magmatic conduit system. the intrusions themselves; (2) as concentrations The mineralisation associated with the within and close to the base of the intrusions; Voisey’s Bay intrusive system has been divided (3) disseminated within footwall rocks to the into four principal zones. From east to west, these intrusions; and (4) in rich veins extending away are the Eastern Deeps, the Ovoid, Discovery Hill from the contact into the underlying footwall and Reid Brook. Disseminated to semi-massive rocks (Naldrett, 1989). sulfide mineralisation in the Eastern Deeps Geochemical and isotopic data suggest that occurs within the Basal Breccia Sequence, an the sulfide ore deposits of the Norilsk region were inclusion-rich unit characterized by gneissic formed by injection of olivine-bearing melts con- and other rock fragments in a gabbronorite to taining immiscible sulfide with the sulfur derived ferrogabbronorite matrix; massive sulfide veins from underlying Devonian evaporite country are also present. The Basal Breccia Sequence is rocks. enclosed by a variable-textured troctolite that locally contains xenoliths and up to 50 per cent Laterites interstitial sulfide. The Ovoid ore body is located to the west of the Eastern Deeps, and consists of Nickel–cobalt laterite deposits contain about 70 a bowl-shaped accumulation of massive sulfide per cent of world nickel resources and account (600 × 350 × 110 m), separated from local gneiss by for about 40 per cent of world nickel production, a thin zone of Basal Breccia. To the west, the but they also contain appreciable concentra- Ovoid ore body narrows into disseminated, tions of cobalt, between 0.025 and 0.18% Co sulfide-bearing troctolitic rocks of the Discovery (Berger et al., 2011). These deposits are the prod- Hill zone dike. West of the Discovery Hill zone, uct of pervasive weathering of ultramafic rocks, sulfide mineralisation is principally in a steeply which initially contain between 0.06 and 0.09% dipping dike in the Reid Brook zone and where Co, exposed in tropical to sub-tropical environ- abundant fragments of country rock are found in ments. The formation of nickel–cobalt laterites breccias around the mineralisation (Evans- is influenced by a number of geological vari- Lambswood et al., 2000). ables including protolith composition, topog- Petrological, geochemical and isotopic studies raphy, structure and the duration of appropriate of the Voisey’s Bay intrusion and associated weathering processes. country rocks have highlighted the importance Nickel–cobalt laterite deposits can be grouped of multi-stage magma contamination in the in three main categories: (1) hydrous silicate Cobalt 131

Depth (metres) 0 Kapshtica profile 1 2 Oxide 3 4 Silicate 5 6 020406004080 048121604812 0.0 1.0 2.0 0 1000 2000 SiO2% Fe2O3% MgO% AI2O3% Ni% Co ppm 0 Treni Mine Transported

2

4 Oxide

6

8 Silicate 01020300 204060 0 5 10 15 20 25 04 812160.0 0.8 1.6 0 1500 3000 SiO2%NFe2O3% MgO% AI2O3% i%Co ppm

Figure 6.5 Geochemistry of a cobalt-bearing nickel laterite profiles. The profiles show the significant loss of SiO2 and MgO at the silicate/oxide boundary and concomitant increases in Ni and Co. The profiles also indicate that the highest concentrations of Ni and Co within the profiles occur in close proximity to the silicate/oxide boundary. (SiO2, silica; Fe2O3, iron oxide; MgO, magnesium oxide; Al2O3, aluminium oxide; Ni, nickel; Co, cobalt.) (After Thorne et al., 2012a.) deposits, where hydrous magnesium–nickel sili- primary nickel–cobalt-bearing minerals, olivine cates occur in the lower saprolite, overlain by and serpentine, and the leaching of silica, oxide laterites; (2) clay silicate deposits, with magnesium and other mobile elements. Different largely smectitic clays developed in the mid or types of nickel–cobalt laterites are developed in upper saprolite; (3) and oxide deposits, also known various climatic regimes. Deposits rich in clay as limonite deposits, comprising largely iron- silicates occur mainly in semi-arid regions with oxyhydroxides overlying altered bedrock, or, in hydrous silicate-rich deposits developing in some examples, hydrous silicate and smectitic humid climates and oxide- dominant deposits clay deposits (Gleeson et al., 2003). formed in a range of climatic conditions (e.g. Climate exerts the major control on nickel lat- Freysinnet et al., 2005). Deposits which are cur- erite formation as relatively high temperatures rently located in cooler climatic regimes e.g. and rainfall facilitate intense weathering of ultra- Bitincke, Albania (Thorne et al., 2012a) and mafic protoliths, with the destruction of the Shevchenko in the Urals (Alexander et al., 2007), 132 stephen roberts and gs gnn or in more arid climates e.g. Murrin Murrin, annual precipitation rates (ca. 90 mm per annum) Western Australia (Gaudin et al., 2005), are con- which are not conducive to nickel laterite deposit sidered to be palaeodeposits, originally formed formation. during warmer or wetter climes in the past. The primary control on nickel–cobalt laterite Manganese nodules and cobalt-rich formation is the exposure of suitable protolith, ferromanganese crusts on the seafloor which is typically an ultramafic rock, for example Although not presently a commercial source of harzburgites and dunites of ophiolite complexes cobalt, nickel–cobalt–manganese nodules and (e.g. Goro, New Caledonia), ultramafic igneous crusts found on the seabed are a potential resource intrusions (e.g. Barro Alto, Brazil) or komatiitic for the future. extrusive rocks (e.g. Murrin Murrin, Western Australia). The majority of nickel–cobalt laterites form by the weathering of ophiolites, such that Manganese nodules about 85 per cent of all nickel laterite deposits are Concretions of manganese and other transitional found within accretionary terranes (Freyssinet metal oxides, including cobalt, can accrete around a et al., 2005.) The instability of the protolith nucleus comprising a rock particle, a mineral grain constituent minerals, such as olivine, at surface or a fragment of pre-existing nodule in marine envi- temperatures and pressures, in conjunction with ronments. Manganese–cobalt nodule formation their elevated contents of nickel and cobalt, can involves the oxidation of a flux of dissolved or chem- result in the development of thick weathering ically absorbed manganese and iron through an profiles which possess important nickel–cobalt oxidation gradient transforming manganese from resources. Mn (II) to the oxidised tetravalent species. Two forms In addition to the importance of a suitable cli- of manganese oxide, birnessite ((Na0.3Ca0.1 K 0.1) mate and protolith, the development of substan- 4+ 3+ (Mn ,Mn )2O4 · 1.5H2O) and todorokite ((Mn,Mg, tial nickel–cobalt laterite also requires stable Ca,Ba,K,Na)2Mn3O12 · 3H2O), are the dominant geological conditions where the rate of chemical manganese species although microcrystalline and weathering is higher than the rate of physical amorphous iron and manganese oxides and oxide- erosion. As ultramafic rocks weather to form a hydroxides may constitute the bulk of the accreted laterite deposit, the weathering process and pro- material. In the Clarion–Clipperton Zone (CCZ) file developed will have a finite duration and of the Pacific Ocean, nodules are abundant, with thickness depending upon geological variables about 10 kg m-2, and growth rates estimated at such as uplift rates and the intensity of fracture 2–8 mm/Ma, although elsewhere estimates vary development in the protolith. Consequently, between 1–24 mm/Ma (Cronan, 1992). Grade esti- where uplift is rapid thick laterite deposits are mates for cobalt in the nodules are 0.74 wt% and unlikely to be preserved. the estimated resource for the CCZ is 78 million Recent data suggest that the majority of peri- tonnes of cobalt (Cronan, 1992). In the Atlantic dotites presently weathering to form nickel– Ocean there appears to be a lower concentration of cobalt laterites experience distinct climatic nodules, probably because of relatively high sedi- conditions characterised by limited seasonality mentation rates. and annual precipitation of more than 1000 mm (Thorne et al., 2012b). The importance of Cobalt-rich ferromanganese crusts minimum annual precipitation is well illustrated by the Oman ophiolite which experiences tem- Ferromanganese oxyhydroxide crusts, up to peratures very similar to modern-day laterite 250 mm thick, are found on hard-rock substrates deposits with a Cold Monthly Mean (CMM) tem- throughout the ocean basins mainly on the perature of 23 °C and Warm Monthly Mean tem- flanks and summits of seamounts, ridges, plateaux perature of 31 °C, but is presently subject to low and abyssal hills which are free of sediment Cobalt 133

Sample 35D193 Layers Thickness Mineralogy Element concentration (%) (mm)

Mn Fe Ni Cu Co P2O5 0–20 Fe-vernadite, Mn-ferroxyhyte, 20.0 18.9 0.33 0.10 0.57 1.4 III quartz, buserite, goethite, haematite, feldspars 20–50 Fe-vernadite, Mn-ferroxyhyte, 16.9 16.2 0.38 0.18 0.38 2.0 goethite, clayey materials, II feldspars, apatite, quartz, calcite, haematite 50–65 Fe-vernadite, Mn-ferroxyhyte, 16.8 13.3 0.31 0.17 0.30 9.3 1-2 apatite 65–105 Fe-vernadite, Mn-ferroxyhyte, 14.6 11.9 0.33 0.09 0.25 8.2 goethite, apatite, asbolane, 1-1 calcite, quartz, feldspar

105–165 Asbolane, vernadite, 8.9 5.8 0.47 0.11 0.13 14.0 R todorokite, ferrihydrite, apatite, calcite, quartz

Figure 6.6 Thin section through a ferromanganese crust showing the mineralogical and geochemical variations within the sample. (Mn, manganese; Fe, iron; Ni, nickel; Cu, copper; Co, cobalt; P2O5, phosphorus pentoxide.) (After Glasby et al., 2007.) accumulation. The cobalt content in most of the but nickel is much more enriched in crusts than crusts from mid-Pacific seamounts and seamounts either copper or zinc. in the Exclusive Economic Zones of island nations Cobalt-rich manganese crusts from the far ranges between 0.3 and 0.8%. western Pacific Ocean are of particular interest Iron–manganese crusts form by hydrogenetic because of their great thickness and high concen- precipitation from cold-ambient bottom waters trations of cobalt and platinum, which make them or by a combination of hydrogenetic and hydro- of potential economic value. Crusts from the thermal precipitation in areas of hydrothermal Magellan Seamount cluster, for example, are venting, such as near oceanic spreading axes, typically up to 15–20 cm thick and have cobalt volcanic arcs, and hotspot volcanoes. Iron– contents up to 0.56% (Glasby et al., 2007). manganese crusts contain sub-equal amounts of Ferromanganese crusts from the Ioan Guyot in iron and manganese, and are especially enriched the Magellan Seamount cluster illustrate changes in manganese, cobalt, lead, tellurium, bismuth in mineralogy and composition during crust and platinum relative to lithospheric crust and growth with manganese, iron and cobalt contents sea water. Cobalt is strongly enriched in hydroge- increasing in the crusts with time, whereas nickel netic crusts because it is oxidised from Co2+ to and copper show no systematic changes. These the less-soluble Co3+ at the crust surface. results demonstrate that high cobalt contents The element concentrations in the crusts gen- occurred in the Magellan seamount crusts during erally reflect their abundance patterns in sea the early Miocene. The age at which manganese water. However, there are many complicating crusts with high cobalt contents began to form in factors: for example, copper, nickel and zinc the oceans is difficult to establish. However, it is occur in comparable concentrations in sea water, suggested that high cobalt contents did not appear 134 stephen roberts and gs gnn in ferromanganese crusts from the far western nickel–cobalt-rich matte and a metal-poor slag. Pacific until the early Miocene, following the The matte is further oxidised in a Peirce–Smith formation of a well-developed Oxygen Minimum converter, while the slag is discarded. This Zone (Glasby et al., 2007). approach yields high recoveries of nickel, cobalt, copper and precious metals, but it is expensive in terms of energy use. Extraction, processing and refining Flash smelting accounts for about 75 per cent of nickel sulfide that is processed by pyrometal- The development of cobalt processing technol- lurgy. In this process roasting and smelting ogies and their use worldwide are reviewed in are carried out in a single furnace at 1300 °C recent publications by Crundwell et al. (2011) by continuously blowing oxygen, air, sulfide and Fisher (2011). The production of cobalt from concentrate and silica flux. The furnace matte primary sources is most frequently linked with produced, containing 20–40% Fe, is oxidised in that of copper and nickel. About half of global Peirce–Smith converters to a low-iron matte with cobalt production is either from the leaching of 0.5–4% Fe. Flash smelting is far more energy effi- nickel-bearing laterites or the smelting of nickel cient than roasting/electric furnace melting, but sulfide ores, while much of the remainder of pri- its main disadvantage is that metal losses to the mary supply is derived from copper sulfide or slag are much greater and an auxiliary settling oxide ores. The process flowsheet for cobalt furnace is generally required to recover metals recovery usually includes initial leaching of from the slag. milled ore, a flotation concentrate or smelter In the refinery the first stage of cobalt recovery matte, depending on the nature of the source involves leaching of the matte using chlorine in material (nickel sulfide, copper–cobalt sulfide or hydrochloric acid, air in ammonia solutions, or laterite ore). The leaching is followed by a copper oxygen in sulfuric acid. This is followed by sol- recovery stage and impurity removal before vent extraction to separate the cobalt and the recovery of the cobalt and, finally, of the nickel, if nickel. Electrowinning or hydrogen reduction is any is present (Fisher, 2011). then used for the recovery of the cobalt metal. Overall recovery of cobalt in nickel sulfide smelters varies between 30–80%, which is much Cobalt from nickel sulfide ores less than the typical recovery rates for nickel Cobalt is nearly always present in nickel sul- (97%) and copper (95%). fide ores, occurring typically in pentlandite (Fe,Ni, Increasing environmental pressures in recent

Co)9S8, with the concentration of cobalt gener- years have led to the development of alternative ally between 0.01 to 0.15%. The standard hydrometallurgical routes for the treatment of recovery of cobalt from sulfides involves produc- nickel-sulfide concentrates (Fisher, 2011). Several tion of a flotation concentrate, where nickel and methods are in commercial operation including cobalt behaviour is highly correlated, followed sulfate–chloride pressure leach at Voisey’s Bay in by smelting the concentrate to a cobalt–nickel– Canada and bio-heap leach, under atmospheric sulfur matte (Figure 6.7). Production of a matte conditions, at Talvivaara in Finland. suitable for refining is carried out either by roast- ing the concentrate followed by smelting or by Cobalt from nickel laterite ores flash smelting of the concentrate (Crundwell et al., 2011). About a quarter of nickel-sulfide Although cobalt is generally concentrated to smelting is carried out by the former method some extent in nickel laterite ores due to its which involves production of a calcine in a relatively high level in the precursor ultramafic fluidised-bed roaster followed by melting the rocks, most nickel laterite is smelted to fer- calcine with a silica flux to produce a ronickel and the cobalt is not extracted. However, Cobalt 135

(a) (b)

Nickel Laterite sulfide ore ore

Crushing, Crushing, milling and milling and screening screening

Flotation Copper concentrate Ore-water Nickel-cobalt slurry concentrate Steam Drying Sulfuric Leaching in acid autoclave at Oxygen Sulfur dioxide 250° C and 40 bar Air Flash smelting for sulfuric acid of concentrate production Silica Slag to matte flux Washing and settling furnace solid-liquid separation Converting of Matte molten sulfide matte with high iron, nickel Slag, recycled for Impure nickel and cobalt nickel recovery and cobalt solution

Molten sulfide matte with high Removal of nickel and cobalt, impurities and low iron addition of hydrogen sulfide Figure 6.7 Generic flowsheet for the production of cobalt by (a) flash smelting Separation of nickel Pure cobalt Precipitation of and converting of nickel sulfide ore; and and cobalt by high purity nickel hydrometallurgical (b) high-pressure acid leaching of laterite and cobalt techniques in refinery Pure nickel sulfide ore. in the 1950s in Cuba the introduction of hydro- Cobalt from copper–cobalt ores in DRC metallurgical techniques for the treatment of and Zambia certain nickel laterite ores made the recovery of cobalt from these materials possible. Since that In the major cobalt-producing region of the time high-pressure acid leach (HPAL) technology Central African Copperbelt, although the deposits has been significantly improved and, with several formed primarily as sulfides, weathering has operational plants in Australia, New Caledonia in many areas transformed the cobalt-bearing and Madagascar, it is now a major global source of sulfide phases, such as carrollite (Co2CuS4), to het- nickel and cobalt. A simplified generic flowsheet erogenite (CoOOH) and sphaerocobaltite (CoCO3), for the production of cobalt by high-pressure which occur in association with malachite leaching of goethite-bearing laterite is shown in (CuCO3.Cu(OH)2) and chrysocolla (Cu.SiO2.H2O). Figure 6.7. The ore mined from these weathered zones 136 stephen roberts and gs gnn

Milling and Sulfuric Washing and Cobalt- and Copper-cobalt Sulfide froth Roasting acid solid-liquid copper-rich sulfide ore concentrate flotation leaching separation pregnant solution

Sulfur dioxide Copper Solvent for sulfuric separation and extraction acid production purification

Impurity removal

Production of cobalt hydroxide and metal

Figure 6.8 Schematic flowsheet for the extraction of cobalt from copper–cobalt sulfide ores. typically contains about 0.3% Co and 3% Cu. It be sold or redissolved and cobalt metal recovered is easily excavated from open pits by surface by electrowinning. Although cobalt metal is more scrapers and diggers, although the stripping ratio valuable than cobalt hydroxide or other is often high. intermediate cobalt salts which are traded, the In the past most cobalt produced in the Central production of large quantities of high-quality African Copperbelt was derived from copper flota- cobalt cathode is both technically challenging and tion concentrates. In this route the sulfide concen- capital intensive (Fisher, 2011). Since the global trates are roasted to produce a soluble sulfate economic recession of 2008 and the consequent calcine which is leached in sulfuric acid. Solvent impact on metal prices, some companies have extraction is used to separate cobalt and copper chosen a conservative approach for new projects and the latter is then recovered by electrowinning and opted to produce an intermediate product, (Figure 6.8). Cobalt is recovered by cobalt-hydroxide rather than cobalt metal. precipitation by addition of magnesia, following The WOL method is dependent on abundant the removal of impurities, chiefly iron,aluminium, and cheap supplies of sulfuric acid, but these costs manganese and copper. The cobalt hydroxide may are more than offset by improved metal recoveries then be dried and sent to the market for use in the (Fisher, 2011). The successful utilisation of WOL production of chemicals. Alternatively, the technology has been an important factor in the hydroxide is redissolved and cobalt metal recov- revival of mining in the DRC in the last decade ered by electrowinning. However, the sulfide and this route is likely to become a major contrib- flotation process is inefficient for cobalt, with utor to global cobalt supply in the near future. recoveries as low as 40 per cent for mixed oxide– sulfide ores. Consequently, direct whole ore leach Other sources of cobalt (WOL), which provides much improved metal recoveries, has been increasingly used in recent As mentioned above, the flotation process rou- years in the DRC for processing of dominantly tinely used for the processing of copper sulfide oxide ores (Figure 6.9). In this process the ores ores does not yield efficient recovery of cobalt. undergo reductive leaching followed by solvent Consequently, tailings from these operations are extraction to separate copper and cobalt. As in the potentially large resources of cobalt and have ‘traditional’ method, cobalt hydroxide is precipi- been considered for exploitation. The Kolwezi tated by addition of magnesia. The hydroxide may Tailings project in the Katanga Province of the Cobalt 137

Weathered Leaching Washing and Cobalt- and Milling copper-cobalt Slurry by sulfuric solid-liquid copper-rich and sizing ore acid separation pregnant solution

Sulfur dioxide gas Copper Solvent separation extraction To market for and purification Cobalt metal cobalt chemical to market production Magnesia

Dissolution High-purity Cobalt Cobalt-rich Cobalt Impurity and cobalt hydroxide aqueous cathode removal electrowinning hydroxide precipitation raffinate

Figure 6.9 Schematic flowsheet for the extraction of cobalt from weathered copper–cobalt ore using whole ore leach (WOL).

DRC is the most advanced of this type with a slags which was implemented at Chambishi in reserve of 1.7 million tonnes of copper and 1998. This involved carbothermic reduction in a 363,000 tonnes of cobalt in the tailings which DC arc furnace of the oxides of cobalt, nickel and grade 1.49% Cu and 0.32% Co. First Quantum copper leaving iron as oxide in the slag. The Minerals (FQM) of Canada spent US$750 million cobalt alloy is then processed by pressure leach- to acquire and develop this project up to ing. The company Groupement du Terril de September 2009 when the government of the Lubumbashi (GTL) is processing part of a slag DRC revoked its exploration licence. After pro- heap, known as Big Boy, derived from mining longed legal proceedings the Eurasian Natural operations near Lubumbashi between 1924 and Resources Corp (ENRC) agreed in early 2012 to 1992, which comprises material from mines purchase all the assets and claims of FQM in the working both copper–cobalt and zinc ores. About Katanga Province including the Kolwezi Tailings one third of this heap, comprising about 4.5 mil- project and its processing facility. ENRC is lion tonnes with a cobalt content of about 2.1 per planning to begin production at Kolwezi in late cent, is being processed using this pyrometallur- 2013 with a targeted production capacity of gical process. This operation has the capacity to 70,000 tonnes per annum of copper and 10,000 produce about 5000 tonnes of cobalt per annum tonnes per annum of cobalt (ENRC, 2012). in an alloy which is transported to Finland for Another source of cobalt which has long refining. attracted interest are the slag stockpiles built up Since ENRC became owners of the Chambishi over many years from copper-smelting operations cobalt operation in 2010 there has been a in both Zambia and the DRC. Cobalt follows iron move away from the expensive DC furnace during the smelting process and consequently processing of slag towards the hydrometallurgical over the years some operations have generated processing for the recovery particularly of copper, large slag dumps which contain significant but also cobalt, from concentrates supplied by amounts of cobalt. For example, the Nkana slag ENRC’s mining operations in the DRC. Cobalt dump on the Zambian Copperbelt comprises production at Chambishi, the world’s largest about 20 million tons of slag grading between 0.3 cobalt metal producer, is expected to reach 6000 and 2.6% Co (Jones et al., 2001). Mintek devel- tonnes per annum by the end of 2012 (Darton oped a process to recover the cobalt from these Commodities, 2012). 138 stephen roberts and gs gnn

Other sources of cobalt which are expected perspective is the north-central equatorial Pacific, to become increasingly important result from where a great many volcanic edifices occur within improvements to conventional heap-leach tech- national jurisdictions which would be appropriate nology which is cheaper to set up than high-pressure targets for exploration. Much smaller regional acid leach or smelter operations. Previously uneco- permissive areas exist in the South Pacific, nomic deposits, such as those at Çaldag in Turkey Atlantic and Indian Oceans (Hein et al., 2009). and Acoje in the Philippines, may be amenable to leaching in this manner. Bioleach technology under ambient pressure and temperature has also World production and trade been successfully implemented at the Talvivaara polymetallic sulfide deposit in Finland which will World mine production of cobalt is presently produce 1800 tonnes of cobalt per annum from dominated by the Democratic Republic of Congo, 2012 onwards. which produces more than 65 per cent of the There is also likely to be a revival of interest in global total and currently extracts more than ten cobalt arsenide ores as a result of the development times as much as China, the second largest of a high-temperature, pressure leach process for producer. Other significant producers, but each their treatment which delivers an environmen- with less than five per cent of world production, tally acceptable stable arsenic residue. For include Zambia, Australia, Canada, Cuba, Russia, example, Fortune Minerals Ltd is planning to New Caledonia and Morocco. Global mine pro- develop an open-pit and underground mine to duction of cobalt has grown dramatically during exploit the NICO cobalt–gold–bismuth deposit in the past ten years (Figure 6.10), despite the global the Northwest Territories of Canada. economic downturn, from about 47,000 tonnes/ Formation Metals Inc. is currently building a year in 2002 to 104,000 tonnes/year in 2010 (BGS, mine and concentrator at its Idaho Cobalt Project 2012). The increase in mine production has near Salmon in Idaho. This will be the only pri- largely been supplied by the DRC, which has mary cobalt operation in the USA and is planning raised its output over the same period from to produce 1500 tonnes per annum of high-purity 14,500 tonnes to 70,000 tonnes (Figure 6.11). cobalt metal over a minimum 10-year mine life. As with mine output, global production of The ore reserves in this deposit are currently 2.64 cobalt metal has grown steadily over the past million tons at 0.56% Co and 0.60% Cu two decades (Figure 6.12). In contrast to mine (Formation Metals, 2012). production, refined metal production is domi- nated by China, which produces more than 40 per cent of the world total (Figure 6.13), with Mining considerations for iron–manganese– other substantive cobalt refining capacity in cobalt crusts and nodules Finland (12 per cent) and Zambia (7 per cent). Although iron–manganese nodules and crusts are Notably, the DRC only refines 6 per cent of world considered only as potential future sources of cobalt metal production. About 36 per cent of cobalt and nickel, speculation can be made refined cobalt production is based on imported regarding the geological, geochemical and ocean- material and processed by countries that have ographic parameters that will ultimately deter- no indigenous cobalt mining production, e.g. mine the economic viability of an iron–manganese Belgium. crust mine. Hein et al. (2009) outlined the Cobalt is traded as a variety of refined prod- assumptions which can be made that will likely ucts, predominantly as speciality products and characterise a sea-floor mine site including water chemicals (chiefly cobalt hydroxide, carbonate, depth, topography, crust thickness, and the oxide and sulfate), but also as broken and cut extraction technique used. Overall, it appears cathodes, coarse powder, briquettes, ingots and that the most permissive area from a global rounds (Darton, 2012). Cobalt 139

120

100

80

60

40

20 Figure 6.10 Global mine production of Thousand tonnes (metal content)Thousand cobalt from 1992 to 2010. (Data from 0 British Geological Survey World

Mineral Statistics database.) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

2% 4% 2% cathode metal production in recent years. China 2% 4% is the largest importer of both cobalt ores and concentrates and refined metal by significant 4% margins (Figure 6.15). In 2009 China imported more than 283,000 tonnes of cobalt in ores 5% and concentrates compared with Zambia, the second largest importer, with about 41,000 tonnes. Ores are shipped across the border from 5% the Katanga province of southern DRC for processing in the Zambian Copperbelt, chiefly at Chambishi. 6%

67% Resources and reserves

According to the US Geological Survey world reserves in 2011 are estimated at 7.5 million Australia Democratic Republic of Congo China tonnes, dominated by the DRC with about 45 per cent of the current total, followed by Australia Morocco Canada Cuba New Caledonia with about 19 per cent (USGS, 2012). Global Russia Zambia Others cobalt resources are estimated to be in the region of 15 million tonnes, with the majority located Figure 6.11 The distribution of global mine production within nickel-bearing laterite deposits. The of cobalt in 2010 by country. (Data from British remaining resources occur mainly in nickel– Geological Survey, 2012.) copper sulfide deposits hosted in mafic and ultra- mafic rocks in Australia (Kambalda district, The DRC is the largest global exporter of Western Australia), Canada (Sudbury district, both cobalt ores and concentrates, followed by Ontario and Voisey’s Bay, Labrador) and Russia Zambia (Figure 6.14). The DRC is also the larg- (Norilsk-Talnakh district), and in the copper– est exporter, by a significant margin, of refined cobalt deposits of the DRC and Zambia. In metal, reflecting the increasing level of cobalt addition, as much as one billion tonnes of 140 stephen roberts and gs gnn

80

70

60

50

40

30

Thousand tonnes Thousand 20

10 Figure 6.12 Global production of cobalt 0 metal from 1992 to 2010. (Data from British Geological Survey World Mineral 1992199319941995199619971998199920002001200220032004200520062007200820092010 Statistics database.)

6% Uses 2% 3% 3% Although cobalt was first isolated in elemental form in 1730, use in pigments remained its only 3% 43% practical application until 1907 when the metal was first used in alloys. Cobalt continues to have 4% few applications in its pure form and is most commonly used as an alloy constituent or 5% chemical compound, where its chemical and wear resistance, magnetic properties and high temperature strength are used in a diverse range 6% of materials with commercial, industrial and mil- itary applications. At present the main uses of cobalt are in: 6% ● Batteries: the largest use of cobalt (30 per cent of total) is in batteries. It is an important 7% component in the three main rechargeable battery technologies: 1) lithium-ion batteries 12% may contain up to 60 per cent cobalt as lithium China Finland Zambia Canada cobalt oxide. However, this depends on the Democratic Republic of Congo Russia chemical construction of the lithium ion Australia Norway Belgium battery. Batteries composed of Li-Ni-Al-Co (NCA) may contain as little as 9 per cent Japan Morocco Others cobalt; 2) nickel-metal hybrid batteries, used in current hybrid electric vehicles, contain up Figure 6.13 The distribution of global production of to 15 per cent cobalt; 3) cobalt oxide or hydroxide cobalt metal in 2010 by country. (Data from British powder is used in nickel–cadmium batteries Geological Survey, 2012.) accounting for one to five per cent of the battery composition. unidentified (hypothetical and speculative) cobalt ● Superalloys and magnet alloys: cobalt is widely resources may exist in manganese nodules and used as an alloying metal in superalloys and crusts on the ocean floor (USGS, 2012). magnets. Superalloys are primarily used in jet Cobalt 141

300

250

200

150 Thousand tonnes Thousand 100

50

0

Cuba USA Zambia Congo Russia Canada Australia D R Congo Netherlands South Africa Other Countries Ores & Concentrates Metal

Figure 6.14 The main exporting countries of cobalt ores and concentrates and refined metal in 2010. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.)

300

250

200

150

Thousand tonnes Thousand 100

50

0

India USA Japan Zambia Finland Norway Germany Netherlands United Kingdom Other Countries

China (inc Hong Kong) Ores & Concentrates Metal

Figure 6.15 The main importing countries of cobalt ores and concentrates and refined metal in 2010. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) 142 stephen roberts and gs gnn engines and turbines, and in automotive and (Kapusta, 2006). It is estimated that 24 per cent of chemical applications. Cobalt is used in magnetic the USA annual consumption of cobalt was alloys, either in high-strength samarium–cobalt sourced from scrap in 2011 (USGS, 2012). magnets or lower powered AlNiCo magnets. Recycling of alloy and hard metal scrap is gen- These magnets are used in a variety of applica- erally operated by and within the superalloy and tions including high-performance electrical metal carbide sectors. Recycling of catalysts and equipment. batteries is also done via the cobalt industry. ● Catalysts: catalytic processes account for 10 These end-of-life products are an increasingly per cent of all cobalt consumption, with one of important source of cobalt supply for the EU the primary uses to increase polymerisation and cobalt industry in particular. The end-of-life oxidisation rates in the manufacture of plastic recycling rate of cobalt is estimated at 68 per resins. Cobalt is also used in gas–liquid technol- cent and the recycled content rate is estimated ogies where natural gas is processed to produce at 32 per cent (United Nations Environmental synthetic diesel fuel. It is also used in the petro- Program, 2011). Overall, the EU target is to chemical industry in the form of cobalt oxide to recycle 40 per cent of batteries by 2016 in order remove sulfur from crude oil in the refinery to reduce the demand for primary cobalt. process. Politically, recycling cobalt has also become ● Other applications: cobalt is used in a variety of more important in order to become less other applications including: as a binder material in dependent on a few primary suppliers, histori- hard materials, such as cemented carbide and dia- cally located in central Africa. Recent events in mond tool applications (13 per cent); as a compo- the DRC, such as the seizure of the Kolwezi nent of high speed steels and other high-strength Tailings Project by the government, have served alloys (5 per cent); in pigments in glass, enamels, to discourage foreign investment in the mining pottery and china (9 per cent); in medical applica- industry in that country. tions as part of cancer treatments, as well as in the alloy vitallium (cobalt–chromium–molybdenum– carbon) used in prostheses systems and dentistry; Substitution and in electronic connectors on integrated circuits (containing up to 15 per cent cobalt). Substitutes for cobalt are constantly being sought Further information on cobalt uses is available mainly due to metal price volatility. However, from the Cobalt Development Institute (CDI, given the unique properties of cobalt, there are 2012a). limited options for substitution and almost all substitutes result in reduced product performance. Nevertheless, potential substitutes for cobalt in Recycling its major end uses include: ● in magnets by barium or strontium ferrites, Price volatility, geopolitics of supply and potential neodymium–iron–boron, or nickel–iron alloys; cost and environmental benefits drive the recy- ● in paints by cerium, iron, lead, manganese, or cling of cobalt. Scrap metal, spent catalysts, and vanadium; rechargeable batteries are the most readily ame- ● in jet engines and petroleum catalysts by nickel nable cobalt-bearing products for recycling, and nickel-based alloys. whereas cobalt recycling from applications in pig- ● in lithium ion batteries, by iron–phosphorus– ments, glass and paint is not possible because it is manganese, nickel–cobalt–aluminium or nickel– dissipated in use. In terms of tonnage, cobalt cobalt–manganese. The cobalt content may recovery from secondary feeds more than doubled ultimately be reduced or replaced by these in the period 1995 to 2005, from an estimated 4200 cheaper metals and alloys, with technology tonnes to more than 10,000 tonnes respectively development reducing the cobalt content from 60 Cobalt 143

13% Measured atmospheric concentrations of cobalt in unpolluted areas are typically <1–2 ng/ 30% m3 and surface and groundwater concentra- tions of cobalt are also low, typically less than 7% 1–10 μg/l (Smith and Carson, 1981; Hamilton, 1994). The mean cobalt concentrations reported in surface waters throughout Europe is 0.333 ±1.01 μg/l [standard deviation] (Salminen 9% et al., 2005). Cobalt concentrations in drinking water are generally less than 1–2 μg/l (Kim et al., 2006). The average concentration of cobalt in soils throughout the world is about 8 ppm (Kim et al., 9% 2006); European topsoil and subsoil concentrations were reported as 10.4 ±13.3 and 11.1 ±10.5 ppm, respectively (mean ± standard deviation) (Salminen 19% 13% et al., 2005). Elevated cobalt concentrations in soils around some mine sites in Ontario, Canada Battery chemicals Super alloys Hardmetals have been reported as high as 6450 ppm (Frank et al., 1976). Catalysts Ceramics/Pigments Magnets Cobalt metal is an essential component of

Others vitamin B12, cobalamin. Neither higher plants nor

animals can synthesize vitamin B12, but both Figure 6.16 Estimated cobalt demand by end-use in require trace amounts. Non-ruminant animals

2011. (Data from Cobalt Development Institute, are unable to synthesize vitamin B12 from inor- 2012a.) ganic cobalt, rather they require cobalt in the

form of vitamin B12. Ruminant animals (e.g. per cent to less than 10 per cent. It is predicted cows), on the other hand, have micro- organisms that the use of existing cobalt cathode materials present in their stomach (rumin) that are able to for Li-ion batteries will drop to approximately 45 synthesize vitamin B12 from elemental cobalt, per cent of current use by 2013 having previously and the vitamin B12 produced by the bacteria accounted for almost all the market (European serves to meet the animals’ cobalt require- Pathway to Zero Waste, 2011). ment. Therefore, ruminant animals have a dietary requirement for elemental cobalt, while non- ruminant animals have a requirement for vitamin

Environmental issues B12 (NRC, 2005). Vitamin B12 deficiency can have various effects on health including infertility, Cobalt in the environment is derived from both increased perinatal mortality, anaemia, fatty natural and anthropogenic sources. Natural sources liver and decreased ability to fight disease. of cobalt include erosion of cobalt-containing Characteristic signs of chronic cobalt overexpo- rocks, seawater, volcanic activity and biogenic sure for most species include reduced food con- emissions. Anthropogenic sources of cobalt include sumption, decreased body weight, blood disorders, mining and processing of cobalt-containing ores, debility and increased disease susceptibility. agricultural application of cobalt-containing fer- Estimates of human exposure to cobalt sug- tilisers and deposition of atmospheric particu- gest that more than 99 per cent is through the lates from combustion of fossil fuels (Smith and ingestion of food, with an estimated daily intake Carson, 1981). of 5–40 μg/day, most of which is inorganic 144 stephen roberts and gs gnn cobalt and almost all of which passes through the next two years, the price of cobalt fell in the body unabsorbed. (Kim et al., 2006). The response to weakened demand given the global recommended Dietary Reference Intake (RDI) economic downturn (Plunkert and Jones, 1999). of Vitamin B12 for adults is 2 μg/day (NRC, Due to concerns over cobalt supply from the 1998) and ingestion of cobalt within normal DRC, the price of cobalt more than doubled in dietary ranges has not been associated with late 1993 – early 1994. Between 1993 and 1995, adverse health consequences. High repetitive despite increased world production of cobalt, oral doses of cobalt, on the other hand, have high levels of demand supported a cobalt price been associated with effects on red blood cells, between $20 and $30 per pound. However, fore- thyroid and heart. Inhalation of high concentra- casts of large increases in nickel demand with tions of cobalt is linked to lung disease such as associated new sources of cobalt production, led asthma and pneumonia, but these effects appear to concerns about potential over-supply and the restricted to workers exposed to high levels of cobalt price fell to approximately $21.50 per cobalt in the air (U.S. EPA, 2000). pound by the end of 1995. From 1995 to 2002, the Additional information about the health, general trend in cobalt prices was downward as safety and environmental effects of cobalt is supply outpaced demand. However, during 2003, available from the Cobalt Development Institute cobalt prices increased sharply in response to (2012b). reduced production and concerns over tightness in global supply. With the commodity boom that followed prices continued to rise until 2008 Prices when, in response to the global financial down- turn, prices fell from near $50 per pound to $15 Unlike other major industrial metals cobalt has per pound. By mid-2011 these prices were stable only recently began trading on the London at around $20 per pound with an early 2012 price Metal Exchange. Prior to this, Western Mining of about $15 per pound. It seems probable that Corporation began selling cobalt on its website price volatility will continue, although this will (the Cobalt Open Sales System – COSS) in 1999 depend on various factors including the timing of and in September 2000 it was joined by the OMG the opening of new mine capacity relative to Group Inc. who also began selling its briquettes mine closures and the pace of demand growth. in this manner. Following these initiatives, other trading companies began to offer a buying and selling service through the internet. In 2008 BHP Outlook (now incorporating Western Mining Corp) sus- pended the COSS and the London Metal Exchange There are many potential new sources of cobalt, (LME) started the trading of cobalt (minimum Co both onshore, in Canada, Western Australia, the content 99.3%) in February 2010. The LME offers a DRC, Zambia and Madagascar, and offshore in fully regulated market with which to trade spot and deep-sea nodules. A major recent development future cobalt contracts. The global contract is has been the increase in mining activity in traded in 1-metric-tonne lots, minimum 99.3% the DRC where significant volumes of ore and cobalt metal, with delivery to warehouses in Asia, concentrate are being mined for refining else- Europe, and the United States. where, mainly China. It appears that future The price of cobalt is linked to supply, demand developments in the DRC will be an impor- and the prevailing political environment of the tant factor in cobalt production. Given the key producer, the DRC (Figure 6.17). For example, importance of supply from the DRC and its in the early 1990s the price of cobalt peaked at relatively complicated and uncertain political around $33 per pound largely due to political and environment it is perhaps not surprising that economic tensions in the DRC. However, during cobalt finds its way onto lists of critical metals, Cobalt 145

60

50

40

30 US$ per lb

20

10

0

Q1, 1992Q4, 1992Q4, 1993Q4, 1994Q4, 1995Q4, 1996Q4, 1997Q4, 1998Q4, 1999Q4, 2000Q4, 2001Q4, 2002Q4, 2003Q4, 2004Q4, 2005Q4, 2006Q4, 2007Q4, 2008Q4, 2009Q4, 2010Q4, 2011

Figure 6.17 The quarterly average price of cobalt, 1992–2011. (High Grade Metal Bulletin free market US$ per lb in warehouse.)

scoring high on economic importance, and with of scenarios for the future deployment of electric significant security of supply risks (European vehicles and the market share of various battery Commission, 2010). types and concluded that the availability of cobalt Since 2002 cobalt consumption in Asia has is more than adequate up to 2025, even without increased significantly, while demand in the West additional supply from the DRC (U.S. DOE, 2011). has remained stable. The growth in demand is due Cobalt supply and demand forecasts suggest chiefly to increased use in chemical applications, that the cobalt market will remain roughly in particularly rechargeable batteries and catalysts. balance for the coming decade (CDI, 2012a). The The demand for rechargeable batteries for portable Cobalt Development Institute (CDI) predicts a electronic devices and in automotive applications growth rate of 2.5 per cent per year, which lies in is likely to continue to grow, although the demand between the high and low growth rates modelled for cobalt will strongly depend on which battery by Öko-Institut of 2.8 and 1.7 per cent (Öko- technology is adopted by the car industry. Cobalt Institut, 2009). In addition to the growing demand is used in both nickel–metal-hydride batteries, as for cobalt in batteries, its use in superalloys is currently utilised in hybrid electric vehicles, and expected to increase in response to continuing in the more powerful lithium-ion batteries used in expansion of the global aerospace market. On the all electric and plug-in hybrid electric vehicles. supply side there is a considerable degree of The US Department of Energy analysed a number uncertainty over the opening of new mines over 146 stephen roberts and gs gnn the next few years. The USGS lists over 80,000 Annels, A.E. and Simmonds, J.R. (1984) Cobalt in the tonnes of potential new mine capacity that is Zambian Copperbelt. Precambrian Research 25, 75–98. scheduled for completion by 2013. Some of this Baja Mining Corp (2010) http://www.bajamining.com/ will augment or replace existing capacity, static/downloads/Resource%20and%20Reserve.pdf although some is unlikely to open or be fully uti- Bending, J.S. and Scales, W.G. (2001) New production in the Idaho cobalt belt; a unique metallogenic province. lised. Nonetheless, the existence of this potential Transactions – Institution of Mining and Metallurgy. mining capacity goes some way to alleviating Section B: Applied Earth Science 110, 81–87. scarcity concerns for cobalt even if demand Berger V.I., Singer D.A., Bliss, J.D. and Moring. B.C. growth is higher than predicted. Substitution of (2011) Ni-Co Laterite Deposits of the World – cobalt by cheaper metals in batteries and Database and Grade Tonnage Models. US Geological increased recycling of cobalt from metal alloys Survey Open File Report 2011-1058, 26pp. and batteries may also have significant impacts Bernau R., Roberts, S., Richards, M., Nesbitt, B., Boyce, on the market in the future. A.J. and Nowecki, J. (2013) The Geology and Geochemistry of the Lumwana Cu (± Co ± U) deposits, NW Zambia. Mineralium Deposita 48, 137–153. Acknowledgements Bookstrom, A.A., Johnson, C.A., Landis, G.P. and Frost, T.P. (2007) Blackbird Fe-Cu-Co-Au-REE deposits. In: Gus Gunn publishes with the permission of the O’Neill J.M. (ed.) Metallogeny of Mesoproterozoic Executive Director of the British Geological Survey. Sedimentary Rocks in Idaho and Montana - Studies by the Mineral Resources Program, U.S. Geological Survey, 2004–2007. U. S. Geological Survey Open- Notes File Report 2007–1280, 11–20. Brems, D., Muchez, P., Sikazwe, O. and Mukumba, W. 1. The amount of time in years required to replace a (2009) Metallogenesis of the Nkana copper-cobalt given ocean constituent from river supply. South Orebody, Zambia. Journal of African Earth 2. Ma, million years before present. Sciences 55, 185–196. British Geological Survey (2012) World Mineral References Production 2006–2010 (Keyworth, Nottingham, British Geological Survey). Ahmed, A.H., Arai, S., and Ikenne, M. (2009) Mineralogy Broecker, W.S. and Peng, T.H. (1982) Tracers in the Sea. and paragenesis of the Co-Ni arsenide ores of Bou Lamont-Doherty Geological Observatory, Palisades, N.Y. Azer, Anti-Atlas, Morocco. Economic Geology 104, Broughton, D.W., Hitzman, M.W. and Stephens, A.J. 249–266. (2002) Exploration history and geology of the Kansanshi Alexander, K., Herrington, R., Kirk, C. and Chlup, D. Cu(-Au) deposit, Zambia. Special Publication Society (2007) Geology and mineralogy of the Shevchenko of Economic Geologists (U. S.) 9, 141–153. Lateritic Ni-Co deposit, Kazakhstan. In: Andrew, C.J. Cailteux, J.L.H., Kampunzu, A.B., Lerouge, C., Kaputo, et al. (ed) Proceedings of the 9th biennial meeting of A.K. and Milesi, J.P. (2005) Genesis of sediment- the Society for Geology Applied to Mineral Deposits, hosted stratiform copper-cobalt deposits, Central Dublin, Ireland, 20th-23rd August 2007. Irish African Copperbelt. Journal of African Earth Sciences Association of Economic Geology, Dublin, 1393–1396. 42, 134–158. Amelin, Y., Li, C., and Naldrett, A.J. (1999) Cobalt Development Institute (2012a). Cobalt Facts: Geochronology of the Voisey’s Bay Intrusion, Cobalt Supply and Demand 2011. http://www.thecdi. Labrador, Canada, by precise U-Pb dating of coexist- com/cdi/images/documents/facts/Cobalt%20Facts- ing baddeleyite, zircon, and apatite. Lithos, An Supply%20%20Demand-2011.pdf International Journal of Petrology, Mineralogy and Cobalt Development Institute (2012b). www.thecdi. Geochemistry 47, 33–51. com/hse Annels, A.E. (1989) Ore genesis in the Zambian Conly, A.G. Beaudoin, G. and Scott, S.D. (2006) Isotopic Copperbelt, with particular reference to the northern constraints on fluid evolution and precipitation sector of the Chambishi Basin. Special Paper – mechanisms for the Boléo Cu–Co–Zn district, Geological Association of Canada 36, 427–452.` Mexico. Mineralium Deposita 41, 127–151. Cobalt 147

Cronan, D.S. (1992) Marine Minerals in Exclusive Gleeson, S.A., Butt, C.R.M. and Elias, M. (2003) Nickel Economic Zones. Volume 5 of Topics in the Earth laterites; A review. Society of Economic Geologists Sciences, Springer, pp. 212. (SEG) Newsletter, 54 (1), 12–18. Crundwell, F.K., Moats, M.S., Ramachandran, V., Groves, D.I., Bierlein, F.P., Meinert, L.D. and Hitzman, Robinson, T.G. and Davenport, W.G. (2011) Extractive M.W. (2010) Iron Oxide Copper-Gold (IOCG) Deposits Metallurgy of Nickel, Cobalt and Platinum-Group through Earth History: Implications for Origin, Metals. Elsevier, Amsterdam, Netherlands, pp. 610. Lithospheric Setting, and Distinction from Other Darton Commodities Ltd (2012) Cobalt Market Review Epigenetic Iron Oxide Deposits. Economic Geology 2011–2012. http://www.dartoncommodities.co.uk/ 105, 641–654. Dietz, R.S. (1964) Sudbury structure as an astrobleme: Hein, J.R., Conrad T.A. and Dunham, R.E. (2009) Journal of Geology 72, 412–434. Seamount Characteristics and Mine-Site Model ENRC (2012) http://www.enrc.com/Media/ press- Applied to Exploration- and Mining-Lease-Block releases Selection for Cobalt-Rich Ferromanganese Crusts. Evans-Lamswood, D.M., Butt, D.P., Jackson, R.S., Lee, Marine Georesources and Geotechnology 27, 160–176. D.V., Muggridge, M.G., Wheeler, R.I., and Wilton, Hu, Z. and Gao, S. (2008) Upper crustal abundances of D.H.C. (2000) Physical controls associated with the trace elements; a revision and update. Chemical distribution of sulfides in the Voisey’s Bay Ni-Cu-Co Geology 253, 205–221. deposit, Labrador. Economic Geology 95, 749–770. John, T., Schenk, V., Haase, K., Scherer, E., and Tembo, European Pathway to Zero Waste (2011) (http://www. F. (2003) Evidence for a neoproterozoic ocean in environment-agency.gov.uk/static/documents/ south-central Africa from mid-oceanic-ridge-type Business/EPOW-recovering-critical-raw-materials- geochemical signatures and pressure-temperature T5v2.pdf. estimates of Zambian eclogites. Geology 31, 243–246. European Commission (2010) Report of the Ad hoc Jones, R.T., Denton, G.M., Reynolds, Q.G., Parker, Working Group on defining critical raw materials. J.A.L. and van Tonder, G.J.J. (2001) Recovery of cobalt http://ec.europa.eu/enterprise/policies/raw-materials/ from slag in a DC arc furnace at Chambishi, Zambia. files/docs/report_en.pdf. Presentation given at Copper Cobalt Nickel and Zinc Fisher, K G. (2011) Cobalt processing developments. 6th Recovery conference, Victoria Falls, Zimbabwe, 16–18 Southern African Base Metals Conference 2011. The July 2001. Southern African Institute of Mining and Metallurgy, Kapusta, J. (2006) Cobalt Production and Markets: A 237–258. Brief Overview. Journal of the Minerals, Metals and Formation Metals Inc. (2012) Formation Metals Materials Society (JOM) 58, 33–36. company presentation, March 2012. http://www. Kampunzu, A.B., Tembo, F., Matheis, G., Kapenda, D. formationmetals.com/s/Home.asp. and Huntsman-Mapila, P. (2000) Geochemistry and Fortune Minerals Limited. (2012) http://www.fortune- tectonic setting of mafic igneous units in the minerals.com/Projects/NICO/NICO-Gold-Cobalt- Neoproterozoic Katangan Basin, central Africa; Bismuth-Deposit/default.aspx. implications for Rodinia break-up. Gondwana Frank, R., Ishida, K. and Suda, P. (1976) Metals in agri- Research 3, 125–153. cultural soils of Ontario. Canadian Journal of Soil Kim, J.H., Gibb, H.J. and Howe, P.D. (2006) Cobalt and Science 56, 181–196. inorganic cobalt compounds. Concise international Freyssinet, Ph., Butt, C.R.M. and Morris, R.C. chemical assessment document; Document 69. (2005) Ore-forming processes related to lateritic World Health Organization, Geneva, Switzerland. weathering. Economic Geology 100th Anniversary Leblanc, M. and Kroener, A. (1981) The Late Proterozoic Volume, 681–722. Opholites of Bou Azzer (Morocco); Evidence for Gaudin, A., Decarreau, A., Noack, Y., and Grauby, O. Pan-African Plate Tectonics. In: Precambrian Plate (2005) Clay mineralogy of the nickel laterite ore Tectonics, Netherlands (NLD): Elsevier, Amsterdam, developed from serpentinised peridotites at Murrin Netherlands, 431–451. Murrin, Western Australia. Australian Journal of Leblanc, M. and Billaud, P. (1982) Cobalt arsenide ore- Earth Sciences 52, 231–241. bodies related to an upper proterozoic ophiolite; Bou Glasby, G., Ren X., Shi X., and Pulyaeva, I. (2007) Co– Azzer (Morocco). Economic Geology and the Bulletin rich Mn crusts from the Magellan Seamount cluster: of the Society of Economic Geologists 77 (1), 162–175. the long journey through time. Geo-Marine Letters Leblanc, M. and Fischer, W. (1990) Gold and platinum 27, 315–323. group elements in cobalt-arsenide ores; hydrothermal 148 stephen roberts and gs gnn

concentration from a serpentinite source-rock (Bou Council, National Academy of Science, The National Azer, Morocco). Mineralogy and Petrology 42, 197–209. Academies Press, Washington, D.C., USA. Lesher, C.M. (1989) Komatiite hosted Ni-Sulphide Oberthur, T., Melcher, F., Henjes-Kunst, F. et al. (2009) deposits. Reviews in Economic Geology 4, 45–101. Hercynian age of the cobalt-nickel-arsenide-(gold) Li, C. and Naldrett, A.J. (2000) Melting reactions of ores, Bou Azer, Anti-Atlas, Morocco; Re-Os, Sm-Nd, gneissic inclusions with enclosing magma at Voisey’s and U-Pb age determinations. Economic Geology Bay, Labrador, Canada; implications with respect to 104, 1065–1079. ore genesis. Economic Geology 95, 801–814. Öko-Institut (2009) Critical metals for future sustain- Lightfoot, P.C., Keays, R.R., Morrison, G.G., Bite, A. able technologies and their recycling potential. and Farrell, K.P. (1997a) Geochemical relation- http://oeko.de/oekodoc/1070/2009-129-en.pdf. ships in the Sudbury igneous complex; origin of Plunkert, P.A., Jones, T.S. (1999) Metals prices in the the main mass and offset dikes. Economic Geology United States through 1998: Cobalt. http://minerals. 92, 289–307. usgs.gov/minerals/pubs/commodity/cobalt/210798. Lightfoot, P.C., Keays, R.R., Morrison, G.G., Bite, A. pdf. and Farrell, K.P. (1997b) Geologic and geochemical Ryan, B., Wardle, R.J., Gower, C.F. and Nunn, G.A.G. relationships between the contact sublayer, inclu- (1995). Nickel-copper sulphide mineralisation in sions, and the main mass of the Sudbury Igneous Labrador: The Voisey’s Bay discovery and its Complex; a case study of the Whistle Mine exploration implications; Current Research, Report Embayment. Economic Geology 92, 647–673. 95-1, Geological Survey, Department of Natural Lightfoot, P.C. and Doherty W. (2001) Chemical Resources, Government of Newfoundland and Evolution and Origin of Nickel Sulfide Mineralisation Labrador, 109–129. in the Sudbury Igneous Complex, Ontario, Canada Salminen R., Batista M.J., Bidovec M. et al. (2005) Economic Geology 96, 1855–1875. Geochemical Atlas of Europe. Part 1 – Background McGowan, R.R., Roberts, S. and Boyce, A.J. (2006) Information, Methodology and Maps. Espoo, Geological Origin of the Nchanga copper-cobalt deposits of Survey of Finland. the Zambian Copperbelt. Mineralium Deposita 40, Selley, D., Broughton, D., Scott, R. et al. (2005) A new 617–638. look at the geology of the Zambian Copperbelt. In: McGowan, R.R., Roberts, S., Foster, R.P., Boyce A.J. and Hedenquist, J.W., Thompson, John F.H.; Goldfarb, Coller, D. (2003) Origin of the copper-cobalt deposits R.J. and Richards, J.P. (eds.) Economic Geology; of the Zambian Copperbelt; an epigenetic view from One Hundredth Anniversary Volume 1905–2005, Nchanga. Geology 31, 497–500. 965–1000. Mudd, G.M. (2010) Global Trends and Environmental Slack, J.F. (2006) High REE and Y concentrations in Issues in Nickel Mining: Sulfides Versus Laterites. Co-Cu-Au ores of the Blackbird District, Idaho. Ore Geology Reviews 38 (1–2), 9–26. Economic Geology 101, 275–280. Muchez, P., Brems, D., Clara, E. et al. (2010) Evolution Smith I.C. and Carson B.L. (1981) Trace Metals in the of Cu-Co mineralizing fluids at Nkana Mine, Central Environment. Ann Arbor, Michigan, Ann Arbor African Copperbelt, Zambia. Journal of African Earth Science Publishers. Sciences 58, 457–474. Trans Continental Exploration and Mining Company Naldrett, A.J. (1989) Magmatic Sulphide Deposits. (2011). http://www.tcemco.com/ Oxford Monographs on Geology and Geophysics 14. Therriault, A.M., Fowler, A.D. and Grieve R.A.F. vii +, New York, Oxford: Oxford University, pp.186. (2002) The Sudbury Igneous Complex: A Differen- Nash, J.T. and Hahn, G.A. (1989) Stratabound Co-Cu tiated Impact Melt Sheet. Economic Geology 97, deposits and mafic volcaniclastic rocks in the Blackbird 1521–1540. mining district, Lemhi County, Idaho. Special Paper – Thorne, R., Herrington, R.J. and Roberts, S. (2012a). The Geological Association of Canada 36, 339–356. Formation and Evolution of the Bitincke Nickel lat- National Research Council (1998) Dietary Reference erite deposit, Albania. Mineralium Deposita 47 (8), Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, 933–947. Folate, Vitamin B12, Pantothenic Acid, Biotin, and Thorne, R., Roberts, S. and Herrington, R.J. (2012b) Choline. Washington, DC: The National Academies Climate change and the formation of Nickel laterites. Press. Geology 40 (4) 331–334. National Research Council (2005) Mineral Tolerance of United Nations (UN) Comtrade (2013) Commodity Animals, 2nd Revised Edition, National Research Trade Statistics Database (http://comtrade.un.org/db/). Cobalt 149

United Nations Environmental Program (UNEP) (2011) Vale (2012) Annual Report 2012. http://www.vale.com/ Recycling Rates of Metals – A status report, A report of EN/investors/Quarterly-results-reports/20F/ the Working Group on the Global Metal Flows to the 20FDocs/20F_2012_i.pdf. International Resource Panel. Graedel, T.E., Alwood, Williams P.J. (1999) Fe-oxide-Cu-Au deposits of the J., Birat, J.-P., Reck, B.K., Sibley, S.F., Sonnemann, G., Olympic Dam/Ernest Henry type. In: Hodgson, C.J., Buchert, M., Hagelüken, C. and Franklin, J.M., (eds.) New developments in the U.S. Department of Energy (2011) Critical materials geological understanding of some major ore types and strategy, December 2011. environments, with implications for exploration: U.S. Environmental Protection Agency (2000). Air Toronto, Prospectors and Developers Association of Toxics Web Site – Cobalt: http://www.epa.gov/ttn/ Canada, 1–43. atw/hlthef/cobalt.html. Xstrata (2012) http://www.xstratanickel.com/EN/ U.S. Geological Survey. (2012) Mineral Commodity Publications/Publications/xni%20brochure%20 Summaries: Cobalt. http://minerals.usgs.gov/minerals/ 2012_2%208_150%20dpi.pdf. pubs/commodity/cobalt/cobalmcs07.pdf. 7. Gallium

THOMAS BUTCHER1 AND TERESA BROWN2

1 Independent Consultant, New York, USA 2 British Geological Survey, Keyworth, Nottingham, UK

Introduction able anisotropy in electrical resistivity dependent upon the orientation of crystals within its struc- The existence of gallium was predicted by Dmitri ture. Other selected properties of gallium are Mendeleev in 1871 as part of his development of shown in Table 7.1. the Periodic Table of elements, but it was actu- ally discovered in 1875 by Paul Émile Lecoq de Boisbaudran. As a patriotic Frenchman, he named Mineralogy and distribution the new element ‘gallia’ after the Latin name for France. Gallium has an average crustal abundance of 19 ppm, which is greater than many better known metals, for example lead (ten ppm) or tin (two Physical and chemical properties ppm). However, it does not occur in nature as a native metal but instead it substitutes for As a pure metal, gallium is silvery-white in other elements in certain minerals, although appearance. It is a relatively soft metal, with a usually at trace quantities. It is generally low melting point but a comparatively high extracted as a by-product of aluminium or zinc boiling point. Combined with a low vapour production because it is rarely, if ever, found in pressure, even at high temperatures, this gives it sufficient quantities by itself to enable economic the longest liquid range of any metal. As a liquid extraction. it will wet glass and skin, and it will readily con- There are a few minerals in which gallium forms taminate other metals by diffusing into their a significant part. The main gallium-bearing min- lattice structures, making it more difficult to eral is gallite (CuGaS2) which is similar to the handle than many other commodities. copper sulfide mineral chalcopyrite (CuFeS2), but When it solidifies from a liquid, gallium will with gallium substituting for iron in the crystal expand and therefore should not be stored in a structure. Gallite is known to occur at Lubumbashi restricted container as this would rupture. As a in the Democratic Republic of Congo and at solid, gallium will fracture conchoidally like Tsumeb in Namibia (Roskill, 2011), albeit in uneco- glass. Gallium is magnetic and a good conductor nomic quantities. Other gallium-bearing minerals of both electricity and heat. It exhibits a notice- have been identified, principally at Tsumeb in

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Gallium 151

Namibia, as shown in Table 7.2, but all remain also appears to be more abundant where the relatively rare (Geier and Ottemann, 1970; Scott, bauxite was derived from alkali source rocks 1971 and Jambor et al., 1996). Gallium is more com- (Weeks, 1989). monly found as a trace element in bauxite (the It is generally accepted that the gallium in main ore of aluminium) and sphalerite (a zinc ore). bauxite originates from minerals such as feld- spar or nepheline. Weathering processes release both aluminium and gallium from these min- Sources of gallium erals and their similar geochemical properties result in the enrichment of both elements in Currently there are no mines worked primarily bauxite (Dittrich et al., 2011). However, in for gallium. However, gallium is recovered as a certain environmental situations gallium also by-product of processing bauxite or sphalerite shows an affinity to iron and becomes decou- and is found in uneconomic quantities elsewhere. pled from aluminium. Hieronymus et al. (2001) concluded that this happens where later remo- Bauxite bilisation occurs as a result of differences in Deposits of bauxite are residual deposits, that solubility between the two elements. In these is, they are formed by the weathering of pre- circumstances gallium can become more con- existing rocks in tropical or sub-tropical condi- centrated in iron-rich horizons associated with tions. They are formed from a wide variety of bauxite. source rocks and their composition varies con- The average gallium content in bauxite siderably (Hill and Sehnke, 2006). The ratio of is reported to be approximately 50 ppm (Jaskula, gallium to aluminium, and therefore the 2011a) although it can vary from 10 to 160 ppm concentration of gallium, in bauxite increases (Mordberg et al., 2001 and Bhatt, 2002). The with greater intensity of weathering. Gallium production of alumina from bauxite ores using the Bayer process results in the concentration Table 7.1 Selected properties of gallium. of gallium in the liquor and this process remains the primary source of the element (Roskill, Property Value Units 2011).

Symbol Ga Atomic number 31 Sphalerite (ZnS) Atomic weight 69.72 Gallium concentrations in the zinc ore, sphal- 3 Density at 25°C 5905 kg/m erite, are known to increase as the temperature of Melting point 30 °C deposition decreases, although it can still be pre- Boiling point 2204 °C sent in intermediate and higher-temperature Hardness (Mohs scale) 1.5 Electrical resistivity at approx. 14.00 µΩ m deposit types (Stoiber, 1940 and Cook et al., 25°C 2009). Analyses from the 1940s revealed gallium Thermal conductivity 29 W/(m °C) contents in sphalerites from the Mississippi Valley province (a low-temperature deposit type)

Table 7.2 Known minerals in which gallium forms a significant part.

Mineral name Chemical formula Ga content (%) Locality

Gallite CuGaS2 35.3 Tsumeb, Namibia and Lubumbashi, DR Congo

Sohngeite Ga(OH)3 57.7 Tsumeb, Namibia

Gallobeudantite PbGa3[(AsO4),(SO4)]2(OH)6 14.6 Tsumeb, Namibia

Carnevallite Cu3GaS4 20.9 Tsumeb, Namibia 152 thomas butcher and teresa brown which were greater than from most other types of possible to extract gallium from fly ash, as yet it deposits (typically 0.01–0.1 per cent), but gallium has not proved to be commercially viable was absent from samples of European origin (Roskill, 2011). ‘Mississippi Valley type’ sphalerites. Stoiber (1940) concluded that although there is clear evi- Apex Mine, St Georges, Utah, USA dence for a correlation between gallium content Possibly the only attempt to open a mine pri- and temperature, the composition of sphalerite marily to extract gallium and germanium from each metallogenic region is distinctive. occurred in 1986/7 at the Apex Mine near the Sphalerite from barytes veins in Central Kentucky town of St Georges in Utah, USA. This histor- was also found to contain up to 0.1% gallium ical mine had been worked many times in the (Stoiber, 1940). past, principally for copper, although the deposit Whilst Moskalyk (2003) reported that analyses also contained small quantities of zinc and sil- of USA zinc deposits have indicated a typical ver (Dutrizac et al., 1986). It was described by content of 50 ppm gallium, earlier work by Hall Bernstein (1986) as having quantities of gallium and Heyl (1968) found gallium contents of sphal- contained in jarosite (a sulfate mineral formed erites from the USA to be erratic, with their anal- from the oxidation of iron sulfides) and limo- ysis ranging from 10 to 320 ppm. Analyses by nite (a mixture of iron oxides and hydroxides). Cook et al. (2009) found gallium contents of up to It did not prove to be economic and the oper- 366 ppm in a few samples of sphalerite from ator went into liquidation towards the end of south-east European epithermal deposits and up 1987 (Kramer, 1988). A subsequent attempt to to 273 ppm from a similar deposit in Japan, re-open the mine also proved to be uneco- although they concluded that values in excess of nomic. 100 ppm are rare. The processes used to extract zinc provide a secondary source of gallium although the actual Recovery methods and refining recovery remains low (less than one per cent) (Roskill, 2011). Primary recovery Most gallium today is produced as a minor Other geological settings by-product of alumina/aluminium production Gallium also occurs in other aluminosilicate from bauxite ores. To a much lesser extent, the minerals, geothermal fields in volcanic zones (for metal is produced as a by-product of the smelting example in Taupo, New Zealand (Christie and and refining of zinc. In Russia, small amounts of Brathwaite, 2002)), in association with phosphate gallium have also been produced in conjunction ores and some coal deposits (Moskalyk, 2003). with the mining of both apatite-nepheline and Rytuba et al. (2003), with reference to two nepheline-syenite. particular deposits in the USA, noted that gallium Whilst a number of exploration companies can be enriched by hydrothermal processes in have found potentially promising deposits of zones of advanced argillic alteration in high gallium minerals, so far none has been worked sulfidation epithermal precious metal deposits; to produce gallium as a primary product. Despite although only one of the samples analysed the price of gallium having hit relatively ele- contained >100 ppm gallium. vated levels over the past couple of years, the Gallium concentrations of 149–320 ppm metal has yet to be discovered in economically have been reported in fly ash from the recoverable concentrations. One of the most combustion of coal (Font et al., 2007). Whilst promising such deposits, the Cordero Property in research has demonstrated that it is technically Humboldt County, Nevada in the USA (owned Gallium 153 by Gold Canyon Resources Inc.) remains unde- Non-proprietary extraction processes include: veloped. Research into other, potentially prom- The de la Bretèque process: since this process ising, sources of the metal, for example, fly ash extracts the gallium (by direct electrolysis) using (Font et al., 2007), waste from elemental mercury as a cathode, it is not very popular. phosphorous plants (Degerstrom, 2011), and The Bielfeld and Laspeyres process: similar to claystone (Orbite Aluminae, 2011) has not yet the de la Bretèque process, however, mercury loss led to any commercially attractive, or exploited, is improved through the use of a sodium–mer- alternatives. cury amalgam. However, over the last couple of decades, The Beja process: this process uses carbon companies have continued to improve gallium dioxide to produce a gallium-bearing precipitate production either by upgrading existing methods from which the gallium is then extracted. of extracting the metal from Bayer liquors, or by The Frary and Pechiney processes: in both developing their own, proprietary, or using these processes, the gallium is recovered through others’, technologies (Moss et al., 2011). Figure 7.1 electrodeposition. outlines the processes involved in schematic form. From zinc processing Once a gallium-bearing zinc ore has been leached From alumina/aluminium processing with sulfuric acid to produce a zinc sulfate solu- Over 95 per cent of all mined bauxite is con- tion (akin to Bayer liquor), and before electrolysis verted into alumina using the Bayer process to extract the metal itself, the impurities, which (Cardarelli, 2008), with an estimated 95 per cent include gallium, are removed through the of all the alumina so produced being smelted addition of antimony trioxide, zinc dust or pro- and refined to aluminium metal, mainly using prietary reagents. The gallium is then extracted the Hall–Héroult process. In the Bayer process, from the resulting separated solids or ‘cement the reactions of aluminium trihydroxide and residues’. aluminium oxide hydroxide with aqueous caustic soda form the so-called ‘Bayer liquor’ (a Secondary recovery sodium aluminate solution). This solution, with equilibrium concentrations of 100–125 Since one of the main uses of gallium is in ppm of gallium, is then recycled on a contin- compound semiconductor wafers, and their fab- uous basis. rication typically generates around 60 per cent To remove the gallium from Bayer liquor, the scrap, recovered gallium from this waste is an stream of liquid is tapped and the crude metal extremely important source of the metal. The most usually extracted using proprietary tech- wastes from the manufacture of gallium arsenide niques, employing ion-exchange resins (Srinivasa (GaAs) and gallium nitride (GaN) wafers, Rao et al., 2003), and then purified further. Both because of their purity and availability, are the the types of resin used, and their longevity, have metal’s most important secondary source. an important bearing on the cost of extracting Whilst most of the scrap materials containing the gallium from the Bayer liquor, as does the gallium are generated prior to the actual manufac- purity of the liquor itself. Extraneous and/or ture of the wafers themselves, both broken and unexpected impurities in the liquor can drasti- damaged wafers are also recycled. At present, cally reduce the life of resins and drive up costs. gallium is not recovered from post-consumer scrap, Since neither all Bayer liquors, nor all bauxites, not least because, in items such as printed circuit are equal, this is an important consideration boards, it is highly dispersed (Öko-Institut e.V. (Moss, et al., 2011). (2009). 154 thomas butcher and teresa brown

Gallium-bearing Gallium-bearing bauxite ores zinc ores

Alumina for smelting into Bayer process Acid leach Antimony trioxide, aluminium zinc dust or Gallium-bearing proprietary reagents Gallium-bearing zinc sulphate bayer liquor solution Liquid for Proprietary techniques Precipitation to electrolysis to using remove ‘impurities’ produce zinc ion exchange resins metal or Gallium-bearing non-proprietary cement residue processes (see text)

Separation ‘New’ scrap recycled from manufacturing 3N gallium processes

Electrolytic refining or washing with hydrochloric acid

4N gallium

Zone refining (growing of crystals)

6N or 7N gallium

Figure 7.1 Schematic diagram of the processes involved in extracting gallium.

From compound semiconductor manufacture, ● residues in GaAs single crystal reactors; the main sources of gallium scrap include (Burke, ● substandard GaAs ingots. 2011): The gallium content in such scrap can be bet- ● boule ends and chunks; ween one per cent (for example, from wastewater ● broken wafers – from all stages of fabrication; filters) and 99.99 per cent (frozen metal) (Kramer, ● cutting and polishing sludges; 2002), depending at which stage in the fabrication ● defective devices; process the scrap is generated. ● dust and residues in wastewater filters; Typically the recovery and recycling of gallium ● epitaxial growth wastes; is economically viable from throughout the pro- ● kerf generated during wafer slicing; duction process of gallium-containing compound ● liquids following etching; semiconductors. However, many of the newer, Gallium 155 and currently burgeoning, uses of gallium provide Gallium in GaAs semiconductors few opportunities for recycling, primarily because A brief description of how GaAs semiconductors they do not generate as much waste. There is, for are typically manufactured will serve as illustra- example, very little waste produced from the tion of how, for one end-use (and, indeed, manufacture of: currently, one of the metal’s largest single end- ● batteries and LEDs made using gallium uses), gallium ends up in the finished product. trichloride; In the simplest terms, a semiconductor consists ● magnets containing gallium metal; of a substrate on which one, or more, very thin ● phosphors containing gallium oxide. surface layers (epitaxial layers) are deposited or For certain end-uses, such as batteries, phos- ‘grown’. Compound semiconductor substrates are, phors and LEDs, there is currently little if any as their name suggests, composed of one or more recycling, particularly end-of-life recycling. different elements, in this case the metals gallium Although there is already some recycling of the and arsenic. Ultrahigh-purity single crystals of gallium used in the solar industry, in CIGS GaAs are first grown and then sliced into wafers (copper indium gallium selenium) technology in (the basis of electrical components) upon which the particular, it is likely to both increase in volume requisite epilayers are grown. Different end-uses and improve in effectiveness over time. Since require different qualities of substrate crystal, with CIGS cells are manufactured with longevity in integrated circuits (ICs) and microwave devices mind, it will, of course, be some time before a requiring the highest quality. large volume of material becomes available for end-of-life recycling. Substrate Three of the most common manufacturing Refining and purification technologies for growing single crystals of GaAs Except for its use in chemicals, where it is usu- substrate are: ally of 4N–5N (99.99–99.999 per cent) purity, and ● Vertical Gradient Freeze (VGF) - including in solar cells (5N), for most of its other end-uses, vertical Bridgman (VB); gallium must be of at least 6N (99.9999 per cent) ● Liquid Encapsulated Czochralski (LEC); through to 7N (99.99999 per cent) purity. The ● Czochralski (CZ). gallium (as does the arsenic) used in optoelec- A fourth method, the horizontal (as opposed to tronic applications usually has to be of at least vertical) Bridgman method (HB) is also used, for 6N purity, and in electronic applications the example, by Sumitomo Electric in Japan. These gallium needs to be of 7N purity. methods are shown diagrammatically in Figure 7.2. Since the gallium from Bayer liquors and The LEC and VGF methods are currently the most scrap will usually only have a purity of up to commonly used, with the various VGF technol- 3N (99.9 per cent), the metal needs to be further ogies dominating the market. All these methods refined. Either electrolytic refining, or washing produce cylindrical single-crystal ingots, or boules, with hydrochloric acid will bring the metal which can be of varying diameters – most usually up to 4N purity. Thereafter, in order to increase two inches, four inches, six inches and eight inches. it to the purities of 6N or 7N, the gallium can, typically, be purified by fractional crystal- VGF: In the VGF method, a crucible containing lisation. In this process melted gallium is chunks of GaAs, together with a seed crystal, seeded and increasingly pure crystals grown is placed vertically in a furnace and, whilst (since impurities tend to remain in the liquid remaining static, a temperature gradient is moved phase they cannot contaminate the growing up the length of the crystal away from the seed. crystal). The single crystal propagates from the seed 156 thomas butcher and teresa brown

Crucible Crucible

Melt Crucible Seed Crystal Melt Crystal Melt

Furnace Crystal Furnace Seed Furnace (c) Horizontal Bridgman (HB) method

(a) Vertical Gradient Freeze (VGF) method seed (b) Vertical Bridgman (VB) method

Seed Seed

Boron Crystal Crystal trioxide

Melt Melt Furnace Furnace Crucible Crucible

(d) Liquid Encapsulated (e) Czochralski (CZ) method Czochralski (LEC) method

Figure 7.2 An illustration of most common manufacturing technologies for growing single crystals of GaAs substrate. crystal in the bottom of the crucible. In contrast, molten layer of ‘encapsulating’ boron trioxide, and in the Bridgman processes, the actual crucible is slowly withdrawn, bringing with it a layer of the moved either vertically (VB) or horizontally (HB) melt, which, as it cools, assumes the crystalline relative to the furnace. structure of the seed. LEC: In the Liquid Encapsulated Czochralski CZ: In this method, since there is no encapsulation, method, named after the Polish chemist Jan the crystal is drawn directly from the melt. Czochralski, either elemental gallium and arsenic, or polycrystalline GaAs chunks, are placed in a Epitaxial layers crucible together with a pellet of boron trioxide, set inside a high-pressure vessel and heated. A small A number of different technologies, classified crystal is then dipped in the melt, through the now according to the phase (for example, liquid or Gallium 157 vapour) of material used to form the epitaxial ● Magnets: the efficiency of neodymium–iron– layer, can be used to lay down or grow the GaAs boron (NeFeB) magnets is increased through the layers on the substrate. Some of the main methods addition of a small amount of gallium. However, of epitaxial growth are: efficiency is not the primary reason for its use in ● MBE: molecular beam epitaxy; NeFeB magnets. Gallium actually helps increase ● MOVPE: metal organic vapour phase epitaxy; the fluidity of the alloy during the production ● LPE: liquid phase epitaxy. process. It acts as a molten lubricant for the hot In MBE, which takes place in a vacuum, forming process. extremely high purity gallium (together with ● Molecular Beam Epitaxy: gallium metal, at equally pure arsenic) is heated and the gases of purities of 7N and, sometimes, above, is used in the two elements are ‘beamed’ at the substrate to the MBE process to grow epitaxial layers of the condense (a physical deposition) on its surface, metal in the production of semiconductors. leaving a layer of gallium arsenide. ● Nuclear weapon pits: alloyed with plutonium, In MOVPE (also commonly referred to in the gallium was used in the pits (the cores of implo- industry as MOCVD – metal organic chemical sion weapons) of the first nuclear weapons. vapour deposition), the substrate is heated and a ● Thermometers: alloyed with indium and tin in, carrier gas ‘transporting’ precursors, for example for example, Galinstan, gallium provides a non- trimethyl-gallium (a chemical compound of toxic substitute to mercury in thermometers. gallium) and arsine, reacts chemically with the sur- ● Thin-film deposition: metallic gallium of 4N face of the substrate, leaving atoms of gallium and purity is often used as the source material for thin arsenic behind on its surface. As opposed to MBE, film deposition. For example, by thermal evapo- this is a chemical deposition of the two elements. ration in the manufacture of solar cells. In LPE, the epitaxial layer forms after a saturated solution is placed in contact with the polished sur- Gallium antimonide face of the substrate. Like gallium arsenide, gallium antimonide (GaSb) is a semiconducting compound (Dutta et al., Specifications and uses 1997). It is used in both electronic and optoelec- tronic devices and, in particular, those func- Gallium is used in a number of different forms. tioning in the infrared. Some of the more The most common are: important (and interesting) include: ● ● gallium metal; forward-looking infrared (FLIR) systems for ● gallium antimonide; night-time navigation; ● ● gallium arsenide; LEDs; ● ● gallium chemicals; missile homing guidance systems; ● ● gallium nitride; thermal imaging; ● ● gallium phosphide. high-speed electronic circuits (Bennett et al., 2005). Gallium metal Gallium arsenide As a pure metal, or alloyed with other metals, gallium is used in a number of applications: Currently, the greatest consumption of gallium ● Eutectic alloys: alloyed with any, or all, of is in gallium arsenide (GaAs) compound semi- indium, selenium and zinc, gallium, with its low conductors. Compound semiconductors, in melting point, is used in devices ranging from particular GaAs semiconductors, can provide fluid unions and heavy current switches, to a number of advantages over other semicon- rocker switches and sphygmomanometers (blood ducting materials, for example, silicon. In ICs, pressure meters). for example, GaAs is significantly more 158 thomas butcher and teresa brown

Table 7.3 Electron mobility in selected semiconductor Cell phones materials, measured in centimetres squared per volt second. (Adapted from Tummala and Morris, 2001.) Probably the most important ICs in a cell phone are the power amplifiers (PAs). The PAs in a cell Semiconductor Mobility (cm2/V-s) phone are the vital components that amplify sig- nals, both voice and data, to the appropriate Silicon (Si) 1500 Germanium (Ge) 3900 power level for them to be transmitted back to Gallium Arsenide (GaAs) 8500 the network base-station. The more advanced the Gallium Antimonide (GaSb) 4000 generation used by the handset, the more PAs it Indium Phosphide (InP) 4600 needs. Whilst 2G handsets contain a single PA, Indium Arsenide (InAs) 33,000 3G handsets can contain up to five PAs. Currently, Indium Antimonide (InSb) 78,000 the vast majority of PAs in cell phones are made using GaAs. It is also used in both the phone’s switch and filter modules. Table 7.4 Electron speed in selected semiconductor materials. (Adapted from Tummala and Morris, 2001.) Military applications Semiconductor Speed (kilometres per hour) GaAs is used in a number of different military applications, for example, in communications, Indium Antimonide (InSb) 3000 Gallium Arsenide (GaAs) 300 night vision, radar and satellite. Silicon Germanium (SiGe) 54–1400 In communications, GaAs is employed in a Indium Phosphide (InP) 180 number of different contexts. In fibre optics, Germanium (Ge) 150 GaAs components, for example sensors, are used N-type Silicon 60 to facilitate the increasingly high speeds required Gallium Phosphide (GaP) 10 in fibre-optic data communications. In military wireless communications, GaAs components are used, amongst other things, in point-to-point efficient as a substrate than silicon, as shown radios, wireless networks and cellular communi- in Tables 7.3 and 7.4. Not only is it faster than cations. Communications devices using GaAs silicon (the electrons in GaAs travel faster than will usually employ the semiconductor in one (or they do in silicon), but it can also operate over sometimes both) of two different ways. In the a much wider range of temperatures. first, GaAs is used electronically in a device in Pure GaAs substrates also have the great which both the input and output are electrical, advantage of being semi-insulating, whilst for example, in a field effect transistor (FET) or an silicon substrates are semiconducting. This is analogue IC, also known as a MMIC (monolithic of particular importance as it allows the microwave integrated circuit), that functions as integration of a number of different devices on an amplifier, filter, phase shifter, frequency con- a single substrate. In addition, in contrast with verter or mixer. Alternatively, GaAs can be used silicon-based semiconductors, GaAs-based optoelectronically in a device in which the output semiconductors operate at higher breakdown or input is light rather than electrical, for voltages and generate less noise at frequencies example, an optoelectronic IC. higher than 250 megahertz. Some of the most In modern night-vision technology, the semi- important uses of GaAs are in: conductor is used specifically in an optoelec- ● cell phones; tronic context. In, for example, goggles using ● military applications; light amplification, Generation (Gen) 3 technology ● Infrared Emitting Diodes (IREDs), Laser Diodes (as opposed to thermal imaging technology), (LDs) and Laser Emitting Diodes (LEDs); GaAs is used to coat the photocathode (photo- ● wireless communications. multiplier) that converts any available light into Gallium 159 the required electrical energy to enable an for ‘touchscreens’ or multi-touch displays, image to be seen on a phosphor screen. In this IREDs are to be found in anything from PC and instance, GaAs is converting photons (light) tablet PC displays, to self-service terminals, into electrons (electrical energy) in order for the ATMs and smart phone devices. wearer to see photons (light) emitted from the LDs are an extremely efficient way of phosphor screen. converting electrical signals into optical sig- Military radars using GaAs MMICs vary in nals. In addition to their primary uses in optical shape and type, and have many purposes, storage and communications, LDs are used in including (Fisher and Bahl, 1995): bar scanners, materials processors and sensors. ● air defence; Whilst GaAs in LEDs is not used as a substrate ● altitude measurement; nor as a dye to produce many colours, the ones ● guidance (Doppler systems); it does emit are distinctive. As GaAs alone, or ● mapping (synthetic aperture); in combination with aluminium as gallium ● missile defence; aluminium arsenide (GaAlAs), it can be used ● surveillance: battlefield and long-range; very effectively in the infrared (940–850 nano- ● weather monitoring. metres). Then, as wavelength decreases, at 660 As an example of a significant current defence nanometres it emits ‘ultra-red’. Descending use, Raytheon’s enormous XBR radar, a phased further down the wavelengths, gallium arsenide array device some 85 metres (280 feet) tall, phosphide (GaAsP) emits ‘High Efficiency Red’ employs more than 45,000 of Raytheon’s own (635 nanometres), ‘Orange’ (605 nanometres), GaAs transmit/receive modules (Defence Industry and ‘Yellow’ (585 nanometres). Daily, 2011). As with radars, military satellites serve an Wireless communications array of purposes, not least surveillance, map- ping, navigation and communications. All sat- In addition to their use in military applications, ellites need power and this can conveniently and cell phones, GaAs MMICs (with such devices be provided by solar cells. Because of its ability in each application operating in specific fre- to operate over a wider temperature range than quency ranges), are used in: silicon, and the fact that it has much higher ● broadband (high speed) satellite services; radiation hardness, GaAs is ideal for use in the ● electronic toll collection systems; solar cells that power military satellites. GaAs- ● Global Positioning Systems (GPS); based solar cells are highly efficient (Savage, ● satellite TV; 2011). However, one of their main drawbacks ● WiMAX; has been their cost and hence their lack of ● wireless LAN (WiFi, Bluetooth, etc.). general commercial exploitation, except in Finally, electronic devices employing GaAs are anything other than aerospace applications, also found in commercial radars, ranging in size especially satellites. from those used in air traffic control and weather monitoring, to those needed in smart cruise con- trols and advanced collision warning systems for IREDs, LDs and LEDs motor cars. In each of IREDs, LEDs and LDs, GaAs is used in an optoelectronic capacity, because of its Gallium chemicals ability to convert electrical input into light output. Functioning in the infrared part of the The most important gallium chemicals currently electromagnetic spectrum (at wavelengths commercially manufactured are: gallium nitrate, longer than those of visible light) IREDs emit gallium trichloride, gallium trioxide, triethyl- infrared radiation. As an enabling technology gallium and trimethyl-gallium. 160 thomas butcher and teresa brown

Gallium nitrate (Ga(NO3)3) electronic devices, and some of the most inter- esting advances in its use have been in these. It is Produced using 4N purity gallium, this compound increasingly being used in military radar, cable of gallium is used in two very different areas. As TV, aerospace applications, utility grids and a pharmaceutical chemical, gallium nitrate is wireless applications such as base stations. used in the treatment of navicular syndrome in Amongst the advantages GaN offers over other horses, and arthritis and hypercalcemia (danger- semiconductors, some of the most important are: ously high levels of calcium in the blood) in ● durability; humans. The chemical is also used in the manu- ● greater electrostatic discharge resistance; facture of catalysts for the petrochemical industry ● greater power densities; and in the production of styrene. ● greater power-added efficiency. In addition, it offers: Gallium trichloride (GaCl ) 3 ● higher current capacity; Typically produced with 4–5N purity gallium, ● higher operating voltages; gallium trichloride is used as a precursor in the ● inherently higher breakdown voltage; production of organometallic gallium compounds, ● linearity; for example, trimethyl- and triethyl-gallium. ● wide operating bandwidth. However, as a compound itself, it reduces passiv- Despite all these advantages, one characteristic ation (where a layer of oxide forms on a surface) of GaN poses significant problems. It is difficult and also improves ion transportation. Because of to make. Whereas GaAs is composed of two these two characteristics, gallium trichloride is metals, GaN is formed from a metal and a gas. used, in particular, as a cathode depolariser in Epitaxial layers of GaN have, for some time, been lithium thionyl chloride batteries. grown, using MBE and MOVPE, on a variety of substrates, for example, glass, sapphire (alumina), silicon carbide and silicon itself. However, Gallium trioxide (Ga2O3) combining the two elements to produce bulk Gallium trioxide is used as an intermediary, and GaN as a substrate has proved considerably more in the manufacture of langasite (lanthanum challenging. At present there are three main gallium silicate) crystals for communications and commercial technologies for growing bulk GaN piezoelectrics, and both LED and PDP phosphors. crystal: ● Ammonothermal: using both high tempera- Triethyl-gallium (TEGa, Ga(C2H5)3) and tures and pressures and an ammonia solution, Trimethyl-gallium (TMGa, Ga(CH3)3) this method uses liquid phase growth to ‘create’ These two organometallic compounds of gallium crystals are used as precursors for the MOVPE growth ● Hydride Vapour Phase Epitaxy (HVPE): GaN of epitaxial layers, although trimethyl-gallium is crystal is grown using existing vapour phase tech- the most commonly used (SAFC, 2011). In nology particular, trimethyl-gallium is used in the depo- ● High Nitrogen Pressure Solution (HNPS): a sition of the metal in thin-film contexts, espe- method pioneered by Polish company TopGan. cially in the production of LEDs. The two main uses of GaN in optoelectronic devices are primarily in LEDs and, to a much lesser extent, in laser diodes. Gallium nitride As a semiconductor, gallium nitride (GaN) has, Laser diodes since the end of the last century, been widely used in both LEDs and LDs. However, it is also GaN is used in the production of both ultraviolet ideal for use in high-frequency and high-power (UV) and blue laser diodes. Probably the most Gallium 161 popular use of such lasers is in Blu-ray devices. In ultraviolet. Although, as with green LDs, creating addition, GaN is a strong contender in the race low-cost, high-performance green LEDs is proving to produce direct emission green laser diodes. a continuing challenge (Rensselaer, 2011). The creation of such diodes poses a significant In general, apart from using considerably less challenge, as they require the use of extremely energy than conventional incandescent electric pure semiconductor material. GaN may now lamps and not containing any of the toxic heavy provide a solution. metals found in compact fluorescent light (CFL) bulbs, LEDs (in particular high brightness (HB) white LEDs) provide the following advantages: LEDs ● durable; In the last couple of years, the use of LEDs in ● fully dimmable; general, and GaN-based LEDs in particular, has ● instantly on/off; exploded. In 2010 alone, the GaN LED market ● colour saturation; grew by 60 per cent according to IMS Research ● longevity (up to 100,000 hours) (Focus Digital, (2011a). GaN-based LEDs have become the LEDs 2011). of choice. Not least because cheap, efficient, Although, currently, more than 80 per cent of white-light LEDs can be made using LEDs emit- LEDs are produced using a sapphire substrate ting blue light (which is possible with GaN) (Öko-Institut e.V, (2009), there are various combined with phosphors as illustrated in research efforts to find alternatives, two of which Figure 7.3. Whilst producing white light from a are glass and silicon. stack of red, green and blue LEDs is theoretically Whilst LEDs are used in a very wide range of the most efficient, green LEDs remain inefficient applications, their principal uses today are in: and therefore white LEDs using this method televisions and monitors, mobile appliances and are more expensive than using a blue LED with solid-state lighting (SSL). Other significant uses phosphors (USDOE, 2011). are in signals (for example, traffic lights) and However, apart from producing white light, motor vehicles. commercially available GaN-based LEDs now For Thin-Film Transistor (TFT) Liquid Crystal cover the spectrum from green, through blue to Display (LCD) televisions and monitors to work,

White light White light

Yellow/green and red phosphors Blue GaN LED Blue light

Blue LED Green GaN LED Substrate

Figure 7.3 Two ways to produce white Red GaAs LED light from GaN LEDs (DenBaars, 2008). 162 thomas butcher and teresa brown there has to be a light source. GaN LEDs are now environments, in the area of power conversion in the light source of choice, replacing cold cathode electric and/or hybrid vehicles (Davis, 2009), and fluorescent lamps (CCFLs). There can be anything the power grid (Bradbury, 2011). In these contexts from hundreds to thousands of individual LEDs GaN can also reduce power loss (EE Times-Asia, in each backlit television. LED televisions and 2011) and improve switching speeds (Briere, monitors are, literally, backlit (with the LEDs 2011). (coloured) directly behind the flat panel of the Perhaps cognizant of the importance of such screen) or side-lit (with the LEDs (white) behind uses of GaN, in November 2011 the Berlin-based the plane of the screen, but ranged around its Ferdinand-Braun-Institut announced the launch perimeter). LEDs are, similarly, also used in of the EU project HiPoSwitch (Ferdinand-Braun- smaller, more portable, devices such as netbooks, Institut, 2011), which will receive significant notebooks, tablet computers, Personal Digital funding from the European Community. This Assistants (PDAs), mobile gaming devices and project will focus on ‘novel GaN-based transis- some mobile phones. tors’ as ‘key switching devices’ in these areas As a major consumer of LEDs, the SSL market amongst others. can be split into a number of segments. The most significant of these are: architectural, commercial, Gallium phosphide consumer portable (for example, torches), Gallium phosphide (GaP) alone is used in a couple industrial, outdoor and residential. of visible spectrum LEDs: ‘High Efficiency Green’ Although the predominant use of GaN is in (565 nanometres) and ‘Pure Green’ (555 nanome- optoelectronics, its use in electronic devices is tres). Compounded with GaAs it is also used in both widening and increasing. GaN is particu- red, orange and yellow LEDs (see above). When larly suitable for use in military radars, not only compounded in the form of indium gallium alu- because of its durability, but also because of its minium phosphide (InGaAlP) and used in LEDs, a considerably greater power output than, GaAs. number of the wavelengths and, hence, colours It is also capable of working across a very broad between ‘Super Red’ (633 nanometer) and ‘Pure range of frequencies from 1–100 Gigahertz Green’ can also be achieved. Compounded with (Roosevelt, 2011). Successful tests have recently indium as indium gallium phosphide (InGaP), it been completed for the U.S. Navy’s Air and is used in the manufacture of electronic devices Missile Defence Radar (AMDR) using GaN mod- such as power amplifiers for WiFi and WiMAX ules (Strategic Defence Intelligence, 2011). applications. However, GaN-based transistors can function at even higher frequencies and have been made to Photovoltaics generate THz radiation (Marso, 2010). In addition to their use in the broadband, cel- Gallium can be used in a number of different lular and WiMAX markets, GaN power transis- forms, for example as a metal or in a chemical tors are ideal for use in aerospace applications compound, in the manufacture of photovoltaic (both military and commercial). Because of its (PV) or solar cells. Whilst GaAs-based solar cells particular properties, GaN is eminently more are preferred in extraterrestrial applications and suitable than either GaAs or silicon to withstand are increasingly being used in conjunction with the high-level radiation environment of space. As concentrator photovoltaics (CPV) technology, the with its use in military radars, it also provides vast majority of solar cells for terrestrial use con- significantly improved power output (Barnes tinue to employ crystalline silicon (c-Si, both et al., 2011). mono- and multi-crystalline) technology. In 2010, Finally, because of both its inherently higher this technology accounted for some 86.5 per cent breakdown voltage and its current capacity, the of the market (Solarbuzz, 2011). The remaining use of GaN is especially suitable in high-voltage 13.5 per cent of the market was accounted for by Gallium 163 solar cells employing thin-film technology, or GaN is the latest development in the field of thin-film solar cells (TFSC). semiconductors and is, if anything, likely to sub- The main TFSC technologies, apart from GaAs, stitute for more silicon-based devices in future, are amorphous silicon, cadmium telluride (CdTe), particularly in high-power electronics because it copper indium selenide (CIS) and CIGS. All is more durable, can operate at greater speeds and involve the deposition of a thin film, only a few cope with higher temperatures (GaN Systems micrometres deep, of semiconducting material on Inc., 2011). various different surfaces. Putting aside the merits of each of these different technologies, and the fact that CIGS technology is still developing, the Environmental aspects quantities of gallium currently used in solar-cell production remain small. Whilst CIGS technology Gallium does not occur in nature as a pure metal. does continue to grow, until it becomes widely As a compound it occurs at trace levels in many adopted, the production of solar cells requires con- areas of the natural environment, including siderably less gallium than is currently consumed watercourses. It also occurs in very small quan- in electronic devices or LEDs. tities in the human body, where it has no proven benefit but is not harmful. Acute exposure (that is, the ingestion of large doses) to certain rare Substitution compounds, e.g. gallium (III) chloride, can result in throat irritation, breathing difficulties and As with many of the specialist metals, gallium chest pains (Lenntech, 2011). GaAs is potentially has some unique properties which make it diffi- more toxic to humans, but this is caused by the cult to substitute in certain circumstances, or arsenic not the gallium (Martin, 1993). Recent which results in poorer performance where research has shown that GaN is non-toxic and substitution occurs. For example, silicon can be potentially suitable for biomedical implant appli- used instead of GaAs in photovoltaic cells but cations (Jewett et al., 2011). the conversion efficiency is reduced from 18–22 As mentioned before, gallium is produced as a per cent to 8–15 per cent (Feneau, 2002). The use by-product from the Bayer liquors associated of gallium as a semiconductor typically results in with the production of alumina from bauxite, increased speed, lower energy consumption and from processing zinc ores and from recycling better resistance to radiation compared to silicon- scrap. The main environmental issues associated based alternatives (Feneau, 2002). with these processes result from the various Despite the advantages of using gallium, there chemicals used. Research has been undertaken, are some specialist applications where other and is continuing, to improve the efficiency of materials can be used. Germanium can be used as gallium recovery processes. Frequently, these a substitute for GaAs in some electronics. developments have associated health and envi- Organic-based liquid-crystal displays can substi- ronmental benefits, for example there has been a tute for gallium in certain LED applications. general shift away from electrolysis of Bayer Indium phosphide can be used for infrared laser liquors using mercury as a cathode. diodes at a limited range of wavelengths and An important aspect to consider is the environ- helium-neon lasers compete with GaAs in visible mental benefits which can be achieved through laser diode applications (Jaskula, 2011a). products which use gallium. For example, gallium However, there is no effective substitute for is an essential component of LEDs, which are an GaAs in integrated circuits for many applications alternative to traditional incandescent or compact (Jaskula, 2011a), although silicon- or silicon– fluorescent electric lamps. It has been estimated germanium-based alternatives have been used in that replacing these traditional light bulbs certain situations for many years. with LED-based lighting can reduce electricity 164 thomas butcher and teresa brown

consumption by between 50 per cent and 75 per Table 7.5 Estimated virgin gallium capacity and cent (Tsao, 2003 and Briere, 2010) and for this reason production in 2010 (Data from Roskill Information the LED market for lighting is likely to grow signif- Services Ltd, 2011 and company accounts.) icantly in the future. The use of gallium in CIGS Capacity (tonnes per Production PVs also has the potential to improve further the Country year) (tonnes) efficiency of this technology (Masters, 2004 and Green et al., 2011). China 141 118 Germany 32 32 Kazakhstan 25 18.7 Ukraine 15 15 World resources and production South Korea 15–20 ≥10 Russia 10 4–7 Most commercial gallium today is produced as a Hungary 8 4 by-product of aluminium processing, with a Japan 10 4 lesser amount being produced as a by-product of zinc processing. Assuming an average gallium Production in 2010 content in bauxites, the minerals from which alu- Obtaining figures for both primary and secondary minium is most usually extracted, of around 50 gallium production capacity, and actual gallium ppm (Jaskula, 2011a), the USGS estimates that production, is notoriously difficult. As with the global resources of gallium contained in many of the critical metals, not only is the market bauxite alone exceed one million tonnes, and relatively small, but its primary production and, that global zinc reserves also contain significant to a lesser extent, its secondary production, are amounts of gallium. Whilst these estimates refer restricted to just a few companies. Because of to the total gallium contents of typical ores, cur- both the nature and size of the market and their rently very little gallium is actually recovered businesses, most of these companies prefer not to from either bauxite or zinc ores. publish any data or only to publish information that is well out of date. With so little information “In 2010, only about 10% of alumina producers available, many figures can only to be estimated extract gallium as a byproduct of alumina processing. The remainder of producers find it or arrived at through discussion with market too expensive to extract the gallium and thus participants. treat gallium as an impurity in the aluminum refining process.” Primary production U.S. Department of Energy, Critical Materials China, Germany and Kazakhstan have both the Strategy, December 2010 largest capacity to produce primary gallium, and are the largest producers (Table 7.5). Based on world mine production in 2010 for bauxite of 211 million tonnes (Bray, 2011) and for China zinc, of 12 million tonnes (Tolcin, 2011), together they could have potentially yielded (based upon a China has, over the past ten years, established hypothetical 100 per cent recovery rate) some itself as, by far, the world’s largest primary 11,170 tonnes of gallium or more (bauxite: 10,550 gallium producer. Production is currently con- tonnes and zinc: 620 tonnes or more). However, trolled by four major concerns. In descending with primary gallium production in 2010 esti- order of estimated production these are: the mated to be 106 tonnes (Jaskula, 2011a), 196 Aluminum Corporation of China Limited tonnes (Neo Material Technologies Inc., 2011) or (Chalco); Zhuhai Fangyuan (owned by Golden 209 tonnes (Table 7.5), this indicates a recovery Harvest); East Hope Mianchi Gallium Industry rate in the range of 1.00–1.98 per cent. Co; and China Crystal Technologies Co Ltd. Gallium 165

Germany below) of gallium from the processing of zinc. The company publishes no figures for gallium The Ingal Stade primary gallium production production. plant in Stade, north-west of Hamburg, and on the same site as the Aluminium Oxid Stade Russia: Formerly owned by SUAL, and now (AOS) alumina refinery, is jointly owned 50:50 owned by Mr Oleg Deripaska’s Basel Cement, the by 5N Plus (formerly by MCP) and Neo Material Pikalevo alumina refinery, in the eponymous Technologies Inc. of Canada. In 2010, each of Russian industrial town, was in 2010 thought to the co-owners received around 16 tonnes of 4N be last and only significant producer of virgin gallium. gallium left in Russia. Kazakhstan Hungary: The Alumina Products Division of Eurasian Natural Resources Corporation PLC MAL Hungarian Aluminium Ltd extracts gallium (ENRC), quoted on the London Stock Exchange, from enriched Bayer liquor at its Ajka plant. operates the Pavlodar Alumina Refinery in northern Kazakhstan, which, in addition to being Jaan: Whilst Japan may be the world’s largest one of the world’s largest producers of alumina, is consumer of gallium, in 2010 it is thought that one of the largest producers of virgin gallium. only Dowa Electronics Materials Co Ltd actually Production in 2010 was the same as 2009: 18.702 produced any virgin metal and is shifting its tonnes. activities away from gallium and towards indium production. Other Countries Secondary production Ukraine: RUSAL’s Nikolaev Alumina Refinery on the Black Sea refines bauxite imported from There continues to be considerably more capacity Guyana and Kindia in Guinea. for recycling and refining gallium than there is for primary production (Table 7.6). However, as South Korea: Korea Zinc Company is one of with primary production, reliable figures for the few producers (another is Dowa in Japan – see both capacity and production are very difficult to

Table 7.6 Estimated secondary gallium capacity in 2010. (Data from Roskill Information Services Ltd, 2011 and authors.)

Country Number of companies Refining (U) Recycling (R) Total capacity (tonnes per year)

Japan 2 R ≤20 4 R/U ≈150 1U 60 USA/Canada 1 R/U ≤60 R ≤30 China 1 R/U 7/8N–15 and 6N–20 2U 60 Slovakia 1 R 8 U25 UK 1 R/U 20 Germany 1 R ≤10 U≤10 France 1 U N/A 166 thomas butcher and teresa brown find. In addition, differentiating between those Elsewhere around the world, gallium was amounts produced from either recycling gallium recycled and/or refined by: CMK in western metal or the waste, for example from wafer Slovakia; Mining and Chemical Products Ltd production, and those produced by upgrading (owned by 5N Plus) at their UK operation in 3/4N materials is extremely difficult. Wellinborough; PPM Pure Metals GmbH (owned Within the GaAs sector, the importance of by France’s Recylex SA) based in Germany; and recycling cannot be overemphasised. Typically, Azelis Electronics in France. some 30–40 per cent of the gallium used by the sector comes from recycled materials. With advances in recycling technology, it is likely that Future supplies there is a recovery rate of around 90 per cent from the gallium-bearing waste generated within the In the foreseeable future, China will continue to sector. be world’s largest producer of primary gallium. At In the epitaxy sector, with recovery rates also the end of 2010, it was thought that all the coun- around 90 per cent, it is estimated that, at any try’s major producers had plans to expand both one time, in the region of 60 tonnes of gallium capacity and production, and during 2011 a can be circulating in the liquid phase epitaxy number of these projects were initiated. In 2012, (LPE) ‘loop’ in Japan. the projected annual capacity of primary gallium Within the CIGS solar cell sector, and until in China alone was approximately 300 tonnes. yields improve significantly, recycling recovery Outside China, in the first such move by a rates remain around 50 per cent. major aluminium producer in many years, Rio By far the largest recycler (as well as the Tinto Alcan has signed a memorandum of under- world’s largest consumer) of gallium is Dowa standing with 5N Plus. Under this memorandum Electronics Materials Co Ltd. Other companies the two would discuss a project to extract gallium that, like Dowa, both upgrade and recycle at Rio Tinto Alcan’s Vaudreuil alumina facility in gallium include: Nichia Corporation, Rasa Quebec, Canada (5N Plus Inc., 2011). Industries Ltd and Sumitomo Metal Mining. In This announcement was significant because, 2010, two companies just refined gallium to for most major alumina/aluminium producers, higher purities: Furukawa Denshi Co Ltd and the risks of tapping into their Bayer liquor streams Nippon Rare Metal Inc. Sumitomo Chemical, to extract gallium are not justified by the rewards whilst undertaking no recycling in 2010, remains from doing so. The concentrations of gallium one the world’s largest producers of high-purity may be quite small and the capital expenditure gallium at its Ehime plant. In the USA and on an extraction unit may be significant, both of Canada, Neo Performance Materials Ltd (for- which reduce its economic viability. Furthermore, merly Recapture Metals Inc.) is the only recycler gallium remains only a by-product and any and refiner of gallium in North America. derived returns are generally small in contrast Perhaps surprisingly, as the world’s largest with those derived from processing the primary producer of primary gallium, China appears to metal. As importantly, for those with Bayer have, comparatively, very little capacity to liquor streams, the risks of disruption to the pri- recycle the metal. In 2010, the largest partici- mary production processes cannot be dismissed. pants in the Chinese secondary gallium market The same arguments can be made in relation to were: Nanjing Jin Mei Gallium Ltd (majority extracting gallium from the residues of processing owned by the US company AXT Inc.), Sumika zinc. Electronic Materials (Shanghai) Co Ltd (owned by In 2007, there were at least ten companies in Sumitomo Chemical) and MCP Jin Shu (Shenzen) the Commonwealth of Independent States with Co Ltd (now owned by 5N Plus and the country’s some capacity to produce gallium. Today, there largest producer of 6N purity gallium). are only three: one each in Kazakhstan, Russia Gallium 167 and Ukraine. It is not known whether the plant at systems record gallium trade combined with the other seven facilities has either been disman- other elements such as hafnium, indium, nio- tled or just mothballed. Some unused capacity, bium and/or rhenium. Individual countries iden- therefore, may remain elsewhere. However, tify some trade movements, but this gives only a gallium extraction may retain some attraction partial picture of the trade that takes place. because at least one company in Russia has For example, the United States Geological recently indicated it is both aware of the gallium Survey reported that gallium (unwrought, waste, contained within its apatite-nepheline ore and scrap) imports for consumption in the USA resources and its potential value (PhosAgro, amounted to 35.9 tonnes in 2009, decreasing 2011). from 41.1 tonnes in 2008. Imports to the USA of Any increase in the supply of gallium from doped GaAs wafers amounted to 117 tonnes in recycling will come from the thin-film solar sec- 2009, decreasing from 165 tonnes in 2008 (Jaskula, tor, in particular CIGS. A great deal of capacity 2011b). The countries supplying those 2009 already exists to recycle the waste from both the imports are shown in Figure 7.4. However, figures manufacture of compound semiconductor wafers, for imports that are not consumed within the and from the liquid phase epitaxy. Recovery rates USA and statistics for exports from the USA are for each are around 90 per cent. In contrast, recy- not reported. cling in the solar market is still growing and Trade in ‘unwrought gallium and powders recovery rates, while still considerably less effec- thereof’ to and from the United Kingdom in 2010 tive at around 50 per cent, are improving. is shown in Figure 7.5. Imports amounted to 73.2 Amongst those who have recently become tonnes in 2010, which is an increase of nearly 350 involved in recycling associated with CIGS, in per cent compared to the 16.4 tonnes imported in Europe, Umicore has opened a new facility to 2009. Exports of gallium in this form were 39.9 recycle production waste from the manufacture tonnes in 2010, an increase of 57 per cent from of CIGS cells at its Hoboken facility, in the 25.5 tonnes exported in 2009 (UK Trade Info, Belgium. In the USA, Indium Corporation has 2011). Details relating to the purity levels of this a recycling facility in New York State and in trade are not available. Information for gallium Japan, a number of companies are already traded in other forms cannot be separated from exploring end-of-life recycling of CIGS cells. other elements. Neo Materials also has significant capacity at its Canadian operations to treat all forms of CIG and CIGS scrap. Prices

Prices for gallium will vary depending on the World trade form and purity in which it is traded. Generally, higher-purity gallium metal will be approxi- Gallium is traded in a variety of forms including mately US$200 per kilogram higher in price than unwrought metal, powders, waste and scrap, lower-purity grades (Jaskula, 2011b). Typical or GaAs wafers (doped or undoped) and other arti- average prices are quoted in trade publications, cles containing the element. It is also traded although usually this is for lower-purity forms of at a variety of degrees of purity with some gallium metal rather than higher-purity metal or countries importing lower-grade materials for gallium compounds. processing into higher-purity forms which are Historically, gallium prices have been very then exported. stable over long periods. It is believed that this Obtaining accurate data on the worldwide general stability was a result of there being trade in gallium is extremely difficult, not least relatively few producers, supplying one main because the internationally recognised trade code substantial market (GaAs wafers), and the 168 thomas butcher and teresa brown

60

50

40

30 Tonnes

20

10

0 Japan Germany Canada United China Taiwan Singapore Italy Ukraine Others Kingdom

Unwrought, waste and scrap Doped GaAs wafers

Figure 7.4 Sources of United States of America imports of gallium in 2009. (Data from Jaskula, 2011b.)

existence of significant spare production capacity demonstrated by the typical price trend for (Mining Journal, 2010). The latter meant that 2006 to 2011 shown in Figure 7.6, which is output could readily be increased in response to based on quarterly averages. The price increase higher prices, leading to an over-supply in the in early 2007 was attributed to Chinese pro- market and the consequent fall in prices. This ducers claiming that there was a shortage in general stability was interrupted in the final supply. As these fears subsided, and as a result quarter of 2000 and first quarter of 2001 when of lower demand during the economic reces- prices for high-purity gallium metal increased to sion, the price reduced to the previous level of more than US$2000 per kilogram as a result of an approximately US$400 per kilogram. During apparent severe shortage in supply. However, this the first half of 2011, prices rose again to more shortage was artificial and was driven by errone- than US$900 per kilogram before reducing to ously high future consumption projections, which below US$700 per kilogram by November 2011. led to hoarding and speculative inventories devel- A price rise was seen in many of the minor oping. These high prices were short lived and by metals, due in part to speculative buying, but July 2001, once forecasts had been re-evaluated, for most of these prices decreased over the prices returned to more normal levels (Kramer, summer as the quantity of trade reduced (Metal 2000 and 2001). Prices remained in the range Bulletin, 2011a). In the case of gallium, prices US$300 to US$400 from 2001 to 2006. continued to fall as increases in capacity were Since 2007 there has been an unusual expected (Metal Bulletin, 2011b) and demand amount of movement in the price of gallium, as remained weak (Metal Bulletin, 2011c). China (inc Hong Kong)

Germany

Ukraine

South Korea

Netherlands

Russia

Japan

United States of America

Others

–30 –20 –10 0 10 20 30 40 50 Tonnes 2010 Imports 2010 Exports

Figure 7.5 Imports and exports of unwrought gallium and gallium powders to and from the United Kingdom in 2010. (Data from UK Trade Info, 2011.)

1000

900

800

700

600

500

400 US$ per kilogram 300

200

100

0 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2Q3 Q4 Q1 Q2 Q3 Q4 2006 2007 2008 2009 2010 2011

Figure 7.6 Quarterly average prices for gallium metal from January 2006 to November 2011. (Calculated from data quoted in Mining Journal, 2006–2011 and Metal Bulletin, 2011d.) 170 thomas butcher and teresa brown

Outlook In the USA, a Defence Advanced Research Projects Agency (DARPA) programme is devel- The estimated demand for gallium in 2010 can oping the semiconductor’s use at ever-higher be divided between its various end-use sectors, frequencies, perhaps even as high as the Terahertz as shown in Figure 7.7, and it is developments bands (DARPA, 2011 and Majumdar, 2011). within these sectors which will determine the However, the use of GaN-based devices is not overall demand for gallium in the future. now confined only to the defence sector. They are Gallium continues to demonstrate its versa- increasingly being used in civilian contexts tility in a wide range of applications. Improvements (Bindra, 2010), for example in CATV, satellites continue to be made in the performance of and wireless infrastructure (ElectroIQ, 2011). As devices in which it is already used, and the the use of GaN further improves power elec- number of devices containing gallium is expand- tronics, in, for example, high-voltage, high-power ing. This expansion includes both its use for switches (ARPA-E, 2010), in terms of energy existing purposes and also for new applications. efficiency alone there is every possibility that its GaN-based defence products, for example potential in a commercial context as a ‘green’ power amplifiers, continue to improve, amongst material will come closer to realisation. other things, in terms of reliability, power output GaAs has for some time been used in a photo- and absence of distortion (Majumdar, 2011). voltaic capacity only in quite specialised circum- stances and environments, e.g. space. However, the increasing interest in concentrator photovoltaics 2% (CPV) technology, in which sunlight is concentrated 6% on solar cells, could indicate a promising area of 4% growth for terrestrial applications. Indeed, the research firm Strategy Analytics believes that, by 6% 2016, the CPV market will account for five per 50% cent of new photovoltaic installations (Strategy Analytics, 2011a). Furthermore, the CPV market 8% could be provided with a considerable boost if solar cells with the new record conversion efficiency of 28.4 per cent (Compound Semiconductor, 2011a), achieved by a USA company in 2011, can be suc- cessfully commercialised. Two new applications for gallium are in coat- ings for LEDs and in transparent conducting oxides (TCOs) for LCD displays. In the former, 24% gallium is substituted for some, or all, of the aluminium in yttrium aluminium garnet (YAG) coatings. In the latter, gallium, as a constituent of an indium, gallium, zinc oxide (InGaZnO (IGZO)) SI (ICs) SC (LEDs, LDs Etc) PV (CIGS) semiconductor has already been used in small- Alloys Magnets Others and medium-sized LCD panels developed by Sharp (Sharp, 2011) and in 3D television screens Catalysts by Samsung (Bourzac, 2011). Further development of technologies using Figure 7.7 Estimated demand for gallium in 2010. gallium-doped zinc oxide (GZO), both in LEDs Notes: SI, semi-insulating; SC, semiconducting; PV, (Liu et al., 2011) and as a substitute electrode photovoltaic. material (both as anode and cathode) for the more Gallium 171 costly indium–tin oxide (ITO), could lead to con- both more environmentally friendly than CdTe siderable expansion of the metal’s use in, for (cadmium telluride) technology, the current example, large area displays (Wang et al., 2010), market leader, and more efficient than silicon solar cells (Ihn et al., 2010) and solid-state lighting. technology, CIGS technology is well placed to As technological advances mean gallium is become a significant participant in the PV used in more applications, so the manufacturing market. If it does, then demand for gallium processes that enable it to be used also advance. from this sector would rise accordingly. For example, in the field of GaN transistors, However, it is difficult to make reliable fore- researchers announced in 2011 that they had, for casts for growth of demand for gallium from the the first time, succeeded in producing high-speed CIGS sector of the solar-cell market as the transistors containing GaN, but grown on a industry is new and still evolving, both com- silicon wafer (Daimler, 2011), thereby offering mercially and technologically. both speed and functionality at high tempera- Whilst announcements of involvement in tures. In the field of GaN LEDs, Samsung of Korea CIGS cell manufacturing by organisations such announced in 2011 that “one of its research teams as of Hyundai Heavy Industries and Saint had figured out a way to grow crystalline gallium Gobain, Samsung, and Taiwan Semiconductor nitride (GaN) LEDs on regular glass” (Yirka, 2011). Manufacturing Corp (TSMC) provide advocates Whilst these are some of the more interesting with a degree of comfort (TSMC, 2011), announce- ways in which gallium may be used in future, ments like that of Solyndra’s bankruptcy demand for the metal is currently driven pri- (amongst others) in the USA (Church et al., 2011) marily by its use in cell-phone handsets, particu- illustrate some of the challenges the industry larly in ‘smart’ phones, and in LEDs. still faces. In mobile wireless communications, not only In trying to reach any conclusions as to future do the higher generations of the wireless standard demand for gallium (and, indeed, availability), require more power amps per handset, but as there still remains a huge unknown – purchases by handsets become more sophisticated and the both traders and speculators. There is no way of demand for data continues to grow, more GaAs discovering what these volumes are. However, components are required. It is estimated that, these participants in the market should always be in 2014, 1.7 billion handsets will be shipped considered, especially as ‘investors’ become increas- (Higham, 2011). In 2010, there was a 30 per cent ingly interested in a wider range of metals. growth in the demand for GaAs epitaxial wafers However, we can be confident that, with those alone (Strategy Analytics, 2011b). resources of gallium that have hitherto been iden- There continues to be significant demand for tified, there is more than sufficient to supply all LEDs for backlighting in televisions, computer the developing technologies for the foreseeable screens, etc., and the GaN LED market grew future. Moreover, since gallium is available from around 60 per cent in 2010 (IMS Research, 2011b; the alumina process and the plants undertaking Compound Semiconductor, 2011b). The use of this process are located in a diverse range of LEDs both in general and street solid-state geographical locations, there continues to be lighting, because of their durability, energy potential to develop gallium production facilities savings and longevity, will become increasingly in alternative areas. important in future (Figure 7.8). After its use in wireless devices and LEDs, the third largest demand for gallium comes Acknowledgements from its use in solar cells, in particular those using CIGS (as opposed to GaAs) technology. Teresa Brown publishes with the permission of However, in 2010 its use in solar cells con- the Executive Director of the British Geological sumes comparatively little of the metal. Being Survey. 172 thomas butcher and teresa brown

14

12

10

8

6 US$ dollars billions

4

2

0 2009 2010 2011 2012 2013 2014 2015 2016

TVs Monitors Mobile phones NBs Other BLUs Automotive

Lighting Signage Other Off-spec

Figure 7.8 Forecast of packaged LEDs by application 2009–2016, revenue by segment. (Data from IMS Research, 2011c.) Notes: TVs, televisions; NBs, notebooks; BLUs, back-lit units.

References Bennett, B.R., Magno, R., Boos, J.B., Kruppa, W. and Ancona, M.G. (2005) Antimonide-based compound 5N Plus Inc. (2011) 5N Plus and Rio Tinto Alcan in semiconductors for electronic devices: A review. Discussions Regarding Development of New Primary Elsevier, Solid-State Electronics 49, 1875–1895. Gallium Source, http://www.5nplus.com/uploads/ Bernstein, L.R. (1986) Geology and mineralogy of Eng_5NPlus_%20RioTintoAlcan.pdf the Apex germanium-gallium mine, Washington Advanced Research Projects Agency – ENERGY County, Utah. United States Geological Survey, (ARPA-E). (2010) Transphorm Inc: High Performance Bulletin 1577. GaN HEMT Modules for Agile Power Electronics. Bhatt, C.K. (2002) Estimation of gallium in a bauxite- http://arpa-e.energy.gov/ProgramsProjects/ADEPT/ ore deposit using an energy-dispersive X-ray fluores- HighPerformanceGaNHEMTModulesforAgilePower. cence technique. Radiation Physics and Chemistry aspx 65, 193–197. Barnes, A., Vitobello, F., Daeubler, J., Hirche, K., Laetti, Bindra, A. (2010) GaN Power Transistors Gain J. and Rostewitz, M. (2011) Europe improves reliability Ground In Military And Commercial Applications. of GaN microwave devices for space applications. MicroWaves&RF, http://mwrf.com/Articles/ArticleID/ Compound Semiconductor Volume 17, No. 7, pp. 31. 22528/22528.html Gallium 173

Bourzac, K. (2011) An Ultra-High-Definition 3-D TV. Davis, S. (2009) SiC and GaN Vie for Slice of the Technology Review. http://www.technologyreview. Electric Vehicle Pie. Power Electronics Technology. com/computing/37691/ http://powerelectronics.com/passive_components_ Bradbury, D. (2011) Transphorm rolls out EZ-GaN packaging_interconnects/resistors/sic-gan-vie-slice- switching transistor. Business Green. http://www. electric-vehicle-pie-20091101/ businessgreen.com/bg/news/2071714/transphorm- Defence Advanced Research Projects Agency (DARPA). rolls-ez-gan-switching-transistor (2011) Nitride electronic next-generation technology Bray, E.L. (2011) Bauxite and Alumina in Mineral (NEXT) program. http://www.darpa.mil/Our_Work/ Commodity Summaries. United States Geological MTO/Programs/ Nitride_Electronic_NeXt-Generation_ Survey. Technology_%28NEXT%29.aspx Briere, M.A. (2010) GaN on Si based power devices: Defence Industry Daily (2011) Floatin’ Smokey: The an opportunity to significantly impact global energy USA’s SBX Radar. http://www.defenseindustrydaily. consumption. http://www.csmantech.org/Digests/2010/ com/Floatin-Smokey-The-USAs-SBX-Radar-04804/ Papers/13.1.066.pdf Degerstrom, N.A., Inc. (2011) Metallurgical – Gallium. Briere, M. (2011) Commercial 600V GaN-Based Power http://www.nadinc.com/metallurgical/gallium. Devices Coming of Age. Power Electronics Technology. html http://powerelectronics.com/discrete-semis/gan-based- DenBaars, S. (2008) What is LED lighting: Technology power-devices-coming-of-age-0711/ overview and introduction. Presentation at the DOE Burke, E. (2011) Waste Minimization, Pollution Prevention SSL Market Introduction Workshop 2008. Solid State and Resource Recovery at a GaAs Manufacturer. Paper Lighting and Energy Centre, University of California, delivered at CS MANTECH Conference, 16–19 May Santa Barbara, USA. 2011, Palm Springs, California, USA Dittrich, T., Seifert, T. and Gutzmer, J. (2011) Gallium – Cardarelli, F. (2008) Materials Handbook: A Concise can future demand be met by geological and techno- Desktop Reference. Springer-Verlag, London, UK, logical availability? Department of Mineralogy, pp. 166. Technische Universität Bergakademie Freiberg, Christie, T. and Brathwaite, B. (2002) Mineral commodity Germany. https://sga.conference-services.net/ report 19 – beryllium, gallium, lithium, magnesium, resources/1054/2590/pdf/SGA2011_0203.pdf uranium and zirconium. Institute of Geological and Dutrizac, J.E., Jambor, J.L. and Chen, T.T. (1986) Host Nuclear Sciences Ltd, The New Zealand Ministry of minerals for the gallium-germanium ores of the Apex Economic Development. Mine, Utah. Scientific Communications in Economic Church, S., Schneider, J. and Sandler, L. (2011) Solyndra, Geology 81, 946–950. Solar-Panel Maker in California, Seeks Bankruptcy. Dutta, P. S., Bhat, H. L. and Kumar, V. (1997) The Bloomberg. http://www.bloomberg.com/news/2011- physics and technology of gallium antimonide: An 09-06/solyndra-solar-panel-maker-in-california-files- emerging optoelectronic material. Journal of Applied for-bankruptcy-protection.html# Physics 81 (9) 1 May 1997, 5822. Compound Semiconductor (2011a) Gallium arsenide solar EE Times-Asia (2011) GaN power transistors eliminate cells break SQ limits. http://compoundsemiconductor. current collapse, cut power loss. http://www.eetasia. net/csc/news-details.php?id=19734213 com/ART_8800639119_765245_NP_ ce147f95.HTM Compound Semiconductor (2011b) LED-TV explosion ElectroIQ (2011) Gallium nitride extends beyond creates a boom in the LED Market. 2 November 2011 defence products. http://www.electroiq.com/articles/ http://compoundsemiconductor.net/csc/news- sst/2011/09/gallium-nitride-extends-beyond-defense- details/id/19734184/name/LED-TV-explosion- products.html creates-a-boom-in-the-LED-market.html Feneau, C. (2002) From Ag to Zn. Umicore. http://www. Cook, N.J., Ciobanu, C.L., Pring, A. Skinner, W., umicore.com/en/aboutUs/fromAgToZn/main.html Shimizu, M., Danyushevsky, L., Saini-Eidukat, B. Ferdinand-Braun-Institut (2011) More efficiency for power and Melcher, F. (2009) Trace and minor elements electronics. Ferdinand-Braun-Institut, Berlin, Germany. in sphalerite: a LA-ICPMS study. Geochimica et http://www.fbh-berlin.com/fileadmin/downloads/Press/ Cosmochimica Acta 73, 4761–4791. PM_FBH_HiPoSwitch_Kickoff_2011_11_10_engl.pdf Daimler Technology (2011) Gallium nitride transistors. Fisher, D. and Bahl, I.J. (1995) Gallium Arsenide IC http://www.daimler-technicity.de/en/gallium- Applications Handbook, Volume 1. Academic Press, nitride-transistors/ Inc., San Diego, California, USA, pp. 81. 174 thomas butcher and teresa brown

Focus Digital Displays, LLC (2011) Why a 100,000 Hour new mineral species from Tsumeb, Namibia and LED Half-Life Rating is Important To You. http:// associated new gallium analogues of the alunite-jarosite www.focusdigitaldisplays.com/100000-hour-led- family. The Canadian Mineralogist 34, 1305–1315. halflife-rating-important Jaskula, B.W. (2011a) Gallium in Mineral Commodity Font, O., Querol, X., Juan, R. , Casado, R., Ruiz, C.R., Summaries. United States Geological Survey. López-Soler, A., Coca, P. and García Peña, F (2007) Jaskula, B.W. (2011b) Gallium in 2009 Minerals Recovery of gallium and vanadium from gasification Yearbook. United States Geological Survey. fly ash. Journal of Hazardous Materials 139 (3), Jewett, S.A., Makowski, M.S., Andrews, B., Manfra, 413–423. M.J. and Ivanisevic, A. (2011) Gallium nitride is bio- GaN Systems Inc. (2011) Why Gallium Nitride? http:// compatible and non-toxic before and after functional- www.gansystems.com/technology/why-gallium-nitride ization with peptides. Purdue University and North Geier, B.H. and Ottemann, J. (1970) New primary Carolina State University. Acta Biomaterialia. vanadium-, germanium-, gallium- and tin-minerals Kramer, D.A. (1988) Gallium. United States Bureau of from the Pb-Zn-Cu-deposit Tsumeb, South West Mines Minerals Yearbook 1988, Volume 1, 407–411. Africa. Mineralium Deposita 5, (1), 29–40. Kramer, D.A. (2000 and 2001) Gallium. In: United Green, M.A., Emery, K., Hishikawa, Y. and Warta, W. States Minerals Yearbook. United States Geological (2011) Solar cell efficiency tables (version 37). Progress Survey. in Photovoltaics: Research and Applications. Issue Kramer, D.A. (2002) Recycling – Metals. In: United 19, 84–92. States Minerals Yearbook. United States Geological Hall, W.E. and Heyl, A.V. (1968) Distribution of Survey. minor elements in ore and host rock, Illinois- Lenntech, B.V. (2011) Health effects of gallium. http:// Kentucky Fluorite District and Upper Mississippi www.lenntech.com/periodic/elements/ga.htm Valley Zinc-Lead District. Economic Geology 63, Liu, H.Y., Li, X., Liu, S. Ni, X., Wu, M., Avrutin, N., 655–670. Izyumskaya, N., Özgür, Ü., Yankovich, A.B., Kvit, Hieronymus, B., Kotschoubey, B. and Boulèque, J. (2001) A.V., Voyles, P.M. and Morkoç, H (2011) GaN based Gallium behaviour in some contrasting lateritic light emitting diodes utilizing Ga doped ZnO as a profiles from Cameroon and Brazil. Journal of highly transparent contact to p-GaN. Physica Status Geochemical Exploration 72, 147–163. Solidi C 8 (5), 1548–1551. http://onlinelibrary.wiley. Higham, E. (2011) Market trends in GaAs RF ICs. com/doi/10.1002/pssc.201000860/abstract Semiconductor Today 6 (3), April/May, 94–95. Majumdar, D. (2011) GaN Revolution. Defense News. Hill, V.G. and Sehnke, E.D. (2006) Bauxite. In: Kogel, http://www.defensenews.com/story.php?i=5819236 J.E., Trivedi, N.C., Barker, J.M. and Krukowski, S.T. Marso, M. (2010) GaN for THz sources. 8th Interna- (eds.) Industrial Minerals and Rocks, Society for tional Conference on Advanced Semiconductor Mining Metallurgy and Exploration Inc: Littleton, Devices and Microsystems (ASDAM), 25–27 October Colorado, USA. 2010, pp. 147–154. ieeexplore.ieee.org/Xplore/login. Ihn, S-G., Shin, K-S., Jin, M-J. , Bulliard, X., Yun, S., jsp?url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel Choi, Y.S., Kim, Y., Park, J-H., Sim, M., Kim, M., 5%2F5653062%2F5666309%2F05666326.pdf%3Far Cho, K., Kim, T.S., Choi, D., Choi, J-Y., Choi, W. & number%3D5666326andauthDecision=−203 Kim, S-W (2011) ITO-free inverted polymer solar cells Martin, R.B. (1993) The elements in the Environment. using a GZO cathode modified by ZnO. Solar Energy In: Downs, A.J. (ed.) Chemistry of Aluminium, Materials and Solar Cells, January 2011. Gallium, Indium and Thallium. Blackie Academic IMS Research (2011a) IMS expects LED market to stall and Professional, Glasgow, UK. in 2011. Compound Semiconductor October 2011, 17 Masters, G.M. (2004) Renewable and Efficient Electric (7) pp. 6. Power Systems. John Wiley and Sons Inc., New IMS Research (2011b) IMS Research Downgrades 2011 Jersey, USA. Packaged LED Market Growth to 1% Despite 29% Metal Bulletin (2011a) Tellurium and gallium fall, other Increase in Lighting Revenues. September 2011 minors may follow. Metal Bulletin magazine, 8 June IMS Research (2011c) Quarterly GaN LED Supply/ 2011. Demand Report. August 2011. Metal Bulletin (2011b) High-tech metals fall as China Jambor, J.L., Owens, D.R., Grice, J.D. and Feinglos, M.N. de-stocks, traders take profits. Metal Bulletin maga-

(1996) Gallobeudantite, PbGa3[(AsO4),(SO4)]2(OH)6, a zine, 16 June 2011. Gallium 175

Metal Bulletin (2011c) Indium, gallium prices fall again Savage, N. (2011) Solar Cell Breaks Efficiency Record. on weak electronics demand. Metal Bulletin maga- IEEE spectrum. http://spectrum.ieee.org/green-tech/ zine, 14 September 2011. solar/solar-cell-breaks-efficiency-record. Metal Bulletin (2011d) Gallium Metal MB free market Scott, J.D. (1971) Crystal structure of a new mineral, prices. http://www.metalbulletin.com sohngeite. In Mineralogical Notes. The American Mining Journal (2006–2011) Minor metals sections. Mineralogist 56, 355. Mining Journal magazine. Sharp (2011) Sharp to Commercialize World’s First Mining Journal (2010) Minor metals: staying strong in Small- and Medium-Size LCD Panels using Oxide summer. Mining Journal magazine, 9 July 2010. Semiconductor. http://sharp-world.com/corporate/ Mordberg, L.E., Stanley, C.J. and Germann, K. (2001) news/110421.html Mineralogy and geochemistry of trace elements in Solarbuzz (2011) Global PV Market. http://www. bauxites: the Devonian Schugorsk deposit, Russia. solarbuzz.com/facts-and-figures/market-facts/global- Mineralogical Magazine 65 (1) 81–101. pv-market Moskalyk, R.R. (2003) Gallium: the backbone of the elec- Srinivasa Rao, K., Sarangi, D., Dash, P.K. and Roy tronics industry. Minerals Engineering 16, 921–929. Chaudhury, G. (2003) Preferential extraction of Moss, R.L., Tzimas, E., Kara, H., Willis, P. and Kooroshy, gallium from Bayer liquor using ion exchange che- J. (2011) Critical Metals in Strategic Energy lating resin containing hydroxamic acid functional Technologies, European Commission Joint Research group. Journal of Chemical Technology and Centre Institute for Energy and Transport, pp. 66. Biotechnology 78 (5), 555–561. Neo Material Technologies Inc. (2011) Presentation Strategy Analytics Inc. (2011a) Strategy Analytics: given at MMTA’s International Minor Metals Accelerating CPV Penetration of Renewable Energy Conference, 4–6 May 2011, Philadelphia, PA, USA. Markets. http://www.businesswire.com/news/home/ Öko-Institut e.V. (2009) Critical Metals for Future 20110912006520/en/Strategy-Analytics-Accelerating- Sustainable Technologies and their Recycling CPV-Penetration-Renewable-Energy Potential. United Nations Environment Programme Strategy Analytics Inc. (2011b) Strategy Analytics: Brisk and United Nations University, pp. 38. 2010 GaAs Epitaxial Substrate Revenue Growth. Orbite Aluminae Inc. (2011) Products and Markets. http://www.businesswire.com/news/home/ http://www.orbitealuminae.com/en/products/ 20111031006235/en/Strategy-Analytics-Brisk-2010- products-markets/ GaAs-Epitaxial-Substrate PhosAgro (2011) Presentation given for the non-deal Stoiber, R.E. (1940) Minor elements in sphalerite. roadshow. VTB Capital Conference “Russia calling!”, Economic Geology 1 June 1940, 35 (4), 501–519. Moscow, October 2011. http://www.phosagro.com/ Strategic Defence Intelligence (2011) US Navy’s AMDR upload/1393/PhosAgro_Russia_Calling_2011_NDR.pdf programme moves ahead. http://www.strategicdefen- Rensselaer Polytechnic Institute (2011) Energy@ ceintelligence.com/article/CPpUgH6rN66/2011/09/20/ Rensselaer: Zeroing in on the Elusive Green LED. us_navys_amdr_programme_moves_ahead/ News and Events, 25 April 2011. http://news.rpi.edu/ Tolcin, A.C. (2011) Zinc in Mineral Commodity update.do?artcenterkey=2860 Summaries. United States Geological Survey. Roosevelt, A. (2011) Raytheon’s GaN Technology Tsao, J. (2003) Roadmap projects significant LED pene- Advances Radar. Defense Daily 249, (31) 1. tration of lighting market by 2010. Laser Focus World. Roskill (2011) Gallium: Global Industry Markets and http://lighting.sandia.gov/lightingdocs/ Outlook, 8th Edition. Roskill Information Services TsaoJYLaserFocus200303.pdf Ltd., London, UK. TSMC Solar Ltd. (2011) TSMC Solar readies CIGS PV Rytuba, J.J., John, D.A., Foster, A., Ludington, S.D. and fab for 1GW future. http://www.electroiq.com/arti- Kotlyar, B. (2003) Hydrothermal enrichment of cles/pvw/2011/10/tsmc-solar-readies-cigs-pv-fab- gallium in zones of advanced argillic alteration – for-1gw-future.html examples from the Paradise Peak and McDermitt ore Tummala, R.R. and Morris, J.E. (2001) The Role of deposits, Nevada. United States Geological Survey Packaging in Microelectronics in Fundamentals of Bulletin 2209-C. http://pubs.usgs.gov/bul/b2209-c/ Microsystems Packaging, McGraw-Hill, 44–79. SAFC (2011) Triethylgallium – TEGa. http://www. UK Trade Info. (2011) UK Trade Information. Revenue safcglobal.com/safc-hitech/en-us/home/overview/ and Customs. Beyond 20/20 Web Data Server, Version products/gallium/triethylgallium.html. 7.4, SP1. https://www.uktradeinfo.com/index.cfm HM 176 thomas butcher and teresa brown

U.S. Department of Energy (USDOE) (2011) LED Basics. Physics. http://jap.aip.org/resource/1/japiau/v107/i4/ U.S. Department of Energy – Energy Efficiency and p043103_s1?isAuthorized=no Renewable Energy. http://www1.eere.energy.gov/ Weeks, R.A. (1989) Gallium, Germanium and Indium buildings/ssl/sslbasics_ledbasics.html Resources. In: Carr, D.D. and Herz, N. (eds.) Concise Wang, L., Matson, D.W., Polikarpov, E., Swensen, J.S., Encyclopedia of Mineral Resources, Pergamon Press, Bonham, C.C., Cosimbescu, L., Berry, J.J., Ginley, Oxford, UK. pp. 133. D.S., Gaspar, D.J. and Padmaperuma, A.B. (2010) Yirka, B. (2011) Samsung researchers announce break- Highly efficient blue organic light emitting device through in growing gallium nitride LEDs on glass. using indium-free transparent anode Ga:ZnO with PhysOrg.com. http://www.physorg.com/news/2011- scalability for large area coating. Journal of Applied 10-samsung-breakthrough-gallium-nitride-glass.html 8. Germanium

FRANK MELCHER1 AND PETER BUCHHOLZ2

1 Federal Institute for Geosciences and Natural Resources (BGR), Stilleweg, Hannover, Germany, Present address: Chair of Geology and Ore Deposit Geology, University of Leoben, 8700 Leoben, Austria 2 Mineral Resources Agency (DERA) at the Federal Institute for Geosciences and Natural Resources (BGR), Dienstbereich Berlin, Wilhelmstraße, Berlin-Spandau, Germany

Introduction Germanium has five naturally occurring iso- topes, 70Ge, 72Ge, 73Ge, 74Ge and 76Ge, the latter Germanium (Ge, atomic number 32) is a chemical being slightly radioactive, decaying by double element in Group 14 of the Periodic Table.1 The beta decay with a half-life of 1.78 × 1021 years. 74Ge existence of an element (temporarily known as is the most common isotope, having a natural ‘ekasilicon’) with properties intermediate bet- abundance of approximately 36 per cent. When ween the metal tin and the non-metal silicon had bombarded with alpha particles, 72Ge will gen- been predicted by the Russian chemist D.I. erate stable 77Se, releasing high-energy electrons Mendeleev in 1871 due to the systematic nature in the process. Therefore, it is used in combination of the Periodic Table of the elements. However, it with radon for nuclear batteries. was the German chemist Clemens Winkler who Germanium is an essentially non-toxic element, first detected it in 1886 as a component of the except for a few compounds. Dissolved in drinking mineral argyrodite (Ag8GeS6) in silver ores from water, germanium in the ppm range may cause the Himmelsfürst mine near Freiberg, Germany. chronic diseases (Gerber and Leonhard, 1997). Typically, germanium is recovered as a by-product from zinc and copper ores and coal. Distribution and abundance in the Earth

Physical and chemical properties Germanium is a rare element in rocks. However, it is present in trace quantities in most rock Germanium is a greyish-white, brittle semi- types because of its siderophile, lithophile, chalco- metal (metalloid) which has a bright lustre. Key phile and organophile characters. The crustal abun- properties are summarised in Table 8.1. Its dance is estimated at 1.5 ppm for oceanic crust and electrical properties are those of a semiconductor, 1.6 ppm for continental crust (Taylor and McLennan, i.e. between a metal and an insulator, which 1985). The average germanium content of the Earth makes germanium potentially useful for many is 13.8 ppm, with 37 ppm germanium in the core technical applications. The oxidation states of and 1.1 ppm in the primitive mantle (Dasch, 1996). germanium are +2 and +4, with a strong tendency The highest enrichment in common rock types is towards quadrivalence. in deep-sea clays that average 2 ppm germanium.

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 178 frank melcher and peter buchholz

Coals and coal ashes carry on average 2.2 and (IMA) and, in addition, some oxides, hydroxides, 15 ppm germanium, respectively (Ketris and sulfates and silicates of germanium are known. Yudovich, 2009). Considerable concentrations of Most common are argyrodite, canfieldite, briar- germanium may be found in copper–zinc–iron–sil- tite, reniérite and germanite, even though they ver sulfides and sulfosalts, as well as in association have been reported from only a limited number with organic matter, e.g. in some coals found in the (<50) of occurrences (Table 8.2). Other phases are Russian Far East and in China. Germanium con- rare, or even unique to one deposit. Kipushi in centrations in waters range from 0.06 parts per bil- the Democratic Republic of the Congo (DRC) is lion (river water) to several 100 parts per billion in the type locality of reniérite (Figure 8.1b) and thermal waters. briartite, and Tsumeb in Namibia of 13 rare germanium-bearing species. The Polkovice mine in the Kupferschiefer of Poland is the type

Mineralogy locality of morozevicite (Pb3Ge1–xS4) and polkov-

icite (Fe,Pb)3(Ge,Fe)1–xS4), whereas barquillite

Germanium does not occur as a free metal in (Cu2CdGeS4) was first described from the Fuentes nature. About 30 germanium minerals are known Villanas Li-Sn mine in Spain. to contain germanium ranging up to 70% in Germanium is a substituting element in argutite, GeO2. Many are sulfides, underlining many sulfide structures, most notably in the strong chalcophile character of germanium. common minerals such as in zinc sulfides (up to There are four germanate minerals approved by 3000 ppm in sphalerite and wurtzite) and copper the International Mineralogical Association sulfides (e.g. enargite, tennantite, bornite and chalcopyrite (Bernstein, 1985, Höll et al., 2007 Table 8.1 Selected properties of germanium. and Cook et al., 2009. Elevated concentrations have also been recorded from willemite (ZnSiO ; Property Value Units 4 up to 4000 ppm, Figure 8.1d; Saini-Eidukat et al., Symbol Ge 2009), cassiterite (up to 3000 ppm), hematite (up Atomic number 32 to 7000 ppm) and goethite (up to 5310 ppm) Atomic weight 72.63 (Bernstein, 1985). Density at 25°C 5323 kg/m3 Silicates may accommodate germanium due Melting point 938 °C to the similarity of its ionic radius and valence to Boiling point 2820 °C those of silicon; the maximum concentrations Crystal structure cubic, diamond recorded are 700 ppm in topaz from pegmatite Hardness (Mohs scale) 6.0 deposits. Quartz may carry up to a few ppm ger- Specific heat capacity at 25°C 0.32 J/(g °C) manium.

Table 8.2 Properties of the most common germanium minerals.

Typical Ge content Name Formula (%) Appearance Crystal structure Type locality

Argyrodite Ag8GeS6 5–7 steel grey with red tint, orthorhombic, Freiberg, Germany tarnishes black pseudocubic

Canfieldite Ag8(Sn,Ge)(S,Te)6 1–2 steel grey with reddish tint, orthorhombic, Colquechaca, tarnishes black pseudocubic Bolivia

Briartite Cu2(Fe,Zn)GeS4 13–18 grey to grey-blue in reflected tetragonal Kipushi, DR Congo light

Reniérite (Cu,Zn)11Fe2(Ge,As)2S16 4–8 orange-bronze, tarnishes reddish tetragonal Kipushi, DR Congo

Germanite Cu13Fe2Ge2S16 5–9 reddish grey, tarnishes dark cubic Tsumeb, Namibia brown Germanium 179

(a) (b)

(c) (d)

10 cm 0.5 mm

Figure 8.1 Photographs of germanium ores and minerals. (a) Solution collapse breccia cemented by different stages of sphalerite (different shades of brown) and galena (black) from the Bleiberg mine, Austria; grey colour is carbonate host rock, white colour is hydrothermal calcite. This represents a typical carbonate-hosted zinc-lead deposit of the ‘Alpine type’ grading 200–300 ppm Ge. (Courtesy of W. Prochaska, Leoben.) (b) Germanium ore from the Kipushi deposit, consisting of reniérite (Ren), tennantite (Tn), chalcopyrite (Cp) and gallite CuGaS2 (Ga), forming myrmekitic intergrowth textures in chalcopyrite. Reected light, oil immersion. (c) Banded Ge-rich sphalerite (grey to black with dark brown interlayers) at the Tres Marias mine, Mexico. Scale bar = 10 cm; photo taken in

2006. (d) Willemite (ZnSiO4) carrying up to 1.5 wt% Ge from Tres Marias, Mexico, showing a bright blue luminescence colour in areas of low trace element concentrations, and weak colours in areas of high trace elements (Ge, Pb). Cathodoluminescence image. Scale bar = 0.5 mm.

Deposit types up to 78 ppm, Smirnov, 1977), manganese nodules (less than 10 ppm), shale-hosted sedi- Germanium is a trace metal in most types of mentary copper deposits (<20 ppm) and por- oxidic and sulfidic metalliferous deposits, phyry copper deposits (10 to 100 ppm). It is including banded iron formations (magnetite- accumulated to economic concentrations in hematite; average probably 20 ppm, in magnetite only a few deposit types (Figure 8.2). In most Figure 8.2 The global distribution of mines, deposits and major occurrences from which germanium may be produced. Locations marked * are producing mines where the deposit contains germanium, but which are not currently recovering the germanium. Germanium 181

Table 8.3 High-grade germanium concentrations and germanium potential in various deposit types. (Based on Höll et al., 2007.)

Typical ore grade Class Deposit type Germanium-bearing species Past production Potential* (ppm Ge)

Sulfide ores 1 Volcanic-hosted Cu-Zn(−Pb) Sphalerite (bornite, low medium <<100 (−300) (−Ba) (Kuroko-type) Ge-sulfides) 2a Porphyry and vein-stockwork Cu-As sulfides, bornite, low low 10–100 Cu-Mo-Au sphalerite 2b Porphyry and vein-stockwork Argyrodite, sphalerite low medium 10–100 Sn-Ag 3 Vein-type (Ag-Pb-Zn) Argyrodite, sphalerite high (until 1993) low 100–1000 4 Sediment-hosted Zn-Pb-Cu Sphalerite, wurtzite high high 10–100 (−Ba) 5a-c Carbonate-hosted Zn-Pb Sphalerite, wurtzite high high 100–1000 (MVT, IRT, APT) 5d Kipushi-type (KPT) Ge-sulfides (sphalerite) high (until 2000) medium 10–1000 polymetallic 6 Sediment-hosted stratiform Sphalerite, pyrite, rare Ge low medium 1–20 Cu deposits sulfides

Oxide ores 5e Oxidation zones of Fe oxides/hydroxides, medium low 100–1000 KPT ores sulfates, arsenates 5f Non-sulfide Zn-Pb Fe hydroxides, willemite, low low 10–100 hemimorphite 2c Oxidation zones of Secondary Sn hydroxide, Sn low low 10–100 Sn sulfides oxide 7 Iron oxide ores Fe oxides/hydroxides none low 10–50

Coal and lignite deposits 8a Coal and lignite Organic matter medium high 100–1000 8b Coal and lignite ash high high 10–1000

*estimated future potential difficult to assess for many deposit types due to a lack of data on germanium grade in concentrates, recovery and reserves/resources. Ag, silver; Au, gold; Ba, barium; Cu, copper; Fe, iron; Mo, molybdenum; Pb, lead; Sn, tin; Zn, zinc. APT, Alpine-type; IRT, Irish-type; KPT, Kipushi-type; MVT, Mississippi Valley-type.

of them, germanium may be recovered as a Accumulation of germanium in sulfide deposits by-product from copper, zinc and lead produc- tion. In general, three modes of occurrence are In volcanogenic massive sulfide (VMS-) deposits, distinguished (Table 8.3): sulfidic, oxidic and germanium concentrations are low to moderate biogenic. Sulfidic concentrations of germanium (<100 ppm); however, due to their large tonnages, are the most widespread and the most variable the refining of VMS zinc–copper–lead–gold–silver type, although the importance of germanium- ores may yield significant cadmium, indium, rich coal deposits is likely to increase in the gallium, tin, antimony, bismuth and also germa- future. nium (class 1, Tables 8.3 and 8.4). 182 frank melcher and peter buchholz

Table 8.4 Key features of the main germanium-bearing deposit classes. (Modified from Höll et al., 2007.)

Class Deposit type Brief description Features Examples

1 Volcanic-hosted massive Lenticular seafloor massive In extensional oceanic Kuroko-type ores (Japan); sulfide Cu-Zn(-Pb) sulfide orebodies, often settings; associated with Neves Corvo Cu-Sn-Zn (-Ba) (VMS) with footwall stringer volcanic (basic to acid) (−Ag-Se-ln-Ge) (Portugal); zones rocks. Ge concentrations Gorevskoe Pb-Zn (Russia); up to 100 ppm in recent Ozernoe Zn-Pb (Russia) ores, up to 370 ppm in Kuroko-type deposits 2a Porphyry and Medium to large, In compressional tectonic Capillitas (Argentina); Bor vein-stockwork low-grade stockwork of settings; Ge-bearing (Serbia); Chelopech Cu-Mo-Au quartz veinlets and sulfides occur in peripheral (Bulgaria); Ladolam disseminations in felsic zones of porphyry systems, (Papua New Guinea) intrusive rocks including late-stage epithermal veins 2b Porphyry and Stockworks and arrays of Mineralization includes Potosi/Bolivian ‘Ag-Sn belt’; vein-stockwork Sn-Ag ore veins in subvolcanic, argyrodite and Ge-bearing Barquilla (Spain) felsic intrusions Sn minerals 3 Vein-type Ag-Pb-Zn(-Cu) Ore veins hosted by Heterogeneous group of vein Freiberg (Germany); Noailhac- sedimentary and deposits found in different Saint Salvy (France); Kirki magmatic rocks tectonic settings (Greece) 4 Sediment-hosted Concordant lenses of Hosted by clastic marine Red Dog (Alaska, U.S.A.); massive sulfides stratiform massive to sediments, including black Jinding/ Lanping (China) Zn-Pb-Cu(-Ba) (SMS) semi-massive sulfides shales, along continental and sulfates (baryte), margins or in intracratonic often with footwall rift settings stringer zones and stockworks 5 Carbonate-hosted base Semi-massive to Stable carbonate platforms; Worldwide from the early metal deposits disseminated sulfide low-temperature Proterozoic to Cenozoic ores (Pb-Zn(-Cu)) in (commonly <300°C) carbonate rocks 5a Mississippi Valley-type Hosted in carbonate Signifcant age difference Elmwood-Gordonsville district (MVT) Zn-Pb-Fe(-Cu) successions since the between host rock and (Tennessee, U.S.A.); Pend (-Ba)(-F) Palaeoproterozoic; very mineralisation; sulfur Oreille (Washington, common type derived from thermal U.S.A.); Tres Marias sulfide reduction (Mexico); Fankou, Huize (China) 5b Irish-type (IRT)Zn-Pb(-Ag) Hosted in Lower Bacteriogenic sulfur source Navan, Tynagh, Lisheen (all (-Ba) Carboniferous carbonate Ireland) rocks 5c Alpine-type (APT) Zn-Pb Hosted in Triassic Bacteriogenic and Bleiberg (Austria); Mežica limestones non-bacteriogenic sulfur (Slovenia); Cave de Predil sources (Italy) 5d Kipushi-type (KPT) Hosted in Neoproterozoic High-temperature fluids Tsumeb (Namibia); Kipushi polymetallic limestones (central and (250-450°C); sulfur (DR Congo); Kabwe Cu-Zn-Pb-Ag-As SW Africa) derived from thermal (Zambia); Ruby Creek sulfate reduction. Highest (Alaska, U.S.A.) Ge grades known

(Continued ) Germanium 183

Table 8.4 Continued

Class Deposit type Brief description Features Examples

5e Oxidation zones of KPT Extensive oxidation of Ge in Fe-hydroxides, and Tsumeb (Namibia); Apex deposits primary sulfide ores forming secondary Ge (Utah, U.S.A.) minerals 6 Sediment-hosted Stratiform disseminated Low-temperature diagenetic Kupferschiefer (Poland, stratiform Cu deposits sulfide mineralisation to epigenetic deposits Germany); Central African associated with formed from oxidized Copperbelt (DRC, Zambia) carbonaceous sediments fluids of low salinity 7 Iron oxide ores Precambrian banded iron Ge in iron oxides from Hamersley Range (Australia), formations; Phanerozoic sedimentary deposits; low Krivoi Rog (Ukraine), volcanogenic- Ge in Fe skarn deposits Atasu type (Kazakhstan) sedimentary deposits 8 Coal and lignite Extensive beds of coal and Up to 3000 ppm Ge in fly Lincang, Wulantuga (China); lignite close to bedrock ash from coal combustion Tarbagataisk, Shotovsk contacts (Russia); Lugansk (Ukraine); Angrensk (Uzbekistan)

Ag, silver; As, arsenic; Au, gold; Ba, barium; Cu, copper; F, fluorine; Fe, iron; In, indium; Mo, molybdenum; Pb, lead; Se, selenium; Sn, tin; Zn, zinc.

Porphyry copper and related vein-stockwork vein districts yielding scattered germanium copper–(molybdenum–gold–silver–tin) deposits values are known from the Freiberg district and host large tonnages of ore, but commonly Harz mountains in Germany, Kutna Hora (Czech low grades of germanium (class 2a and 2b). Republic), Sardinia (Italy), Kirki (Greece) and Nevertheless, germanium minerals are reported South Korea, but their economic potential is from late-stage epithermal veins in Argentina, limited (Höll et al., 2007). Peru and Bolivia, and from several porphyry A major share of today’s global germanium pro- copper deposits where germanium is hosted by duction is from large-tonnage, but low-grade sedi- bornite, chalcopyrite and pyrite, and also present ment-hosted zinc-lead deposits of the ‘SEDEX class’ as microphases. In tin–silver stockwork deposits (class 4). The sphalerite of the giant Red Dog deposit, of Bolivia and Peru (Potosi-type), germanium is Alaska (proven and probable reserves in 2006 were an accessory component with ore shoots contain- 85 million tonnes at 18.2% Zn, 5.6% Pb) averages ing argyrodite associated with silver phases. about 100 ppm Ge (Kelley et al., 2004) which is However, no reliable information exists on partly recovered as a by-product in the Teck smelter average grades and expected tonnages in any of in Trail, British Columbia, Canada. In 2007, the the above-mentioned deposits (Höll et al., 2007). Trail facility produced about 40 tonnes of germa- Polymetallic tin–silver and silver–lead–zinc nium from zinc concentrates sourced from Red Dog vein deposits (class 3) contributed significantly to (600,000 tonnes per year zinc concentrate) and Pend the global germanium production until the clo- Oreille, Washington (83,000 tonnes per year zinc sure of the Noailhac-Saint Salvy deposit, French concentrate) (Guberman, 2008). Since then, no pro- Massif Central, after 18 years of production, in duction data have been released; however, Teck 1993. The cumulative production was 500 tonnes announced that 25 per cent of the zinc concentrates germanium together with 0.35 million tonnes produced at the Red Dog mine are transported to the (Mt) zinc and 280 tonnes silver; zinc concentrates Trail smelter for further treatment. graded 700–800 ppm Ge, and sphalerite carried Carbonate-hosted zinc-lead deposits, com- up to 2500 ppm Ge (Cassard et al., 1996). Other monly referred to as ‘Mississippi Valley-type’ 184 frank melcher and peter buchholz

(MVT, class 5), constitute a class of potentially cadmium–vanadium–molybdenum–tungsten) economically interesting germanium deposits. reected in a complex mineralogical composi- There is discussion on alternative classifications tion, and by the presence of discrete germanium or the application of MVT subsets (e.g. Leach et phases. At the type localities, Kipushi in al., 2005); in addition to strictly epigenetic MVT Katanga, DRC (past production 1925–1993: 60 sensu stricto (class 5a), an ‘Irish-Type’ (IRT, million tonnes at 6.8% Cu, 0.9% Pb, 11% Zn, on class 5b), an ‘Alpine-type’ (APT, class 5c; average 100–200 ppm Ge) and Tsumeb in the Figure 8.1a) and a ‘Kipushi-type’ (KPT, class 5d; Otavi Mountainland, Namibia (past production Figure 8.1b) are distinguished by Höll et al. (2007) 1947–1996: 28 Mt at 4% Cu, 12% Pb, 5% Zn, based on the relationships between ores and average 50–150 ppm Ge), the sulfide ores are host rocks and geochemical characteristics of the composed of tennantite, sphalerite, galena, chal- ores and ore uids (Table 8.4). In all carbonate- copyrite, bornite, pyrite, and a large number of hosted zinc–lead sulfide deposits, germanium is rare phases including the germanium minerals invariably substituted into the sphalerite and germanite, reniérite and briartite (Figure 8.1b; wurtzite structures; with discrete germanium Melcher, 2003, Schneider et al., 2007 and minerals usually absent. High germanium Kampunzu et al., 2009). The ores are medium- to concentrations are reported from the Tri-State dis- low-temperature (400–200 °C) hydrothermal in trict (Tennessee, Missouri, Arkansas) on the origin and frequently occupy breccia zones in Viburnum trend, USA (60–400 ppm in zinc ore carbonate successions. Similar deposits are concentrates), Bleiberg, Austria (on average known in the Neoproterozoic polymetallic ore 300 ppm; 126 tonnes germanium produced), Cave districts of the Otavi Mountainland, Namibia, del Predil, Italy (250–450 ppm), Huize, China (up and the Katangan Copperbelt, Zambia and DRC, to 354 ppm), and Niujiaotang, China (up to including the small, but high-grade Khusib 546 ppm) (Bernstein, 1985, Höll et al., 2007 and Ye Springs deposit in the vicinity of Tsumeb with et al., 2011). Only some of the deposits are actively average germanium concentrations of 100 ppm mined at present, and only a few contribute to the hosted by enargite, tennantite and germanium- global germanium production. Huize, for example, bearing colusite (Melcher et al., 2006). None of has expanded its annual production to 100,000 the ‘Kipushi-type’ deposits are mined at present, tonnes zinc and ten tonnes germanium. Some and the high-grade zones in the known deposits deposits have exceptional concentrations of ger- that graded over 1% Ge have been mined out. manium, e.g. on average 1000 ppm germanium in However, for Kipushi, remaining resources sphalerite at the small, high-grade Tres Marias down to the 1500-metre level are estimated at > 5 deposit in northern Mexico (Figures 8.1c and d; million tonnes zinc, >500,000 tonnes copper, Saini-Eidukat et al., 2009). The potential for ger- and >100,000 tonnes lead from ores averaging manium recovery from carbonate-hosted zinc- 21.4% Zn, 2.1% Cu and 0.88% Pb with traces of lead deposits is large, due to the widespread germanium (68 ppm), cadmium, cobalt and sil- occurrence of these ores in carbonate platforms ver (Kampunzu et al., 2009). ranging from the Palaeoproterozoic to the Tertiary. Although there is a potential for new KPT dis- However, local factors that are not well under- coveries, base-metal exploration at present stood at present control the germanium distribu- focuses on SMS, VMS and MVT-type zinc–lead tion in these deposits. ores, on porphyry copper ores, on sandstone- The highest germanium concentrations are hosted copper ores, and on sediment-hosted found in carbonate-hosted polymetallic hydro- copper ores of the Kupferschiefer type that are thermal deposits of the ‘Kipushi-type’ (KPT, usually low in germanium (class 6). Kupferschiefer class 5d). These enigmatic deposits are charac- from Mansfeld in the Harz mountains of Germany terised by a complex elemental association contains 8–15 ppm Ge, which was partly recov- copper–zinc (–lead–silver–arsenic–germanium– ered prior to mine closure in 1990. Germanium 185

Oxidised portions of germanium-rich sulfide germanium is present in humus (Yang et al., deposits (class 5e) may carry exceptionally high 2003), mainly bound to humin (86–89 per cent) germanium grades. However, as such deposits are and humic acid (10–12 per cent). A process to rare and most of them mined out, their economic recover germanium from coal ash was developed potential is limited. Germanium may be present in the UK and used by Johnson Matthey between as secondary germanium phases (Tsumeb, 1950 and 1974 utilising ashes containing between Namibia), or absorbed on iron hydroxides and 0.5 and 1% Ge. The coals from Durham and oxides (up to 2.5% Ge at Tsumeb; 0.5% at Apex, Northumberland contained approximately USA). The little available data on germanium 300 ppm Ge. Flue dusts from combustion were concentrations in supergene non-sulfide zinc– collected from coal gas works and power stations. lead deposits (class 5f) suggest only minor signifi- However, when coal gas was replaced by natural cance for germanium recovery, although some of gas and power stations were converted to oil the deposits formed from germanium-bearing power, germanium recovery ceased (Roskill, sphalerite. Some giant iron oxide deposits (class 7), 1988). In the former USSR, 850 tonnes of germa- e.g. Archaean and Palaeoproterozoic banded iron nium was mined from germanium-rich coals formations, carry appreciable germanium locked from the Novikovsk deposit (Sachalin) between up in iron oxides and hydroxides (up to 100 ppm). the 1960s and the 1980s, and were burnt with less However, the lack of appropriate technology and germanium-rich coals from the Tarbagataisk high costs will probably deter industry from deposit at the Chita Heat Electropower Station recovering trace elements such as germanium. (Kats et al., 1998). Germanium-rich ashes were Iron ore wastes might constitute a future source subsequently transported to Angren (Uzbekistan) of germanium; it was estimated that stockpiled for upgrading and from there to Krasnoyarsk for wastes in the Ukraine contained at least 200 mil- further processing. Other germanium-rich coals lion tonnes of iron and 100,000 tonnes of germa- from eastern Russia were processed in the nium (Levine and Wallace, 2000); this, however, Ukraine. would imply on average 500 ppm Ge in the Today, 30 to 50 per cent of the primary germa- wastes, which seems unreasonable. nium production is from lignite deposits in Appreciable quantities of germanium are China, Russia and Uzbekistan (Seredin and contained in slag produced in the past from the Finkelman, 2008, Bleiwas, 2010 and Seredin, smelters at Tsumeb (Namibia), Lubumbashi and 2012). The first germanium-rich coal deposit in Kolwezi (DRC). The Tsumeb smelter processed the former USSR was in the Angren valley polymetallic ores from the Otavi Mountainland (Uzbekistan) in the 1950s, followed by a number (Tsumeb, Kombat, Khusib Springs, Berg Aukas of deposits in various parts of Russia and China, and others) and imported copper ores, whereas with most of them in north-eastern China and the Congolese smelters processed ores from the the Russian Far East. According to Seredin and Congolese Copperbelt (sediment-hosted strati- Finkelman (2008) and Seredin et al. (2013), the form copper ores) in addition to germanium-rich largest known germanium-rich coal deposit is in KPT-type ores from Kipushi. Wumuchang (Yumin / Yimin coal field, Inner Mongolia) with an estimated 4000 tonnes of ger- manium resources, followed by Bikinsk (eastern Enrichment of germanium in lignite and coal Siberia) with 2600 tonnes of germanium. Average Since V.M. Goldschmidt (Goldschmidt, 1930 and germanium grades in ash range from 30 ppm Goldschmidt and Peters, 1933) found high germa- (Wumuchang) to >1000 ppm (Shkotovsk, Russian nium concentrations in coal ashes from the Far East), and the thickness of individual germa- Durham Coalfield, U.K. (up to 1.1% Ge), coal and nium-bearing beds ranges up to 15–20 metres lignite deposits have attracted both researchers (Seredin and Finkelman, 2008). The Wulantuga and industry. In lignite, 97 per cent of the deposit in the Shengli Coalfield, Inner Mongolia 186 frank melcher and peter buchholz

(China) is the largest germanium producer at pre- Extraction methods, processing sent (Du et al., 2009, Dai et al., 2012 and Seredin, and beneficiation 2012). The germanium-rich part of the deposit contains 12.3 million tonnes of lignite with Extraction >30 ppm Ge, and 1700 tonnes of germanium Germanium is a by-product in some sulfide ores extractable from coal with Ge content >100 ppm and in some coals; therefore, surface and under- (Du et al., 2009). The Lincang lignite deposit ground mining methods that are commonly used (1060 tonnes germanium resources, Hu et al., to extract base metals (copper, zinc and lead) and 2009) in Yunnan, southern China, is another coal are employed. Large sediment-hosted significant germanium producer (16 tonnes per deposits, such as Red Dog, Alaska, and many lig- year of high-grade GeO ). Evaluation of the Kas- 2 nite deposits operate in open-pits, whereas many Symskaya lignite deposit in western Siberia MVT-type deposits and carbonate-hosted base- yielded resources of 11,000 tonnes germanium at metal deposits have extensive underground work- an average grade of 205 ppm Ge in dry lignite ings, e.g. employing room and pillar techniques. (Evdokimov et al., 2004). Most germanium-rich The Apex mine in Utah was the only mine that lignite deposits are hosted by small fault-con- produced primary germanium and gallium from trolled coal basins in Mesozoic or Cenozoic sedi- an ancient underground copper mine with lead, mentary successions. Germanium was introduced zinc and silver as by-products (Taylor and Quinn, by hydrothermal uids from the basement, often 1987); the valuable metals were concentrated in connected with magmatic activity. Hydrothermal iron oxyhydroxides that were leached with alteration is widespread, and metasomatic rocks sulfuric acid and sulfur dioxide, followed by the associated with coal beds may carry significant precipitation of cement copper and solvent germanium concentrations. In addition to germa- extraction of the pregnant solution to separate nium, the coals often carry elevated concentra- the germanium and gallium. The mine and plant tions of tungsten, uranium, niobium, caesium, operated from 1986 to 1987, but closed due arsenic, antimony, mercury and thallium. to decreasing germanium prices. A subsequent A review of trace-element data in coal prov- attempt to re-open the mine in 1990 also proved inces up to 1990 indicated that germanium to be uneconomic. concentrations mostly range from 0.5 to 50 ppm (Swaine, 1990). However, research in some coal Processing provinces has identified higher germanium concentrations, e.g. in Bulgaria (Gouin et al., After mining, the ores are processed to increase 2007). Germanium-bearing coal seams occur in their base metal and germanium contents using a carbonaceous sandstone deposit with interca- conventional mechanical (grinding, sieving, lated kaolinite at Lang Bay near the Pacific magnetic separation) and otation methods. Three Coast, Canada (Queneau et al., 1986). A com- methods of germanium extraction were used: (1) prehensive study of coal fields in India identi- from germanite and reniérite ores, (2) from zinc ore fied elevated germanium concentrations only concentrates, (3) from coal and y ashes. Extraction locally, e.g. in Assam (600 ppm Ge; Banerjee et from slag material will become important in the al., 2000). Germany´s lignites and coals seem to future (e.g., AUSMELT technology). A simplified carry low germanium concentrations (e.g. ow diagram illustrating the consecutive steps of 3–10 ppm Ge in ashes from Rhenish lignite; germanium processing is presented in Figure 8.3. 1–2 ppm in hard coal from the Ruhr Basin, Originally, the Eagle-Pitcher smelter at Feiser, 1966). However, elevated values of up to Henrietta, Oklahoma, operated a Waelz kiln to 50 ppm Ge in coal have been reported from fume zinc, cadmium, germanium and gallium some middle-German hard-coal basins (Leutwein from low-grade residues from the 1950s until and Rösler, 1956). 1968. GeO and GeS begin to sublime at 700 °C, Germanium 187

Germanium-bearing Germanium-bearing scrap ores, coals, flue dusts (0.001–0.01%)

Grinding, sieving, magnetic separation, flotation

Germanium concentrate Pyrometallurgy (5–30%) hydrometallurgy

Chlorination, distillation, purification

Germanium Fibre optics tetrachloride (GeCl4)

Hydrolysis Crystals (BGO) PET catalysts Germanium γ-ray detection dioxide (GeO2) Medicine Phosphors Reduction

First reduction Alloys, metal compounds

Zone refining

Intrinsic germanium metal Optical casting (polycrystalline)

Crystal growing/ Infrared lenses pulling Figure 8.3 Simplified germanium processing ow diagram. (BGO, bismuth Single-crystal germanium germanium oxides (Bi4Ge3O12); PET, polyethylene teraphthalate.) (Based on Single-crystal X-ray mono- Butterman and Jorgenson, 2005 and Solar cells wafers chromator crystals Naumov, 2007.) whereas the boiling point of germanium tetra- leaching the distillation residue with hydro- chloride (GeCl4) is 85 °C. Today, germanium chloric acid, (5) hydrolysis of germanium tetra-

is mostly obtained from zinc smelter flue sys- chloride into germanium dioxide (GeO2) and (6)

tems and dusts of smelters that process zinc optional reduction of GeO2 into metallic germa- ores. This process involves (1) refining zinc, nium with hydrogen at 760 °C (Global Industry (2) distillation under non-oxidising condition, Analysts, 2010). Metal powder is melted and (3) recovery of the distillation residue, moulded into metal bars, from which highly (4) formation of germanium tetrachloride by purified metal may be produced by zone refining 188 frank melcher and peter buchholz

polycrystalline processes. Fractional crystallisa- Germanium dioxide, GeO2 tion of volatile GeCl (‘fuming’) is also used to 4 GeCl is hydrolysed with deionised water to pre- separate germanium from other metals. 4 cipitate germanium hydroxide. From this GeO is The recovery of germanium from lignite 2 filtrated and vacuum baked. GeO is the most includes burning the coal at 1200 °C and pyro- 2 widely traded germanium product and is used as metallurgical treatment after filtering the a catalyst for the production of high-quality ashes, followed by sulfuric acid leaching polyester (PET). Both crystalline and amorphous (Seredin, 2012). Experiments to recover germa- GeO (99.5%) can be used. Other crystalline nium from lignite using micro-organisms 2 grades are used for special applications, e.g. uo- yielded a germanium recovery of up to 85 per rescence grade for the production of phosphors cent (Yang et al., 2003). and BGO grade (99.999%, 5N2) for the crystal Sirosmelt, a technique developed by CSIRO growth of bismuth germanate (BGO, Bi Ge O ) (Australia), was tested to extract germanium 4 3 12 by the Czochralski technique. Electronic-grade from sandstone-hosted lignite at Lang Bay, GeO (5N) is best suited for the production of ger- Canada (Queneau et al., 1986) due to its effective- 2 manium metal. ness for small-scale fuming of wet particulate feeds. Flotation technology was also developed using a cyanamid collector for concentration of First reduction metal weathered coal, giving germanium recoveries of Reduction of GeO2 with hydrogen in ultra-clean 75 per cent. graphite boats at 760 °C yields germanium metal An innovative, economic and environmentally powder. Metal bars are produced by melting of friendly germanium recovery process from waste the powder at 1100 °C. products of optical fibres has been developed by Bell Labs (Global Industry Analysts, 2010). Production of zone-refined metal (‘intrinsic’ metal) Specifications First reduction metal is purified to yield pure polycrystalline germanium known as ‘intrinsic’ Germanium is used and traded in a variety of germanium metal by a process known as zone forms, including zone-refined crystalline germa- refining, where a succession of small molten nium, germanium tetrachloride, high-purity zones are created by passing the metal bar oxides, first-reduction ingots, single-crystal bars through a series of inductive coil heaters, con- and castings. The most important are shown in centrating impurities in the molten zone and Figure 8.3 (for further information see PPM Pure thus accumulating at the end of the bar. Metal is Metals, 2012 and Umicore, 2012). offered as zone-refined polycrystalline ingots, granules and powder with purities ranging from 4N to 13N, the purity level needed for gamma Germanium tetrachloride, GeCl 4 X-ray detection.

GeCl4 is recovered from germanium-bearing concentrates by dissolution in hydrochloric acid Single crystals and purification through a series of fractional dis- tillations. GeCl4 is used to profile the refractive Single crystals are required as optical compo- index of optical fibres for telecommunication. It nents in infrared optical systems and electronics. is produced in two qualities, fibre-optic grade and Crystals up to 300 mm in diameter are grown by ultrahigh-purity grade with less than five parts the Czochralski vertical pulling process, using a per billion metal impurities, used as a dopant for monocrystalline seed dropped in a bath of molten fibre optical products. germanium. In the horizontal pulling process, a Germanium 189 melting zone moves from a monocrystalline seed have continuously changed over the years along a polycrystalline rod. without showing a major trend. Today, the three sectors, fibre optics (30 per cent), infrared optics (25 per cent) and catalysts for colorless PET (25 Uses per cent) account for 80 per cent of the global end- use sectors (Guberman, 2013). For these three Germanium owes its usefulness to six properties: sectors, applications tend to vary with region, e.g. ● It is an intrinsic semiconductor, particularly the PET sector is not important in the USA and effective at high frequencies and low voltages. Canada, but in Europe (29 per cent of global ● It is transparent to infrared light. volume sales), Asia-Pacific (25 per cent), and ● It is a glass-former, e.g. able to form three- Japan (21 per cent) (Global Industry Analysts, dimensional networks of germanium-O tetrahedra. 2010). Fibre optics are mainly used in the USA ● It has a high refractive index. (42 per cent of global volume sales) and the EU ● It has low chromatic dispersion. markets (21 per cent) followed by Asia-Pacific (12 ● It has an ability to catalyse the polymerisation per cent). The regional end-use distribution in the production of plastic (polyethylene tere- pattern of infrared optics is almost identical to phthalate (PET) mainly used for plastic bottles) the regional end-use pattern for fibre optics without undesirable coloring. (Global Industry Analysts, 2010). Germanium was first used in industry after Compared to the dominant three end-use sec- World War II when Karl Lark-Horovitz from tors, various applications in the electronic/solar Purdue University discovered its properties as a industry consume about 15 per cent of the germa- semiconductor (Haller, 2006). The point-contact nium produced, e.g. for silicon chips (SiGe chips), diodes for radar pulse detection had already been high-speed integrated circuits such as wireless com- used during the war. From 1948 until the 1970s, munication devices or high-efficiency multi- germanium transistors played a vital role in junction photovoltaic cells (gallium arsenide on solid-state electronics, but then were replaced by germanium substrate) for space applications. high-purity silicon that has superior electrical Although in small amounts only, other uses of properties. New applications arose in the late germanium include gamma ray detectors, X-ray 1950s with the boom in nuclear physics and the monochromators, thermo-photovoltaics, medical need to develop nuclear radiation spectrometers applications such as chemotherapy, and metallurgy with good energy resolution. (alloys with tin to increase hardness or silver to Global sales volumes of germanium at manu- prevent tarnishing (Angerer et al., 2009). facturer level, which basically represent the ger- manium content in end-use sectors, were about 115 tonnes germanium in 2010. The major ger- Recycling, re-use and resource efficiency manium users by global volume sales are the USA (37 per cent), Europe (21 per cent) and Asia- Due to its high dispersion in most products and Pacific (14 per cent). Within Europe (including application in very low quantities, little germa- Russia), the largest volume sales of germanium in nium is recovered from post-consumer scrap (‘old end-use applications are in France (33 per cent), scrap’). Recycling from ‘new scrap’, however, is Germany (24 per cent), and the UK (17 per cent); more widespread in the production of germa- (Global Industry Analysts, 2010). nium-containing fibre-optic cables and infrared Over the past two decades, the major uses for imaging devices. In the US economy, about 50 per germanium changed from dominantly infrared cent of recycled material was from optical fibres optics in 1990 to fibre optics in 2000 and new (mainly new scrap). Most of the scrap is sent to applications for PET (Table 8.5, Figure 8.4). germanium processors for recycling which are Relative proportions of these three applications able to recycle material containing a minimum of 190 frank melcher and peter buchholz

Table 8.5 Summary of the major end-use markets of germanium. (Data from Guberman, 2011.)

Market sector End use (2010)* Usage Germanium product

Fibre optics 30% Communication networks GeO2 dopant within the cores of optical glass fibres; Ge is deposited with silica on the inside of a pure quartz tube that is drawn into fine fibres Infra-red optics 25% Night vision systems; optical instruments Polycrystalline and single crystals; n-type such as camera lenses, IR spectroscopy; Ge doped with P or Sb; Ge lenses and

military, active car safety systems, windows; GeO2 is used as a component satellite systems, fire alarms of glasses in wide-angle camera lenses and microscope objectives

Polymerisation catalysts 25% Plastics (polyethylene terepthalate, PET) GeO2 is used to make PET heat resistant for water bottles (Asia) Electronics and solar 15% Wireless devices, optical communication SiGe-based chips compete with GaAs and electrical applications systems, hard disk drives, GPS Si chips (since 1997) Semiconductors Transistors, rectifiers, lasers Germanium doped with Sb, As, P (n-type semiconductor), Ge doped with Al, B, Ga, In (p-type semiconductor); amorphous semiconductors Diodes Light-emitting diodes (LEDs) Ge substrate for high brightness LEDs Transistors Supercomputer Si-Ge heterojunction bipolar transistors; silicon-on-insulator technology Solar cells Space-based photovoltaic solar cells Polished Ge wafers for multilayer solar cells (improved efficiency compared to silicon) Radiation detectors 5% Detection of gamma radiation, e.g. air Single crystals of ultra-pure Ge traffic control Superconductors Melting Ge and other metals to produce

superconducting alloys, e.g. Nb3Ge; vacuum sintering and annealing; thermal dissociation; electrolysis; chemical vapour deposition methods Medicine Chemotherapy, dietary supplement, Ge-organic compounds (Ge-132) found in antioxidant, super nutrient, diagnosis medicinal plants; Bis-beta carboxyethyl of diseases Ge sesquioxide (Ge-Oxy 132); colloidal Ge

*estimated share of global annual sales volumes of germanium.

Al, aluminium; As, arsenic; B, boron; Ga, gallium; GeO2, germanium dioxide; In, indium; Nb, niobium; P, phosphorus; Sb, antimony; Si, silicon.

2% Ge, including solutions, fibres, dust and filter 2005). According to UNEP (2011) the global end- mats from air-conditioning systems. Processing of-life recycling rate is less than one per cent of scrap normally involves dissolution into ger- (recycled germanium in old scrap as share of ger- manium tetrachloride. The quality of products is manium metal content of end-of-life products). not affected by recycled material (Kammer, 2009). Infrared optics: about 30 per cent of the germa- On a global scale, about 25–35 per cent of total nium consumed for this end use is produced from germanium used is processed from recycled scrap, recycled materials that accumulate during the mainly new scrap (Butterman and Jorgenson, manufacture of optical devices such as broken Germanium 191

1990 2000 waste accumulates as new scrap which is recy- 5% 5% cled. Old computers which are dismantled or 13% 8% shredded contain only minute amounts of germa- nium in their electronic parts. The major 19% 25% 50% challenge is the effective collection of old devices and cost-effective separation of metals. Polymerisation catalysts: estimated world 60% 15% polyester production in 2009 was about 50 million tonnes per year (Chemsystems, 2011). About a third 2007 2010 of this production volume was used for PET bottle 5% grade. In 2006, about 350,000 tonnes per year was 10% 24% produced by using germanium catalyst as a substi- 15% 30% 12% tute for antimony catalyst, mainly in Japan. Applying germanium instead of antimony results in very low levels of antimony contamination in 25% 31% 23% bottled drinking water. The average germanium 25% metal consumption in polyester production is about 80 ppm per tonne PET (Thiele, 2006). The ger- manium remains in the product, mainly PET bot- Fibre-optic systems Infrared optics Polymerisation tles, grading 10–70 ppm per tonne (Oakdene Hollins catalysts Ltd, 2011). PET bottles are normally re-used or Electronics and solar Phosphors, metallurgy, shredded and sold for other applications. As PET electric applications medicine post-consumer scrap, germanium is totally lost.

Figure 8.4 Comparison of global end-uses of Important producers of catalyst grade GeO2 include germanium in 1990, 2000, 2007 and 2010. (Data from Teck Metals Ltd. and Umicore. Guberman, 2011.) lenses and glasses from night-vision devices. This Substitution compares favourably with the five per cent to ten per cent of broken lenses and glass from night- Since the beginning of the industrial use of ger- vision devices that were recycled as old scrap in manium, attempts have been made to substitute the past (Jorgenson, 2006). this expensive and rare metal. In each of the Fibre optics: fibre-optic cables for telecommu- applications listed in Table 8.5, substituting nication contain up to 60 per cent of recycled technologies have been developed. The use of material and the recovery rate of germanium germanium in fibre-optical systems is challenged from fibres is up to 80 per cent (Jorgenson, 2006). by the invention of photonic crystal fibres con-

Optical silica fibres contain about 4% GeO2, but sisting of an array of glass capillaries and solid only 0.03 per cent to 2 per cent of the cable is the rods stacked together. optical core. For 2.5 km of optical fibre, about one Zirconium–indium-based uoride glass for gram of GeO2 is used (Oakdene Hollins Ltd, fibre-optic applications might also be a successful 2011). Fibre-optic cables also have a long residence substitute, but research is at an early stage time in their applications, thus recycling vol- (Angerer et al., 2009). umes for old scrap are rather low. However, in the Silicon is used as a substitute in certain future, they may have considerable post-con- electronic applications. Competing materials in sumer recycling potential. infrared technologies are zinc selenide (ZnSe) and Electronics and solar electrical applications: GASIR, an infrared-transmitting chalcogenide during solar-cell production, about 50 per cent of glass developed by Umicore. In the PET market 192 frank melcher and peter buchholz with its low-cost products, replacement of expen- regions. More than 80 per cent of the germanium sive GeO2 is possible using titanium-, antimony- reserves and resources are located in Guangdong, and aluminium-based catalysts. Substitution for Yunnan, Jilin, Sichuan and Shanxi. In the Russian germanium substrates in space solar cells comes Federation, about 4000 tonnes germanium from Inverted Metamorphic Cells or advanced reserves have been calculated for 21 deposits double junction cells grown on gallium arsenide (Kats et al., 1998), with over 50 per cent of the (GaAs), and in LEDs by copper and silicon (blue total reserves accumulated in germanium-rich LEDs) and gallium phosphide (GaP) and copper lignite deposits of the Russian Far East (Primorsky (red LEDs). Silicon–germanium (SiGe) chips may Krai and Sachalin), and 40 per cent in lignite be substituted by gallium arsenide chips that are fields in the Kemerovo and Chita regions. The less expensive, but SiGe may achieve higher oper- remaining reserves are hosted in sulfide deposits ational frequencies than its competitor products in the Urals, Altai and Caucasus regions. (Global Industry Analysts, 2010). Following considerable exploration, the esti- mated reserves of germanium-rich coals in the Russian Far East and the Inner Mongolia Province Environmental aspects of the life cycle of China are now estimated as 6000–7000 tonnes of germanium and its products and 5600 tonnes, respectively (Seredin and Finkelman, 2008). Höll et al. (2007) give a conser- Germanium has little or no effect upon the envi- vative estimate of “a few thousand tonnes” of ronment because it usually occurs as a trace recoverable germanium from low-temperature element in rocks, ores and most products. Of the hydrothermal zinc deposits and polymetallic materials consumed or produced as a by-product Kipushi-type deposits, including slag at Tsumeb during processing of germanium-bearing mate- and Kipushi. They also estimate that at least a rials, arsenic and cadmium may present potential few thousand tonnes of germanium are present in problems. However, these metals are separated the coal ash and ue dust produced annually at out at the smelter stage and are readily controlled coal-fired power stations. However, extraction of in the refineries. Acids and bases used in germanium from these waste products is not gen- processing are neutralised and held in tailings erally economic under present conditions. Data ponds (Roskill, 1988). Germanium compounds on germanium resources in volcano-sedimentary also have a low order of toxicity, except for ger- magnetite-hematite deposits are not available. manium tetrahydride, which is considered toxic A compilation of published and estimated (Roskill, 1988). reserves for active and semi-active mines and advanced exploration projects by Elsner et al. (2010) amounted to more than 27,000 tonnes ger- Resources and reserves manium reserves and resources, including sulfide ores, coal and slag. With the addition of further Global resource and reserve data for germanium projects to that database, e.g. the Kas-Symskaya are difficult to obtain, because details relating to coal field in Siberia with 11,000 tonnes germa- trace-metal concentrations in many sulfide and nium, the combined reserves and resources are coal deposits are not readily available, or are of estimated at 13,000 tonnes germanium from sul- poor quality. fide deposits and associated slags, and 25,000 The United States Geological Survey lists 450 tonnes from germanium-rich coals (Tables 8.6 tonnes germanium reserves for the United States and 8.7). The amount of germanium potentially (USGS, 2013). The combined germanium reserves recoverable from coal ash is unlimited, but the and resources in China are estimated to be about commercial recovery is currently not viable 3782 tonnes (Xun, 2002), distributed in 33 germa- except for germanium-rich coals from Russia and nium ore districts in 11 provinces or autonomous China. The same holds for the extraction of Table 8.6 Active and potential producers of by-product germanium from sulfide ores and selected projects under development. Annual germanium production capacity is the installed capacity, except where the value is in parentheses when it is potential capacity.

Annual Ge production Ge resources Ge grade (maximum capacity) Status of Deposit Country Mining Company1 Type (tonnes) (ppm) in tonnes per year operation

Huize China Yunnan Huize Lead Sulfide ore 500–600 40 10 (30) production and Zinc Mine2 Jinding (Lanping) China Sichuan Hongda Sulfide ore 3000 10–100 10 production Co., state3 Fankou China Zhongjin Lingnan Sulfide ore 600 100 15 production Nonfemet, Guangdong state4 Red Dog AK, USA Teck Resources5 Sulfide ore (SMS) 1200 15 40 (200) production Lubumbashi DR Congo GTL6 Slag 2250 100–250 2 (20) production Kipushi DR Congo Unclear7 Sulfide ore (KPT) 1500 68 care and maint. (1993) Andrew and Darcy Yukon, Overland Resources Sulfide ore (SMS) 88 18 exploration Zn deposits Canada Ltd.8 Tres Marias Mexico War Eagle Mining Sulfide and 150 150 (10) exploration Company9 oxide ore (MVT) Pend Oreille WA, USA Teck Resources10 Sulfide ore (MVT) 300 10–100 (50) care and maint. (2009) Gordonsville/ TN, USA Nyrstar, Tennessee Sulfide ore (MVT) 800 20 (35) care and Elmwood Valley Resources11 maint. Kolwezi DR Congo ENRC12 Slag 500 ? development, suspended 09/2009 Tsumeb Namibia Emerging Metals Slag 530 260 (10) exploration Ltd.13 Current production ca. 80 Potential future ca. 380 production

1Refineries and smelters not included 2Ge resource from Zhang (2003); Ge grade is taken as an average of seven analyses published by Han et al. (2007) 3Ore reserves + resources of 163.2 Mt grading 6.9% Zn (Metals Economics Group, 2011); Ge reserve calculated for a (conservative) estimate of 30 ppm in Zn ore concentrate 4Probably a major Ge source; sphalerite contains 30–170 ppm Ge (Höll et al., 2007). Ore reserves are 42 Mt at 9.97% Zn, production in 2008 was 117,200 tonnes Zn in concentrate (Metals Economics Group, 2011) 5Based on proven and probable reserves of 85 Mt grading 18.2% Zn and 4.6% Pb, and an estimated Ge content of 15 ppm. 6Groupement de Terril de Lubumbashi; 2009 production was 4590 tonnes Co, 3000 tonnes Cu, 20,000 tonnes Zn oxide 7Ge reserves calculated from 23.4 Mt remaining ore resources down to 1500 m level (Kampunzu et al., 2009). 8Mineral resource (May 2009) 8.95 Mt at 7.5% Zn equivalent; 132 ppm Ge in Zn concentrate (58% Zn). Indicated resources 4.1 Mt at 18.5 g/t Ge at 3% Zn cut-off (April 2008). http://www.overlandresources.com/pdfs/InvestorPresentationRoundUp2011Vancouver-28Jan11.pdf 9Ge reserve estimated for 1 Mt ore grading 150 ppm Ge in the Tres Marias mine; total reserves on the property may be significantly bigger. Previous production (1949–1992) was 62 tonnes Ge (Saini-Eidukat et al., 2009). 10Reserve estimate based on proven and probable reserves of 4.7 Mt grading 7% Zn and 1.2% Pb, and 100 ppm Ge in the Zn concentrate 11Reserve estimates for the mines are given as 39 Mt at 3.4% Zn and 20 ppm Ge. Returned residues from Zn production at the Clarksville, Tennessee, smelter were expected to contain up to 45 tonnes per year Ga and 35 tonnes per year Ge (Guberman, 2008). 12Eurasian Natural Resources Corporation (ENRC) purchased this deposit from First Quantum in 2012 13Purchased from Ongopolo in 2008 194 frank melcher and peter buchholz

Table 8.7 Active and potential producers of by-product germanium from germanium-rich lignite and coal, and selected projects under development.

Ge resources Ge grade in Ash Deposit Country Company Type (tonnes) coal (ppm) content % Status of operation

Lincang China Lignite 1060 850 5–20 production 1 Wulantuga China Tongli Ge Refine Lignite 1700 2 240 2 21 production 1 Co., Ltd. Wumuchang China Lignite 4000 (?) 30–50 exploration Novikovsk (Central Russia SUEK 3 Lignite 1665 700 closed; produced 850 and Southern) tonnes Ge from 1960 to 1980s Luchgorsky/Bikinsk Russia CJSC 4 Lignite 2600 300 42 exploration Tigninskiy/ Russia SUEK 3 Lignite 340 53 closed Tarbagataisk Pavlovsk (Spetzugli + Russia SUEK 3 Lignite 1015 450 18–27 production Luzanovsk) Smolianinovsky/ Russia SUEK 3 Lignite 880 1043 21 exploration Shkotovsk Rakovsk Russia Lignite 380 230 Kas-Symskaya Russia Krasnoyarskaya Lignite 11,000 5 205 5 17 exploration mining co. Angren Uzbekistan JSC Uzbekugol Hard coal 180 30 2–6 currently no Ge extraction Church deposit, ND, USA Entrée Gold 6 Lignite 165 40–70 exploration Sentinel project

1Resources and average Ge contents from Seredin & Finkelman (2008), unless otherwise stated 2Ge production from coal in China is about 30 tonnes per year, with 16 to 25 tonnes from the Lincang deposit (Mikolajzak, 2011) Seredin et al. (2013) 3SUEK, Siberian Coal and Energy Company 4CJSC, Luchegorsky Fuel Energy Company 5Evdokimov et al. (2004) 6Exploration project for uranium-molybdenum in lignite-bearing sedimentary strata

germanium from coking coal and iron ores, e.g. at China, Russia, Ukraine and possibly Uzbekistan the newly developed Yakovlevskoye deposit in (see Tables 8.6 and 8.7). Estimated global germa- the Belgorod Oblast, Russia (Kats et al., 1998), nium mine production in 2010 was probably with a targetted capacity of 4.5 million tonnes more than double the reported refined germa- iron ore per year (Mining Journal, 2006). nium production (Mikolajczak, 2011; more than 300 tonnes in residues), which means that major amounts of germanium are not being extracted Production from residues. Major zinc mines outside China, which con- As with data for reserves and resources, data for tain significant amounts of germanium, are Red global germanium mine and refinery production Dog (Alaska, USA), intermittently Pend Oreille are not readily available or are of poor quality. (Washington, USA), and zinc–lead mines in Germanium is produced as a by-product from Tennessee (USA). Minor production of about two zinc mining. It is also extracted from coal ash in tonnes germanium was from the DRC (period Germanium 195

140,000

120,000

100,000

80,000

60,000

40,000 Germanium (kilograms)

20,000

0 1965 1970 1975 1980 1985 1990 1995 2000 2005

Austria Belgium Canada China DR Congo Germany Japan Namibia Russia Remaining east. World Remaining west. World Spain USA USSR/CIS

Figure 8.5 World germanium production, by country, 1962–2009. (Data from BGR database.)

2003 to 2009). This production was probably from 2009; Mikolajczak, 2011; Tse, 2010): Yunnan the remelting of slags from zinc–copper mining in Lincang (ca. 25 tonnes germanium from coal ash, the Lubumbashi area (Yager, 2009). Yunnan); Nanjing Germanium Co. Ltd. (ca. 20 World refinery production in 2012 is estimated tonnes germanium, partly from coal ash, Jiangsu), at 128 tonnes germanium, produced in China (70 Yunnan Chihong Zinc-Germanium Co. Ltd (ca. per cent), USA (ca. 2 per cent) Russia (ca. 4 per 18 tonnes germanium, Yunnan); Shaoquan cent), and other countries including Canada, Smelter (ca. 7 tonnes germanium, Guangdong); Spain, India, Finland and Australia (Guberman, Nei Mongol Xilingol Tongtai Germanium Refine 2011, 2013, and Mikolajczak, 2011). Production Co. Ltd. (Nei Mongol, Xilinhot); the latter may levels are highly volatile and the market is quite produce 100 tonnes germanium annually, if in opaque. The currently installed global extraction full production (Seredin, 2012). The total Chinese and refining capacity is estimated at about 180 production capacity is thus estimated to be 130 tonnes per year (Mikolajczak, 2011). to 200 tonnes germanium. In Russia, only the Before 1978, DRC, Namibia, Japan and USA Spetzugli lignite deposit in the Pavlovsk area is dominated world refinery production of 60–80 currently in production. According to Seredin tonnes per year. This was followed by a period of et al. (2013), the designed annual capacity of the over-production (more than 100 tonnes per year) three high-germanium (240–850 ppm) coal led by USA and other western producers. China deposits (Spetzugli, Russia; Lincang and started to enter the market around 1985. Since Wulantuga, China; Table 8.7) is 150–170 tonnes 2001, China has continuously taken the lead in of germanium. global refinery production (Figure 8.5). The major germanium producer outside China In China, five smelters/refineries account for is Teck Metals Ltd. with its metallurgical plant in the estimated 70 to 100 tonnes of germanium and Trail (Canada) with reported production of 40 germanium compounds produced (Guberman, tonnes germanium in 2007 (including germanium 196 frank melcher and peter buchholz from local and imported zinc concentrates, e.g. sing on zinc sulfide ores from large SMS and VMS from Red Dog, Alaska). deposits and on non-sulfide zinc ores. About 100 In addition to primary germanium produc- zinc exploration projects are in the conceptual tion and refining, Umicore (Belgium) and stage with little information about germanium Recylex (France) with its subsidiary PPM Pure grades. Another 50 exploration projects are in the Metals GmbH (Germany) are large global recy- prefeasibility and in the feasibility stage which clers and refiners of zinc with combined germa- are being studied in greater detail (Raw Materials nium production. They also have special Group, 2011). Although most of these zinc expertise in manufacturing and offer a wide deposits have low germanium contents, it is pos- range of germanium-based products for various sible that germanium could be extracted from applications. OJSC Germanium Krasnoyarsk, residues of the smelting and refining stages. which operates the Krasnoyarsk Non-Ferrous Carbonate-hosted zinc–lead ores that are ger- Metals Plant, is among the leading germanium manium enriched are currently of minor interest producers and the only reported major supplier to exploration companies. With the restart of from Russia (Global Industry Analysts, 2010). mining carbonate-hosted (MVT-type) zinc ores Xstrata, Vedanta Resources, Minerals and Metals carrying 15–20 ppm germanium at Gordonsville, Group – all among the five leading zinc pro- Tennessee, USA, by Strategic Resources Acqui- ducers outside China – operate several zinc sition (SRA, Toronto, Canada) in 2007, the refineries and have the potential to extract Gordonsville mining complex was expected to germanium as a by-product. become one of the world’s largest producers of Special germanium-based technologies are germanium and gallium (Table 8.6). However, produced by various companies. Prominent the mines were placed under care and mainte- examples include: Applied Materials Inc. (the nance in 2008 and SRA filed for bankruptcy in world’s largest manufacturer of semiconductor 2009. The mining complex produced a low-iron production equipment, USA); Soitec SA (the zinc concentrate, which was further smelted at world’s leading manufacturer and supplier of Nyrstar’s Clarksville smelter. The leachate Silicon-On-Insulator wafers for high-performance contained 0.3–0.8% Ga and 0.2–0.5% Ge that microprocessors, France); AXT Inc. (lighting was recovered through a series of process steps display applications, wireless and fibre-optic com- that include leaching, solution purification, munication, USA); and Germanium Corporation concentration via precipitation, and sequential of America (USA). recovery. An SRA-owned germanium–gallium recovery plant in Tennessee was planned to reach completion sometime in early 2010. Future supplies Another project of similar scale is the Tres Marias zinc mine in Mexico which grades 20% There are no known high-grade deposits that Zn and 150 ppm Ge (in some sections up to would be economic for production of germanium 245 ppm Ge). Other potential germanium-bear- as a major element (the only historic exception is ing deposits under exploration are Sierra Mojada the gallium–germanium Apex mine in Utah, (Mexico), Mehdiabad (Iran), and various MVT USA, which was in production between 1985 and deposits in Ireland (Tatestown, Ballinalack, 1987). Future germanium supplies will, therefore, Keel; Navan Group). continue to depend upon the availability of ger- The germanium-rich polymetallic sulfide ores manium-bearing residues from the processing of of the Kipushi-type (KPT) are currently operating zinc ores, from germanium-bearing coal ash and at low production levels; residues from smelting, from recycling. especially slags, constitute important future Exploration and mine expansions for zinc have resources of germanium in such deposits been at a high level in recent years, mainly focus- (Lubumbashi, Kolwezi smelters; Table 8.6). Germanium 197

Although there is potential to identify further germanium-bearing slags, ashes and residues is deposits of this class, especially in the DRC and constrained by the Chinese government by in Namibia, their remoteness and the political applying an export tax of ten per cent for these instability in the area are currently deterrents to goods. The traded volumes and transport costs investment. However, about 300 potential per unit are low, thus refined germanium and germanium-bearing sites have been reported for manufactured germanium compounds are Africa alone, excluding coal (Deschamps et al., easily shipped around the globe. Detailed trade 2006). routes are not available. Germanium may be recovered from copper Between 2000 and 2006, the volumes of global ores in porphyry deposits and in sediment-hosted annual sales did not vary significantly; they stratiform copper ores (Kupferschiefer/Germany, ranged between 88 tonnes (2003) germanium to Central African Copper Belt of DRC/Zambia). 125 tonnes germanium (2006) (Global Industry For example, if appropriate production capacities Analysts, 2010). The share of sales volumes in were developed, germanium could be extracted these countries did not vary either. These data from the residues of electrolytic zinc production. indicate that the market and trade have been Germanium-bearing coal and ue dusts con- relatively stable in the past. stitute the most important germanium resource World trade has also been inuenced by sales by volume (Seredin et al., 2013). The resources and purchases of stocks from the Defense are huge (e.g. more than 25,000 tonnes germa- Logistics Agency, DLA Strategic Materials (for- nium in germanium-rich lignite in Russia and merly Defense National Stockpile Center, USA). China, Table 8.7). In addition, large germanium However, the DLA suspended sales of germa- reserves are related to coking coals in the nium in 2011 (DLA, 2011). No sales of germa- Kemerovo region, Siberia, that carry 1.6 to 3 g/t nium metal were reported during fiscal year 2010 germanium (Kats et al., 1998). However, produc- (USGS, 2011). tion from this source will depend on the development of germanium prices and technolog- ical improvements, especially in western coun- Prices tries. Within coal deposits, germanium is enriched in distinct zones of the coal seams. Since the 1950s, germanium prices were mainly Selective mining, processing, transport and inuenced by its usage in high-tech applications. thermal use of coal from such zones would lead Before the 1970s, germanium was essential for to increased germanium grade in ashes. To install the production of diodes, rectifiers and transis- this selective production route would be a major tors, but subsequently substituted by electronic- challenge and selective mining of germanium- grade silicon. This trend was offset by the massive rich zones would not be practical in most coal demand of fibre optics in the telecommunica- deposits. In addition, recycling of germanium tions sector, in infrared night-vision devices and will remain a major source of supply in future. as a polymerisation catalyst. Due to these trends and tight supply, the first drastic price increase occurred at the beginning of 1979 when germa- World trade nium prices tripled (Brown, 1998). Between 1982 and 1994, germanium prices stagnated despite The major trade for germanium is related to the massive purchases of germanium by the US main zinc production and refining sites, notably National Defence Stockpile Programme. By mid- within North America and China, and the over- 1995, germanium metal prices doubled, and those seas trade of zinc concentrates. For recycled for germanium dioxide rose fourfold as a result of material, the dominant markets are North the strong increase in demand for fibre optics in America, Europe and Asia-Pacific. Trading of IT applications, increased use for PET production 198 frank melcher and peter buchholz

1600

1400

1200

1000

800

600 US$ per kilogram 400

200

0 Jan 92 Jan 93 Jan 94 Jan 95 Jan 96 Jan 97 Jan 98 Jan 99 Jan 00 Jan 01 Jan 02 Jan 03 Jan 04 Jan 05 Jan 06 Jan 07 Jan 08 Jan 09 Jan 10 Jan 11

Figure 8.6 Germanium dioxide price (minimum 99.99%, monthly averages), 1992–2011. (Data from BGR database, 2012.)

and anticipated increase in satellite communica- The future price of germanium will depend tion (Figure 8.6). Following strong downward on the Chinese market, breakthrough technologies price movement up to 2003, prices for germa- (see below) and various substitution possibilities nium dioxide climbed and dropped due to the already developed. boom and bust movement of the global economy. The most recent price peak in 2011 is a function of global economic growth combined with a Outlook strong demand of germanium dioxide for fibre optics and supply shortages. The supply short- Supply challenges ages are caused by the closure of the Chinese Shaoguan dioxide plant due to political clamp- Major germanium supply will be driven by pri- down on pollution- and power-intensive indus- mary zinc mine production and subsequent tries and strong demand from the Chinese defence refining, and by extraction from coal ashes recov- sector. Fears of export restrictions by Chinese vered from burning of germanium-rich coal. authorities supported higher prices. In 2012, Processing costs versus germanium price will germanium dioxide prices increased by 49 per determine the production of germanium as a cent from US$925 in mid-March to US$1375 per by-product. New demand could probably be met kilogram in late September. This can be explained through the expansion of existing capacities and by the interplay of several factors, namely the installation of new processing capacities, e.g. Chinese export taxes, shutdown of three Chinese at Red Dog, Pend Oreille, and Gordonsville/ germanium dioxide plants and speculation of Elmwood (all USA). Further capacities could be increased demand for germanium used in poly- built up in China, Russia, Mexico, the DRC and merisation catalyst following the 2011 earth- Namibia (Table 8.6). The total future production quake and tsunami in Japan (Guberman, 2013). capacity up to 2020, based on new zinc Germanium 199 exploration projects and announced production (Smirnov, 1977, Kats et al., 1998), but information capacities of existing facilities, is estimated to be regarding extraction methods or the germanium about 380 tonnes per year, which is three times content of pig iron slags using iron ores from the size of today´s refined germanium production such regions is not available. (Table 8.6). Recycling of old scrap will not be a significant Demand drivers contribution to world supply in the near future. However, post-consumer scrap may increase In fibre optics, germanium is used in Local Area slightly when solar fibre-optic cables or military Network (LAN) cables and for more sensitive devices are being replaced or if germanium-based video and audio transmissions. Programmes solar cells are widely adopted and are recycled at such as “Fibres-To-The-Home” (FTTH) in the the end of their lifetime after 15 to 25 years. USA may boost the demand of the fibre-optics Recycling of new scrap is already an important technology (Buchert et al., 2009). Zirconium– source of germanium and an individual recycling indium-based uoride glass is being studied as a rate of 70–80 per cent was reached for optical- substitute in fibre-optic applications and could fibre production in the USA (Jorgenson, 2006). impact on germanium demand, but the research Advances in recovering germanium from coal is at an early stage (Angerer et al., 2009). ash at coal-fired power plants, which is already an Future infrared optic applications, which important source in China and Russia, could could boost germanium demand, may include make a difference to future supply. The current infrared night-vision systems or car-safety sys- annual production capacity of 150–170 tonnes tems, such as adaptive cruise-control systems in germanium from three producing mines (Seredin the automotive industry, or fire rescue and detec- et al., 2013) could be considerably enhanced if tion systems. additional germanium-rich coal deposits enter Although germanium is an ideal substitute into production (e.g. Wumuchang in China and for antimony, cheaper catalysts have been devel- Bikinsk in Russia). Commercial recovery of the oped (Thiele, 2006). According to Global Industry metal appears not to be economically viable for Analysts (2010) the use of germanium in PET many plants burning germanium-poor coal production is expected to decrease in future (Bleiwas, 2010), but it appears that few samples because of other low-cost catalysts based on of ash from power plants and of coal deposits titanium as a substitute for germanium and anti- have been studied in detail for germanium. In mony. However, new technical applications for 2008, War Eagle Mining Co. entered a research germanium-based PET production may be of project with Sask Power International and interest because of the high clarity and low crys- Saskatchewan Power Corp. to sample y ash tallisation speed of germanium-based polyesters. from their plants. If successful, they have already This includes applications for TV screens or arranged an offtake agreement. Similarly, War technical yarn. Eagle Mining Co. built a small-scale laboratory A wide range of electrical applications use ger- pilot plant for y-ash germanium extraction manium, e.g. detectors and semiconductors from the Elcogas S.A. coal gasification plant in which include transistors and diodes. As an Puertollano (Spain) (y-ash production of 12,000 example of a growing market, germanium-based tonnes per year; Mining Engineering, 2008). A gallium arsenide solar cells may replace silicon weak acid solution recovered up to 83 per cent solar cells in future applications. If the tech- of the germanium from f1y ash samples over a nology becomes a commercial success in solar- two-hour leach period. Other potential primary cell production, demand for germanium could sources include germanium-rich magnetite increase significantly. Germanium is also used in from magnetite-hematite volcanic-sedimentary optoelectronics as a carrier substrate in certain deposits in Ukraine, Kazakhstan and Russia high-end LEDs that are increasingly applied in 200 frank melcher and peter buchholz displays, printers, High Definition Television Acknowledgments (HDTV), traffic lights, etc. To increase the number of transistors on a We kindly acknowledge the input of Kristof silicon chip, germanium is also used in small Dessein (Umicore), Vladimir V. Seredin (IGEM, amounts to improve their performance. New Moscow), Sandro Schmidt, Doris Homberg- generations of superchips may require more Heumann, Elke Westphale, Maren Liedtke and germanium for supercomputers based on silicon– Maria Sitnikova (BGR). We also like to thank germanium. They are also useful for chips used in Prof. Dr.-Ing. Dr.h.c. mult. F.-W. Wellmer, Kristof wireless communication products where they are Dessein (Umicore), and Teresa Brown and Gus suitable for integrating mobile phone, e-mail, and Gunn (BGS) for their reviews that greatly internet access functions. Recently developed improved the manuscript. optical discs for video recording systems by Sony Corp., based on a germanium–antimony–tellu- Notes rium (GeSbTe) alloy, have a storage capacity of eight gigabytes (Global Industry Analysts, 2010). 1. Much information, if not otherwise quoted, is taken Forecasts predict wide applications of germanium from the reviews on germanium by Butterman and in further microelectronic applications (Depuydt Jorgenson (2005) and from Höll et al. (2007) that et al., 2007). resulted from an extensive literature study conducted by Kling et al. (2000) for the BGR. 2. Metal purity is often expressed by the ‘N’ notation, Supply and demand scenario e.g. 3 N stands for 99.9% metal, and 5 N stands for The future demand for germanium is expected to 99.999% metal. be driven by the production of fibre-optic cables, which today account for about a third of total ger- manium demand. Between 2006 and 2030, the References demand may increase eight-fold from 28 tonnes to 220 tonnes in 2030 or 8.6 per cent annually Angerer, G., Erdmann, L., Marscheider-Weidemann, (Angerer et al., 2009). This additional demand F. et al.(2009) Rohstoffe für Zukunftstechnologien. Einuss des branchenspezifischen Rohstoffbedarfs could be met by the estimated 380 tonnes of new in rohstoffintensiven Zukunftstechnologien auf production capacities from zinc refining together die künftige Rohstoffnachfrage. ISI-Schriftenreihe with additional contributions from coal ash and “Innovationspotenziale“, Fraunhofer Institut für from recycling. However, as technological break- System- und Innovationsforschung. Fraunhofer throughs of potential applications, which are cur- IRB-Verlag. rently in the research or development stages, Banerjee, N.N., Ghosh, B. and Das, A. (eds.) (2000) Trace cannot be foreseen, additional end-use sectors Metals in Indian Coals. Allied Publishers Ltd., New could further increase the demand for germa- Delhi. pp. 100. nium up to 2030. The most promising applica- Bernstein, L.R. (1985) Germanium geochemistry and tions are infrared night-vision systems, car-safety mineralogy. Geochimica et Cosmochimica Acta 49, systems and fire rescue and detection systems. 2409–2422. BGR database (2012) BGR database of commodity Other growth markets are likely to include ger- prices. Federal Institute for Geosciences and Natural manium-based polyesters for TV screens and ger- Resources, Hannover, Germany. manium-based gallium arsenide solar cells. In Bleiwas, D.I. (2010) By-product mineral commodities general, supply and demand was in balance in the used for the production of photovoltaic cells. USGS past. Furthermore, the history of germanium pro- Circular 1365. duction has demonstrated that in the past the British Geological Survey (2012) World Mineral existing production capacities were not used Production 2006–2010 (Keyworth, Nottingham: completely. British Geological Survey.) Germanium 201

Brown, R.D. (1998) Germanium. In: Metal prices in the steigender Bedarf bei unzureichender Versorgungslage? United States until 1998, United States Geological Commodity Top News, 33. http://www.bgr.bund.de/ Survey. DE/Themen/Min_rohstoffe/Produkte/produkte_ Buchert, M., Schüler, D. and Bleher, D. (2009) Critical node.html#ctn metals for future sustainable technologies and their Evdokimov, A.P., Ozerskii, A.Y. and Ehanin, A.G. (2004) recycling potential. United Nations Environment Germanium-bearing lignites of the southeastern rim Programme and Öko-Institut. of the West Siberian Platform. Razvedka i ohrana Butterman, W.C. and Jorgenson, J.D. (2005) Germanium: nedr 6, 26–29. Mineral Commodity Profiles. United States Feiser, J. (1966) Nebenmetalle. Die Metallischen Geological Survey, Open-File Report 2004–1218. Rohstoffe, ihre Lagerungsverhältnisse und ihre Cassard, D., Chabod, J.C., Marcoux, E., Bourgine, B., wirtschaftliche Bedeutung. 17, Band, Ferdinand Enke Castaing, C., Gros, Y., Kosakevich, A., Moisy, M. and Verlag Stuttgart, 61–90. Villefond, L. (1996) Mise en place et origine des Gerber, G.B. and Leonard, A. (1997) Mutagenicity, minéralisations du gisement à Zn, Ge, Ag, (Pb, Cd) de carcinogenicity and teratogenicity of germanium Noailhac-Saint Salvy (Tarn, France). Chronique de la compounds. Mutation Research 387, 141–146. Recherche Minière 514, 3–37. Global Industry Analysts Inc. (2010) Germanium – a Chemsystems (2011) Polyester shows rapid recovery. global strategic business report. MCP-2637. Abstract, Petrochemical Market Dynamics – Polyester Goldschmidt, V.M. (1930) Über das Vorkommen des & Intermediates, Nexant Limited. http://www.chem- Germaniums in Steinkohlen und Steinkohlenprodukten. systems.com/about/cs/news/items/PPE%20 Nachrichten von der Gesellschaft der Wissenschaften PCMD%20Aromatics%2020101.cfm zu Göttingen. Mathematisch-Physikalische Klasse, 28, Cook, N.J., Ciobanu, C.L., Pring, A., Skinner, W., 398–401. Shimizu, M., Danyushevsky, L., Saini-Eidukat, B. and Goldschmidt, V.M. and Peters, C. (1933) Zur Geochemie Melcher, F. (2009) Trace and minor elements in sphal- der Germaniums. Nachrichten von der Gesellschaft erite: A LA-ICPMS study. Geochimica et der Wissenschaften zu Göttingen. Mathematisch- Cosmochimica Acta 73, 4761–4791. Physikalische Klasse 31, 141–166. Dai, S., Wang, X., Seredin, V.V. et al. (2012) Petrology, Gouin, J., Lerouge, C., Deschamps, Y. et al. (2007) mineralogy, and geochemistry of the Ge-rich coal Petrological and geochemical characterisation of ger- from the Wulantuga Ge ore deposit, Inner Mongolia, manium-bearing coals from the eastern Rhodopes, China: New data and genetic implications. Bulgaria. Proceedings of the Ninth Biennial SGA International Journal of Coal Geology 90–91, 72–99. Meeting, Dublin, 1521–1524. Dasch, E.J. (ed.) (1996) Encyclopedia of Earth Sciences, Guberman, D.E. (2008) Germanium. USGS 2007 Volume I. MacMillan Reference USA, New York. Minerals Yearbook, United States Geological Survey. Depuydt, B., De Jonghe, M., De Baets, W. et al. (2007) Guberman, D.E. (2009) Germanium. USGS 2008 Germanium materials. In: Claeys, C. and Simoen, E. Minerals Yearbook, United States Geological Survey. (eds.) Germanium Based Technologies, from Materials Guberman, D.E. (2011) Germanium. USGS 2009 to Devices. Elsevier. 11–40. Minerals Yearbook, United States Geological Survey. Deschamps, Y., Milési, J.P., Hocquard, C. et al. (2006) Guberman, D.E. (2013) Germanium: Mineral Germanium potential in Africa. A predictive Commodity Summaries 2013, United States approach methodology. CAG21 – Colloquium of Geological Survey, 64–65. African Geology, 21st Maputo, Mozambique, Haller, E.E. (2006) Germanium: from its discovery to 03-05/07/2006, pp. 283. SiGe devices. Materials Science in Semiconductor Defense Logistics Agency (DLA) (2011) Germanium, Processing 9, 408–422. BOA sales. Strategic Materials. https://www.dnsc. Han, R.S., Liu, C.Q., Huang, Z.L., Chen, J., Ma, D.Y., dla.mil/pgm.asp?Commodity=Germanium Lei, L. and Ma, G.S. (2007) Geological features and Du, G., Zhuang, X., Querol, X. et al. (2009) Ge distribution origin of the Huize carbonate-hosted Zn-Pb-(Ag) in the Wulantuga high-germanium coal deposit in the District, Yunnan, South China. Ore Geology Reviews Shengli coalfield, Inner Mongolia, northeastern China. 31, 360–383. International Journal of Coal Geology 78, 16–26. Höll, R., Kling, M. and Schroll, E. (2007) Metallogenesis Elsner, H., Melcher, F., Schwarz-Schampera, U. and of germanium – A review. Ore Geology Reviews 30, Buchholz, P. (2010) Elektronikmetalle – zukünftig 145–180. 202 frank melcher and peter buchholz

Hu, R.Z., Qi, H-W., Zhou, M-F. et al. (2009) Geological Melcher, F., Oberthür, T. and Rammlmair, D. (2006) and geochemical constraints on the origin of the giant Geochemical and mineralogical distribution of ger- Lincang coal seam-hosted germanium deposit, manium in the Khusib Springs Cu-Zn-Pb-Ag sulfide Yunnan, South West China: A review. Ore Geology deposit, Otavi Mountain Land, Namibia. Ore Geology Reviews 36, 221–234. Reviews 28, 32–56. Jorgenson, J.D. (2006) Germanium recycling in the United Metals Economics Group (2011) Mine Search Database. States in 2000. USGS Circular 1196–V, Flow studies for Metals Economics Group, Canada. recycling metal commodities in the United States. Mikolajczak, C. (2011) The economic linkage between Kammer, U. (2009) Recycling von seltenen Metallen Zinc and Germanium. International Lead and Zinc und deren Verbindungen. In: Thomé-Kozmiensky, Study Group, Lisbon, Portugal, September 29, 2011, K.J. und Goldmann, D.: Recycling und Rohstoffe. presentation, pp. 24. Band 2, TK-Verlag, Neuruppin, 647–655. Mining Engineering (2008) War Eagle to sample y ash Kampunzu, A.B., Cailteux, J.L.H., Kamona, A.F., for germanium. Volume. 60 (2), page 35. Intiomale, M.M. and Melcher, F. (2009) Sediment- Mining Journal (2006) Russia. http://www.mining-jour- hosted Zn-Pb-Cu deposits in the Central African nal.com/reports/russia---2006-last-updated-septem- Copperbelt. Ore Geology Reviews 35, 263–297. ber-2 Kats, A.Ya., Kremenetsky, A.A., Podkopaev, O.I. (1998) Naumov, A.V. (2007) World market of germanium and The germanium mineral resource base of the Russian its prospects. Russian Journal of Non-Ferrous Metals Federation. Mineral Resources of Russia, Economics 48, 265–272. and Management 1998 (3), 5–9. PPM Pure Metals (2012) Germanium. http://www.ppm- Kelley, K.D., Leach, D.L., Johnson, C.A. et al. (2004) puremetals.de/products/germanium/germanium. Textural, compositional, and sulfur isotope varia- php?inhalt=inhalt_gr.htm tions of sulfide minerals in the Red Dog Zn-Pb-Ag Oakdene Hollins Ltd. (2011) Study into the feasibility deposits, Brooks Range, Alaska: implications for ore of protecting and recovering critical raw materials formation. Economic Geology 99, 1509–1532. through infrastructure development in the south east Ketris, M.P, Yudovich, Ya.E. (2009) Estimations of of England. European Pathway to zero waste, Clarkes for Carbonaceous biolithes: World averages Environment Agency, UK. for trace element contents in black shales and Queneau, P.B., Avotins, P.V. and Farris, L.F. (1986) coals. International Journal of Coal Geology 78, Germanium recovery from Fargo Oil´s Lang Bay prop- 135–148. erty. CIM Bulletin 79 (886), 92–97. Kling, M., Höll, R. and Schroll, E. (2000) Literaturstudie Raw Materials Group (2011) Raw materials data. Raw über Germanium in Gesteinen, Mineralien, Erzen Materials Group, Sweden. und Lagerstätten. Berichte zur Lagerstätten- und Roskill Information Services Ltd. (1988) The economics Rohstoffforschung, 39. Bundesanstalt für Geowissen- of germanium. schaften und Rohstoffe, Hannover. Saini-Eidukat, B., Melcher, F. and Lodziak, J. (2009) Leach, D.L., Sangster, D.F., Kelley, K.D., et al. (2005) Zinc-germanium ores of the Tres Marias Mine, Sediment-hosted lead- zinc deposits: A global per- Chihuahua, Mexico. Mineralium Deposita 44, 363– spective. Economic Geology 100th Anniversary 370. Volume 561–607. Schneider, J., Melcher, F. and Brauns, M. (2007) Levine, R.M. and Wallace, G.J. (2000) The mineral indus- Concordant ages for the giant Kipushi base metal tries of Armenia, Azerbaijan, Belarus, Georgia, deposit from direct Rb-Sr and Re-Os dating of sul- Kazakhstan, Kyrgyzstan, Moldova, Russia, Tajikistan, fides. Mineralium Deposita 42, 791–797. Turkmenistan, Ukraine, and Uzbekistan. 2000 Minerals Seredin, V.V. (2012) Proizvodstvo germaniya iz buryh Yearbook, United States Geological Survey, 1–48. uglei mestorozhdeniya Ulantuga. Energetik 2012, Leutwein, F. and Rösler, H.J. (1956) Geochemische N3, 19–23 (in Russian) Untersuchungen an paläozoischen und mesozoischen Seredin, V.V. and Finkelman, R.B. (2008) Metalliferous Kohlen Mittel- und Ostdeutschlands. Freiberger coals: A review of the main genetic and geochemical Forschungshefte, C19, 1–196. types. International Journal of Coal Geology 76, Melcher, F. (2003) The Otavi Mountain Land in 253–289. Namibia: Tsumeb, germanium and Snowball Earth. Seredin, V.V., Dai, S., Sun, Y., Chekryzhov, I.Yu. (2013) Mitteilungen der Österreichischen Mineralogischen Coal deposits as promising sources of rare metals for Gesellschaft, 148, 413–435. alternative power and energy-efficient technologies. Germanium 203

Applied Geochemistry (2013), 31, 1–11 doi: http:// UNEP. (2011) Recycling rates of metals – a status report. dx.doi.org/10.1016/j.apgeochem.2013.01.009 A Report of the Working Group on the Global Metal Smirnov, V. I. (1977) Deposits of Germanium. In: Flows to the International Resource Panel. Smirnov, V.I. Ore Deposits of the USSR. Vol III. U.S. Geological Survey, USGS. (2011) 2010 Minerals Pitman Publishing Pty. Ltd. 455–462. Yearbook: Germanium. Swaine, D.J. (1990) Trace Elements in Coal. U.S. Geological Survey, USGS. (2013) Mineral Butterworths, London, pp. 278. Commodity Summaries: Germanium. Taylor, S.R. and McLennan, S.M. (1985) The Continental Xun, Z. (2002) Mineral Facts of China. Science Press, Crust: its Composition and Evolution. Blackwell, Beijing, China. Oxford. Yager, T.R. (2009) The Mineral Industry of Congo Taylor, A. and Quinn, A.J. (1987) Musto Explorations – (Kinshasa). 2009 Minerals Yearbook, United States case study of financing of a small, high-technology Geological Survey. mine. Transactions of the Institution of Mining and Yang, X., Yun, Z., Yuxia, G. and Banghui, G. (2003) A Metallurgy, Section A, Mining Industry 96, 23–29. novel biotechnological process for germanium Thiele, U.K. (2006) Polyester Catalysts: A critical anal- recovery from brown coal. Proceedings Inter- ysis of current technology and available alternatives. national Biohydrometallurgy Symposium IBS The European PET Conference, 11–12th October 2003, Athens. 2006, conference presentation, Barcelona. http:// Ye, L., Cook, N.J., Ciobanu, C.L. et al. (2011) Trace and www.polyester-technology.com/Publication/ minor elements in sphalerite from base metal BARCELONA-SCRIPT.pdf deposits in South China: A LA-ICP-MS study. Ore Tse, P-K. (2010) The mineral industry of China. 2009 Geology Reviews 39, 188–217. Minerals Yearbook, United States Geological Survey. Zhang, M. (2003) The geochemical feature and enrich- 1–25. ment mechanism of Yunnan Huize Pb-Zn- Umicore. (2012) Electro-optic materials. http://eom. germanium-Cd deposits. Masters Thesis. Chengdu umicore.com/en/eom/ University of Technology. 9. Indium

ULRICH SCHWARZ-SCHAMPERA

Federal Institute for Geosciences and Natural Resources (BGR), Stilleweg, Hannover, Germany

Introduction these properties at very low temperatures approaching absolute zero, making it ideal for Indium was discovered in 1863 by two German cryogenic and vacuum applications. Indium does chemists, Ferdinand Reich and Hieronymous not workharden, endures considerable deforma- Theodor Richter, who were testing zinc ores from tion through compression, and it is easily cold- the mines around Freiberg, Saxony. They named it welded. Indium metal is not oxidised by air at ‘indium‘ from the indigo blue colour seen in its ordinary temperatures, but burns to trioxide spectrum. In 1924, indium was found to have a (In(III)) at high temperatures. On heating, indium valuable ability to stabilise non-ferrous metals, reacts directly with metalloids (arsenic, anti- which was the first significant use for the element. mony, selenium, tellurium) and with halogens, However, the early applications of indium were sulfur and phosphorus. It dissolves in mineral few, the most important being in light-emitting acids and amalgamates with mercury but is not diodes and in coating bearings in high-speed motors affected by alkalis, boiling water and most organic such as aircraft engines. Indium-containing semi- acids. The chemistry of trivalent indium is char- conductors became important from the 1950s acterised by covalent bonding. Indium is fre- onwards, while the widespread use of indium-con- quently used for glass coatings: as indium metal taining nuclear control rods increased demand dur- it forms a mirror surface with reflective prop- ing the 1970s. Since 1992 the major application of erties equal to that of silver and with greater cor- indium has been in the form of indium–tin oxide rosion resistance; or in alloys to form transparent (ITO) in liquid-crystal displays. This use now dom- and conductive coatings. Selected key character- inates the market accounting for more than half of istics and physicochemical properties of indium total indium consumption. are listed in Table 9.1. Indium is a post-transition metal of Group 13 of the Periodic Table falling between gallium and Physical and chemical properties thallium. The geochemical properties of indium are such that it tends to occur in nature with base- Indium is a soft, lustrous, silver-white metal, metal Groups 11 (Cu, Ag), 12 (Zn, Cd), 14 (Sn, Pb), with a face-centred tetragonal crystalline struc- and 15 (Bi) of the Periodic Table. Indium has two ture. It is very malleable and ductile and retains main oxidation states, +3 (III) and +1 (I). Naturally

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Indium 205 occurring indium consists of two isotopes, 113In (4.3 Mineralogy per cent of total) and 115In (95.7 per cent of total). Indium minerals are rare in natural systems. Twelve indium mineral phases have been Abundance in the Earth’s crust defined (Table 9.2). Roquesite is the most impor- tant indium mineral representing a trace compo- The Earth’s continental crust is estimated to con- nent in the principal ore-forming minerals like tain about 0.05 ppm In, with 0.072 ppm In in the bornite, chalcopyrite and sphalerite (Figure 9.1). oceanic crust, about the same as silver. Indium is More often, indium substitutes for elements with a highly volatile chalcophile element which similar ionic radii, especially those having tetra- behaves in a moderately to highly incompatible hedral coordination with respect to the principal manner during mantle melting. The meteoritic metal ion, in base-metal sulfides. High indium abundance of indium is 0.08 ppm for chondrite concentrations usually occur in sphalerite, chal- (CI), while seawater contains 0.2–0.7 ppb In. copyrite, stannite, tin-sulfosalts, tennantite and cassiterite. The incorporation of indium and the formation of indium-bearing sulfides may occur Table 9.1 Selected properties of indium. by (i) diadochic replacement and/or coupled Property Value Units substitution of iron, copper, tin and arsenic form- ing solid-solution series, (ii) the incorporation Symbol In into the lattice of tetrahedral coordinated sul- Atomic number 49 fides, and (iii) the formation of sub-microscopic Atomic weight 114.82 inclusions of indium minerals (e.g. roquesite, Density at 25°C 7290 kg/m3 Hardness (Mohs scale) 1.5 indite) in sulfides or cassiterite. Melting point 157 °C Sphalerite is the most important indium- Boiling point 2072 °C bearing mineral and the source of most indium Specific heat capacity at 25°C 0.23 J/(g °C) currently mined. High indium concentrations in Electrical resistivity at 25°C 0.08 µΩ m sphalerite are closely associated with elevated Thermal conductivity 82 W/(m °C) copper contents due to the formation of a dis-

Table 9.2 Compilation of defined indium minerals.

Mineral Formula Indium content (%) Specific gravity Appearance (colour, lustre)

Roquesite CulnS2 47.35 4.80 bluish-grey, metallic

Laforetite AglnS2 40.03 4.92 brown, metallic

Indite Feln2S4 55.50 4.59 iron-black, metallic

Sakuraiite (Cu,Zn,Fe,Ag)3 (In,Sn) S4 24.35 4.34 greenish-grey, metallic

Petrukite (Cu,Fe,Zn)3 (Sn,In) S4 6.05 4.07 grey, brown, metallic

Abramovite Pb2SnlnBiS7 11.41 9.0 silver-grey, metallic

Cadmoindite Cdln2S4 49.58 4.85 black, dark brown, adamantine

Dzhalindite In(OH)3 69.23 4.37 yellow brown

Yanomamite InAsO4·2H2O 39.62 3.92 blue green, light yellow green, vitreous

Yixunite Pt3ln 16.40 18.33 white, metallic

Damiaoite Ptln2 54.07 10.94 white, metallic Native indium In 100 7.29 yellow grey, metallic 206 ulrich schwarz-schampra

(a) (b)

(c) (d)

Figure 9.1 (a) Example of indium-rich ore (600 ppm In), Maranda J volcanic-hosted massive sulfide (VHMS) deposit, South Africa and indium rods (99.95% In); (b) Native indium metal from the former Kidd Creek processing plant, Canada (courtesy of Falconbridge Ltd); (c) Chalcopyrite-sphalerite ore with roquesite inclusions in sphalerite; Kidd Creek, Canada; (d) Sphalerite-roquesite solid solution series as light zonal enrichments in sphalerite grains from the Vai Lili hydrothermal vent field, southern Lau basin, south-west Pacific.

continuous binary solid solution of CuInS2 of metals such as tin, copper, zinc, lead, silver, (roquesite) forming solid-solution series mem- bismuth, selenium and arsenic.

bers of [Zn2–2xCuxInxS2] composition (Figure 9.1). Co-precipitation of indium with copper, zinc and iron in chalcopyrite accounts for the ele- Major deposit classes vated concentrations in copper ores. The highest indium concentrations commonly occur in the Indium occurs in different types of ore deposits of all chalcopyrite-rich high-temperature cores or at ages, from the presently forming deposits at modern, the base of the individual deposits. The miner- actively spreading ridges and fumarole precipitates alogy of indium-bearing deposits is character- of active volcanoes to deposits in the Archaean ised by complex intergrowths and replacement volcanic strata of greenstone belts, e.g. in Canada textures containing significant concentrations and South Africa (Table 9.3). In decreasing order of Figure 9.2 The global distribution of indium mines, deposits and major occurrences. 208 ulrich schwarz-schampra

Table 9.3 Size and grade of the major types of indium deposits (grades and tonnages are very variable between deposits and figures given are indicative only).

Deposit size Typical grade Estimated indium metal content Deposit type range (tonnes) (indium g/t) of known deposits (tonnes) % of total

Volcanic-hosted massive sulphide 106–10 8 20–200 13,750 43 Sediment-hosted massive sulphide 106–10 8 20–200 11,750 37 Epithermal 104–10 6 10–800 4500 14 Tin-tungsten porphyry 105–10 7 10–400 1250 4 Vein-type 104–10 6 10–350 520 2 Total 31,770 100

significance indium-rich deposits are represented by to intermediate volcanic and intrusive host rocks, volcanic- and sediment-hosted exhalative massive often in a volcanic cauldron setting, which display sulfide deposits, epithermal deposits, polymetallic certain types of alteration (silicification, sericitisa- base metal vein deposits, granite-related tin–base- tion, chloritisation, greisenisation) of the original metal deposits, skarn deposits and porphyry copper host rocks. Volcanic host rocks are usually subma- deposits. These deposits are commonly associated rine, subaerial and pyroclastic deposits of rhyolite, with active oceanic or continental-plate margins dacite and andesite. Subvolcanic to hypabyssal and orogenic belts with steep geothermal gradients intrusive host rocks comprise granite–granodio- due to enhanced magmatic activity. rite–quartz–monzonite differentiates. Indium has Indium deposits are most commonly associ- an affinity for ore deposits that have some mag- ated with base-metal hydrothermal ore-forming matic contribution or indicate magmatic-derived systems enriched in zinc, copper, lead and tin, components, as demonstrated by the occurrence of accompanied by trace metals such as bismuth, a number of epithermal and skarn-type deposits. cadmium and silver. Indium is usually found They are commonly related to multiple and highly primarily in zinc and copper sulfide ores, and in variable magma sources and combined magmatic- tin ores. Typical indium-rich ores have zinc meteoric element precipitation processes including concentrations between 10 and 22 wt% and magma degassing, wall rock leaching and mag- copper concentrations at or above 2 wt%. Because matic-meteoric fluid mixing (Schwarz-Schampera of its relatively low concentration in these ores and Herzig, 2002). indium can only be economically extracted as a Indium-bearing ores occur over a wide range of by-product under appropriate technical and geological time and are found in a variety of deposit processing conditions. types and provinces. Collision tectonics, mobilisa- The most important deposits are volcanic- and tion from the subducting slab, magmatic recycling sediment-hosted base-metal sulfide deposits, and metallogenic processes within the continental which are generally characterised by high metal crust or continental fragments are most impor- abundance and large tonnages (Table 9.3). The tant. The majority of indium-bearing ore deposits concentration of indium in these ores is in the are associated with the subduction-related west- range 20–200 ppm. Even in a zinc concentrate, ern Pacific plate boundaries, especially in east and representing the most common commercial south-east Asia. A second indium province can be source, the indium concentration is relatively delineated at the Nazca–South American plate low at 70–200 ppm, but may reach 500–800 ppm. boundary in Bolivia and Peru with a possible Indium-bearing deposits are closely associated continuation along the western plate margins of with calc-alkaline to peralkaline, porphyritic felsic North America. Other provinces are related to Indium 209

different metallogenic epochs in central Europe siliciclastic sediments, with minor intercalations covering the Hercynian and Alpine belts. Indium- of carbonate layers. SHMS deposits occur in ter- rich deposits also occur in the Mesozoic strata ranes dominated by sedimentary strata. Volcanic of Patagonia, Argentina, in the Caledonian- rocks (lava, tuff) may be a minor component of the Appalachian belt (New Brunswick, Canada) and in associated strata, and synsedimentary intrusions Archaean greenstone belts (Canada, South Africa). (mafic sills, dikes) may be present regionally. Zinc, lead and silver are the primary metals recov- Base-metal sulfide deposits ered from these deposits, but tin and indium may be important by-products. Base-metal sulfide deposits originate at oceanic Active seafloor hydrothermal systems, which spreading and rift zones, both at mid-ocean are modern analogues of ancient VHMS and SHMS spreading axes and in back-arc rift zones. Volcanic- deposits, may have elevated indium concentra- hosted massive sulfide (VHMS) deposits occur in tions (Schwarz-Schampera and Herzig, 2002). The submarine volcanic rocks of all ages, from the highest known values in active hydrothermal sys- presently forming deposits in modern, actively tems were detected in copper-rich sulfides associ- spreading ridges to deposits in the pre-3400 Ma ated with back-arc spreading centres propagating volcanic strata of the Pilbara Block in Australia into island-arc crust (Lau Basin, Eastern Manus (Galley et al., 2007 and Schwarz-Schampera et al., Basin). Back-arc spreading centres are character- 2010). They are major sources of base and precious ised by bimodal volcanism and the different vent metals, and are the most important indium pro- fields are closely associated with highly evolved ducers in the world. Polymetallic examples of the magmas, derived from partial melting of an older zinc–lead–copper group are found in Canada at oceanic crust. Assimilation of older crust may Kidd Creek, Ontario and in deposits of the Iberian have been an important factor in developing highly Pyrite Belt in Spain and Portugal. Deposits of the fractionated arc-like magmas primarily enriched zinc–lead–copper group are characteristically dom- in incompatible elements such as indium. inated by felsic host rocks, with mafic volcanic rocks rare or absent. A distinct metallogenic zona- Polymetallic vein-type deposits tion of the orebodies is characteristic of the majority of the ore deposits. An iron-rich sulfate- Indium was first discovered in hydrothermal base and quartz-rich exhalite top is typically underlain metal vein deposits in the Erzgebirge region of south- by a massive zinc–iron–lead–tin (sphalerite– east Germany (Schwarz-Schampera and Herzig, pyrite–galena–cassiterite) orebody. A copper-rich 2002; Seifert and Sandmann, 2006). Polymetallic keel and vein-type mineralization (chalcopyrite) vein-type deposits are structurally controlled and underneath represent the high-temperature feeder occur in faults, fault systems and vein breccia zones zones of fluid upflow (referred to as the ‘stockwork of clastic metasedimentary or magmatic-dominated zone’). The deposits contain significant and recov- terranes. The hydrothermal fluids were derived from erable amounts of trace elements like silver, tin, basinal brines and may contain contributions of bismuth, cobalt and indium; bismuth, cobalt, and magmatic volatiles in magmatically active domains indium typically indicate elevated formation tem- in syn- to post-collisional orogens. The deposits peratures. Modern examples of this type have been include lead-zinc-silver-, copper-zinc-lead-silver-tin-, discovered in the south-west Pacific (Lau basin, and tin-tungsten-base-metal-dominated mineralisa- Manus basin). tion. In the case of magmatic host rock lithologies, Stratiform sediment-hosted base metal sulfide intermediate to felsic sub-volcanic intrusions clearly (SHMS) deposits are concordant, massive to semi- predominate. Polymetallic vein deposits may con- massive accumulations of sulfide and sulfate min- tain considerable and often recoverable indium con- erals that formed by exhalative processes on or centrations and were among the most important immediately below the seafloor (Goodfellow and indium producers worldwide until the 1980s, for Lydon, 2007). The ore is syngenetic with the host example in Japan and Bolivia. 210 ulrich schwarz-schampra

Base-metal-rich tin–tungsten important by-products, especially in deposits with and skarn deposits high silver grades. Indium deposition also occurs in active mag- Vein-stockwork deposits of tin and tungsten matic systems. Indium mineralisation is reported occur in a wide variety of structural styles that from fumaroles of high temperature fluids (500– include individual veins, multiple vein systems, 940 °C) forming sulfide segregations in altered vein and fracture stockworks, skarns, breccias wall rocks (Kovalenker et al., 1993). Relative to and replacement zones in altered wall rocks adja- the source magma, fumarole precipitates and cent to veins. The deposits generally occur in or gases are enriched by factors greater than 104 in near granitic intrusions which have been indium and other trace metals. emplaced at relatively shallow levels in the The key characteristics and examples of the Earth’s crust. The associated intrusions are highly major types of indium deposits are given in Table 9.4. fractionated and typically enriched in lithophile as well as highly volatile elements. Porphyry tin deposits usually formed in areas of great Extraction methods and processing continental thickness related to collisional tec- tonic settings, although the deposits generally Mining post-date the collision event. Vein stockwork tin–tungsten as well as porphyry tin deposits The mining methods employed to extract indium are important hosts for indium mineralisation are related to the recovery of other base metals in (Sinclair et al., 2006). massive sulfide deposits. The exclusive extraction of indium is uneconomic and consequently primary indium is obtained solely as a by-product in the Base-metal-rich epithermal deposits smelting of polymetallic ores containing zinc, copper Epithermal deposits originate at active subduction and tin. The strict co-enrichment with zinc and zones and are closely associated with volcanic copper makes copper-bearing zinc ores the most processes above metalliferous intrusive rocks. The favourable ores for the indium recovery. The appro- development of epithermal deposits is typically priate mining method typically depends on the linked to the shallow crust emplacement of mag- physical and chemical properties of the minerals, matic porphyry copper systems. Two principal sub- tonnage and grade, the geometry and the depth of the types of epithermal deposits are distinguished on orebody. The most common extraction technique the basis of alteration and ore mineral assemblages: for massive base-metal sulfide deposits is usually a the quartz-(kaolinite)-alunite or high-sulfidation combination of surface and underground mining, sub-type and the adularia-sericite or low-sulfida- although all economically significant indium- tion sub-type (Heald et al., 1987 and Hedenquist, bearing ores are mined by underground methods. 1987). Differing acidity and oxidation state largely The mining of zinc-rich massive sulfide determine the end-member hydrothermal environ- deposits usually starts as open pit mining, if the ments of these two deposit types. Temperatures of deposit is near surface. Overburden, altered formation for epithermal deposits range from about volcanic or sedimentary wall-rock, and waste 100 °C for hot-spring or steam-heated deposits to rock may be deposited and stored for later back- about 350–400 °C for deeper vein and replacement fill. Massive sulfide deposits are usually a few deposits. Pronounced changes in the physical, hundred metres in diameter and comprise clearly thermal and chemical properties of the hydro- defined orebodies. Open pit mining may reach thermal solutions may occur over short distances considerable depths but rarely exceeds a few hun- in these systems and promote ore deposition. Gold dred metres before going underground. and silver are the principal commodities of epither- If surface mining results in high production mal deposits. Copper and other base metals (zinc, costs and geological conditions are complex, then the lead, tin) and rare metals (bismuth, indium) are deeper parts of a deposit may be mined underground. Indium 211

Table 9.4 Key characteristics and examples of the major types of indium deposits.

Deposit type Brief description Features Examples

Volcanic-hosted massive Concordant lenses of pyrite-sphalerite- Polymetallic examples of Neves-Corvo (Portugal), Kidd sulfide chalcopyrite-galena in submarine the Zn-Pb-Cu group dominated Creek, Gai (Russia), Deerni, volcanic sequences. Syngenetic by felsic host rocks; distinct Hongtoushan (China). volcanogenic exhalative origin. metallogenetic zonation; significant amounts of trace elements like Ag, Sn, Bi, Co, In. Sediment-hosted massive Concordant stratiform lenses Massive polymetallic Red Dog (USA), Broken Hill, sulfide of pyrite-sphalerite-galena- accumulations of sulfide and HYC (Australia), Dachang (chalcopyrite) in submarine sulfate minerals; replacement (Pb-Zn; China), Brunswick sedimentary siliciclastic sequences. processes immediately below #12 (Canada). Syngenetic exhalative origin. the seafoor; minor intercalations of carbonate layers or volcanic rocks; synsedimentary intrusions may be present; In may be important by-product. Epithermal Medium to large, low-grade stockwork- Poylmetallic examples of Toyoha (Japan), Equity Silver, type quartz-sulfide veinlets and the adularia-sericite or Hemlo (Canada), Kirki disseminations in felsic volcanic low-sulfidation subtype; (Greece), Lepanto sequences in a subduction zone wallrock equilibration of (Phillippines), Cerro and island arc setting. Syngenetic magmatic fluids; In is Vanguardia (Chile), exhalative origin. associated with Au-poor McLaughlin (USA). but Ag-rich Zn-Pb-Cu ore. Porphyry Medium to large, low-grade vein- Base metal-rich ores in concentrically Mount Pleasant (Canada), stockwork deposits of Sn-W and zoned altered and mineralized Altenberg (Germany), Zn-Cu-Pb in subvolcanic felsic granites; highly host rocks Dachang (Sn-W; China), intrusive rocks and replacement typically enriched in lithophile Qibaoshan (China). zones in altered wall rocks adjacent and highly volatile elements; In is to veins. Post-collisional tectonic associated with the polymetallic settings. Zn-Cu-Pb-Sn zone. Polymetallic Vein Small to medium polymetallic Polymetallic Pb-Zn-Ag-, Cu-Zn-Pb- Freiberg (Germany), deposits; structurally controlled in Ag-Sn-, Sn-W-base metal ores; Colquechaca, Pulacayo, post-collisional vein breccia zones densely intergrown ore minerals Bolivar (Bolivia), Akenobe of clastic metasedimentary or (telescoping); In hosted by (Japan), West Shropshire magmatic-dominated terranes; sphalerite, partly replaced by (Great Britain). mineralisation by basinal brines. chalcopyrite; quartz gangue more prospective than carbonate gangue. Hot spring exhalative Siliceous precipitates High temperature fluids forming Hydrothermal vent fields: Vai deposited by hydrothermal fluids, sulfide segregations in altered Lili, Whitechurch (Lau basin), hot springs, fumaroles; volcanic intermediate to felsic wall rocks; Pacmanus (E. Manus basin); activity. In co-enrichment with Zn and magmatic fumaroles at Cu (Cd, Te, Se). Kudryavyi, Kuril islands (Russia), Merapi (Indonesia).

Zn, zinc; Pb, lead; Cu, copper; Ag, silver; Sn, tin; Bi, bismuth; Co, cobalt; In, indium; W, tungsten; Te, tellurium; Se, selenium; Cd, cadmium. 212 ulrich schwarz-schampra

Copper-rich massive sulfide deposits, which are roasting is leached to provide a ‘pregnant’ zinc-rich the major hosts for indium, and copper-rich ore- solution from which cadmium is removed. The bodies often occur in the deeper sections of an purified zinc-rich solution is sent to the electrol- individual deposit. Block caving and stoping are ysis plant where zinc is plated onto aluminium underground mass mining methods appropriate cathodes. The jarosite process (formation of iron to large-scale production from massive orebod- sulfate) for the zinc refining and iron fixation scav- ies with large vertical and horizontal dimen- enges up to 60 per cent of the indium and germa- sions. Stoping is applicable when the deposit lies nium from solution. Remains from the jarosite within stable hanging and footwall rocks, process may contain up to 600 ppm In. As a includes competent ores and host rocks, and has consequence of the electrochemical series, the regular ore boundaries. Stoping techniques dom- insoluble components (impurities), such as indium, inate the mining of most underground massive settle as solids at the bottom of the electrolysis sulfide deposits. cells forming anode slime. In some ores this metal- rich slime may be an important potential source of indium and a number of other trace metals such as Processing, beneficiation silver, cadmium and selenium. Where the appro- and conversion to metal priate technology is installed the pure metals are Indium is produced almost exclusively as a isolated from the anode slimes by further processing by-product of zinc mining and processing. It is involving a combination of dissolution and precip- extracted from the residues of concentrating itation methods. and smelting zinc ores and from the recycling of Indium can be dissolved from anode slime or dusts and gases produced during the smelting the processed intermediates from silver and of zinc (Alfantazia and Moskalyk, 2003). A small cadmium recovery with hydrochloric acid or proportion, less than 5 per cent of the total, is sulfuric acid. Anode slime residues may also be extracted from the residues of copper and tin processed in lead smelters. The ’imperial smelt- treatment. Base-metal sulfides are generally sepa- ing process’ (ISP) enriches about 50 per cent of rated by flotation techniques with indium typi- the indium with the lead fraction (bullion). The cally enriched in zinc concentrates, comprising remaining indium stays with the waste blast- mainly sphalerite, with elevated copper concen- furnace slag and may be recovered by roasting in trations at or above 2 wt%. a rotary furnace and selective vaporisation (slag Techniques for indium recovery are capital fuming). Indium is separated from the lead intensive. In the past, when indium consumption fraction during slagging and by electrothermal was small, demand was satisfied by processing very processes which also produce lead, antimony and large amounts of low-grade indium-bearing zinc tin. Electrolytic operations are employed to pro- ore. Low commodity prices and problems with sep- duce an indium anode and a slime containing aration from elements like copper, lead, tin, anti- 20–25% In. The slime is chemically treated to mony and arsenic made the recovery of indium and provide a crude indium metal of about 99% other trace metals costly, inefficient and generally purity. Solvent liquid extraction leads to the uneconomic. However, progress in the treatment of electrolytic refinery of indium metal (Demarthe process slimes since the end of the 1990s has et al., 1990). improved the beneficiation and reduced costs. The Kidd Creek zinc smelter at Timmins, Impure zinc and copper concentrates, contain- Ontario, represented a significant indium producer ing more than 80–90 per cent sphalerite and chal- and developed a successful process for the recovery copyrite respectively, are normally produced at or of indium from zinc ores of the nearby Kidd Creek near the mine site using flotation techniques. The VHMS deposit (Figure 9.3). The average indium zinc concentrates are roasted, and the sulfur concentration in the concentrate was approxi- dioxide produced is used in the manufacture of mately 270 ppm (Jorgensen and George, 2004). The sulfuric acid. The zinc calcine derived from the indium flowsheet starts with the roasting of the Indium 213

Cottrell dust Zinc concentrate from copper Indium smelter Sulfur dioxide leaching Roast for production Copper cake of sulfuric acid to copper smelter Calcine Silver-lead residue Leach Solvent extraction Pregnant solution Indium metal Solid residues Purification Iron as jarosite To tailings Solution Cadmium metal

Electrolysis

Zinc sheets

Melting, casting Zinc metal

Figure 9.3 Generalised beneficiation flow diagram, based on the Kidd Creek operations. (Modified after http:// www.mining-technology.com/projects/kidd_creek/kidd_creek7.html.) sulfidic zinc concentrate which produces a zinc and Moskalyk, 2003). The process used coke calcine. The oxidation process is followed by a reduction of the lead bullion and the addition of two-stage dissolution process. The first stage lead and zinc chlorides to recover indium chlo- selectively removes iron as jarosite from solution. ride (InCl3) in the slag. The solution was purified The second stage introduces metal-rich Cottrell by removal of tin and lead, and indium was recov- ash from the copper refinery to the solution and ered as a sponge by adding powdered zinc. silver and lead are enriched in a residue. The Indium is generally concentrated and recov- indium-bearing solution is then treated by solvent ered at a purity of greater than 99%. This extraction and complex precipitation–dissolution low-grade indium standard is further refined to processes result in the recovery of pure indium. 4 N (99.99%) and higher purity levels. Zone At the Umicore facilities in Hoboken, Belgium, smelting and electrolytic refining are used for the ‘Harris Process’ is used to recover indium obtaining ultrapure indium and for the produc- during the refining of lead. The process involves tion of single crystals. slag composition manipulation by selective oxidation using alkali-metal hydroxides and Indium production from copper ores nitrites, and removal of impure compounds and elements found in lead bullion during pyrometal- The production of indium from copper concen- lurgical processing. Indium was formerly pro- trates is poorly developed despite the fact that duced at the La Oroya smelter in Peru from there are relatively high indium concentrations concentrates derived from Centromin SA, Peru’s in the copper ores from many VHMS deposits. leading producer of silver, lead and zinc (Alfantzi This is due chiefly to the very fine grain size of 214 ulrich schwarz-schampra these ores which makes the extraction of indium (Roskill, 2010). Production in 2011 was estimated at uneconomic. Urals Mining and Metals Co (UMMC), 1000 tonnes of tin and 25 tonnes of indium. Russia’s second largest copper producer, was reported to be starting indium recovery from the Indium recovery from secondary sources residues of the copper production circuit in 2006 (USGS, 2010). The project aimed at the produc- The Xstrata Zink GmbH in Nordenham, Germany, tion of five tonnes of indium per year but no uses the electrolytic zinc refinery process for further information on the process is available. imported products and recycled materials. The trace metals and impurities are separated from the zinc sulfate solution prior to electrolysis. Any Indium production from tin ores trace metals are precipitated with the iron sulfate. Indium is recovered from some tin concentrates They are then either sold to copper and lead which contain about 100 ppm In. Bolivian tin smelters with the minor metal concentrate (copper ores are reported to contain elevated indium con- concentrate, lead/silver concentrate) or discarded centrations, although in 2010 none of the indium as waste. The level of production of trace metals was being recovered domestically (Anderson, such as indium, germanium and gallium from 2012). The indium-bearing concentrates were minor metal concentrates in copper and lead exported and processed in other countries, most smelters is not publically reported. The German likely in Japan and China. Aurubis, Europe’s largest copper smelter, co- At Capper Pass in the United Kingdom, which produces precious and trace metals such as gold, closed in 1991, the process of tin recovery involved silver, tellurium and selenium. However, indium the chlorination of electrorefined tin to produce is not produced due to the very low concentrations a tin chloride slag containing less than 3% In present in the primary copper concentrates. (Alfantazi and Moskalyk, 2003). The majority of the tin was precipitated in two neutralisation stages in series, including recycling from the second stage. Specifications and uses The recovery of indium involved cementation with zinc to produce a sponge material. The 95% purity Indium is traded as indium metal in a wide variety sponge was then electrorefined to produce a 99.5% of forms including pellets, powder, plates, sheets, purity indium cathode that, in turn, was recast as wire and foils. The standard quality is 99.99% the anode to yield 99.97% purity indium metal in a (termed ‘4 N’), but higher purities up to 5 N and second stage of electrolysis. 7 N (99.99999%) are also standard specifications. An alternative vacuum refining technique at Indium is valuable in a diverse range of prod- Novosibirsk Tin (NOKN), Russia`s single tin ucts and applications because of its unique producer, produces a lead–tin alloy with indium properties: concentrations up to 2000 ppm In. The indium-bear- ● low melting point; ing alloy is melted in a 20-tonne vessel and the ● workability at very low temperatures down to indium is recovered selectively from solution prior absolute zero (−273 °C); to further refining. This process is also applicable to ● ability to wet glass; other refinery residues with indium concentrations ● suppleness, softer than lead; of at least 1000 ppm and may also be appropriate for ● tight adhesion to other metals; other by-products such as lead, bismuth and gallium. ● dissolution in acids; The Novosibirsk Tin Plant is one of the largest pro- ● amphoteric character; ducers of pure tin in Europe. The plant is treating tin ● no reaction with water, boron, silicon or carbon; concentrates from the Perevalnoye, Molodezhnoye ● reaction with oxygen only at higher und Pravourminskoye (Khabarovsk), Khinganskoye temperatures; (Autonomous Jewish Region), Deputatskolovo ● oxidation by halogens or oxalic acid, to give (Yakutien) und Syryzhdatskoye (Kyrgyzia) deposits indium(III) compounds. Indium 215

The first large-scale application for indium was 8% as a coating for bearings in high-performance 3% aircraft engines during World War II. Subsequently production gradually increased as new uses were 5% 56% found in fusible alloys, solder and electronics. In the 1950s, tiny indium beads were used for the emitters and collectors of PNP alloy junction tran- 8% sistors. In the middle and late 1980s, the development of indium phosphide semiconductors and indium–tin oxide (ITO) thin films for liquid 6% crystal displays (LCD) aroused much interest. Indium is used for numerous purposes in a variety of applications. However, the end uses of 4% pure indium metal are limited, with the majority used in some form of chemical compound. The major uses are in high-purity compounds of low- 10% temperature alloys, (soft) solder and thin films (Figure 9.4). Other uses are in the manufacture of batteries, electrical components, semiconductors and in research. Flat panel displays Solders The size of the indium market has increased Alloy/compounds Thermal interface materials rapidly in the past 20 years with most growth Photovoltaics occurring in thin-film coatings such that this Batteries Compound application accounted for 56 per cent of total con- semiconductors Miscellaneous sumption in 2010 (Mikolajczak and Harrower, and LEDs 2012). Exceptional growth in demand has come from Japan, which accounted for about 60 per Figure 9.4 The main end-uses of indium in 2010. cent of indium consumption in 2011, followed by (Modified after Mikolajczak and Harrower, 2012, the Republic of Korea, China and Taiwan. More courtesy of Indium Corporation.) than two-thirds of the indium demand from these four largest consumers is used in applications for cent of the visible light passing through. This indium–tin oxide (ITO) and semiconductors. unusual combination of properties makes ITO useful in a number of applications. Transparent Indium–tin oxide (ITO) heat-reflecting ITO films are used in glass coat- ings, solar collectors and streetlights. ITO films Indium oxide is a transparent conducting oxide are also used for electrophoretic displays, electro- which combines the three properties transpar- luminescent displays, plasma display panels, ency, heat reflection and electrical conduction. electrochromic displays, field emission displays Transparent materials which are also electrically and as coatings on architectural glass, solar col- conducting are quite rare because the mecha- lectors, windscreen glass and cathode ray tubes. nisms of conduction and the absorption of light ITO is also used to improve the performance of energy are both determined by the free electron low-pressure sodium vapour lamps. density of a material. Doping indium oxide with approximately 10 per cent tin oxide to form Alloys and solders indium–tin oxide increases both the electrical conductivity and heat reflectivity without signif- Indium forms low melting point alloys with icantly affecting transparency. A film of ITO various metals including bismuth, tin, lead and which is 5 μm thick will absorb less than 20 per cadmium. These alloys are used as solders 216 ulrich schwarz-schampra and fusible alloys in the electronics industry (LEDs) and laser diodes, photodetectors, fibre- and in a variety of specialised applications, e.g. optic telecommunications and optoelectronic metal forming, optical grinding, fusible links integrated circuits. Indium antimonide, arsenide and liquid metal switching devices. Lead-free or phosphide compounds are used as a substrate. solders consisting of tin, indium and silver are a Several indium-bearing compounds are used as significant growth market for indium. These the epitaxial layer, such as indium gallium solders benefit from limited crack propagation arsenide. Indium-based LEDs are used to and improved resistance to thermal fatigue. optically transmit data in displays. Indium- They also inhibit the mobilisation of gold com- based laser diodes are used in fibre-optic com- ponents in electronic devices. Alloys of gold and munications. Ultrapure metalorganics of indium palladium used in dentistry often include include high-purity trimethylindium (TMI), indium to improve their casting and machining which is used as the semiconductor dopant in properties. III-VI compound semiconductors. In the optical industry, low-melting-point alloys are applied to lenses and act as a surface for Others machine tools to grip during the polishing pro- Indium metal and indium alloys are used for a cess. The indium property of tight adhesion is number of surface-coating applications. The most also applied in alloys which are used as bonding important of these is plating automobile and agents between non-metallic materials, such as aircraft engine bearings. Other uses include glass, glazed ceramics and quartz. In the nuclear decorative plating, corrosion inhibitors and plat- industry an alloy of silver (80% Ag), indium (15% ing on aluminium. Indium plates readily from In) and cadmium (5% Cd) is used in nuclear con- either acid or alkaline solutions onto a wide trol rods to control nuclear chain reactions by variety of metallic substrates. The most com- absorbing free thermal neutrons. monly used plating solutions are indium sulfa- Global consumption of indium in low melting mate, indium cyanide, indium fluoroborate point alloys and solders was estimated to be and indium sulfate. Indium sulfamate is an about 200 tonnes in 2011, equivalent to approxi- easily controlled, low-maintenance plating agent. mately 14 per cent of total indium consumption Indium cyanide gives even bright deposits and (Mikolajczak and Harrower, 2012). This is a con- exhibits excellent throwing power (measure of siderable increase from the consumption levels of the efficiency of electrodeposition). Indium fluo- the early 1980s when less than 10 tonnes per year roborate platings are usually used for applying was used for these purposes. heavy (25–75 μm), dense deposits of indium on items such as seal rings. Semiconductors Indium is also used for electrodeless lamps, Demand for very high purity indium is increasing mercury alloy replacements and in alkaline bat- rapidly in the photovoltaic industry and in the teries in which indium prevents the zinc from production of I-III-VI compound semiconductors corroding and the build-up of hydrogen gas within

(such as CuInSe2, CuInGaSe2 (CIGS), CuInS2 sealed casings. Phosphors, like indium–boron

(CIS), CuInGaS2) for solar cells. Recent research oxide (InBO3) and manganese–indium–based indicates that there is potential for enhanced products, are used for monochrome cathode ray efficiency at decreasing production costs for thin tubes. Small amounts of indium are also used in films based on indium for the mass production of aluminium alloy sacrificial anodes (for salt-water solar cells (ZSW, 2012). applications) to prevent passivation of the alu- Indium is also used in III-V compound semi- minium. Indium wire is applied as a vacuum seal conductors (InP, InSb, InAs, InN, InxGa1–xN) and and a thermal conductor in cryogenics and ultra- III-VI compound semiconductors (InS, InSe, high-vacuum applications. Indium is an ingre- InTe) for the production of light-emitting diodes dient in the gallium–indium–tin alloy Galinstan, Indium 217 which is liquid at room temperature but is not a resource of 95,000 tonnes and a reserve of toxic like mercury. 12,500 tonnes of indium metal. Additional con- Indium is also used as a thermal interface tributions to global indium resources and reserves material in personal computers. Pre-shaped may also be estimated from global copper foil sheets are fitted between the heat-transfer resources. Table 9.5 provides estimates of global surface of a microprocessor and its heat sink. indium resources and reserves derived from the The application of heat partially melts the foil most recent USGS estimates for zinc and copper and allows the indium metal to fill in any (USGS, 2012a and 2012b). microscopic gaps and pits between the two sur- It is important to note that such estimates of faces, removing any insulating air pockets that indium resources and reserves are not well con- would otherwise compromise heat-transfer strained because they treat all zinc and copper efficiency. ores as if they had the same indium contents. In Indium leukocyte scintigraphy, using the fact, the figures derived in this manner signifi- radioactive isotope 111In, has a variety of medical cantly underestimate global indium resources. If applications (van Nostrand et al., 1988). These the average size and zinc grades of reserves in include early-phase drug development and the massive sulfide (VHMS and SHMS) deposits are monitoring of activity of white blood cells in examined it can be shown that known large areas of infection. VHMS deposits worldwide contain 13,750 tonnes of indium metal, based on an average indium content of 50 g/t. If the indium content is 20 g/t the indium reserves in these deposits is 5500 Resources and reserves tonnes (Schwarz-Schampera and Herzig, 2002 and Alfantazi and Moskalyk, 2003). Similarly, it Reliable estimates of global indium resources are can be calculated that in the largest SHMS not available. However, given that more than 95 deposits there is an additional 11,750 tonnes of per cent of indium is recovered as a by-product from zinc refineries, estimates of primary indium resources are generally derived from zinc resources worldwide. A minor proportion of global indium production is derived from the treatment of sul- Table 9.5 Estimated global resources and reserves of indium (tonnes) calculated from global zinc and copper fur-rich copper and tin ores, with each accounting resources and reserves reported by USGS (2012a and for less than five per cent of global supply. The 2012b). close association of indium with copper-rich ore-forming systems and its co-enrichment with tonnes zinc and copper, however, make copper ores Resources in zinc ores1 95,000 from massive sulfide deposits a potentially Reserves in zinc ores2 12,500 important indium resource. Slimes and other Resources in copper ores3 30,000 residues (e.g. anode slime, Cottrell ash) from the Reserves in copper ores4 6300 copper production process in indium-bearing ore Total indium resources 125,000 deposits have been introduced into indium refin- Total indium reserves 18,800 eries and contribute to enhanced production of 1 indium, e.g. at the Kidd Creek smelter until its calculated from global zinc resources with indium concentration of 50 g/t closure in 2010. 2calculated from global zinc reserves with indium concentration of Identified global zinc resources are about 1.9 50 g/t billion metric tonnes, with zinc reserves esti- 3calculated from global copper resources with indium mated at 250 million tonnes (USGS, 2012a). concentration of 10 g/t Using a conservative estimate of 50 grams of 4calculated from global copper reserves with indium concentration indium per tonne of zinc ore, this is equivalent to of 10 g/t 218 ulrich schwarz-schampra indium if the zinc ore contains 50 g/t In, or 4700 Production tonnes at 20 g/t In. Among the large-tonnage massive sulfide Global indium production was only about 40–50 deposits (i.e. those with more than 25 million tonnes in the early 1970s, but rapidly growing tonnes of base-metal massive sulfide ore), six demand over the following three decades led to deposits in China account for some 317 million major increases in primary indium production. tonnes of ore with an average of about 3.2% zinc World production of primary indium increased by (VHMS) and about 188 million tonnes averaging five per cent from 609 tonnes in 2010 to an esti- about 5.6% zinc (SHMS), respectively. Given an mated 640 tonnes in 2011, while between 2000 and estimated average indium grade of 50 grams per 2011 mine production almost doubled (USGS, tonne this equates to 1100 tonnes of indium 2012c). The contributions to primary mine produc- reserves in large-tonnage Chinese base metal sul- tion of indium from individual countries is not fide deposits (VHMS, SHMS). The small size of available, although more than half of global indium these reserves is inconsistent with China being production comes from China, where indium is the largest world producer of indium, having con- produced mainly from tin–base-metal sulfide tinuously increased its production since 1998. In deposits (e.g. Dachang district), massive sulfide addition to supplies from domestic mines, other deposits (Yunnan) and from secondary sources (res- sources, such as stocks and residues from over- idues; Schwarz-Schampera and Herzig 2002). seas, have to be inferred. The mining of indium alone as a primary com- A study by the Indium Corporation of America modity is uneconomic, and indium is a typical (Mikolajczak, 2009) estimated primary indium by-product of mining a major industrial metal, in reserves and resources of 26,000 tonnes of indium this case zinc (Schwarz-Schampera and Herzig, metal in western base-metal sulfide mines. China, 2002 and Niederschlag and Stelter, 2009). More Russia and other CIS member states account for than 95 per cent of global indium production another 23,000 tonnes of indium metal. An addi- comes from electrolytic zinc refining. In 2011 tional 15,000 tonnes of indium may come from global zinc mine production grew by four per cent tailings and residues (Mikolajczak, 2009). to 12.4 million tonnes, with China, Peru and The most recent USGS assessment of global Australia the largest producers (USGS, 2012a). indium reserves in 2008 was 11,000 tonnes The largest producers of refined zinc were (USGS, 2008). This was a major increase over the China, Canada and India. According to the USGS 2006 figure when reserves were estimated at 6000 (2012a) China is the largest refinery producer of tonnes of indium metal. China accounted for indium (340 tonnes, 53.1 per cent of total), fol- more than 75 per cent of global indium reserves lowed by the Republic of Korea (100 tonnes), in 2008. As a result of ongoing exploration pro- Japan (70 tonnes), Canada (65 tonnes) and Belgium grammes, Chinese reserves increased greatly in (30 tonnes). Other producers accounted for an only one year, from 280 tonnes in 2006 to 8000 additional 35 tonnes. It is evident that the tonnes of indium in 2007. Peru, Canada, USA and majority of refinery production comes from zinc Russia together had more than eight per cent of smelters using electrolytic refining and the global indium reserves (USGS, 2008). Global treatment of zinc-refinery slimes. However, not indium resources in 2008 were 49,000 tonnes, all zinc smelters process their own residues to less than half the amount estimated in 2012 recover indium and other trace commodities. In (Table 9.5). these cases the metal-rich residues are generally It is evident that a lack of reliable data has led sold to secondary producers which possess the to wide variation in global estimates of indium appropriate facilities. It should be noted that resources and reserves. However, it is clear that many of the large primary indium producer coun- these estimates have increased in size over time, tries do not have significant indium-bearing ore in tandem with growing worldwide demand. deposits. Indium 219

700

600

500

400

300

200

100

Production of primary indium (tonnes) 0 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

Figure 9.5 Global primary indium production, 1994–2011. (Data from USGS, 2012c.)

Table 9.6 Estimated indium mine production in 2009 from base metal zinc deposits. (Data from Roskill, 2010.)

Mine production Estimated indium concentration Zinc mine production zinc concentrate and quantity in the concentrate Share

(million tonnes (million tonnes Country contained zinc) sphalerite) (ppm) (tonnes) %

Australia 1.3 1.9 15 29 6.2 Canada 0.7 1.1 37 40 8.6 China 2.8 4.2 50 210 45.0 Mexico 0.5 0.8 20 16 3.4 Peru 1.5 2.2 20 44 9.4 USA 0.7 1 20 21 4.5 Other countries 3.6 5.4 20 107 23.0 Total 11.1 16.6 29 467 100

Roskill (2010) concluded that about 470 tonnes During 2010 more than 90 producers worldwide of primary indium were recovered from zinc min- delivered pure indium metal and indium com- ing in 2009. It was estimated that worldwide zinc pounds (Roskill, 2010). The majority of these are ore (sphalerite) concentrates contain an average independent smelters and recycling companies of 15 to 50 g/t In. The Indium Corporation esti- which possess the technical knowledge and equip- mated global production of primary indium in ment for trace-metal recovery. There are only a 2010 to be about 500 tonnes (Indium Corp., 2010). few vertically integrated producers which have The levels of global indium production from their own mining and smelting operations for primary sources between 1994 and 2011 are indium recovery. These include Xstrata and Teck shown in Figure 9.5, while the estimated indium Resources producing in Canada, and state-owned refinery production in 2009 from zinc-rich base- consortia of mining companies and smelters metal deposits is shown in Table 9.6. in China. The number of indium producers and 220 ulrich schwarz-schampra

Table 9.7 Important global indium producers. (Data from Roskill, 2010 and USGS, 2009.)

Country Refinery capacity (tonnes) Secondary capacity (tonnes)

Nanjaing Germanium Factory China 150 Huludao Zinc China 50 Zhuzhou Smelter Group China ? Dowa Metals & Mining Co Japan 70 150 Asahi Pretec Corp Japan 200 Mitsubishi Mat. Group Japan 96 Korea Zinc Rep. of Korea 100 Umicore SA Belgium 30 Teck Resources Ltd Canada ~75 Xstrata plc Canada ?

specialised indium refiners (producing indium of capacity has been installed in South Korea, purity greater than 4 N) has increased sharply since Japan, Canada, Bolivia, Brazil and Peru. 2000, especially in industrialised countries like Numerous Chinese companies improved their Japan, USA, Republic of Korea and Germany. In treatment lines to recover indium from residues China too, the number of primary indium pro- containing less than 0.5% In, while other Chinese ducers has grown rapidly with a total of 21 producers focused on high-purity indium metal. operating in 2010. In 2006 the Canadian company Teck Resources The Chinese raw material industry has the Ltd completed the expansion of its indium pro- technical know-how and operating capacity for duction capacity to more than 75 tonnes per the recovery of various rare metals from differ- year at its Trail smelter in British Columbia ent ore deposit types. Accordingly China cur- (Teck Cominco Ltd., 2006). Teck accounted rently dominates the international markets for for all of Canada’s indium production in 2011 most of these metals, including indium. In con- (65 tonnes). trast, a number of refineries outside China have Kazzinc, the leading zinc and lead producer in abandoned trace-metal production or now focus Kazakhstan, also produces indium as a by-product on the production of higher indium purities. (Kazzinc, 2012). The major zinc mines in Globally important companies with a significant Kazakhstan include Maleevsky, Shubinsky, share of indium refinery production are given in Tishinsky and Shaimerden. The recovery and Table 9.7. refining of trace metals is well developed at the The annual refinery capacities for indium are Ust-Kamenogorsk smelter and Kazzinc also pro- estimated at 727 tonnes (Roskill, 2010). The duces selenium, tellurium, thallium, mercury shares are: China, 360 tonnes; Republic of Korea, and antimony. 120 tonnes; with 50 tonnes each in Japan, Canada and Belgium and 40 tonnes in the USA. Russia, Production from residues and scrap Ukraine, Kazakhstan, Germany and Peru have smaller installed capacities. The worldwide The recovery of indium from secondary sources refinery capacity is estimated to be 17–25 per like residues (new scrap) and end-of-life scrap (old cent higher than the annual production of 540 to scrap) exceeds the production from primary 600 tonnes indium (Roskill, 2010; Mikolajczak sources by at least 250 per cent (Roskill, 2010). and Harrower, 2012). The recycling of indium compounds and the A number of indium refineries have been recovery of secondary indium metal are particu- expanded in recent years. For example, new larly important in some producing countries such Indium 221 as Japan, China, Republic of Korea, UK, Canada, The Australian company Crusader Holdings Philippines, Taiwan and Germany. In Japan the acquired exploration licences in the prospective annual secondary indium production from treat- Goiás tin district in Brazil, which includes the ing ITO production waste is now more than 450 indium-rich Mangabeira deposit (Moura et al., tonnes (Roskill, 2010). In Germany companies 2007). like PPM Pure Metals and Haines and Maassen Exploration and resource assessment are specialise in the recycling of indium-bearing continuing at the polymetallic Mount Pleasant compounds. PPM produces and refines high- deposit, New Brunswick. The companies Geodex purity indium (up to 7 N). Minerals and Adex Mining have confirmed significant potential for a wide variety of metals, including indium, and plan to start mining at Projects under development Mount Pleasant in 2015 (Adex Mining Corp., The growing markets for LCD and photovoltaics 2012). The Canadian company Lithic Resources have led to an increase in exploration activities is exploring for zinc, copper, silver, indium and for indium-bearing ore deposits. Current projects gold at its Crypto project in Utah. Total indicated are focusing on magmatic base-metal-rich deposits, and inferred indium resources are 283 tonnes and hydrothermal base-metal-vein deposits, and, 516 tonnes, respectively (Lithic Resources, 2009). most significantly, on massive base-metal sulfide The South American Silver Corp. was developing deposits. Promising results with high trace metal the Malku Khota silver–indium–gallium project and indium concentrations have been reported in Bolivia until August 2012 when it was nation- from the well-known Mount Pleasant deposit in alised by the Bolivian government. The deposit New Brunswick, Canada and from a number of contains an indicated resource of 230.3 million deposits in Argentina, Australia, Brazil, Bolivia, ounces of silver and 1481 tonnes of indium, and Finland, Portugal and the USA. an inferred resource of 140 million ounces silver Argentex Mining is working on the definition and 935 tonnes indium (South American Silver of the mineral resources at Pingüino, Santa Cruz, Corp., 2011). Patagonia. Exploration revealed numerous base- Several exploration targets in south-east metal veins with local indium concentrations in Finland have been studied with respect to indium excess of 1000 g/t. Resource estimates total 7.32 concentrations (Cook et al., 2011). Ores occur million tonnes of ore grading 16.26 g/t indium. as massive indium-bearing magnetite–sphalerite Inferred resources add another 35.4 million veins (Getmossmalmen), as greisen-style veins tonnes of ore with an average of 8.89 g/t indium (Jungfrubergen) and as polymetallic quartz veins (Argentex Mining Corporation, 2009). The (Korsvik-1 and −2, Sarvlaxviken area). The occur- Canadian company Marifil Mines Ltd. is exploring rences have bulk indium concentrations between the lead–zinc–silver–gold–indium San Roque 28 and 1160 ppm; roquesite is associated with project in Argentina`s Rio Negro province. The copper-rich quartz veins. Indium is a significant drill programme indicated 143 metres of ore con- trace metal in the ores of Lagoa Salgada, a 4 mil- taining 0.74 g/t gold, 27.2 g/t silver, 0.37% lead lion tonne unexploited massive sulfide deposit in and 0.55% zinc and including 11.7 metres of the Iberian Pyrite Belt, Portugal (de Oliveira 174 g/t indium (Marifil Mines, 2007). In 2011, et al., 2011). Indium concentrations reach up to Kagara Mining announced mining at the aban- 93 ppm in ores and are hosted by sphalerite, with doned Baal Gammon mine in Queensland, a maximum of 0.8 wt% In. Australia for copper, with potential by-products tin, silver and indium. The Baal Gammon deposit Abandoned production contains an indicated and inferred resource of 5.5 million tonnes grading 0.8% copper, 29 g/t silver, In 2010 the Kidd Creek smelter in Timmins, 29 g/t indium and 0.2% tin (Kagara Mining, 2012). Ontario, which had an annual production 222 ulrich schwarz-schampra capacity of 50 tonnes of indium metal, closed World trade permanently. The smelter treated the copper and zinc ores from the nearby Kidd Creek VHMS Most indium is traded in the form of semi-refined deposit. The indium production came from the unwrought indium with 99.97% or 99.95% purity zinc concentrates, from residues from the copper produced by electrolytic zinc refineries. Some production (Cottrell ash) and from imported con- zinc refineries ship the anode slime to other refin- centrates. The smelter produced indium ingots eries which are equipped to extract and purify with a 99.97% purity. The production of indium indium. Indium scrap and powders are also at Doe Run Peru’s La Oroya metallurgical facility traded. High-purity metal (5 N–7 N) is produced was suspended in June 2009 due to high sulfuric in only few countries including Japan, USA, UK acid emissions and governmental restrictions. and Germany. The trade of unwrought indium is Some metal production at La Oroya restarted closely associated with zinc mining and refining in 2012. capabilities, and is dominated by Canada, China, The abandoned lead–zinc–indium–silver Bolivia, Russia, Kazakhstan, Ukraine and Brazil. Toyoha deposit, Japan, contained an average of It is difficult to obtain accurate data specific to 150 to 250 ppm In and produced about 500,000 indium trade because the trade codes record tonnes of ore concentrates each year (Ohta, indium trade combined with other metals, such 2005). Annual production was about 50 to 55 as gallium, germanium, hafnium, niobium, thal- tonnes of indium metal (USGS, 2002) and lium and vanadium. accounted for one third of the entire global pro- Since 2005 China has been the principal global duction. Production from the silver-lead-zinc epi- source of indium. China’s indium exports to the thermal deposit, which is underlain by an active international markets between 2005 and 2011 are magmatic system, was finally suspended in 2006 estimated at 1854 tonnes (Roskill, 2010; USGS, when ore reserves were exhausted (Mining 2012c). This quantity represents more than 40 Journal, 2007). A number of other ore deposits in per cent of the estimated total indium which Hokkaido and Honshu, which are similar in style entered the trade market. However, since 2007 and mineralisation, are not economic. A recent export restrictions, designed to improve the envi- resource evaluation of indium in Japanese ore ronmental performance of the indium sector, deposits identified the Toyoha mine as the largest have reduced the quantity of indium available indium deposit in Japan with a total of 4700 from China. From about 700 tonnes in 2005 it fell tonnes of indium, followed by the Ashio mine to only 88 tonnes in 2009, although the figure for with an estimated total of 1200 tonnes of indium 2010 and 2011 was 340 tonnes (USGS, 2012c). (Ishihara et al., 2006). The number of companies that received export The largest European copper mine, Neves- licences dropped to 18 in 2011 (USGS, 2012c). Corvo in Portugal, has a tonnage of 270 million The other principal indium exporters include tonnes of massive base-metal sulfides (Galley Canada, USA, Belgium, Brazil, Japan, Republic of et al., 2007). This hybrid VHMS/SHMS deposit Korea and Russia. Belgium, France and the has very rich indium reserves which are associ- Republic of Korea increased their exports in 2008 ated with copper-, zinc- and tin-rich sulfide ores. and 2009 (Roskill, 2010). Mining of the deposit commenced in 1989 when Japan, USA and UK have been the major Somincor SA began to work copper and tin ores. importers of indium materials since 2005. In In 2006 the present owner, the Lundin Mining 2011 the USA`s imports reached 150 tonnes, an Corp., started zinc production but this was increase of 43 per cent over 2009 (USGS, 2012c). suspended in 2008 due to the low zinc price. The main suppliers to USA were China (31 per However, zinc production resumed in 2010 and a cent, 46.5 tonnes), Canada (25 per cent, 37.5 new zinc plant was commissioned in 2011 tonnes), Japan (16 per cent) and Belgium (9 per (Lundin Mining, 2012). cent). Most of the imports were refined into 5 N Indium 223 purity qualities and to indium compounds and an annual average of 57 US$/kg in 1973 to 547 alloys. Japan imports large quantities of indium US$/kg in 1980. However, the installation of new from China, Canada, Republic of Korea and extraction capacity, combined with decreasing Taiwan. In 2011 Japanese indium imports reached global demand after the Three Mile Island a record high, increasing by 18 per cent to 494 accident and economic recession in the early tonnes from 418 tonnes in 2010. Indium is mainly 1980s, resulted in indium oversupply and imported into Japan in the form of residues, decreasing prices. The average annual price in wastes, scrap and semi-refined metal. 1986 was only 84 US$/kg. Renewed interest in indium uses, largely from the Japanese elec- tronics industry, started in 1988 when the annual Prices average indium price reached 319 US$/kg. Since the mid-1980s, the development of indium phos- Smelters usually produce indium of 99.95% phide semiconductors and thin-film ITO com- purity which is further refined in the smelter or pounds for LCDs have dominated the indium by secondary processors and tertiary manufac- markets. Low production costs in 1993 and 1994 turers to higher purities. Indium metal (99.99%) resulted in a supply shortfall and subsequently is traded in US$ per kilogram. Unlike the major led to a strong price increase in 1995 (annual traded metals indium is sold by private contracts, average 388 US$/kg). Variable demand for LCDs, often long term, between buyer and seller. The technological advances which reduced the estimated value of primary indium metal con- amount of indium in LCD screens and the avail- sumed in 2011, based upon the annual average ability of supply from stockpiles in the Ukraine New York dealer price, was about US$ 82 million led to a price decline in the following years (USGS, 2012c). (Figure 9.6), with the annual average indium price The costs of indium exploration, mining and falling to 97 US$/kg in 2002. Subsequently, a metallurgy are closely linked to those for the number of factors led to a marked rise in the main commodities zinc, copper and tin, as well as indium price, which reached an annual average additional trace metals (e.g. cadmium). The all-time high of more than 946 US$/kg in 2005. majority of the costs are included within the This increase was a response to improving costs of the electrolytic zinc refining and indium global economic conditions, especially in emerg- production costs cannot readily be calculated ing economies, strong demand for flat screens, separately. However, the recovery of indium from technical innovations in thin-film photovoltaic zinc-refinery residues requires significant addi- cells, the closure of established indium producers, tional investments which include technological problems in global zinc production and specula- facilities, chemicals, electricity and environ- tion of a worldwide shortage of indium supply. mental compliance. The direct production costs As a consequence, production capacities were for indium in these facilities are in the range of 50 expanded and recycling began to make a to 100 US dollars per kilogram. significant contribution to global supply for the Since 1970 indium prices have followed an first time. Subsequently, increased production almost cyclical trend. The years 1975, 1980, 1988, combined with weakening demand from Asia 1995 and 2003 were characterised by rapid (Republic of Korea, Japan) and economic reces- significant increases in global indium prices, fol- sion slowed down the markets and the indium lowed by a slow gradual decrease (Figure 9.6). price decreased gradually. In 2009, the average Between 1973 and 1980 demand increased for annual price was 500 US$/kg indium. From mid- indium use in nuclear control rods. At the same 2009 increasing ITO demand led to a general time, supply from easily accessible indium increase in prices. In 2011 the indium price rose deposits, such as those in Bolivia, gradually approximately 43 per cent over 2010, to an annual ceased. The indium price increased sharply from average of 693 US$/kg. 224 ulrich schwarz-schampra

1200

1000

800

600

400

200

0 Figure 9.6 Annual average indium Annual average indium price (US$/kg) prices between 1936 and 2011. (Data

1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1999 1992 1996 2000 2004 2008 2011 from BGR database.)

Recycling and substitution Yamasue et al., 2009). New technologies and increased capacities, together with improved col- The recovery of indium from production waste lection of electronic scrap, may further reduce (‘new scrap’) and, to a lesser extent, from the the recovery costs. recycling of indium-bearing electronic scrap have Price volatility and supply concerns have become increasingly important in the last decade triggered research into the development of ITO as the price of indium has risen. About two thirds substitutes (USGS, 2012c). However, given that (1000 tonnes) of global indium supply comes the cost of ITO is a very small proportion of the from recycling. total cost of a LCD panel and that there has Most ITO producing countries have increased been considerable recent investment in ITO their recycling rates significantly in recent years. sputtering technology and capacity, it is likely The recovery of indium from the indium–tin that ITO will remain the predominant trans- oxide (ITO) production and the sputter process parent conductive coating for most LCD appli- has grown in parallel with the growth in produc- cations for the foreseeable future. Antimony tin tion of flat screens and probably accounts for oxide coatings, which are deposited by an ink- more than 50 per cent of the worldwide indium jetting process, are one possible alternative, supply. The sputter process deposits ITO parti- while another is zinc oxide with aluminium cles as a thin-film coating onto a glass substrate doping (AZO) which has good transparency and in a vacuum chamber. This process is highly electrical properties. Carbon nanotube coatings, inefficient and more than 70 per cent of the ITO applied by wet-processing techniques, have target material is not deposited on the glass, with been developed as an alternative to ITO coat- some of it remaining on the walls of the chamber. ings in flexible displays, solar cells and touch These large quantities of excess material, screens (Glatkowski, 2005). Poly(3,4-ethylene together with the grinding sludge, make its dioxythiophene; PEDOT) has been studied as a recovery economic. Indium recycling from substitute for ITO in flexible displays and electronic scrap is technically difficult because organic light-emitting diodes (Galagan et al., of its low indium content and only becomes 2011). Graphene quantum dots may replace ITO economic at prices above US$ 400 per kilogram. electrodes in solar cells and have also been A process to reclaim indium directly from scrap explored as a replacement for ITO in LCDs (Wu LCD panels has been developed (Sharp process; et al., 2008). Gallium arsenide can substitute for Indium 225 indium phosphide in solar cells and in many stopes thereby virtually eliminating surface semiconductor applications. contamination. Given the rapid increase in the use of indium in recent years and the anticipated future growth Environmental aspects in consumption in electronic, photovoltaic and LED applications it is important to understand The environmental impacts of the mining and the natural and industrial cycling of indium and processing of sulfide-rich indium-bearing ores its toxicology. This will provide a basis for are dependent on a wide range of factors. Most improving knowledge of its environmental VHMS deposits are small in size and have limited behaviour and human health effects and thus effects on the regional hydrological regime. contribute to its safe and sustainable use. Indium However, soluble sulfate salt minerals derived has no known physiological function in humans from the weathering and oxidation of sulfide or animals and is not known to be used by any minerals in mine dumps and tailings piles are a organism. Although the toxicity of indium metal potential source of metal contamination and is low and harmful effects are not reported, there generation of acidic drainage. The significance of is evidence suggesting that some indium com- this process and the related impacts depend on pounds are toxic. There is no drinking water stan- many factors including the size of the miner- dard for indium in the United States but alised surface outcrop, the nature and extent of workplace exposure limits have been set for the mining operations and the disposal of waste indium and its compounds in some jurisdictions products. The type of host rock, sedimentary or (0.1 mg/m3 time-weighted average; NIOSH, 2011; volcanic, does not seem to affect these charac- CAREX, 2012). There is little data available on teristics. The most acidic and metal-rich mine environmental exposure to indium or its com- water, draining from the base of well-consolidated pounds. Very low concentrations of indium have tailings piles from VHMS deposits, has a pH of been reported in air, seawater, rain water and 2.6 to 2.7 and contains high concentrations of a food, although higher levels have been detected range of metals, including some with well-known in seafood from contaminated waters near metal toxic effects, up to 21,000 μg/l Fe, 3600 μg/l Cu, smelters (IARC, 2006; White and Hemond, 2012 ). 220 μg/l Pb, 3300 μg/l Zn, 30 μg/l Co, 10 μg/l Cd, The average daily intake of indium is estimated and 311 mg/l sulfate (Taylor et al., 1995 and refer- to be low, in the range 8–10 μg per day (Smith ences therein). et al., 1978). Potential environmental concerns associated Indium phosphide, used to make semiconduc- with the processing of base-metal ores are related tors for laser diodes, solar cells, LEDs and chiefly to tailings ponds which may contain high integrated circuits, was the most commonly used abundances of lead, zinc, cadmium, bismuth, indium compound until the last decade. Laboratory antimony and other reactants used in flotation studies of rats and mice have provided clear evi- and recovery circuits. Fine-grained and inter- dence that indium phosphide is carcinogenic in grown sulfide ores may require very fine grinding animals, although it has not been adequately for effective beneficiation and can thus lead to studied in humans (National Toxicology Program, the generation of particularly reactive fine- 2001). On this basis indium phosphide was classi- grained tailings. Today, tailings ponds are rou- fied as ‘probably carcinogenic’ to humans (Group tinely lined with impermeable barriers, but in the 2A) by the International Agency for Research on past such measures were not common and there Cancer (IARC, 2006). was potential for significant contamination of Given that indium–tin oxide (ITO) is the surface water and shallow groundwater. Today, predominant form in which indium is currently some underground mines dispose of most of their used, there is considerable interest in establishing tailings by backfilling and cementing mined the risks associated with its use. Potential 226 ulrich schwarz-schampra exposure to particles of ITO is greatest in the natural sources, such as volcanic eruptions and industrial workplace, particularly during the weathering, and it is therefore important that preparation of ITO targets, ITO film deposition these are monitored to better understand indi- and recycling. Most evidence for ITO toxicity has um’s behaviour and transport and its toxic effects, been derived from occupational inhalation especially those due to long-term low-level exposure, but whether or not it is a potential exposure (White and Hemond, 2012). Base-metal health hazard to workers is not fully resolved mines, smelters and refineries treating indium- (Tanaka et al., 2010). A number of clinical and epi- bearing materials are employing environmental demiological studies, mostly from Japan, but also management principles, practices and equipment from China and the United States, have reported which allow them to meet or exceed relevant adverse effects on workers exposed to ITO environmental requirements. in indium processing and recycling facilities (National Toxicology Program, 2009; Omae et al., 2011). These investigations indicate that exposure Outlook to ‘hardly soluble indium compounds’, such as particulate ITO, can potentially cause a several The outlook for the indium market is healthy, types of lung damage. However, measures such as with demand predicted to grow at a rate of 5–10 proper ventilation and the wearing of masks is per cent per annum. Demand for LCD panels may sufficient to prevent health hazards in production. increase more slowly, but it will continue as the A number of studies to investigate the toxicity main use of indium, especially for applications of indium in laboratory animals (mice, rats, such as personal computers, tablets and mobile hamsters, rabbits) have shown that indium com- phones. The ITO sputtering process onto glass pounds exhibit systemic toxicity in mammals has reached maturity and plant has been installed with adverse effects on kidneys, liver, blood, lungs widely and is performing well. New installations and the reproductive and developmental systems will probably use rotary ITO targets (instead of (Hoet at al., 2012 and references therein; White planar) for increased efficiency. Another market and Hemond, 2012; Tanaka et al., 2010; Nakajima area with significant potential for growth is in et al., 2008). thin-film photovoltaic cells where copper– Much remains unknown about indium’s natural indium–gallium–selenide alloy (CIGS) is one of a and anthropogenic cycling, but, given its rapidly number of technologies that might become increasing use, more research is required to investi- widely adopted as photovoltaic systems are gate the extent of its toxic and ecotoxic effects and increasingly used for electricity generation. the factors that control them. The behaviour of The fact that indium is a by-product metal has inhaled indium in the human body and its poten- caused some concern because the availability of tial effects on carcinogenesis and developmental primary supply is dependent on the extraction, and reproductive systems are poorly understood smelting and refining of base metals, chiefly zinc. and should be given particular attention. Although Operators of a zinc smelter are very unlikely to releases of indium from the semiconductor and increase their zinc production simply to deliver electronics industries are currently small, about larger quantities of a by-product. It has been esti- twenty times less than that released by mining and mated that only about 25–30 per cent of the smelting activities (White and Hemond, 2012), indium actually mined each year ends up as there is a need for long-term monitoring of the refined indium. The remainder is either lost in environment and the workplace to understand and residues or goes to refineries that do not have the quantify the risks associated with the increased technology to recover indium. Nevertheless, use of indium. The quantity of indium released worldwide capacity should be sufficient to into the environment from mining, smelting and accommodate the forecast steady growth in industrial activities already exceeds that from indium demand. Additional supply can be Indium 227 expected to come from increased zinc and Cook N.J., Sundblad K., Valkama M., Nygård R., copper mine production, particularly from ores Ciobanu C.L. and Danyushevsky, L. (2011) Indium containing high indium concentrations. Global mineralisation in A-type granites in southeastern mine production of base metals increased by 24 Finland: Insights into mineralogy and partitioning between coexisting minerals. Chemical Geology 284, per cent over the five-year period 2006–2011 62–73. and, given the worldwide revival of exploration Demarthe, J.M., Rousseau, A.M. and Fernandez, F.L. activities for base-metal sulfide deposits in (1990) Recovery of Specialty Metals, Mainly known ore districts and the current interest in Germanium and Indium, from Zinc Primary critical metals, it is likely that new indium Smelting. Lead-Zinc ’90, 151–160. reserves will be identified. Further contribu- Galagan, Y, Rubingh, J.E.J.M., Andriessen, R., Fan, C.C., tions to supply will also come from the installa- Blom, P.W.M., Veenstra, S.C., Kroon, J.M. (2011) ITO- tion of new or expanded indium extraction free flexible organic solar cells with printed current plants at the base-metal smelters, from improved collecting grids. Solar Energy Materials and Solar recovery efficiency at existing indium plants Cells 95, 1339–1343 and from improvements in indium-recycling Galley, A.G., Hannington, M.D. and Jonasson, I.R. (2007) Volcanogenic massive sulphide deposits. In: technology. Goodfellow, W.D. (ed.) Mineral deposits of Canada: The possible influence of geopolitical factors A synthesis of major deposit-types, district metallog- on supply security cannot be discounted. Export eny, the evolution of geological provinces, and explo- restrictions have been imposed by some producer ration methods. Geological Association of Canada, countries. For example, China has imposed Mineral Deposits Division, Special Publication No. export quotas and plans to increase domestic pro- 5, 141–161. duction of LCD screens. Stockpiling by govern- Glatkowski, P.J. (2005) Carbon nanotube based trans- ments and private institutions may also have parent conductive coatings. www.walkingitaly.com/ some impact on global indium supply. Some tuserg/tuserg_metallo/conductive_coatings.pdf governments, notably in Korea, Japan and China, Goodfellow, W.D. and Lydon, J.W. (2007) Sedimen- tary exhalative (SEDEX) deposits. In: Goodfellow, have announced their intentions to stockpile W.D. (ed.) Mineral deposits of Canada: A synthesis of indium although only in Korea is an official major deposit-types, district metallogeny, the evolu- stockpile in place. tion of geological provinces, and exploration methods. Geological Association of Canada, Mineral Deposits Division, Special Publication No. 5: 163–183. References Heald, P., Hayba, D.O. and Foley, N.K. (1987) Comparative anatomy of volcanic-hosted epithermal Adex Mining Corp. (2012) Mount Pleasant: property deposits: acid-sulfate and adularia-sericite types. overview. http://www.adexmining.com/overview.php Economic Geology 82, 1–26. Alfantazi, A.M. and Moskalyk, R.R. (2003) Processing Hedenquist, J.W. (1987) Mineralization associated with of indium: a review. Minerals Engineering 16 (8), volcanic-related hydrothermal systems in the cir- 687–694. cum-Pacific Basin. In: Horn, M.K. (ed.) Transactions Anderson, S.T. (2012) The mineral industry of Bolivia. of the Fourth Circum-Pacific Energy and Mineral http://minerals.usgs.gov/minerals/pubs/country/2010/ Resources Conference, Singapore. American myb3-2010-bl.pdf Association of Petroleum Geologists, 513–524. Argentex Mining Corporation (2009) Resource estimate Hoet, P., De Graef, E., Swennen, B. et al. (2012) for the Pinguino silver, zinc, lead, indiu, gold occur- Occupational exposure to indium; what does bio- rence. (http://www.argentexmining.com/i/pdf/ATX_ monitoring tell us? Toxicology Letters 213, 122–128. Pinguino_Mineral_Resource_Estimate_October_2011. IARC (2006). International Agency for Research on pdf.) Cancer, IARC Monographs on the Evaluation of CAREX (2012) Indium phosphide and other indium Carcinogenic Risks to Humans. Volume 86: Cobalt in compounds. http://www.carexcanada.ca/en/indium_ Hard Metals and Cobalt Sulfate, Gallium Arsenide, phosphide/ Indium Phosphide and Vanadium Pentoxide. 228 ulrich schwarz-schampra

Ishihara S., Hoshino K., Murakami H. and Endo Y. de Oliveira D.P.S., Matos J.X., Rosa C.J.P. et al. (2011) (2006) Resource evaluation and some genetic aspects The Lagoa Salgada Orebody, Iberian Pyrite Belt, of indium in the Japanese ore deposits. Resource Portugal. Economic Geology 106, 1111–1128. Geology 56 (3), 347–364. Ohta, E. (2005) Toyoha Mine, the largest indium producer Jorgensen, J.D. and George, M.W. (2004) Indium. Open- in the world. http://staff.aist.go.jp/ohta-e/toyoha.htm File Report 2004–1300: 1–24 [http://pubs.usgs.gov/ Omae, K., Nakano, M., Tanaka, A., Hirata, M., Hamaguchi, of/2004/1300/2004-1300.pdf]. T. and Chonan, T. (2011) Indium lung – case reports and Kagara Mining (2012) http://www.kagara.com.au/irm/ epidemiology. International Archives of Occupational ShowStaticCategory.aspx?CategoryID=251andHide and Environmental Health 84, 471–477. TopLine=Trueandmasterpage=45. Roskill Information Services Ltd. (2010) Indium: Global Kazzinc (2012) http://kazzinc.vestnik.com.kz/#en/ industry markets and outlook (9th ed.): London, Operations/Products/Indium/. United Kingdom, Roskill Information Services Ltd. Kovalenker, V.A., Laputina, I.P., Znamenskii, V.S. and Schwarz-Schampera, U. and Herzig, P.M. (2002) Indium: Zotov, I.A. (1993) Indium mineralization of the Great Geology, Mineralogy, and Economics. Springer-Verlag, Kuril Island Arc. Geology of Ore Deposits 35, 491– 495. Berlin. Lithic Resources (2009) Crypto project: resource estimate. Schwarz-Schampera, U., Terblanche, H. and Oberthür, http://lithicresources.com/properties/resources_ T.H. (2010) Volcanic-hosted massive sulphide estimate/. deposits in the Murchison greenstone belt, South Lundin Mining (2012) http://www.lundinmining.com/s/ Africa. Mineralium Deposita 45/2, 113–145. Neves-Corvo.asp. Seifert, T.H. and Sandmann, D. (2006) Mineralogy and Marifil Mines (2007)http://www.marifilmines.com/s/ geochemistry of indium-bearing polymetallic vein- news.asp?daterange=2007/01/01…2007/12/31. type deposits: Implications for host minerals from the Mikolajczak, C. (2009) Availability of Indium and Freiberg district, Eastern Erzgebirge, Germany. Ore Gallium: 4 S: www.indium.com/_dynamo/download. Geology Reviews 28 (1), 1–31. php?docid=5529. Sinclair, W.D., Kooiman, G.J.A., Martin, D.A. and Mikolajczak C. and Harrower, M. (2012) Indium Sources Kjarsgaard, I.M. (2006) Geology, geochemistry and and Applications. Minor Metals Conference, February mineralogy of indium resources at Mount Pleasant, 2012. New Brunswick. Ore Geology Reviews 28, 123–145. Mining Journal (2007) Report Japan 2005–6, updated Smith, I.C., Carson, B.L. and Hoffmeister, F. (eds). (1978) October 2007. https://www.mining-journal.com/ Trace Metals in the Environment. Volume 5: Indium. reports/japan—20056-last-updated-october-25 Ann Arbor, Michigan; Ann Arbor Science. Moura, M.A., Botelho, N.F. and Carvalho de Mendonca South American Silver Corp. (2011) Economic assessment F. (2007) The indium-rich sulphides and rare arsenates highlights for Malku Khota silver-indium project. http:// of the Sn-In-mineralized Mangabeira A-type granite, www.soamsilver.com/may-16-2011-news-release.asp central Brazil. Canadian Mineralogist 45, 485–496. Taylor, C.D., Zierenberg, R.A., Goldfarb, R.J., Kilburn, J.E., Nakajima, M, Usami, M, Nakazawa, K, Arishima, K Seal II, R.R. and Kleinkopf, M.D. (1995) Volcanic- and Yamamoto, M. (2008). Developmental toxicity of associated massive sulfide deposits. http://pubs.usgs. indium: Embryotoxicity and teratogenicity in experi- gov/of/1995/ofr-95-0831/CHAP16.pdf mental animals. Congenital Anomalies 48, 145–150. Tanaka, A., Hirata, M., Homma, T. and Kiyohara, Y National Toxicology Program (2001) Abstract for TR-499 – (2010) Chronic pulmonary toxicity study of indium- Indium Phosphide 9CASRN 22398-80-7). Toxicology tin oxide and indium oxide following intratracheal and Carcinogenesis Studies of Indium Phosphide in installations into the lungs of hamsters. Journal of

F344/N rats and B6C3F1 Mice (Inhalation Studies). Occupational Health 52, 14–22. National Toxicology Program (2009) Chemical Teck Cominco Ltd. (2006) Annual report 2005: Vancouver, Information Profile for Indium Tin Oxide [CAS No. British Columbia, Canada, Teck Cominco Ltd. 50926-11-9]. June 2009. U.S. Geological Survey (2002) Minerals Yearbook Niederschlag, E. and Stelter, M. (2009) 145 Jahre Indium - 2002: Indium. http://minerals.usgs.gov/minerals/ Ein Metall mit Zukunft? World of Metallurgy - pubs/commodity/indium/indimyb02.pdf Erzmetall 62, 17–23. U.S. Geological Survey (2008) Mineral commodity sum- NIOSH (2011) NIOSH Pocket Guide to Chemical Hazards. maries: Indium. http://minerals.usgs.gov/minerals/ Indium. http://www.cdc.gov/niosh/npg/npgd0341.html pubs/commodity/indium/mcs-2008-indiu.pdf. Indium 229

U. S. Geological Survey (2009): Minerals Yearbook white blood cell uptake in noninfected closed fracture 2008: Indium. http://minerals.usgs.gov/minerals/pubs/ in humans: prospective study. Radiology 167, 495–498. commodity/indium/myb1-2008-indiu.pdf White, S.J.O. and Hemond, H.F. (2012) The U.S. Geological Survey (2010) Minerals Yearbook Anthrobiogeochemical Cycle of Indium: A Review of 2010: Indium. http://minerals.usgs.gov/minerals/ the Natural and Anthropogenic Cycling of Indium in pubs/commodity/indium/myb1-2010-indiu.pdf the Environment. Critical Reviews in Environmental U.S. Geological Survey (2012a) Mineral commodity Science and Technology 42, 155–186. summaries: Zinc. http://minerals.usgs.gov/minerals/ Wu, J., Becerril, H.A., Bao Z., Liu Z., Chen Y. and pubs/commodity/zinc/mcs-2012-zinc.pdf Peumans P. (2008) Organic solar cells with solution- U.S. Geological Survey (2012b) Mineral commodity processed graphene transparent electrodes. Applied summaries: Copper. http://minerals.usgs.gov/minerals/ Physics Letters 92, 263–302. pubs/commodity/copper/mcs-2012-coppe.pdf. Yamasue, E., Minamino, R., Numata, T. et al. (2009) U.S. Geological Survey (2012c) Mineral commodity Novel Evaluation Method of Elemental Recyclability summaries: Indium. http://minerals.usgs.gov/minerals/ from Urban Mine—Concept of Urban Ore TMR. pubs/commodity/indium/mcs-2012-indiu.pdf Materials Transactions 50, 1536–1540. Van Nostrand, D., Abreu, S.H., Callaghan, J.J., Atkins, F.B., ZSW–Zentrum für Sonnenenergie – und Wasserstoff- Stoops, H.C. and Savory, C.G. (1988) In-111-labeled Forschung (2012) Results – Annual Report 2011, pp. 88. 10. Lithium

KEITH EVANS

Independent Consultant, San Diego, California, USA

Introduction Brande and Humphrey Davy. Its name is derived from the Greek word ‘lithos’, which means ‘stone’. Lithium was discovered in 1817 by Johan Arfvedson when he analysed a sample of petalite and found that it contained “silica, alumina and Properties and abundance in the Earth an alkali”. However, it was a year later when the pure metal was isolated independently by William Due to its high reactivity, lithium only occurs in nature in the form of compounds such as silicates in igneous rocks, in a number of clay minerals Terminology and generally as chloride in brines. The principal Concentration levels of lithium in pegmatites properties are shown in Table 10.1. (and for the one known occurrence of jadarite) The average crustal abundance of lithium is and ore concentrates are normally reported as 17 ppm, but it ranges from approximately 30 ppm in igneous rocks to an average of 60 ppm in sedimen- percentage lithia (Li2O). Brine grades are nor- mally reported as parts per million (ppm), mil- tary rocks. Seawater has an average concentration ligrams per litre (mg/lt) or weight per cent of 0.18 ppm. lithium (wt%Li). Global reserves are normally reported as tonnes Li. Because a great range of chemicals are pro- Mineralogy and deposit types duced, production tonnages are often expressed as tonnes of lithium carbonate equivalents Commercially viable concentrations of lithium (LCEs). Lithium carbonate, much the largest are found in pegmatites, continental brines, geo- volume chemical, contains 0.188% Li. Carbo- thermal brines, oilfield brines, the clay mineral nate is, by definition, one LCE and other major hectorite and the newly discovered mineral, products are metal (5.32 LCEs), bromide jadarite. Locations for the extraction operations (0.425), hydroxide monohydrate (0.88), chloride and advanced stage projects mentioned are (0.87) and butyllithium (0.576). shown on the world map in Figure 10.1.

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Figure 10.1 Location of producing lithium mines and selected advanced-stage lithium projects. 232 keith evans

Table 10.1 Selected properties of lithium. Other lithium minerals in pegmatites have little economic significance but zinnwaldite, Property Value Units mined for chemical production by Metall- Symbol Li gesellschaft on the German/Czech border before Atomic number 3 the second world war, is receiving renewed Atomic weight 6.94 attention. Density at 25 °C 533 kg/m3 Hardness (Mohs scale) 0.6 Melting point 181 °C Atomic radius 145 pm Continental brines Ionic radius of Li+ 90 pm The term continental brines refers to brines in 6 Electrical conductivity 11.7 × 10 S/m enclosed (endorheic) basins where in owing Electric potential 3.04 V surface and sub-surface waters contain modest Specific heat capacity at 3.58 J/(g °C) quantities of lithium which has been released 25 °C from surrounding volcanic rocks as a result of weathering (Ide and Kunasz, 1989). Lithium becomes concentrated along with other elements of economic interest, particularly potassium Pegmatites and boron, as a result of high evaporation rates. Pegmatites are coarse-grained igneous rocks A description of the geology, climate and hydro- formed by the crystallisation of late magmatic chemistry of the Andean region can be found in uids. Lithium-containing pegmatites are Risacher et al. (2003). In China, some occur- relatively rare and frequently also contain tin and rences are open lakes (Zheng, 1989) but in tantalite. South America the brines occur at shallow Spodumene, a lithium–aluminium silicate depths in aquifers composed predominantly

(LiAlSi2O6), is the most common mineral con- of salt or of mixed sediments. In the Andes, the taining a theoretical 7.9% Li2O (or 3.7% Li). In abundance of lithium may originate from North Carolina, the principal area for lithium the Altiplano-Puna Magma Body at depth (de production prior to the development of brines in Silva et al., 2006 and Houston et al., 2011) Argentina and Chile, ore grades averaged 1.4% (Figure 10.2).

Li2O and the richest known deposit, at the Table 10.2 indicates the variability of the Greenbushes operation in Western Australia, active and proposed projects in the Andean has substantial reserves grading up to 3.9% Li2O region. Of note is the very high grade of the (Talison Lithium, 2011a). feed to SQM’s potash and lithium pond system,

Petalite (LiAlSi4O10) is less common and the the very low magnesium/lithium (Mg/Li) ra- only substantial production is from the Bikita tios at the FMC, Lithium One, Orocobre and operation in Zimbabwe where the ground pet- Lithium America’s proposed operations and alite typically grades 4.2% Li2O. the high and very high Mg/Li ratios at the Lepidolite, a lithium-containing mica is Rincon Lithium (Salar de Rincon) and Comibol equally rare. Bikita Minerals was a major (Salar de Uyuni) proposed operations. A high producer of both lump material and otation ratio increases production costs as lithium concentrate but the presence of uorine in the and magnesium concentrate together in solu- product reduced its attractiveness to the glass tion. Figure 10.2 shows the location of these industry. It is used on a small scale for chemical Andean basins, known as salars, and feedstock in China and this yields a by-product Figure 10.3 shows the Salar de Atacama in of caesium and rubidium salts. more detail. Lithium 233

West Cordillera 18°S

Altiplano East Cordillera

Precordillera 20°S

7

22°S

5 1

6 Subandean Ranges

Longitudinal Valley 24°S 4 3 2

Puna 26°S

Coastal Cordillera

Limit of the Altiplano-Puna magma body at depth 28°S shown by dotted lines 100 km (de Silva et al, 2006)

72°W 70°W 68°W 66°W 64°W 1 Salar de Atacama 2 Salar de Hombre Muerto 3 Salar de Diablillos 4 Salar de Rincon 5 Salar de Olaroz 6 Salar de Cauchari 7 Salar de Uyuni

Figure 10.2 Location of salars in the Andean region where lithium extraction is active or proposed. (Compiled by John Houston.)

Foote Mineral Company (now Rockwood/ In 1980, Foote negotiated an agreement with Chemetall) commenced production of lithium CORFO, a Chilean government agency, to eval- from brines at its property in Clayton Valley, uate and develop brines at the Salar de Atacama. Nevada in 1966 and, although grades have The project came on stream in 1984 and the declined significantly, it is still in operation. company’s spodumene-based operation in North Lithium was also recovered at Searles Lake, America ceased production. Two other brine California, where it was considered as a contami- operations commenced in the Andean region in nant in the company’s other products. As a result the late 1990s. The first developed by Sociedad of a process change the recovery of lithium ceased. Quimica y Minera de Chile (SQM) located in 234 keith evans

Table 10.2 Partial analysis of Andean brines. All figures are wt%. (Compiled by the author from published information.)

Salar de Salar de Salar Salar de Salar de Salar de Hombre Hombre Salar de Salar de Salar de Salar de Salar de name Atacama Atacama Atacama Muerto Muerto Rincon Olaroz Cauchari Uyuni** Diablillos

Lithium Lithium Rincon Americas Company Chemetall SQM SQM FMC One Lithium Orocobre Corp Comibol Rodinia

(MOP)* (SOP)* K 1.8 2.97 1.49 0.617 0.62 0.656 0.477 0.57 0.72 0.517 Li 0.147 0.305 0.11 0.062 0.057 0.033 0.057 0.064 0.045 0.046 Mg 0.96 1.53 0.82 0.085 — 0.303 — — 0.65 — Ca 0.031 0.04 0.02 0.053 — 0.059 — — 0.046 —

SO4 1.46 0.88 2.19 0.853 — 1.015 — — 0.85 — B 0.058 0.065 0.068 0.035 — 0.04 — — 0.02 — Mg/Li 6.53 5.02 7.45 1.37 2.19 9.29 2.4 2.34 18.6 3.68

K, potassium; Li, lithium; Mg, magnesium; Ca, calcium; SO4, sulphate; B, boron. Notes: *MOP & SOP analyses are in respect of the feed to the potassium chloride/lithium chloride pond system and the potassium sulfate/boric acid pond systems **Initial production from the Salar de Uyuni will probably commence from an area with a lithium grade of approximately 0.15% Li containing an estimated 400,000 tonnes Li

the northern portion of the Salar de Atacama Materials LLC is proposing production based on a and FMC’s (formerly Lithium Corporation of group of wells which will co-produce electrical America) operation at the Salar de Hombre power. Muerto in Argentina. Potassium chloride and boric acid are actual and potential Oilfield brines co-products and SQM also produces potassium The Smackover Formation extending through sulfate, a specialised fertiliser. east Texas, Arkansas, Oklahoma, Wyoming and Brines in eastern China and Tibet are also North Dakota contains brines grading up to found at high altitudes with high net evaporation 700 mg/lt Li (Collins, 1976). These brines are a rates but complex chemistries have delayed major source of bromine recovered by two com- development compared with earlier projections. panies in Arkansas. Moderate grade brines also exist in the Paradox Basin in Utah and in west Geothermal brines Alberta, Canada several companies are evalu- ating Devonian-age formation waters over an Lithium-containing geothermal brines are known area of 4000 km2. at Wairakei in New Zealand (13 ppm) at the Reykanes Field in Iceland (8 ppm) and El Tatio Hectorite north of and draining into the Salar de Atacama in Chile (47 ppm). A very much higher Hectorite clay (0.53% Li), a magnesium– concentration occurs in the Salton Sea Known lithium smectite, is present in the sedimentary Geothermal Resource Area (KGRA) in southern sequence in the Clayton Valley but large ton- California (Duyvesterin, 1972), where Simbol nages have been identified in the McDermott Lithium 235

CORFO mining claims North

Buffer zone

Lithium isocontours for the top 40 m depth in weight percent 20 Km (Minsal, 1987)

Pond system

0.1 SQM Potassium sulphate-boric acid pond system 0.1 0.2 Proposed NX Uno project 0.3 0.35 SQM Potassium chloride- lithium pond system

Chemetall Potassium Chloride-lithium pond system

Figure 10.3 Landsat 7 ETM + image of the Salar de Atacama showing the outline of the mining claims held by SQM and Chemetall, together with the buffer zone between these claims and the lithium isocontours in red. Background is Landsat 7 ETM+, December 2001 (ETM+, Enhanced Thematic Mapper Plus). (Compiled by John Houston.)

caldera on the Nevada–California border. Jadarite Similar clays have been reported in Argentina, A newly discovered mineral, jadarite (LiNaSiB3O7 Turkey and Mexico. Western Lithium USA Inc. (OH)), with a chemical composition of B2O3 is developing the northern Nevada hectorite 47.2%, SiO2 25.6%, Na2O 15.0%, Li2O 7.3% and deposit. H2O 4.3%, was discovered in Serbia in 2004 by 236 keith evans

Rio Sava Exploration, a subsidiary of Rio Tinto. density separation. The concentrate is then con- The jadarite-containing beds are in three stacked verted from its natural alpha form, which is not layers. amenable to acid leaching, to its beta form by decrepitation in a kiln at a temperature of 1150 °C. This is then attacked by sulfuric acid to produce a Extraction methods and processing number of sulfates including lithium sulfate. This is concentrated, purified and reacted with sodium Currently lithium is produced from pegmatite carbonate to produce lithium carbonate. The sale and continental brine sources. Photographs of of by-product sodium sulfate partly offsets produc- these two widely varying extraction methods are tion costs. Nemaska Lithium is proposing a differ- shown in Figures 10.4 (Talison Ltd, Greenbushes ent technology to take advantage of low electricity mine, Australia) and 10.5 (SQM, Salar de Atacama, costs and high soda ash costs in Quebec and pro- Chile). pose the recovery of both hydroxide and carbonate It appears that most, if not all, spodumene-based from lithium sulfate by electrolysis in a plant to be pipeline projects will use an acid leach process to located near Montreal. produce lithium carbonate – a process developed by Details of the processes proposed for recov- Lithium Corporation of America for its former ering lithium from geothermal brine (Simbol North Carolina operation. Following mining the Materials LLC) and oilfield brines (Albermarle ore is crushed, ground and spodumene and gangue Corp) have not been published. are separated by otation. Some projects where the Western Lithium initially had a choice for spodumene is coarse can dispense with the ota- recovering lithium from hectorite of two tion step and produce an acceptable product with processes – one developed by Chevron and the

Figure 10.4 C3 Pit at the Greenbushes pegmatite in Western Australia, the world’s leading source of spodumene. (Photograph courtesy of Talison Lithium Ltd.) Lithium 237 other by the U.S. Bureau of Mines. They chose solar evaporation to recover the economic prod- the latter with some modifications and it involves ucts. This involves the precipitation of unwanted the thermal decomposition of a mixture of ore, elements such as sodium chloride, excess calcium, anhydrite and gypsum, water leaching and react- sulfate and magnesium but avoiding the precipi- ing the uid with sodium carbonate. The process tation of chemicals of interest in an undesirable results in the co-production of a substantial ton- form such as potash in the form of carnallite nage of potassium sulfate. rather than as sylvinite or the precipitation of Rio Tinto’s preliminary proposal for recov- lithium in the form of a complex salt. Excess ering lithium carbonate from jadarite involves magnesium in the concentrated brine that remains underground mining (room and pillar), multi- in solution increases lithium carbonate recovery stage crushing and wet scrubbing, digestion in costs by requiring larger quantities of sodium hot concentrated sulfuric acid followed by gangue carbonate in converting the lithium chloride to removal by leaching, boric acid production from lithium carbonate. the liquor by crystallisation, magnesium and In a simple case, brine is pumped into a first calcium removal from the boric acid plant liquor set of ponds where the concentrating brine and precipitation of lithium carbonate by sodium precipitates unwanted sodium chloride. At the carbonate addition. For each tonne of lithium appropriate level of concentration the brine is carbonate produced, 4.5 tonnes of boric acid will transferred to a second set of ponds in which syl- be co-produced. vinite is precipitated. This is a mixture of sodium Methods of processing continental brines vary chloride and potassium chloride and when considerably depending on the overall chemistry. harvested the two components are separated in a The basic aim is to concentrate the brines by otation plant. The brine continues to evaporate

Figure 10.5 SQM’s solar evaporation pond system at the Salar de Atacama covering an area equivalent to 7000 football pitches and in excess of 30 km2. (Photograph courtesy of Sociedad Quimica y Minera de Chile S.A. (SQM).) 238 keith evans and, in the case of the Salar de Atacama, this reaches a concentration of six per cent lithium 22% chloride and the material is transferred to 27% chemical plants located near Antofagasta, where, following a series of purification steps, the lithium chloride is reacted with sodium car- bonate to produce lithium carbonate. At the Chemetall operation the feed into the ponds is a 3% mixture of two brine types. 1% At FMC’s Hombre Muerto operation, solar 3% evaporation plays a minor role and the process is basically an ion-exchange one with the selective 3% extraction of lithium chloride. The process is 4% being modified and will allow the recovery of 14% potassium chloride. 6% At the Salar de Rincon a first set of solar ponds raises the lithium content to 2.5 grams per 8% 9% litre. It then passes to a reactor where both hydrated lime and sodium sulfate are added to Li-ion batteries Greases precipitate magnesium hydroxide and calcium Glass-ceramics Air treatment sulfate. Brine is then returned to a second set of Ceramic and enamel frits Polymers solar ponds for further concentration to chemical Aluminium smelting Metallurgical plant feed grade. At Orocobre’s operation at the Salar de Olaroz there is a similar treatment of Primary batteries Pharmaceuticals brine at an early stage and SQM operates two Other pond systems – one for a high-sulfate brine and the other for a high-potash, high-lithium brine Figure 10.6 Lithium chemical uses, 2010. (Data from further south. Roskill Information Services Ltd, personal communication.)

Specification and uses There are only two major sources – approxi- The market for lithium is divisible into two mately 80,000 tpa of low iron content spodumene major segments. The first is for mineral concen- concentrate from Talison’s Greenbushes pegma- trates with spodumene concentrate the dominant tite in Australia and approximately 50,000 tpa of product and lesser tonnages of petalite and lepid- petalite from Bikita in Zimbabwe. olite. The second segment is for a large range of Chemical demand – approximately 200 lith- lithium chemicals and metal (further detail is ium-containing products are marketed. As is the available from Harben, 2002). case with non-chemical demand, Asian demand Non-chemical demand – glasses, ceramics and dominates, accounting for more than 50 per cent glass ceramics account for a high percentage of by value, with Europe and the Americas dividing demand where the principal value of lithia most of the rest almost equally. addition is to reduce melting temperatures and to A breakdown by principal applications is provide thermal shock resistance. The concen- shown in Figure 10.6 and the distribution by trates contain silica and alumina – important in major products approximates to lithium many glass and ceramic formulation. The other carbonate 42 per cent, hydroxide 14 per cent, significant use is in continuous steel casting. butyl-lithium five per cent, metal four per cent, Lithium 239 chloride four per cent and others nine per cent. Nissan Leaf (24 kilowatt hour) requires 14.4 kg. Carbonate is the precursor chemical for all Currently, the main battery types have graphite other chemicals with the exception of metal anodes and variable lithium-containing cathodes and metal derivatives which are derived from of lithium cobalt, lithium manganese spinel, lithium chloride. lithium iron phosphate and tri-element blends. Lithium carbonate – the main current use for Current lithium-ion batteries have storage capac- carbonate is in glasses and ceramics, glass ities of 200 watt hour/kg with a theoretical poten- ceramics, enamels and glazes where quality and tial of double this. The ultimate potential for other factors rule out the use of mineral concen- lithium batteries is projected to be in a lithium– trates. The same factors apply as with mineral sulfur and lithium-oxygen battery with a lithium concentrates – the strong uxing power of lithia metal anode. Both are in the distant future. In the based on its small ionic radius reducing the meantime the search goes on for greater energy melting point, lowering the viscosity and thermal storage (greater range) and faster charging expansion and increasing density, workability, time. Typically, the cost of carbonate in a lith- chemical durability and hardness. In the case of ium-ion battery is less than three per cent of the glass ceramics the thermal expansion is reduced battery cost. to near zero and is used in cooktops, cookware In addition to that for motor vehicles, demand and large telescopic lenses. for lithium-ion batteries could be considerable In aluminium electrolysis, carbonate is added for the storage of intermittent sources of energy, to the cell converting to lithium uoride to lower such as solar and wind, as well as load-levelling the liquefaction temperature, thus reducing the in a wide range of applications from regional grids operating temperature and reducing emissions. to individual factories and residences. Substantial tonnages of carbonate are used as an Lithium hydroxide monohydrate is used prin- accelerator for quick-setting, high- and medium- cipally in multi-purpose lubricating greases. alumina cements. These greases are effective over a wide tempera- Carbonate is the principal chemical used for ture range and have excellent water resistance. In the production of cathodes for lithium-ion bat- its anhydrous form, hydroxide is most frequently teries although hydroxide is increasingly being used as a carbon-dioxide absorbent in closed sys- used in this application. tems such as submarines and spacecraft. Carbonate is sold in a variety of grades depend- Following the explosion of one of its oxygen ing on purity, particle size, and particle-size dis- tanks, a rigged system using the chemical saved tribution. Industrial grade is normally guaranteed the lives of the Apollo 13 crew. at 99 per cent purity, with battery grade at 99.5 Lithium chloride as a solid is used as a ux, per cent purity. Several producers claim purity particularly in welding aluminium, and as a solu- levels of 99.9 per cent. tion is used for controlling humidity in food Mobile phones require three grams of lithium processing, pharmaceutical manufacturing and carbonate for their batteries, while notebooks hospitals where it has a sanitising effect. and power tools require 30–40 g. Motor vehicles Lithium bromide has a high solubility in require 0.6 kg per kilowatt hour, with a mild water, and brines with a 54 per cent concentration hybrid (one kilowatt hour) requiring two kg, a are used in large scale absorption-refrigeration plug-in hybrid (16 kilowatt hour) requiring 15 kg systems. and an all-electric vehicle (25 kilowatt hour) Lithium metal is produced by the electrolysis requiring 22 kg of lithium carbonate (or of a mixture of potassium and lithium chloride, equivalent). These figures vary somewhat with normally in a conventional sodium metal cell. differing battery chemistry, but, as examples, the Production is in the form of ingots, rods, gran- Chevrolet Volt and its Opel equivalent (16 kilo- ules, foils and powder and generally at several watt hour) on this basis requires 9.6 kg and the levels of purity based on sodium content. Lithium 240 keith evans metal is highly reactive. It is used in primary regulations. All the following process steps need (non-rechargeable) battery anodes having the to clearly separate the materials by battery chem- highest electrochemical potential with the lowest istry. This separation is facilitated by labels with mass and a very long shelf and operating life. A battery chemistry information, which should be major use in metal is in the production of a wide harmonised worldwide and almost impossible to range of organolithium compounds, the main remove from the battery casing. The next step is one being butyllithium used in the manufacture the removal of the battery case. The BMS is of synthetic rubber. Increasing quantities of brought into the existing electronics recycling lithium–aluminium alloys are being used in scheme (for example, in Europe, under the Waste aircraft, principally for weight saving. Electrical and Electronic Equipment (WEEE) Directive) and the cells can be removed from the battery modules. The cells can feed a smelter for Recycling the pyrometallurgical treatment (see above) or feed a shredder for the hydrometallurgical route. Potentially, after a period of probably eight to ten For the hydrometallurgical process, after shred- years following the large-scale introduction of ding, the resulting powders must be carefully sep- large batteries, there will be a need to recover and arated. This is done by means of sophisticated recycle the valuable components. In many separation steps, in which copper, aluminium as countries it will be mandatory as a result of well as cathode material are separated. legislation. The refined cathode material is the head feed Electric vehicle (EV) batteries differ signifi- for Chemetall’s extraction process. A lithium- cantly from existing batteries in respect of weight, containing fraction as well as a transition metal volume, energy content, modular structure, bat- fraction is obtained. The lithium solution is tery management system (BMS) and cell chem- further refined and subsequently transformed by istry. Consequently, new recycling processes an electrochemical process to high-purity have to be developed or existing systems used for lithium hydroxide. This salt can be used directly portable batteries need to be modified. In general for the production of new active cathode mate- there are two processes applicable for EV battery rials. The transition metal salts can also be used material recycling, each with fundamentally for producing the same product without further different approaches – pyrometallurgical and refining. hydrometallurgical. The hydrometallurgical process can be applied The focus with pyrometallurgical processes to all lithium-ion battery chemistries, including lies in the recycling of transition metals like those without cobalt and nickel. valuable cobalt and nickel. After smelting the battery cells, an alloy of cobalt, nickel and copper is purged out for further hydrometallurgical Substitution refinement. Lithium will remain in the slag and is lost for high-value applications as extracting Most of the applications of lithium have existed lithium from the slag is probably not economi- for decades and have grown at rates comparable cally viable. However, with hydrometallurgical with those of the world’s economy. In glasses, processes it is possible to get high-purity lithium glass ceramics and ceramics other uxes can be salts and transition metal salts back from the EV used but none increases thermal shock resis- battery stream. tance, for example, as much as with lithium car- The recycling chain for EV batteries is fairly bonate. The market share held by lithium stearate complex. Starting with dismounting the battery in multi-purpose greases has not declined in the from the car, the battery must be discharged to a face of alternative formulations. In aluminium certain level for meeting transport and handling smelting the main competition has been the Lithium 241 result of improvement in capturing emissions – for recycling large lithium-ion batteries is dis- particularly uorine. cussed elsewhere. In glasses, glass ceramics and In primary batteries the competition is from ceramics, the lithium is entrained in the products other alkaline cells, while lead acid batteries from which it cannot be leached after disposal. are currently dominant in rechargeable batteries. Lithium products are used in the synthesis of In applications where weight and energy density many drugs, in small quantities in the treatment are important lithium-ion is progressively replac- of manic depression and as a sanitiser in food- ing nickel cadmium and nickel metal hydride manufacturing facilities, laundries, swimming varieties and there is no current competition for pools and hospitals. It presents few environ- lithium-ion in most electric vehicles. The com- mental problems in these applications. petition here is the continued use of oil, natural gas, hydrogen and biofuels. World resources and production Environmental factors Reserves and resources Any new project and existing projects are faced In the mid-1970s the National Research Council with numerous environmental hurdles. For in the United States created a committee to example, in the case of a new mining and examine a range of issues including an processing project in west United States the assessment of world reserves and resources of company needs permits from the local authority, lithium. The result was subsequently published the State and the Federal governments. This (Evans, 1978) which estimated a total of 10.65 requires intensive environmental baseline sur- million tonnes of lithium in the western world, veys including those for vegetation, wildlife, as little data were available concerning Russia threatened or endangered species, cultural and China. resources, air quality and water resources. This Subsequently, there have been major discov- will result in a document in compliance with the eries, particularly of brines in western China and National Environment Policy Act. Additionally, the Andes. However, the development of the any company will require a host of permits potentially large-scale use of lithium-ion bat- regarding the mining and processing of the ore teries in motor vehicles has caused lithium avail- and air quality will be a major issue with processes ability to become a major issue and concerns requiring roasting or calcining ore. have been raised suggesting that resources are Any pegmatite development will, in most inadequate to support the large-scale electrifica- jurisdictions, require similar permits as with tion of vehicles. any new mining operation and when onward The reserves and resources assessment has processing from concentrate to carbonate is pro- been updated several times (e.g. Kunasz, 2007; posed the operation of a kiln will have an impact Evans, 2008; and Yaksic and Tilton, 2009) and on air quality. the current version is shown in Table 10.3. In conventional brine-based projects, most of With minor exceptions the tonnages are in situ. the energy used to remove water and increase the The only exceptions are in respect of the totals concentration of elements of interest is solar in for North Carolina undeveloped and the origin. The predominant waste product is sodium extremely large pegmatites in the Democratic chloride, which can be safely deposited back on Republic of Congo where the tonnages used in the salar’s surface. the National Research Council report are Used batteries containing lithium in cell quoted, which allowed for estimated mining phones and computers are, or should be, collected and processing losses. The listing is of concen- along with other electronic waste. The potential trations of lithium of economic grade with a 242 keith evans

Table 10.3 Estimated world lithium resources compiled by the author based on previous work and published sources.

Tonnes Li

Pegmatites North Carolina, USA *undeveloped 2,600,000 Manono, *D.R. Congo 2,300,000 Greenbushes, Australia (Talison) 1,500,000‡ Russia, numerous 1,000,000 China, numerous 750,000‡ Canada, others 430,000 North Carolina, USA (Former Operations) 230,000 Quebec Lithium, Canada (Canada Lithium) 230,000 Whabouchi, Canada (Nemaska) 187,000 Karalpa, Austria (E. Coast Minerals) 134,000 Lithium One, Canada (James Bay) 130,000 Mibra, Brazil (CIF Mineracao) 100,000 Mount Marion, Australia (Reed) 93,500 Mount Cattlin, Australia (Galaxy) 90,000‡ Brazil, other 85,000 Bikita, Zimbabwe 56,700‡ Lantiar, Finland (Keliber Oy) 14,000 Total Pegmatites 9,930,000‡ Continental Brines Salar de Uyuni, Bolivia (Comibol) 8,900,000 Salar de Atacama, Chile (SQM & Chemetall) 6,900,000‡ China & Tibet, numerous 2,600,000‡ Salar de Cauchari, Argentina (Lithium Americas) 1,520,000 Salar de Rincon, Argentina (Rincon Lithium) 1,400,000‡ Salar de Olarez, Argentina (Orocobre) 1,200,000 Sal de Vida, Argentina (Lithium One) 1,020,000 Salar de Hombre Muerto, Argentina (FMC) 850,000‡ Salar de Diablillos, Argentina (Rodinia) 530,000 Salar de Maricunga, Chile 200,000 Silver Peak, Nevada, USA (Chemetall) 40,000‡ Total Continental Brines 25,160,000‡ Others Hectorite Kings Valley, U.S.A (Western Lithium) 2,000,000 La Ventana, Sonora, Mexico (Bacanora) 180,000 Geothermal Brines Brawley, California (Simbol) 1,000,000 Jadarite Jadar, Serbia (Rio Tinto) 950,000 Oilfield Brines Smackover Formation, U.S.A (Albermarle) 850,000 Total Others 4,980,000 Total Overall 40,070,000

*Adjusted for mining losses ‡Includes reserves at producing operations (where published) Lithium 243 wfreasonable and realistic prospect of economic make estimates of recoveries through the various extraction. stages of production from most source types. The United States Geological Survey in 2011 In the case of pegmatites, ore is generally estimated resources (a combination of reserves clearly distinguishable from the host rock (as and resources) as 33 million tonnes Li (USGS, shown in Figure 10.7) and mining recovery 2011), up from 25.5 million tonnes in 2010 should normally be high: Canada Lithium, for (USGS, 2010) and 13.8 million tonnes in 2009 example, estimates 80–85 per cent. Recovery (USGS, 2009). Clarke and Harben (2009) in a through otation and chemical conversion is ‘Lithium Availability Wall Map’ estimated a estimated at 67 per cent (Canada Lithium, 2011a). ‘Broad Base Total Reserve’ of 39.4 million tonnes With a hectorite feed Western Lithium estimate Li (including both producers and developing pro- a recovery of more than 87.2 per cent through the jects). In a 2011 update of the Wall Map, Clarke entire process (Western Lithium, 2011). Rio (2011) quoted a reserve of 38.3 million tonnes Tinto, with its proposed underground operation and resources (all categories) of nearly 62 mil- in Serbia, estimates a 40 per cent recovery from lion tonnes Li. SQM estimated global reserves mining using a room-and-pillar system, 75 per and resources at 18.8 and 56.4 million tonnes Li, cent recovery through beneficiation and 80 per respectively (Solminihac, 2009). cent recovery in the chemical plant (Kellie, Unfortunately, many commentators have 2009). ignored the fact that reported figures are not Significant losses can occur with a brine feed always recoverable resources and, particularly – a combination of losses in the pond system due because of the publication of scoping, pre-feasi- to brine entrainment in salts precipitated in the bility and feasibility studies, it is now possible to solar ponds and in the chemical plant. Losses at

Figure 10.7 The Whabouchi spodumene pegmatite cutting through host rocks in the James Bay region, Quebec, Canada. (Photograph courtesy of Nemaska Lithium Inc.) 244 keith evans the Salar de Atacama are thought to be in the acts (CEOLs) similar in type to those held by range of 30–35 per cent. Orocobre estimate such Rockwood and SQM. The government decided on losses at the Salar de Olaroz at 26 per cent the second option and in September received (Houston and Gunn, 2011). FMC’s losses at the three bids for 20-year concessions to produce up Salar de Hombre Muerto should be lower but to 100,000 tonnes of lithium over 20 years, in there are no published data. Lithium One esti- return for an up-front payment and royalty mates process recoveries at its Sal de Vida project payments of seven per cent on sales. SQM’s bid of at 60 per cent. US$40 million appeared to be the clear winner The other major factor regarding most brine but it soon emerged that the company was operations concerns the recovery from the host engaged in litigation with several Government aquifer. Houston et al. (2011) suggest that a con- agencies – grounds for exclusion from the process. servative estimate of one third should be assumed Li3 Energy was the second highest bidder with an as there are no known examples of a salar-type offer of US$17.5 million for a project based on a aquifer of known initial volume being drained to relatively small portion of the Salar de Maricunga the point where further abstraction ceases to with an estimated measured and inferred resource become viable either technically or economi- grading 1250 g/lt totaling 120,000 tonnes Li. cally. This loss would not exist where the feed is Subsequently, the Government abandoned the from saline lakes. Recoveries from other poten- concept of granting additional leases. tial lithium sources from geothermal and oilfield Companies other than Rio Tinto are exploring brines are not known. for additional jadarite deposits in Serbia and the Offsetting these reductions in recoverable second bromine producer from the Smackover lithium is the fact that exploration activity is at a Formation could follow Albermarle’s lead. In high level. Numerous pegmatites in Namibia, Alberta, Canada, numerous companies are fol- Mozambique, Ireland and particularly Canada, lowing up on a report by the Provincial Research where many pegmatites were partially explored Council and the Geological Survey that oil field during an earlier lithium exploration ‘boom’ in brines of Devonian age contain a resource of the 1960s, are currently estimated to contain approximately 500,000 tonnes Li (Bachu et al., 450,000 tonnes Li and most remain open along 1955). Concentration levels, though, are low at strike and at depth. about 140 ppm. In Argentina numerous other salares including Despite the very low concentration of lithium Rio Grande, Arizaro, Mariana, Antofalla, Ratones, in sea water, Korean companies and research Pozuelos, Pocitos, Huayatayoc and Salinas Grandes organisations are funding a three-year project are known to be lithium containing. In Chile, the with a value of US$30 million to evaluate pro- Geological Survey has identified numerous salares duction economics from this source. that could possibly be viable. These include If the massive-scale electrification of motor Pedernales, La Isla and Quisquiro with lithium vehicles actually occurs, the batteries are samples grading between 423 and 1080 mg/lt Li and expected to have a life of eight to ten years. The Punta Negra, Aquas Calientes, Pajonales, Aquilar, lithium is not consumed in the battery and, if it Tara, Parinas and Pujsa with concentrations of bet- is economically viable to recycle, the demand for ween 220 and 620 mg/lt Li (Solminihac, 2010). virgin material will be greatly reduced. In early 2012 the Chilean government announced its intention to expand lithium Production production. As the mineral is classified as stra- tegic under the 1973 Mining Code there were two Worldwide lithium production (Li content) options. The first was to change the law which dropped in 2009 by 31 per cent compared to 2008 required parliamentary approval and the second to less than 12,700 tonnes. However, output in was to award Special Lithium Operation Contr- 2010 recovered to 20,000 tonnes of Li content, Lithium 245

22 20 18 16 14 12 10 8 6 4 Figure 10.8 Worldwide annual Thousand tonnes Li content 2 production of lithium 1996–2010 (Li content). (Data from World Mineral 0 Statistics Database, British Geological

Survey.) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

6% 2008 (Figure 10.8). Chile was again the world’s 4% largest producing country in 2010, when mea- sured by Li content, with output amounting to 8% more than 9900 tonnes Li content, or 50 per cent of the world’s total. Australia, the USA and Argentina are the next largest producers and together the top four represented 89 per cent of the world’s total in 2010 (Figure 10.9). 14% 50% Current producers Rockwood Holdings/Chemetall In the early 1980s Foote Mineral Co. reached an agreement with the Government of Chile which 18% allowed it to evaluate and develop a block of claims at the southern end of the Salar de Chile Australia USA Atacama (Kunasz, 1983). Initially, the government through CORFO, a development agency, was a Argentina China Others partner but they progressively sold their interest to Foote. Production commenced in 1984 with a Figure 10.9 Lithium producing countries 2010 (based capacity of 13,000 tonnes per annum of lithium on tonnes Li content). (Data from British Geological carbonate and production from its spodumene- Survey, 2012.) based operation in North Carolina ceased. Foote was later acquired by Chemetall which, in turn, was acquired by Rockwood Holdings. which is just one per cent down compared to The combined capacity of both brine opera- 2008. The trend over the previous decade has tions in Nevada and Chile is currently 38,000 been for increasing production with worldwide tonnes per annum of carbonate and hydroxide. A output rising by 79 per cent between 1999 and phased expansion to 65,000 tonnes per annum 246 keith evans by 2020 has been announced. The Atacama Carolina closed. Current capacity is approxi- project also co-produces potassium chloride. The mately 17,500 tonnes per annum but an expan- company produces an extensive range of down- sion to 23,000 tonnes per annum has been stream lithium chemicals in North America and announced. The company produces a full range Germany. of downstream chemical products in the United States. Sociedad Quimica y Minera (SQM) In 1983 the Chilean Government agreed to put Talison Minerals the remaining CORFO claims up for bid. Amax The company acquired the Greenbushes peg- Exploration and a Chilean partner won the right matite in Australia in 2007. The weathered to enter into negotiations with the Government surface material was first mined for tin start- taking the position that the new operation would ing in 1888 and for tantalite commencing in be restricted to producing potash and boric acid the early 1940s. Lithium was discovered in (Evans, 1986). However, the right to co-produce unweathered pegmatite in 1949 and the operation lithium was eventually conceded. After com- has grown to be the world’s largest producer of pleting its feasibility study, Amax decided not to lithium which is contained in spodumene con- proceed and the project was acquired by SQM. centrates grading between 4.8% and 7.5% Production commenced in 1996. SQM reduced Li O. the carbonate price by approximately 50 per cent 2 A series of pits has been developed along the because of the high concentration of lithium in 3.5-km strike length of the pegmatite and in 2009 the brine and their wish to make a rapid entry the company announced an overall resource into the business estimate of 1.5 million tonnes Li subdivided into In late 2011, SQM announced that it was a ‘lithium resource’ of 35.5 million tonnes grad- increasing carbonate capacity from 40,000 ing 3.31% Li O and a ‘tantalite resource’ of 190.8 tonnes per annum by 50 per cent. Further expan- 2 million tonnes grading 1.1% Li O. More recent sions are possible at relatively low cost because 2 drilling results in the higher-grade area produces with potash production at a very high level the an estimate of proved and probable reserves of volume of lithium in the pumped brine is so 31.4 million tonnes grading 3.1% Li O, with mea- high that approximately 400,000 tonnes per 2 sured and indicated resources of 70.4 million annum of carbonate equivalent in a concentrated tonnes grading 2.6% Li O (Talison Lithium, brine is returned to the aquifer through a series 2 2011a). of injection wells. The company produces car- In addition to size and exceptionally high bonate, hydroxide and chloride at a chemical grade, the pegmatite contains a substantial ton- plant located close to that of Chemetall near nage of ore which yields a concentrate with a Antofagasta. very low ferric oxide content (less than 0.1 per cent). This makes it an attractive source of FMC Corporation lithium for the glass and ceramic industry and After an apparently successful negotiation with the only significant competition is the petalite the Bolivian government regarding possible produced by Bikita Minerals in Zimbabwe. development of the Salar de Uyuni the agreement However, the bulk of spodumene concentrate was strongly rejected by the local population. produced by the company, grading 6% Li2O, is The company withdrew and eventually acquired shipped to China for conversion into lithium the rights to develop the Salar de Hombre Muerto chemicals by a number of companies. Talison is in Argentina. The project came on stream in increasing its concentrate capacity to 740,000 1997 and its pegmatite-based operation in North tonnes per annum, most of which will be for Lithium 247 chemical conversion and consideration is being Spain and Portugal given to developing the company’s own chemical The British Geological Survey reports a substan- production capacity by 2015. tial production of lepidolite in Spain and Portugal In early January 2012 Talison announced com- (5000 and 40,609 tonnes, respectively, in 2010; pletion of its major expansion at Greenbushes BGS, 2012). It is believed that these tonnages are and later in the month received a takeover offer of a quartz–feldspar–lepidolite mixture used in from Rockwood at a 53 per cent premium over the local ceramics industry and not entering the Talison’s then current price. Two months later international trade. Chengdu Tianqui Industry Co offered a higher price of A$792 million and this was accepted. Brazil Tianqui is the major convertor of spodumene into chemicals and has been Talison’s main Companhia Brasileira de Litio (CBL) in Brazil had customer. Conceivably, Tianqui will undertake the capacity to produce approximately 2000 some conversion to chemicals in Australia with tonnes per annum of LCE from locally sourced a saving in freight costs; a possibility that Talison spodumene and Companhia Industrial Fluminense was considering. Earlier in the year the company (CIF) is developing a pegmatite source at Mibra had acquired a significant interest in Nemaska with a tantalite grade of 300 grams per tonne Lithium in Quebec. which also contains a lithium resource of 21.3

million tonnes grading 1.0% Li2O. Bikita minerals China The Bikita pegmatite in Zimbabwe is unique in many respects. The area developed initially Lithium-containing pegmatites have a wide dis- was a classic zoned pegmatite with a massive tribution and include the deposits Maerkang, lepidolite core which was the feedstock for a Daoxian, Jiajika, Kokalay, Jinjuan, and Ningdu third of the USA stockpile of lithium hydroxide and resources could total an estimated 750,000 purchased by the Atomic Energy Commission tonnes lithium. Most appear to be high cost or commencing in 1953. Approximately half the have concentrates of poor quality, as there seems material was subjected to an isotope separation a distinct preference amongst the major chemical process to breed tritium from the 6Li for the producers for Australian imports despite the high hydrogen bomb programme. The entire stock- freight costs. Domestic spodumene production pile of both virgin and depleted hydroxide was satisfies less than 25 per cent of internal demand. disposed of by the US Department of Energy in China has a number of major brine lake the mid-1990s. sources. A consortium of Tibet Zhabuye Further north, the pegmatite becomes a mix- Lithium Industry High Tech Group, Tibet ture of large crystals of petalite together with Mineral Development, Yuxin Trading and BYD, spodumene and quartz. Initially, production was the car manufacturer, have rights to the based on picking of coarsely crushed run-of-mine Zhabuye Salt Lake in Tibet. Qinghai Salt Lake ore but current production is based on density Industry Group and Western Mining Group separation. Petalite, the principal product, grades operate at the Dongtai Salt Lake, also known as between 4.0 and 4.2% Li2O and has an Fe2O3 East Taijinar. Qinghai National Security Co and content of 0.03% thus making it attractive to CITIC have rights to West Taijinar also known glass and ceramic customers. Production over the as Xitai Salt Lake. The brines are generally com- last three years (2010–2012) has totalled 47,000, plex with relatively low production levels in 48,000 and 53,000 tonnes of a variety of grades of comparison with their reserves but one projec- petalite. tion is that the two Lake Taijinar producers will 248 keith evans increase capacity to 30,000 tonnes carbonate and carbonate from its Whabouchi pegmatite in over the next five years. A new entrant will be Quebec. Figures released are in respect of an 18 Zong Chuan International Mining at the years life of mine project with an initial capital Dangxiongcuo (DXC) salt lake in Tibet. cost of US$454 million. Production of 3.8 million Producers of lithium chemicals in China are tonnes of 6% concentrate will allow the produc- numerous, many producing a full range together tion of 366,000 and 177,000 tonnes of battery with metal. The major technical grade (generally grades of hydroxide and carbonate, respectively, 99.3 per cent purity) carbonate producers in and average costs of US$3400 per tonne and China are Sichuan Tianqi Lithium, Qinghai US$3500 per tonne. Annual production tonnages Guoan, Qinghai Lithium Industry, and Tibet are estimated at approximately 20,700 and 10,000 Minerals with a joint production of 8500 tonnes for the two products but the plant will be designed in 2010. Convertors to battery grade (generally to be exible. Product will be recovered from 99.5 per cent purity) carbonate are Sichuan lithium sulfate by electrolysis (Nemaska Tianqi, Sichuan Ni & Co, Ronghui and Xinyu Lithium, 2012). Ganfeng. Production in 2010 approximated to In Australia, Talison estimates the capital cost 7000 tonnes. Convertors to lithium hydroxide for its expansion from 315,000 tonnes per annum were Sichuan Tianqi, Sichuan Ni & Co, Kinjian to 740,000 tonnes per annum of concentrates at Non-ferrous, ABA Guansheng and Minfeng with between A$65 and A$70 million (Talison Lithium, a 2010 production of 10,800 tonnes. Sichuan 2011b). Tianqi is the sole producer of lithium chloride at Also in Australia, Reed Resources announced 1500 tonnes in 2010. in October 2012 the results of a new Pre- feasibility Study giving more emphasis to Production costs hydroxide production rather than carbonate production from a chemical plant to be located Existing producers have not disclosed their pro- in Malaysia. The project entails importing duction costs but a general assumption is that 147,000 tpa of spodumene concentrate grading those incurred by the two producers at the 6% Li O, to produce 10,000 and 8800 tpa Salar de Atacama are the lowest due to the high 2 hydroxide and carbonate, respectively, from its grades of the brine and exceptional climatic Mount Marion pegmatite in Australia. The conditions. capital cost is estimated at US$83 million with The recent completion of feasibility studies production costs of US$3828 for hydroxide and for a number of projects based on pegmatites, US$4538 for carbonate. Spodumene concentrate continental brines and hectorite reveal cost delivered to the plant is priced at US$350/tonne estimates but no such data are available yet on (Reed Resources, 2012). geothermal brine, oilfield brine or jadarite. Altura Mining intends developing the Pilgangoora project also in Western Australia. A Pegmatites Scoping Study estimates a capital cost of A$96.3

Canada Lithium, with a feed grade of 0.85% Li2O, million to produce 150,000 tpa of 6% concentrate. estimates a total capital cost of US$202 million Cash operating costs are estimated at A$16 per for the mine, mill and chemical plant to produce tonne ore and A$90 per tonne concentrate 20,000 tonnes per annum of battery-grade car- ex-plant (Altura Mining Ltd, 2012). bonate. Average cash operating costs, after a major revision, are estimated at US$3,164 per Hectorite tonne LCE (Canada Lithium, 2011a). In October 2012, Nemaska Lithium announced Western Lithium USA estimates capital costs for the completion of a Preliminary Economic Asses- the initial production of 13,000 tonnes per sment for the production of lithium hydroxide annum carbonate at US$248 million rising to Lithium 249

US$409 million when production is increased to in October 2011. For a target production of 25,000 26,000 tonnes per annum. Cash costs for car- tonnes per annum of lithium carbonate and bonate are estimated at US$3472/tonne but with 107,000 tonnes per annum of potassium chloride, co-product credits, principally from potassium capital costs are estimated at US$356 million and sulfate, these are reduced to US$1,967/tonne production costs for carbonate, FOB Antofagasta (Western Lithium, 2011). at US$1537/tonne. The company points out the In January 2013, Bacanora Minerals announced significant contribution that potash makes to the completion of a Preliminary Economic project economies where sales revenues are Assessment for its La Ventana hectorite deposit sufficient to more than cover total operating in Sonora, Mexico, with a resource of 600 expenses (Galaxy Resources, 2013a). million tonnes grading 3000 ppm Li. At a pro- In late 2011, Rodinia Lithium announced duction rate of 35,000 tpa carbonate and initial the results of its Preliminary Economic capital costs of US$114 million, it estimates Assessment for its proposed operation at the average operating costs of US$1958 per tonne Salar de Diablillos. Two potential scenarios are (Verley and Vidal, 2013). The announcement did considered. The first is for the production of not disclose a processing method and the cost 15,000 tonnes per annum lithium carbonate estimates differ greatly from those of Western and 51,000 tonnes per annum potassium chlo- Lithium. ride with a capital cost of US$144 million and a cash cost of $1519/tonne for carbonate. The second option is to produce 25,000 tonnes per Continental brines annum carbonate and 85,000 tonnes per annum In Argentina, studies have recently been published potash. Capital for the larger case is estimated regarding anticipated costs at four somewhat at $220 million with a cash production of similar brine-based projects. $1486/tonne for carbonate. In both cases potash At the Salar de Cauchari, Lithium Americas revenues are said to cover total operating costs Corp, plans on producing 40,000 tonnes per (Rodinia Lithium, 2011). annum carbonate with development in two No cost estimates are available for other phases. Capital for Phase 1 with a production of projects. 20,000 tonnes per annum on stream in 2014 is estimated at US$217 million with an additional US$181 million required to double capacity three Future supplies years later. Cash operating costs are estimated at US$1434/tonne (Lithium Americas, 2011). The prospect of a major increase in demand for At the adjacent Salar de Olaroz, Orocobre lithium, particularly battery-grade carbonate has intends producing 16,400 tonnes per annum of resulted in a high level of exploration activity carbonate at a capital cost of US$207 million and several projects have reached an advanced and a cash operating cost of US$1512/tonne. stage. Approximately three years into the project the co-production of 10,000 tonnes per annum Pegmatite-based projects of potash will commence. This will require an additional investment of US$14.5 million Canada Lithium Corp has acquired the Val d’Or and the credit from this will reduce the car- property which was formerly operated as an bonate production cost to US$1230/tonne underground operation between 1955 and 1965. (Orocobre, 2011). The latest resource estimate for measured and A preliminary economic assessment of what indicated resources is 33.2 million tonnes grading was formerly Lithium One’s Sal de Vida project 1.19% Li2O (Canada Lithium, 2011b). Plant (now acquired by Galaxy Resources) was released construction has commenced. 250 keith evans

Nemaska Lithium, also in Quebec, is devel- Copiapo (a nitrate and iodine producer) and oping a moderately high-grade pegmatite with Korean interests proposed to develop the NX Uno measured and indicated resources of 19.6 million project on the western margin of the Salar de tonnes grading 1.49% Li2O. Initially the company Atacama with an initial target of 200,000 tpa of planned to produce only spodumene concentrate, potash and 20,000 tpa lithium carbonate. Feed to but the new target is to produce both, hydroxide the operation would have approximated to 0.15% and carbonate (Nemaska Lithium, 2012). Li. At the Salar de Maricunga Li3 Energy has A third Quebec project near James Bay was announced an indicated resource of 120,000 initially developed by Lithium One with 11.75 tonnes Li grading 1250 mg/lt in claims covering 2 Mt at 1.3% Li2O of indicated resources and 10.47 14 km . Neither project can now proceed due to

Mt at 1.2% Li2O inferred (Galaxy Resources, the cancellation of the Government’s proposal to 2013b). The property was acquired by Galaxy award additional contracts for the production of Resources at the same time as they purchased the lithium. Numerous other exploration pro- company’s brine prospect in Argentina. grammes have, almost certainly, ceased in the In Europe, Keliber Oy plans to produce 6000 country. tonnes per annum carbonate from a group of In Argentina, the evaluations of four salares pegmatites but production, originally scheduled have reached an advanced stage. Galaxy’s for 2010, has been delayed. (formerly owned by Lithium One) Sal de Vida In Western Australia, Galaxy Resources is project is adjacent to FMC’s existing operation at mining the pegmatite at Mt. Catlin with upgraded the Salar de Hombre Muerto. The inferred resources (in all categories) announced in March resource comprises a clastic (sediment-filled)

2011 of 197,000 tonnes of contained Li2O (Galaxy body containing 1.02 million tonnes lithium Resources, 2011). The target is to produce 127,000 grading 695 mg/lt Li and 11.0 million tonnes of tonnes per annum of concentrate grading six potash (Galaxy Resources, 2013a). A later dis- per cent and ship it to a company-owned chem- covery was a salt basin (similar to FMC’s source) ical plant in an industrial zone near Shanghai. with significantly higher grades. Carbonate (17,000 tonnes per annum) will be Rodinia Lithium at the Salar de Diablillos used in the production of e-bike batteries. At claims a recoverable inferred resource grading capacity, production will total 17,000 and 4000 556 mg/lt Li and 6206 mg/lt K containing 2.82 tpa carbonate and hydroxide, respectively. million tonnes of lithium carbonate and 11.27 At Mount Marion in Western Australia, Reed million tonnes of potassium chloride (Rodinia Resources plans to produce 147,000 tpa of spodu- Lithium, 2011). The company is considering mene concentrate from a resource of 10.5 Mt two levels of production and potash sales are grading 1.4% Li2O and ship it to a plant proposed expected to cover all cash operating costs for the for Malaysia to produce both hydroxide and car- project. bonate (Reed Resources, 2012). At the Salar de Rincon (a salt-filled salar) Altura Mining, in a relatively recent discovery Rincon Lithium, which is 100 per cent owned by south of Port Headland in Western Australia, has the Sentient Group, is reportedly in the early announced a resource of 25.2 Mt grading 1.23% stages of commercial production with a reputed

Li2O. Current planning is to produce 150,000 tpa target of 15,000 tonnes per annum carbonate. of 6% spodumene concentrate (Altura Mining Orocobre, at the Salar de Olaroz and Lithium Ltd, 2012). Americas Corp at the Salar de Cauchari (adjacent clastic salares) have somewhat similar deposits. Orocobre’s resources are estimated to contain Continental brines 1.2 million tonnes Li and 19.0 million tonnes of A number of new projects are proposed in the K grading 690 mg/lt Li and 5730 mg/lt K Andes. In Chile, a joint venture between Minera (Orocobre, 2011). Total resources, including Lithium 251 inferred, at Cauchari are 1.52 million tonnes Li zons with a total salt thickness of 170 metres. It at a grade of 627 mg/lt Li and 13.3 million tonnes is claimed that the salt has high porosity K at 5417 mg/lt K (Lithium Americas, 2011). As throughout and that the lithium grades persist at with all the Argentinian clastic salares the depth. Comibol, the project manager, claims that magnesium/lithium ratio in the brine is low, this would indicate a resource of lithium in facilitating lithium recovery. Lithium Americas excess of 100 million tonnes. plans initial carbonate production at 20,000 Finally, as far as brine developments are tonnes per annum and doubling the tonnage concerned, Zong Chuan’s DXC project in Western within a few years. Orocobre plans initial pro- China is expected to produce 5000 tonnes per duction at 16,400 tonnes per annum with potash annum carbonate within a few years. production starting two years after start up. Construction commenced in late November 2012 (Orocobre, 2013). Geothermal brine The Salar de Uyuni, in Bolivia, has attracted The brine in the Salton Sea Known Geothermal considerable attention in the world’s press and Resource Area (KGRA) in southern California has is the world’s most politicised potential source a uniquely high concentration of lithium and other of lithium (Mares, 2010 and Wright, 2010). potentially recoverable elements. Correspondence A strategic plan published by the government in dated February 2008 states that one existing October 2010 included a commitment of US$900 228-megawatt facility with a lithium concentration million to develop a project producing 30,000 of 200 ppm Li pumps 84,000 tonnes per annum tonnes per annum carbonate by 2014 with sales of LCEs (W. Bourcier, personal communication). from a US$60 million pilot plant commencing in Simbol Materials, owned principally by private 2011. The overall sum included the cost of a lith- investors, plans initial production of 16,000 tpa ium-ion battery plant (Achtenberg, 2010). Various carbonate plus other battery materials based on countries have been courting the government for zinc and manganese with feed from a 50-megawatt rights to participate in the development of the power plant. Longer-term plans are to develop a salar but, despite numerous memoranda of under- total of four similar facilities. standing, nothing concrete has emerged except that in July 2011 the government and Korean Oilfield brine interests agreed to a joint venture to produce rechargeable battery parts. A later statement Collins (1976) estimated a possible resource of referred to the joint development of the country’s 0.75 million tonnes Li in one tenth of the area lithium resources and Posco (Korea) is to be underlain by the Smackover Formation which allowed to undertake testwork on a potential extends through North Dakota, Wyoming, new process for lithium recovery at Uyuni. In Oklahoma, east Texas and Arkansas. Currently, January 2013 Bolivia commenced sales with a the brine is exploited for its bromine content capacity of 40 tpm of lithium carbonate and 1000 and one company, Albemarle Corporation, has tpm of potash. announced its intention to recover lithium The latest published estimate of resources from its brine feed at Magnolia, Arkansas. at Uyuni is for 8.9 million tonnes grading 0.045% Laboratory testing is complete and a pilot plant Li (Risacher, 1989) which includes an area of was under construction in mid-2011. The avail- 240 km2 bounded by the 1000 mg/lt isoconcentra- able brine feed ranges from 200 to 300 ppm Li tion line containing 430,000 tonnes Li at a grade and the current volume will be sufficient to of 0.15% Li. However, the aquifer is very thin and produce 20,000 tonnes per annum carbonate. the area oods seasonally. Two deep holes drilled The company says its production will be com- nearer the centre of the salar have indicated a salt petitive in terms of price and quality with the and impermeable clay sequence with 12 salt hori- South American producers. Production of up to 252 keith evans

3000 tonnes per annum could start within 12 of 13,000 tonnes per annum carbonate com- months and reach capacity by 2015. mencing in 2015, increasing to 26,000 tonnes per annum in 2019 from a proved and probable reserve of 27.1 million tonnes grading 0.395% Hectorite Li and 3.89% K in the southernmost lens Western Lithium USA is developing the Kings (Western Lithium, 2011). In 2012, probably in Valley project located on the Nevada–Oregon recognition of potential oversupply in the short border. The deposit comprises five hectorite- term and the project’s relatively high costs, the rich clay lenses at shallow depth within sedi- company delayed its development plans for mentary and volcanic rocks in the moat of a lithium production and is evaluating the pro- caldera (Figure 10.10). When first drilled by duction of low-grade lithium clays as a compo- Chevron the resource was estimated to contain nent in drilling muds. 2.0 million tonnes of lithium grading between In Sonora, Mexico, Bacanora Minerals is 0.31% and 0.37% Li. In a revised pre-feasibility evaluating its La Ventana lithium deposit with study the company plans the initial production an estimated 60 million tonnes of hectorite

Tba Tba Basalt and andesite flows with local sediments and Tql agglomerates Tba Th Th Lake sediments Tt

Oregon Tql Quartz latite flows Nevada Tql Undifferentiated tuffs Tt Tt Th Tql and ash flows

Mg Mesozoic granodiorite

Stage 5 lens Caldera margin Stage 4 lens

Mg Stage 3 lens Tt Caldera ring fractures

Stage 2 lens Tt Tql Tt Lithium lenses

N Stage 1 lens Claim blocks

024 miles

Figure 10.10 Geological map of Kings Valley showing the five hectorite-rich clay lenses. (Modified from Chmelauskas, 2010.) Lithium 253 clay grading approximately 3000 ppm Li – per annum, the production would approximate approximately 180,000 tonnes Li. to 27,000 tonnes per annum lithium carbonate and 133,000 tonnes per annum boric acid (Kellie, 2009). Jadarite Rio Tinto owns the Jadar Valley project located 100 km south-west of Belgrade in Serbia. An World trade exploration programme, which started in 1998, aimed at finding colemanite penetrated three Data for the trade in lithium-containing min- vertically stacked zones containing a newly erals are difficult to obtain because under discovered mineral subsequently named most trade-code systems these minerals are jadarite. The lowest and thickest zone was grouped with other commodities. However, 2 data for trade in lithium oxides and hydrox- estimated to contain, over an area of 4.5 km , ides can be obtained, as can data for lithium 114.6 Mt grading 1.8% Li2O and 13.1% B2O3. Test work indicates that, after mining and carbonate. The major importing and exporting processing losses, this horizon at a depth bet- countries for these are shown on Figures 10.11 ween 300 and 600 metres and with a thickness and 10.12. of 9 to 20 metres will yield 6.4 million tonnes Talison Minerals in Western Australia cur- of boric acid and 1.4 million tonnes of lithium rently dominates the lithium minerals concentrate market with Asia accounting for 50 carbonate. Jardarite itself contains 7.3% Li2O. Work on the project is continuing but, per cent of demand with Europe being the second assuming ore production of one million tonnes largest market and North America using less

25

20

15

10 Thousand tonnesThousand 5

0

USA India Japan Spain Belgium Canada France Germany Hong Kong) Rep. of Korea United KingdomOther Countries

China (inc

Oxides & Hydroxides Carbonates

Figure 10.11 The main importing countries for lithium carbonate, oxides and hydroxides, 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) 254 keith evans

25

20

15

10 Thousand tonnes 5

0

USA Chile Japan Spain Belgium Argentina Germany Switzerland Netherlands United Kingdom Other Countries

China (inc Hong Kong)

Oxides & Hydroxides Carbonates

Figure 10.12 The main exporting countries for lithium carbonate, oxides and hydroxides, 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.)

than 10 per cent. Although Talison does not lithium in the production of synthetic rubber and disclose the breakdown of its various products it bromide in absorption chillers. is possible that about 80,000 tonnes per annum of Both Japan and South Korea, the other leading its concentrates have a low iron content, allow- battery producers, lack domestic lithium sources. ing its use in glasses and ceramics. The bulk of Historically, Japan has evaluated seawater as a production (with a higher iron content) is des- potential source and South Korea is doing so tined for conversion to lithium chemicals in currently. Japan, South Korea and China are China. Talison has completed a major expansion participating in many overseas advanced and to 740,000 tpa of concentrate and most of this is exploration projects. destined for chemical conversion. Galaxy’s ship- Chile and Argentina currently dominate pri- ments to its own chemical plant in China and mary chemical production from brine sources but concentrates produced by Reed Resources and present demand is significantly below capacity. Altura Mining will add significantly to Australian Two of the leading producers produce their down- exports. stream products in facilities outside South China, using imported spodumene together America – in the United States and Germany in with much lower quantities of domestic spodu- particular. mene and domestic brine, is the leading producer of lithium-containing glasses and ceramics. It is also either the number one or two in lithium Prices batteries, number one in grease production, the leading user of carbonate in aluminium produc- Talison reported an average sales price for all tion, the leading consumer of spodumene in grades of spodumene concentrates for the last cast-steel production, the leading user of butyl- three months of 2012 of US$367 per tonne. Lithium 255

Elsewhere, price indications are published by ally recovered and the same company cut prices Industrial Minerals (IM) magazine. in response to the major reduction in chemical In early February 2013, prices for the highest demand in 2009 (70,000 tonnes) down from grade spodumene, with a minimum of 7.5% Li2O, 91,500 tonnes in 2008. Again, prices are ranged between US$720–770 per tonne CIF1 Asia recovering. and between US$750–800 per tonne CIF Europe. IM also indicates US$300–400 per tonne for 5%

Li2O spodumene concentrates CIF Asia and Outlook US$440–490 per tonne CIF Europe. Finally, IM also shows a price of US$165–260 per tonne for Estimating future lithium demand is complicated 2 4.2% Li2O petalite concentrates, FOB Durban by the extreme difficulty in estimating future (Industrial Minerals, 2013). battery demand. Many estimates exist and in the Prices for lithium carbonate for large contracts case of vehicle demand, most estimators present delivered to the USA are indicated by IM to be in ranges covering both ‘ultra-green’ and conserva- the range US$2.5–3.0 per pound. Lithium tive scenarios. Difficulties stem from timing, the hydroxide prices are quoted by IM as US$6.5–7.5 rate of market penetration of battery-powered per kilogram for 56.5–57.5% LiOH delivered vehicles, vehicle types (whether hybrids (HEVs), under large contracts to Europe or the USA and plug-in hybrids (PHEVs) or pure electric (EVs)) US$6.0–6.6 per kilogram for Chinese lithium and battery chemistries. hydroxide of the same grade delivered under large Chemical demand in 2010 totalled between contracts to Europe (Industrial Minerals, 2013). 102,000 and 105,000 tonnes of LCEs. Table 10.4, A leading producer in China, in mid-2011, based principally on an SQM presentation in quoted prices in China, all in tonnes as follows August 2010 concerning possible changes to the – technical grade carbonate US$4700, battery- grade carbonate US$5300, hydroxide monohy- Table 10.4 Lithium chemical production capacities in drate US$5760, chloride US$4850 and metal at 2010. (Data from Solminihac, 2010.) US$52,000. A further indication of prices is in SQM’s Company Country Source Tonnes LCE Annual Report for 2011 in which the company Continental reports sales of 40,700 tonnes (a mixture of car- Brines bonate, hydroxide and possibly chloride) for reve- SQM Chile Atacama 40,000 nues of US$183.4 million – an average of US$4506 Chemetall Chile Atacama 38,000 per tonne. Nevada Silver Peak Comments have been made by various FMC Argentina Hombre Muerto 17,500 observers regarding the ‘dramatic’ rise in lithium CITIC China Taijinaer L 5000 prices. The facts do not support this. Looking QLL China Taijinaer L 2000 back in time to when Chilean brine-based car- Tibet China Zhabuye L 2500 bonate entered the market in 1985 it was priced Pegmatites at US$1.45 lb (US$3200/tonne). United States ABA China Maerkang 2500 in ation, as reported by the Bureau of Labor Jianxi China Ningdu 2000 Statistics, from that time to the end of 2010 is Minfeng China Maerkang 2000 calculated to be 71 per cent. Applying this to the Ni & Co China Maerkang 5000 XLP China Kekeluhai /Talison 5500 1985 price would give a current price of US$5500 Panasia China Talison 4000 tonne, which is in the same range as the figures Tianqi China Talison 9500 quoted above. CBL Brazil Cachoeira 2300 A major reduction in price occurred when Total 137,800 SQM entered the market, but they have gradu- 256 keith evans

Table 10.5 Expanded production and new projects. Mining Law in Chile, estimates chemical produc- (Compiled by the author from published sources.) tion capacity in 2010 at 138,000 tonnes of LCEs indicating significant excess capacity of approxi- Tonnes LCE mately 33,000 tonnes. Continental Brines Table 10.5 lists expansions and targeted pro- FMC, Hombre Muerto, Argentina 5500 Expansion duction levels announced by would-be producers Chemetall, Atacama (Chile) & 12,000–27,000 Expansion announced through January 2013. Some of the Silver Peak (Nevada) expansions have been completed and some are SQM, Atacama, Chile 20,000 Expansion under construction. All aim to be in production Chinese brine expansion ? prior to 2020. If all were to proceed, total Zong Chuan, China 5000 chemical capacity would increase to between Orocobre, Olaroz, Argentina 16,400 Sentient, Rincon, Argentina 15,000 593,000 and 643,000 tpa of lithium carbonate Lithium Americas, Cauchari, 20,000–40,000 equivalents. Argentina Demand estimates for 2020 vary. Evans (2012) Lithium One, Sal de Vida, Argentina 25,000 listed those of current producers and other Rodinia, Diablillos, Argentina 15,000–25,000 responsible organisations at the fourth Lithium Comibol, Uyuni, Bolivia 20,000 Supply & Markets Conference in Buenos Aires. Pegmatites They included those from FMC (dated 2011) of Talison, Greenbushes, Australia 83,000* 278,000 tonnes, General Motors (dated 2010) of Galaxy, Mount Catlin, Australia/ 21,000 215,000 to 243,000, SQM (dated 2011) of 200,000 China to 270,000, Roskill Information Services (dated Reed Resources, Mt. Marion, 19,000** 2011) of 215,500 to 264,500 and Rockwood Australia quoting an average of six estimates (dated 2011) Altura, Pilgangoora, Australia 19,000** of 220,000 to 320,000 tonnes. Canada Lithium, Quebec, Canada 20,000 As can be seen from Figure 10.13, demand is New & Expansions, China 11,000 Keliber Oy, Finland 6500 estimated to increase rapidly after 2020 with both Nemaska, Quebec, Canada 30,000** high- and low-demand scenarios (Chemetall, Lithium One/Galaxy, Quebec, ? 2010). Other estimates exist but this one is used Canada for illustration as it is based on six market studies. Geothermal Brine Using this estimate, demand in 2030 should be Simbol, Salton Sea, USA 16,000 between 400,000 and 750,000 tonnes per annum. Oilfield Brine The lower tonnage can be easily met by current Albermarle, Arkansas, USA 20,000 and planned production, but to meet the higher scenario will require further expansions or addi- Hectorite tional sources. Western Lithium, Kings Valley, USA 13,000–27,700 Bacanora, La Ventana, Mexico 35,000 Finally, it should be said that there is no short- age of sceptics on the issue of vehicle electrifica- Jadarite tion with criticisms of costs, the lack of range and Rio Tinto, Jadar, Serbia 27,000*** slow charging with many claiming that electric TOTAL (excluding Lithium One/ 455,400–515,100 vehicles will only occupy a niche market. Most Galaxy and major Chinese brine motor manufacturers do not appear to share these expansions) opinions. *Assumes 90% of estimate production is converted to carbonate Current sources, when the huge upside poten- at 85% recovery tial at the Salar de Atacama is taken into **Assumes concentrate production is converted to carbonate at consideration, appear to be adequate to meet 85% recovery demand for very many decades. This though, has ***Not announced. Potential estimate if mined at 1.0 million tpa not reduced the enthusiasm of the Koreans in Lithium 257

800

700

600

500

400

300

200

100 Lithium carbonate equivalent (thousand tonnes)

0 2010 2015 2020 2025 2030

Figure 10.13 Forecast of the demand for HEV+PHEV+EV HEV+PHEV+EV Portable lithium to 2030, with two different (high scenario) (low scenario) batteries scenarios relating to the uptake of electric vehicles. (Courtesy of Rockwood Other Lubricants Glass and ceramics Lithium.) particular and, to a lesser extent the Chinese and although large battery costs are projected to Japanese, from acquiring interests in developing halve as volume increases, the initial cost of a projects. These interests range from sales agency battery-powered vehicle is still high in agreements for percentages of production to comparison with one powered by an internal financing feasibility studies, to providing access combustion engine and government subsidies to low-cost loans and to acquiring significant are necessary to promote sales. The biggest equity interests. Each of these countries, together threat to vehicle electrification appears to be the with a French vehicle manufacturer, have successful development of biofuels, hydrogen heavily courted the Bolivian Government for a and possibly natural gas. right to participate in the development of the The requirement for the production of fusion Salar de Uyuni. energy is not included in any demand projections The new projects all claim, by including co- although potential demand from this application product credits in some cases, to have produc- was the main reason for establishing a National tion costs that are competitive at current prices. Research Council committee in the United States With potential over-production until at least in the mid-1970s. Any generating system will 2020, there seems little likelihood of significant almost certainly employ a deuterium-tritium price increases over and above in ationary (DT) reaction, with the tritium obtained from costs. lithium’s 6Li isotope. All demand estimates have large batteries Two major research projects are underway: accounting for a high percentage of the total and, the International Thermonuclear Experimental 258 keith evans

Reactor (ITER), located in the south of France; and Bachu, S., Brulotte, M. and Yuan, L.P. (1955) Resource the National Ignition Facility (NIF) at Livermore, estimates of industrial minerals in Alberta formation California. The former aims to generate fusion waters. Alberta Research Council. reactions in a doughnut-shaped vessel call a British Geological Survey (2012) World Mineral tokomak and the other proposes igniting DT pel- Production, 2006–10. (Keyworth, Nottingham, British Geological Survey.) lets by means of 192 lasers. Lithium requirements Canada Lithium Corp. (2011a) Feasibility Study Update will be determined by the ultimate design(s) and NI 43–101 Technical Report, Quebec Lithium chosen for reactors and commercialisation still Project, La Corne Township, Quebec. appears to be decades in the future. Canada Lithium Corp. (2011b) Canada Lithium increases measured and indicated mineral resource to 33.2 million tonnes grading 1.19%. http://www. Acknowledgements canadalithium.com/s/Home.asp or www.sedar.com Chmelauskas, J. (2010) Kings Valley Lithium Project, The author would like to thank Chemetall for Shareholder Presentation. provision of the section on recycling in this Clarke, G.M. and Harben, P.W. (2009) Lithium Availability chapter. The author would also like to thank the Wall Map – Lithium Availability a Question of Demand. AABC, Long Beach, California, USA. following for the provision of photographs and Clarke, G.M. (2011) Lithium Availability Wall Map assistance with figures: John Houston, Talison created and published as a part of “Lithium avail- Lithium Ltd, Sociedad Quimica y Minera de ability: prospective demand being answered.” Chile S.A. (SQM), Roskill Information Services, Advanced Automotive Battery Conference, Mainz, Nemaska Lithium Inc., Western Lithium Corpo- 2011. Copies available on application to: gmclarke1@ ration and Rockwood Lithium. ntlworld.com Collins, A.G. (1976) Lithium Abundance in Oilfield Waters. United States Geological Survey Professional Notes Paper 1005. de Silva, S.L., Zandt, G., Trumball, R., Viramonte. 1. CIF, Cost, Insurance, Freight. The seller’s price J.G. Sales. G. and Jiminez, N. (2006) Large ignimbrite includes the cost of the goods, the insurance of eruptions and volcano-tectonia depressions in the the goods to their destination port, and the Central Andes: a thermomechanical perspective. In: Troise, C., De Natale, G. and Kiburn, C.R.J. (eds.) cost of freight. Mechanisms of Activity and Unrest at Large Calderas. 2. FOB, Free on Board. The seller is responsible Geologist Society, London. Special Publication No. for the cost of delivering goods to the ship. The 269, 47–63. buyer is responsible for transportation and Duyvesterin, W.P.C. (1972) Recovery of Base Metals insurance costs from that point. from Geothermal Brines. Geothermics Journal of Geothermal Research and its Applications 21 (5/6). Evans, R.K. (1986) Further Developments at the Salar de References Atacama. Industrial Minerals, International Congress. Achtenberg, E. (2010) Bolivia bets on state-run lithium Evans, R. K. (1978) Lithium Reserves and Resources industry. North American Congress on Latin Energy 3 (3). America. https://nacla.org/node/6799 Evans, R. K. (2008) Know Limits. (An Abundance of Altura Mining Ltd. (2012) Positive results from Lithium) Industrial Minerals. Pilgangoora lithium project scoping study. ASX Evans, R.K. (2012) An overabundance of lithium? announcement, 19 November 2012. http://alturamin Lithium supply and markets conference, Buenos ing.com/files/announcements/2012/2012%2011%20 Aires, January 2012. 19%20-%20Pilgangoora%20Lithium%20Project%20 Galaxy Resources Ltd. (2011) Galaxy extends Mt Cattlin Scoping%20Study%20Positive%20Results.pdf mine life after resource upgrade. ASX announcement / Lithium 259

Media release, 22 March 2011. http://www.galaxyre carbonate plant. Press release 2 October 2012. http:// sources.com.au/investor_asx.shtml nemaskalithium.mwnewsroom.com/press-releases/ Galaxy Resources Ltd. (2013a) Sal de Vida, Argentina. nemaska-lithium-announces-positive-preliminary- http://www.galaxyresources.com.au/projects_sal_ eco-tsx-venture-nmx-201210020823456001 de_vida.shtml Orocobre Ltd. (2011) Orocobre, the next low cost Galaxy Resources Ltd. (2013b) James Bay, Quebec, lithium producer. Investor Presentation 14 December Canada. http://www.galaxyresources.com.au/projects_ 2011. http://www.orocobre.com.au/PDF/16Dec2011_ james_bay2.shtml Investor%20Presentation.pdf Harben, P.W. (2002) Lithium Minerals and Chemicals. Orocobre Ltd. (2013). Salar de Olaroz. http://www.oro- In: The Industrial Minerals HandyBook, 4th Edition. cobre.com.au/Projects_Olaroz.htm Metal Bulletin plc, London, 184–192. Reed Resources Ltd. (2012) Shareholder agreement and Houston, J., Butcher, A., Ehrens, P., Evans, K. and PFS results Mt Marion lithium. 17 October 2012. Godfrey, L. (2011) The evaluation of brine prospects http://www.reedresources.com/announce-blog.php? and the requirement for modifications to filing id=656andprojpg=mtmar standards. Economic Geology, 106, 1225–1239. Risacher, F. (1989) Economic Study of the Uyuni Salar Houston, J. and Gunn, M. (2011) Technical Report for UMSA-ORSTOM-CIRESU PACT. the Salar de Olaroz lithium-potash project, prepared Risacher, F., Alonso, H. and Salazar, C. (2003) The Origin for Orocobre Ltd. www.sedar.com of Brines and Salts in Chilean Salars-a Hydrochemical Ide, Y.F. and Kunasz, I.A. (1989) Origin of Lithium in review. Earth Science Reviews 63, 249–293. the Salar de Atacama, Chile. In: Ericken, G.E., Cañas Rodinia Lithium Inc. (2011) Rodinia Lithium files Pinochet, M.T. and Reinemund, J.A. (eds.) Geology favourable preliminary economic assessment of the Andes and its relation to hydrocarbon and technical report. http://www.rodinialithium.com/ and mineral resources. Circum-Pacific Council for www.sedar.com Energy and Mineral Resources. Earth Science Series Solminihac, P. (2009) SQM Lithium Resources and 11, 165–172. View of the Lithium Industry. Presentation to Industrial Minerals. (2013) Industrial Minerals Prices. Lithium Supply Markets conference, Santiago, Chile, http://www.indmin.com/Prices/Prices.aspx January 2009. Kellie, R. (2009) Jadar Lithium Project, Serbia. Proceedings Solminihac, P. (2010) Resources de Lithio en el Mundo of Lithium Supply and Markets Conference, Santiago, y Chile. Presentation by SQM concerning possible Chile. 28 January 2009. Chief Geologist, Rio Tinto changes in the Mining Law in Chile. Exploration. Talison Lithium Ltd. (2011a) Annual Information Form Kunasz, I.A. and Pavlovic, P.Z. (1983) The Salar de for year ended 30 June 2011. http://www.talisonlith Atacama - a new centre for lithium production. In: ium.com/home and www.sedar.com Cope, B.M. and Clarke, G.M. (eds.) Proceedings of the Talison Lithium Ltd. (2011b) Talison Lithium commences 5th Industrial Mineral International Congress. Metal expansion to double capacity at the Greenbushes Bulletin plc, London lithium operations. News Release. Kunasz, I. (2007) Lithium Resources. In: Kogel, J. (ed.) UN Comtrade (2013) United Nations Commodity Trade Industrial Minerals and Rocks 7th Edition, Society Statistics Database, Department of Economic and for Mining, Metallurgy and Exploration Inc. (SME), Social Affairs/ Statistics Division, http://comtrade. Englewood, Colorado, USA, 599–613. un.org/db/ Lithium Americas. (2011) NI 43–101 Technical Report, U.S. Geological Survey (2009). Mineral commodity Preliminary Assessment and Economic Evaluation of summaries 2009, Lithium. http://minerals.usgs. the Cauchari-Olaroz Lithium Project. In: WorleyParsons, gov/minerals/pubs/commodity/lithium/mcs-2009- A.R.A (ed.) www.sedar.com lithi.pdf Mares, D.R. (2010) Lithium in Bolivia – can resource U.S. Geological Survey (2010). Mineral commodity nationalism deliver for Bolivians and the world? In: summaries 2010, Lithium. http://minerals.usgs.gov/ James A. (ed.) Baker III Institute for Public Policy. minerals/pubs/commodity/lithium/mcs-2010-lithi. Nemaska Lithium Inc. (2012) Nemaska Lithium pdf announces positive preliminary economic assessment U.S. Geological Survey (2011). Mineral commodity (PEA) for Whabouchi mine and lithium hydroxide/ summaries 2011, Lithium. http://minerals.usgs.gov/ 260 keith evans

minerals/pubs/commodity/lithium/mcs-2011-lithi. supports low cost lithium production in Nevada. pdf. News Release. Verley, C.G. and Vidal, M.F. (2013) Preliminary Wright, L. (2010) Lithium Dreams. The New Yorker. Economic Assessment for the La Ventana lithium Yaksic, A. and Tilton, J. (2009) Using the Cumulative deposit, Sonora, Mexico for Bacanora Minerals. 24 Availability Curve to Assess the Threat of Mineral January 2013. http://www.bacanoraminerals.com/ Depletion: The Case for Lithium Policy 34. reports/pdf/pea_ventana.pdf Zheng, M. (1989) Saline Lakes on the Quinghai-Xizang Western Lithium Corp. (2011) Western Lithium Pre- (Tibet) Plateau. Beijing Scientific and Technical feasibility Study indicates $550 million NPV and Publishing House, Beijing. 11. Magnesium

NEALE R. NEELAMEGGHAM1 AND BOB BROWN2

1 ‘Guru’, Ind LLC, 9859 Dream Circle, South Jordan, Utah, USA 2 Publisher, Magnesium Monthly Review, Prattville, Alabama, USA

Introduction development of two main-stream production tech- nologies, unlike many other metals which use Magnesium has a variety of special properties minor variations of a single process technology. which make it useful in broad range of applications. Concern over the security of supply of On account of its high strength and light weight it magnesium has arisen since the late 1990s, not is the most commonly used structural metal after due to constraints on the availability of raw steel and aluminium. The fact that the density of materials or technology, but rather due to magnesium (1738 kg/m3, Table 11.1) is only 63 per changes in the pattern of the global supply cent of that of aluminium and 23 per cent of that of chain. Over the past 20 years or so the produc- steel has promoted its use in many transportation tion of magnesium metal has increased applications where whole life-cycle energy consid- in China, whilst production in many other erations are important and may be a critical factor countries has declined or ceased. During the for energy conservation in the 21st century. About twentieth century western world production half of the magnesium used today is in the produc- increased to about 400,000 tonnes per year, tion of alloys, most commonly with aluminium. while Chinese production was reported to be These have important applications in the food about 3000 tonnes per year in 1991. However, industry (for example, in drinks cans) and in the by 2008, the primary magnesium metal produc- aerospace and military sectors. Magnesium metal is tion capacity of the western world, including also unique in that it can be used as the reducing Russia, had decreased to about 120,000 tonnes agent in the production of several special metals, per year while in China capacity increased to such as beryllium, zirconium and uranium, which more than 600,000 tonnes per year. are strategically important for defence and nuclear reactor applications, and titanium, which is used in the chemical and aerospace industries. Physical and chemical properties Magnesium is the only structural metal that can be obtained economically from both the lithosphere Magnesium (chemical symbol Mg) has an and hydrosphere. Magnesium can be produced from atomic number 12, an atomic weight of 24.32 oxide resources or from chloride raw materials orig- and it exhibits a valency of two. It occurs in inating from saline water. This has resulted in the three stable isotopes, 24Mg, 25Mg and 26Mg, of

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 262 neale r. neelameggham and bob brown

Table 11.1 Selected properties of magnesium. Distribution and abundance in the Earth

Property Value Units Magnesium is the fifth most abundant element in the hydrosphere and the most abundant Symbol Mg Atomic number 12 structural metal ion in the ocean. It is therefore Atomic weight 24.31 unique among the structural metals in that it Density solid Mg at 25°C 1738 kg/m3 can be extracted from either the hydrosphere or Density liquid Mg at 700°C 1584 kg/m3 the lithosphere. In contrast, aluminium is only Melting point 650 °C sparingly soluble in the oceans, and is extracted Boiling point 1090 °C from lithospheric resources only. In the upper Hardness (Mohs scale) 2.5 crust magnesium is the eighth most abundant Specific heat capacity at 25°C 1.05 J/(g °C) element (2.5% MgO by weight (Rudnick and Latent heat of fusion 368 kJ/kg Gao, 2004)). It is an important constituent of Latent heat of vaporisation 5251 ± 251 kJ/kg major rock-forming silicate minerals, such as Thermal conductivity 156 W/(m °C) Electrical resistivity at 25°C 44.8 nΩ m pyroxene and olivine. Consequently, rocks rich Young’s modulus (at 20°C, 44.7 GPa in these minerals, such as gabbro and basalt, 99.8% pure Mg) have higher magnesium contents, commonly exceeding 5% MgO. Irrespective of the source of the raw material, which the former accounts for about 79 per an energy input is required to effect the conversion cent of the total. It makes strongly ionic of the mineral into metal. The nature and cost of compounds with common halide, sulfate or the energy source is a particularly important phosphate anions as well as making weakly consideration in the choice of the process route ionic but covalent compounds of magnesium for the production of primary magnesium. oxide and nitride. Magnesium, in its structural applications, is typically used as an alloy casting or as a Mineralogy wrought alloy. The mechanical properties of these materials can vary quite widely, with The solid mineral phases from which magnesium tensile strength in the range 150 to 400 mega- is extracted from the lithosphere are: dolomite pascals (MPa). The cast alloys have lower (CaCO3.MgCO3), magnesite (MgCO3), periclase tensile strength and elasticity than wrought (magnesium oxide, MgO), hydromagnesite alloys or shapes made by powder metallurgy (3MgCO3.Mg(OH)2.3H2O), brucite (MgO.H2O), using nanotechnology. However, the properties and various silicates of magnesium, such as of die cast magnesium shapes are adequate for olivine ((Mg,Fe)2SiO4), serpentine (3MgO. 2SiO2. many automotive components and for portable 2+ 2H2O) and biotite (K(Mg,Fe )3AlSi3O10(OH,F)2). electronic devices where ‘light-weighting’ is The solid mineral phases of hydrospheric important. origin, found in evaporite deposits, from which Magnesium can be used as a reducing agent magnesium is extracted are: magnesium sulfate because of its high position in the electrochem- (epsomite, MgSO4.7H2O), kieserite (MgSO4.H2O), ical series. It is one of the most electronegative langbeinite (K2SO4.2MgSO4), kainite (KCl. elements, only being surpassed by the alkali MgSO4.3H2O) and carnallite (KCl. MgCl2.6H2O) metals, higher alkaline earth metals and sev- and bischofite (MgCl2.6H2O). Many of these were eral rare earth metals. It forms more stable initially identified in Stassfurt, Germany in the compounds (oxides, sulfates, carbonates, mid-nineteenth century. It should be noted that nitrates, halides, etc.) than most other common magnesium is the second most abundant cation metals. in solution in the oceans and terminal lakes. Magnesium 263

Sodium is present in larger amounts usually the 1980s or the emergence of the global economy providing the ionic balance for the chloride ions, since the 1990s, and through the recent com- with the ionic balance for magnesium provided modity rise and fall between 2006 and 2009. In by sulfate and chloride ions. 1935, John A. Gann, Chief Metallurgist of The Dow Chemical Company, noted the following – “.. our light metals, occur only in the form of com- Deposit types pounds so stable that their discovery, isolation, commercial production, and use were forced to Of the potential lithospheric sources dolomite await some of the modern advances in chemistry (calcium–magnesium carbonate) is the most and engineering. Under such conditions, the evo- commonly used. Thus, calcined dolomite, MgO. lution of a new industry is often a romance in CaO, with a minimum magnesium content of 8 which scientific and industrial difficulties and per cent, is one of the main sources of magnesium. near failures adds to the thrill of success.” (Gann, Alternative sources are magnesite and magnesium 1935). This statement has been verified time and silicates, which have been used as raw material again in the development of magnesium produc- for magnesium production during the past two tion processes and also in the development of new decades. The magnesium silicate can be in the alloys and uses of magnesium. form of olivine or from the tailings of asbestos There are two main routes for making extraction where it is present as serpentine magnesium metal: minerals. 1. The electrolytic method – magnesium chlo- In saline lake water or brines magnesium is ride is the feed material to this process. The present at lower concentrations, typically less magnesium chloride is derived from magnesium than 0.7 per cent by weight, while in sea water oxide sources (by chlorination of magnesia or the magnesium content is 0.13 per cent. Two magnesite) or from chloride sources (by dehydra- major undesirable components, water and sodium tion of brines or hydrous carnallite) (Figure 11.1). chloride, must be removed from these resources 2. Metallothermic reduction – the dolomite feed before attempting to extract magnesium. Other material is calcined and then reduced, either by major contaminants are potassium and sulfate. silicon or one of its alloys, or by other metals such as aluminium. One of the most important processes of this type, which uses ferrosilicon as Extraction methods, processing the reductant, is known as the Pidgeon process and beneficiation (Figure 11.2). All magnesium metal production utilises the A historical review is given to provide back- following general steps, although the specific ground to the development of production physical and chemical features of the source raw processes that are currently in use. Brown (2000, materials lead to considerable variation in the 2003a) provides an in-depth history of magnesium, process technology employed: including a review of production methods, up ● Raw material upgrading: in the electrolytic pro- to 2003. Neelameggham (2013) provides detailed cess this involves removal of sulfates, and other information on the fundamental chemistry impurities from the magnesium chloride. For the and production technology needed to produce metallothermic process the proportion of dolo- magnesium from different raw materials. mite in the feed should be as high as possible The processes used in the production of with the least amount of non-reacting compo- magnesium have gone through several major nents, such as silicates. changes which almost mirror changes to the global ● Removal of impurities: for example, in the supply chain over the same period – whether it be magnesium chloride feed preparation all non- during the two World Wars, the Cold War through chloride anions (such as sulfates, borates, etc.) are 264 neale r. neelameggham and bob brown

Solar evaporation

Impure magnesium chloride brine

Removal of impurities, Calcium chloride mostly sulphate & boron

Concentration

Spray drier

Magnesium chloride with water Hydrochloric & oxychlorides acid

Carbo- Carbon chlorination

Final purification

Molten magnesium chloride (MgCl2)

Chlorine Electrolysis

Molten magnesium

Casting

Cold magnesium metal

Figure 11.1 Schematic owsheet for U.S. Magnesium’s electrolytic process in Utah. (Modified from Holywell, 2005.) Magnesium 265

Ferrosilicon Dolomite (10 cm lumps)

Crushing and sizing Crushing and milling Calcination in kilns –65 mesh powder Calcined dolomite

Milling

Mixing

Briquetting in press

Briquettes

Reduction in retorts

Sleeves with crowns

Crown remover

Crowns

Melting and refining

Pure magnesium metal

Figure 11.2 Schematic owsheet for the production of magnesium by the Pidgeon process. removed and also most of the water by preparing nents which do not contribute to the process. either close to a monohydrate of magnesium This is accomplished by using high-grade dolo- chloride (MgCl2.1.3H2O) or essentially anhydrous mite as the raw material feed, thereby reducing magnesium chloride, depending on the electro- the proportion of other components which do not lytic process chosen. In the case of the calcined take part in the reaction and reduce retort pro- dolomite reduction process, the contents of duction capacity. calcium oxide and magnesium oxide are maxi- ● Removal of impurities undesirable in the mised by reducing the content of other compo- finished metal: for example, one of the most 266 neale r. neelameggham and bob brown undesirable impurities in the finished metal is small laboratory cell using molten anhydrous nickel, which should be kept below 0.001 per magnesium chloride feed. In 1852 Bunsen dem- cent to avoid the product magnesium from onstrated that it is easier to dehydrate magnesium becoming corrosive. In practice, impurity chloride in a potassium chloride bath. This dis- control is achieved mainly through measures covery subsequently led to the use of naturally aimed at avoiding the incorporation of occurring carnallite as a source for making impurities into the process. The unalloyed magnesium. Commercial production of magnesium magnesium normally marketed has several on a larger scale was initiated in 1886, about the grades, depending on the customer require- same time as the beginnings of Hall–Heroult cell ment, normally in the range from 99.8 per cent for aluminium. pure to 99.95 per cent. About 30 to 40 per cent Molten dehydrated carnallite (KCl.MgCl2) was of the magnesium is marketed as alloys electrolysed to magnesium metal in 1886 by the containing 91 to 97 per cent magnesium. Aluminium und Magnesium Fabrik, Germany. ● Separation from other process products: for This was further developed by Chemische Fabrik example, in the Pidgeon process a calcium sili- Griesheim–Elektron starting in 1896. Production cate residue containing iron remains in the retort of magnesium metal by electrolysis of molten after removal of the magnesium product. This carnallite continues to the present day. by-product, which is mostly calcium silicate, may be used as a substitute cement in construction either by itself or with other ingredients added. Commercial magnesium production Different by-products may be derived from the processes of the twentieth century electrolytic process. Typically, these include chlorine gas and excess mixed chloride electro- Given the variety of raw materials potentially lyte from the electrolysis. suitable for the production of magnesium metal a ● Melting, refining and casting metal and alloys: wide range of processes has been developed for the typical product is magnesium in the form of this purpose. Changing global economic condi- ingots of the pure metal, although some plants tions and demand patterns have determined supplying magnesium for chemical applications which technologies have been adopted at differ- supply magnesium as a powder. ent times. Local conditions and the specific fea- tures of the available source materials have contributed to the development of the particular Nineteenth-century magnesium process route employed in any one area. production processes In 1938 Haughton and Prytherch (1938) noted Humphrey Davy was the first person to isolate that the extraction of magnesium was carried out magnesium. In 1808 he passed a current through by three main processes: the electrolysis of fused moistened magnesium sulfate and produced chlorides; the electrolysis of the oxide in solution magnesium at a mercury cathode. He also con- of molten uorides; and the carbo-thermal pro- verted red hot magnesium oxide with potassium cess which involves direct reduction of the oxide vapours collecting the magnesium into mercury. by carbon in an arc furnace with a hydrogen In 1828, Bussy reduced magnesium chloride atmosphere followed by re-distillation in an inert with potassium metal in a glass tube and, when atmosphere. During World War II, the silico-ther- the potassium chloride was washed out, small mic reduction of oxide technology was developed globules of magnesium were present. which gave rise to the important Pidgeon process, In 1833 Faraday electrolysed impure magnesium which was named after its inventor, Lloyd chloride in a molten state to get magnesium Pidgeon (Pidgeon and Alexander, 1944). metal. However, it took two more decades before The Pidgeon process (Figure 11.1) uses a bri- Robert Bunsen made commercial quantities in a quetted charge of finely ground calcined dolomite Magnesium 267 and ferrosilicon that is placed in a horizontal which were in operation in the US between alloy retort with a diameter of ten inches. The 1930–1960, is presented in detail in the U.S. retort is externally heated to 1200 °C in a furnace Bureau of Mines Reports of Investigations (Elkins and a vacuum is applied to it. The silicon of the et al., 1968). Some of the technical aspects of ferrosilicon reduces the magnesium oxide in the these analyses are still valid 40 years later. charge and the resulting magnesium vapours ow The carbo-thermal process developed by to a condenser in the portion that extends from Hansgirg was in commercial production only the furnace. This is a cyclic process with the between 1930–1950. Efforts to revive it in an retorts being charged and discharged several economically viable form have continued to the times per day. The Chinese were introduced to present day, mostly in the laboratory, but some- the Pidgeon process in the late 1980s. Since then times on a pilot-plant scale (Hansgirg, 1932 and they have built hundreds of silico-thermic plants Brooks et al., 2006). In the 1930s, magnesium based on this process and China is now the oxide was converted to magnesium using world’s largest magnesium producer. calcium carbide as a reductant, by the Murex Between 1977 and 2002 Alcoa produced process (Beck, 1939 and Emley, 1966). Although magnesium by Magnetherm technology, a varia- it is not used commercially today, this process tion of silico-thermic reduction (Jarrett, 1981 and may be applicable whenever the calcium car- Faure and Marchal., 1964). This process was used bide required for reduction is cheaper than at Addy in Washington, USA, and in Marignac in ferrosilicon. France, where its use was discontinued in 2002. The demand for magnesium fell sharply after The Rima Group in Brazil, the only producer of World War II, giving rise to competition between magnesium metal in the southern hemisphere, the available production technologies. Carbo- uses another variation of silico-thermic reduction, thermal reduction and uoride-uxed magnesium known as the Bolzano process, which was used oxide melt electrolytic processes were no longer in Bolzano, Italy, up to 1992. In this method viable on account of the high costs associated briquettes of dolime (calcined dolomite) and fer- with specific technical issues. The silico-thermic rosilicon are processed in a larger retort which is process continued to take a backseat to the elec- internally heated. trolytic conversion of magnesium chlorides until Electrolytic magnesium technology, chiey the mid-1990s. However, broader Chinese engage- developed in Germany, has been described in ment in the global economy from the early 1990s detail by Strelets (1977). This technology was has had a major impact on the supply chain for subsequently improved in the Soviet Union by magnesium metal. The availability of very low- VAMI (the Soviet Aluminium and Magnesium cost ferrosilicon in China led to a revival of the Institute) and at the Soviet magnesium plants in silico-thermic method, making it the dominant Solikamsk, Russia, Zaporozhye, Ukraine, and process of the first decade of the twenty-first Kazakhstan (Muzhzhavlev et al., 1965 and 1977). century. Improved versions of these electrolytic methods are still in use – at the Dead Sea Magnesium Plant in Israel since 1995 and at Solikamsk in Russia. Specifications and uses Both use carnallite as the starting material, which is derived from mining evaporite deposits For structural applications, magnesium is alloyed in Russia and from solar evaporation in the with other metals to give suitable properties for Dead Sea. castings and for wrought or formed material. The An economic comparison of the carbo-ther- alloy elements used are identified by the use of mal, silico-thermal and electrolytic processes prefixes (A for aluminium, M for manganese, Z was made by the U.S. Bureau of Mines (Dean, for zinc, etc.). The tensile properties of cast 1965). Information on these three processes, magnesium are much lower than wrought 268 neale r. neelameggham and bob brown magnesium. Typical casting alloys, such as alloys as well as effective design, for example, by AM50, AM60 and AZ91, contain 5 to 9% alu- isolating joints between dissimilar metals. minium, 0.5 to 0.7% zinc, and 0.2 to 0.5% Magnesium metal has a long history of use in manganese. Typical wrought alloys contain less the automotive and aerospace sectors dating back aluminium, such as AZ31 and AM30, as well as to the 1920s. Each of the 24 million Volkswagen the M1 alloy (99.3% Mg and 0.5–0.7% Mn). Beetles manufactured had a magnesium trans- In terms of volume the largest use of mission and engine block. All the engines in magnesium has for many years been as an German aircraft during the 1930s and throughout additive to impart stiffness to aluminium World War II were joined to the fuselage by the alloys. For example, aluminium foil, which is use of forged magnesium engine bearers. made from essentially pure aluminium, can be Many new uses of magnesium and magnesium rolled into very thin, imsy sheets. However, alloys were triggered by the need to reduce the with the addition of magnesium or other weight of the automotive to improve fuel economy. alloying additives, aluminium becomes a struc- The special properties of some wrought alloys have tural metal. Adding as little as one per cent resulted in the development of specific niche appli- magnesium provides aluminium alloys with cations. For example, the AZ31 alloy has been used enough stiffness and good forming properties to make printing plates for over sixty years making making it suitable for the body of beverage use of masked chemical etching techniques. cans. The beverage can top is stiffer with three However, the complexity of the conventional per cent magnesium, and the closures are stiffer multi-step process for making wrought alloys kept still with 4.5 per cent magnesium. Magnesium the growth of the market for wrought magnesium is present in most of the aluminium alloys used low for a long time. However, the twin roll casting in structural extrusions and sheet. Since the technique, which was originally developed in the global production of aluminium alloys, from late 1970s, reappeared in a modified form in the primary and secondary sources, is about 50 mil- early 2000s, initially in Australia, later in Taiwan lion tonnes per year it is easy to see how this and recently in South Korea. This method with use consumes the largest proportion of fewer processing steps in making magnesium sheet magnesium metal production. is now facilitating more applications for wrought Design engineers select materials to per- magnesium. As a result, there is now potential for form specific functions in large volume applica- growth in this market in automobiles with a tions such as in the automotive and the 3-C magnesium frame and body. (computers – cameras – communications) indus- Interest in using magnesium alloys for mili- tries based on the following criteria, as described tary applications has recently been revived, for by Avedesian (1999): example, in plates in military armour, and (a) High strength to weight ratio to minimise in lightweight applications, including signal life-cycle energy consumption. system electronic and radio machinery (Jones (b) Good damping capacity. and DeLormo, 2008; Jones and Kondoh, 2009). (c) Suitability for high-speed production In addition, magnesium powder has found a processes. new application in the ‘magnesium ready to (d) Long life for the tooling and machinery parts eat’ (MRE) food packages used by military per- used in the fabrication process, e.g. the wear of sonnel in field situations, where the reaction steel moulds by molten magnesium is low com- between metallic magnesium and water pro- pared with molten aluminium. vides the energy needed for warming the food. (e) Electromagnetic shielding properties. A significant market has developed over the (f) Long lifetime of the finished magnesium parts as past 30 years for using magnesium powder to shown by low corrosion rates in each application. reduce the sulfur content of molten iron to the This depends on the composition of the magnesium low levels required for making high-quality steel Magnesium 269 sheet. The high electronegativity of magnesium 6% provides the potential to reduce oxides which cannot be reduced by carbon or aluminium. 11% Magnesium is also essential to the production of some strategic metals used in specialist defence 43% and nuclear technologies, such as beryllium, uranium and zirconium. Magnesium is also the preferred reductant in the production of titanium from titanium chloride. In the organic chemicals industry magnesium has been used in making Grignard reagent, an important organic synthesis intermediary, for over a century. Magnesium metal is the preferred reducing agent in the preparation of metals such as titanium, zirconium, uranium and beryllium from the respective chlorides or uorides. In one 40% such application, the Kroll Process, magnesium is used as the reducing agent in the production of titanium metal from titanium tetrachloride. The system is designed so that the magnesium chlo- Aluminium alloys Structural applications ride generated in the reduction process is recov- Desulfurisation of iron and steel Other ered and is utilised to produce magnesium, making this process essentially a ‘closed-loop’ system in respect to magnesium. Global produc- Figure 11.3 The main sectors of magnesium use in the USA in 2011. (Data from United States Geological tion of titanium metal is estimated to be about Survey, 2012.) 120,000 tonnes per year of which more than 90 per cent is made by the Kroll process. Since one tonne of magnesium is required to produce one tonne of titanium in the Kroll process it is esti- permanent mould casting, sand castings, wrought mated that about 120,000 tonnes per year of products, cathodic protection, desulfurisation of magnesium is produced and re-used in this pro- iron, nodular iron production, as a reducing agent cess. Based on the volumes of magnesium they for titanium, and other metal production. buy, it appears that titanium producers purchase In 2011 in the United States the main use of about one to five per cent of their magnesium magnesium metal (43 per cent of total) was in requirement to replace process losses. aluminium-based alloys used for packaging, trans- A review of the world usage divided among dif- portation, and other applications (USGS, 2012). ferent categories with emphasis on North America Structural uses of magnesium (castings and wrought was presented by Slade (2011). This author sug- products) accounted for 40 per cent of primary gested that the outlook for magnesium was brighter metal consumption, with desulfurisation of iron at in 2011 than in 2010. In terms of usage by region, 11 per cent and other uses at six per cent (Figure 11.3). China is projected to continue to be the largest user with over one third of the total. North American consumption in 2011was estimated to be 135,000 Recycling, re-use and resource efficiency tonnes with a similar level of demand in Europe. Other important consumers are in Asia (Japan, It is very important to utilise natural resources, Korea, etc.), Russia, and Brazil. The main market particularly metals, in a sustainable manner. For segments are aluminium alloying, die casting, magnesium, it is particularly important to optimise 270 neale r. neelameggham and bob brown

Table 11.2 Magnesium scrap classification. (After Fechner et al., 2009.)

Type of Scrap Description

Class 1: Sorted clean returns Start up shots, gates, runners, sprues, biscuits, trimmings, rejected parts Class 2: Sorted clean returns with inserts Rejected castings containing metal inserts Class 3: Sorted oily painted returns Post-consumer parts or parts rejected after painting/coating operations, may contain non metallic inserts Class 4: Sorted clean chips Generated during dry machining of magnesium products Class 5: Sorted oily/wet chips Generated during machining of magnesium products using oil or oil/water emulsions Class 6: Dross Salt free furnace cleaning products, mostly oxides from the melt surface, also residues from the bottom of the crucible Class 7: Sludge Salt containing cleaning residues mostly from the bottom of the crucible Class 8: Mixed and off-grade returns Mixed magnesium—including post consumer scrap—off grade magnesium and items not included above

the use of the energy invested to produce the metal. Magnesium is a metal that is easily recycled Production of primary magnesium using an electro- in many of its primary forms. It is also capable lytic process can consume up to 35 kilowatt-hours of being recycled from secondary sources when per kilogram (kWh/kg), including the thermal it appears as an alloy with aluminium. The energy equivalent used in the feed preparation. By recycling of aluminium cans (Used Beverage comparison, re-melting magnesium metal scrap Containers, (UBC)) has become a highly technical uses less than ten per cent of the energy used to and controlled process. The average magnesium make primary metal. content of the UBC scrap is about two per cent. The low density of magnesium makes it By careful handling, most of the magnesium is attractive as a material of construction, particu- recovered in the recycled aluminium. In the larly for transport applications. However, the author’s estimate, in 2010, approximately 740,000 low weight of magnesium can also be a penalty tonnes of cans were recycled, saving nearly according to some methods of calculation that 15,000 tonnes of magnesium in the recovered are based on weight rather than volume. alloy. Albright and Haagensen (1997) noted that the Magnesium scrap classification follows a energy consumption (at the primary plant site system devised by the magnesium alloy producers only) to produce magnesium, aluminium and which was standardised by the International steel are estimated at 35, 30 and 11 kWh/kg, Organisation for Standardisation (ISO) in 2005 respectively. However, when these data are con- (Table 11.2). This classification was developed in sidered on a unit volume basis, which is the the 1980s to ensure the purity of magnesium common design concern (that is, components metal and alloy derived from secondary sources must be designed or packaged into a given space and thereby to reduce problems associated with or volume), the life-cycle advantage of corrosion. magnesium is clearer. On this basis, the above Class 1 scrap is the most frequently and most numbers correspond to 63, 81, and 87 kWh/litre easily recycled. This material consists of gates and for magnesium, aluminium and steel, respec- runners from die casting, together with scrap cast- tively. For recycled materials, the relative ings and used parts that have been isolated from energy used is even more favourable for other scrap. Class 1 magnesium scrap recovery is magnesium, with recycled magnesium energy mostly carried out by special smelters and refiners consumption about one kWh/litre. who take the scrap from the die casters and return Magnesium 271 clean ingot with closely specified alloy chemistry. scrap of known composition. This material is re- This is called ‘toll melting’ since the customer is melted, refined and re-used for manufacturing only charged a conversion fee. magnesium components. The metal from the It is estimated by the author that the total other scrap classes can be recovered, but it is magnesium recycling capacity in the world is mainly used for the production of iron desulfuri- about 200,000 tonnes per year with actual recy- sation reagents or as an alloy element in an cling of about half that amount in 2010. North aluminium alloy production process. America has a Class 1 recycling capacity of The processing of scrap with a low metal 75,000 tonnes per year. In addition, it has been content and of sludges from the melting and recy- estimated that 30,000 tonnes per year could be cling process still poses major problems. As the recycled internally by the die casting companies. utilisation of magnesium alloys in the automo- Class 1 scrap recycling capacity in Europe has bile industry grows, and more attention is been estimated to be 75,000 tonnes per year directed to the management of scrap from end-of- (Fechner et al., 2009). Asia, particularly China, life vehicles, more technical developments will has many primary plants and many of these uti- be required to cover the treatment of all magnesium lise scrap returned by their clients. scrap classes. Several large European, North There are two basic methods of recycling American and Asian research organisations are magnesium: ux-free recycling and recycling working on these problems. with ux. The uxes used are chloride-based salts which have been used for many years to separate oxide and other solid inclusions from the molten Substitution magnesium. In recent years, specially designed furnaces have been developed to enable melting Magnesium alloys are used in many applications to be done without ux. Instead they use inert including the aerospace, automotive and the 3-C atmospheres to prevent the molten magnesium industries. The properties of magnesium, such as from oxidising during melting and refining. its strength-to-weight ratio, its low fabrication The quality and characteristics of scrap may costs and other special properties, such as vary widely, even from one producer of electromagnetic shielding, thermal and electrical magnesium die castings to another. Even Class 1 conductivity, and damping characteristics, com- scrap can vary and, consequently, recycling bine to make it difficult to find a direct substitute contracts are typically based on a preliminary for magnesium. evaluation of representative material. This is From time to time, some plastics or compos- followed by a detailed quality plan produced in ites have been proposed as potential replacements collaboration with the generator of the scrap to for magnesium especially for applications with ensure continuing conformance to the agreed short product lives. However, where longer lives specification. and higher durability are required magnesium The value and cost to the die caster are highly parts generally deliver superior performance – for dependent on the quality of the metal recovered. example, in notebook computer cases and auto- Major components of the recycling costs are: motive instrument panels. 1. value of input material – related chiey to its Some potential problems when magnesium price and to transportation costs; components are being joined can be addressed 2. processing – related to the costs involved in by choosing fasteners with similar thermal pre-treatment, melting and casting; and expansion coefficients, such as aluminium 3. residue disposal – costs associated with the or magnesium fasteners. Neelameggham (2013) digestion and disposal of waste residues. notes that the use of fasteners made of aluminium The industry recycling programmes for or magnesium would also overcome the problems magnesium are applied to clean and compact caused by the inappropriate use of steel fasteners 272 neale r. neelameggham and bob brown which are the root cause of many galvanic controls are not regulated in a consistent manner corrosion issues with magnesium, as well as worldwide. The development of regulations often-publicised creep problems. When a steel and controls started in the US and Europe in fastener is used, its lower thermal expansion 1970, and producers in other countries are now against the increased expansion of magnesium being asked to perform to similar standards by causes stresses when the temperature of the part automotive industry consumers. increases, and when the part cools the bolt tension gets reduced, which is termed creep. When a steel fastener is used for holding Non-greenhouse-gas regulations – electrolytic magnesium alloy base and top part, whenever an magnesium production aqueous solution gets between the steel and magnesium a galvanic cell is created causing Many of the major magnesium production plants magnesium to corrode more rapidly However, if that were based on electrolytic technologies have an aluminium or magnesium fastener is used been shut down. These include Dow Chemical in these problems are avoided and the longevity of the US, Hydro Magnesium in Norway and Canada, the parts increased. and Noranda Magnesium, also in Canada. Some of the original LCA studies were done on the Hydro Magnesium plants in Porsgrunn, Norway Environmental aspects and Becancour, Québec, Canada (Albright and Haagensen, 1997). The Porsgrunn plant in Norway As the lightest available structural metal, the use used sea water and dolomite as raw material of magnesium provides the benefit of reducing which was converted to magnesium oxide and the weight of automobiles and thereby reduces then to magnesium chloride which was used to gasoline consumption. This results in substantial feed the electrolytic cells. At times, the Porsgrunn reduction in the carbon dioxide emissions from an plant also utilised magnesium chloride brine from automobile during its lifetime. Carbon dioxide Germany as a source of magnesium. In Becancour

(CO2) from vehicle engines is one of the major the magnesium chloride was produced by treating sources of anthropogenic emissions of ‘Greenhouse magnesite with hydrochloric acid. Gases ‘(GHG). The environmental regulations of the Notwithstanding the benefits from the use of 1980s were devoted to issues which controlled magnesium in reducing carbon dioxide emis- chlorinated hydrocarbons in water, especially sions, the emissions of GHG during the produc- polychloro biphenyl compounds (PCBs) and tion and processing of magnesium alloys must dioxin. This required Hydro Magnesium to also be considered through the use of tools such modify the electrolytic process in Porsgrunn to as Life-Cycle Analysis (LCA). The climate con- develop the use of hydrogen to convert the anode trol aspects of GHG release to the atmosphere gas chlorine into hydrogen chloride and hydro- from metal production has already prompted pro- chloric acid. This process was selected for the ducers to use sulfur dioxide and other reagents plant built in Becancour which operated from instead of sulfur hexauoride to provide the cover 1991 until 2006. gas used in the production of magnesium alloys. The published estimates for the production The cover gas provides an inhibitive and tena- capacities of each of these plants were about cious film of uoride and/or sulfur compound to 40,000 metric tonnes per annum (Burstow, 1999). prevent reaction of the molten alloy with oxygen. Chlorinated hydrocarbon (CHC) emissions were The issue of GHG emissions has dwarfed actually reduced by half at both plants. Dioxin other production-related pollution control issues emissions also decreased substantially. Albright relating to air, water and solid-waste emissions and Haagensen (1997) also discussed some of the over the past two decades. Even basic pollution global warming issues which came to prominence Magnesium 273

Table 11.3 Life-cycle inventory for magnesium reducing emissions. Although the construction production. (After Albright and Haagensen, 1997.) of electrolytic plants continues to be slow, the research and development of the new equipment Mg Alloy Item Pure Mg AZ91 and methods should be encouraged on a global basis. Total energy (MJ/kg metal) 144 151

Global warming effect (kg CO2 eq /kg) 19 19 Acidification (kg /kg metal) 0.02 0.025 Non-greenhouse-gas regulations – thermal Winter smog (kg /kg metal) 0.015 0.017 magnesium Solid waste (kg /kg metal) 0.5 0.5 Thermal magnesium plants had to address Dioxins to air (µg /kg metal) 0.24 0.21 common environmental regulations on airborne Chlorinated hydrocarbon to air 13.7 12.4 (mg /kg metal) particulate emissions from calciners and ferrosil- icon furnaces. In addition, they had to address disposal of solid wastes comprising mostly reduction furnace residues and casting sludges in the early 1990s following the Kyoto Protocol. containing uxes. Recent thermal processes tend Albright and Haagensen (1997) produced a LCI to make by-products for sale from the solid (Life Cycle Inventory) for Hydro Magnesium residue from retorts in an effort to provide added plants. Separate calculations were made for both revenues. pure magnesium production and for magnesium alloy production (Table 11.3). In regard to the emissions and wastes at Hydro Magnesium Greenhouse-gas emission studies plants, it was reported that “The solid wastes Casting cover-gas issues from Hydro’s plants consists mostly of inorganic salts and minerals from the raw materials. This For most of the twentieth century magnesium waste is disposed of at authorized sites and does production, both electrolytic and thermal, used not cause any negative environmental impact. either sulfur or sulfur dioxide as a cover gas while The treatment of gas and waste water on site, casting magnesium to prevent oxidation. When along with incineration of residues containing this was done in an uncontrolled fashion sulfur chlorinated hydrocarbons, form a portion of the dioxide was released into the work area. The (LCI) analysis as well.” early work by Reimers (1934) was further devel- Designers of Australian Magnesium’s process oped by Hanawalt, and others from Dow, leading also spent considerable effort in meeting the to the use of non-toxic sulfur hexauoride (SF6) environmental regulations in Australia. This pro- as cover gas and this became an acceptable cess, using magnesite feed material, was tested in alternative. Accordingly from the 1970s onwards a 1500 tonne per year pilot operation, but the magnesium companies worldwide started switch- commercial plant was not constructed. ing to using sulfur hexauoride. In 1997 when Noranda was permitting its Following Kyoto Protocol studies, the

Magnola magnesium project, utilising an asbestos advantage of SF6 as a cover was questioned waste (serpentine) as raw material, the company because of its global warming potential. SF6 was was required to meet strict controls on emissions found to have a global warming potential 23,900 of chlorinated hydrocarbons, dioxins and chlo- times that of carbon dioxide (Gjestland et al., rine. Similar concerns must be addressed in the 1997). This problem was addressed by a partner- design and operation of electrolytic magnesium ship between the U.S. Environmental Protection reduction plants today (Brown, 2003b). Improved Agency (US EPA) and the magnesium community scrubbing technology, together with various of producers and users (Bartos, 2001 and Bartos, process modifications, has made great progress in 2002). Several alternative uorine-containing 274 neale r. neelameggham and bob brown cover gases were developed as a result, but most The earliest LCA papers were prompted of the companies started reverting back to using by the environmental work that was done to sulfur or sulfur dioxide in a controlled fashion. get construction permits for new electrolytic magnesium plants planned for Canada and Australia in the late 1980s. The work focused on Life-cycle analysis the GHG impact, including the effects of SF6 and

When comparing the use of various structural CO2 emissions from furnaces. Some of the work metals in automobiles, it has become customary was conducted as part of investigations into the to normalise the advantages and disadvantages by advantage of lighter weight magnesium cars in carrying out Life-Cycle Analyses of the material regard to Life-Cycle emissions. usage with respect to reduction in fuel usage. The initial calculations were part of a study to There have been several life-cycle analyses of assess the environmental impact of a magnesium magnesium production along with downstream engine block supply chain. This study was broad- fabrication of parts for use in automobiles. ened to include an engine block made from Most of the major GHG-related environmental magnesium alloy ingots produced at Becancour problem areas in the production of the metal and secondary ingots produced in the USA. The were studied showing the effects of the electro- study concluded that “the use stage of a passenger lytic magnesium and the thermal magnesium car contributes significantly to the total GHG production methods. Further sub-division assessed impact of magnesium components over their the preparation of the magnesium-containing entire life cycle. A significant reduction in the feed materials and the actual processing of this GHG impact during the use stage, and hence over material to produce magnesium metal. All of the entire life cycle, may be achieved from a these areas are impacted by secondary and upstream significant mass reduction of a car by using processes which impact on the magnesium pro- magnesium components” (Tharumarajah and duction. For example, the method of onsite Koltun, 2005). electricity generation for the production of the Casting operations in China mostly used very input materials may be responsible for significant small amounts of SF6 in the ingot casting opera- emissions, e.g. if coal is used for the electricity tions, and this is an advantage for these processes. required to make 75 per cent ferrosilicon used in The majority of the plants built in China during the Pidgeon process. the 1990s used very simple melting and casting Environmental concerns related to magne- operations with sulfur powder directly dusted on sium production and magnesium processing first the cast ingots to prevent oxidation during solid- appeared in the early 1990s when Life-Cycle ification. When the plants increased in size, most Analysis articles started to be published. of the operations retained the sulfur dusting for Subsequently various studies have compared convenience and simplicity. the environmental impacts of the two main Most of the small Chinese Pidgeon process magnesium production processes. As the mea- plants that used coal as the main energy source surement technology improves and the informa- and large quantities of hand labour are now disap- tion is better understood so the environmental pearing. Larger, more efficient plants, with challenges can be more clearly identified. The advanced technology and equipment, are being pressures from the climate control aspects might built. The Chinese are approaching the problem impact on the status of the differing processes of addressing the global warming potential by used for magnesium production. Academic improving the production plants and working studies on the global warming impacts of pres- hard in downstream processing to produce light- ent-day magnesium production processes have weight magnesium products for use in bicycles, been published (Ramakrishnan and Koltun, motorcycles, cars, trains and airplanes. China has 2004a, 2004b and Cherubini et al., 2008). been carrying out industrial restructuring for the Magnesium 275 purpose of optimising its magnesium industry calcination of dolomite and ferrosilicon produc- and upgrading its competitiveness. Some of these tion are carried out using charcoal instead of measures have been announced in the Chinese coal, and using sulfur dioxide instead of SF6 in 5-year plans, as noted in China Magnesium their casting process (Fernando 2011). For these Industry and Market Bulletin (Sunlight Metal reasons, in combination with their use of hydro- Consulting (Beijing) Co. Ltd., 2011). electric power, Rima’s process probably has The Chinese government has announced new the lowest carbon footprint for magnesium regulations requiring existing smelters to con- production in the world. sume a maximum of 5.5–6.0 tonnes of coal for Dubreuil et al. (2010) made a cradle-to-grave one tonne of magnesium smelting, and for new comparison of magnesium, aluminium and steel smelters, a maximum of five metric tonnes of usage in automobiles. He concluded that coal. Chinese magnesium smelters, over the past magnesium is the preferred metal in terms of its few decades, have been consuming as much as overall carbon dioxide emissions. 11–18 metric tonnes of coal for each tonne of It is also important to note that the raw mate- magnesium smelting, although some smelters in rials used in making magnesium, such as dolo- the last two years were able to cut the usage to mite and magnesium chloride, are not toxic. All around 5–6 metric tonnes coal, following the plants have magnesium in the centre of the chlo- state’s advocacy of a low-carbon economy, rophyll molecule essential for photosynthesis according to Chinese Magnesium Association through the green pigment in the leaves. (CMA) (Shukun et al., 2010). Magnesium is therefore a common additive to Much of the fuel-based pollution in China is fertilisers. Magnesium is also important to all currently being addressed and, wherever it is living cells where it plays a major role in the available, natural gas is being designed into many functioning of numerous enzymes. of the newer plants. It should be noted that many of the Pidgeon plants in the USA and Canada used natural gas as a fuel for their retorts between World resources and production 1940 and 1970. Cherubini et al. (2008) concluded that world Global resources of magnesium, on land and in the magnesium production contributes to the global sea, are huge compared to current production and warming potential with an emission of about usage. Figure 11.4 shows the world production of

25.5 million tonnes of CO2 equivalent per year. magnesium since 1992 and Figure 11.5 shows the This will be lowered with improved energy distribution of production by country in 2010. efficiency and as SF6 is increasingly substituted. There have been a number of closures of for- In the EU car manufacturers are addressing this merly important magnesium producers in the issue as a matter of urgency. They are cutting West in the last two decades. These include carbon dioxide emissions faster than expected Pechiney in France, which stopped producing and will reach the European Union targets ahead magnesium in 2001 from its thermal magnesium of time (D’Errico et al., 2010). plant. In 1997 Dow Magnesium ceased produc- Magnesium is produced by a highly modified tion of magnesium in the USA after a period of 83 vertical furnace process by Rima Metallurgical years. Dow initially used brine in Michigan as the in Brazil. This is a silico-thermic process which source of magnesium and latterly sea water in uses an electrical resistance furnace for reducing Freeport, Texas. At its peak Dow produced about the briquettes of calcined dolomite with ferro- 100,000 tonnes per year, utilising partly hydrated silicon, instead of the fuel-fired retorts of the magnesium chloride cell feed in its electrolytic Pidgeon process. Not only does Rima produce process. For more than 25 years up to 2001, Alcoa its own ferrosilicon, but the company also operated a thermal magnesium production plant states that its process is very ‘green’, noting that in Addy, Washington, USA, with a production 276 neale r. neelameggham and bob brown

900 800 700 600 500 400 300 Thousand tonnes 200 Figure 11.4 World production of 100 magnesium metal, 1992–2010. 0 (Data from British Geological Survey World Mineral Statistics 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 database.)

3% 1% It is of historical interest to note that magnesium 3% 4% was produced at Clifton Junction, UK during and prior to World War II, and also that the Magnesium 5% Elektron company, with its headquarters in the UK, has operated in downstream wrought magnesium applications for over 70 years. Other countries which formerly produced magnesium include Canada, Italy, Norway and Yugoslavia (Serbia). The magnesium plant in Becancour, Quebec, Canada, which operated between 1991 and 2006, produced 40,000 to 50,000 tonnes per year magnesium. The Magnola Metallurgie, Noranda magnesium plant, also in Canada, oper- ated for a short time between 1999 and 2001. Both were electrolytic magnesium producers. The only continuing magnesium producers in 84% the West are: 1. US Magnesium, which has a present capacity of about 50,000 tonnes per year, producing from China USA Russia the Great Salt Lake brines since 1972. Israel Kazakhstan Other 2. Rima Metallurgical, Brazil, which has been in production since the late 1970s and currently produces about 20,000 tonnes per year of thermal Figure 11.5 World production of magnesium metal by magnesium. country in 2010. (Data from British Geological Survey, 2012.) 3. Dead Sea Magnesium, Israel, which has been producing about 30,000 tonnes per year by elec- trolysis of carnallite since 1995. capacity of 27,000–36,000 tonnes. In Japan, Ube 4. Solikamsk, Russia, which has been in Industries stopped its thermal magnesium pro- operation since 1936. Solikamsk’s cumulative duction at the end of 1994. magnesium production passed the one million Magnesium 277 tonne mark in 2009. Its annual production is Molycor Gold Corp., with its high-grade dolo- typically between 15,000 and 20,000 tonnes. mite property at Tami-Mosi in Nevada, com- Total world magnesium production was pleted a resource analysis and a preliminary estimated at 775,000 tonnes in 2010, up from economic assessment for a 30,000 tonnes per year 619,000 tonnes in 2009. 84 per cent of the world’s magnesium plant for the North American magnesium in 2010 was produced by China market, and is in the process of raising funds (Figure 11.5). The demand for magnesium grew at (Molycor Gold Corp., 2011). Molycor, which was an estimated rate of 16 per cent worldwide, renamed Nevada Clean Magnesium Inc. in 2012, including estimates for the magnesium used plans to use a silicothermic reduction method via within China and CIS. About 40 to 45 per cent of an updated and automated Bolzano process. usage continues to be in aluminium alloying. The global economic downturn has led to a world- wide decrease in the die-casting market, which World trade now accounts for between 28 and 35 per cent of total usage. The principal market for die castings Global import and export data for magnesium are continues to be the automotive industry. illustrated in Figures 11.6 and 11.7, respectively. The status of the Netherlands as the largest importer of magnesium is a function of the way Future supplies in which theses data are collected and can be A new thermal reduction plant using the Pidgeon explained because Rotterdam is a free port process was commissioned in Perak, Malaysia in through which many of the transactions pass. 2010 (CVM Minerals, 2010). It has an initial USA, Japan and Germany are the leading design capacity of 15,000 tonnes per year and was importers, which reects the importance of developed with the assistance of a Chinese magnesium use in the automotive and alu- company. minium sectors in these countries. South Korea has recently started primary The export statistics indicate that China magnesium production from dolomite by the exports about 180,000 tonnes per year of their thermal method. Posco, the South Korean steel annual production of 700,000 tonnes per year, producer, will be the main producer, with a pro- indicating the high level of domestic consump- jected output of 10,000 tonnes per year and a tion in China of over 500,000 tonnes per year. plant start-up in 2012. Posco plans two further The high level of exports from Austria is due expansions to bring the magnesium plant’s mainly to recycled magnesium – as Austria is not capacity to 100,000 tonnes a year by 2018 (Eun- a producer of primary magnesium or its alloys. Seong Min, 2011). They have demonstrated the production of a 16-tonne coil of wrought magnesium alloy sheet using the twin roll casting Prices machinery originally developed by Fata Hunter. Posco plans to utilise the sheet magnesium in the It should be noted that magnesium is not traded automotive industry. on any metal exchanges. As a result, prices are Since 2006 the Ural Asbestos Company has the product of negotiated contracts which reect been studying the construction of a magnesium variations in the supply–demand balance and plant using asbestos waste converted to local conditions in different parts of the world. magnesium chloride for an electrolytic process. An internal export tax of ten per cent is The parent company is interested in this project levied on magnesium and magnesium alloys and has been involved in raising capital for the that are shipped out of China. Furthermore, project since 2009 but there were no major there is steady upward pressure on production announcements in 2011. costs in China due to increases in the price of 200

180

160

140

120

100

80 Thousand tonnes 60

40

20

0

USA Italy Japan India Spain Brazil Canada France Mexico Germany Norway Australia Netherlands Rep. of Korea United Kingdom Other Countries

Figure 11.6 The top importing countries of magnesium metal in 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.)

200

180

160

140

120

100

80 Thousand tonnes 60

40

20

0

Italy Israel USA Spain Austria Russia Mexico Japan Germany Romania Belgium Netherlands Czech RepublicUnited Kingdom Other Countries

China (Inc Hong Kong)

Figure 11.7 The top exporting countries of magnesium metal in 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) Magnesium 279

3.50 3.00 2.50 2.00

US$/lb 1.50 1.00 0.50 0.00 1960 1970 1980 1990 2000 2010

Figure 11.8 Year-end prices of magnesium in USA, 1960 – 2011. (Prices are U.S. spot Western price for 99.8%-pure magnesium ingot, compiled from United States Geological Survey (1999) for 1960–1998, and from United States Geological Survey Annual Mineral Commodity Summaries for 1999 to 2011.)

electricity and labour and the need to comply imported supplies. However, demand slumped in with environmental regulations. Since the mid- 2008 as a result of the global economic recession, 1990s prices in the USA have been higher than leading to a fall in the price to about US$2.35 per in the rest of the world mainly because of anti- lb in the USA at the end of 2011. dumping duties levied against metal produced in China, Russia and Canada. According to the World Trade Organisation (WTO), when any Outlook imported product is proven in a court of the International Trade Council to be sold at less The 21st century may see changes in magnesium than the cost of production, the importing production processes not because of lack of raw countries can levy an anti-dumping duty on the materials or lack of process know-how but rather item being imported. from the continued pressure to balance market The long-term price trend for magnesium needs and environmental costs of process compo- metal in the USA is shown in Figure 11.8. nents, as it did in the nineteenth and twentieth Following a long period of abundant cheap centuries. There is continuing international supplies magnesium prices rose steadily from the research and development into reducing the GHG early 1970s after the oil crisis. In the following emissions associated with the production of decade the price of magnesium more or less tre- magnesium. In China, while the older Pidgeon bled in response to escalating energy costs and process produced emissions of about 36 kg CO2 increased demand. New suppliers in North per kg of magnesium, this is being reduced to

America and Russia entered the market from about 20 kg CO2 per kg of magnesium by better early 1990 causing the price to fall back, stabilis- energy management, as noted above. In contrast, ing at just below US$1.50 per lb until about 1994. the electrolytic magnesium plant in Canada, However, increasing demand in the USA, together which was operated with hydroelectric power with the effects of the anti-dumping duties, led to and used magnesite as the starting material, pro- rapid price escalation during 1995 (USGS, 1999). duced 6.9 kg CO2 per kg magnesium (Das, 2010). In the following year imports from Russia and The other western electrolytic process in the US Israel resumed resulting in lower prices in the and elsewhere will have higher emissions due to USA. Prices began to rise sharply again in 2006 in the use of calcium carbonate for removing sulfate the face of increasing demand and restrictions on impurity from the brine, and for evaporation and 280 neale r. neelameggham and bob brown dehydration of the purified brine. The use of elsewhere and primary metal production con- power derived from fossil fuels also contributes tinues in the US, whilst producers in Canada, to higher emissions using this technology. France, Italy and Norway have shut down. If Alternative raw materials, such as sulfate min- demand for magnesium in China grows faster erals (gypsum, epsomite, etc.), are currently being than production the volume of Chinese exports investigated by the authors for use in the produc- will inevitably fall, which could limit global tion of magnesium where they may lead to market growth – unless other producers expand further significant reductions in carbon dioxide output or enter the market. The fact that most of emissions. the world now uses lower-priced Chinese metal The unique properties of magnesium make it has limited the growth of the magnesium market. essential for some important applications. In Improved recycling of magnesium from end-of- particular, the low density of magnesium alloys life products will contribute to improved supply means that they are especially useful when ‘light- security and to the reduction of the environ- weighting’ is an important factor, such as in the mental impacts associated with magnesium use. transport sector. The high reactivity of magnesium A significant portion of U.S. demand for is important in some applications which together magnesium depends on aluminium alloys. currently account for less than 20 per cent of total Aluminium production in the United States is magnesium production. These include magnesium increasing as aluminium companies are opening use in the desulfurisation of iron and steel, in the new capacity and upgrading rolling mills to make production of titanium and other reactive metals car body sheet to help make lighter cars. In mid- from their compounds and in the production of 2010, only about 50 per cent of the US primary specialised ares for military applications. Most aluminium production capacity was in use. A of the magnesium used in the production of recent report on the aluminium sheet business by titanium is recycled and is not counted in the Novelis, a large USA aluminium sheet maker, total world production. The total volume of stated that the can sheet market is mature and magnesium utilised each year is growing and is usage remains level in the USA (Martens, 2011). likely to continue to do so, especially as new In contrast, in Europe and Japan the use of alu- applications are developed. The main growth minium cans is increasing. It was also noted that areas are likely to be in the structural uses of the automotive industry will see an acceleration magnesium, either in die-cast metal components of aluminium use, and that, in the USA there will or wrought-alloy sheet components. be a significant increase in body structures made It is important to note that the quantity of up of some percentage of aluminium by the end magnesium present in seawater, brines, dolomite of this decade. and other minerals is very high in relation to present The level of use of magnesium die casting varies and anticipated future demand for magnesium. around the world. In the past 25 years diecast However, due to a variety of other factors the magnesium has been used mainly in the number of locations where primary magnesium is construction industry for lightweight hand tools produced has decreased since the 1980s. China now and in the automotive industry. Magnesium die dominates global output using a technology that casting usage in China has grown markedly in may not have the lowest environmental footprint, recent years compared to other countries, with the although China has made major improvements in largest growth in the 3-C industries and in the its environmental controls and energy reduction in automotive sector. Over the past six years, the its process industries in recent years. Chinese automotive industry has seen an increase The US Government currently imposes an of more than 500 per cent in the number of cars anti-dumping duty on magnesium metal and produced. alloys imported from China with the result that In 2011, it is forecast that auto production the price for magnesium in the US is higher than demand in the United States for die castings will Magnesium 281 be up ten per cent, partially due to the increase in research centre is the Magnesium Research the US Fuel Economy standards from 27 to 35 Institute (MRI) in Beer-Sheva, southern Israel, miles per gallon. The use of new magnesium part which was established in the mid-1990s. The designs that incorporate crash energy management MRI is concentrating on developing new prod- and improved fit and finish over assembly of steel ucts, processes and markets for magnesium use stampings provides some examples of the move and applications. towards increased magnesium use. The inner lift gate for the Lincoln automobile is a 7.7 kg die casting which replaces six steel stampings weigh- References ing 13.6 kg. General Motors has developed a 10-kg magnesium die cast battery tray for the Albright, D. and Haagensen, J. (1997) Life Cycle Chevrolet Volt. This is a single piece casting that Inventory of Magnesium, Proceedings of the 54th replaces a steel assembly which was twice as Annual World Magnesium Conference (IMA), heavy. Toronto, Canada, June 1997, 32–37. Governments and industry worldwide are Avedesian, M. (1999) Magnesium and Magnesium currently funding major research programmes Alloys. In: Avedesian, M. and Baker, H. (eds.) ASM into developing new uses for magnesium. There Specialty Handbook, ASM International, Materials is particular interest in the automotive sector Park, Ohio, USA, 1–4. where manufacturers already use magnesium Bartos, S.C. (2001) United States Environmental Protection Agency SF6 emission reduction partner- for individual components, but where new ship for the magnesium industry: an update on early alloys and processing methods are needed before success. In: Hryn, J. Magnesium Technology, TMS it can become economically and technologi- Annual Meeting, 43–48. cally feasible as a major automotive structural Bartos, S.C. (2002) Building a Bridge for Climate material. Protection: United States Environmental Protection An international collaboration between Canada, Agency and the Magnesium Industry, Proceedings of China and the United States, called the Magnesium the 59th Annual World Magnesium Conference (IMA), Front End Research and Development (MFERD) Montreal, Canada. May 2002, 22–24. programme, was initiated in 2007. The goal of Beck, A. (1939) The Technology of Magnesium and Its this programme is to advance the manufacture of alloys. F.A. Hughes and Co. Ltd, London, 1–19. magnesium-intensive vehicles, by developing tech- Brooks, G., Trang, S., Witt, P., Khan, M.H.N. and Nagle, M. (2006) The carbo thermic route to magnesium. nologies and knowledge that will lead to a vehicle JOM The Member Journal of TMS (The Minerals, front-end body structure that is 50 to 60 per cent Metals and Materials Society) 58 (05), 51–55. lighter, equally affordable, more recyclable, and of Brown, R.E. (2000) Magnesium Industry Growth in the equal or better quality than today’s vehicles. 1990 Period. In: Kaplan, H.I., Hryn, J. and Clow, B. Toward this end, databases are being developed (eds.) Magnesium Technology 2000, Nashville, TN, that will enable further alloy and manufacturing TMS Annual Meeting, 3–12. process optimisation, and a life-cycle analysis of Brown, R.E. (2003a) A History of Magnesium the magnesium alloys is being conducted. Production. http://www.magnesium.com/w3/data- In Europe there are several active research bank/index.php?mgw=196. groups including the Magnesium Innovation Brown, R.E. (2003b) M-Cell Modernization Improves US Magnesium Process and Environmental Centre (MagIC) in Geesthacht, Germany and Performance. Light Metal Age, June 2003, 6–12. BCAST (the Brunel Centre for Advanced Burstow, C. (1999) A Strategic Analysis of the Global Solidification Technology) in London. These Magnesium Industry, Metal Bulletin Research, London. groups focus on twin roll casting, although Cherubini, F., Raugei, M. and Ulgiati, S. (2008) LCA of MagIC is also one of the lead institutes in the Magnesium Production: Technological Overview and relatively new field of developing magnesium Worldwide Estimation of Environmental Burdens. alloys for medical implants. Another important Resources, Conservation and Recycling 52, 1093–1100. 282 neale r. neelameggham and bob brown

CVM Minerals Ltd. (2010) Company profile, Message Holywell, G.C. (2005) Magnesium: the first quarter mil- and Vision. http://www.cvmminerals.com/ceo.html lennium. JOM The Member Journal of TMS (The

Das, S. (2010) Life cycle energy and CO2 Analysis of Minerals, Metals and Materials Society) 57 (07), 25–33. Solid Oxide Membrane Primary Magnesium Jarrett, N. (1981) Advances in the Smelting of Production Technology, U.S. DOE Report No. Magnesium. American Institute of Mining and DE-AC5-00OR22725. Metallurgy, Englewood, Colorado, USA. Dean, K.C. (1965) Economic and Technical Evaluation Jones, T. and DeLorme, R.D. (2008) Development of a of Magnesium Production Methods - Part 1. Metallo Ballistic Specification for Magnesium Alloy AZ31B. Thermic. USBM R.I. 6656. Technical Report / ARL-TR-4664, Army Research Dubreuil, A., Bushi, L., Das, S., Tharumarajajah, A. and Laboratory. Xianzheng, G. (2010) A Comparative Life Cycle Jones, T. and Kondoh, K.(2009) Initial Evaluation Assessment of Magnesium Front End Autoparts. of Advanced Powder Metallurgy Magnesium Alloys http://aluminumintransportation.org/downloads/ for Armor Development, Technical Report/ARL- MFRED-LCA-Study.pdf TR-4828, Army Research Laboratory. Elkins, A., Dean, K.C. and Rosenbaum, J.B. (1968) Martens, P. (2011) http://www.novelis.com/en-us/ Economic Aspects of Magnesium Production. In: Pages/Video-Gallery.aspx. Henrie, T.S. and Baker Jr, D.H. (eds.) Electrometallurgy, Molycor Gold Corp. (2011) Positive preliminary Proceedings of the Extractive Metallurgy Division economic assessment study completed for Tami- Symposium on Electrometallurgy, AIME, Cleveland, Mosi magnesium project. Molycor Gold Corp. http:// OH, 1968 – Dec, 173–184. www.molycor.com/positive-preliminary-economic- Emley, E.F. (1966) Principles of Magnesium Technology, assessment-study-completed-for-tami-mosi-magne- London, Pergamon. sium-project/ D’Errico, F., Fare, S. and Graces, G. (2010) The Next Muzhzhavlev. K.D., Lebedev, O.A., Frantas’ev, N.A., Generation of Magnesium Based Material to sustain Olyunin., G.V., Sheka, T.S. and Dolgikh, T.K. (1965) the IPCC Policy. Unpublished paper from Italy/Spain. Results of tests on individual elements in the design of Eun-Seong Min. (2011) Vision of POSCO Magnesium – magnesium electrolyzers. Tsvetn. Metal 38 (5), 57–60. from Raw Materials to Products. International Muzhzhavlev, K.D., Kostrev, S.P., Schegolev, V.I., Ivanov, Magnesium Association Annual Meeting, Prague, A.B., Romanenko, O.N., Yazev, V.D. and Vasiliev, A.V. May 2011. (1977) Diaphragmless electrolyzer for producing Faure, C. and Marchal, J. (1964) Magnesium by the magnesium and chlorine. USA Patent 4058448. Magnetherm Process. Journal of Metals (1964 Neelameggham, N.R. (2013) Magnesium metal pro- September), 721–723. duction technology. In: Pekguleryuz, M., Kainer, K. Fechner, D., Hort, N., Blawert, C., Dietzel, W. and Kainer, and Kaya, A. (eds.) Fundamentals of Magnesium K.U. (2009) Magnesium Recycling: State-of-the-Art Alloy Metallurgy, Woodhead Publishing, Cambridge, Developments, Part I. Light Metal Age (4), 46–50. UK. Fernando, F. (2011) Rima’s Process: Green Magnesium Pidgeon, L.M. and Alexander, W.A. (1944) Thermal pro- From a Fully Integrated Plan. Proceedings of the 68th duction of magnesium – pilot plant studied on the Annual World Magnesium Conference (IMA), Prague, retort ferrosilicon process, Transactions American Czech, May 2011, 80–84. Institute of Mining and Materials Engineers 159, Gann, J.A. (1935) Magnesium in Modern Uses of Non- 315– 352. ferrous Metals. In: Mathewson, C.H. (ed.) Proceedings Ramakrishnan, S. and Koltun, P. (2004a) A comparison of American Institute of Mining and Metallurgy. of the greenhouse impacts of magnesium produced by Gjestland, H., Westengen, H. and Phahte, S. (1997) Use electrolytic and Pidgeon process. In: Luo, A. (ed.)

of SF6 in the Magnesium Industry: An Environmental Magnesium Technology 2004, Charlotte, NC, TMS Challenge. Proceedings of the Third International Annual Meeting, 173–178. Magnesium Conference, The Institute of Materials, Ramakrishnan, S. and Koltun, P. (2004b) Global London, 33–41. Warming Impact of the Magnesium Produced in Hansgirg (1932) Carbothermal production of China Using the Pidgeon Process. Resources, magnesium. U.S. Patent, 1884993, 1932 - October. Conservation and Recycling 42 (1), 49–64. Haughton, J.L. and Prytherch, W.E. (1938) Magnesium Reimers, H.A. (1934) Method for inhibiting the and Its Alloys. Chemical Publishing Co, pp. 170. oxidation of readily oxidizable. USA Patent 1,972,317. Magnesium 283

Rudnick, R.L. and Gao, S. (2004) Composition of the http://www.lightmetal.com.cn/en/content. Continental Crust. In: Holland, H.D. and Turekian, asp?id=4. K.K. (eds.) Treatise on Geochemistry, Volume 3, The Tharumarajah, A. and Koltun, P. (2005) Lifecycle Crust. Elsevier, Pergamon. Assessment of Magnesium Component Supply Shukun, M., Xiuming, W. and Jinxiang, X. (2010) China Chain. Proceedings of the 62nd World Magnesium Magnesium Development Report in 2009. IMA 67th Conference (IMA), Berlin, Germany May 2005, Annual world magnesium Conference, Hong Kong, 67–73. 2010-May, 3–10. UN Comtrade (2013) United Nations Commodity Slade, S. (2011) Presentation at Magnesium Technology Trade Statistics Database, Department of Economic 2011. TMS Annual Meeting, San Diego, CA. and Social Affairs/ Statistics Division, http:// Strelets Kh, L. (1977) Electrolytic Production of comtrade.un.org/db/ Magnesium. (Translated from Russian by J. United States Geological Survey (1999) Metal Prices in Schmorak.) Israel Program for scientific Translations, the United States through 1998. In: Plunkett, P.A. NTIS, Springfield, VA. and Jones, T.S. (eds.) Magnesium, 79–81. Sunlight Metal Consulting (Beijing) Co. Ltd. (2011) United States Geological Survey (2012) Mineral China magnesium bulletin industry and market. Commodity Summaries: Magnesium Metal. 12. Platinum-group metals

GUS GUNN

British Geological Survey, Keyworth, Nottingham, UK

Introduction data storage, medical implants and renewable energy. The six chemical elements normally referred to The name platinum is derived from the Spanish as the platinum-group elements (PGE) are, in ‘platina’ meaning little silver. Archaeological evi- order of increasing atomic number: ruthenium dence indicates that there is a very long history (Ru), rhodium (Rh), palladium (Pd), osmium of platinum working in South America (Scott (Os), iridium (Ir) and platinum (Pt). Metallurgists and Bray, 1980). The platinum, which was used and engineers more commonly refer to them as for making ornaments, was derived from sands platinum-group metals or PGM. Minerals in and gravels in rivers draining the Pacific low- which one or more of the PGM comprise an lands of Ecuador and Colombia. Its discovery is essential part are referred to as platinum-group often credited to the Spanish military leader minerals, also often abbreviated to PGM. In this Don Antonio de Ulloa who brought it to Europe chapter PGM is used to refer to platinum-group from South America around 1750. Palladium metals. was discovered in 1803 by the English chemist The PGM are rare precious metals, although William Hyde Wollaston. It was named after the unlike gold, they are used in a diverse range of asteroid Pallas, identified in 1801, and which industrial applications as well as in jewellery. itself was named after the Greek goddess of Today, platinum and palladium are the most wisdom. commercially important of the PGM with their largest application being in the automotive industry where they are used to reduce harmful Properties and abundance in the Earth emissions from vehicle exhaust systems. Rhodium, which is also used in autocatalysts, The PGM are transition metals, which together is the third most important PGM, albeit with with iron, cobalt and nickel, are found in Groups consumption an order of magnitude less than 8, 9 and 10 of the Periodic Table. Osmium, that of platinum and palladium. Ruthenium iridium and ruthenium are sometimes referred and iridium are used in smaller amounts to as the IPGM or iridium sub-group, and although they are becoming increasingly impor- platinum, palladium and rhodium as the PPGM tant in a variety of new technologies such as or palladium sub-group, on account of their

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Platinum-group metals 285

Table 12.1 Selected properties of the six platinum-group metals (PGM) compared with gold (Au).

Property Value Units

Name Platinum Palladium Rhodium Iridium Ruthenium Osmium Gold

Symbol Pt Pd Rh Ir Ru Os Au Atomic number 78 46 45 77 44 76 79 Atomic weight 195.08 106.42 102.91 192.22 101.07 190.23 196.97 Density at 25 °C 21450 11995 12420 22550 12360 22580 19281 kg/m3 Melting point 1769 1554 1960 2443 2310 3050 1064 °C Electrical resistivity at 25 °C 106 105 45 51 71 92 22 nΩ m Hardness (Mohs scale) 4–4.5 4.75 5.5 6.5 6.5 7 2.5–3 similar geochemical behaviour under magmatic concentrations of approximately five parts per conditions. Together with gold and silver, the billion (ppb), by weight. Rhodium, iridium and PGM have long been referred to as noble or pre- ruthenium are even scarcer at about one ppb. cious metals. Relative to other rock types the PGM are enriched All six PGM are chemically similar with in ultramafic lithologies, such as peridotite, in strong siderophile and chalcophile tendencies, which platinum and palladium concentrations preferentially bonding with iron, nickel, copper are commonly 10–20 ppb. and sulfur rather than with oxygen. The distri- bution of PGM in the Earth is thus controlled by the presence of metallic phases, being strongly Mineralogy concentrated in the Earth’s core after planetary accretion. They do not dissolve in strong acids, In nature, the PGM are chiefly held either in but they do react with oxygen at high tempera- base-metal sulfide minerals, such as pyrrhotite, tures to form volatile oxides. The physical prop- chalcopyrite and pentlandite, or in PGM-bearing erties of the PGM vary considerably (Table 12.1) accessory minerals. Cabri (2002) provides miner- but all display properties typical of metals, alogical and chemical data for more than 100 dif- including the ability to form alloys, to conduct ferent platinum-group minerals, although many heat and electricity, and some degree of mallea- other poorly characterised forms are also known. bility and ductility. Platinum, iridium and Some are very rare and are known only from one osmium are the densest known metals, being deposit. significantly denser than gold. Platinum and pal- The PGM rarely occur as native metal but ladium are highly resistant to heat and to corro- commonly form a wide variety of alloys with one sion, and are soft and ductile. Rhodium and another or with other metals, notably with iron, iridium are more difficult to work, while ruthe- and less commonly with tin, copper, lead, mer- nium and osmium are hard, brittle and almost cury and silver. In other platinum-group minerals unworkable. All PGM, commonly alloyed with the PGM are bonded to sulfur, arsenic, antimony, one another or with other metals, can act as cat- tellurium, bismuth and selenium (Table 12.2). alysts which are exploited in wide range of industrial applications. Platinum and palladium are of major commercial significance, while, of Major deposit classes the other PGM, rhodium is the next most important. Enrichment of PGM occurs in deposits of several The PGM are very rare in the Earth’s crust types developed in a limited range of geological with platinum and palladium present at similar settings, commonly associated with nickel and 286 gus gunn

Table 12.2 A selection of the more common platinum-group minerals.

Group Name Formula Group Name Formula

Alloys Isoferroplatinum Pt3Fe Arsenides Sperrylite PtAs2

Osmiridium Oslr Stillwaterite Pd8As3

Rustenburgite (Pt,Pd)3Sn Antimonides Genkinite (Pt,Pd,Rh)4Sb3

Tulameenite Pt2FeCu Isomertieite Pd11As2Sb2

Sulfides Braggite (Pt,Pd)S Stibiopalladinite Pd5Sb2

Cooperite PtS Geversite Pt(Sb,Bi)2 Hollingworthite RhAsS Tellurides Kotulskite PdTe

Laurite (Ru,Os)S2 Merenskyite (Pd,Pt)(Te,Bi)2

Bismuthides Froodite PdBL2 Moncheite (Pt,Pd)(Te,Bi)2

copper. A map showing the distribution of major ● Merensky Reef type; deposits, mines and districts, and other selected ● chromitite reef type PGM occurrences worldwide is shown in ● contact type; Figure 12.1. Magmatic PGM deposits, found in ● dunite pipes. mafic and ultramafic igneous rocks, are of two All of these are typically developed in the principal types (Table 12.3): PGM-dominant Bushveld Complex in South Africa, which is deposits which are associated with sparse, dis- the largest layered igneous complex in the persed sulfide mineralisation (Maier, 2005); and world. nickel–copper sulfide deposits in which the PGM occur in association with sulfide-rich ores Bushveld Complex (Naldrett, 2010). A variety of other, less econom- The Bushveld Complex (ca 2054 Ma1) hosts about ically important deposit types is also known 90 per cent of all known platinum resources in (Table 12.4). the world and accounts for more than 75 per cent In simple terms PGM-sulfide deposits are of global platinum production. It extends about derived from magmatic processes of crystallisa- 450 km east–west and 350 km north–south, and is tion, differentiation and concentration. The about 9 km thick (Figure 12.2). The Bushveld magma becomes saturated with sulfur and an Complex is divided into three main sectors: the immiscible sulfide liquid separates from the Western Limb, the Eastern Limb and the magma as disseminated droplets. On account of Northern Limb (Cawthorn et al., 2002). It is sau- their chalcophile behaviour the PGM are concen- cer-shaped with the layered mafic and ultramafic trated strongly in the sulfide liquid and are, there- rocks dipping into the centre of the intrusion fore, scavenged from the silicate liquid. An ore where they are covered by felsic lithologies. The deposit may form if this PGM-enriched sulfide platinum-bearing horizons are located within liquid accumulates in sufficient quantities at a part of the complex, known as the Critical Zone, particular location. which comprises packages, or cyclic units, of repetitively layered chromitite, pyroxenite, PGM-dominant deposits norite and anorthosite on a scale of centimetres In these deposits the PGM are the main economic to tens of metres. products with minor nickel and copper derived Platinum is currently mined from three from sparsely disseminated (up to five volume horizons in the Bushveld Complex: the per cent) sulfides, chiefly pyrrhotite, pentlandite Merensky Reef, the UG2 Chromitite Reef, and and chalcopyrite. Four classes of PGM-dominant the Platreef. The Merensky Reef has been deposits are recognised: exploited for many years with PGM production Figure 12.1 The distribution of the main platinum-group minerals mining districts, mines and deposits. Selected occurrences of other types are also shown (some nickel mines which produce by-product platinum-group minerals are omitted). 288 gus gunn

Table 12.3 Key characteristics and examples of the major platinum-group minerals deposit types.

Deposit type Description Typical grades Major examples

PGM-dominant Merensky type Extensive, laterally continuous, thin 5 − 7 g/t Pt + Pd; Pt/Pd = 3 in Merensky Reef, Bushveld Complex, South layers, or ‘reefs’, of ultramafic rocks Merensky Reef. Major source of Africa; Great Dyke, Zimbabwe; J-M Reef, in large layered mafic-ultramafic Pt, Pd and Rh. Stillwater Complex, USA. intrusions; with minor disseminated nickel and copper sulfides. Chromitite type Similar morphology to Merensky 4 − 8 g/t Pt + Pd; Rh 0.3 − 0.6 g/t UG2, Bushveld Complex, South Africa; type but comprising thin layers in UG2. Pt/Pd = 2.5 in UG2. Lower Chromitites, Stillwater Complex, of massive chromite with sparse Major source of Pt, Pd and Rh. USA. base metal sulfide minerals. Contact type Extensive (km) discontinuous PGM 1–4 g/t Pt + Pd in Platreef, with Platreef, Bushveld Complex, South Africa; mineralisation with low-grade Pt/Pd ca. 1; by-product Ni and Duluth Complex, USA; East Bull Lake, nickel and copper in basal contact Cu. Major PGM resources. Canada; Portimo, Finland. zones of layered intrusions. Commonly heterogeneous, brecciated, xenolithic. Dunite pipes High grade platinum mineralisation 3–2000 g/t. Largely worked out Onverwacht, Driekop and Mooihoek, in discordant pipe-like bodies of and no longer mined. Bushveld Complex, South Africa. dunite, up to 1 km in diameter. Nickel-copper-dominant (with by-product PGM) Associated with Nickel-copper sulfide deposits in 1–10 g/t Pt + Pd, Pt/Pd < 1. Sudbury, Canada. meteorite impact melt rocks and underlying Numerous deposits worked for impact radial and concentric fracture and Ni, with lesser Cu and PGM. breccia zones. Related Nickel-copper sulfide deposits in 2–100 g/t PGM. Average 7.31 g/t Norilsk, Russia; Jinchuan, China. to rift- and sub-volcanic sills which feed flood Pd and 1.84 g/t Pt in Noril’sk. continental- basalts associated with flood basalts intracontinental rifting. Komatiite related Nickel sulfide deposits related to Commonly a few hundred ppb, Kambalda, Western Australia; Pechenga komatiitic (magnesium-rich) locally greater than 1 g/t. district, Russia; Thompson Belt, volcanic and intrusive rocks of Pt/Pd generally <1. Canada; Ungava Belt, Quebec, Canada. Archaean and Palaeoproterozoic age.

Cu, copper; Ni, nickel; Pd, palladium; Pt, platinum; Rh, rhodium.

starting in 1928 in the Rustenburg area. The from the Bushveld. The Platreef is restricted to UG2 Chromitite Reef was not mined on a the Northern Limb of the complex where min- commercial scale until 1985 on account of met- ing commenced in 1926 but lasted only a few allurgical problems related to the high chrome years. Open-pit mining restarted in 1992 at and low sulfide contents. However, the UG2 is Mogolakwena and higher PGM prices have now worked at several localities and accounts spurred considerable exploration on the Platreef for a significant proportion of PGM production in the last decade. Platinum-group metals 289

Table 12.4 Key characteristics and examples of other platinum-group minerals deposit types.

Deposit type Description Typical grades Examples

Alaskan or Alaskan-Ural type PGM enrichment in small zoned Generally low grade, locally very Nizhny Tagil, Urals; Tulameen ultramafic complexes. high. Complex, Canada. Placer An accumulation of dense Generally low grades with Urals, Russia; Choco River, platinum-group minerals, sporadic rich pockets. Colombia; Goodnews Bay, normally in sand and gravel Alaska; Southland, New deposits in a river system. Zealand. Ophiolites PGM enrichment chiefly with Values in g/t range reported Unst, UK; Zambales, Phillippines; podiform chromite deposits from Unst. Veria, Greece; Kempirsai, in lower sections of ophiolite Russia; Al Ays, Saudi Arabia. complexes. Hydrothermal processes may enrich PGM. Laterites Residual enrichment of PGM ca 0.5 g/t Pt. Yubdo, Ethiopia. caused by intense tropical weathering of ultramafic rocks. Hydrothermal Polymetallic veins and breccias Up to tens of g/t Pt + Pd. New Rambler, USA; Waterberg, and/or hydrothermal Hundreds of g/t at South Africa; Rathbun Lake, remobilisation of nickel, Waterberg. Canada; Goodsprings, Nevada. copper and PGM in mafic-ultramafic rocks. Unconformity-related Hydrothermal PGM enrichment 0.19 g/t Pt, 0.65 g/t Pd, Coronation Hill, Australia; Serra with gold ± uranium in 4.31 g/t Au at Coronation Pelada, Brazil. metasedimentary rocks at, Hill. Tens/hundreds g/t Pt or close to, an unconformity and Pd at Serra Pelada. with younger, oxidised sedimentary rocks. Commonly focused along faults. Porphyry deposits Local PGM enrichment with Generally less than 1 g/t PGM. Elatsite, Bulgaria; Skouries, gold in porphyry style copper Greece; Santo Tomas II, deposits hosted by igneous Phillippines. intrusions. Shales and other sedimentary PGM enrichments in Generally less than 1 g/t, but Kupferschiefer, Poland; Yukon, rocks metalliferous black shales. local very high values in the Canada; Sukhoi Log, Siberia. Generally thin but may be Kupferschiefer. laterally extensive. Carbonatite and alkaline PGM enrichment in small Few hundred ppb in Palabora; Palabora, South Africa; Catalão complexes carbonatite-phoscorite few g/t in Brazilian and Ipanema, Brazil; Kovdor, igneous complexes. examples. 0.76 g/t Pd, Russia. Coldwell Complex, Disseminated copper- 0.23 g/t Pt and 0.27% Canada; Marathon, Canada. palladium-rich sulfide Cu in Marathon deposit. mineralisation in gabbro at Coldwell.

Au, gold; Cu, copper; Pd, palladium; Pt, platinum. 290 gus gunn

BOTSWANA NORTHERN LIMB N NAMIBIA

Pretoria Johannesburg

SWAZILAND Boikgantsho POLOKWANE Mogalakwena Durban SOUTH EASTERN LIMB AFRICA LESOTHO MOKOPANE Lebowa Bokoni Volspruit Limpopo Ga-Phasha

Cape Town Twickenham Port Elizabeth Marula Modikwa BURGERSFORT Tumela Dishaba Zondereinde Kennedy’s Vale Smokey Hills Two Rivers Spitzkop Union Der Brochen WESTERN LIMB Mareesburg Booysendal Mototolo Magazynskraal Blue Ridge Everest Pilanesberg Frischgewaagd-Ledig Sheba’s Ridge Styldrift

Bafokeng- Thembelani Rasimone Khomanani Crocodile Western Leeuwkop River Bushveld JV Eland Impala Khuseleka RUSTENBURG Bathopele Pandora PRETORIA Lonmin Kroondal Marikana Siphumelele 0 20 km JOHANNESBURG

Major divisions of the Bushveld Complex Mines Approximate outcrop of Upper zone Advanced projects Merersky Reef and UG2 Main zone Reef Critical zone Lower zone Faults Marginal zone Towns

Figure 12.2 The geology of the Bushveld Complex showing the location of platinum mines and advanced projects.

Merensky Reef type is a major source of global production of platinum The Merensky Reef is the type example of a PGM- and palladium. dominant deposit associated with a minor amount These deposits occur as thin, laterally exten- of sulfide mineralisation. Other important exam- sive, stratiform zones of sparsely disseminated ples are found in the Great Dyke, Zimbabwe and sulfides within major layered mafic-ultramafic the Stillwater Complex, USA. This deposit class igneous intrusions. In the Bushveld Complex Platinum-group metals 291 these deposits are typically located about 2 km up to 300 m thick, that can be traced about above the base of the intrusion. There is no con- 35 km along the base of the Northern Limb of sensus on the processes responsible for their the Bushveld Complex (Kinnaird and MacDonald, genesis but the most widely accepted model 2005). It has generally higher base-metal con- involves mixing of a residual magma after partial tents than Merensky-type deposits, particularly crystallisation with a new pulse of magma. This nickel and copper, up to about three volume per leads to sulfide saturation and separation of cent. The Platreef is a major PGM resource with immiscible sulfide droplets which scavenge mineralisation in both the pyroxenite and the PGM from the silicate magma (Lee, 1996 and underlying rocks. The mineralisation is vari- Maier, 2005). able in thickness and is best developed along a The Merensky Reef comprises a pegmatoidal 25-km section of strike length. It varies consid- pyroxenite layer about 1 m thick, bounded on erably within this interval, in part according to both sides by thin chromitite layers and contain- the nature of the footwall rocks which include ing up to three per cent disseminated sulfides. It extensive sections of granite-gneiss basement, can be traced laterally for hundreds of kilometres dolomite, banded iron formation, chert, and is found on both the Eastern and Western quartzite and shale. The mineralisation in the limbs of the Bushveld Complex. The grade of the Platreef is less laterally continuous than the Merensky Reef averages about 5–7 g/t Pt + Pd reef styles and the PGM contents are also gen- (Maier, 2005), although the mineralogy is quite erally lower, typically in the range 1–4 g/t. The variable and includes a wide range of alloys, sul- PGM mineralogy of the Platreef is dominated fides, tellurides and arsenides. by tellurides, bismuthides and arsenides, but this varies significantly both laterally and vertically. Chromitite Reef type The PGM-dominant mineralisation in the The UG2 Reef, one of many thin near-massive Platreef may be classified as a contact-type PGM chromite layers in the Critical Zone of the deposit, which comprise basal accumulations of Bushveld Complex, is generally regarded as the magmatic sulfides associated with abundant largest PGM resource in the world. The ruthe- xenoliths and brecciation, and coarse pegmatitic nium and rhodium contents of the UG2 are also textures. Other examples of this class are recog- significantly higher than the Merensky Reef, and nised in the USA, Finland, the Kola Peninsula of hence it is an important source of these metals. Russia and in Ontario, Canada (Table 12.3), The UG2 comprises a chromitite layer between although these deposits are not mined for PGM at 0.7 and 1.3 m in thickness with sparse dissemi- present. nated sulfide mineralisation. The UG2 has a sim- ilar lateral extent to the Merensky Reef, occurring below it at depths from 30 m up to about 400 m. Dunite pipes The average grade of the UG2 varies from approx- imately 4–8 g/t Pt + Pd. The platinum-group min- These are discordant pipe-like bodies, up to 1 km erals in the UG2 comprise mainly sulfides and in diameter, of varying composition, which are alloys, including laurite, cooperite, braggite and found widely within the Bushveld Complex Pt–Fe alloy. (Scoon and Mitchell, 2004). However, in only four of them, located in a 20-km belt on the Eastern Limb to the north-west of Burgersfort, has Contact type significant PGM mineralisation been discovered. This type of deposit is exemplified by the High-grade platinum ore, locally up to 2000 g/t Platreef in the Bushveld Complex. The Platreef Pt, occurs in discrete zones of the pipes within is a layer of ultramafic rocks, chiefly pyroxenite, unmineralised dunite. Discovered in 1924 and 292 gus gunn largely mined out by 1930, there is little modern processes responsible for its formation remain information on these deposits. unclear (Maier, 2005). Open-pit mining com- menced at Lac des Iles in 1993, with underground operations following in 2006. Other PGM-dominant deposits Nickel–copper-dominant deposits The Great Dyke (2575 Ma) is a layered mafic- ultramafic intrusion up to 11 km wide that can Magmatic nickel–copper deposits are the most be traced for 550 km in a north-north-easterly important source of nickel worldwide (Naldrett, direction across central Zimbabwe (Oberthu r, 2010). Copper, cobalt and the PGM, mainly palla- 2002). The Great Dyke is a major PGM resource dium, are important by-products. Gold, silver, and also contains important deposits of chro- chromium, sulfur, selenium, tellurium and lead mite. The platinum mineralisation is associated are also recovered from some deposits. with two layers: a 30–60-m zone of dissemi- The dominant ore minerals are sulfides, nated mineralisation in the Lower Sulfide Zone pyrrhotite, pentlandite and chalcopyrite, which (LSZ); and a 2–8-m PGM-rich interval in the generally constitute more than ten per cent by Main Sulfide Zone (MSZ). Currently only the volume of the host rock. Nickel grades typically MSZ is economic to exploit. The Hartley range from 0.5–3.0% Ni, with attendant copper Complex, one of five zones within the Great in the range 0.2–2.0%. PGM contents vary widely Dyke, is by far the largest and economically from a few ppb up to, exceptionally, 10 g/t. The most significant and contains approximately 80 size of the deposits ranges from a few hundred per cent of its PGM resources. thousand tonnes of ore up to a few tens of million The Stillwater Complex (2700 Ma) in Montana, tonnes. Globally two nickel–copper districts are USA, is another of the world’s major layered predominant, Sudbury in Ontario, Canada and intrusions that hosts important PGM mineralisa- Norilsk-Talnakh in the polar region of Russia, tion (Zientek et al., 2002). The PGM-bearing J-M with each containing more than ten million Reef can be traced for 45 km along strike and tonnes of nickel metal. approximately 1.6 km down-dip. The reef con- Magmatic sulfide deposits occur in diverse geo- sists of a 1–3-m thick pegmatitic peridotite and logical settings in rocks ranging in age from troctolite containing sparsely disseminated sul- Archaean to Permo-Triassic. Eckstrand and Hulbert fides (0.5–1 volume per cent) and platinum-group (2007) recognised four principal classes: minerals. In contrast to deposits in the Bushveld 1. A meteorite-impact mafic melt with basal sul- Complex and the Great Dyke, the J-M Reef is fide ores. Sudbury is the only known example. relatively enriched in palladium with a Pt/Pd 2. Rift- and continental-flood basalt, with associ- ratio of about 0.3. The reef contains an average of ated dykes and sills. Important examples include 20–25 g/t Pd + Pt over a thickness of about two the Norilsk-Talnakh district in Russia and metres. Most of the palladium is contained Jinchuan in China. within pentlandite, while the platinum occurs 3. Komatiitic (magnesium-rich) volcanic flows mainly in moncheite, braggite, cooperite and and related intrusions. Important examples Pt–Fe alloy. include Kambalda in Australia, Thompson and The Lac des Iles Intrusive Complex (2700 Ma) Raglan in Canada and Pechenga in north-west comprises a suite of mafic-ultramafic rocks situ- Russia. ated approximately 85 km north of Thunder 4. Other mafic-ultramafic intrusions. Although Bay, Ontario, Canada. The deposit, mined deposits of this class contain significant resources of chiefly for palladium, is similar in some respects nickel and copper (e.g. Voisey’s Bay, Labrador, to contact-type PGM mineralisation, but the Canada), their PGM contents are generally very low. Platinum-group metals 293

The origin of these deposits involved deriva- world’s largest palladium producer and the second tion of the metals from primitive magmas which largest platinum producer (Johnson Matthey, ascended to high levels in the crust along major 2012). The resource in the region has been esti- faults where they were contaminated by sulfur- mated at more than 1.3 billion tonnes of ore bearing sedimentary rocks, such as evaporites or averaging 1.77% Ni, 3.57% Cu, 0.06% Co, 1.84 g/t carbonaceous shales. This led to segregation of an Pt and 7.31 g/t Pd (Naldrett, 2004). The Norilsk– immiscible sulfide liquid that settled to the lower Talnakh deposits are associated with the huge part of the magma chamber scavenging chalco- Permian flood basalt suite known as the Siberian phile elements such as nickel, copper and PGM. Traps, which developed in major intracontinental Where these sulfide droplets became sufficiently rift zones. The nickel–copper–PGM ores occur as concentrated they formed a magmatic sulfide flat-lying sheets at the base of gently dipping gab- deposit, either in a basal or marginal setting broic sills, in the underlying sedimentary rocks (Naldrett, 2010). and in proximity to the upper contacts of the sills. Another major deposit of this type is Jinchuan Deposits related to meteorite impact in China which contains more than 500 million tonnes of ore with an average grade of 1.1 wt% Ni The origin of the nickel–copper–PGM deposits of and 0.7 wt% Cu (Su et al., 2008). It has been mined the Sudbury Igneous Complex (SIC) is related to a for nickel and copper since 1964, although in meteorite impact at 1850 Ma (Lightfoot et al., recent years PGM production has become increas- 2001). The impact crater was originally about ingly important and Jinchuan currently accounts 200 km in diameter with the mineralisation for about 95 per cent of China’s PGM production. located at the base and in the footwall of the impact melt. Following later deformation the SIC now occupies an elongate basin 65 km long and Deposits related to komatiitic rocks 27 km wide. The nickel deposits, which are pre- Nickel–copper deposits with low PGM con- sent on both the north and south flanks of the tents associated with komatiitic volcanic flows basin, are hosted by a basal mafic noritic unit and sills are widespread in Archaean and containing abundant fragmental material, Palaeoproterozoic terranes in Australia, Canada, referred to as the ‘sublayer’. Radiating and con- Brazil, Finland and Zimbabwe (Lesher and Keays, centric fracture and breccias zones, known as ‘off- 2002). Both small high-grade deposits (1.5–4.0% sets’, in the underlying footwall locally host Ni) and much larger low-grade deposits (ca. 0.6% sulfide-rich orebodies enriched in copper and Ni) have been mined at many localities in the PGM at depths up to 400 m below the sublayer. Eastern Goldfields of the Archaean Yilgarn The sulfide ores have typical magmatic sulfide Craton of Western Australia (Hoatson et al., mineralogy and textures. A wide range of plati- 2006). Other important komatiite-associated num-group minerals is present of which the most deposits which are mined for nickel and abundant are tellurides (michenerite and mon- by-product PGM include the Thompson Nickel cheite) and arsenides (sperrylite). Belt of Manitoba, Canada and the Pechenga dis- trict of the Kola Peninsula in Russia. Deposits related to rift- and continental-flood basalts Other deposit types The Norilsk–Talnakh district in the Taimyr Peninsula of Russia is the world’s largest producer High PGM concentrations are known in several of nickel. The ores here are also very rich in PGM other geological environments. Although very and consequently this area is the source of the high grades may occur locally, most deposits are vast majority of Russia’s PGM and is also the small and mining of PGM is not currently 294 gus gunn economic. Alluvial placer deposits of PGM, underground. This method typically involves mainly related to Alaskan-type intrusions, were removing the overburden, digging the ore or worked for many decades in Colombia, Alaska, blasting with explosives, then removing the ore British Columbia, the Urals and central New by truck or conveyor belt for further processing. South Wales (Tolstykh et al., 2005). They were the world’s principal source of platinum until the Processing discoveries at Sudbury and in the Bushveld After mining, the ores are processed to increase Complex, but now significant PGM production their PGM content. Concentration is normally from alluvial deposits is restricted to the Russian carried out at, or close to, the mine site and Far East (Johnson Matthey, 2012). involves crushing the ore and separating platinum- bearing and gangue minerals, using a range of physical and chemical processes. The concentrate Extraction and processing may be transported further afield for smelting and refining to produce pure metals. Different proce- Extraction methods dures are used for processing sulfide-poor ores (e.g. Merensky and UG2) and sulfide-dominant ores The method used to mine PGM-bearing deposits (e.g. Norilsk) due to their contrasting chemical, is dependent on their size, grade, morphology and mineralogical and physical properties (Cole and the value of any co-products. Underground extrac- Ferron, 2002). tion uses a variety of standard mining methods The processing flowsheets for PGM-dominant depending on the characteristics of the orebody. and nickel–copper-dominant ores are shown The deepest currently operating platinum mine is schematically in Figure 12.3. Zondereinde, owned by Northam Platinum Ltd, on the western limb of the Bushveld Complex where mining takes place at a maximum depth of 2.2 km. PGM-dominant ore Due to the high temperatures at this depth, ‘hydro- Comminution and concentration power’ is now widely used in the deep-level mines on the Bushveld. This technology, pioneered by Crushing, followed by fine grinding (milling), is Northam Platinum, uses water, cooled on the sur- used to facilitate separation of the ore minerals face to 5 °C, to power its drilling machines and cool from the gangue (Merkle and McKenzie, 2002). working places. Typically, these underground oper- Some mines employ additional processing stages ations use labour-intensive drilling and blasting after comminution to optimise PGM recovery. techniques, though increased mechanisation is Where the PGM occur within the crystal struc- being introduced into the workplace. Ore is blasted ture of pyrrhotite, then pyrrhotite may be mag- using explosives then transported to the shaft using netically separated, collected and treated an underground rail system. Old workings are separately. Dense media separation may be used backfilled with waste material to improve ventila- to remove lighter silicate minerals from the tion by forcing air to travel through only those denser chromite and platinum-group minerals areas that are being worked, as well as providing after initial crushing and prior to milling. This more roof support. Underground mining is also significantly reduces the volume of ore to be employed in the Norilsk-Talnakh, Sudbury and milled and further treated without reducing PGM Stillwater deposits. recoveries. Surface mines are generally cheaper and safer Froth flotation normally follows milling. In to operate than underground mines. Open-pit this process, water is added to the powdered ore mining is used where the ore is near-surface (typ- to produce a suspension and air is blown upwards ically less than 100 m) and for lower-grade ore through the tanks. Chemicals are added to the bodies, which would not be economic to mine mix, making some minerals water-repellent and Platinum-group metals 295

Ore

Crushing & grinding

Froth flotation

PGM and base metal concentrates

Roasting, smelting, converting Sulfur dioxide (used for sulfuric acid)

PGM-bearing nickel-copper matte

Matte from PGM-dominant ores Matte from nickel and copper dominant ores

Copper solution Reverse leaching Wet grinding Nickel solution PGM concentrate Cobalt concentrates Pressure oxidisation leaching Nickel concentrates Selective solvent extraction Copper concentrate containing PGM

Platinum Palladium Gold, silver and other PGM Smelting

Copper blister containing PGM

Electrolysis Copper cathodes

PGM-bearing anode slimes

Smelting

PGM-bearing matte

Selective solvent extraction

Platinum Rhodium Gold, silver and and and other PGM palladium ruthenium

Figure 12.3 Schematic flowsheet for the processing of PGM-dominant and nickel–copper-dominant ores.

causing air bubbles to stick to their surfaces. rich (up to about three volume per cent base-metal These minerals collect in the surface froth and sulfides), while UG2 ore generally comprises are removed as a metal concentrate. 60–90 per cent chromite. Many operations employ UG2 and Merensky ores are normally treated two stages of milling and floating conducted in separately on account of their contrasting and var- series, known as mill-float- mill-float (MF2) opera- iable mineralogy. Merensky ore is the more sulfide tions, in order to maximise PGM recovery. 296 gus gunn

alloy. After separation from the slag the iron is Smelting removed from this alloy prior to PGM refining. Matte-smelting of the flotation concentrate pro- ConRoast technology ensures that chromium is duces a silicate melt (slag) from which an immis- dissolved in the slag, thereby allowing chromite- cible sulfide melt (matte) containing the PGM rich materials not amenable to the traditional separates on account of its greater density (Jones, matte-smelting process to be treated (Jones, 2009). 2005). Matte smelting takes places in electric furnaces at temperatures of about 1350 °C, although Refining higher temperatures may be needed for UG2 concentrates. Fluxes such as limestone are added Refining to produce high-purity platinum products to the smelter to reduce the melting temperature is a lengthy and complex process the details of of the platinum and base-metal sulfides which which are commonly closely guarded commercial accumulate in the matte. UG2 ore, which is low in secrets. However, in general, the process consists sulfur and high in chromite, is usually blended with of a series of hydrometallurgical operations. Merensky concentrate to ensure effective smelting. Following removal of most of the copper, nickel, The percentage of chromite in the smelted ore must cobalt and iron by reverse leaching in the base- be carefully managed to minimise PGM losses. metal refinery, the PGM-bearing residue, contain- Once the matte has been tapped off, the liquid ing over 60 per cent PGM, is transferred to the metal undergoes a process known as converting. precious metal refinery for separation and purifica- This involves blowing air, or oxygen, into the tion of the PGM. This is achieved either through a matte for several hours to oxidise contained iron series of dissolution and precipitation stages and sulfur. Silica is added to the matte to react involving PGM salts, or by using a technique with the oxidised iron to form a slag that can be known as total leaching, which is followed by easily removed, while the sulfur is collected to sequential metal separation with the aid of solvent produce sulfuric acid. The converter matte con- extraction. sists of copper and nickel sulfide with smaller quantities of iron sulfides, cobalt and PGM. This Nickel–copper-dominant ore is usually cast into ingots which are processed in the base metal refinery. However, Anglo American In recent years considerable research has been Platinum, the largest platinum producer in the undertaken to improve PGM recovery from these world, uses a different slow-cooling process for ores which previously were regarded chiefly as the recovery of PGM. In this method, about 95 per sources of nickel. Several different processing cent of the PGM are concentrated in a small routes are utilised to optimise the treatment of volume of copper–nickel–iron alloy, which is sep- ores with different sulfide contents and textures. arated from the matte by magnetic methods after At Norilsk, the ore is crushed and milled before crushing. This alloy is sent directly to the pre- undergoing gravity concentration to recover up to cious metal refinery, without passing through the 40 per cent of the PGM. Froth flotation is then used base-metals refinery, thereby reducing processing to produce a metal concentrate which is sent to the times considerably. smelter where it is roasted, smelted and converted A potentially important new PGM smelting to produce a copper–nickel–PGM matte, together process, known as ConRoast, has been developed with slag as a waste product. The matte is granu- by Mintek (Jones, 2005). In this process the sulfur lated, wet-ground and then treated in the base- is removed prior to smelting by roasting the sul- metal refinery where pressure oxidation leaching is fide concentrates in a fluidised bed. The derived used to produce concentrates of nickel, copper and sulfur dioxide is efficiently collected and the cobalt. The copper concentrate, which includes sulfur-free product is then smelted in a DC arc almost all the PGM and gold, is then smelted in furnace in which the PGM are collected in an iron furnaces to produce copper blister (98.5% to 99.5% Platinum-group metals 297

Cu) with PGM and gold. The copper is then further catalysts. Other important properties that make refined using electrowinning to produce pure PGM useful for many additional purposes include cathode copper with the PGM remaining in the their high strength, high melting point and resis- anode slimes. The anode slimes are then combined tance to corrosion. with other slimes produced from the production of Because of the wide range of applications and nickel, and with the gravity concentrate produced the differences in the contributions of recycling after the initial crushing and milling. This mixture to PGM supply between sectors it is important to is smelted again to produce a PGM-bearing matte. distinguish between net and gross demand for The matte is then pressure-leached to produce a PGM. Net demand is defined as gross demand silver concentrate and two PGM concentrates, one less the amount of PGM recovered from recy- containing platinum and palladium and the second cling. This means that in applications where the other PGM. recycling rates are very low then net demand is Treatment of the matte produced by smelting to approximately equal to gross demand. Conversely, extract the PGM is carried out in various ways. At where recycling rates are high then net demand is Vale’s Sudbury operations, the matte is treated by small. The main sectors in which PGM are used flotation and magnetic separation to produce sepa- and the net demand in each are shown in rate nickel, copper and precious-metal concen- Table 12.5 and Figure 12.4. trates. The precious-metal concentrates are treated in a two-stage pressure leaching process to dissolve Uses of platinum, palladium first the nickel and cobalt, and then the copper, and rhodium selenium and tellurium. This concentrate, con- taining 60–80% PGM + gold, together with small The largest net use of platinum, palladium and quantities of silver and base metals, is then refined rhodium is in autocatalysts, which are used to to produce high-purity individual elements. convert noxious emissions (carbon monoxide, The copper concentrates derived from many oxides of nitrogen and hydrocarbons) from car copper ores may also contain small quantities of exhaust systems to harmless non-toxic products. PGM. In processing these ores the PGM will be Demand for PGM in autocatalysts has grown present, together with gold and silver, in the markedly from the 1970s onwards as many coun- anode slimes at the copper refinery. Most large tries followed the lead of USA and Japan and copper refineries therefore operate a precious- introduced legislation setting standards for emis- metals recovery plant to treat these slimes. sions from motor vehicles. The quantities and proportions of platinum and palladium used in autocatalysts have varied Specifications and uses considerably over time in response to technolog- ical changes and price variations. Currently palla- PGM are sold in many forms including pure dium is the major constituent of catalysts used in metal and a wide variety of compounds, solutions gasoline-powered vehicles while platinum is still and fabricated products. These are produced by a the chief active component in catalysts and small number of specialist companies including particulate filters fitted to diesel-powered vehi- Johnson Matthey, BASF, Umicore, Heraeus and cles. Consequently, this is the main contribution Tanaka Kikinzoku. to platinum demand in Europe where more than Demand for PGM has grown rapidly in recent 50 per cent of all new cars have a diesel engine. years, with annual mine production doubling since Although global demand for rhodium in 2011 1980 and increasing by an order of magnitude since (906,000 troy ounces gross or 28.2 tonnes) was 1960. The catalytic properties of the PGM are the little more than ten per cent of that for palladium basis of their most important applications in (Johnson Matthey, 2012), rhodium is a critical emission control systems and in industrial process component of autocatalysts which account for 298 gus gunn

Table 12.5 The main uses of platinum and palladium, and global consumption by sector in 2011. (Data from Johnson Matthey, 2012.)

Platinum Palladium

Consumption Consumption Sector Main uses (thousand troy ounces) Main uses (thousand troy ounces)

Autocatalyst Catalysts for vehicle exhaust 3105 Catalysts for vehicle exhaust 6030 emission control. emission control. Investment Exchange traded funds, ingots, 460 Exchange traded funds, (565)* bars, coins. coins. Jewellery Fabrication of of platinum 2480 Fabrication of palladium 505 jewellery. jewellery, white gold. Chemical Catalysts for production of nitric 470 Catalysts for production of 445 acid and other bulk and nitric acid and other bulk speciality chemicals. and speciality chemicals. Electrical Hard disk coatings, 230 Multi-layer ceramic chip 1380 thermocouples. capacitors, hybrid integrated circuits, plating. Petroleum Catalysts for petroleum refining 210 — — and production of petrochemicals. Glass Vessels for glass manufacture, 555 — — bushings for fibre glass production. Dental, medical Anti-cancer drugs, implants such 230 Alloying agent, mainly 550 and biomedical as heart pace-makers. with minor platinum, gold, silver. Other Spark plug tips, oxygen sensors 355 Emission control catalysts 105 for engine management, in stationary applications. fuel cells. Total global consumption 8095 Total global consumption 8450

1000 troy ounces is approximately 31.1 kilograms. *in 2011 palladium investment was negative and 565,000 ounces were supplied back to the market. about 80 per cent of its use. Other areas where and palladium through exchange-traded funds rhodium is important include glass manufacture (ETFs) which were launched in Europe in 2007 and the chemicals sector. and in USA in 2010. Total net holdings of Platinum has long been widely used in jewel- platinum and palladium in ETFs at the end of lery, especially in Japan, and in recent years it has 2011 were estimated to be 1.32 million troy become increasingly popular in Europe and ounces of platinum and 1.7 million troy ounces China. In 2011 jewellery was the second most of palladium (CPM Group, 2012). important use of platinum accounting for about Platinum and palladium are widely used in cata- 31 per cent of total net consumption. lysts in the production of chemicals and pharma- Like gold and silver, there is a demand for ceuticals. However, net PGM demand for these platinum as a safe haven for storing wealth. purposes is relatively small because the catalysts are Various platinum coins, ingots and bars have recycled with a very high degree of efficiency and been sold over the past thirty years, but, today, new supply is needed only to replace catalysts at the investors are increasingly purchasing platinum end of their useful lifetime or when additional Platinum-group metals 299

Platinum Palladium

3% 4% 6% 1% 3% 15%

30% 38%

6%

67% 5%

6% 6% 7% 3%

Autocatalyst ChemicalDental Electrical Glass

Investment Jewellery Medical Petroleum Other

Figure 12.4 The main applications of platinum and palladium in 2011. (Data from Johnson Matthey, 2012.)

capacity is installed. Most important is the use of Platinum and platinum–rhodium alloys are platinum in the conversion of ammonia to nitric used in equipment that hold, channel and form acid for the manufacture of fertilisers and explo- molten glass. These materials are able to withstand sives. Platinum-based catalysts are also indispens- the abrasive action of molten glass and do not react able to the refining of petroleum and the production with it at high temperatures. Palladium and, to a of petrochemical feedstocks used in the manufac- much lesser extent, platinum are also important ture of plastics, synthetic rubber and polyester components of dental alloys used in inlays, bridges fibres. Some palladium is also used in petrochemical and crowns. On account of its chemical inertness, catalysts but the main use is in fine chemistry (i.e. high electrical conductivity and its ability to be the manufacture of vitamins, dyes, antibiotics, etc.) fabricated into very small complex components, and in pharmaceuticals (Hagelüken, 2008). platinum is an important component of various The electronics industry is the second largest biomedical devices that are implanted in the body. consumer of palladium where it enjoys a Platinum electrodes are used in heart pacemakers significant cost advantage over rivals such as and in implantable cardioverter defibrillators (ICD) gold. A large proportion of this palladium is used used in the treatment of irregular heartbeat. in multi-layer ceramic capacitors (MLCC), while The ability of platinum, in certain chemical other important applications are in hybrid forms, to inhibit cell division is the basis of its use integrated circuits, used mainly in the automo- in a range of anti-cancer drugs. The first of these, tive sector, and in plating of connectors inside cisplatin, was introduced in 1977 and is widely computers. Platinum is also used as a component used in the treatment of testicular, ovarian, of the magnetic coating on computer hard disks bladder and lung cancers. Several other platinum- for which demand has escalated in recent years. based drugs are now in use or undergoing testing. 300 gus gunn

developing countries and use of inappropriate Uses of ruthenium, iridium and osmium treatment technology (UNEP, 2011). The main use of ruthenium is in perpendicular It is important to distinguish ‘closed-loop’ and magnetic recording technology which is employed ‘open-loop’ applications as their differences fun- in hard disk drive manufacture. Ruthenium is also damentally impact on PGM recycling rates used as a catalyst in the chemical and electro- (GFMS, 2005). In closed-loop applications owner- chemical industries. Iridium is used in the ship of the metal effectively stays in the same chemical industry, for crucibles in the electronics hands throughout its life cycle and it is continu- sector and in spark plug tips. Osmium is used in ally recycled and reused in the same application small amounts in an alloy with platinum and with very transparent flows, ensuring high rates iridium, which is used in pen tips, electrical con- of recycling. In contrast, in open-loop applica- tacts, filaments in light bulbs and in medical tions, such as autocatalysts and electronics, the implants. flow of end-of-life products is less clear: recycled and refined PGM are sold back into the market, supplementing primary supply but with less Recycling, re-use and resource efficiency assurance that they will be recycled when the products in which they are used reach the end of The durability of PGM in use, combined with their their lives. recent high prices, has ensured a growing interest In some industrial applications closed-loop in using them efficiently and recycling them when- recycling is the norm. For example, it is estimated ever possible. Furthermore, PGM production from that 95–98 per cent of PGM used in glass manu- secondary sources, such as scrap and end-of-life facture are recycled and re-used in glass making products, reduces the need to mine primary ores (Johnson Matthey, 2011). High recycling rates are thereby contributing to improved supply security. also achieved in the manufacture of chemicals Recycling also produces less carbon dioxide than and in oil refining where the spent catalysts are mining and processing primary ores. returned directly to a recycling company. In 2010 the proportions of PGM supply from For many consumer applications PGM recy- recycling increased as economies began to recover cling rates are much lower due mainly to ineffi- from the global economic recession and PGM cient collection and losses during product prices rose. For platinum, approximately 23 per dismantling. Although the technology required to cent (1.84 million troy ounces) was derived from handle these materials is both technically efficient recycling, while the corresponding figures for pal- (more than 95 per cent recovery) and environmen- ladium and rhodium were 19 per cent (1.85 mil- tally friendly, it is highly specialised and not lion troy ounces) and 27 per cent (236,000 troy widely available. UNEP (2011) reports that end-of- ounces), respectively, (Johnson Matthey, 2011). life recycling rates for platinum and palladium are The largest contributions came from autocata- 80–90 per cent in industrial applications, but, lysts, stimulated by national car-scrappage where the products do not enter an appropriate schemes, from platinum jewellery and from the closed-loop recycling chain, they are much lower recovery of palladium from electronic scrap. In at about 50–55 per cent in vehicle applications and the EU the Waste Electrical and Electronic less than ten per cent in electronic goods. Equipment (WEEE) Directive (2003), which is In some uses, such as medical and biomedical, intended to promote the recovery of electrical spark plugs, sensors and hard disks PGM are scrap, has had a positive impact on PGM recy- effectively dissipated in use and not available for cling. However, only a small proportion of the recycling. In contrast, recycling rates for PGM in recycling potential of PGM from this source is jewellery and investment items are very high. currently achieved in Europe because of low col- Overall, if gross demand and recycling rates in lection rates, high exports of old equipment to closed-loop applications (chemicals, oil refining Platinum-group metals 301 and the glass industry), which are not normally The greatly increased use of PGM in recent included in published statistics, are considered years has raised the level of PGM emissions into recycling contributes more than 50 per cent of the environment. The largest contribution has total PGM supply. been from vehicle emissions following the intro- duction of PGM-bearing autocatalysts from the 1970s onwards. High PGM concentrations have Substitution been recorded in dust, silt, soils and waters close to major highways and in urban areas (Ravindra In many applications one PGM can substitute for et al., 2004). Furthermore, complexation of PGM another. This has happened in the past in response derived from autocatalysts has been shown to to price differentials between palladium and give rise to mobile species with possible associ- platinum. For example, in recent years palladium ated increase in bioavailability (Colombo et al., has been increasingly used in three-way catalysts 2008). Several studies suggest a variation between for petrol cars with manufacturers also moving to individual PGM in environmental mobility and increase the proportion of palladium to platinum uptake rate in plants and animals, with palla- in diesel catalysts which formerly used platinum dium greater than rhodium, which is, in turn, as the sole catalytic metal. greater than platinum (Ek et al., 2004). Further In response to the price spikes in 2001 there research is required to provide a better under- was also significant substitution of palladium in standing of PGM emissions from autocatalysts, MLCC electronics applications by nickel and their dispersion and bioavailability and their copper, albeit with some reduction in performance effects on animals and plants. However, it is level. Consequently, in many high-tech applica- important to note that continual improvements tions palladium remains the material of choice. to the design of catalytic converters and vehicle The prevailing high prices have stimulated exhaust systems have led to significant reduction research in thrifting of the PGM, i.e. using less of specific PGM emissions, provided fuel of material in a particular application with little or appropriate quality is used and vehicles are well no reduction in performance. For example, as tech- maintained. nology has advanced autocatalysts have become In general, the levels of PGM in the environ- more efficient and smaller quantities of PGM are ment are so low that health effects are limited. required to achieve the same performance. However, the allergic effects of platinum salts, However, at the same time, emission standards termed platinosis, to occupationally exposed around the world have become increasingly strin- workers, such as those employed in PGM refin- gent and so the amounts used have remained eries, are well known. The symptoms include nearly constant. There is also considerable interest irritation of the nose and upper respiratory tract, in reducing the amount of PGM used in fuel cells sometimes accompanied by asthmatic symp- in order to reduce unit costs and stimulate uptake toms. Some cases of contact dermatitis due to of this technology (Johnson Matthey, 2011). palladium allergy have been reported, while sen- sitisation to palladium is also a concern to miners, dental technicians and chemical workers Environmental issues (Kielhorn et al., 2002). Further research is required into the occupational health effects on platinum In metallic form the PGM are generally regarded miners and workers involved in the manufacture as inert, non-toxic and non-allergenic. However, and recycling of autocatalysts. some of their compounds, particularly certain The main environmental issues associated PGM-chlorinated salts, are both highly toxic and with the mining and processing of sulfide-bearing allergenic, and DNA damage due to PGM ores of any type, including PGM ores, include: the exposure has been reported (Gagnon et al., 2006). generation of acid mine drainage from mine work- 302 gus gunn ings, ore dumps, treatment plant and tailings; significant potential to mine at greater depths and release of associated trace elements which may be there are several other geological units that could harmful to human health and the environment be exploited for PGM if market conditions were (e.g. arsenic, lead, antimony); discharge of chemi- favourable. There are also considerable PGM cals used in the mining and processing of the ores resources in other countries, notably in the Great and concentrates; and gaseous emissions (chiefly Dyke in Zimbabwe, the Stillwater Complex in the sulfur dioxide) associated with the smelting and USA, and in nickel–copper deposits, for example refining processes. The South African government in the Norilsk-Talnakh district of Russia and the has begun to implement a programme of increasing Sudbury area of Canada. environmental regulation aimed at delivering The rise in the price of platinum over the last 20 improved environmental performance, including years has led to increased levels of exploration, both more stringent controls on air and water quality on new targets and on several deposits previously and emissions of greenhouse gases. The platinum thought to be uneconomic. The latter category mining industry in South Africa has been proac- includes the Skaergaard Complex in Greenland, the tive in its environmental management and report- Duluth Complex in the USA, and the Penikat and ing of sustainability issues including water and Keivitsa complexes in Finland. At the Marathon energy utilisation and greenhouse gas emissions copper–PGM deposit in western Ontario the (Mudd, 2012). Environmental management sys- construction of a new open-pit mine and processing tems operated by the major companies, certified operation is planned, which, subject to permitting, against international standards, share common could produce up to 200,000 troy ounces of PGM goals. These are aimed at minimising pollution of annually (Marathon PGM Corporation, 2011). all types, promoting efficient use of energy, Proven and probable ore reserves in the reducing waste, and minimising the footprint on Norilsk-Talnakh area are about 55 million troy the land, including effective remediation and reha- ounces of palladium and 13 million troy ounces of bilitation at closure. platinum (Norilsk Nickel, 2010). In the Stillwater Complex in Montana, USA, proven and probable reserve of approximately 19.9 million troy ounces combined palladium and platinum have been World resources and production reported (Stillwater Mining Company, 2011). Resources and reserves Production The Merensky, UG2 and Platreef in South Africa dominate world platinum resources. It is difficult In 2010 there was a modest increase in global to estimate the total quantity of platinum in the mine production of platinum (total in 2010, Bushveld Complex but a recent study suggested 193,000 kg or 6.8 million troy ounces) (Figure 12.5). that the proven and probable reserves of PGM In South Africa there had been six years of steady (platinum, palladium and gold) exceed 250 million growth from 2001 in response to increased troy ounces (Cawthorn, 2010). This amount alone demand for autocatalysts until it fell back in 2007 is adequate to meet global demand at current and 2008 with onset of the global economic reces- levels for several decades with existing technology sion. In 2010 South African platinum production and current prices. However, the true availability was at the same level as in 2009. is almost certainly very much greater because South Africa is the largest platinum producer, companies only report reserves that are required to accounting for about 76 per cent of world produc- underpin their short-and medium-term plans. tion in 2010 (Figure 12.6). Production is domi- Cawthorn (2010) suggests that there is about 800 nated by two companies, Anglo American million troy ounces in the upper two kilometres of Platinum and Impala Platinum, who together the Bushveld Complex. In the longer term there is control about two thirds of global production of Platinum-group metals 303

300

250

200

150

100

Thousand kilograms (metal content) 50

0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Platinum Palladium Other PGM

Figure 12.5 World mine production of platinum, palladium and other PGM, 1992–2010. (Data from British Geological Survey World Mineral Statistics database.) (1 kilogram equals 32.151 troy ounces.)

350

300

250

200

150

100

50 Thousand kilograms (metal content)

0 South Africa Russia Zimbabwe USA Canada Other Other PGM 57,292 2200 1550 0 400 0 Palladium 82,222 115,700 6968 11,600 6200 4205 Platinum 147,790 26,500 8703 3450 3600 3076

Figure 12.6 World mine production of platinum, palladium and other PGM in 2010 by country. (Data from British Geological Survey, 2012.) (1 kilogram equals 32.151 troy ounces.) 304 gus gunn platinum. Russia is the other major supplier of costs now account for more than 50 per cent of platinum, with minor production from operating costs), currency fluctuations, strikes, Zimbabwe, Canada (Lac des Iles mine and safety stoppages, shortages of skilled labour and Sudbury), USA (from two mines in Montana) and the rising prices of fuel, steel and other raw mate- Australia (as a by-product of nickel mining). rials. At the same time electricity supplies from Relatively small amounts of platinum are also Eskom, the South African state power utility, have produced at Jinchuan in China, where PGM are become increasingly unreliable and expensive. also extracted from ores mined in Australia, Platinum and palladium supplies from Spain and Zambia. Zimbabwe have increased markedly in recent Over the last decade mine production of palla- years with expansion of mining and concen- dium has declined (Figure 12.5) and levels remain trating capacity at three locations on the Great significantly lower than those recorded before the Dyke by South African-owned companies. A new recession. There was a sharp fall in 2002 as a mine built by Anglo American Platinum at Unki result of the high price, leading to substitution of East was commissioned in 2011 and is expected palladium by platinum in autocatalysts. However, to contribute 60,000 troy ounces of platinum per as the price differential was reversed and vehicle annum from 2012. However, recent government sales continued to rise, so manufacturers again legislation requiring the transfer of majority own- began to use more palladium in autocatalysts. ership to indigenous investors has cast a shadow Production of palladium is dominated by over possible further expansion of the PGM Russia and South Africa (Figure 12.6). In Russia, industry in Zimbabwe. PGM are by-products of nickel and copper min- ing, with Norilsk Nickel the largest global producer accounting for more than 40 per cent of World trade mine supply. World palladium mine production increased in 2010 to 227,000 kg (approximately 8 Platinum is traded in many forms, generally in million troy ounces), with higher output from the either unwrought (ores, concentrates, sponge, main producing countries and from smaller pro- powders) or in manufactured or semi-manufac- ducers such as Zimbabwe and Botswana. tured forms (e.g. ingots, wire, mesh). Platinum Following the economic recession in 2008 and trade is complex as significant quantities of 2009 PGM production in South Africa began to unwrought materials are processed in countries recover in 2010 with mothballed operations com- other than those in which they have been pro- ing back on stream, improved performance at some duced. For example, precious-metal concentrates operations and several major new mines and mine- from Canada are processed in the UK and Norway expansion projects well advanced. However, mine to produce high-purity PGM. production decreased in 2011 by about three per Imports of platinum and palladium are domi- cent (120,000 troy ounces), chiefly due to stoppages nated by the leading industrialised nations, such imposed by the government’s Department of as Japan, Switzerland, USA, Germany and UK, Mineral Resources as a result of increased moni- and the major emerging economies of China and, toring and enforcement of safety conditions at the for platinum only, India (Figure 12.7). The majority mines (Johnson Matthey, 2012). South Africa’s of imported metal is used to make catalytic con- platinum mining industry continues to face these verters for vehicles, where the market has grown and other major challenges to ensure its future rapidly in recent years, especially in China. profitability and growth. In recent years operating However, significant proportions are also used in costs have escalated, increasing from under US$300 electronics, jewellery and for investment. per troy ounce in 2003 to just above US$800 per Exports of platinum are dominated by South troy ounce in 2010 (CPM Group, 2011). This is due Africa, which is also the world’s leading producer to a range of factors including rising wages (labour (Figure 12.8). South Africa exported about 200

180

160

140

120

100

80 Thousand kilograms 60

40

20

0

India Italy Japan Germany Singapore Switzerland United States United Kingdom Korea (Rep. of) Other countries

China (inc Hong Kong) Platinum Unwrought Palladium Unwrought Other PGM Unwrought

Figure 12.7 Major importers of platinum metal, concentrates and intermediate products, 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) (1 kilogram equals 32.151 troy ounces.)

200

180

160

140

120

100

80

Thousand kilograms 60

40

20

0

Japan Germany Thailand Belgium South Africa Switzerland United States United Kingdom Korea (Rep. of) Other countries Russian Federation Platinum Unwrought Palladium Unwrought Other PGM Unwrought

Figure 12.8 Major exporters of platinum metal, concentrates and intermediate products, 2009. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) (1 kilogram equals 32.151 troy ounces.) 306 gus gunn

180,000 kg of platinum in 2009, more than three quarter of the year. During 2009 and 2010 the times the quantity from the next biggest exporter platinum price recovered strongly, peaking at Switzerland, which is not a primary producer. US$1500 per troy ounce in December 2010, in Other leading exporters include the UK, USA, response to improved industrial and autocatalyst and Germany. However, of these countries, only demand and increased investor interest. The the USA is a primary producer. platinum price continued to trade at high levels, Russia is the leading exporter of unwrought between US$1700 and US$1900 in the first eight palladium (ca 54,000 kg in 2009), followed closely months of 2011. However, in late August the price by Switzerland and the UK. USA, South Africa began to decline rapidly due to the slowdown in and Germany are also major palladium exporters. economic growth, especially in Europe, and falling investor confidence. The price recovered briefly in the first quarter of 2012 as a result of concerns Prices over supply due to strikes in South Africa, but it fell back in the second quarter. A number of daily market prices are quoted for The price of palladium peaked at over US$1000 the pure (minimum 99.9 per cent) PGM in US per troy ounce in early 2001 and was almost dou- dollars per troy ounce. These include the European ble the price of platinum at the time (Figure 12.9). free market price, the Johnson Matthey base price This was due to the high level of use by vehicle and the London Platinum and Palladium Market manufacturers in catalytic converters combined prices. In the spot market platinum and palladium with restrictions on exports from Russia. Since are sold for cash and immediate delivery in then palladium has traded mostly between sponge, plate or ingot form. US$200 per troy ounce and US$400 per troy Unlike gold, platinum and palladium are ounce. However, in 2008, mirroring the price important and widely used metals in industry. trend for platinum, there was considerable price Consequently, their prices are predominantly volatility in the first half of the year, followed by determined by normal supply and demand rela- a sharp dip to end the year at just under US$200 tionships, although movements in the gold price per troy ounce. The palladium price rallied do have some effect and periodic speculation steadily in 2009 and showed an even more sub- increases price volatility. Since 2007 Exchange stantial increase in 2010, almost doubling its Traded Funds have also invested heavily in PGM. price to end the year at US$803 per troy ounce, in The price of platinum has experienced periods response to increased investment interest and a of volatility over the last decade (Figure 12.9). resurgence of physical demand chiefly in the Throughout the 1990s, the price hovered around autocatalyst sector. In 2011 the average price was US$400 per troy ounce, approximately US$250 per about 39 per cent higher than in 2010, but by the troy ounce lower than the highest price for end of the year it had fallen back and it traded in platinum in the early 1980s. However, following the range US$600–700 in the first half of 2012. the introduction of emission-control legislation, demand for platinum and palladium in catalytic converters increased markedly contributing to a Outlook substantial increase in the price of platinum from US$420 per troy ounce in August 2001, to just Following many years of increased demand, over US$1500 per troy ounce at the end of 2007. fuelled by new technologies and economic The price reached an all-time record high in March growth, particularly in the developing world, 2008 of US$2276 per troy ounce. However, in the PGM consumption fell in 2008 with the onset of second half of 2008 the price fell back dramatically the global economic recession. However, in 2010 with the onset of the global financial crisis and industrial demand was revitalised, especially in associated massively reduced demand, falling to the automotive sector as vehicle production levels below US$800 per troy ounce during the last exceeded pre-recession levels. In 2011 and the Platinum-group metals 307

2200

2000

1800

1600

1400

1200

1000

US$ per troy ounce 800

600

400

200

0 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2 Q4 Q2Q4 Q2 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012

Platinum Palladium

Figure 12.9 The prices of platinum and palladium between 2000 and end June 2012. (Johnson Matthey base price (unfabricated) US$ per troy ounce, quarterly average prices from Metal Bulletin.) first half of 2012 global economic growth slowed developing countries so demand for PGM in auto- again and the high levels of industrial demand for catalysts will increase significantly. At the same PGM have not been sustained. time increasingly stringent environmental legis- China uses large and increasing quantities of lation is being introduced in many countries to PGM to support its rapidly growing economy, the control vehicle emissions. This will require modernisation of its industries and the spread of increased PGM loadings in autocatalysts and prosperity among its population. China is the hence further raise demand. PGM market growth largest vehicle market in the world and conse- is also expected in emission control systems for quently a major consumer of palladium in its pre- stationary applications, such as diesel generators, dominantly gasoline-powered vehicle fleet. China gas turbines and manufacturing processes. is also the largest market for platinum jewellery Given the continuing global economic uncer- and has major glass and electronics manufac- tainties, the increased demand for platinum and turing sectors. However, China meets only a palladium from the investment sector seen in small proportion of its PGM demand from 2010 may well continue in the short and medium domestic sources (Figure 12.10). In 2009 domestic term. The increased availability of exchange- production accounted for less than six per cent of traded funds (ETFs) makes investment in platinum total platinum demand, with approximately equal and palladium simpler and easier than in the past. proportions from mining and recycling. For palla- Considerable growth in demand for platinum is dium, a larger proportion (almost 32 per cent) was anticipated in the fuel-cell sector as the require- domestically sourced although almost 90 per cent ment for clean energy assumes ever-growing global of this was from recycling. importance. Today fuel-cell manufacturing capacity Per capita vehicle usage in China and India is small and the units are relatively expensive. The remains very low compared with the USA. As largest markets are in the provision of auxiliary or vehicle ownership increases in these and other back-up power in camper vans and boats, in 308 gus gunn

3000

2500

2000

1500

1000 Thousand troy ounces

500

0 2005 2007 2009

Platinum Palladium Demand-supply Demand-supply imbalance imbalance Domestic supply Domestic supply

Figure 12.10 The consumption of platinum and palladium in China. (Data from CPM Group, 2011.) combined heat and power supply for houses and given rise to concerns about supply security businesses, and in materials-handling vehicles, (House of Lords, 1982). However, these concerns such as forklift trucks. The largest markets for subsided following the break-up of the Soviet future use of fuel cells are likely to be light-duty Union in 1991 and the end of the apartheid era in vehicles, although development of the supporting 1994. More recently, the continued concentration hydrogen infrastructure may constrain growth to of PGM production in the same two countries some extent. The potential fuel cell market in and the burgeoning demand has once again given consumer electronics, such as portable computers, rise to worries about the security of future tablet devices and mobile phones, is huge. supplies. Of particular significance for the Demand for platinum and rhodium in glass long-term future of the South African mining manufacturing has been growing rapidly since industry is black empowerment. The Youth 2000, driven largely by new technologies. Further League of the African National Congress has been growth is anticipated in the production of fibre highly vociferous in calling for nationalisation of glass, glass for solar photovoltaic applications the mining industry. Although it is generally and for LCD displays, especially in portable accepted that, in principle, exploitation of South electronic devices. Other growth areas include African mineral wealth should benefit its citi- the use of PGM in medical and biomedical appli- zens, nevertheless it is widely considered that cations and in specialist alloys. For example, any move towards nationalisation would damage ruthenium, together with rhenium, is increas- South Africa’s reputation as a safe destination for ingly being used in nickel-based superalloys in international investment and would threaten the turbine blades for jet engines. future of the industry. The recent and ongoing The high level of concentration of PGM pro- labour disputes affecting South Africa’s mining duction in South Africa and Russia has in the past industry are very likely to have short-term Platinum-group metals 309 impacts on global platinum supply as well as lon- practices have prevailed. The economic and envi- ger-term effects on investor confidence. ronmental benefits of efficiently recovering the The physical availability of PGM resources is PGM resource tied up in vehicles still on the road not likely to be a constraint on economic growth will become increasingly important as the global in any sector in the long term. New supplies from market continues to expand. Hagelüken and both primary (mining) and secondary (recycling) Grehl (2012) have suggested that globally up to sources are likely to be available with existing 2010 there may be as much as 3000 tonnes of technology. In addition to the possibility of min- PGM in the autocatalysts of vehicles being used, ing to greater depths on the Merensky and UG2 equivalent to more than six years of the combined Reefs, alternative cheaper sources are expected to global mine production of platinum, palladium make increasing contributions to PGM supply and rhodium in 2010. However, it is important to from South Africa. These include the development stress that open-loop systems are highly complex of new open-pit mines, including some exploiting with many points at which PGM losses may chromite seams at higher levels in the Bushveld occur. Consequently, improvements in recovery Complex, and the recovery of PGM from chro- rates cannot be expected to be easily achieved and mite mine tailings. Full commercialisation of the such systems are not likely to be adopted widely ConRoast smelting technology will also increase or quickly without legislative intervention. the range of materials that can be smelted, while As demand increases there will be technical also providing important environmental benefits challenges related to mining and processing new through reduced pollution. ore types and lower grade ores, to mining at greater Mine production of PGM from Russia has been depths, and to achieving more efficient recycling. relatively stable in recent years despite falling Given timely and adequate investment in educa- grades. Supplies have been regularly supple- tion and research, especially in South Africa, tech- mented by major sales, of up to a million troy nological innovation will, in the long term, ounces of palladium per annum, from state stocks. continue to ensure sustainable PGM supplies with Market analysts have recently predicted that improved environmental performance throughout supply from Russian state sales is likely to decline the supply chain. However, given the major prob- significantly in the near future, contributing to lems currently faced by the mining industry in possible future shortfalls in palladium supply. South Africa, it is almost inevitable that there will In contrast, if nickel prices remain high, PGM be some impact on global platinum supply, even if supplies from nickel mines in Canada, Australia, only in the short term. PGM resources in other Botswana and elsewhere might be expected to countries will become more attractive to mining increase. In addition, new supply could be derived companies if supplies from South Africa are con- from mining PGM-dominant ores in several coun- strained and if demand from industry, especially tries including Canada, Greenland, Australia and the automobile sector in developing countries, Brazil, although bringing new PGM production continues to grow. capacity on stream is a lengthy and expensive pro- cess. As with other critical metals there may also be increased vertical integration in the PGM supply Acknowledgements chain with industry users investing in mining and processing capacity to ensure security of supply. Gus Gunn publishes with the permission of the An increasing proportion of PGM supply is Executive Director of the British Geological Survey. expected to be derived from secondary sources, stimulated by increased regulation and fiscal incentives from governments to promote recy- Note cling and new initiatives to encourage closed-loop recycling in sectors where hitherto open-loop 1. Ma, million years before present 310 gus gunn

References Journal of Environmental Science and Health, Part A, 41, 397–414. British Geological Survey (2012) World Mineral GFMS. (2005) Materials Flow of Platinum Group Production 2006–2010 (Keyworth, Nottingham: Metals. Umicore, Öko-Institut, GFMS, London. British Geological Survey.) Hagelüken, C.J. (2008) Recycling of spent catalysts. In: Cabri, L.J. (2002) The Platinum-Group Minerals. In: Ertl, G., Knötzinger, H., Schüth, F. and Weitkamp, J. Cabri, L.J. (ed.) The geology, geochemistry, miner- (eds.) Handbook of Heterogeneous Catalysis, 2nd alogy and mineral beneficiation of platinum-group edition, Wiley-VCH, Weinheim, Germany, 1846–1863. elements, Canadian Institute of Mining, Metallurgy Hagelüken, M. and Grehl, M. (2012) Recycling and loop and Petroleum, Special Volume 54. concept for a sustainable usage, In: Sehrt, U. Cawthorn, R.G., Merkle, R.K.W. and Viloen, M.J. and Grehl, C. (eds.) Precious Materials Handbook, (2002) Platinum-group element deposits in the Umicore AG and Co KG, Hanau-Wolfgang, Germany. Bushveld Complex, South Africa. In: Cabri, L.J. (ed.) 38–87. The geology, geochemistry, mineralogy and mineral Hoatson, D.M., Jaireth, S.A. and Jacques, A.L. (2006) beneficiation of platinum-group elements, Canadian Nickel sulfide deposits in Australia: Characteristics, Institute of Mining, Metallurgy and Petroleum, resources and potential. Ore Geology Reviews 29, Special Volume 54. 177–241. Cawthorn, R.G. (2010) The Platinum Group Element House of Lords (1982) Strategic Minerals. 20th report of Deposits of the Bushveld Complex in South Africa. the House of Lords Select Committee on the European Platinum Metals Review 54, 205–215. Communities, Session 1981–82, HMSO. Cole, S. and Ferron, C.J. (2002) A review of the benefici- Johnson Matthey (2011) Platinum 2011. ation and extractive metallurgy of the platinum-group Johnson Matthey (2012) Platinum 2012. elements, highlighting recent process innovations. Jones, R.T. (2005) An overview of South African PGM In: Cabri, L.J. (ed.) The geology, geochemistry, miner- smelting. Nickel and Cobalt 2005: challenges in alogy and mineral beneficiation of platinum-group extraction and production. 44th Annual Conference of elements, Canadian Institute of Mining, Metallurgy Metallurgists, Calgary, Canada. http://www.mintek. and Petroleum, Special Volume 54. co.za/Pyromet/Files/2005JonesPGMsmelting.pdf Colombo, C., Oates, C.J., Monhemius, J. and Plant, J.A. Jones, R.T. (2009) Towards commercialisation of (2008) Complexation of platinum, palladium, and Mintek’s ConRoast process for platinum smelting. rhodium with inorganic ligands in the environment. Nickel and Cobalt 2009: Advances in the processing Geochemistry: Exploration, Environment, Analysis of nickel, cobalt and PGMs using pyrometallurgical 8, 91–101. techniques. 48th Annual Conference of Metallurgists, CPM Group (2011) The CPM Platinum Group Metals Calgary, Canada. http://www.mintek.co.za/Pyromet/ Yearbook 2011. Files/2009Jones-ConRoast.pdf. CPM Group (2012) The CPM Platinum Group Metals Kielhorn, J., Melber, C., Keller, D. and Mangelsdorf, I. Yearbook 2012. (2002) Palladium – a review of exposure and effects to Eckstrand, O.R. and Hulbert, L. (2007) Magmatic human health. International Journal of Hygiene and nickel-copper-platinum group element deposits. In: Environmental Health 205, 417–432. Goodfellow, W.D. (ed.) Mineral Deposits of Canada: Kinnaird, J.A. and MacDonald, I. (2005) An introduction A Synthesis of Major Deposit-Types, District to the mineralisation in the northern limb of the Metallogeny, the Evolution of Geological Provinces, Bushveld Complex. Applied Earth Science (Tran- and Exploration Methods. Geological Association of sactions of the Institution of Mining and Metallurgy) Canada, Mineral Deposits Division, Special 114, B194–198. Publication No. 5, 205–222. Lee, C.A. (1996) A review of mineralization in the Ek, K.H., Morrison, G.M. and Rauch, S. (2004) Bushveld Complex and some other layered mafic Environmental routes for platinum group elements intrusions. In: Cawthorn, R.G. (ed.) Layered Intrusions, to biological materials - a review. Science of the Total Elsevier Science, Amsterdam, 103–146. Environment 334–335, 21–38. Lesher, C.M. and Keays, R.R. (2002) Komatiite-associated Gagnon, Z.E., Newkirk, C. and Hicks, S. (2006) Impact Ni-Cu-PGM deposits: geology, mineralogy and geo- of platinum-group metals on the environment: a tox- chemistry. In: Cabri, L.J. (ed.) The geology, geochemistry, icological, geotoxic and analytical chemistry study. mineralogy and mineral beneficiation of platinum-group Platinum-group metals 311

elements. Canadian Institute of Mining, Metallurgy and Ravindra, K., Bencs, L. and Van Grieken, R. (2004) Petroleum, Special Volume 54. Platinum group elements in the environment and Lightfoot, P.C., Keays, R.R. and Doherty, W. (2001) their health risk. The Science of the Total Environment Chemical evolution and origin of nickel sulfide min- 318, 1–43. eralization in the Sudbury Igneous Complex, Ontario, Scoon, R.N. and Mitchell, A.A. (2004) The platiniferous Canada. Economic Geology 96 1855–1875. dunite pipes in the eastern limb of the Bushveld Maier, W.D. (2005) Platinum-group element (PGM) Complex: Review and comparison with unminer- deposits and occurrences: Mineralisation styles, ge- alised discordant ultramafic bodies. South African netic concepts, and exploration criteria. Journal of Journal of Geology 107, 505–520. African Earth Sciences 41, 165–191. Scott, D.A. and Bray, W. (1980) Ancient platinum tech- Marathon PGM Corporation (2011) Project Overview. nology in South America. Platinum Metals Review http://www.marathonpgmproject.com/Project- 24, 147–157. Overview.html Stillwater Mining Company (2011) Annual report for Merkle, R.K.W. and McKenzie, A.D. (2002) The 2010. geology, geochemistry, mineralogy and mineral Su, S., Li, C., Zhou, M-F., Ripley, E.M. and Qi, L. (2008) beneficiation of PGE. In: Cabri, L.J. (ed.) The Controls on variations of platinum-group element geology, geochemistry, mineralogy and mineral concentrations in the sulphide ores of the Jinchuan beneficiation of platinum-group elements. Canadian Ni-Cu deposit, western China. Mineralium Deposita Institute of Mining, Metallurgy and Petroleum, 43, 609–622. Special Volume 54. Tolstykh, N.D., Sodorov, E.G. and Krivenko, A.P. (2005) Mudd, G.M. (2012) Sustainability reporting and Platinum-group element placers associated with platinum group metals: a global mining industry Ural-Alaska type complexes. In: Mungall, J.E. (ed.) leader. Platinum Metals Review 56, 2–19. Exploration for platinum-group element deposits. Naldrett, A.J. (2004) Magmatic Sulphide Deposits: Geology, Mineralogical Association of Canada, Short Course Geochemistry and Exploration. Springer, Berlin. Series Volume 35, 113–143. Naldrett, A.J. (2010) From the mantle to the bank: the UNEP. (2011) Recycling rates of metals – a status report. life of a Ni-Cu-(PGE) sulphide deposit. South African A Report of the Working Group on the Global Metal Journal of Geology 113, 1–32. Flows to the International Resource Panel. Norilsk Nickel (2010) Mineral Reserves and Resources United Nations (UN) Comtrade (2013) Commodity Trade Statement. http://www.nornik.ru/en/our_products/ Statistics Database (http://comtrade.un.org/db/). MineralReservesResourcesStatement/ Zientek, M.L., Cooper, R.W., Corson, S.R. and Geraghty, Oberthu r, T. (2002) Platinum-group element mineraliza- E.P. (2002) Platinum-group element mineralization in tion in the Great Dyke, Zimbabwe. In: Cabri, L.J. (ed.) the Stillwater Complex, Montana. In: Cabri, L.J. (ed.) The geology, geochemistry, mineralogy and mineral The geology, geochemistry, mineralogy and mineral beneficiation of platinum-group elements. Canadian beneficiation of platinum-group elements. Canadian Institute of Mining, Metallurgy and Petroleum, Special Institute of Mining, Metallurgy and Petroleum, Special Volume 54. Volume 54. 13. Rare earth elements

FRANCES WALL

Head of Camborne School of Mines and Associate Professor of Applied Mineralogy, Camborne School of Mines, University of Exeter, Penryn, UK

Introduction (rare) and to denote that the REE are usually stable as oxides (earths) rather than metals. The The rare earth elements (REE) are defined first rare earth, a mixture of REE called ‘yttria’, according to the International Union of Pure and was isolated from a mineral found at Ytterby, Applied Chemistry (IUPAC) as the 15 lanthanides Sweden by Finnish chemist, J. Gadolin in 1794; (also called lanthanoids) together with yttrium the others followed over the next 153 years, until and scandium. In practice, the element scandium promethium was found in 1947 (Gupta and behaves rather differently in nature to the other Krishnamurthy, 2005, and Kaltosoyannis and rare earths and its inclusion in the official defini- Scott, 1999). Separation of REE from each other is tion is not particularly helpful. Moreover, one of still a significant challenge in the processing and the lanthanides, promethium has a short half-life refining of rare earth ores. The term didymium, and is thus exceptionally rare. This leaves a group used originally for an ‘element’ that proved to be of 15 elements comprising 14 lanthanides and a mixture of praseodymium (Pr) and neodymium yttrium (Y), which do form a coherent group in (Nd), is still used today, particularly in the USA, nature as well as in technological uses. The to refer to the REE, mainly La, Nd and Pr, remain- IUPAC definition divides the REE into the light ing after extraction of cerium (Ce). The United REE (lanthanum (La) – samarium (Sm)) and heavy States Geological Survey (USGS) uses didymium REE (europium (Eu) - lutetium (Lu)). Yttrium to refer to 75 per cent neodymium, 25 per cent behaves as a heavy REE. The terms ‘mid’, ‘middle’ praseodymium. and ‘medium REE’ are also sometimes used to denote Sm – dysprosium (Dy). Together with the actinides, the lanthanides are members of the ‘f Physical and chemical properties elements’ or ‘f block’ in the Periodic Table. Some authors prefer to use the abbreviation REY instead The REE are soft, silver-coloured metals that tar- of REE to denote Y and the lanthanides. nish quickly in air and have high melting points The name rare earths is older than the term (Table 13.1). The lanthanide series results from lanthanides and was used to describe the diffi- filling of the ‘f’ electron shell in the atoms, with culty experienced by nineteenth-century chem- configurations ranging from 5d6s2 for La and ists in separating the elements from each other 4f15d16s2 for Ce to 4f145d 6 s2 for Lu at the end of

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Rare earth elements 313

Table 13.1 Selected properties of the rare earth elements.

Chemical Atomic Atomic Density at Melting Cation Colour of +3 cation Element name symbol number weight 25 °C (kg/m3) point (°C) radius (pm) in solution scandium Sc 21 44.96 2992 1541 87.0 Colourless yttrium Y 39 88.91 4475 1522 107.5 Colourless lanthanum La 57 138.91 6174 918 121.6 Colourless cerium Ce 58 140.12 6711 798 119.6 Colourless praseodymium Pr 59 140.91 6779 931 117.9 Yellow-green neodymium Nd 60 144.24 7000 1021 116.3 Rose promethium Pm 61 144.91 7220 1042 – Pink samarium Sm 62 150.36 7536 1074 113.2 Yellow europium Eu 63 151.96 5248 822 112.0 Colourless gadolinium Gd 64 157.25 7870 1313 110.7 Colourless terbium Tb 65 158.93 8267 1356 109.5 Pale pink dysprosium Dy 66 162.50 8531 1412 108.3 Pale yellow-green holmium Ho 67 164.93 8797 1474 107.2 Yellow erbium Er 68 167.26 9044 1529 106.2 Pink thulium Tm 69 168.93 9325 1545 105.2 Greenish tint ytterbium Yb 70 173.04 6966 819 104.2 Colourless lutetium Lu 71 174.97 9842 1663 103.2 Colourless

Compiled from Gupta and Krishnamurthy (2005); pm, picometres. the series. This gives rise to the magnetic and and, therefore, as the atomic number (and thus spectroscopic properties that make REE so useful number of protons in the nucleus) increases, in many applications. These inner electrons are the electrons are pulled close to the nucleus. shielded and so maintain the distinct elemental The Y3+ cation is most similar in size to Ho properties in various bonding situations. The (Figure 13.1), whereas Sc3+ is much smaller at electronic properties give the REE distinct and just 0.87Å. sharp absorption and emission spectra, including In nature, REE almost always occur in the +3 distinct colours in solution (Table 13.1). Most valency state (i.e. they form the oxide REE2O3) REE are strongly paramagnetic and the strong but there are two important exceptions in natural magnetism that results from their combination environments. In oxidising environments, such with transition metals such as iron and cobalt is as weathered deposits and seawater deposits, Ce 4+ one of their most important features. forms Ce (CeO2), a much smaller cation, and in Although the physical and chemical prop- reducing environments, Eu forms the larger Eu2+ erties of the REE are similar, many of their prop- cation (Figure 13.1). In the laboratory, +2 valen- erties change systematically and smoothly cies are also known for Sm, Tm, Yb; Pr6O11 is a through the lanthanide series. One of the most mixed synthetic oxide (4PrO2⋅Pr2O3). important properties in determining chemical behaviour is cation size and, conversely to what might be expected, the size of lanthanide cations Distribution and abundance in the (and atoms) decreases smoothly with increasing Earth’s crust atomic number (Figure 13.1). This phenomenon is called the lanthanide contraction and arises The rare earth elements range in abundance in because the f electrons do not screen the other the crust from about the same level as copper and electrons from the positive pull of the nucleus lead down to the same levels as bismuth. They 314 frances wall

1.4

Eu2+ 1.3

REE3+ 1.2

1.1 Y3+ Cation radius (Å) 1 4+ Ce Figure 13.1 The lanthanide contraction demonstrated by plotting the radii of 3+ cations in nine coordination. The 0.9 Sc3+ radii of the cations of the other most common lanthanide oxidation states are also shown, as are the radii of Y3+ 0.8 and Sc3+. (Modified after Gupta and La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Krishnamurthy, 2005.)

65 64

55

45

35 30 26 25 22 Figure 13.2 Crustal abundances of rare

Crustal abundance (ppm) earth elements (Data from Taylor and 15 McClennan, 1985). The light rare 7.1 earths, lanthanum (La), cerium (Ce) and 4.5 3.8 5 3.5 2.3 2.2 neodymium (Nd), are the most 0.88 0.64 0.8 0.33 0.32 abundant rare earth elements in the La Ce Pr Nd Pm Sm Eu Gd Tb Dy Ho Er Tm Yb Lu Y Earth’s crust. are considerably more abundant than precious usual practice to remove it from graphical presen- metals such as gold and the platinum group ele- tations of REE concentrations in rocks and min- ments. Light REE are up to 200 times more abun- erals by chondrite normalising the data, which dant than heavy REE and odd atomic number involves dividing the absolute REE concentra- REE are more abundant than the even number tions by the accepted REE values in chondrite elements (Figure 13.2). This ‘zig-zag’ odd–even meteorites (e.g. McDonough and Sun, 1995). pattern is called the Oddo Harkins effect and it is Crustal abundances determine which REE are Rare earth elements 315 most abundant in ore deposits and which are only the series and there are no geochemical processes ever found in small quantities subsidiary to the that can separate and concentrate these other REE. For example, Tm is heavy and odd individual REE sufficiently for them to form atomic number and thus has a low abundance their own minerals. (Figure 13.2); Ce is light and even atomic number Monazite-(Ce) is one of the most common and and is the most abundant of the REE, usually pre- widespread REE minerals. It occurs as an dominant in minerals. Heavy to light REE ratios accessory mineral in granite, gneiss and other are often used and are best calculated using La to igneous and metamorphic rocks. It is highly resis- represent the light REE and either Y or Yb to tant to weathering, robust during transport and represent the heavy REE. thus survives to become incorporated into sedi- mentary rocks and concentrated in mineral sand placer deposits. Monazite also occurs in hydro- Mineralogy thermal deposits and may form during weathering. It is almost always strongly enriched in the light There are over 200 rare earth minerals (Miyawaki REE (Figure 13.3). Examples of monazite compo- and Nakai, 1996) approved by the International sition in a variety of rock types are given by Mineralogical Association (2012). There are also Förster (1998), Smith et al. (2000), Wall and many others that can contain significant substitu- Mariano (1996), Wall and Zaitsev (2004). The tions of REE in place of their main constituents. other principal constituents of monazite are tho- However, many REE minerals are rare and the list rium, calcium and silicon. Thorium substitution of minerals likely to be encountered as major con- into monazite must be coupled with either stituents of an ore deposit is rather less extensive calcium or silicon in order to main charge (Table 13.2). balance. Coupled substitution of thorium and Rare earth minerals are named with a suffix silicon for REE moves the composition towards that indicates the predominant REE in the REE the end member huttonite (ThSiO2), whilst cou- site (Bayliss and Levinson, 1988). Monazite-(Ce), pled substitution of calcium and thorium changes for example, is the most common form of mona- the composition towards cheralite (CaTh(PO4)2). zite. If La, Nd, or Sm are dominant in the REE Monazite derived from granite (e.g. monazite in site instead of Ce, they form different mineral placer deposits) tends to have higher thorium and species called monazite-(La), monazite-(Nd) or uranium than monazite from carbonatite. For monazite-(Sm). In practice, almost all light REE example, van Emden et al. (1997) found an average minerals will have the suffix-(Ce), because Ce of 8.79 and standard deviation of 0.08 wt% ThO2 has the highest crustal abundance, but they will in 500 monazite grains from Western Australian also contain significant La, Pr and Nd. Likewise, mineral-sand deposits. Y is almost always the predominant REE in The other most common light REE minerals heavy REE minerals because of its higher crustal that are important in REE ore deposits are the abundance compared with the other heavy REE. uor carbonate group of minerals including bast- Only in rare cases have other heavy REE species näsite-(Ce), synchysite-(Ce) and parisite-(Ce) been described, e.g. keivyite-(Yb). It is possible to (Figure 13.3). These minerals have layered find more than one mineral species within a structures and are commonly syntaxially inter- single mineral grain if the mineral is zoned. Most grown with the each other; they also frequently minerals are either light or heavy REE-enriched occur as sheaves of fine needle-like crystals. rather than equally rich in all REE because of the The most common heavy REE mineral is xeno- variation in cation size. Pr, Eu, Gd, Tb, Dy, Ho, time-(Y), which occurs as an accessory mineral in Er, Tm and Lu are almost never dominant in any granite, gneiss, and other igneous and metamor- mineral because their absolute abundance is too phic rocks. It is also resistant to weathering and low compared with neighbouring members of robust enough to be incorporated into sedimentary 316 frances wall

Table 13.2 The most common rare earth minerals, including an indication of the rare earth oxide (REO), thorium (Th) and uranium (U) contents and beneficiation process.

Mineral Formula Wt % REO Th, U Other REE variants Beneficiation

CARBONATES AND FLUORCARBONATES . ancylite-(Ce) SrCe(CO3)2(OH) H2O 43 – La HCl acid dissolution (exp) bastnäsite-(Ce) CeCO3F 75 – La, Nd, Y F huanghoite-(Ce) BaCe(CO3)2 F40– parisite-(Ce) CaCe2(CO3)3 F 2 50 – Nd F synchysite-(Ce) CaCe(CO3)2 F 51 – Nd, Y F PHOSPHATES cheralite Ca,Th(PO4)2 variable M – churchite-(Y) YPO4.2H2O 51 V Nd – florencite-(Ce) (Ce)Al3(PO4)2(OH)6 32 – Sm – monazite-(Ce) CePO4 70 V La, Nd, Sm F or GME xenotime-(Y) YPO4 61 V Yb GME OXIDES aeschynite-(Ce) (Ce,Ca,Fe,Th)(Ti,Nb)2(O,OH)4 32 V Nd,Y – cerianite-(Ce) CeO2 100 V – loparite-(Ce) (Ce,La,Nd,Ca,Sr)(Ti,Nb)O3 30 – Chlorination yttropyrochlore-(Y) (Y,Na,Ca,U)1–2Nb2(O,OH)7 e.g. 17 V – SILICATES 2+ allanite-(Ce) CaNdAl2Fe (SiO4)(Si2O7)O(OH) 38 V La, Nd, Y – britholite-(Ce) (Ce,Ca,Sr)2(Ce,Ca)3(SiO4,PO4)3 (O,OH,F) e.g. 23 V Y – exp eudialyte Na15Ca6Fe3Zr3Si(Si25O73)(O,OH,H2O)3(Cl,OH)2 e.g. 9 – Have leached REE 2+ fergusonite-(Ce) CaNdAl2Fe (SiO4)(Si2O7)O(OH) 53 –? Nd, Y, β-fergusonite- Hot caustic (Ce), Nd,Y digestion and acid dissolutionexp 2+ gadolinite-(Ce) Ce2Fe Be2O2(SiO4)2 60 V Y treated with acid (small scale) 2+ gerenite-(Y) CaNdAl2Fe (SiO4)(Si2O7)O(OH) 44 – – kainosite-(Y) Ca2Y2(SiO3)4(CO3).H2O38– keiviite-(Y) Y2Si2O7 69 – Yb – 2+ 3+ exp steenstrupine-(Ce) Na14Ce6(Mn )2(Fe )2Zr(PO4)7Si12O36(OH)2.3H2O 31 V F and leach FLUORIDES fluocerite-(Ce) CeF3 83 – La –

Minerals currently mined commercially are in bold. F = flotation; GME = combination of gravity, magnetic and electrostatic processes; exp = technique not yet applied on a working mine. U, Th contents: M = usually has major (wt%) Th; V = Th and U are variable from <1 wt% to major (e.g. 10 wt%) amounts. Other minerals usually have <1 wt% Th and U but almost all will contain trace quantities of these elements. rocks and concentrated in placer deposits such as (Table 13.2), where Ce is dominant and mineral sands. Compositions are always enriched concentration of the other light REE are very low. in the heavy REE (Figure 13.3). In reducing environments, Eu2+ substitutes for Ca Cerium, when oxidised to its +4 state, can in feldspar and fractionates from the other REE. form separate Ce minerals, such as cerianite This gives rise to Eu anomalies with respect to Rare earth elements 317

100.00 × × × × × × × × × × × × × × ×× × × × × × × × 10.00 × × × × × × × 1.00 × × × × Mineral/chondrite Figure 13.3 Compositions of the most 0.10 × × common light REE minerals, bastnäsite- × (Ce) and monazite-(Ce), and the most common heavy REE mineral, xenotime-(Y). Following usual precedent, 0.01 values are divided by values of a chondrite La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu meteorite (McDonough and Sun, 1995) in order to remove the Oddo-Harkins Bastnäsite Bayan Obo Monazite Bayan Obo ‘zig-zag’ effect. (Data from Förster (1998), Monazite carbonatite Monazite mineral sand Smith et al. (2000), van Emden et al. Xenotime mineral sand Xenotime granite (1997), Wall and Mariano (1996), Wall Xenotime carbonatite et al. (2008).) the concentrations of neighbouring REE in min- Deposit types erals but does not result in discrete Eu minerals. The most abundant non-REE mineral with There are many potential REE ore deposits in a significant REE substitution is the calcium phos- wide variety of rocks. Orris and Grauch (2002) phate, apatite ( uorapatite is the official mineral listed 822 occurrences of REE divided into 14 dif- name for the most common uorine-bearing ferent deposit types and Figure 13.4 illustrates 78 variety), which occurs in many different rock deposits divided into nine different categories. types. In most, it is present in minor quantities but For the purposes of the discussion here, the REE in carbonatites and alkaline igneous rocks apatite deposits are divided into carbonatite-associated can be one of the main rock-forming minerals. It is deposits, including weathered carbonatite; also the main constituent of sedimentary phos- alkaline igneous rocks, including alkaline gran- phorites. The REE can substitute into the two Ca ites; other hydrothermal deposits; ion adsorption sites in apatite, often in a coupled substitution deposits; placer deposits and sea oor deposits. with silicon or sodium, and can reach levels of REE are also produced as by-products of other tens of weight per cent (wt%), although REE con- minerals and can be recovered from waste. The tents of about 1 wt% are more normal in alkaline various types of deposits have particular charac- rocks and carbonatites, with lower levels in other teristics of size and grade (Figure 13.5). rocks, such as phosphorites. Carbonatites tend to be medium to large tonnage Zircon and uorite are also important hosts and high grade, whereas alkaline rock deposits for REE. Zircon favours the heavy REE but only are generally larger tonnage but lower grade, contains trace or minor quantities. Fluorite usu- tending to have higher proportions of heavy REE. ally only contains trace quantities of REE, but it Mineral sands are low grade but REE minerals are can occasionally take up either light or heavy by-products; ion adsorption deposits are small REE and has yttro uorite and cer uorite species. and low grade but relatively rich in heavy REE. Figure 13.4 Global distribution of rare earth deposits (Source: British Geological Survey, 2011). Rare earth elements 319

Total REO (thousand tonnes) 1 10 100 1000 10,000 100,000 CARBONATITE Bayan Obo X Maoniuping X Weishan X Mountain Pass X Mount Weld X Dong Pao X Zandkopsdrift X Bear Lodge X Wigu Hill X Kangankunde X ALKALINE ROCKS Khibiny X Lovozero X Dubbo zirconia X Nechalacho, Thor Lake X Kvanefjeld X Strange Lake X Kutessay II X ION ADSORPTION Chenxian County X Guidong X Jianghua X OTHER HYDROTHERMAL Steenkampskraal X Nolans project X Hoidas Lake X PLACER Chavara X IOCG Olympic Dam X 0 2 4 6 8 101214161820 Grade (%REO)

Total amount REO X Grade (%REO) Figure 13.5 Comparison of the size (thousand tonnes) and grade of rare earth mines and Resource with significant advanced development projects. (Data amount of HREE from Table 13.3.)

Hydrothermal deposits in an advanced stage of economic source of REE (Figure 13.4, Table 13.3). development are high grade and medium size. Although they are igneous rocks, carbonatites are not typical of mid-ocean ridges and plate collision zones where most volcanoes form, but instead are Carbonatite-related REE deposits found in the middle of continents, often associated Carbonatites, which are igneous rocks composed of with extensional plate tectonic activity, such as more than 50 per cent carbonate, are the main rift valleys or continental breakup. Intrusive 320 frances wall

Table 13.3 Rare earth mines and advanced projects

Notes on company, mining, Name, Location Geology Ore minerals* Grade and size processing

Carbonatite BAYAN OBO, Inner Metamorphosed and bastnäsite, monazite 750 Mt at 4.1% REO Baotou Steel Rare Earth Mongolia, China metasomatically (Group) Hi Tech Co Ltd. altered carbonatite Open pit. Flotation, sulphuric acid extraction. REE by-product of iron ore. MAONIUPING, Mianning Carbonatite related. bastnäsite 1.2 Mt at 2.89% REO Jiangxi Copper Sichuan Rare County, Sichuan, China hydrothermal Earth Company. Open pit. Flotation. DALUCAO, Liangshan Carbonatite/alkaline bastnäsite 1.86 Mt ?REO Dechang Houdi Rare Earth Autonomous Prefecture, rocks related Mining Co., Ltd. Open pit. Sichuan, China WEISHAN, Shandong Carbonatite and bastnäsite 12.75 Mt REO, grade >1.6% Shandong Weishan Lake Rare Province, China alkaline rocks (Wu et al. 1996) Earth Co. Ltd. Open pit. MOUNTAIN PASS, Carbonatite bastnäsite (monazite 20 Mt at 8.24% REO, 5% Molycorp Inc. Open pit. California, USA planned) cut-off Flotation. Processed to REE products on site. Bear Lodge, Crook County, Carbonatite dykes and ancylite, 15.88 Mt at 3.454% REO, 1.5% Rare Element Resources (RER). Wyoming, USA veins in alkaline bastnäsite-group cut-off. 0.5484 Mt REO Open pit. Crushing, igneous complex minerals attrition, screening and hydrochloric acid leaching. MT WELD, Western Lateritic weathered monazite 17.49 Mt at 8.1% REO, 1.416 Lynas Corp. Open pit. Australia carbonatite Mt REO# Flotation. Concentrate sent to Kuantan, Malaysia. Dong Pao, North Vietnam Lateritic weathered bastnäsite, synchysite, 7.4 Mt at 5.22% REO, 3% Toyota Corp, Sojito, Vinacomin carbonatite (minor monazite) cut-off, for No. 3 orebody joint venture. Zandkopsdrift, Weathered monazite with 21 Mt at 1.99% REO at 1% Frontier Minerals. Open pit. Namaqualand, northern carbonatite crandallite and cut-off, 415,000 t REO Flotation. REE extraction at Cape, South Africa some churchite Saldanha Bay.

Alkaline rocks KHIBINY, Kola Peninsula, Agpaitic nepheline fluorapatite 9 Mt REO O&G Apatit GOK Open Pit and Russia syenite underground. By-product of fertiliser manufacture if nitric acid used. LOVOZERO, Kola Agpaitic nepheline loparite 3.4 Mt REO#, 0.8–1.5% REOO&G Lovozerskiy Mining Company. Peninsula, Russia syenite Underground mine. REE by-product of Nb. Dubbo Zirconia, Toongi, Alkaline trachyte dyke bastnäsite, ancylite 73.2 Mt at 0.75% REO. 0.549 Alkane Resources Ltd. central west NSW, Mt total REO. Sulphuric acid leaching Australia followed by solvent extraction. Also Zr, Nb, Ta Nechalacho, Thor Lake, Peralkaline layered bastnäsite, allanite, 107.59 Mt at 1.26–1.48% Avalon Rare Metals Inc. North West Territories, nepheline syenite parisite, monazite, REO, 4.3 Mt REO Undergound mine. Also Nb. Canada with hydrothermal synchysite. HREE in upgrade of REE fergusonite and zircon. Rare earth elements 321

Notes on company, mining, Name, Location Geology Ore minerals* Grade and size processing

Kvanefjeld, Ilimaussaq Agpaitic nepheline steenstrupine, 619 Mt at 10,585 ppm Greenland Minerals and Gardar Province, syenite (eudialyte) REO, 6.55 Mt REO Energy Ltd and Westrip Greenland Holdings. Open pit. Carbonate pressure leaching to recover U, precipitation of REE. Strange Lake, NE Peralkaline granite yttropyrochlore, 114.8 Mt at 0.999 % Quest Rare Metals. Quebec/NW Labrador (hypersolvus, kainosite, allanite, REO, 43% HREE border, Canada subsolvus and gadolinite, gerenite, pegmatite) with keiviite: additional (fluorocarbonates & hydrothermal monazite) concentration of REE Zeus, Kipawa alkaline Alkaline compex with eudialyte, Y-bearing 67,200 t REO, significant Matamec Explorations. Open complex, syenite and granite titanite, britholite HREE pit. Mild crushing and Témiscamingue, leaching. Also Zr. Quebec, Canada Kutessay II, Aktyuz Ore 3 mineralised zones, monazite, xenotime-(Y), 16.27 Mt at 0.264% REOBGS Stans Energy Corp. Former field, Chu Oblast, (I, II, III). Massive bastnäsite-(Y), 50:50 LREE:HREE REE mine. Produced all Kyrgyz Republic pockets of REE parisite- (Y)?, 15 REE. carbonates synchysite-(Y), yttrofluorite, fluorcerite

Ion adsorption Ion adsorption deposits, Clay-rich weathered REE are adsorbed most deposits <10,000 t, Multiple companies. In-situ 214 deposits in Jiangxi, granites and other onto kaolinite and 0.03–0.35% REOBGS, leaching or small scale Hunan, Guangdong, silicate rocks other clays HREE rich mining and leaching. Guangxi,and Fujian, China

Hydrothermal Steenkampskraal, Western Hydrothermal monazite 249,500 t at 17% REO, Rareco (Great Western Cape Province, South monazite in 29,400 t REO)# Minerals Group). Africa metamorphosed Underground mine. Gravity gneisses separation and flotation. Nolans project, Northern Sub-vertical dykes, REE-bearing 30.3 Mt at 2.8% REO, 848 kt Arafura Resources. Open pit. Territories, Australia veins and fluorapatite REO Heavy media. REE stockworks of (especially Nd and carbonate recovery from fluorapatite in Eu), cheralite acid leach. gneiss. Hoidas Lake, REE in apatite and apatite, allanite 2.8 Mt at 2.139–2.568% REO Great Western Minerals Group. Saskatchewan, Canada allanite veins Placer Manavalakurichi, Tamil Marine placer? monazite total reserves in India amount Indian Rare Earths Ltd (IREL), Nadu, India to 10Mt monazite.# Federal Department of Atomic Energy.

(continued) 322 frances wall

Table 13.3 Continued.

Notes on company, mining, Name, Location Geology Ore minerals* Grade and size processing

Chavara, Kerala, India Marine/alluvial monazite IREL placers Orissa, Orissa, India Alluvial placers monazite IREL Chhatapur, Ganjam Alluvial placers on monazite Toyotsu Rare Earths (IREL and district, Orissa, India coast Toyota Tsusho).

Other SARECO, Kasatomprom’s Y-rich U ore tailings, U ? ? Sumitomo, Kazatomprom. nuclear facilities, ores, in situ Processing at Kazatomprom Ust-Kamenogorsk, leaching solutions facility. Kazakhstan** and REE deposits Pitinga, Amazon Region, Processing tailings at xenotime-(Y) 8.5% REO in tin tailings Neo Material Technologies, Brazil a tin mine Mitsubishi.

OPERATING MINES in capitals # source Roskill, company websites **also a second source in Kazakhstan but too few details known to list *all minerals are –(Ce) varieties unless specified O&G Orris & Grauch, 2002. Nb, niobium; Ta, tantalum; U, uranium; Zr, zirconium.

carbonatites occur as plugs, ring dykes, dykes and carbonatites at Mountain Pass, USA and Bayan veins; they are frequently surrounded by an alkali Obo, China (Table 13.3) are different and atypical in metasomatic aureole called fenite and are com- this respect. monly (but not necessarily) associated with alkaline The world’s largest REE deposit at Bayan Obo, or ultrabasic igneous silicate rocks. A typical Inner Mongolia, China has produced REE since igneous carbonatite assemblage of calcite, apatite, the late 1980s and currently accounts for the magnetite with accessory pyrochlore is unlikely to majority of global REE supply. The ore minerals constitute an economic REE deposit although it are bastnäsite-(Ce) and monazite-(Ce) set in an may contain REE minerals as late-stage minor com- iron oxide-rich metamorphosed stratiform dolo- ponents. The weight per cent levels of light REE mite together with a variety of other REE min- required for an economic deposit commonly occur erals, notably aeschynite, uorite and aegirine. in magnesium- and iron-rich carbonatite dykes and There are three ore bodies (main, east and west) veins emplaced in pegmatoid-type, uid-rich envi- along strike of the H8 dolomite horizon. The ronments late in the development of the igneous exact nature of Bayan Obo has been controver- complex. The original minerals have commonly sial, and, although most researchers now agree on been replaced although their shapes may have been a carbonatite-related origin, there is no doubt maintained, forming ‘pseudomorphs’. Such deposits that the deposit has been highly altered and can be high grade but frequently consist of narrow reworked (Smith and Wu, 2000 and Wu, 2008). dykes and veins. Minerals include bastnäsite, mon- The other best known REE deposit is the azite, and ancylite. Kangankunde, Nkombwa Hill, Mountain Pass carbonatite, USA (Table 13.3) Wigu Hill, Songwe, Lofdal, Bear Lodge, Qeqertaasaq (Castor, 2008). This is a large dyke-like igneous and Tikiusaaq are examples (Figures 13.4 and 13.5, intrusion of carbonatite, which is associated with Table 13.3). The two most famous REE-rich an igneous silicate rock called shonkinite. The Rare earth elements 323

(a) (b)

Monazite

Feox

Clay 1 cm Monazite

0.5 mm

Figure 13.6 (a) Backscattered electron image of monazite in weathered carbonatite at Mount Weld, Australia. There is a narrow band of later monazite on the main monazite grain. The associated minerals are iron oxides (Feox) and clays. The grains are set in resin. (Sample courtesy of Lynas Corporation). (b) Steenstrupine crystals from a sodalite- and feldspar-rich late vein in nepheline syenite. Taseq Slope, Ilimaussaq alkaline complex, Greenland. (Sample and photo courtesy Henrik Friis.) carbonatite contains bastnäsite, and other uor- sodium and potassium contents compared with carbonate minerals (e.g. synchysite and parisite) silicon and aluminium and contain characteristic together with some monazite. It has been pro- mineral suites, including complex sodium, posed that much of this deposit consists of igneous titanium, zirconium silicates. Some of these min- minerals (all light REE enriched) that precipitated erals can host significant quantities of REE, directly from the carbonatite magma, rather than although few of them have been beneficiated on a being subject to the more common late and post- commercial scale. These minerals may form magmatic processes in carbonatites. during the original crystallisation of the magma Weathering or alteration by hydrothermal or during later hydrothermal alteration that uids dissolves the soluble carbonates and thus upgrades the original magmatic REE concentra- concentrates the less-soluble REE. The monazite tions. The rocks may contain polymetallic deposit at Mount Weld, Western Australia is pro- deposits with zirconium, niobium and tantalum duced by weathering of carbonatite, as is the as well as REE (Table 13.3). Examples include the Zandkopsdrif deposit in South Africa. These agpaitic nepheline syenite complex at Illimaussaq deposits can have fine-grained and complex min- (Kvanefjeld), Greenland (Table 13.3) which con- eralogy (Figure 13.6) but tend to have low Th. tains steenstrupine (Figure 13.6), The perovskite They are of similar size and grade to other car- mineral, loparite, is mined primarily as a source bonatite deposits (Figure 13.5) but richer in REE of niobium from the layered agpaitic nepheline than the original unweathered carbonatite. syenite at Lovozero, Kola Peninsula, Russia but contains about 1.04–1.25% REO, which can be extracted (Petrov, 2004) (Table 13.3). Eudialyte in Alkaline igneous rocks a separate intrusion at Lovozero has also been Alkaline igneous rocks are also typical of intrac- considered as a potential source of REE. The ontinental tectonic settings. Variants termed agpaitic nepheline syenite at Khibiny, Kola ‘per alkaline’ and ‘agpaitic’ have particularly high Peninsula, Russia contains the world’s largest 324 frances wall igneous apatite mines. The apatite contains with ‘moderate’ light to heavy REE-enrichment 0.66 wt% REE, which can be extracted during fer- but negative Eu anomalies. The REE content of tiliser production if a nitric acid process is used apatite in other iron ores is about 1000 ppm rather than the more usual sulfuric acid produc- (Frietsch and Perdahl, 1995). tion route (Petrov, 2004). The Kutessay II deposit in Kyrgyzstan is notable for its Y-variety minerals Placer deposits (mineral sands) (Table 13.3) and high proportion of heavy REE. Its geology is poorly described but it appears to be Monazite and xenotime are among a set of highly altered and associated with granitic intru- resistate minerals that survive weathering, sions, although the REE minerals are mainly car- erosion and transport and can be redeposited in bonates (Stans Energy Corp., 2012). river, estuarine or shallow marine environments. Sedimentation processes tend to concentrate the Other hydrothermal veins heavier minerals together and thus deposits of quartz with elevated concentrations of zircon, A number of deposits have apparently formed by ilmenite, rutile, monazite and xenotime can be precipitation of REE-bearing minerals from hot- formed. There are well-known deposits along the water-based solutions without any direct rela- coast of southern India in Kerala, and Orissa; in tionship to the emplacement of igneous rocks. It Western Australia north and south of Perth; and is notable that most of these contain apatite, in South Africa at Richards Bay. in addition to REE minerals. They can be as large Elliot Lake in Ontario, Canada is a palaeopla- as some of the carbonatite deposits and are of cer uranium and REE deposit in conglomerate. variable grade (Figure 13.5). For example, the monazite-apatite vein at Steenkampskraal, South Africa (Andreoli et al., 1994) is related to Ion adsorption deposits local geological structures but has no apparent Ion adsorption deposits are formed on weathered igneous source. granites, and, to a lesser extent on pyroclastic rocks and lamprophyres, in the sub-tropical cli- mates of Jiangxi, Hunan, Guangdong, Guangxi, Iron-oxide–apatite deposits, including iron- and Fujian provinces, south of 28°N in southern oxide–copper–gold (IOCG) deposits China (Bao and Zhao, 2008). There are 214 The Olympic Dam iron-oxide–copper–gold deposits recorded over an area of 90,000 km2. deposit, South Australia, which is mined cur- Individually, each deposit tends to be small, with rently for copper, uranium, gold and silver, also less than 10,000 tonnes of ore in each deposit, contains a large, low-grade deposit of REE in its and low grade, varying from 300 to 3500 ppm steeply dipping, dyke-like bodies of hematite REO (British Geological Survey, 2011). The breccias five km long and one km deep within weathering profiles are typically 15 to 35 m thick fractured granite. Hydrothermal bastnäsite, or- (Bao and Zhao, 2008). The deposits contain REE encite, monazite, xenotime and britholite host an adsorbed onto the surface of clays as well as REE average of 5000 ppm REE with variable light to minerals. A key controlling factor in their heavy REE ratios (Oreskes and Einaudi, 1990), formation is the presence of REE in the fresh giving a large but low-grade deposit (Figure 13.5). granite in minerals that are easily weathered such Bayan Obo has also been proposed as an IOCG- as parisite (Table 13.2) (Bao and Zhao, 2008) rather type deposit but has higher grades of REE (Wu, than resistant minerals such as monazite. The 2008). Other iron-ore deposits that contain apa- radioactivity of these deposits is low. Furthermore, tite are also potential deposits of REE. For the deposits, being clay, are easy to mine and in example, apatite in apatite–magnetite ores at some cases the adsorbed REE can be released Kiruna, Sweden contains 2000 to 7000 ppm REE from the clays in situ by ion exchange. Rare earth elements 325

Seafloor deposits Extraction methods, processing and beneficiation Deep-ocean manganese nodules, iron-manganese Mining crusts and deep sea muds are potential marine sources of REE, of similar size to deposits such Most REE mines, such as Bayan Obo, Mountain as Mountain Pass. The Clarion–Clipperton Pass and Mount Weld, are open-cast operations, Manganese Nodule Zone (CCZ) in the north-east involving conventional blast, load and haul tech- Pacific contains a resource of 211 million tonnes niques. No underground mines have ever been at 0.1% REO or 21 million tonnes REO. The designed for the exclusive production of REE but Prime Iron–Manganese crust zone (PCZ) in the there is, or has been, production of REE from a central Pacific contains 7500 million tonnes at few underground mines. For example, the loparite 0.3% REO or 23 million tonnes REO. These (niobium) mines at Lovozero, Kola Peninsula, deposits have higher proportions of heavy REE Russia, the former thorium mine at Steenkampskraal, (e.g. 6.5 to 10 per cent) than carbonatite-related South Africa, now being re-opened primarily for deposits and have very low thorium contents of REE, and the former uranium mines at Elliot Lake, 11–14 ppm (Hein et al., 2011). Deep-sea sedi- Ontario, Canada. ments in the Pacific Ocean that have REE held in Different mining techniques are used for the iron oxyhydroxides and phillipsite have also been beach sand placer deposits because they are generally proposed as large, low-grade, low-thorium REE much less consolidated than carbonatites, alkaline deposits (Kato et al., 2011). rocks or hydrothermal deposits. They are also often under water. Mining techniques include dredging By-products, co-products and waste products and excavation by bucket wheel or by excavator. Some crushing may be required (e.g. beach placers, An important point when considering future now inland in Western Australia). There are few deposits, is that there is potential for production details available of mining techniques for ion adsorp- of REE as by-products and co-products of other tion deposits in China but many are small-scale commodities. In the deposits described above, operations, with much of the mining done by manual REE at Bayan Obo are a by-product of iron ore, labour. The clay deposits are excavated and leached monazite from mineral sands is a co-product with to extract REE or are leached in situ. ilmenite, zircon and rutile, and xenotime is pro- duced from tin concentrate. REE can be extracted Beneficiation from loparite during processing for niobium and from apatite during fertiliser production (e.g. at In most mining operations, it is more economic Lovzero and Khibiny – see Alkaline igneous to concentrate the ore mineral and remove as rocks, above). There is more potential in other many of the waste (gangue) minerals as possible phosphate deposits used for fertiliser production. before attempting to extract the elements of Simandl et al. (2011) calculated that, assuming an value. This rule applies in most REE mining oper- average REE content of 460 ppm in phosphorite, ations, where both bastnäsite and monazite are the world’s annual phosphate production of 170 liberated from accompanying phases by crushing million tonnes represents over 70,000 tonnes of and grinding and then concentrated by otation. contained REE. At the high REE prices of mid- The otation process demands a very fine grain 2011, the REE could have a higher market value size of about 50 μm and is thus energy intensive. than the phosphate in some deposits (Simandl At Mountain Pass the ore is crushed and ground et al., 2011). Bauxites and waste from their and then processed in long series of otation processing also contain REE. The potential for stages. The reagents used are hot, which is extracting REE from the red mud wastes has been unusual in otation processes (Figure 13.7). At considered by various researchers (Red Mud Mount Weld, the ore is crushed in a ball mill and Project, 2006). then sent to the otation circuit. At Bayan Obo, 326 frances wall

Crushing, Chemical & Ore slurry Bastnäsite Froth grinding, steam (30–35% ore flotation screening conditioning solids)

Bastnäsite concentrate

REE Bastnäsite Thickening, concentrate Acid Calcination concentrate filtering, (85–90% leach (60% REO) drying Figure 13.7 Schematic summary of the REO) beneficiation of bastnäsite ores.

Gravity Heavy mineral Water separation Screening Slurry Concentrate sands added by jigs, spirals, tables

Induced Non- Non- Washing, Electrostatic Magnetic Dry magnetic conductive magnetic dewatering, separation separation concentrate separation minerals minerals drying

Zircon Conductive Magnetic minerals minerals e.g. leucoxene e.g. ilmenite

Weakly Xenotime concentrate Gravity magnetic separation minerals Monazite concentrate

Figure 13.8 Schematic summary for the extraction of monazite and xenotime from heavy mineral sands. the ore is more complex and the beneficiation cir- usually produced as co-products. The combinations cuit also produces magnetite, uorite, hematite and owsheets vary in detail according to the min- and niobium oxide as by-products (Gupta and erals present and products desired. Krishnamurthy, 2005). However, the rare earth Processing of ion adsorption clays is distinct minerals, bastnäsite and monazite, are again from other REE deposits in that no beneficiation recovered in a series of otation stages. is required and the REE are leached directly from There is an alternative beneficiation route for the ore, in situ or in ponds, in neutral or slightly REE minerals produced from mineral sands deposits. acid solutions with an ion-exchange agent such Since the grains are unconsolidated (or much less as ammonium sulfate or EDTA (ethylenedi- consolidated), these deposits are amenable to aminetetraacetic acid) (Bao and Zhao, 2008). In gravity, magnetic and electrostatic (high tension) situ leaching was developed in Ganzhou, Jiangxi separation methods (Figure 13.8). Many, especially in an attempt to avoid the despoliation caused by contemporary beach deposits, require no crushing pond leaching. However, it is more suited to the or grinding. Monazite and xenotime are separated softer ionic clays in the areas around Dingnan from quartz, which is the main constituent of the and Longnan. Pond leaching is still used for the sands; ilmenite, leucoxene, rutile and zircon are also harder granite clay deposits in Guangdong. Rare earth elements 327

Bastnäsite concentrate

Acid solution contains Dissolve REE La, Ce and lesser amounts minerals of the other REE

Heavy REE

Organic Organic Organic Organic

Acid MixtureAcid MixtureAcid Mixture Acid Figure 13.9 Schematic owsheet for the extraction of rare earths from bastnäsite Light REE at Mountain Pass. (After Molycorp Inc., Separate the REE by multi-stage solvent extraction 2011b.)

The exploitation of new types of deposits will such as sulfates. Multiple stages are then used to require different methods of concentration. separate the REE. Selective oxidation can be used There are no commercial-scale processes demon- for Ce, and also for Pr and Tb, which have a poten- strated for any of the REE minerals apart from tial +4 oxidation state, whilst selective reduction monazite, xenotime, bastnäsite and loparite. works for Eu, Sm and Yb because they have a Other minerals have been mined on a much potential +2 valency. Otherwise, the processes smaller scale or separated in laboratory tests exploit the property of the decreasing cation size (Table 13.2). During development of new across the REE series (i.e. the lanthanide contrac- deposits, consideration of geometallurgy and tion) during processes of fractional crystallisation, potential extraction methods is critical. For fractional precipitation, ion exchange and solvent example, otation tests on ancylite from the extraction. Bear Lodge deposit, USA were unsuccessful. A At the Molycorp operation at Mountain Pass, owsheet involving crushing, scrubbing (attri- USA, bastnäsite concentrate is roasted in air at tion), hydrochloric acid leaching to dissolve the 620 °C to drive off carbon dioxide and oxidise Ce3+ ancylite (and presumably other carbonates), then to Ce4+. This calcined product is then treated a separate stage to remove uranium and thorium, with hydrochloric acid to dissolve the non- and final production of REE oxalate product has cerium REE. The cerium concentrate can be sold. been devised (Richardson et al., 2010). Further processing is then used to separate Eu (originally for cathode ray tubes) and other REE (Gupta and Krishnamurthy, 2005). The processing Extraction and separation of the REE and extraction operation is now being updated Separation of the individual REE is the most diffi- to make it more water, resource and energy effi- cult part of the extraction process. There are many cient (Molycorp Inc., 2011a). In the new process, routes for dissolving the various REE minerals the bastnäsite will be dissolved in acid and then and then separating the REE from each other subjected to counter-current solvent extraction (Gupta and Krishnamurthy, 2005). Most involve using immiscible organic and acid solvents an initial acid attack to create soluble REE salts (Figure 13.9). 328 frances wall

In China, bastnäsite concentrate is processed samarium–cobalt (SmCo5) is also an important with 98 per cent sulfuric acid at 500 °C in a magnetic material. Small quantities of Dy and Tb rotary kiln. This converts REE to sulfates. They are substituted into Nd2Fe14B magnets in order to are then converted to hydroxides and dissolved improve their properties and performance at in hydrochloric acid before separation and higher temperatures. purification using solvent extraction (Gupta and REE make good phosphors, and this is Krishnamurthy, 2005). another high-value sector. In addition to Eu, Y Xenotime is treated with concentrated (93 per is used in various phosphor applications, cent) sulfuric acid to make water-soluble REE including televisions and computer monitors. sulfates (Gupta and Krishnamurthy, 2005). Yttrium is also important in yttrium aluminium garnet (YAG) lasers in high-temperature super- conductors, and in specialist alloys. Other Specifications and uses applications of REE in ‘green’ technologies include their use in low-energy uorescent REE have a wide variety of uses and, although lighting, their use as the ‘metal’, e.g. La, in they are often used in small quantities, they metal hydride batteries used in rechargeable have become pervasive in many technologies, batteries, including in hybrid vehicles, and use especially those for improving energy efficiency in fuel cells. Magnetic refrigeration may be an and in digital technologies (Table 13.4). Cerium important future application. oxide polishing powders, still used today for vir- tually all polished glass products, and the Ce metal alloys in lighter ints were some of the Recycling, re-use and resource efficiency early applications of REE in the 1950s. Europium gave the red colour to colour televisions and was Low prices have been the main contributing one of the main products from the Mountain factor to a low rate of recycling or re-use of REE Pass mine during the 1970s and 1980s. The products. Less than one per cent of REE, mainly small headphones first used in the ‘Walkman’, magnets, were thought to have been recycled and in all portable music devices since, are from end-of-life products in 2010 (Kara et al., thanks to the small high-strength REE-bearing 2010). The recent rising prices and supply prob- permanent magnets, which are now essential lems have encouraged new research and components in all computer disk drives and development in this area. There is also research many other technologies. The uses of REE can taking place to reduce the amounts of REE used be grouped into eight categories (Table 13.4). in various processes. Recycling is often difficult Catalysts form one of the largest market sectors because of the way that REE are incorporated as by volume (Table 13.4). For example, Ce is a small components in complex items or are part of component of catalytic converters in cars complex materials. The processes required are because of its ability to form non-stoichiometric energy intensive and complex (Schüler et al.,

CeO2, which helps oxidise unburned hydrocar- 2011). Some REE are incorporated in components bons. Lanthanum and cerium halides help to that typically have short lifetimes such as mobile stabilise the zeolite structures used to crack phones, computer disk drives and low-energy petroleum. light bulbs but others are in products with lifes- The use of REE in permanent magnets is the pans of 10–20 years or more, such as wind tur- highest value market sector and is forecast to bines and vehicles. increase in size, not least because of the use of The advantages of recycling are that the Nd-bearing magnets in large wind turbines. REE are already separate from radioactive ele-

Neodymium–iron–boron (Nd2Fe14B) is the stron- ments (and thus one of the main problems in gest permanent magnetic material known and mining raw materials is avoided), the energy Rare earth elements 329

Table 13.4 Uses of rare earth elements, estimated global rare earth oxide (REO) demand in 2012 and forecast REO demand in 2016. (Data from Kingsnorth, 2013.)

Forecast Forecast Demand 2012 demand demand Principal Demand 2012 (market 2016 2016 (market Category Application elements used (tonnes REO) share %) (tonnes REO) share %)

Magnets Motors Nd, Pr, Tb, 22,500 20 33,000 21 Disc drives Dy, Sm Power generation Actuators Microphones and speakers MRI Automotive parts Communication systems Electric drive Frictionless bearings Magnetic refrigeration Metal alloys Hydrogen storage (NiMH Ce, La, Pr, 22,000 19 30,000 19 batteries, fuel cells) Nd, Sm, Steel Sc, Y Aluminium/magnesium Cast iron Superalloy Catalysts Catalytic converter Ce, La, Pr, 22,000 19 26,000 15 Chemical processing Nd, Y Diesel additives Petroleum refining Polishing Polishing compounds Ce, La, Nd 19,000 16 25,000 15 Glass Optical glass La, Ce, Pr, 7,500 7 9,000 6 UV resistant glass Nd, Eu, Thermal control mirrors Gd, Dy, Colourisors Ho, Er, Y Phosphors and Display phosphors Eu, Tb, Y, La, 8,500 7 9,000 7 pigments Medical imaging Dy, Ce, Pr, Lasers Gd, Nd, Fibre optics Ce, Er, Eu Fluorescent lighting Optical sensors Pigments LEDs Ceramics Capacitors Y, Ce, La, 6,500 6 8,000 5 Sensors Pr, Nd Colourants Scintillators Other Water treatment Gd 7,000 6 20,000 12 Fertiliser Medical tracers Coatings Nuclear reactors Total 115,000 100 160,000 100 330 frances wall and resources requirement can be lower for accumulate in the bone, liver, heart, and lung recycling than mining, and the dependency on (U.S. Environmental Protection Agency, 2009) imports of REE is reduced (Schüler et al., and prolonged exposure has been associated 2011), particularly important for Europe, Japan with lung problems such as interstitial lung and USA. The disadvantage is that recycled disease or pneumoconiosis. Cerium poisoning materials, such as magnets, may not perform causing endomyocardial fibrosis has been so well as new materials (British Geological reported in Kerala, India, where the local Survey, 2011). Re-use is also potentially pos- population eat root vegetables and thus ingest sible, for example, of rare earth magnets. The soil containing high levels of Ce. However, a problem is that product design changes rapidly study of an area in Uganda with elevated Ce in and new sizes and styles of magnets are usu- soil in Uganda did not find a similar link ally required, so this kind of re-use would (Smith, 1998). Cerium is common in other require a ‘cradle-cradle’ design philosophy products but the exposure is much less, e.g. as from the outset. catalysts for auto-emissions control, decolour- isers for glass. There is now some evidence of other REE from anthropogenic sources in the Substitution natural environment. Gadolinium anomalies have been found in river waters following use For many applications of the REE, there are either of Gd in medical imaging (Kulaksız and Bau, no known substitutes or the substitutes do not 2007). These levels of Gd have not been corre- perform as well (Haxel et al., 2002). There is also lated with any adverse health effects but the the problem that some substitutes may be as Gd-based contrast agents administered during expensive as the REE. An example of substitution medical imaging can cause nephrogenic is the use of less strong samarium–cobalt magnets systemic fibrosis in patients with impaired or, where size is not so important, iron magnets in kidney function (U.S. Food and Drug some applications of neodymium–iron–boron. Administration, 2010). Limits for levels of REE The hardest applications for which to find substi- permitted in discharges are generally lacking tutes are the uses that require the optical and and there is little regulation of factories that chemical properties of REE, such as colouring and use REE. In China, REE have been used in fer- decolourising agents, catalysts and phosphors. tilisers for many years and also fed to livestock However, reduction of the amount of REE used is as a growth promoter, with the conclusion that proving possible in certain applications. For there is no harmful effect to humans or ani- example, the amount of REE in some catalysts mals (Baotou National Rare-earth Hi-tech used to crack petroleum was reduced by up to a Industrial Development Zone, 2011). REE- half in 2011 as prices escalated in response to con- bearing fertilisers have been distributed on cerns about the security of supplies from China. 67,000 km2 of land in China thus providing an opportunity for a future case study of REE dis- tribution in the environment. Environmental aspects The main environmental concerns regarding REE mines and processing facilities are not usu- In general, the toxicity of the rare earths them- ally the REE themselves but the presence of the selves is thought to be low. However, there radioactive elements, thorium and uranium, and are some health risks documented (U.S. the chemicals used to extract and separate the Environmental Protection Agency, 2012). The REE from the ore minerals. Thorium, and to a main reported concern is the use of cerium pol- lesser extent uranium, may be present as discrete ishing powders in lenses and other glass objects mineral phases that can be separated from the and jewellery. Once in the body, Ce tends to REE minerals at the beneficiation stage. However, Rare earth elements 331 there will almost always be some thorium, and 2011). The new rules govern chemicals used in lower amounts of uranium, incorporated into the the processing and production of rare earths, such REE mineral lattice and hence present in any as an emission cap on ammonia nitrogen in water final concentrate and in the factory where extrac- of 25 mg/l from 1 January 2012, reduced from tion of REE takes place. Processing of monazite previous limits of 300–5000 mg/1. In addition to from mineral sands, which can contain about these considerations, processing uses consider-

10 wt% ThO2, has been banned in some coun- able energy to grind ores fine enough for ota- tries, such as Australia. The ore at Mountain Pass tion. Processes that rely on acid dissolution may contains 0.02 wt% Th and 0.002 wt% U; the tail- also have high carbon footprints as a result of the ings at Mount Weld will contain 0.05 wt% ThO2 energy used to produce the acid and require and 0.003 wt% UO2 (Schüler et al., 2011). Areas appropriate safety measures. All of the processing around Chinese rare earth mines have been routes require water. polluted and the Chinese government is now try- ing to regulate its industry and encourage larger companies to take over the small REE leaching World resources and production operations on ion adsorption deposits, which have low thorium levels but have caused chemical The best estimates of world reserves of REE are pollution. China’s Ministry of Environment about 114 million tonnes of REO (Cordier, 2011) Protection brought new rules into effect in (Table 13.5). China has the largest proportion (48 October 2011 to limit pollution from rare earth per cent) of reserves, mainly in the Bayan mines (China Daily, Europe, China Daily Europe, Obo deposit, but also in other carbonatite and

Table 13.5 Production of rare earth oxide (REO) concentrates in 2010 and estimates of reserves. (Production data from Roskill Information Services Ltd; reserve data from Cordier, 2011.)

Country Comment REO production (tonnes) REO reserves (tonnes)

China Mainly from bastnäsite from Baotou 120,000 55,000,000 and ion adsorption clay from southern provinces stockpiled ore in Sichuan ca.10,000 ‘unoffcial’ sources ca.15,000 Russia (reserve: CIS) In chemical concentrates from mine 1,898 19,000,000 output in 2009 USA From stockpiled ore. Mining restarted 1,883 13,000,000 in December 2010 India Chemical concentrates from minerals 25–50 3,100,000 sands tailings Brazil Monazite stockpiled by-product from 550* 48,000 rutile production Australia Some mining has taken place at the ? 1,600,000 Lynas operation at Mount Weld — the ore was stockpiled until the concentrator was completed. Others Small amounts of monazite and ? 22,030,000 xenotime from south-east Asia, probably shipped to China for processing Total 113,778,000 332 frances wall alkaline rock deposits. It is significant to note markedly during this time. Early production of that a large proportion of the world’s heavy REE REE came from mineral sands in Western reserves is in ion adsorption deposits (Tables 13.3 Australia and India; then the Mountain Pass and 13.5) and also some mineral sands in China. deposit in USA was discovered and became the The USA has reserves at Mountain Pass and in world’s main supplier of REE during the 1960s various projects under development such as Bear and 1970s, set up particularly to produce Eu for Lodge, Wyoming (Tables 13.3 and 13.5). CIS, televisions. In the 1980s Bayan Obo in China including Russia, has significant reserves in its became the main supplier, joined by other large alkaline complexes at Khibiny and Lovozero Chinese producers, to reach the current position and at Kutessay II in Kazakhstan (Tables 13.3 and where China supplies some 97 per cent of the 13.5, Figure 13.4). Additional resources include world’s REE (Figure 13.10). As recently as 1992 the weathered carbonatite deposit at Tomtor though, there was a wider variety of producers, in Siberia. Other reserves and resources are in with USA, Australia and Russia significant, and the large alkaline complexes in Canada and China producing just 38 per cent of the world’s Greenland. There are many carbonatite complexes rare earths (Figure 13.10). By 2010, only Russian in Africa that contain REE (Figure 13.4), although loparite and Indian mineral sands were still con- they tend to be smaller than the carbonatites cur- tributing any appreciable REE to the total. rently mined. The reserve figure will vary with Estimates of current production (Table 13.5) time, as new exploration projects are completed vary slightly because of factors such as unofficial and as economics and new technologies permit (illegal) production from China and inconsistency the consideration of additional deposit types in in recording production that is stockpiled rather the reserve category. There are also potentially than sold (e.g. monazite in Brazil is a waste prod- large amounts of REE in lower-grade deposits uct of the benefication of rutile and is stockpiled). such as Olympic Dam, for example (Table 13.3, Other potential producers such as Australia do Figures 13.4 and 13.5) and in alkaline complexes not record their stockpiles as ‘production’ and so as well as carbonatites (Table 13.3, Figures 13.4 figures are harder to compare. Some mineral sand and 13.5). The reserve figure does not include producers return monazite to the ground, diluted novel sources of REE such as ocean- oor deposits. with waste material rather than stockpile radio- An estimate of the long-range extractable active material. geological resource is difficult to obtain. One As noted above there are many REE resources hypothetical approach is to consider apatite, outside China. If demand is sufficient and prices which is much more abundant than the common high enough, the other deposit types could be rare earth minerals and therefore is probably host used, including apatite as described above, and, in to a higher proportion of crustal REE. In alkaline the longer term, sea oor deposits. By-product igneous rocks and carbonatites, apatite com- REE from other ores and extraction from waste monly contains 1 wt% total REE. Beneficiation are also potential future sources of supply. and extraction technologies are well established for apatite. Assuming there are about 2500 alkaline complexes at the Earth’s surface, Future supplies averaging five km2 in area and containing 3 wt% apatite, with an REO content of 1 wt%, the total Exploration activity for REE has increased greatly REO content of these complexes would be 10,000 in recent years. Many deposits are now under million tonnes of REO. This is equivalent to active exploration, including most of those 50,000 years supply of REO at a projected annual shown in Figure 13.4, and many more. A new consumption rate of 200,000 tonnes per year. trend may be that non-carbonatite sources Production of REE has increased consistently become mines, especially targeting heavy REE. since the 1950s but the producers have changed With the higher prices, production of REE as a Rare earth elements 333

160,000

140,000

120,000

100,000

80,000

60,000

40,000 REE production (tonnes REO) 20,000

0 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

China USA Russia Australia Other

Figure 13.10 Rare earth production between 1992–2010. (Data from British Geological Survey World Mineral Statistics database.)

by-product may also become economic in more clear if there is still any REE production from cases. Recycling is also increasingly economic this source now that the Silmet factory has been and, together with Government and company sold to Molycorp. funding to reduce reliance on China, it is likely that more recycling will take place. Two new projects have recently started pro- World trade duction. The re-opening of the Mountain Pass mine is the first major development. The second China is by far the main exporter of REE metals project, at Mount Weld in Western Australia, and compounds (Figure 13.11), re ecting its commenced mining in 2011 and the new REE dominance in the mining of REE. China has facility in Malaysia, which processes the concen- pursued a strategy of building up its own REE trates from Mount Weld, began operation in early industry so that it manufactures the high-value 2013. Market analysts list up to 20 projects on a REE intermediate products and attracts foreign variety of deposit types that are at an advanced companies to operate in China rather than stage of exploration and development (Table 13.3). exporting raw materials. The USA, Germany, It is noteworthy that although the first two France and Austria are the largest importers. mines opening (Mountain Pass and Mount Weld) None of these countries has mined REE in are in carbonatite, there is a wider range of recent years and so they are totally reliant on deposits in the next phase of projects likely to Chinese exports. This situation will change in reach fruition. REE have been produced from the USA as Mountain Pass re-opens but Europe will niobium ore, loparite, mined at Lovozero, Kola still need to import practically all of its REE. Peninsula, Russia, with REE production taking The imports to Estonia are related to the Silmet place at the Silmet factory in Estonia. It is not factory. This factory has now been bought 334 frances wall

40

30 Exports 20

10

0 Thousand tonnes

10 Imports 20 USA Brazil China Japan Russia France Austria Estonia Belgium Germany Sri Lanka Kazakhstan Rep. of Korea

Figure 13.11 Imports and exports of rare earth metals and compounds, 2009. (Data from British Geological Survey World Mineral Statistics Database and UN Comtrade, 2013.) by Molycorp, the operator of Mountain Pass, so to produce these elements it will be necessary to that it now has three sites that can process also produce more La and more Ce or Y. REE: Mountain Pass (California, USA), Tolleson (Arizona, USA) and Sillamae (Estonia, previ- ously AS Silmet). Prices There is no international metals exchange for REE and many mines and deposits are now verti- Prices for REE vary considerably (Figure 13.12). The cally integrated into the supply chain, with large main controlling factors are the nature and purity of companies such as Toyota and Siemens under- the product, the abundance of the particular REE, taking joint ventures to source raw materials and and demand from current uses. The cheapest prod- thus gain security of supply. Mines such as Mount ucts are mixtures of the REE, such as mischmetal Weld have their own processing factories so that (La 35%, Ce 65%) or mixed REE carbonates they will sell individual REE oxides and metals (Figure 13.12). Generally, the oxides are easier and rather than intermediate products. REE are sim- cheaper to produce than the metals. The cheapest ilar to industrial minerals in this respect. As a pure REE are the most abundant elements, La and result of this, there is little trade in REE mineral Ce, which are used in many applications. Although concentrates or leach products (carbonates or Nd is abundant it is more expensive because of the oxalates). demand for Nd2Fe14B magnets. The most expensive Demand for REE is predicted to continue REE are Eu, Dy and Tb because of their low abun- to rise. The U.S. Department of Energy (2011) dance in the majority of REE ores and high demand highlights five REE (neodymium, europium, dys- for these elements in phosphors and magnets. prosium, terbium and yttrium) as critical in the There is no international metal exchange for near- (up to 2015) and medium-term (2015 to 2025) REE. Prices are determined by the actual trades for use in energy technologies. It is notable that that take place and are available on various none of the most critical REE are the main constit- commercial and company web sites, quoted in uents of any common REE minerals and in order Chinese currency and US dollars FOB1 China. Rare earth elements 335

Mischmetal La 35% Ce 65% Ce carbonate 45% REO Yttrium Terbium Samarium Praesodymium Neodymium Lanthanum Gadolinium Europium Figure 13.12 Prices of rare earth metals, Dysprosium oxides, mischmetal and mixed light REE Cerium carbonate in January 2013. Prices are FOB China, 99% metal min., except Y 99.9% 1 10 100 1000 10000 min.; oxide: FOB China, 99% min., except Price (US$/kg) Eu 99.9%, Y, 99.999%. (Data from Metal Oxides Metals Other Pages, www.metal-pages.com.)

350

300

250

200

150 US$ per kilogram per US$ 100

50

0 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2006 2007 2008 2009 2010 2011 2012

Ce La Nd Pr

Figure 13.13 Quarterly average prices for selected LREE. Prices are for oxide FOB China 99% min. (Data from Metal Pages, www.metal-pages.com). Ce, cerium; La, lanthanum; Nd, neodymium; Pr, praseodymium.

These are export prices but between 60–70 per Prices of all of the REE increased very rap- cent of all rare earth consumption is now in idly during 2010 and 2011 owing to supply China and here the REE may be traded at differ- problems as China restricted its export quotas ent internal prices, which may impact on the (Figures 13.13 and 13.14). Demand reduced and price of part-processed or finished goods for prices fell at the end of 2011 and continued to export. decline throughout 2012. Nevertheless, prices 336 frances wall

6000

5000

4000

3000

US$ per kilogram 2000

1000

0 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2006 2007 2008 2009 2010 2011 2012

Dy Eu Tb Y

Figure 13.14 Quarterly average prices for selected HREE and Y. Prices are for oxide FOB China 99% min. for Dy and Tb, 99.9% min. for Eu, and 99.999% min. for Y. (Data from Metal Pages, www.metal-pages.com). Dy, dysprosium; Eu, europium; Tb, terbium; Y, yttrium. for most REE at the end of 2012 remained signif- industry. This kind of vertical integration is icantly higher than the levels in 2008 prior to the likely to remain important in the market as a onset of the global economic recession. For whole, with more end users buying into the example, at the beginning of 2008 the price of supply chain to secure the resources they need. dysprosium oxide was about US$93/kg compared However, it will be interesting to see if a market with US$615/kg at the end of 2012. Over the for intermediate mixed carbonate leached prod- same period the prices of La, Ce and Tb oxides ucts or mineral concentrate does develop because more than doubled. this would permit more exibility and allow smaller players to produce REE, including processing by-products and wastes. Outlook In southern China developments are taking place outside Jiangxi and Guangdong, tradition- There is a wide variety of rare earth deposits, con- ally the main centres for HREE production. State- taining plenty of REE for the foreseeable future. owned companies and research institutes (e.g. The challenge is to diversify supply quickly Grirem) are involved in developing new sites for enough to overcome the current market prob- mining, separating and downstream processing in lems and allow the current and new uses of REE Hunan, Fujian and Guangxi. In Guangxi new to continue and expand. China is building zones mining licences have been granted that could around the lesser known Maoniuping and replace exhausted HREE resources in other parts Weishan mines for ‘deep processing’ of rare earths of the south. similar to that at Baotou near the Bayan Obo The desire of end users to prevent another mine. It aims to make a vertically integrated supply crisis may help to increase the number Rare earth elements 337 of REE producers, including the encouragement British Geological Survey (2011) Rare Earth Elements of projects in Europe. Considerable research is Mineral Profile, British Geological Survey, required to establish production from new types Nottingham, United Kingdom. http://www.bgs.ac. of deposits and to conduct more REE separation uk/mineralsuk/search/home.html Castor, S.B. (2008) The Mountain Pass rare-earth car- and purification outside China. The most attrac- bonatite and associated ultrapotassic rocks, tive deposits to find will be those with high pro- California. The Canadian Mineralogist 46, 779–806. portions of the ‘critical REE’, i.e. the most Cordier, D.J. (2011) Rare Earths, Mineral Commodity highly sought after heavy REE and Nd (i.e. Nd, Summaries. United States Geological Survey, Eu, Dy, Tb, Y). 128–126 China is bringing in new environmental regu- China Daily Europe (2011) China’s rare earth industry lations in order to reduce pollution from rare faces reshuffle: experts. http://europe.chinadaily. earth mines but has long-standing problems to com.cn/china/2011-03/06/content_12126297.htm overcome. With a diverse range of deposit types Förster, H-J. (1998) The chemical composition of available, once there are alternatives to China, REE Y-U-rich accessory minerals in peraluminous responsible sourcing from the most environmen- granites of the Erzgebirge-Fichtelgebirge region, Germany. Part II: Xenotime. American Mineralogist tally friendly mines may also become an issue. 83, 302–1315. Recycling rates will increase if the price for REE Frietsch, R. and Perdahl, J-A. (1995) Rare earth elements remains sufficiently high. in apatite and magnetite in Kiruna-type iron ores and some other iron ore types. Ore Geology Reviews 9, 489–510. Gupta, C.K. and Krishnamurthy, N. (2005) Extractive Note Metallurgy of Rare Earths. CRC Press, Boca Raton, USA. 1. FOB, Free on Board. The seller is responsible for the Haxel, G.B., Hedirk, J. B. and Orris, G.J. (2002) Rare cost of delivering goods to the ship. The buyer is Earth Elements – Critical Resources for High responsible for transportation and insurance costs Technology, USGS Fact Sheet 087-02. US Department from that point. of the Interior, United States Geological Survey. Hein, J.R., Conrad, T. and Koschinsky, A. (2011) Comparison of land-based REE ore deposits with References REE-rich marine Fe-Mn crusts and nodules. Mineralogical Magazine 75, 1000. Andreoli, M.A.G., Smith, C.B., Watkeys, M., Moore, International Mineralogical Association (2012). http:// J.M., Ashwal, L.D. and Hart, R.J. (1994) The geology www.ima-mineralogy.org/Minlist.htm of the Steenkampskraal monazite deposit, South Kaltosoyannis, N. and Scott, P. (1999) The f Elements. Africa; implications for REE-Th-Cu mineralisation in Oxford Chemistry Series, Oxford University Press, charnockite-granulite terranes. Economic Geology Oxford, United Kingdom. 89, 994–1016. Kara, H., Chapman, A., Crichton, T., Willis, P. and Bao, Z. and Zhao, Z. (2008) Geochemistry of minerali- Morley, N. (2010) Lanthanide Resources and sation with exchangeable REY in the weathering Alternatives. A report for Department for Transport crusts of granitic rocks in South China. Ore Geology and Department of Business, Innovation and Skills, Reviews 33, 519–535. Oakdene Hollins Research and Consulting. Baotou National Rare-earth Hi-tech Industrial Kato, Y., Fujinaga, K., Nakamura, K. et al. (2011) Deep- Development Zone (2011) Rare Earth: an introduc- sea mud in the Pacific Ocean as a potential resource tion. http://www.rev.cn/en/int.htm Downloaded for rare-earth elements. Nature Geoscience, Advance 11.12.11. Online Publication, DOI: 10.1038/NGEO1185 Bayliss, P. and Levinson, A.A. (1988) A system for Kulaksız, S. and Bau, M. (2007) Contrasting behaviour nomenclature of rare-earth mineral species: Revision of anthropogenic gadolinium and natural rare earth and extension. American Mineralogist 73, 422–423. elements in estuaries and the gadolinium input into 338 frances wall

the North Sea. Earth and Planetary Science Letters Smith, B. (1998) Cerium and Endomycardial fibrosis in 260, 361–371. tropical terrain. Project summary report. Technical Kingsnorth, D.J. (2013) Rare earths forecast update. report WC/98/26, British Geological Survey. Presentation to AusIMM 2013 Critical Minerals Smith, M. and Wu, C. (2000) The Geology and Genesis Conference, Rare Earths – Is Supply Still Critical, of the Bayan Obo Fe-REE-Nb Deposit: A Review. In: Perth, Western Australia, 4–5 June 2013. Porter, T. M. (ed.) Hydrothermal Iron Oxide Copper- McDonough, W.F. and Sun, S. s- (1995) The composition Gold and Related Deposits: A Global Perspective, of the Earth. Chemical Geology 120, 223–253. Volume 1, PGC Publishing, Linden Park, South Miyawaki, R. and Nakai, I. (1996) Crystal chemical Australia, Australia, 271–281. aspects of rare earth minerals. Rare Earth Minerals: Smith, M.P., Henderson, P., Campbell, L.S. (2000) Chemistry, origin and ore deposits. In: Jones, A.P., Fractionation of the REE during hydrothermal Wall, F. and Williams, C.T. (eds.) Mineralogical processes: Constraints from the Bayan Obo Fe-REE-Nb Society Series 7, Chapman and Hall, London, deposit, Inner Mongolia, China. Geochimica et 21–40. Cosmochimica Acta 64, 3141–3160. Molycorp Inc. (2011a) The Many Uses of Rare Earths in Stans Energy Corp. (2012) Kutessay II Geology, Advanced Technologies. http://www.molycorp.com/ Historical Geological information of the Kutessay II GreenElements/RareEarthsManyUses.aspx. REE-Th-Nb-Ta deposit. http://www.stansenergy.com/ Molycorp Inc. (2011b) Separating rare earths at projects/?page_id=258 Mountain Pass http://www.molycorp.com/resources/ Taylor, S.R. and McClennan, S.M. (1985) The science-center Continental Crust: Its Composition and Evolution. Oreskes, N. and Einaudi, M.T. (1990) Origin of rare Blackwell Scientific Publications, Oxford. earth element-enriched hematite breccias at the United Nations (UN) Comtrade (2013) Commodity Trade Olympic Dam Cu-U-Au-Ag deposit, Roxby Downs, Statistics Database (http://comtrade.un.org/db/). South Australia. Economic Geology 85, 1–28. U.S. Department of Energy (2011) Critical materials Orris, G.J. and Grauch, R.I. (2002) Rare Earth Element strategy, December 2011, DOE PI-0009. Mines, Deposits, and Occurrences. United States U.S. Environmental Protection Agency (2009) Geological Survey Open-File Report, 02–189. Toxicological review of cerium oxide and cerium Petrov. S.V. (2004) Economic deposits associated with compounds (CAS No. 1306-38-3) In: Support of the alkaline and ultrabasic complexes of the Kola Summary Information on the Integrated Risk Peninsula. Phoscorites and Carbonatites from Mantle Information System (IRIS) 2009, EPA/635/R-08/002F, to Mine: the Key Example of the Kola Alkaline Washington, DC. Province. In: Wall, F. and Zaitsev, A.N. (eds.) U.S. Environmental Protection Agency (2012) Rare Mineralogical Society Series 10, Mineralogical Earth Elements: A Review of Production, Processing, Society, London, 469–490. Recycling, and Associated Environmental Issues. Red Mud Project (2006) The Red Mud Project. http:// EPA/600/R-12/572, Cincinatti, OH. www.redmud.org/Bibliography.html#rareelements. U.S. Food and Drug Administration (2010) FDA Drug Richardson, M.P., Noble, A.C., Roman, R. and Clark, Safety Communication: New warnings for using J.G. (2010) Technical report: preliminary economic gadolinium-based contrast agents in patients with assessment (scoping study) of the Bear Lodge Rare- kidney dysfunction. http://www.fda.gov/Drugs/ earths project - a national instrument 43-101 report. DrugSafety/ucm223966.htm#sa Crook County, Wyoming. Prepared for Rare Element van Emden, B., Thornber, M.R., Graham, J. and Resources Ltd. Lincoln, F.J. (1997) The incorporation of actinides in Schüler, D., Buchert, M., Liu, R., Dittrich, S. and Merz, C. monazite and xenotime from placer deposits in (2011) Study on Rare Earths and their Recycling. Western Australia. The Canadian Mineralogist 35, Öko-Institute. V., Final Report for The Greens / EFA 95–104. Group in the European Parliament. Wall, F. and Mariano, A.N. (1996) Rare earth minerals in Simandl, L., Simandl, G.J. and Fajber, R. (2011) Rare carbonatites: a discussion centred on the Kangan— Earth Elements (REE) recovery as a by-product of kunde carbonatite, Malawi. Rare Earth Minerals: fertiliser production from sedimentary Phosphate Chemistry, origin and ore deposits. In: Jones, A.P., Wall, deposits – Conceptual evaluation. Mineralogical F. and Williams, C.T. (eds.) Mineralogical Society Series Magazine 75, 1878. 7, Chapman and Hall, London, 193–225. Rare earth elements 339

Wall, F. and Zaitsev, A.N. (2004) Rare earth minerals in dykes at Lofdal, Namibia: unusually low light to Kola carbonatites. In: Wall, F. and Zaitsev, A.N. (eds.) heavy REE ratio in carbonatite, and the first dating of Phoscorites and Carbonatites from Mantle to Mine: xenotime overgrowths on zircon. Canadian the Key Example of the Kola Alkaline Province. Mineralogist 46, 1251–1267. Mineralogical Society, 43–72. Wu, C. (2008) Bayan Obo Controversy: Carbonatites Wall, F. Niku-Paavola, V.N., Storey, C., Mueller, A. and versus Iron Oxide-Cu-Au-(REE–U). Resource Geology Jeffries, T. (2008) Xenotime-(Y) from carbonatite 58 (4), 348–354. 14. Rhenium

TOM A. MILLENSIFER1 , DAVE SINCLAIR2 , IAN JONASSON2 AND ANTHONY LIPMANN3

1 Executive Vice President and Technical Director of Powmet, Inc., Rockford, Illinois, USA 2 Formerly research scientist at Geological Survey of Canada, Ottawa, Ontario, Canada 3 Managing Director, Lipmann Walton & Co Ltd, Walton on Thames, Surrey, UK

Introduction Physical and chemical properties

Rhenium is reported by many sources to be the Naturally occurring rhenium, element number last natural element to be discovered (Habashi, 75 in the Periodic Table, consists of two isotopes: 1996). Its discovery in 1925 is credited to Ida 187Re, which accounts for 62.6 per cent of the Tacke, Walter Noddack and Professor Otto Berg. total, and 185Re, which makes up the balance of They also claimed the discovery of masurium, 37.4 per cent. 187Re is radioactive with a half-life Ma, the neighbour of rhenium, which today is of approximately 4.3×1010 years. However, the known as technetium, Tc. However, this was dis- beta radiation emitted is very weak, 0.3 MeV, and puted and the discovery of technetium was not cannot penetrate human skin. The chemical confirmed until 1936. properties of rhenium resemble the metals in the Much of the early work developing rhenium manganese group (Group 7) of the Periodic Table. processing was done in the USA by Kennecott and The physical properties, however, are much more many patents related to rhenium recovery from similar to those of the refractory metals of Groups molybdenite processing were granted to this 5 and 6, particularly molybdenum and tungsten. company. Shattuck Chemical of Denver Colorado It was for this reason that the Noddacks used subsequently licensed the Kennecott processes and molybdenum concentrates as their source for began rhenium recovery in 1960 from the roasting larger quantities of rhenium (Habashi, 1996). of molybdenite concentrates derived from mines Rhenium is considered a refractory metal because operating in western USA. Since then rhenium of its high melting point of about 3200 °C, with production has taken place in several other coun- only tungsten having a higher melting point. tries with Chile now the dominant supplier. In However, in contrast to other refractory metals, recent years production of rhenium from primary rhenium does not form carbides. Platinum and sources has grown substantially to current levels of osmium have greater specific gravities. Selected about 44.8 tonnes (MMTA, 2012a). Today the main properties of rhenium are listed in Table 14.1. use of rhenium is in superalloys for gas turbines in Rhenium exhibits several valences from –1 to aircraft and in land-based applications, while its +7, with the most common being +7, +6, +5 and use in petroleum-reforming catalysts accounts for +4. It easily changes from one valence to another, about 10 per cent of total consumption. a property which makes it ideal for use as a

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Rhenium 341

Table 14.1 Selected properties of rhenium. (Modified Table 14.2 Rhenium-bearing minerals. Most rhenium after Treichel, 2000.) is produced from molybdenite.

Property Value Units Name Formula Rhenium content

Symbol Re Rheniite ReS2 74%

Atomic number 75 Tarkianite (Cu,Fe)(Re,Mo)4S8 49 to 56%

Atomic weight 186.21 Dzhezkazganite ReMoCu2PbS6 22% 3 Density at 25°C 21023 kg /m Molybdenite MoS2 <10 ppm to 11.5%

Melting point 3180 °C Castaingite CuMo2S5 up to 1%

Boiling point 5926 °C Uraninite UO2 up to 2700 ppm 2+ Hardness (Mohs scale) 7.0 Gadolinite Y2Fe Be2Si2O10 up to 1 ppm Specific heat capacity at 25°C 0.14 J/(g °C) Electrical resistivity at 25°C 18.40 nΩ m Thermal conductivity 48 W/(m °C) ranging from 0.2 to 2 ppb Re. The rhenium Young’s modulus 463 GPa content of common igneous rocks, from basalts to granites, is typically in the range of 0.2 to 1 ppb Re (e.g. Morris and Short, 1969). Comparable con- catalyst. In the 1960s it was observed that when centrations of rhenium occur in most sedimen- rhenium is alloyed with molybdenum or tung- tary rocks, except in reduced, organic-rich rocks, sten, the resultant alloy exhibits or retains the which can have rhenium contents several orders ‘best’ properties of both metals individually of magnitude greater than the crustal average (e.g. and none of the ‘poor’ properties of either Lipinski et al., 2003). (J. Port, Cleveland Refractory Metals, personal communication, 1965). For example, heating tungsten above its recrystallisation temperature, Mineralogy 1200 °C, and then cooling it to room temperature causes it to become brittle. The addition of rhe- Because rhenium is such an extremely rare nium to tungsten and molybdenum significantly element, rhenium minerals are also rare and reduces the brittle characteristics, increasing the relatively few minerals concentrate rhenium to recrystallisation temperature, the ductility and a significant degree (Table 14.2). It should be the ultimate tensile strength of the alloys stressed that minerals rich in rhenium are gener- (Leonhardt, 2009). ally found in very small quantities and are not Rhenium does not have a ductile-to-brittle commercially viable sources of rhenium at the transition temperature. It retains its ductility present time. For example, rhenium occurs at ele- from low sub-zero temperatures to very high vated levels in some minerals deposited around temperatures, making it ideal for use in space high-temperature volcanic fumaroles and, in at propulsion applications. Rocket-thruster nozzles least two locations, it is present as the sulfide have been reported to withstand more than mineral rheniite. One of these locations is Usu 100,000 thermal fatigue cycles from room tem- volcano, Hokkaido, Japan (Bernard and Dumortier, perature to above 2225 °C without evidence of 1986); the other is Kudriavy volcano on Iturup failure (Wooten and Lanshaw, 1989). Island in the Kurile volcanic arc, western Pacific ocean (Korzhinsky et al., 1994) (Figure 14.1). At the Kudriavy volcano, rheniite is being actively Distribution and abundance deposited as a condensate from volcanic gases at a rate of up to 20 tonnes rhenium per year (Naumov, Rhenium is one of the most dispersed of the nat- 2007). However, any commercial recovery of this urally occurring elements in the Earth’s crust, rhenium in the near future is unlikely due to with estimates of average crustal abundance the remoteness of the area and the extreme 342 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

copper ores of the Dzhezkazgan deposit, Kazakhstan (Genkin et al., 1994) and tarkianite, identified in the Hitura nickel–copper–PGE deposit, Finland and other nickel–copper–PGE deposits (Kojonen et al., 2004). Although rhenium has an affinity for sulfide phases, its concentration in most sulfide minerals is relatively low. However, the ionic radius of Re4+ is very close to that of Mo4+, which allows for a limited substitution of rhenium for molybdenum in molybdenite and other molybdenum minerals such as castaingite. Furthermore, rhenium is sim- ilar geochemically to molybdenum, which it com- monly accompanies through magmatic and related hydrothermal processes, and is commonly concentrated in molybdenite associated with var- ious types of granite-related deposits, particularly porphyry deposits. The rhenium content of most molybdenites is generally within a range of a few ppm to several thousand ppm Re (Table 14.3), although contents as high as 4.7 wt% Re occur in molybdenite at the Pagoni Rachi prospect (Voudouris et al., 2009) and up to 11.5 wt% Re has been recorded in molybdenite deposited in high- temperature fumaroles (Bernard et al., 1990). However, molybdenites, from which rhenium is currently recovered, commercially contain in the order of 200 to 1000 ppm Re. Other minerals in which rhenium may be concentrated in significant levels (i.e. >1 ppm Re) include urani- nite, which can contain up to 2700 ppm Re (Min et al., 2005), and gadolinite (Noddack and

Figure 14.1 Crystals of rheniite, ReS2, about 1 mm in size, Noddack, 1931). from a fumarole on Kudriavy volcano, Kurile Islands, western Pacific. (Photo courtesy of R. Lavinsky, iRocks.com.) Deposit types conditions under which the rheniite is being Porphyry deposits deposited (up to 900 °C). Rheniite has also been identified in other types Porphyry deposits are large, low-grade deposits in of occurrences, including the Pagoni Rachi por- which ore minerals occur in extensive zones of phyry-type molybdenum–copper–tellurium–silver– fracturing and brecciation associated with por- gold prospect in northern Greece (Voudouris phyritic granitic intrusions. They are the world’s et al., 2009) and the Phoenix nickel–copper–plati- most important source of copper and molybdenum, num-group element (PGE) deposit in Botswana and are major sources of gold and silver. They also (Maier et al., 2008). Other rhenium sulfide min- account for roughly 85 to 90 per cent of primary erals include dzhezkazganite, recognised in the rhenium production, which is recovered as a Rhenium 343

Table 14.3 Rhenium content of molybdenite in various deposit types worldwide.

Re (ppm)

Country Deposit Average Range References

Porphyry copper deposits Canada Lornex-Valley 330 294–350 Sinclair et al. (in prep.) Canada Gibraltar 443 238–750 Sinclair et al. (2009) Canada Island Copper 1784 1704–1863 Sinclair et al. (2009) Chile Chuquicamata 220 194–250 Giles and Schilling (1972); Mathur et al. (2001) Chile Collahuasi 410 368–448 Mathur et al. (2001) Chile El Salvador 630 Giles and Schilling (1972) Chile El Teniente 390 182–1154 Mathur et al. (2001) Chile Escondida 1355 Mathur et al. (2001) Kazakhstan Kounrad 664 24–1930 Ivanov et al. (1969) Mongolia Erdenet 535 Gerel and Munkhtsengel (2005) Peru Cerro Verde 3280 3061–3497 Mathur et al. (2001) United States Bagdad 460 330–642 Barra et al. (2003) Porphyry copper-gold deposits Armenia Kadzharan 340 80–1610 Ivanov et al. (1972) Armenia Agarak 820 57–6310 Magakian et al. (1984) Canada Snowfield 3579 Pretium (2011) Iran Sar Cheshmeh 1000 900–1160 Shariat and Hassani (1998) Mongolia Oyu Tolgoi 1500 H. Stein (pers. comm., 2007) United States Bingham 360 120–2000 Giles and Schilling (1972) United States Pebble 1130 900–2260 Rebagliati et al. (2009) Uzbekistan Kal’makyr-Dalnee 1200 150–2100 Ivanov et al. (1969); Golovanov et al. (2005) Porphyry copper-molybdenum deposits Canada Schaft Creek 590 Bender et al. (2007) Peru Toquepala 790 387–1496 Giles and Schilling (1972); Mathur et al. (2001) United States Sierrita-Esperanza 600 90–1800 Giles and Schilling (1972) Porphyry molybdenum deposits Canada Endako 35 15–67 Sinclair et al. (2009) United States Climax 35 11–80 Giles and Schilling (1972) United States Henderson 14 7–18 Giles and Schilling (1972) United States Thompson Creek 120 Sinclair et al. (in prep.) Porphyry tungsten-molybdenum deposits Canada Northern Dancer (Logtung) 22 Sinclair et al. (2009) Canada Sisson 9 6–12 Sinclair et al. (in prep.) Kazakhstan Verkhnee Qairaqty 57 47–66 Ivanov et al. (1972) Vein molybdenum deposits Australia Merlin 1062 300–2098 Horton (2010) Canada Playter 402 185–1047 Kilpatrick and Grieco (2010) Japan Daito 132 116–188 Ishihara (1988) Norway Knaben 14 1–28 Fleischer (1959); Giles and Schilling (1972) 344 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann by-product from the processing of molybdenite average rhenium-in-molybdenite contents from concentrates. porphyry copper–molybdenum deposits range from The average rhenium content of molybdenite about 90 ppm Re to 790 ppm Re. For example, from porphyry deposits varies widely, from less molybdenite from the Sierrita-Esperanza deposit in than 10 ppm to nearly 4000 ppm Re (Table 14.3), Arizona averages about 600 ppm Re and in the and is inversely related to the molybdenum content Collahuasi deposit in Chile it contains 410 ppm Re. (Figure 14.2). Porphyry copper and copper–gold Porphyry molybdenum deposits have the deposits have the lowest molybdenum grades, in highest molybdenum grades¸ ranging from 0.07% the order of 0.02% Mo or less, but molybdenites Mo (Endako, British Columbia) to as high as from these deposits have the highest average rhe- 0.24% Mo (Climax, Colorado)¸ but have the nium contents, ranging as high as 3858 ppm Re at lowest rhenium contents in molybdenites, typi- the Kemess South porphyry copper–gold deposit in cally in the range of 10 to 100 ppm Re. For British Columbia and 3280 ppm at the Cerro Verde example, molybdenite from both Endako and porphyry copper deposit in Peru. Molybdenite from Climax contain about 35 ppm Re. Molybdenites the now-closed Island Copper porphyry copper from porphyry molybdenum deposits with rhe- deposit on Vancouver Island, British Columbia, nium contents in excess of 100 ppm Re, such as averaged nearly 1800 ppm Re. In comparison, at Mount Tolman in Washington (182 ppm Re)

1

Climax

Chuquicamata Porphyry Mo and W-Mo deposits Bingham 0.1 Toquepala

Endako Pebble Island copper Mo (per cent) Cerro Verde Porphyry Cu-Mo 0.01 deposits Kal’makyr-Dalnee

Porphyry Cu deposits Porphyry Cu-Au deposits Erdenet r = –0.40 0.001 1 10 100 1000 10000 100000 Re (ppm) in molybdenite

Porphyry Cu Porphyry Cu-Au Porphyry Cu-Mo Porphyry Mo Porphyry W-Mo

Figure 14.2 Rhenium content of molybdenite versus molybdenum grade in porphyry deposits. (Revised after Sinclair et al., 2009.) (Cu, copper; Mo, molybdenum; Au, gold; W, tungsten) Rhenium 345 and at Thompson Creek in Idaho (120 ppm Re) as disseminated grains in the metasedimentary are exceptional. Rhenium content of molybde- host rocks. Also, whereas most vein molybdenum nite from tungsten-dominant porphyry deposits deposits, such as Knaben and Playter, consist of is also relatively low. molybdenite disseminated in quartz veins, the Merlin deposit is distinguished by the absence of quartz. Also, in contrast to molybdenum deposits Vein deposits in quartz veins, which typically contain an average Vein deposits occur in various structural set- of 0.1 to 0.3% Mo, the grade of the identified tings such as faults, fault systems and breccia mineral resource at the Merlin deposit averages zones and, in some cases, include replacement 1.3% Mo (Horton, 2010). Such a high content of zones in associated host rocks. Like porphyry molybdenum, combined with the high rhenium deposits, they are commonly associated with content of the molybdenite, makes Merlin an granitic intrusions. Unlike porphyry deposits, exceptionally rich molybdenum–rhenium deposit. they are typically small, but are highly varied in This, along with the geology, particularly the size and metal contents. In the past, they have absence of quartz, makes it unique among vein- been an important source of molybdenum in type deposits. some countries, including Canada, Norway and Japan, but up to the present time there has been Sediment-hosted copper deposits no recorded production of rhenium from a vein deposit. However, this will change if the Merlin Sediment-hosted copper deposits consist of dis- deposit in Australia can be brought into seminated to veinlet copper sulfides that occur in production. zones more or less concordant with the stratifica- In many vein deposits, molybdenite has rhe- tion of their sedimentary host rocks such as the nium content comparable to molybdenite from Kupferschiefer deposits at Lubin in Poland and porphyry molybdenum deposits, typically less the redbed-type deposits at Dzhezkazgan in than 100 ppm Re. For example, the rhenium Kazakhstan. They account for nearly one-quarter content of molybdenite from vein molybdenum of the world’s production and reserves of copper deposits in the Preissac-Lacorne area in Canada and are important sources of silver and cobalt. averages 25 to 30 ppm Re and molybdenite from They also account for 10 to 15 per cent of world the Knaben deposit in Norway contains an rhenium production. average of 14 ppm Re. However, rhenium-in- The nature of the occurrence of rhenium in molybdenite contents in vein deposits are highly sediment-hosted copper deposits is not well varied, ranging as high as 188 ppm Re in Japanese understood. Rhenium is recovered from the vein molybdenum deposits. At the Playter deposit processing of copper concentrates and some rhe- in Ontario, rhenium content of molybdenite nium may be present in the main copper min- averages about 400 ppm Re, although it ranges as erals, such as chalcopyrite, bornite and chalcocite. high as 1047 ppm. Molybdenite from the Merlin However, rhenium is also present in other min- deposit averages 1062 ppm Re. erals. For example, at Dzhezkazgan rhenium is The recently discovered Merlin deposit in present in the sulfide mineral dzhezkazganite, Australia is an unusual vein-type deposit. The which is closely associated with the copper ores molybdenum mineralisation is broadly strat- (Genkin et al., 1994) and, in the Lubin deposit, abound within metamorphosed Proterozoic sedi- rhenium-bearing minerals include castaingite mentary rocks, but the molybdenite mineralisation and molybdenite (Kucha, 1990). post-dates the metamorphism. The host rocks are The rhenium content of the ore at Dzhezkazgan brecciated and molybdenite occurs as fine-grained, averages 1.4 ppm Re (Hitzman et al., 2005). In the semi-massive to massive concentrations filling Lubin deposit it averages approximately 1 ppm Re open spaces and replacing host-rock fragments, and (H. Kucha, personal communication, 2007). 346 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

Uranium deposits ue dusts resulting from the smelter operations associated with the Noril’sk deposits in Russia, For a short time, beginning in about 1969, although apparently none of this rhenium is significant amounts of rhenium were recovered recovered at the present time. as a by-product from the mining of uranium Rhenium has many properties in common deposits in sedimentary rocks in the Falls City with PGE, in particular a strong affinity for area south of San Antonio in Texas (Millensifer, metallic or sulfide phases, and is present at ele- 1997). At the Palangana deposit, also in southern vated levels in many nickel–copper–PGE Texas, recovery of rhenium was investigated deposits. Data on grades of rhenium in nickel- (Goddard, 1984), although no production was copper-PGE are generally unavailable. However, reported. Currently, some rhenium is recovered rhenium contents of typical ore samples anal- from in situ solution mining of sediment-hosted ysed for Re–Os dating ranged from 0.01 to uranium deposits in the Central Kyzylkum region 0.52 ppm Re (average 0.12 ppm Re) for the of Uzbekistan (Chekmarev et al., 2004), although Noril’sk-Talnakh deposits in Russia (data in this probably accounts for no more than about Walker et al., 1994a) and from 0.07 to 0.31 ppm one per cent of current world production. Re (average 0.19 ppm Re) for Sudbury, Ontario The nature of the occurrence of rhenium deposits (data in Walker et al., 1994b). In some in uranium deposits is not well understood. At deposits, the rhenium occurs as tarkianite and Palangana and other sediment-hosted uranium unnamed rhenium-bearing sulfide minerals deposits, rhenium is closely associated with (Dare et al., 2010); in others it is concentrated molybdenum and is likely concentrated in jord- mainly in base-metal sulfides such as pyrrhotite, isite, an amorphous molybdenum mineral with pentlandite and pyrite (Dare et al., 2011). the same composition as molybdenite. However, at the Sun Valley uranium mine in Arizona, rhe- nium is also associated in part with uranium World resources and production (Petersen et al., 1959) and some rhenium may also be present in uranium minerals such as The location of significant rhenium-bearing uraninite. Rhenium can also be sequestered deposits in the world is shown in Figure 14.3. in carbonaceous material in sediment-hosted World mine production of rhenium from the deposits. main known sources in 2012 has been estimated at The rhenium content of uranium deposits varies about 44.8 tonnes (MMTA, 2012a), most of which from <1 to 5 ppm Re. In the uranium deposits in the was produced as a by-product of molybdenum pro- Central Kyzylkum, the rhenium content is gener- duction associated with porphyry copper and ally in the range of 0.5 to 2 ppm Re (Venatovskij, copper–gold deposits. The largest portion (28 tonnes 1993). The Anna Lake uranium deposit in Labrador Re) was produced by Molymet in Chile, primarily contains 0.18 ppm Re at a cut-off grade of 0.015% from large porphyry deposits such as El Teniente, U O (Fraser and Giroux, 2009). The rhenium grade 3 8 Escondida, Los Pelambres and Collahuasi and, to a at the Palangana deposit is approximately 5 ppm Re, lesser extent, Chuquicamata (Figure 14.5). However, while selected samples from the Sun Valley mine Molymet also sourced molybdenite concentrates contained as much as 1000 ppm Re. from deposits in other countries, including Mexico, Peru and the United States. The United States was Magmatic nickel–copper–platinum-group the second largest producer (7 tonnes Re), with pro- element (PGE) deposits duction primarily from Sierrita-Esperanza and Magmatic nickel–copper–PGE deposits represent Bagdad, although the facility at Sierrita also treated a potentially significant source of rhenium. other molybdenite concentrates on a toll basis. According to Besser et al. (1997), as much as Other porphyry deposits from which rhenium 400 kg of rhenium is concentrated annually in the was recovered include Kadzharan and Agarak in Figure 14.3 The global distribution of significant rhenium-bearing deposits. 348 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

0.7t 0.5t Future supplies 2.5t Porphyry copper and copper–gold deposits are 3t likely to continue as the main source of primary rhenium production for the foreseeable future, despite the fact that overall rhenium grades are low 3t compared to other types of deposits. For example, rhenium grades for porphyry deposits fall mainly in the range of 0.1 to 0.5 ppm Re, compared to 3.5t about 1 ppm Re for sediment-hosted copper deposits such as Lubin and Dzhezkazgan, and 23 ppm Re for the Merlin vein-type deposit (Figure 14.5). However, because of the large size of porphyry deposits, in the range of hundreds of millions to billions of tonnes of ore, the resources 24.6t 7t of rhenium in molybdenite associated with por- phyry deposits are substantial. The Pebble deposit in Alaska, with an estimated resource of more than 4000 tonnes contained rhenium, is probably the Chile USA Poland Kazakhstan largest currently unexploited rhenium resource in the world. When it is developed, the Pebble deposit China Korea Armenia Uzbekistan is projected to produce 1200 tonnes rhenium in the first 45 years of operation (Northern Dynasty Figure 14.4 The main sources of primary rhenium Minerals Ltd, 2011), equivalent to 26.7 tonnes rhe- production in 2012 (tonnes). (Data from MMTA, nium annually, which is more than 50 per cent of 2012a, courtesy of Lipmann Walton & Co Ltd.) current world annual production. Additionally, there are numerous porphyry deposits currently Armenia, Erdenet in Mongolia, Kal’makyr in producing molybdenite with significant rhenium Uzbekistan and Kounrad in Kazakhstan. In China, that is unaccounted for, including Valley-Lornex, Jiangxi Copper lists rhenium as one of the products Gibraltar and Huckleberry in British Columbia, from the Dexing porphyry copper–molybdenum Cerro Verde in Peru and Sar Cheshmeh in Iran. deposit. Although a significant amount of rhenium Although sediment-hosted copper deposits contain is present in porphyry molybdenum deposits such rhenium resources as large as those in porphyry as Climax in Colorado (127 tonnes), the rhenium deposits (3000 to 4000 tonnes each in the Lubin content of the molybdenite at Climax and other and Dzhezkazgan deposits), production from these porphyry molybdenum deposits is generally too deposits is unlikely to increase significantly from low to be recovered economically. their combined current production of about five to Rhenium production from sediment-hosted six tonnes rhenium per year. copper deposits in 2011 is estimated to have been In the near future, rhenium production from the six to seven tonnes, about 3.5 tonnes from the Merlin vein-type deposit, if successfully commer- Lubin deposit in Poland and a lesser amount from cialised, could contribute substantially to world rhe- the Dzhezkazgan deposit in Kazakhstan. In nium production. Merlin has an identified resource addition, an unspecified amount, probably about equivalent to 150 tonnes rhenium and over a pro- 0.5 tonnes or less, was recovered at the Navoi jected 15-year mine life, rhenium production from metallurgical plant in Uzbekistan from the leach the Merlin deposit is expected to be on the order of solutions associated with the in situ mining of 7.5 tonnes rhenium per year, allowing for about 75 sandstone-hosted uranium deposits. per cent recovery (SRK Consulting, 2010). However, Rhenium 349

100 1000 t 10 000 t 100 000 t

Merlin

10 Palangana Cerro Verde Island Playter Copper Dzhezkazgan 1 Lubin Pebble Climax Toquepala Anna Lake Bingham

El Teniente 0.1 Chuquicamata Re (ppm)

Kal’makyr- Knaben Endako Dalnee Sudbury Erdenet 0.01

Norilsk- Talnakh 100 t

0.001

0.001 t 0.01 t 10 t 0.1 t 1 t

0.0001 0.01 0.1 1 10 100 1000 10000 100000 Size of deposit (millions of tonnes)

Porphyry Cu Porphyry Cu-Au Porphyry Cu-Mo Porphyry Mo Porphyry W-Mo Vein Mo Sediment-hosted Cu Uranium Ni-Cu-PGE

Figure 14.5 Rhenium grade versus tonnage of significant rhenium-bearing deposits; diagonal lines represent tonnes (t) of contained rhenium. (Revised after Sinclair et al., 2009.) (Cu, copper; Mo, molybdenum; Au, gold; W, tungsten; Ni, nickel; PGE, platinum-group elements)

Merlin is an unusual molybdenum and rhenium- jordisite, the amorphous variety of molybdenite, rich vein-type deposit. Rhenium resources in other which is generally soluble in the leach solutions vein deposits, such as the Playter deposit in Ontario used for in situ mining. This rhenium is readily (about 1.5 tonnes rhenium), are relatively small and recoverable using existing technology and repre- much lower in grade, and are unlikely to contribute sents a potentially important source of rhenium significantly to world rhenium production except, in the future, even though the rhenium content of possibly, on a relatively short-term basis. individual deposits is relatively small (Figure 14.5). Although rhenium production from uranium Rhenium resources in other types of deposits, mining is limited at the present time, rhenium such as nickel–copper–PGE deposits, is is present in many sediment-hosted uranium significant, although no rhenium production deposits, although probably unrecognised (not from them has been reported. For example, the analysed) in many cases. It typically occurs in Sudbury nickel–copper–PGE deposits originally 350 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann contained in the order of 300 tonnes rhenium, where the company has installed plant facilities those in the Noril’sk-Talnakh district are esti- to recover rhenium from the smelting of copper mated to have contained about 200 tonnes rhe- ores. Offgases are scrubbed as is done where nium (Sinclair et al., in preparation). However, molybdenite is roasted. The sulfuric acid liquids the technology to recover the rhenium in these are then sent to filtration and solid-bed ion- deposits¸ from the ue dust associated with the exchange columns where the rhenium is recov- smelting and refining of the PGE-rich ores, exists ered as ammonium perrhenate. KGHM has also and presumably could be implemented if war- recently installed equipment to produce rhenium ranted by economic conditions. Resources in car- metal powder and pellets. bonaceous materials, such as coal and oil shale, There are rhenium recovery plants where a may be substantial (Chekmarev and Troshkina, combination of liquid ion exchange (LIX), and 1997; Lippmaa et al., 2011), and possibly repre- solid-bed ion exchange is used with LIX preceding sent resources for the future, but recovery of rhe- the resin bed. This allows stripping of the rhenium nium (and other metals) from these materials is from the LIX as an ammonium solution. This unlikely to be economic as long as other sources increases the concentration of the rhenium in of rhenium are available. solution and eliminates the need to neutralise the acidic solution coming from the wet scrubbers. Rhenium metal powder is produced by tradi- Extraction methods, processing tional powder metallurgy techniques (Figure 14.7). and beneficiation Ammonium perrhenate is reduced using hydrogen in common boats-in-tubes type furnaces. Boats When rhenium-bearing molybdenite is roasted, are filled with ammonium perrhenate and pushed converting the molybdenum sulfide to moly- through tubes counter-currently to a ow of bdenum trioxide (MoO3) and to sulfur dioxide (SO2), hydrogen gas. The tubes are heated externally to rhenium is oxidised to volatile rhenium heptoxide the appropriate temperature. Depending on the

(Re2O7) which exits the roaster with the sulfur particle size of the rhenium powder required, the dioxide. Scrubbing the exit gases with water dis- reduction may be a one- or two-stage process and solves the rhenium heptoxide as crude perrhenic the ammonium perrhenate may be specially acid, HReO4. This solution, which also contains ground before reduction. Rhenium for alloy pro- some sulfuric acid and other impurities, is duction is produced by pressing the powder into treated to prepare for rhenium recovery by either pellets of varying sizes, some as large as 15–20 mm solvent extraction (liquid ion exchange) or solid-bed in diameter and 8–10 mm thick, and others ion exchange (Figure 14.6). From either method, 5–8 mm in diameter and 3–5 mm in thickness. rhenium is stripped and crystallised as ammo- The pellets then are sintered to improve physical nium perrhenate, NH4ReO4. Generally, repeated integrity as well as to reduce gases further, espe- recrystallisation is required to achieve the required cially oxygen. Metal products such as wire and ammonium perrhenate purity of 99.95 per cent, plate are made by pressing rhenium powder into metal basis. bars or rods followed by resistance heating to New methods of production include the new sinter. The sintered rods or bars then are drawn Kennecott MAP (Molybdenite Autoclave Process) into wire or rolled into sheets or plates. Rhenium facility which involves high-pressure oxidation of also can be deposited on various types of parts and molybdenite to achieve better recovery and more moulds using chemical vapour deposition. The pure products of molybdic oxide and ammonium production of catalysts may require the use of per- perrhenate. This has been underway since 2010 rhenic acid, HReO4. This is a solution of rhenium according to numerous reports and press releases. heptoxide (Re2O7) in water with a rhenium Since 2007, rhenium has also been produced in concentration of 35–50 weight per cent rhenium Poland by KGHM Ecoren from copper operations contained. Rhenium 351

Rhenium Molybdenium -bearing Roasting trioxide molybdenite calcine Sulfur dioxide & rhenium heptoxide

Sulfur dioxide recovery and Wet scrubbing flue gas cleaning processes

Perrhenic acid

Liquid ion exchange process Solid bed ion exchange process

Sludge to Settling Settling Sludge to recycling recycling

Raffinate Liquid ion exchange 1 Oxidation and Sodium hydroxide to waste neutralisation Sodium hypochlorite Molybdenum-rhenium -bearing solution Solid bed ion Eluant for exchange stripping Molybdenum Conditioning and recovery liquid ion exchange 2 Effluent solution Rhenium-bearing Rhenium-bearing discard solution solution

Evaporation and Evaporation, crystallisation, crystallisation resin stripping

Ammonium perrhenate crystals (NH4ReO4)

Figure 14.6 Typical rhenium recovery process by liquid ion exchange and solid-bed ion exchange.

Molybdenum concentrates Copper concentrates

Spent catalyst Ammonium perrhenate recycle

Perrhenic acid Rhenium powder

Catalysts Pellets Powder metallurgy

Mill products, Figure 14.7 Summary of the production Super alloys wire, alloys of rhenium and products. 352 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

Specifications and uses production of alumina base platinum–rhenium bimetallic reforming catalysts used to produce Minor metals are often traded as an intermediate high-octane gasoline, and added to other catalysts product, which is then further treated prior to for petrochemical catalysis. incorporation into a final component or product. The Minor Metals Trade Association (MMTA) In the case of rhenium the main form in which publishes specifications for numerous rare and it is traded is ammonium perrhenate (APR), minor metals to facilitate trading in those metals sometimes called ammonium tetraoxorhenate and their chemicals (MMTA, 2012b). Basic-grade

(NH4ReO4). In this form, a salt of free-owing APR is specified by the MMTA to contain a white crystals, rhenium is stable, non-hazardous minimum of 69.00% Re and less than 0.05% and suitable for shipment by air. by weight of Si, Fe, K, Mo, Al, Ca, Cu, Mg, Mn, Rhenium in APR is ready to be processed Ni, Na, P and S, together with a maximum water further for use in one of its two main industries – content of 0.10%. superalloy for casting into single-crystal turbine An example of ‘catalyst’-grade purified blades for aero-engines or solutions for the manu- ammonium perrhenate suitable for the manufac- facture of reforming catalysts (Figure 14.8). For ture of bimetallic reforming catalysts provided by superalloys, rhenium powder is produced by a principal consumer is shown in Table 14.4a.The hydrogen reduction of APR which is then sin- limit of certain impurities is set to avoid those tered into small pellets of rhenium metal (Re elements that could be poisonous to the function 99.9%). For the petroleum and petrochemical of the ultimate catalyst. Because rhenium is industry, APR is purified into Catalyst Grade for paired with the much more expensive platinum in bimetallic catalysts (more than 13 times higher than the price of rhenium in October 2012) the 4t aim is to avoid contamination and costly failure. When the ultimate customer for rhenium metal is the aircraft industry, typically, rhenium 5t at 3% to 6% is added to complex nickel-base alloys. These are used for precision casting into single-crystal turbine blades for the High-Pressure Turbine (HPT). In this use, the control of contam- inants is also critical and a typical specification is given in Table 14.4b. An example of the typical composition of a rhenium-bearing superalloy is Cannon Muskegon’s CMSX-4 (Table 14.5). As a consequence of the high price of rhenium and its limited production and availability, alter- natives to it are continually being sought. Possibilities include coating blades and end tips 45t with rhenium (to avoid the need for rhenium in the alloy itself), fibre-reinforced superalloy com- ponents using rhenium–tungsten fibres, ceramic Super alloy Catalysts Others matrix composites (CMCs) and metal matrix composites (MMCs). It has been suggested that a Figure 14.8 The main sectors of rhenium consumption price ceiling for rhenium, at which users of rhe- in 2012 (tonnes). (Data from MMTA, 2012a, courtesy nium might cease to use rhenium, could be in the of Lipmann Walton & Co Ltd.) range US$5000 to 6000/kg. However, even when Rhenium 353

rhenium did reach US$12,000/kg in August 2008, Table 14.4 (b) Specification for rhenium metal usage was not reduced. It should be noted, how- pellets, superalloy grade. (Data courtesy of MMTA, ever, that GE Aviation has been working for 10 2012b.) years to lessen its dependence on rare metals including rhenium. This is being accomplished Maximum permitted through a combination of innovative component content (ppm, unless Element otherwise stated) designs, advanced manufacturing processes and the development of new alloys, as well as recy- Re 99.9% minimum cling (General Electric, 2012.) Se 5 Rhenium has proven successful in aircraft Te 5 gas turbines, accordingly manufacturers of land- Tl 5 based gas turbines used for primary, as well as H50 emergency, power generation, are looking closely Mg 100 Ca 5 at the use of rhenium in turbine blades for those Na 10 Ag 5 As 10 Table 14.4 (a) Specification for catalyst grade ammonium Bi 1 perrhenate. (Data courtesy of MMTA, 2012b.) Cd 50 Maximum permitted Ga 50 content (ppm, unless In 50 Element otherwise stated) Pb 5 Sb 50 Re 68.5–69.8% minimum Th 20 Sb 30 Sn 10 As 30 Zn 50 Ca 50 Cl 50

Cl2 100 K10 Cr 50 C30 Co 50 Si 20 Cu 50 S20 Ir 75 O 300 Fe 50 N50 Pb 30 Fe 100 Mg 50 Mo 500 Mn 50 W 500 Mo 100 Mn 20 Ni 50 P 200 Pd 50 Al 30 P 200 B20 K 200 Co 100 Rh 50 Cr 100 Si 50 Cu 70 Ag 50 Hf 200 Na 50 Nb 200 Sn 50 Ni 100 Zn 50 Ta 200 As, Ca, Fe, K, Mg, Na, Pb, Sn 200 total Ti 200 As, Co, Cr, Cu, Mn, Ni, Sb, Si, Sn 200 total V 200

H20 0.10% Zr 200 354 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

Table 14.5 Composition of Cannon Muskegon’s Based on published compositions of some CMSX-4 superalloy. (Data from C-M Group, 2012.) alloys and recent rhenium prices, one can see the price sensitivity of those superalloys to the price Element Content (wt per cent) of rhenium (Table 14.6). When the rhenium price Cr 6.5 exceeds about US$2200 per kg the rhenium Co 9.0 content of these alloys accounts for at least two Mo 0.6 thirds of the price of the alloy. W 6.0 Ta 6.5 Re 3.0 Al 5.6 Recycling and re-use Ti 1.0 Hf 0.1 Rhenium provides a fascinating study in regard to Ni 61.7 re-use and recycling. Here, perhaps unsurpris- ingly, the determining factor has been value rather than concern about the environment, although a applications. There have been reports that rhe- positive environmental outcome has resulted. nium is already in use in those turbine blades. Rhenium’s twin markets of catalysts and superal- However, a significant problem with these blades loys provide two different case histories. is their size and the quantity of rhenium that would be consumed. The present limited supply Catalysts of rhenium could limit this use. Other recent developments include new cata- In the case of catalysts it was, and still is, the lysts in petrochemical production. One has been value of platinum with which rhenium is allied the addition of rhenium to the silver catalyst in at the rate of Pt 0.3% and Re 0.3% which pro- ethylene oxide production or even replacing the vides the incentive for recycling. A recycling silver in some cases. Several companies have industry developed originally to recover the developed new catalysts which contain rhenium valuable platinum. Recovery of rhenium devel- for the production of intermediates and precur- oped later as rhenium began to be used in the sors for several types of elastomeric materials for reforming catalysts. Those that recover platinum a variety of consumer products. and rhenium from spent reforming catalysts are Rhenium has also been investigated for use in Heraeus Precious Metals GmbH & Co. KG, who catalysts for the gas to liquids, GTL, process. The recycle Pt–Re catalysts at their plants in Germany term gas to liquids is applied to various processes and the USA, and Gemini in the USA. It is which have been developed to convert natural gas thought that about 15 tonnes of rhenium is recov- to liquid fuels (National Petroleum Council, 2007). ered per year by this method, which remains in a These are all based on the Fischer–Tropsch process continuous closed loop within the catalyst developed by Fischer and Tropsch in Germany dur- industry sector (and therefore does not appear in ing the 1920s to convert low-grade coal to liquid supply/demand figures). In Europe, the recovery fuels. These processes all make use of catalysts of rhenium did not begin as soon as it did in the consisting of iron, cobalt or nickel, although it was United States. In Europe it was regarded as a determined recently that rhenium added to the cat- nuisance during platinum recovery while in the alyst improved the efficiency. Another potential US rhenium was recovered along with the application of rhenium is that of rhenium as rhe- platinum by the spent catalyst processor and nium boride, ReB2, as an abrasive tool for cutting or stockpiled for later sale as long as the owner of polishing. Although this use was discovered many the spent catalyst did not want the rhenium years ago it has apparently not developed due to the returned or held for them. There was a time when price and limited availability of rhenium. the catalyst user could purchase virgin rhenium Rhenium 355

Table 14.6 The inuence of the price of rhenium on the cost of superalloys as a percentage of the total cost of input raw materials. (The source of the prices used in these calculations is Metal Prices.com on 28 September 2011. The rhenium contents of the superalloys are from various published sources.)

Rhenium price (US$/kg)

$1215 $2205 $5952

Rhenium Alloy Alloy % alloy Alloy % alloy Alloy % alloy content cost cost cost due cost cost due cost cost due Alloy (%) (US$/kg) (US$/kg) to Re content (US$/kg) to Re content (US$/kg) to Re content

CMSX-3 0 5.15 CMSX-4 3 66.95 55 95.99 69 208.42 86 CMSX-10 6 103.24 72 162.63 81 387.51 92 Rene N5 3 66.95 54 96.65 68 209.09 85

at a lower price than for the cost of recovery from precipitated as APR. (The processes and those the spent catalyst. That has not been the case for using them are confidential and require Non- quite some time. Disclosure Agreements for detailed information.) Feed required by recyclers for the various processes includes used blades, casting scrap Superalloys (runners and risers) as well as grindings generated In the case of complex superalloys, with typical either by cutting virgin alloy ingots or surface- compositions as mentioned in Table 14.5, the finishing blades. At the time of writing, it is growth of a recycling industry has been slow and thought that as much as (but not more than) five spasmodic. The main reason for this was that tonnes of rhenium is recovered by this method during the first two decades of rhenium’s com- per year, while another five tonnes of rhenium mercialisation in single-crystal turbine blades returns into new alloy via internal and external from the 1980s, the rhenium price was relatively ‘revert’, i.e. clean alloy that may be re-melted low – as low as US$ 300/kg 1994 and rarely higher without the need for recycling. than US$1500/kg between 1980 and 2000. In other words, prices were mostly below the cost of recovery. During this period rhenium was regu- Substitution larly wasted when the alloy blades, containing 60 per cent nickel, were sold for their nickel value Presently, rhenium has no substitute in its main only and disposed of into the stainless steel use as a three per cent constituent in complex industry as a nickel addition. nickel-based alloys for single-crystal turbine The incentive for greater efficiency in recycling blades. Here, the main purpose of rhenium is to rhenium-bearing alloys came when rhenium increase the presence of the gamma prime (γ1) reached US$12,000/kg in August 2008. At a phase which increases creep resistance. With tur- stroke, recovery became economically viable and bine inlet temperatures estimated at 1600 °C or two approaches were employed by recyclers: a) above in the current generation of large gas-tur- chemical decomposition of the complex alloy bine engines (e.g. Rolls-Royce Trent 1000, lead followed by an ion-exchange process to recover engine for the Boeing 787 Dreamliner), no other rhenium in the form of APR, and b) a unique pyro- technology has yet been commercialised that has phoric process where rhenium is evaporated as the similar defect tolerance and at the same time heptoxide and subsequently captured and then meets the requirements of temperature ranges 356 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann from sub-zero to 1600 °C with minimal deforma- serve this purpose. With forecasts of 27,000 tion. Substitute materials that have been consid- aircraft deliveries between 2009–2018, and a ered include ceramic matrix composites and further 38,000 between 2019–2028 (Rolls Royce, metal matrix composites, which are being studied 2010), airlines are competing with each other to at leading institutes such as Imperial College commission the most efficient gas turbines. In London and at leading engine makers such as order to save cost from expensive fuel and also General Electric, Rolls-Royce and Pratt & comply with ever more stringent laws this trend Whitney. It is thought that these composites may looks likely to continue. As an example, the EU’s one day be applied for static parts but are unlikely latest law directed towards the airline industry to be approved for rotating sections. seeks to set carbon limits per airline that, if In the catalyst industry, mono-metallic exceeded, will be penalised via the enforced pur- catalysts, consisting solely of platinum as the chase of carbon credits. The development of more precious metal, are regularly used but, in most efficient engines, which is reected in rhenium applications, rhenium is preferred alongside demand, will continue to drive this market. With platinum because it increases efficiency. The rhenium’s value now recognised, due to the price of rhenium compared to platinum is so attention it received when prices peaked in much cheaper that removal of rhenium has little August 2008, it now means that a strong incen- cost benefit. tive exists to recover rhenium from end-of-life blades and casting scrap. The recovery of rhenium units from the aerospace industry will never be as Environmental issues efficient as the catalyst industry due to the wide dispersal of rhenium units in end-of-life blades, Rhenium is radioactive in all forms because of the the cost of recovery and the length of time (some- content of the isotope 187Re. However, 187Re decays times up to 12 months) from the recycling of used by emission of very low energy beta particles blade to production of new rhenium pellet. (electrons) that pose minimal risk to human health. There are few if any concerns related to radioactivity during the mining of copper– World trade molybdenum ore or in the processing of the con- centrates as the concentration of rhenium in these The international trade in rhenium is partly processes is extremely low. Only after the rhe- determined by the geological/geographical loca- nium is recovered could there be any potentially tions in which it arises and by the development harmful exposure. However, even at this stage, of the industries that have commercialised its huge quantities of rhenium, of the order of several use. During the cold war in the 1950s, when rhe- tonnes per annum, would need to be ingested for nium was just beginning to be commercialised in an occupational worker to exceed recommended catalysts, rhenium produced in Kazakhstan and annual limits. Armenia was used to supply the USSR and no The quest for fuel efficiency affects many criti- exports were made to the West. During the same cal metals, none more than rhenium. In so far as period, the first commercial production was at rhenium contributes to high operating tempera- Shattuck Chemical in Denver, Colorado and tures, increased fuel efficiency of gas turbines, Kennecott Copper, in Garfield, Utah and enough reduction of nitrous oxide emissions to the upper was produced to supply the oil-refining industry atmosphere, and blade longevity, it may be said in the western world. that rhenium is a green metal. With law-makers The key change to the structure of the trade ever more concerned to reduce emissions, aero- occurred in the late 1970s when, due to a large engineering and material science tends towards build-up of stock at that time, (as much as 12 the use of elements and techniques which will tonnes in 1974), producers were discouraged out Rhenium 357 of the business and ceased production. The Finally, with the maturing of this market, it is well-known trading company, Philipp Brothers, now thought that Molymet will in future con- then bought the technology for the recovery of struct sales contracts based on objective pub- rhenium from ue dusts from Shattuck Chemical lished prices rather than the previous fixed-price and transported the system to Santiago in Chile. system. One of the motives for this change is Here, at the heart of the copper industry, where pressure from mining companies to receive a by-product rhenium-bearing molybdenite arose, credit for the rhenium content within their it was logical for rhenium to be recovered. This molybdenite. A number of trading parties and was the origin of Carburo y Metalurgia (now stockholders, who are knowledgeable about rhe- known as Molybdenos y Metales, Molymet), nium, are to be found amongst members of the today’s leading producer of rhenium. Minor Metals Trade Association (MMTA). Today, as rhenium’s rarity value, use and unsubstitutability has become more recognised, Molymet’s dominance has emerged by virtue of Prices this strategic position, controlling the main world supply of rhenium via the recovery of rhe- At present, daily and weekly prices for rhenium nium units from the ues at their large molybde- (APR and metal pellets) are available by subscription nite roasters. Production of rhenium here in 2012 to several publications, such as Metals Week, was 24.68 tonnes (more than 55 per cent of world Metal-Pages and Metal Bulletin. primary supply, Figure 14.4). However, in keeping From the 1970s, at the inception of the wider with the history of initial over-production in the trade in rhenium, prices ranged from US$3000/kg 1970s, Molymet, in the 1980s, pursued a policy of Re contained, descending to a low of about long-term fixed price contracts in order to US$300/kg in the mid-1990s, then peaking at encourage the commercialisation of rhenium in US$12,000/kg in August 2008 (Figure 14.9). the aerospace industry. During this period The prices in the 1970s reected first-time buy- Molymet supplied more than 70 per cent of the ing of rhenium as raw material for bimetallic cata- world’s rhenium and the free trade in rhenium lysts, prior to the recycling that made the industry was extremely limited. efficient. The next peak came in the era following All this changed in the early 1990s with a the OPEC oil crisis of the mid-1970s. This resulted coincidence of two factors, a) the break-up of the in wider use of small cars and worldwide laws to Soviet Union, and b) the rapid development of remove lead from petrol. The proliferation in rhenium’s use in modern jet engines. For the first demand for high-octane petroleum increased the time, Soviet stockpiles of rhenium owed to the demand for bi-metallic reforming catalysts. The west from Kazakhstan, which gave life-blood to trough in prices in the early 1990s, when prices the free market, allowing companies in Europe sank to near US$300/kg for almost five years, such as HC Starck GmbH and WC Heraeus resulted from the disposal of redundant Soviet/ GmbH (now called Heraeus Precious Metals Kazakhstan stocks onto the Western market. The GmbH & Co. KG) to manufacture rhenium metal slow recovery from the late 1990s reected the pellets suitable for the superalloy industry using emerging use of rhenium in modern gas-turbine non-South American rhenium raw material. engines, while peak prices in 2008, as high as At the time of writing (August 2011) there is US$12,000/kg, were attributable to the fact that much greater plurality in the trade of rhenium, demand from the aerospace industry exceeded led by more availability of information and Molymet’s ability to supply. The subsequent reporting of prices within numerous publica- decline mirrored the Lehman Brothers financial tions. The need to price rhenium content within crisis. scrap at a discount to a published base price of With growing demand in aerospace as the main APR, has also given life to these quotations. driver for rhenium consumption, it is expected 358 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

Global recession and signs of slow 12 000 down in aerospace industry

Quest for fuel efficiency 10 000 in aero-engines makes rhenium un-substitutable

8000 End of USSR Dispute in Kazakhstan and Excess production based and de-stocking expected demand from on early over expectation GTL (Gas to Liquids) 6000 Introduction of Growth of use of rhenium in single lead-free petrol crystal blades for aerospace Price (US$/kg) and industrial gas turbines 4000 First aerospace applications 2000

0 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Year

Figure 14.9 Rhenium price trend, 1971–2010. (Data from USGS up to 1993 and from Lipmann Walton & Co Ltd after 1993.) that prices are unlikely to decline, as such an out- about how much rhenium might be recovered come would remove the incentive for recycling, from this operation as it would be the first pri- thus cutting off much needed units to the market mary molybdenum ore body with significant and sowing the seeds for further shortage. levels of rhenium. Prior to the discovery of Merlin, primary deposits such as the Climax and Henderson mines in Colorado have contained Outlook only small amounts of rhenium, well under 100 ppm, which is insufficient to justify recovery The outlook for the use of rhenium is mixed. In efforts. It has been only the co- or by-product the 1970s it was appreciated that there were copper mines that had molybdenite concentrates many potential uses for rhenium although their with sufficient rhenium content, above 150 ppm, development and uptake were dependent on a to justify rhenium recovery. The Pebble project reasonable price and adequate supply. Since then in Alaska is an enormous deposit with much rhenium has been incorporated with silver in the potential, but which faces very high levels of ethylene oxide catalyst and also used in more environmental opposition. It is likely that it will than one proprietary petrochemical catalyst. Its be many years before rhenium is produced from application in the Fischer–Tropsch gas-to-liquids this project. catalyst has also been investigated, but the scar- Meanwhile, the reduction in rhenium use by city of rhenium has curtailed that use. the aircraft engine producers has resulted in There seems to be more rhenium coming increased rhenium availability for some of the available now with offers from such as Uzbekistan other applications such as land-based gas turbines for 500 kg per month of APR, recovery from for use in power generation. The use in X-ray Mongolian ores being toll roasted, as well as the tube target production seems to be one area where promise of more from deposits yet to be mined rhenium demand is on the increase, but there is such as Merlin in Australia. There are doubts little doubt that rhenium’s use in nickel-based Rhenium 359 superalloys for the production of single-crystal General Electric (2012) Rhenium Reduction Program: turbine blades is set to remain rhenium’s main using less of a rare material. http://citizenship. market for many years to come. geblogs.com/rhenium-reduction-program-using-less- of-a-rare-mineral/# Genkin, A.D., Poplavko, E.M., Gorshkov, A.I., References Tsepin, A.I. and Sivtsov, A.V. (1994) New data on dzhezkazganite – rhenium-molybdenum-copper-lead Bernard, A. and Dumortier, P. (1986) Identification of sulfide – from the Dzhezkazgan deposit (Kazakhstan).

natural rhenium sulfide (ReS2) in volcanic fumaroles Geology of Ore Deposits 36, 481–489. from the Usu volcano, Hokkaido, Japan. Proceedings Goddard, J.B. (1984) Recovery of rhenium from uranium of the XIth International Congress on Electron in-situ leach liquor. Society of Mining Engineers of Microscopy (Kyoto), 1691–1692. AIME Transactions 274, 1996–2000. Bernard, A., Symonds, R.B. and Rose, W.I. (1990) Habashi, F. (1996) Rhenium Seventy Years Old, In: Volatile transport and deposition of Mo, W and Re in Bryskin, B.D (ed.) Rhenium and Rhenium Alloys, high temperature magmatic uids. Applied TMS, Warrendale, PA, USA. Geochemistry 5, 317–326. Hitzman, M., Kirkham, R., Broughton, D., Thorson, J. Besser, A.D., Peredereev, A.V. and Tarasov, A.V. (1997) and Selley, D. (2005) The sediment-hosted stratiform The perspectives of the development of the rhenium’s copper ore system. In: Economic Geology 100th production in Russia. In: Bryskin, D.D. (ed.) Rhenium Anniversary Volume, 609–642. and Rhenium Alloys. The Minerals, Metals and Horton, J. (2010) Mount Dore-Merlin Deposit, NWQld, Materials Society, Warrendale, PA, USA, 59–65. Australia. NI 43-101 Technical Report prepared for Chekmarev, A.M. and Troshkina, I.D. (1997) Rhenium Ivanhoe Australia Ltd., 162. recovery from non-traditional resource. In: Bryskin, Kojonen, K.K., Roberts, A.C., Isomäki, O.-P., Knauf, V.V., D.D. (ed.) Rhenium and Rhenium Alloys. The Johanson, B. and Pakkanen, L. (2004) Tarkianite, (Cu,Fe)

Minerals, Metals and Materials Society, Warrendale, (Re,Mo)4S8, a new mineral species from the Hitura mine, PA, USA, 99–107. Nivala, Finland. The Canadian Mineralogist 42, 539–544. Chekmarev, A.M., Troshkina, I.D., Nesterov, Y.V., Korzhinsky, M.A., Tkachenko, S.I., Shmulovich, K.I., Maiboroda, A.B., Ushanova, O.N. and Smirnov, N.S. Taran, Y.A. and Steinberg, G.S. (1994) Discovery of a (2004) Associated rhenium extraction in complex pure rhenium mineral at Kudriavy Volcano. Nature processing of productive solutions of underground 369, 51–52. uranium leaching. Chemistry for Sustainable Kucha, H. (1990) Geochemistry of the Kupferschiefer, Development 12, 113–117. Poland. Geologische Rundschau 79, 387–389. C-M Group (2012) Vacuum melt alloy index, Nickel Leonhardt, T. (2009) Processing and Properties of Base Single Crystal. http://www.c-mgroup.com/ Tungsten 25% Rhenium with and without Hafnium vacuum_melt_index/nickel_base_sx.htm Carbide, W-Re 2009 Plansee, Seminar Proceedings, Dare, S.A.S., Barnes, S.-J., Prichard, H.M. and Fisher, Reutt/Tirol. P.C. (2010) The timing and formation of platinum- Lipinski, M., Warning, B. and Brumsack, H.-J. (2003) group minerals from the Creighton Ni-Cu-platinum- Trace metal signatures of Jurassic/Cretaceous black group element sulfide deposit, Sudbury, Canada: shale from Norwegian Shelf and the Barents Sea. early crystallization of PGE-rich sulfarsenides. Palaeogeography, Palaeoclimatology, Palaeoecology Economic Geology 105, 1071–1096. 190, 459–475. Dare, S.A.S., Barnes, S.-J., Prichard, H.M. and Fisher, Lippmaa, E., Maremäe, E. and Pihlak, A.-T. (2011) P.C. (2011) Chalcophile and platinum-group element Resources, production and processing of Baltoscandian (PGE) concentrations in the sulfide minerals from the multimetal black shales. Oil Shale 28, 68–77. McCreedy East deposit, Sudbury, Canada, and the Maier, W. D., Barnes, S.-J., Chinyepi, G., Barton, J.M., origin of PGE in pyrite. Mineralium Deposita 46, Jr., Eglington, B. and Setshedi, I. (2008) The composi- 381–407. tion of magmatic Ni–Cu–(PGE) sulfide deposits in the Fraser, R.D. and Giroux, G.H. (2009) Form 43–101 Tati and Selebi-Phikwe belts of eastern Botswana. Technical Report on the Anna Lake Uranium Project, Mineralium Deposita 43, 37–60. Central Mineral Belt, Labrador, Canada; Prepared for Millensifer, T.A. (1997) Rhenium Background and Bayswater Uranium Corp., pp. 94. Markets. In: Bryskin, D.D. (ed.) Rhenium and 360 tom a. millensifer, dave sinclair, ian jonasson and anthony lipmann

Rhenium Alloys. The Minerals, Metals and Materials Sinclair, W.D., Jonasson, I.R., Kirkham, R.V., and Society, Warrendale, PA, USA, 37–47. Soregaroli, A.E. (in prep.) Rhenium in Canadian Min, M., Chen, J., Wang, J., Wei, G. and Fayek, M. (2005) mineral deposits. Geological Survey of Canada, Mineral paragenesis and textures associated with Open File. sandstone-hosted roll-front uranium deposits, NW SRK Consulting (2010) Independent Technical Report China. Ore Geology Reviews 26, 51–69. on the Merlin Project, Queensland. 43–101 Report Minor Metals Trade Association, MMTA (2012a) on Merlin. SRK Consulting (Australasia) Pty Ltd., Rhenium Supply and Demand. http://www.mmta. pp. 168. co.uk/rhenium-market-overview Treichel, P.M. (2000) Rhenium and rhenium com- Minor Metals Trade Association, MMTA (2012b) pounds. In: Encyclopaedia of Chemical Technology, Metal and assaying norms. http://www.mmta.co.uk/ posted online December 4, 2000. metals/Re/ Venatovskij, I.V. (1993) Study of the geotechnical condi- Morris, D.F.C. and Short, E.L. (1969) Rhenium. In: tions of uranium deposits of Uzbekistan during Wedepohl, K.H. (ed.) Handbook of Geochemistry, Vol. exploration work. In: Uranium in situ leaching. II/5. Springer-Verlag, Berlin and Heidelberg GmbH & Proceedings of a Technical Committee Meeting held Co. K. in Vienna, 5–8 October, 1992. International Atomic National Petroleum Council (2007) Working Document Energy Agency, TECDOC-720, 95–103. of the NPC Global Oil and Gas Study, Topic Paper #9, Voudouris, P.C., Melfos, V., Spry, P.G. et al. (2009) Gas to Liquids. Rhenium-rich molybdenite and rheniite in the Pagoni Naumov, A.V. (2007) Rhythms of rhenium. Russian Rachi Mo-Cu-Te-Ag-Au prospect, northern Greece: Journal of Non-Ferrous Metals 48, 418–423. implications for the Re geochemistry of porphyry- Noddack, I. and Noddack, W. (1931) Die Geochemie des style Cu-Mo and Mo mineralization. The Canadian Rheniums. Zeitschrift für Physikalische Chemie Mineralogist 47, 1013–1036. (Leipzig) 154, 207–244. Walker, R.J., Morgan, J.W., Horan, M.F., Czamanske, Northern Dynasty Minerals Ltd (2011) News Release, G.K., Krogstad, E.J., Fedorenko, V.A. and Kunilov, February 23, 2011. http://www.northerndynastyminer- V.E. (1994a) Re-Os isotopic evidence for an enriched- als.com/ndm/NewsReleases.asp?numCols=1andpage=2 mantle source for the Norils’k-type, ore-bearing Petersen, R.G., Hamilton, J.C. and Myers, A.T. (1959) intrusions, Siberia. Geochemica et Cosmochimica An occurrence of rhenium associated with uraninite Acta 58 (19), 4179–4197. in Coconino County, Arizona. Economic Geology 54, Walker, R.J., Morgan, J.W., Hanski, E. and Smolkin, V.F. 254–267. (1994b) The role of the Re-Os system in deciphering the Rolls-Royce (2010) Rolls-Royce Market Outlook, 2010. origin of magmatic sulfide ores: A tale of three ores. In: http://www.aarcorp.com/mro/files/Rolls_Royce_ Lightfoot, P.C. and Naldrett, A.J. (eds.) Proceedings of Forecast.pdf. the Sudbury-Norils’k Symposium. Ontario Geological Sinclair, W.D., Jonasson, I.R., Kirkham, R.V. and Survey, Special Volume 5, 343–355. Soregaroli, A.E. (2009) Rhenium and other platinum- Wooten, D.J.R. and Lanshaw, P.T. (1989) Proceedings group metals in porphyry deposits. Geological Survey Joint Army-Navy-NASA/Air Force Publication of Canada, Open File 6181, 1 sheet. (JANNAP) Propulsion Meeting, Laurel, Maryland. 15. Tantalum and niobium

ROBERT LINNEN1 , DAVID L. TRUEMAN2 AND RICHARD BURT3

1 Robert W. Hodder Chair in Economic Geology, Department of Earth Sciences, University of Western Ontario, London, Ontario, Canada 2 Consulting Geologist, Richmond, British Columbia, Canada 3 GraviTa Inc., Elora, Ontario, Canada

Introduction only one stable isotope 93Nb, but has eight radio- genic isotopes that range in mass from 89Nb to Tantalum derives its name from king Tantalus in 97Nb. Tantalum has two stable isotopes: 99.988 per Greek mythology. He stood in a pool of water cent occurs as 181Ta and 0.012 per cent as 180Ta. It beneath a fruit tree. When he tried to grasp some also has six radiogenic isotopes that range in mass fruit the branches moved and when he tried to from 177Ta to 183Ta. Niobium and tantalum both drink the water receded, which gives rise to the have a valency of +5 under most natural redox con- word tantalise. His daughter was Niobe, after ditions within the Earth and have nearly the same which the element niobium is named. In the geo- ionic radius. These elements have a high charge to logical literature, these elements are not referred to ionic radius and, because of this, they are insoluble as father and daughter, but rather as ‘geochemical in most geological fluids, are strong Pearson acids twins’ because their behaviour is very similar. that are only complexed by strong ligands such as Charles Hatchett discovered niobium, first called O2–, OH– and F– (Wood, 2005). However, these ele- columbium, in 1801 from a sample that was sent ments are soluble at weight per cent levels in sili- from Connecticut, USA. Anders Gustaf Ekeberg cate melts, particularly alkaline melts (Linnen and first discovered tantalum in 1802 but it was diffi- Cuney, 2005), and can attain even higher solubil- cult to distinguish tantalum from niobium. This ities in carbonatite melts (Mitchell, 2005). issue was not resolved until Heinrich Rose in 1844 and Jean Charles Galissard de Marignac in 1866 were able to demonstrate that niobium and tan- Distribution and abundance in the Earth talum were two different elements (Winter, 2011). The estimated abundances of niobium and tan- talum in the upper continental crust are 12 and Physical and chemical properties 0.9 ppm, respectively, which is enriched relative to the bulk continental crust, 8 and 0.7 ppm, Niobium sits above tantalum in the Periodic Table respectively (Rudnick and Gao, 2004). These and both have very high melting temperatures, values are much higher than the estimated con- 2468 and 2996 °C, respectively. A summary of centrations in primitive mantle, 548 ppb Nb and their properties is given in Table 15.1. Niobium has 40 ppb Ta (Palme and O’Neill, 2004). Both

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 362 robert linnen, dave trueman and ricard burt

Table 15.1 Selected properties of niobium and tantalum.

Property Value Units

Name Niobium Tantalum

Symbol Nb Ta Atomic number 41 73 Atomic weight 92.91 180.95 Density at 25 °C 8578 16670 kg/m3 Melting point 2468 2996 °C Boiling point 4930 5425 °C Hardness (Mohs scale) 6.0 6.5 Electrical resistivity at 25 °C 144 134 nΩ m Crystal structure Body-centred cubic Body-centred cubic Ionic radius (six-fold coordination) 64 64 pm

Table 15.2 Selected niobium and tantalum minerals and indicative contents of

Nb2O5 and Ta2O5.

Mineral name Formula Nb2O5 (%) Ta2O5 (%)

Columbite (Fe,Mn)(Nb,Ta)2O6 78.72 na

Tantalite (Fe,Mn)(Ta,Nb)2O6 na 86.17

Pyrochlore (Na,Ca)2Nb2O6(O,OH,F) 75.12 na

Microlite (Na,Ca)2Ta 2O6(O,OH,F) na 83.53

Tapiolite (Fe,Mn) (Ta,Nb)2O6 1.33 83.96

Ixiolite (Ta,Nb,Sn,Mn,Fe)4O8 8.30 68.96

Wodginite (Ta,Nb,Sn,Mn,Fe)O2 8.37 69.58

Loparite (Ce,La,Na,Ca,Sr)(Ti,Nb)O3 16.15 na

Lueshite NaNbO3 81.09 na

Euxenite (Y, Ca, Ce, U, Th)(Nb, Ti, Ta)2O6 47.43 22.53

Strüverite (Ti,Ta,Fe)O2 11.32 37.65

Ilmenorutile Fex(Nb,Ta)2x4Ti1-xO2 27.9 na elements are highly incompatible (partition in with oxygen; in fact, most niobium–tantalum min- favour of the melt over minerals), are enriched in erals are oxides, rather than silicates, sulfides or alkaline magmas and are characteristically phosphates. The most abundant niobium–tantalum depleted in calc-alkaline magmas (e.g. Winter, minerals are listed in Table 15.2. The two most 2010). Apart from minerals of niobium, tantalum important groups of minerals by far are colum- and tin, the highest concentrations of niobium botantalite and pyrochlore. Columbotantalite has and tantalum are observed in titanium-bearing an orthorhombic crystal structure, with the end- minerals, notably rutile and titanite. members being columbite-(Fe), columbite-(Mn), tantalite-(Fe) and tantalite-(Mn). However, tanta- Mineralogy lite-(Fe) is very rare because the Fe–Ta end-member is typically tapiolite, which has a tetragonal crystal Niobium and tantalum do not occur naturally as structure. Columbotantalite group minerals also free metals. The charge to ionic ratio of tantalum range from completely ordered to more disordered, and niobium results in strong bonds being formed and ixiolite is the completely disordered oxide Tantalum and niobium 363 phase. The pyrochlore group of minerals has deposits, but it is rarely the dominant mineral in numerous end-members because of the large an ore body. number of potential substitutions into its crystal Another source of tantalum is as a by-product structure (cubic). In terms of economic importance of tin mining. Cassiterite can contain weight per the most important are pyrochlore, the Nb end- cent levels of tantalum and niobium and also member and microlite, the Ta end-member. There commonly has inclusions of columbotantalite is considerable overlap in the mineralogy of nio- and rutile–strüverite–ilmenorutile. bium and tantalum deposits, although each com- modity has its own characteristic ore mineralogy. Niobium deposits are predominantly hosted by Deposit types carbonatites, where pyrochlore group minerals are by far the most important ore minerals. The All primary niobium and tantalum deposits are world’s largest niobium deposit is Araxá, which associated with igneous rocks and can be classi- contains primary pyrochlore. However, most of fied on the basis of the associated igneous rocks. the deposit consists of enriched, secondary laterite Three types of deposits are recognised: ore (Pell, 1996) with the main niobium mineral 1. carbonatite-hosted deposits (niobium and being pyrochlore-(Ba). Other niobium minerals industrial minerals); present in carbonatites include columbite-(Fe), 2. alkaline to peralkaline granites and syenites perovskite group minerals (loparite and lueshite) (niobium, yttrium, rare earths, zirconium, tin, and rare silicate minerals, such as niocalite. tantalum); However, these minerals generally are not eco- 3. peraluminous pegmatites and granites (tan- nomically important. The niobium mineralogy of talum, niobium, tin, tungsten, caesium, lithium). alkaline intrusions is similar to that of carbon- Tantalum and niobium minerals are also resis- atites, being dominated by pyrochlore and perov- tant to mechanical and chemical weathering and skite group minerals. However, a wide range of have high specific gravities. These properties ‘exotic’ minerals is also present (Salvi and Williams- favour the accumulation of these minerals as Jones, 2005). placer and alluvial deposits, which provide an Tantalum deposits are hosted by peraluminous important component of global production, espe- pegmatites and granites, where three mineral cially during periods of tight supply from primary groups, columbotantalite, wodginite and micro- mines. A number of tin deposits also contain nio- lite, dominate the ore mineralogy. The colum- bium and tantalum, particularly in south-east botantalite group minerals are the most common Asia. Processing of tin has led to accumulation of and contain both tantalum and niobium. Because of tin slags in the smelting process, which can be this, ore, particularly from central Africa, is com- re-processed and thus are another important monly referred to as ‘coltan’. The crystallisation of source of tantalum. The global distribution of granite-pegmatite melts results in an evolution these deposit types is shown in Figure 15.1. from niobium-rich columbite to tantalum-rich tantalite; deposits of the latter are associated with Carbonatite deposits the most evolved melts. Typically, there is also a progression from iron-rich columbite to manga- Carbonatites are igneous rocks that contain more nese-rich tantalite, although a number of different than 50 per cent carbonate minerals (calcite, iron–manganese crystallisation paths have been dolomite or ankerite). Most carbonatites occur in recognised (Černý et al., 1986). Wodginite is also rift settings, but it is clear that there are several an important tantalum ore mineral and tapiolite different types of carbonatites, many of which are is less abundant, but is nevertheless present in unmineralised and do not contain anomalous nio- many tantalum deposits. Microlite is a common bium concentrations. Carbonatites are genetically tantalum mineral that occurs in many ore related to strongly alkaline silicate rocks, but it is Figure 15.1 Global distribution of niobium and tantalum mines, selected deposits and occurrences. (Modified after Shaw and Goodenough, 2011.) Tantalum and niobium 365 not clear whether these are related by fractional are the Araxá and Catalão I. Both of these deposits crystallisation or by silicate–carbonatite melt occur in Late Cretaceous alkaline ultramafic– immiscibility (most likely both are important, carbonatite complexes of the Alto Paranaiba depending on the individual carbonatite). igneous province. However, both the alkaline silicate and carbon- The Araxá deposit originally contained approxi- atite melts are interpreted to be the product of mately 808 million tonnes of ore with an average melting of a metasomatised mantle. According grade of about 2.3% Nb2O5 (CBMM, personal com- to Mitchell (2005), niobium mineralisation munication, 2013). It lies within the Barreiro car- is associated with the melilitite family of bonatite complex, a circular intrusion approximately carbonatites. 4.5 km across (Figure 15.2). The surrounding The most important carbonatite-hosted nio- Proterozoic quartzites have been affected by fenite bium deposits in the world are in Brazil, which alteration (consisting of alkali feldspar, eckerman- produce approximately 92 per cent of the world’s ite, arfvedsonite, aegirine–augite and dolomite) for a niobium. The mines currently producing niobium distance of 2.5 km from the intrusive contact. In

N

012 km

Calcite carbonatite Host quartzite Glimmerite and schist dominant

Dolomitic carbonatite Mixed dolomitic Fault dominant carbonatite-glimmerite

Figure 15.2 Geology of the Araxá carbonatite complex. (Modified after Pell, 1996.) 366 robert linnen, dave trueman and ricard burt addition to these metasomatic minerals, the and carbonatite partially encircles the inner wallrock also contains hematite, goethite, anatase, carbonatites (Pell, 1996). The ore consists of both rare earth phosphate, titanite and pyrochlore pyrochlore and columbite. Separate rare earth (Nasraoui and Waerenborgh, 2001). The pyrochlore element (REE) zones with bastnäsite and monazite deposit is approximately 1.8 km across, and is asso- are also present. A past niobium producer is also ciated with dolomitic carbonatite that is surrounded located in Quebec at Oka. Between 1961 and 1976 by glimmerite, a rock that consists dominantly of this deposit produced 2000 tonnes of Nb2O5 from phlogopite, with dolomite and accessory magnetite 6.3 million tonnes of ore. At the time of the mine and apatite. Dolomite is the dominant mineral in closure proven reserves of 23 million tonnes of the carbonatite with subordinate calcite and 0.44% Nb2O5 remained (Pell, 1996). The Oka com- ankerite and accessory barite, apatite, magnetite, plex is a Cretaceous intrusion six km long and two perovskite, quartz, pyrite, phlogopite and sodic km wide which contains two intrusive carbon- amphibole. It is also noteworthy that the highest- atite centres that alternate with alkaline silicate grade niobium mineralisation is associated with rocks (Pell, 1996). There are several other carbon- phoscorite, a magnetite–apatite–phlogopite–car- atite complexes in Canada, particularly in British bonate rock. An important feature of the deposit is Columbia, that are being explored for niobium and that it has been deeply weathered, forming a thick rare earth elements. Of note is the Upper Fir lateritic cover, locally in excess of 100 m. Niobium deposit at Blue River, which has an Indicated was concentrated during the weathering process and Mineral Resource of 36.4 million tonnes at a grade primary pyrochlore was altered to secondary of 195 ppm Ta2O5 and 1700 ppm Nb2O5 (Chong and barium- pyrochlore (Pell, 1996). Postolski, 2011). The Catalão deposits are located approximately One of the largest deposits outside Brazil is the 200 km north-north-west of Araxá. They are also Lueshe carbonatite in the Democratic Republic lateritic deposits, with current resources of approx- of the Congo (DRC). The 120 million tonne imately 58 million tonnes at about 1.1% Nb2O5 deposit, intruded into the late Proterozoic (Anglo American, 2013). The geology of these Bugandian period country at around 500 Ma (Pohl, deposits is broadly similar to Araxá; pyrochlore 1994), stands out from the country rocks as a mineralisation is associated with dolomitic car- 200–300 m high plug approximately two km in bonatite and phoscorite (in particular a nelsonite diameter. The busorite core is surrounded by phase of magnetite–apatite with subordinate sovite, which is the host for the apatite and pyro- phlogopite (Cordeiro et al., 2010). Another chlore, which are themselves surrounded on the Brazilian deposit of note is Morro dos Seis Lagos, west by fenitised sediments with an outlier of located in Amazonas State. This deposit is reported rauhaugite containing up to 90 per cent magne- to contain 2.9 billion tonnes at a grade of 2.85 wt% site. The deposit is highly weathered, with a lat-

Nb2O5 (Pell, 1996). However, the environmental eritic cap (Philippo et al., 1997), with the sovite problems associated with attempting to mine in a horizons being up to 150 m in thickness. national park of virgin rain forest will almost cer- In Europe and Asia there has been past produc- tainly mean that the deposit will not be exploited. tion from the Fen carbonatite in Norway (roughly

Currently the only other niobium producer 1.4 million tonnes of 0.3% Nb2O5 (Pell, 1996)). exploiting carbonatites is the Niobec deposit at St. Pyrochlore is also present in carbonatite and phos- Honoré, Quebec, Canada, which contains 12.2 corite at the Kovdor mine in the Kola Peninsula, million tonnes at an average grade of 0.66% Nb2O5 Russia, and the Sokli carbonatite, Finland. However, (Pell, 1996). The St. Honoré Complex is a 6.5 by the former is an iron mine and the niobium min- 8 km elliptical intrusion that consists of dolomitic eralisation is not economic, whereas the latter is and ankeritic carbonatite surrounded by calcite primarily a phosphate deposit. Lastly, Tomtor is a carbonatites with magnetite and locally pyroxene large, weathered carbonatite in northern Siberia or micas. An outer ring of syenite, ijolite, urtite (Kravchenko and Pokrovsky, 1995). Tantalum and niobium 367

is disseminated and is interpreted to be magmatic Alkaline to peralkaline (Bastos Neto et al., 2009). However, the granite is granites and syenites also altered and albite and greisen alteration are Significant concentrations of niobium and tan- observed. talum occur in alkaline to peralkaline granites The Strange Lake deposit, along the Quebec– and syenites. These intrusions also contain Labrador boundary in Canada, is a well-studied high concentrations of zirconium, yttrium and example of a peralkaline granite-hosted deposit. rare earth elements and generally occur in rift Current grade and tonnage estimates of the or failed rift tectonic settings, although there Indicated Resource at the B-Zone of this deposit are exceptions such as the Lovozero and are of 140.3 million tonnes grading 0.933% total

Khibiny intrusions in the Kola Peninsula of REE, 1.93% ZrO2, 0.18% Nb2O5, 0.05% HfO2 and Russia. This style of mineralisation has received 0.08% BeO (Quest Rare Minerals Ltd. 2011). considerable interest recently because of related There is little doubt that magmatic processes high yttrium and rare earth element contents. concentrated zirconium, niobium and REE, but One of the characteristics of this deposit type is ore grade increases with the extent of hydro- the wide range of zirconium, niobium and rare thermal alteration and Salvi and Williams-Jones earth element minerals that are present, (2005) propose that ore metals were enriched dur- including hydrous or anhydrous silicate, phos- ing these hydrothermal events. phate, oxide and mixed silicate–phosphate min- The Meponde alkaline complex, just north of eral phases (Salvi and Williams-Jones, 2005). the Mozambique–Malawi border and 3 km east The dominant mineral of niobium and tan- of Lake Niassa, primarily consists of leuocratic talum is pyrochlore, but other complex phases syenites and nepheline syenite. The complex is such as eudialyte and secondary fergusonite are deformed, and appears to be a synform with limbs also important carriers of niobium. There are that dip generally westwards in conformity with no current producing mines of this type, but the enclosing rocks. Pyrochlore, including urani- they are potential sources of a variety of metals ferous pyrochlore (hatchettolite), is the most in the future. abundant niobium-bearing mineral phase, with

An important peralkaline granite-hosted the ore zone containing up to 8800 ppm Nb2O5. In deposit is the Pitinga tin–niobium–tantalum addition, a columbite zone contains grades gener- deposit in Amazonas State, Brazil. The deposit is ally exceeding 1% Nb2O5 and 500 ppm Ta2O5. associated with the ~1820 Ma1 Madeira A-type Other minerals include zircon, fergusonite, mon- granite, emplaced into a post-collisional exten- azite and other rare earth element minerals sional setting (Bastos Neto et al., 2009). This (Mroz, 1983). This 50 million tonne deposit has deposit contains 424 million tonnes of measured not been developed to date. and indicated resources (Salles, 2000) at a mining There are also several large peralkaline granites grade of 0.17% Sn (as cassiterite) with niobium in the Arabian Shield, the most important of and tantalum (pyrochlore and columbotantalite) which is the Ghurayyah deposit. The Ghurayyah recovered as by-products. The granite is unusual stock is approximately 900 m across and contains in that it has a massive cryolite deposit in its fine-grained disseminations of pyrochlore and core, ten million tonnes of 31.9% Na3AlF6. The Y-columbite. It contains an Inferred Resource of granite consists of four facies, a metaluminous 385 million tonnes of 245 g/t Ta2O5, 2840 g/t biotite–hornblende syenogranite, a peraluminous Nb2O5, 8915 g/t ZrO2, 1270 g/t Y2O5 and 140 g/t biotite–alkali feldspar granite and younger albite- U3O8 (Küster, 2009). Rare earth element concen- enriched and border albite-enriched granites. trations are also high, but grades have not been The albite-enriched granites contain riebeckite, reported. In terms of contained metal, this is the cryolite, and polylithionite, which indicate a per- largest tantalum deposit in the world, but it is not alkaline composition. Cassiterite in the main ore clear whether the tantalum can be economically 368 robert linnen, dave trueman and ricard burt separated from niobium (which is an order of mag- with peraluminous S-type granites and occur as nitude higher in concentration) or from the other late syn- to post-tectonic intrusions in collision heavy minerals. The Ghurayyah has not been belts, typically in association with shear zones. dated, but its age is likely around 580 Ma1 (Küster, Other elements of economic significance include 2009). It contains arfvedsonite and aegirine as well lithium, caesium, beryllium, rubidium, niobium as a wide range of accessory minerals including and tin. These pegmatites are assigned to the pyrochlore, zircon, smarskite, aeschynite-(Y), lithium–caesium–tantalum (LCT) family of pegma- columbotantalite and cassiterite. tites (Cˇ erný and Ercit, 2005). A complete discussion The second type of deposit hosted by peralka- of pegmatite classification is beyond the scope line rocks are those hosted by syenites. There are of this article, but tantalum pegmatite deposits a variety of different types of nepheline syenites are most commonly Complex-Type pegmatites, that are mineralised and commonly these intru- although Albite and Albite–Spodumene-Type peg- sions are well layered. Deposits of note are the matites, e.g., Wodgina, Mount Cassiterite, respec- Khibiny and Lovozero intrusions of the Kola tively, also host economic tantalum mineralisation Peninsula in Russia, the Khaldzan-Buregtey (Sweetapple and Collins, 2002). The Complex deposit in Mongolia, the Ilímaussaq and Motzfeldt Type is further subdivided into spodumene, pet- Complexes in Greenland and the Nechalacho alite, lepidolite, elbaite and amblygonite subtypes deposit at Thor Lake in the Northwest Territories on the basis of their dominant lithium, boron and of Canada. These deposits are primarily of interest phosphorus mineral species. for their rare earth element mineralisation and The best studied of these is the Tanco deposit, will not be discussed in detail. The most common which is a Petalite Subtype. This pegmatite is niobium mineral in these deposits is pyrochlore, approximately 1600 × 820 m and in the central por- although at Nechalacho the dominant niobium tion is greater than 100 m thick. It intruded mineral is fergusonite (see Salvi and Williams- metavolcanic rocks and has been dated at 2640 Ma1 Jones, 2005, for a review). The Motzfeldt Complex (Cˇ erný, 2005). The complex zoning consists of in Greenland is an exception, and here, niobium eight major zones (Figure 15.3) with the economic and tantalum are the primary commodities. Little mineralisation, Li, Cs and Ta, occurring in different has been published on this deposit, but according parts of the pegmatite. The pre-production grades to the Ram Resources website (2011), the Geological at Tanco were 2.07 million tons of 2160 ppm Ta2O5,

Survey of Greenland estimated a mineralised 7.3 million tons of 2.76% Li2O and 0.35 million ˇ zone of 500 million tonnes with average grades tons of 23.3% Cs2O (Cerný et al., 1996). Tantalum between 1320–1480 ppm Nb and 110–130 ppm mineralisation is dominantly hosted by aplitic and Ta, whereas Tukiainen (1988) concluded that the metasomatised (muscovite replacement) K-feldspar nepheline syenite could be host to 50 million pegmatite (Van Lichtervelde et al., 2007). Figure 15.4 tonnes grading between 300–1000 ppm Ta2O5. shows mineralised aplite from the Tanco deposit, Pyrochlore is the primary ore mineral for both where tantalum–niobium oxide minerals occur in niobium and tantalum. primary magmatic layers. By contrast, lithium mineralisation is in Intermediate Zones and the caesium mineralisation is within a separate Peraluminous pegmatites Pollucite Zone (Cˇ erný, 2005). Peraluminous pegmatites have historically been Two of the largest historic tantalum producers the most important source of tantalum. Many of are the Greenbushes and Wodgina deposits in these pegmatites have been mined intermittently Western Australia. The Greenbushes pegmatite, and may or may not be in current production. in the Yilgarn Craton, contains Measured, They are described here because they are well Indicated and Inferred Resources of 135.1 million documented and may resume tantalum produc- tonnes at 220 ppm Ta2O5. It is a Spodumene tion in the future. They are commonly associated Subtype pegmatite that intruded into a shear Tantalum and niobium 369

1100 1000 900 800 700 600 500 400 0 100 200 300 400 500 Elevation in feet Distance in feet above sea level

Zone 20 Wall zone Zone 70 Quartz zone Ta Zone 30 Aplitic Albite zone Cs Zone 80 Pollucite zone Zone 40 Lower intermediate zone Ta Zone 90 Lepidolite zone Li Zone 50 Upper intermediate zone Amphibolite Xenolith Ta Zone 60 Central intermediate zone

Figure 15.3 Longitudinal east–west section through the Tanco pegmatite. (Modified after ˇ ernýC et al., 1996). (Cs, caesium; Li, lithium; Ta, tantalum.)

Figure 15.4 Banded aplite ore from the Tanco deposit. Tantalum mineralisation is contained in the banded aplite (red and blue-grey at the bottom). A massive quartz zone (dark grey) is at the top and coarse, crystalline white beryl separates the aplite and quartz. The scale is approximately 2 m across. (Courtesy of R. Linnen.)

zone and consists of four layers, a Li-Zone, is also an important ore mineral and early formed K-Zone, Na-Zone and Border Zone. The tantalum tantalum minerals occur as inclusions in cassit- mineralisation is associated with a massive erite and tourmaline (Partington et al., 1995 and albite–quartz-rich unit, and like Tanco, the Fetherston, 2004). This pegmatite crystallised at lithium is mined from a different unit. Cassiterite relatively high temperatures and pressures, from 370 robert linnen, dave trueman and ricard burt

750°C and five kbar to a greisen and metasomatic from pegmatites is currently being produced in stage at 620 °C and five kbar (op. cit.). the DRC. Tantalite Valley in Namibia is also a At Wodgina and Mount Cassiterite tantalum former tantalum-producing area. occurs in Albite- and Albite–Spodumene-Type Tantalum-bearing pegmatites in Brazil are pegmatites that are related to a suite of 2890 to numerous and widely distributed, but are typi- 2830 Ma1 post-tectonic granite plutons in the cally relatively small (less than one million North Pilbara Craton. The last published reserves pounds tantalum in size) and are generally mined (Sons of Gwalia, 2002) for the Wodgina mine was on a semi-industrial or even artisanal scale.

63.5 million tonnes of 370 ppm Ta2O5. The main However, the Mibra (Volta) mine, near the city of dyke is approximately one km long and up to Nazareno in the state of Minas Gerais, hosts a 40 m wide. It lacks the zoning that characterises series of pegmatites, with a total reported (indi- Complex pegmatites and consists of primary cated) resource of 25.5 million tonnes grading layers of massive albite with rare quartz, spessar- 290 ppm Ta2O5 (Resende, 2011). Mining is cur- tine and muscovite, and a central banded aplite rently concentrated on the 6.32 million tonne that is largely granitic in composition. Tantalum pegmatite ‘A’ with grades of 375 ppm Ta2O5, minerals are primarily in the massive albite, 92 ppm Nb2O5 and 283 ppm Sn (Mining Journal, where they are interpreted to be primary, but 2010). These pegmatites are Early Proterozoic in metasomatic styles of mineralisation are also age and are hosted by an Archean greenstone belt observed (Sweetapple and Collins, 2002). The at the southern border of the São Francisco Craton Mount Cattlin deposit, also in Western Australia, (Lagache and Quéméneur, 1997). There are a is a lithium deposit that produces a minor amount number of flat, subhorizontal sills, the largest of of by-product tantalum (Fetherston, 2004). which measured 1000 by 150 by 20 m before Current tantalum production elsewhere is pri- exploitation. The pegmatites show internal zon- marily from pegmatites in Africa and Brazil, but ing with an aplitic wall zone, an intermediate less information on these pegmatites has been zone that contains 20 to 30 per cent spodumene published, at least in recent years. In Africa, most and a core zone that is comprised of 60 to 80 of the tantalum pegmatite deposits are associated per cent spodumene crystals (Lagache and with the Pan-African orogeny. One of the most Quéméneur, 1997). important tantalum pegmatites in Africa is Kenticha in Ethiopia. This is a Spodumene-Type Peraluminous granites pegmatite that occurs as a dyke over two km long and 400–700 m wide that intruded at 550 Ma1 The last class of niobium–tantalum deposits (Küster, 2009). It is complexly zoned but Küster considered here are those hosted by peralumi- et al. (2009) simplifies the zones into three units. nous granites. They account for only a minor Most of the tantalum mineralisation occurs in amount of tantalum production, but geochemi- the Upper Zone, which also contains most of the cally are similar to LCT pegmatites, being spodumene mineralisation and is thought to rep- enriched in lithium, fluorine and phosphorus. resent the most evolved unit (bottom-to-top crys- The Yichun deposit, China, reported to contain tallisation; Küster et al. (2009). Other pegmatites approximately 175 million tonnes (personal in Africa include Morrua and Marrapino in communication, 1996), is a vertically zoned Mozambique, the latter of which is deeply weath- granite occupying the top of the south-eastern ered. In Central Africa, particularly the DRC, the section of the Yashan granite batholith. The Manono deposit in Katanga was the largest his- upper portion of the granite, which carries the torical producer. Here, most production is from majority of the tantalum values, is an albitic shallow deposits and from the many eluvial and granite high in lepi dolite and topaz, relatively alluvial placer deposits that are pegmatite derived low in quartz, and depleted in dark minerals. (Fetherston, 2004), although some primary ore The lower portion, which has sub-economic Tantalum and niobium 371

tantalum values, is high in biotite mica rather Pyrochlore than lepidolite (Yin et al., 1995). e.g. Niobec mine The granites at Beauvoir, France (Raimbault et al., 1995) and in Egypt (Abu Dabbab and Nuweibi) Crushing, milling and sizing stages contain disseminated cassiterite–columbotantalite Undersize to mineralisation that is apparently magmatic, as waste well as peripheral tungsten–tin mineralisation that Coarse desliming Fine desliming is hydrothermal. These deposits have not pro- duced, but Abu Dabbab contains approximately Coarse carbonate Fine carbonate

40 million tonnes of 243 g/t Ta2O5 and Nuweibi flotation flotation

98 million tonnes of 146 g/t Ta2O5 (Küster, 2009). Granites which have formerly produced minor Desliming Slimes quantities of tantalum are the Orlovka deposit in Low intensity Russia (Badanina et al., 2004), and at Podlesí and Magnetite Cínovec in the Zinnwald area of the Czech magnetic separation Republic (Breiter et al., 2007 and Rub et al., 1998). Diamine collector Pyrochlore flotation Tailings These deposits have several features in common: most are associated with late, multi-phase intru- Xanthates Sulfide flotation Sulfides sions; magmatic-style columbotantalite is dis- seminated in the granite; and there is also a Hydrochloric Concentrate leaching hydrothermal stage, typically with cassiterite– acid wolframite. The granites are also vertically zoned, Concentrate (54% Nb2O5) in particular with respect to the variety of lithium mica present. The Kougarok tin–tantalum– Figure 15.5 Niobium production flowsheet for the Niobec mine. (Modified after Shaw and Goodenough, niobium deposit in the Seward Peninsula of 2011.) Alaska is another granite-hosted deposit of note (Puchner, 1986). Columbotantalite ores almost exclusively uti- lise standard wet gravity concentration equip- Extraction methods and processing ment including jigs, spirals and tables, at least for production of a low- to medium-grade ‘heavy min- Tantalum and niobium minerals are recovered eral’ concentrate. While the trend will be toward through industry-standard open pit, underground ever-finer recovery by gravity using centrifugal- and placer mining methods. In the case of pri- type concentrators, there is no reason to assume mary sources or lode deposits, ores are usually that any other technology will take over, although mined by standard methods, e.g. drill, blast, and tantalum flotation has been used commercially in muck cycles, prior to further comminution and at least one operation. The heavy mineral concen- concentration. In deposits where weathering trates are often further concentrated in ‘dry’ plants processes are advanced, initial processing may be consisting of such units as dry gravity concentra- facilitated through simple screening. tors, magnetic and electrostatic concentrators and Pyrochlore ores are concentrated by flotation, occasionally removal of sulfides by flotation preceded by removal of irrecoverable ultrafines to (Figure 15.6). achieve acceptable concentrate grades – generally Artisanal mining of both alluvial and heavily weathered ores utilises the very simplest tech- in excess of 50% Nb2O5 – and rejection of unwanted contaminants. Several flotation stages niques. Primary concentration generally consists as well as electrical separation are required of nothing more than a ‘ground sluice’ and pos- (Figure 15.5). sibly winnowing. The primary concentrates are 372 robert linnen, dave trueman and ricard burt

Columbotantalite Crushing, milling ore & sizing Ultrafines to waste

Coarse gravity Fine gravity concentration concentration using jigs, spirals using centrifugal & tables concentrators

Heavy mineral concentrate dewatering Light minerals & drying to waste

Air tables, electrostatic & magnetic separation

Tantalum concentrates Tin concentrates Other heavy minerals 20–40% Ta2O5

Figure 15.6 Schematic summary of the processes involved in the preparation of tantalum concentrates. sold to central secondary processing plants sim- increase the tantalum content to a level suitable ilar to those used in larger-scale industrial plants. for chemical processing. The latter are generally Pyrochlore concentrates are converted by alu- referred to as ‘syncons’. While the former are gener- minothermic reduction, followed by electron ally sold direct from tin smelter to processor, the beam refining, to ferro-niobium, which is the greatest source of the latter are ‘old’ slags dumped product that dominates the niobium market. over the decades, and later dug up and stockpiled, Tantalum concentrates generally contain bet- for treatment when tantalum prices peak. ween 25–35% Ta2O5, much reduced from the Production of tantalum metals and chemicals

1970s ‘norm’ of 45–50% Ta2O5. While lower is a multi-stage process (Figure 15.7), ideally grades can be treated, the relative chemical suited to batch or semi-batch style production, as processing cost is usually regarded as prohibitive. some of the procedures are quite time consuming. Those concentrates from central Africa com- Many of the products are made to order, which monly contain a similar or higher content of can be as long as six months from receipt of ore to

Nb2O5, whereas those from Canada, Brazil and final dispatch of a specific product.

Australia have a significantly lower Nb2O5 The first stage is to convert tantalum concen- content (Burt, 2011). trates to an intermediate chemical – generally

Tin slags, primarily from south-east Asia, are potassium tantalum fluoride (K2TaF7) or tan- the other main feedstock to tantalum processors, talum oxide (Ta2O5). The ‘standard’ method, essentially in two ways. Higher-grade tin slags which is essentially unchanged for decades, is to

(those containing more than 2% Ta2O5) are suitable digest the ore at high temperature in a sulfuric for processing directly, whilst lower grade slags acid – hydrofluoric acid mix (or ‘neat’ HF) and,

(commonly with 1–1.5% Ta2O5) require treatment – after filtering out the insoluble minerals, further generally by an aluminothermic process – to processing via solvent extraction using methyl Tantalum and niobium 373

Radioactive Concentrates, High fluorine waste & tin slags, scrap & acid waste insolubles

Extraction- Acid digestion crystallisation

Heavy metal & ‘KTaF’ acidic waste

Tantalum Finishing Sodium reduction powder

Processing/ Forging, rolling, Mill products, Melting Cleaning, sintering cutting annealing targets, etc

Scrap

Figure 15.7 Schematic flowsheet for the production of tantalum from concentrates, tin slags and scrap. (‘KTaF’, potassium fluortantalate).

isobutyl ketone (MIBK) or liquid ion exchange titanium chloride gas. The niobium–tantalum using an amine extractant in kerosene. This pro- oxichloride gas is then further chlorinated to duces highly purified solutions of tantalum and produce niobium and tantalum chlorides, NbCl5 niobium, from which tantalum is crystallised, as and TaCl5. These are fractionally distilled and potassium fluortantalate (‘KTaF’) by reaction the niobium chloride subsequently reacted with with potassium fluoride. Where the level of steam to produce the hydroxide, which is cal- niobium in the feed warrants, it is recovered as cined to oxide. The tantalum chloride is reacted niobium oxide (Nb2O5) via neutralisation of the with ammonium hydroxide to produce the oxide niobium-fluoride complex with ammonia to (Tantalum-Niobium International Study Center form the hydroxide, followed by calcination to (T.I.C.), 2011). the oxide. Metal powder, including the precursor to Alternative methods do exist and are used capacitor-grade powder, is produced by sodium when they are better suited to particular local reduction of the potassium tantalum fluoride in a conditions. One used for a titanium–niobium– molten-salt system at high temperature. The tantalum–rare earth mineral concentrate metal can also be produced by the carbon or alu- involves blending the crushed concentrate with minium reduction of the oxide or the hydrogen or coke and passing this through a chlorination alkaline-earth reduction of tantalum chloride. stage which separates out the rare earths and The choice of process is based on the specific other elements including most of the thorium. application and whether the resultant tantalum The temperature of the resulting titanium–nio- will be further consolidated by processing into bium–tantalum oxichloride gas is then decreased, ingot, sheet, rod, tubing, wire and other fabri- which causes the iron, thorium and alkali metals cated articles. to precipitate out. The cleaned titanium–nio- The consolidation of metal powder for ingot bium–tantalum oxichloride gas is next cooled to and processing into various metallurgical prod- a liquid and distilled to separate out low-boiling ucts begins with either vacuum arc melting or 374 robert linnen, dave trueman and ricard burt electron beam melting of metal feed stocks, tronics industries. These applications continue comprising powder or high-purity scrap where to account for over 50 per cent of all tantalum the elements with boiling points greater than consumption, with the most significant seg- that of tantalum are not present. Double- and ment being the tantalum capacitor, used not triple-melt ingots thus achieve a very high level just in consumer electronics, such as com- of purification. puters and smart-phones (contrary to popular belief, modern cellphones utilise very little, or even no, tantalum), but also in many automo- Specifications and uses bile components (e.g. anti-lock braking sys- tems, airbag activation, GPS and Close to 90 per cent of niobium production is engine-management modules), in aircraft and used to make high-strength low-alloy steels for medical appliances, such as pacemakers, (T.I.C., 2011). The addition of ferro-niobium to defibrillators and hearing aids. While the steel increases strength and toughness as well growth rate in this sector, in terms of metal as reduces the weight. High-strength low-alloy usage, has been minimal, or even negative, steels are used to manufacture oil and gas pipe- rapid increases in performance (in terms of lines, car and truck bodies, tool steels, ship capacitance per gram of tantalum) and conse- hulls and railroad tracks (T.I.C., 2011). The quent miniaturisation means the growth in electrical resistance of niobium–titanium and number of units used has continued to grow. niobium–tin alloy wire drops to virtually zero Elsewhere in electronics, tantalum metal is at or below temperature of liquid helium used in hard-disk drives and ink-jet printer (–268.8 °C). Consequently, these alloys are uti- heads. A significant growth area is for sputter- lised in superconducting magnetic coils in ing targets, with tantalum being used as the magnetic resonance imagery (MRI), magneto- diffusion barrier between interconnects on encephalography, magnetic levitation transport copper-based semiconductors. Lithium tanta- systems, and particle physics experiments. late is used in surface acoustic wave filters Elsewhere, niobium is used for alloys and where the electronic signal wave dampening chemicals including carbides. Niobium oxide results in better audio and video output in mo-

(Nb2O5) is used to manufacture lithium niobate bile phones, audio systems and televisions.

(LiNbO3) for surface acoustic wave filters, Tantalum oxide has a high refractive index and camera lenses, coating on glass for computer hence lenses in spectacles can be made lighter screens and ceramic capacitors. Niobium car- and brighter, as can lenses in mobile phones bide is used for cutting tools and niobium metal and the high-end digital camera market. and alloys have various speciality applications Tantalum carbide-based cutting tools are now (T.I.C., 2011). a relatively minor segment of the tantalum The physical and chemical properties of tan- industry. talum have been long known; such properties Other uses for tantalum include corrosion- having made its discovery difficult in the first resistant objects, prosthetic devices and high-tem- instance. Its use in light-bulb filaments hails perature alloys for turbines. Single-crystal alloys, from over 100 years ago, while it has been used containing three to eleven per cent tantalum, are to improve the performance of optical glasses used in jet (and some land-based) engine turbines for more than 75 years. Its widespread usage, where they offer greater resistance to hot gases however, arose with the advent of the tran- than other alloys, allowing higher operating tem- sistor and solid-state avionics in the Korean peratures and hence improved fuel efficiency War, in a complementary role with the selenium (T.I.C., 2011). rectifier and transistor. By the 1980s close to 70 Overall, there was a rapid increase in tantalum per cent of tantalum was destined for the elec- demand (as opposed to supply) in all its forms Tantalum and niobium 375

7500 5 O

2 5000

2500 Figure 15.8 Tantalum demand, shown as Thousand lbs Ta

tantalum pentoxide (Ta2O5) equivalent, based upon tantalum processor shipments, 1971–2012. (Data from various T.I.C. 0 bulletins, trade magazines, government 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 statistics and industry observers.) during the 1990s, but the growth rate has slowed lower-cost capacitor, that only operated in lower since 2000 to about a one per cent Compound peak-temperature ranges. Accordingly, it failed to Growth Rate (Figure 15.8). make military specifications in many critical end uses. Tantalum capacitors have a relatively small, Recycling, re-use and resource efficiency but specific portion of the overall capacitor market, with aluminium, ceramic and wet elec- According to the USGS (2011a), niobium is recy- trolytic capacitors having a substantially greater cled when niobium-bearing steels and superalloys market share. The tantalum industry regards are recycled, but scrap recovery specifically for that, where substitution of tantalum with an niobium is negligible. Statistics for the amount of alternative is possible, it has already occurred, recycled niobium are not available, but this could and where tantalum capacitors are currently be as high as 20 per cent of apparent consumption. used, their potential for substitution is low. Tantalum is recycled from electronic components Likewise, it is expected that other energy-storage and from tantalum-bearing cemented carbide and devices will continue to have a greater role than superalloy scrap – about 25 per cent originates tantalum in future green technologies – for from material recycled from the manufacturing example, in electric or hybrid vehicle charging stage of these components (i.e. ‘new scrap’). systems where volume is of lesser consequence (T.I.C., 2011). However, statistics are not available and electrical storage capability is paramount. for the amount of post-consumer tantalum that is However, in airbag triggers, the tantalum capac- recycled and much of that of electronic origin is itor is a must – it is the only capacitor, to date, ignored because of the small amounts involved in that can operate at both high temperatures and miniaturised capacitors. sub-zero temperatures with essentially no varia- tion in performance. Tantalum-carbide machining, cutting and Substitution grinding tools have longer operating lives and higher operating temperatures, which allow As tantalum prices rose in the early 2000s, nio- faster machining times and finer finishing. These bium was more widely substituted into electronic advantages over tungsten carbides are surren- capacitors. Niobium, having a lower dielectric dered as tantalum prices rise, and industry falls capacity, resulted in a physically larger but back on convention until tungsten prices rise. 376 robert linnen, dave trueman and ricard burt

Environmental aspects of niobium acids – generally sulfuric acid with at least and tantalum stoichiometric quantities of hydrofluoric acid. The acids are neutralised with alkali salts and the Niobium and tantalum ores do not pose any spe- resulting mass has thorium and uranium concen- cial environmental problems, and standard indus- trations similar to the feed raw material. try-wide safety and environmental precautions Nevertheless, appropriate, fully licensed, discharge during mining and mineral concentration are sat- containment is required. Later stages of the process isfactory. Flotation reagents, where used, are gen- require appropriate safety procedures but other- erally biodegradable and, because of a lack of wise do not pose any environmental hazards. sulfide minerals associated with the deposits, the The solid forms of tantalum and niobium do tailing impoundments pose little environmental not pose special environmental problems. There hazard such as acid mine drainage. is no reported information on toxicity of the Tantalum concentrates produced from pegma- metals and alloys, and the only associated health tites generally contain minute quantities of natural hazards are due to powders, which, like any other thorium and uranium, and, as a consequence, the powder, can be an irritant. As the key to increased concentrates are described as naturally occurring tantalum capacitor efficiency is the fineness of radioactive materials (NORMs). Some concen- the powder, care must be taken to ensure a static- trates, from Mozambique, central Africa and and ignition-free environment as, like many Brazil, have higher levels and are classified as other very fine powders, they are pyrophoric and ‘Class 7’ under ’Dangerous Goods’ under the may explode if handled incorrectly. International Maritime Dangerous Goods (IMDG) Comprehensive regulation of materials is now Code and other modal regulations (IMO, 2011). in place in the European Union, called REACH Consequently, minor requirements in terms of (Registration, Evaluation, Authorisation and packaging, labelling and staff training have to be Restriction of Chemicals). All companies wishing met. NORMs pose a very low radiological risk dur- to produce substances in the EU or import them ing transport and, further, the regulations are into the EU must ensure they meet their registra- designed so that safety is provided by passive safety tion obligations if they are to continue with their inherent in the package. Next-generation concen- activities; further guidance can be obtained from trates, especially those produced from alkaline and the European Chemical Agency website (ECHA, peralkaline deposits, will almost certainly have 2011) or their national competent authority. higher levels of these radioactive elements, and Dossiers for some of the many niobium and tan- some form of on-site processing prior to shipment talum substances are being developed. Along sim- may well be required. This could include ilar lines to REACH, the US government is by-product production of uranium chemicals drafting legislation to have more comprehensive (uranium concentrates are reportedly produced regulation of chemicals. from ores containing as little as 0.05% U), or simply suitable impoundment at site. Likewise, conversion of niobium concentrates Geopolitical aspects to ferro-niobium poses no special environmental problem, as long as standard safety procedures are From 1998 to 2003 central Africa was embroiled in in place. The pyrochlore concentrates used to what has become known as ‘Africa’s World War’. produce ferro-niobium also contain thorium and The underlying reasons for this conflict are outside uranium, which report in the ferro-niobium slag. the scope of this chapter, but one undoubted effect This slag contains elevated levels of thorium and of it was the involvement of illegal forces in the uranium and is generally stored on site. mining and transportation of minerals, especially The first step in treatment of tantalum from eastern DRC, enabling them to continue concentrates requires their digestion in strong funding their war – and subsequent rebel activities. Tantalum and niobium 377

While tantalum has had only a very minor role finance or benefit armed groups in the DRC or an (generally less than five per cent of all such fund- adjoining country’. Other countries, including ing was from tantalum) it rapidly became the Canada as well as the EU, are considering similar cause célèbre for many activist groups. The tan- legislation. talum industry may have appeared slow to react, In addition, the International Conference on the but several of the larger processors had disen- Great Lakes Region (ICGLR), which is composed of gaged by 2003. However, it was the 2008 report eleven GLR states, has developed regulations to be by the United Nations Group of Experts that put in place by all member states. While embracing really challenged the tantalum, and other, indus- the iTSCi Programme as one programme, the tries to act (Stearns et al., 2008). By mid-2009, a ICGLR is also supporting the development of a tin industry-led Due Diligence initiative (the ‘forensic’ fingerprinting method for columbite– ‘iTSCi’ Programme) was being put in place, with tantalite concentrates, so that the origin may be the tantalum industry becoming partners in early traced. While not, as yet, in commercial use, min- 2010. By mid-2011 it had been activated throughout eral chemistry data is available for a number of Rwanda and the Katanga province in the DRC. African pegmatites through the efforts of the Concurrently, the electronics and telecommu- German Federal Institute for Geosciences and nications industries, through the Electronic Natural Resources (BGR) (Melcher et al., 2008 and Industry Citizenship Coalition, Global e-Sustain- Graupner et al., 2010). The fingerprinting combines ability Initiative (EICC/GeSI), developed the a number of methods including age dating, mineral ‘Conflict Free Smelter’ programme which, since liberation analysis and major- and trace-element 2010, requires processors to provide credible evi- analysis of tantalum minerals. dence of their conflict-free sourcing (Electronic Industry Citizenship Coalition, 2011). Governments and intergovernmental agencies World resources and production also reacted, and in early 2012 the Organisation for Economic Cooperation and Development Current significant resources and supplies of nio- (OECD) published a set of ‘guidelines’ that had bium and tantalum are shown in Figure 15.1 and been endorsed by governments, NGOs and quantified in Tables 15.3 and 15.4, respectively. industry, and these are being piloted in 2011–12 Although niobium deposits are globally widely (OECD, 2011). The United Nations have also dispersed, in excess of 90 per cent of the world’s published guidelines, these closely following the niobium is produced from two mines in Brazil: OECD guidelines. CBMM’s operations at Araxá, and Anglo American’s At the same time, two Senators sponsored leg- Catalão operations. Canada’s Niobec Mine, oper- islation that was eventually included in the US ated by Camet Metallurgy, produces between five Financial Stability Act (commonly known as the and ten per cent of global supply and the remainder ‘Dodd Frank Act’) promulgated in July 2010. The comes from loparite concentrates from the Securities and Exchange Commission (SEC) pub- Karnasursk Mine in the Russian Kola Peninsula. lished the appropriate regulations in mid 2012. In the ten-year period from 1997 to 2007 produc- Essentially the law places a legal obligation tion increased rapidly, from just 20,000 tonnes of on all US companies that report to the SEC to niobium in concentrates in 1997 to the level of declare the use of ‘conflict minerals’ (cassiterite, approximately 100,000 tonnes from 2007 onward, wolframite, columbotantalite and gold) and metals apart from the decline due to the 2008 economic originating from the DRC or adjoining countries. recession (Figure 15.9). The law does not prohibit the use of these ‘conflict Tantalum deposits are similarly widely dis- minerals’: it states that ‘… a product may be labeled persed, and, like niobium, there are few mines in as ‘DRC conflict free’ if the product does not con- production. Annual tantalum concentrate pro- tain conflict minerals that directly or indirectly duction climbed at an eight per cent annual Table 15.3 Estimated global reserves and resources of niobium pentoxide, Nb2O5. (Data from numerous company websites and other published sources.)

Contained Nb2O5 in proven and probable Contained Nb2O5 in measured and indicated Source reserves (thousand tonnes) resources (thousand tonnes)

Australia 165 164 Brazil 44 78,133 Canada 1810 3005 China – 2200 Egypt – 4 Malawi – 174 Mozambique – 52 USA – 129 Total 2019 81,662

1 Inferred resources are also reported in Brazil, Gabon, Kenya, Canada, Tanzania, Ethiopia, Saudi Arabia, Spain, Angola, Mozambique and USA. 2 Some deposits are omitted because no reliable reserve or resource data are available.

Table 15.4 Estimated global tantalum reserves and resources of tantalum. (Data from USGS, 2011b; Burt, 2010; DNPM, 2011.)

Most likely resource base

Source (tonnes Ta2O5) Percentage of resources Reserves (tonnes Ta)

Brazil 129,274 40 87,360 Australia 65,771 21 40,560 China and Southeast Asia 33,112 10 7800 Russia and Middle East 31,298 10 — Central Africa 28,576 9 3120 Other Africa 21,318 7 12,480 North America 5443 2 1500* Europe 2268 1 — Total 317,060 100 152,820

*Deemed uneconomic at 2010 prices (USGS, 2011b).

125,000

100,000 5 O

2 75,000

50,000 Tonnes Nb

25,000

Figure 15.9 Global niobium production, 0 1990–2011. (After Schwela, 2011, with 1990 1995 2000 2005 2010 additional data from T.I.C. Bulletins.) Tantalum and niobium 379 growth rate through the later 1990s, resulting in exacerbated by governments’, and industry’s, major expansion at both Greenbushes and efforts to break the link between mining and the Wodgina in Australia, and the development of financing of illegal forces in central Africa. other, smaller, mines in Africa and Australia. This new production peaked in 2002, at approxi- mately 2300 tonnes, prior to stabilising at approx- Future supplies imately 2100 tonnes until the 2008/9 recession. Mine production declined dramatically to a little The Araxá niobium mine has a history of increasing over 1000 tonnes due to the temporary closure production to meet long-term market demand and of several of the larger industrial mines. The it will therefore remain the major producer of nio- downstream industry recovered rapidly, mined bium. Nevertheless, new or alternative supplies supplies tightened, prices rose accordingly and do exist and could be brought on stream if dormant supplies are returning to production and required. Worldwide known resources of niobium mined production is returning toward pre-2008 are very large in comparison with current global levels (Figure 15.10). This tightening was partially consumption.

6000

5000

4000 5 O 2

3000

Thousand lbs Ta 2000

1000

0 2000 20012002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 Estimated

Secondary Other Africa Central Africa South America

North America Europe Asia Australia

Figure 15.10 Global production of tantalum from concentrates and tin slags, 2000–2012. (Data from various trade documents, government statistics, and industry sources and estimates.) The ‘Secondary’ category includes low-grade tin slags and shipments of old stocks from the US Defence Logistics Agency ‘stockpile’. 380 robert linnen, dave trueman and ricard burt

With high tantalum prices in 2011, some of the Price data are usually shown in a form other idle mine capacity in Australia (the Wodgina than that of elemental metal. For example, tan- mine), Canada and Mozambique came back on talum concentrates are quoted in the oxide form stream. The Mibra Mine, in Brazil, having cap- Ta2O5 – there are no data available for ‘down- tured market share during the economic down- stream’ products, i.e. those from the processors to turn, may continue to produce in excess of 20 per manufacturers. Of this, the tantalum oxide, cent of the world’s demand. Other mines that will approximately 82 per cent is tantalum metal. The continue to produce in excess of 100,000 lbs Ta2O5 contained metal price is therefore about 122 per per annum include Marrapino (Mozambique), cent of that shown in the oxide form. Niobium Kenticha (Ethiopia), Pitinga (Brazil) and Yichun prices were historically quoted in terms of pyro- (China), while small, industrialised mines in Asia chlore or columbite concentrate. However, since and South America will continue to contribute to the early 1990s when pyrochlore was no longer overall supply. sold on the open market, they are now normally Artisanal production in South America, south- quoted in terms of ferro-niobium – either as ‘ferro- east Asia, Nigeria and elsewhere in Africa will niobium’ itself or based upon the niobium content. continue. Although production in individual Historically, niobium and ferro-niobium prices mines is minor, in total it accounts for approxi- had held essentially constant, or declined once mately 20 per cent of global production. As inflation had been taken into account. Production industry incorporates appropriate due diligence capacity increased to meet demand and with rap- into its central African supply chain, ‘conflict- idly increasing demand, a price correction was free’ production will return to, and may even inevitable and occurred in 2007 (Figure 15.11). exceed, historic levels in the Great Lakes Region, Ferro-niobium prices rose significantly and especially in the DRC. continued to rise throughout the global financial There are several large deposits currently crisis, with the benchmark price for standard- being examined for potential niobium and tan- grade Brazilian ferro-niobium experiencing a talum production. These include the Ghurayyah three-fold increase in four years (Naidoo, 2011). deposit in Saudi Arabia, the Motzfeldt Complex in This upturn in niobium prices in general corre- Greenland, the Zashikhinsky deposit in Russia’s sponds to that of all of the rare metals, and is Irkutsk region, the Abu Dabbab deposit in Egypt, likely related to the rapid expansion of Chinese and the Blue River and Nechalacho (Thor Lake) industry and other emerging economies. deposits in Canada. Tantalum prices have been decidedly more volatile (Figure 15.12). Over the last forty years, ‘normal’ tantalum pricing has essentially kept Prices pace with inflation, but there has been a roughly ten-year cycle of ‘spikes’. The 1980 spike was Neither niobium nor tantalum are exchange- caused by (unfounded) rumours that a major traded commodities, with most transaction prices supplier was about to run out of ore, whereas the between buyers and sellers subject to long-term 2000 spike was due to panic over-buying and confidential contracts. Only a limited amount of building inventories during the ‘dot.com boom’ on concentrate is traded against the ‘spot’ market, the assumption that mines could not keep up with and even here prices are little more than ‘best esti- demand. When clearer heads prevailed in 2002, it mates’ that are available through trade-journal was generally admitted that there was never a true subscription (e.g. Metal Pages, Asian Metals) or shortage of any significance, partially due to from annually compiled data, such as that of the increased artisanal mining especially in central US Geological Survey or the Brazilian government. Africa (this is now accepted as being ‘conflict The industry association, the Tantalum-Niobium mining’ and was a result of – rather than the cause International Study Center, by mandate, cannot of – the civil war in the DRC). While the ‘spot’ collect or publish price information. prices in 2000 were far in excess of the 1980s peak, Tantalum and niobium 381

25

20

15

10 US$/lb Nb US$/lb

5 Figure 15.11 Average annual ferro- niobium price per pound of contained niobium, 1991–2012. (Based upon various 0 technical and commercial sources.) 1991 1995 2000 2005 2010

250 5 O

2 200

150

100

Figure 15.12 Average annual price for tantalum pentoxide contained in 50 US$ per pound contained Ta

30% tantalum pentoxide (Ta2O5) concentrate, 1977–2012. (Based upon various technical and commercial 0 sources.) 1975 1980 1985 1990 1995 2000 2005 2010 2015 it actually reflected far less of the market, and the a steady enough level) to satisfy the industrial spot + contract price peak was substantially lower. miners while at the same time not resulting in The 2010/11 spike is the most complex: its cause equipment re-design to use alternatives without actually started as early 2003 when mine supply such price volatility as tantalum, even if they are outstripped demand, and inventories throughout less technically effective. the supply chain grew to excessive levels, forcing prices down and the closure of several big, higher cost, mines by 2007/8. The dramatic, short-term Outlook reduction in demand in 2008/9 masked this, but while demand could and did rapidly return to The known global resources of niobium are normal levels, mine supply could not. As a result exceptionally large and geological availability inventories declined, and prices once again spiked. is, therefore, unlikely to be a significant problem While prices are declining again in late 2011, it is in the foreseeable future. More problematic, how- too early to determine whether they will return to ever, is a lack of suppliers and, at present, con- the historic trend, or will remain at a level (and at sumers are essentially dependent on the policies 382 robert linnen, dave trueman and ricard burt of a single country, Brazil. If, in the unlikely event Badanina, E.V., Vekslerb, I.V., Thomas, R. Syritsoc, L.F. that the Brazilian sources were to become and Trumbull, R.B. (2004) Magmatic evolution of restricted, there are adequate resources elsewhere Li–F, rare-metal granites: a case study of melt inclu- that would be brought on stream. sions in the Khangilay complex, Eastern Transbaikalia A major end use of niobium is in steels and other (Russia). Chemical Geology 210, 113–133. Bastos Neto, A.C., Pereira, V.P., Ronci, L.H., de Lima, alloys essential to the development of infrastructure, E.F. and Frantz, J.C. (2009) The world-class Sn, Nb, Ta i.e. railways, pipelines and automobiles. Niobium is F (Y,REE,Li) deposit and the massive cryolite associ- therefore of importance to the BRIC countries and ated with the albite-enriched facies of the Madeira they can be expected to be principal consumers as A-type granite, Pitinga mining district, Amazonas growth occurs. Elsewhere, niobium can be expected State, Brazil. Canadian Mineralogist 47, 1329–1357. to be consumed in developed countries where Breiter, K., Škoda, R. and Uher, P. (2007) Nb-Ta-Ti-W- ‘ageing’ infrastructure has become a concern. Sn-oxide minerals as indicators of a peraluminous P- The role of niobium and its utility in electronics and F-rich granitic system evolution: Podlesí, Czech is still in the early stages of development. It now Republic. Mineralogy and Petrology 91, 225–248. finds its way into computer chips and supercon- Burt, R. (2010) Tantalum – a rare metal in abundance? T.I.C. Bulletin 141, March 2010, Tantalum-Niobium ductors for medical technologies. Research is not International Study Center, Brussels. for a replacement metal, but for expansion, and Burt, R. (2011) Trends in the tantalum industry – a ret- novel functions, of the metal. rospective. Keynote address at ZrTa 2011 conference. Tantalum demand, after a rapid increase in the South Africa. late 1990s, has slowed to about a 1 per cent Companhia Brasileira de Metalurgia e Mineração compound growth rate throughout the last decade (CBMM) (2011). Mines and operations – production. (ignoring the 2008/9 dip). As noted earlier, demand http://www.cbmm.com/ in the historically largest market sector – capaci- Cˇ erný, P. (2005) The Tanco rare-element pegmatite tors – has declined, while in other sectors it has deposit, Manitoba: Regional context, internal increased. It will remain a critical component in anatomy, and global comparisons. In: Linnen, R.L. and Samson, I.M. (eds.) Rare element geochemistry electronic end uses, and there is no reason to pre- and mineral deposits. Geological Association of sume the ‘new’ uses will decline. However, in the Canada Short Course Notes, 17, 127–158. short to medium term, the outlook for supply is Cˇ erný, P. and Ercit, T.S. (2005) The classification of gra- somewhat unclear. The tantalum industry, as nitic pegmatites revisited. Canadian Mineralogist 43, with most sectors of the mining industry, has a 2005–2026. habit of relatively quickly ensuring supply meets Cˇ erný, P., Ercit, T.S. and Vanstone, P.T. (1996) Petrology (and often exceeds) demand. The large tonnage and mineralisation of the Tanco rare-element pegma- low-grade deposits, identified globally, have yet to tite, southeastern Manitoba. Geological Association play a role in the tantalum supply chain, and these of Canada Fieldtrip Guidebook. ˇ may await demand for their potential co-products, Cerný, P., Goad, B.E., Hawthorne, F.C. and Chapman, R. or new beneficiation techniques, to economically (1986) Fractionation trends of the Nb- and Ta-bearing oxide minerals in the Greer Lake pegmatitic granite capture ultrafine tantalum mineralisation. and its pegmatite aureole, southeastern Manitoba. American Mineralogist 71, 501–517. Note Chong, A. and Postolski, T. (2011) Blue River Ta-Nb Project NI 43–101 Technical Report. 1. Ma, million years before present Cordeiro, P.F.O., Brod, J.A., Dantas, E.L. and Barbosa, E.S.R. (2010) Mineral chemistry, isotope geochem- istry and petrogenesis of niobium-rich rocks from the References Catalão I carbonatite-phoscorite complex, Central Brazil. Lithos 118, 223–237. Anglo American (2013) Fact Book 2012/13. http://media. DNPM. (2011) Sumário Mineral Brasiliero 2011. Tântalo. investis.com/A/Anglo-American-plc/2013/ https://sistemas.dnpm.gov.br/publicacao/mostra_ AA-Factbook-2012-2013.pdf imagem.asp?IDBancoArquivoArquivo=6394 Tantalum and niobium 383

European Chemical Agency, ECHA. (2011) REACH: Nasraoui, M. and Waerenborgh, J.C. (2001) Fe specia- Regulations. http://echa.europa.eu/web/guest/regu- tion in weathered pyrochlore-group minerals from lations/reach. the Lueshe and Araxá (Barreiro) carbonatites by 57Fe Electronic Industry Citizenship Coalition, EICC. (2011) Mössbauer spectroscopy. Canadian Mineralogist 39, www.eicc.info/ 1073–1080. Fetherston, J. M. (2004) Tantalum resources in Western Organisation for Economic Co-operation and Devel- Australia. Geological Survey of Western Australia, opment, OECD. (2011) OECD Due Diligence Mineral Resources Bulletin 22. Guidance for Responsible Supply Chains of Minerals Graupner, T., Melcher, F., Gäbler, H.-E., Sitnikova, M., from Conflict-Affected and High-Risk Areas, Brätz, H. and Bahr, A. (2010) Rare earth element geo- OECD Publishing. http://dx.doi.org/10.1787/9789 chemistry of columbite-group minerals: LA-ICP-MS 264111110-en data. Mineralogical Magazine 74, 691–713. Palme, H. and O’Neill, H.St.C. (2004) Cosmochemical International Maritime Organisation, IMO. (2011) estimates of mantle composition. In: Carlson, R.W., www.imo.org Holland, H.D. and Turekian, K.K. (eds.) Treatise on Kravchenko, S.M. and Pokrovsky, B.G. (1995) The Tomtor Geochemistry 2, 1–38. alkaline ultrabasic massif and related REE-Nb deposits, Partington, G.A., McNaughton, N.J. and Williams, I.S. northern Siberia. Economic Geology 90, 676–689. (1995) A review of the geology, mineralization, and Küster, D. (2009) Granitoid-hosted Ta mineralization in geochronology of the Greenbushes pegmatite, Western the Arabian-Nubian Shield: Ore deposit types, tec- Australia. Economic Geology 90, 616–635. tono-metallogenetic setting and petrogenetic frame- Pell, J. (1996) Mineral deposits associated with carbon- work. Ore Geology Reviews 35, 68–86. atites and related alkaline igneous rocks. In: Mitchell, Küster, D., Romer, R.L., Tolessa, D., Bheemalingeswara, R.H. (ed.) Mineralogical Association of Canada Short D.Z.K., Melcher, F. and Oberthür, T. (2009) The Course Volume 24, 271–312. Kenticha rare-element pegmatite, Ethiopia: internal Philippo, S., Naud, J. and Verkaeren, J. (1997) Geochemical differentiation, U–Pb age and Ta mineralization. evaluation of the Lueshe niobium deposit (Zaire) by Mineralium Deposita 44, (7), 723–750. Rietveld quantitative X-ray diffraction. Applied Lagache, M. and Quéméneur, J. (1997) The Volta Grande Geochemistry 12, 175–180. pegmatites, Minas Gerais, Brazil: An example of rare- Pohl, W. (1994) Metallogeny of the northeastern Kibara element granitic pegmatites exceptionally enriched belt, central Africa – recent perspectives. Ore Geology in lithium and rubidium. Canadian Mineralogist 35, Reviews 9, 105–130. 153–165. Puchner, C.C. (1986) Geology, alteration and minerali- Linnen, R.L. and Cuney, M. (2005) Granite-related rare- zation of the Kougarok Sn deposit, Seward Peninsula, element deposits and experimental constraints on Alaska. Economic Geology 81, 1775–1794. Ta-Nb-W-Sn-Zr-Hf mineralization. In: Linnen, R.L. Quest Rare Minerals Ltd. (2011) www.questrareminer and Samson, I.M. (eds.) Rare element geochemistry als.com. and mineral deposits. Geological Association of Raimbault, L., Cuney, M., Azencott, C., Duthou, J.L. Canada Short Course Notes, 17, 45–68. and Joron, J.L. (1995) Geochemical evidence for a Melcher, F., Sitnikova, M.A., Graupner, T. et al. (2008) multistage magmatic genesis of Ta-Sn-Li mineraliza- Fingerprinting of conflict minerals: columbite- tion in the granite at Beauvoir, French Massif Central. tantalite (“coltan”) ores. SGA News 23, 1–14. Economic Geology 90, 548–576. Mining Journal (2010) Tantalum Supplement. Ram Resources Inc. (2011) www.ramresources.com.au. Mitchell, R.H. (2005) Mineralogical and experimental con- Resende, I. (2011) AMG Mining Mibra Mine Update. straints on the origins of niobium mineralization in car- Presentation at 52nd Annual Assembly of the Tantalum- bonatites. In: Linnen, R.L. and Samson, I.M. (eds.) Rare niobium International Study Center, Almaty, Kazakhstan, element geochemistry and mineral deposits. Geological October 2011. Association of Canada Short Course Notes, 17, 201–215. Rub, A.K., Štemprok, M. and Rub, M.G. (1998) Mroz, J.P. (1983) Inventory of minerals in the extreme Tantalum mineralization in the apical part of the north of Mozambique IN Library report #1205 Cínovec (Zinnwald) granite stock. Mineralogy and (extended extracts in Hunting 1988). Petrology 63, 199–222. Naidoo, B. (2011) http://www.miningweekly.com/ Rudnick, R.L. and Gao, S. (2004) Composition of the print-version/ferro-niobium-price-stability-benefits- continental crust. In: Rucnick, R.L., Holland, H.D. and malawi-project-says-globe-2011-06-07. Turekian, K.K. (eds.) Treatise on Geochemistry 3, 1–64. 384 robert linnen, dave trueman and ricard burt

Salvi, S. and Williams-Jones, A.E. (2005) Alkaline gran- Community. In: Boissonnas, J. and Omenetto, P. (eds.) ite-syenite deposits. In: Linnen, R.L. and Samson, Springer-Verlag, Berlin. 230–246. I.M. (eds.) Rare element geochemistry and mineral U.S. Geological Survey, USGS. (2011a) Mineral Commodity deposits. Geological Association of Canada, Short Summaries: Niobium. http://minerals.usgs.gov/min- Course Notes, 17, 315–341. erals/pubs/commodity/niobium/mcs-2011-niobi.pdf Salles, J.J. (2000) Production of niobium and tantalum U.S. Geological Survey, USGS. (2011b) Mineral from the Pitinga hard rock tin mine. T.I.C. Bulletin Commodity Summaries: Tantalum. http://minerals. No. 101, March 2000. usgs.gov/minerals/pubs/commodity/niobium/ Schwela, U. (2011) Tantalum-Niobium International Study mcs-2011-tanta.pdf Center (T.I.C.), Fifty-second General Assembly, Almaty, Van Lichtervelde, M., Linnen, R.L., Salvi, S. and Didier, 16–19 October 2011, http://tanb.org/webfm_send/166. B. (2007) Textures and chemical evolutions in tan- Shaw, R.A. and Goodenough, K.M. (2011) Niobium- talum oxides: a discussion of magmatic versus meta- tantalum. Mineral Profile. British Geological Survey. somatic origins for Ta mineralization in the Tanco http://www.bgs.ac.uk/mineralsuk/statistics/ Lower Pegmatite, Manitoba, Canada. Economic mineralProfiles.html Geology 102, 257–276. Sons of Gwalia (2002) Sons of Gwalia Annual Report, Winter, M. (2011) WebElements: the periodic table on 28–35. the web: Home of the periodic table. WebElements, Stearns, J., Mahtani, D., Diallo, M.K., Danssaert, P. and The University of Sheffield and WebElements Ltd, Finardi, S. (2008) Final report of the Group of Experts UK. www.webelements.com. on the Democratic Republic of Congo S/2008/773. Winter, J.D. (2010) Principles of Igneous and United Nations Security Council. Metamorphic Petrology. Prentice Hall, Upper Saddle Sweetapple, M.T. and Collins, P.L.F. (2002) Genetic River, New Jersey, USA. Framework for the Classification and Distribution of Wood, S.A. (2005) The aqueous geochemistry of zirco- Archean Rare Metal Pegmatites in the North Pilbara nium, hafnium, niobium and tantalum. In: Linnen, Craton, Western Australia. Economic Geology 97, R.L. and Samson, I.M. (eds.) Rare element geochem- 873–895. istry and mineral deposits. Geological Association of Tantalum-Niobium International Study Center (T.I.C.). Canada Short Course Notes, 17, 217–268. (2011) www.tanb.org Yin, L., Pollard, P.J., Shouxi, H. and Taylor, R.G. (1995) Tukiainen, T. (1988) Niobium-tantalum mineralisation Geologic and geochemical characteristics of the in the Motzfeldt Centre of the Igaliko Complex, South Yichun Ta-Nb-Li deposit, Jiangxi Province, South Greenland. Mineral Deposits within the European China. Economic Geology 90, 577–585. 16. Tungsten

TERESA BROWN AND PETER PITFIELD

British Geological Survey, Keyworth, Nottingham, UK

Introduction highest of all elements behind carbon. Of all pure metals, tungsten has the lowest coefficient Tungsten, also known as wolfram, was dis covered of expansion and the highest tensile strength in 1781 when Carl Wilhelm Scheele produced tung- at temperatures over 1650 °C (Christie and stic acid from scheelite. In 1783 brothers Juan José Brathwaite, 1996). Tungsten is also known for and Fausto de Elhuyar discovered that the same its high density, which is similar to gold, and its acid could also be produced from the mineral wol- high thermal and electrical conductivities. It framite and subsequently they managed to isolate has excellent corrosion resistance, does not the metal by reducing this acid with charcoal. The react with air or water at room temperature name tungsten comes from the Nordic words ‘tung’ (although fresh surfaces will oxidise) and is and ‘sten’ meaning ‘heavy’ and ‘stone’. The name largely unaffected by most acids. Key properties ‘wolfram’ has older roots and is believed to have are summarised in Table 16.1. been used because the yield of tin during smelting was reduced if tungsten minerals were present (for more details see Schubert and Lassner, 2009). The main applications for tungsten are in ‘hard Distribution and abundance metals’, i.e. tungsten carbide and cemented car- in the Earth’s crust bides, used for cutting, drilling and wear-resistant parts or coatings. Tungsten is also used to add The average abundance of tungsten in the Earth’s hardness and strength to steel alloys, particularly continental crust is estimated to be 1.0 parts per where heat resistance is also required. million (ppm) (Rudnick and Gao, 2003). The upper crust contains about 1.9 ppm tungsten, whereas the middle and lower crust is estimated to con- Physical and chemical properties tain 0.6 ppm. The abundance of tungsten in the oceans is estimated to be 0.1 parts per billion Tungsten is a hard, very dense, steel-grey to (ppb). The average concentration in workable ores greyish-white metal. It has the highest melting is usually between 0.1 and 1.0 per cent tungsten point of all non-alloyed metals and the second trioxide (WO3).

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 386 teresa brown and peter pitfield

Mineralogy exploration and mining. The colour of the fluorescent light is influenced by the molyb- Tungsten does not occur in nature as a free date content and changes from blue to cream metal, but only in the form of chemical com- and then to pale yellow and orange with pounds with other elements. Although several increasing molybdenum content (Lassner and tungsten-bearing minerals are known, most are rare Schubert, 1999). or very rare. Only scheelite and the wolframite Wolframite is a general term for iron–manganese group are abundant enough to be considered ores tungstate and is a solid-solution series between

(Table 16.2). two end members: ferberite (FeWO4, with less

Scheelite, a calcium tungstate (CaWO4), is than 20 per cent manganese) and hübnerite typically white to yellowish in colour, and (MnWO4, with less than 20 per cent iron). In has blue-white fluorescence in ultraviolet practice, the name wolframite is often used for light; a property which is especially utilised in the intermediate mineral between these two end members. The wolframite group exhibit typically Table 16.1 Selected properties of tungsten. tabular morphology, and are usually black, dark grey or reddish-brown in colour (Lassner and Property Value Units Schubert, 1999). Symbol W Secondary tungsten minerals, such as

Atomic number 74 hydrotungstite (H2WO4 · H2O) or cerotungstite

Atomic weight 183.84 (CeW2O6(OH)3), can be produced by alteration Density at 25 °C 19254 kg ⁄ m3 processes or weathering and may cause problems Melting point 3422 °C during processing leading to reduced recovery of Boiling point 5555 °C tungsten (Schmidt, 2012). Hardness (Mohs scale) 7.5 Specific heat capacity at 25 °C 0.13 J ⁄ (g °C) Electrical conductivity 18.2 × 106 S ⁄ m Deposit types Coefficient of linear thermal 4.5 × 10−6 ⁄ °C expansion Tungsten deposits usually occur within, or near Tensile strength at 20 °C 1000 MPa to, orogenic belts resulting from subduction- Tensile strength at 1650 °C approx 100 MPa related plate tectonics. All major deposit types Thermal conductivity 174 W ⁄ (m °C) are associated with granitic intrusions or with

Table 16.2 Properties of the most common tungsten minerals.

Wolframite Group

Scheelite Ferberite Wolframite Hübnerite

Chemical formula CaWO4 FeWO4 (Fe,Mn)WO4 MnWO4

Tungsten trioxide content (WO3 %) 80.6 76.3 76.5 76.6 Specific gravity (g/cm3) 5.4–6.1 7.5 7.1–7.5 7.2–7.3 Colour Pale yellow to orange, green to dark Black Dark grey to black Red-brown to black brown, pinkish-tan, dark blue to black, white or colourless Lustre Vitreous or resinous Submetallic to Submetallic to Submetallic to metallic metallic adamantine Hardness (Mohs scale) 4.5–5.0 5.0 5.0–5.5 5.0 Crystal structure Tetragonal Monoclinic Monoclinic Monoclinic Tungsten 387 medium- to high-grade metamorphic rocks. The In general, vein and stockwork deposits tend locations of selected major tungsten mines and to be low grade. However, even with grades as deposits are shown in Figure 16.1. low as 0.1 per cent tungsten trioxide they can Werner et al. (1998) classified major tungsten still be exploited economically by bulk mining deposits into seven types: vein/stockwork, methods, as demonstrated at Mount Carbine skarn, disseminated, porphyry, stratabound, mine in Australia (De Roo, 1988), which oper- placer, and brine/evaporite. Four additional ated successfully at this low grade between types of relatively minor economic interest were 1973 and 1986. Other notable examples of vein/ also identified as: pegmatite, breccia, pipe, and stockwork deposits are found at Verkhne- hot-spring deposits. In reality the categorisation Kayrakty, Kazakhstan (Rubinstein and Barsky, of individual localities can be complicated by 2002); Xihuashan in the Nanling tungsten–tin multiple stages of formation creating deposits province, China (Elliott, 1992; Guiliani et al., that could be ascribed to more than one of these 1988; Wang et al., 2011); Bolsa Negra and types. Table 16.3 contains a summary of typical Chicote Grande, Bolivia (Cox and Bagby, 1986) sizes and grades for the major producing deposit and Hemerdon, United Kingdom (Mining types. Magazine, 1979; Keats, 1981). The tungsten mineralisation at Hemerdon in Devon, is hosted in sheeted greisens-bordered vein systems and Vein and stockwork deposits stockworks in the apex of a steeply dipping, Vein and stockwork deposits are genetically dyke-like, granite body (known as Hemerdon related to the development of fractures that occur Ball) on the south-west side of the early in or near granitic intrusions during emplace- Permian-age Dartmoor pluton. Mineralisation ment and crystallisation. These fissures are fre- extends to depths of at least 400 metres below quently filled with quartz and can be up to several surface. metres in width. Large vein deposits may contain The Ta’ergou tungsten deposit, located in several individual veins, while stockworks com- the western part of the Qilian orogen in Gansu prise swarms of parallel, or near parallel, veins Province, north-western China, consists of with interconnecting veinlets (Werner et al., scheelite skarn bodies and wolframite–quartz 1998). veins. The deposit is genetically related to the The veining is commonly bordered by greisen Caledonian Teniutan granodiorite emplaced (a form of endoskarn alteration) and is often spa- within Proterozoic rocks bordering the tially associated with disseminated greisen and Archaean North Chine Platform. The tungsten porphyry-style tungsten mineralisation. veining overprints the calcic skarns (Zhang The mineralogy of vein deposits range from et al., 2003). the simple, consisting almost entirely of quartz The tungsten vein system in the Nanling and wolframite, to the complex, as at Pasto Range, South China, has been described using a Bueno in Peru (Landis and Rye, 1974) or so-called “five-oor building” model (Gu, 1982; Panasqueira in Portugal (Kelly and Rye, 1979) Liu and Ma, 1993; Li et al., 2011). Within this where more than 50 vein-forming minerals have model there are broadly five vertical zones: a been identified. The wolframite series is the thread or stringer zone at the top, over a veinlet main tungsten-bearing mineral but scheelite also zone, thin vein or mixed zone, large vein zone occurs in some deposits of this type. Tin, copper, and finally a thin-out or extinction zone molybdenum, bismuth and gold may also be pre- (Figure 16.2). Large veins can also be sparsely sent in economic quantities. In addition, developed in the underlying granite where typi- uranium, thorium, rare earth elements and phos- cally they are weakly mineralised. The large-vein phate minerals may also occur (Elliott et al., zone and the thin-vein or mixed-vein zones are 1995). the most economically valuable. Figure 16.1 The location of selected major tungsten mines and deposits. Note: In China there are many other deposits noted in the literature and it is not possible at this scale to include them all. Tungsten 389

Table 16.3 Typical size and grade of major producing tungsten deposit types.

Deposit size range Typical grade

Deposit type (metric tonnes) (WO3 %) Examples

Vein/stockwork <105 to 108 0.1 to 0.8 Panasqueira (Portugal), Pasto Bueno (Peru), Hemerdon (UK) Skarn <104 to 107 0.1 to 1.5 Mactung and Cantung (Canada), Xintianling and Yaogangxian (China) Disseminated <107 to 108 0.1 to 0.5 Akchatau (Kazakhstan), Krásno (Czech Republic) Porphyry <107 to 108 0.08 to 0.4 Xingluokeng and Yangchulin (China), Northern Dancer (Canada) Stratabound <106 to 107 0.2 to 1.0 Mittersill-Ferbertal (Austria), Damingshan (China)

Morphological Zones Average Density Mineral association zones thickness vein width (veins (metres) (metres) per metre) I Stringer or 100–200 0.001– 0.05–5 muscovite, cassiterite, wolframite, thread zone 0.01 tourmaline

II Veinlet zone 50–150 0.02–0.1 5–20 cassiterite, wolframite, chalcopyrite, bismuthine, beryl, muscovite, pyrite III Thin vein or 150–250 0.05–0.5 1–8 wolframite, cassiterite, scheelite, mixed zone chalcopyrite, pyrite, molybdenite, bismuthine, beryl, galena, sphalerite

IV Large vein zone 200–450 0.2–2 0.03–2 wolframite, chalcopyrite, molybdenite, potash, feldspar, pyrite, galena

V Extinction or >50 0.05–2 0.05–0.1 wolframite, molybdenite, chalcopyrite, thin-out zone pyrite, potash, feldspar

Figure 16.2 Schematic model of ‘five-oor building’ vein-type tungsten deposit. (Data sourced from: Gu, 1982; Huang and Xiao, 1986.)

Skarn deposits aureoles. They can also occur along faults, major Skarns are coarse-grained rocks dominated by shear zones, in shallow geothermal systems calc-silicate minerals that have formed by meta- and in metamorphic rocks at lower crustal depths somatic processes in sequences containing car- (Meinert et al., 2005). Less common types of bonate-bearing rocks such as limestone (Einaudi skarns form in contact with sulfidic or carbona- et al., 1981). Most are found adjacent to plutons ceous rocks such as banded iron formations and are often associated with hornfels, skarnoid, or black shales. Calcic skarns are characterised marble and other similar rocks in the thermal by calcium- and iron-rich silicates (andradite, 390 teresa brown and peter pitfield hedenbergite, wollastonite); magnesian skarns by several tungsten-bearing skarn deposits in China. calcium- and magnesium-rich silicates (forster- It is spatially associated with two-stage granite ite, diopside, serpentine); and aluminous skarns emplacement into Carboniferous dolomitic lime- by aluminium- and magnesium-rich calc-silicates stones over a 10 Ma interval in the mid–late Jurassic. (grossularite, vesuvianite, epidote). Dolomitic rocks Skarn ore formation is related to the older stage tend to inhibit the development of tungsten- granite. It is a typical example of the Jurassic bearing skarns; consequently magnesian tung- tungsten–tin ore-forming event in the Nanling sten-bearing skarns are uncommon (Ray, 1995). Range of south China (Zhang et al., 2011). Newberry and Einaudi (1981) distinguished two types of skarn on the basis of host-rock composi- Disseminated or greisen deposits tion and relative depth: reduced skarns such as the In disseminated or greisen deposits wolframite or Cantung and Mactung deposits in the North West scheelite are disseminated in highly altered (grei- Territories, Canada, and oxidised skarns, such as senised) granite or granitic pegmatite. Greisen King Island, Tasmania, Australia (Kwak and Tan, comprises mainly quartz and mica and is formed 1981; Kwak, 1987). In general, oxidised tungsten- by post-magmatic metasomatic replacement of the bearing skarns are smaller than reduced tungsten- primary granite minerals. Disseminated deposits bearing skarns. The highest grades in both systems are distinguished from the greisen-bordered veins are associated with hydrous minerals and retro- and stockworks by the pervasive nature of the grade alteration (Meinert et al., 2005). alteration and the absence of uid pathways. In Scheelite is the principal tungsten mineral and reality, these deposit types commonly coexist. this may occur as disseminated grains or fracture Disseminated greisen deposits usually occur fillings. Copper, molybdenum, tin, zinc and near to the upper parts of intrusions that are bismuth may also be present and can be econom- emplaced at depths of between 0.5 and 5 km, ically recoverable. Economically exploitable skarn where uids can boil but are prevented from deposits usually contain between 0.1 per cent and escaping to the surface. Tungsten is usually 1.5 per cent tungsten trioxide (Werner et al., 1998). present as wolframite although some deposits Mactung in the Yukon Territory, Canada, is the also contain scheelite. Tin, molybdenum, bismuth one of the largest tungsten-bearing skarn deposits, and other base metals may also be present, along with a NI43-101 compliant resource estimate with quartz, topaz, white mica, tourmaline and including 33 million tonnes @ 0.88% WO in the 3 uorite. Tungsten grades are generally low but indicated category and a further 11.9 million exploitation can be economic as a by-product of tonnes @ 0.78% WO in the inferred resource 3 tin extraction. Examples of disseminated deposits category (Narciso et al., 2009). The deposits occur include the Akchatau and Kara Oba deposits in in the thermal aureole of a late Cretaceous felsic central Kazakhstan (Zaraisky and Dubinina, intrusion, which was emplaced into a dominantly 2001), the Torrington district of New South pelitic, Lower Paleozoic sequence along the eastern Wales, Australia and the Hub stock at Krásno in margin of the Selwyn basin. The mineralisation is the Czech Republic (Jarchovsky, 2006). stratabound and confined to four individual beds (Dick and Hodgson, 1982). Porphyry deposits Other tungsten-bearing skarn deposits include Los Santos in Spain (Tornos et al., 2007), Porphyry deposits are extensive, low-grade deposits Tyrnyauz and Vostock-2 in Russia (Soloviev formed following the separation of metal-rich and Krivoschchekov, 2011) and Xintianling and uids from a crystallising wet magma. Tungsten Yaogangxian in China (Zhao et al., 1990; Chang, tends to be concentrated in stockwork zones and 2005). The Xintianling scheelite deposit, located fractures either in or near to the upper parts of gra- on the north-east side of the Qitianling batholith nitic intrusions emplaced at shallow depths. in Hunan Province, is one of the largest amongst Mineralised breccia zones may also be present. Tungsten 391

Tungsten occurs either as wolframite or schee- W–Sn–Mo–Bi–Be mineralisation associated with lite, and sometimes both are present. Molybdenum, a later granite phase which is superimposed on bismuth and tin often occur and may represent an the early stage massive skarn-greisen zone. On opportunity for co-production. Tungsten-bearing this basis some researchers have described this porphyry deposits tend to be large in size but may deposit as porphyry in style (Li et al., 2004). not be economic due to their low grade. Never- theless, important examples include Northern Breccia deposits Dancer (formerly known as Logtung), Sisson Brook Breccia deposits are composed of angular, broken and Mount Pleasant in Canada (Brand, 2008; fragments of rock located within, above or marginal Snow and Coker, 1986; Kooiman et al., 1986) and to the apex of an intrusion. They are formed either the Xingluokeng, Shizhuyan, Lianhuashan and by magmatic/hydrothermal hydraulic fracturing Yangchulin deposits in China (Liu, 1980; Zhaolin or by explosive interactions between water and Zhongfang, 1996; Werner et al., 1998). and magma. Many vein/stockwork and porphyry The Northern Dancer porphyry of the Western deposits have breccia zones associated with them. Cordillera, which straddles the boundary between However, some tungsten-bearing breccia bodies Yukon Territory and British Columbia Province, appear to have formed inde pendently of other comprises multiple, mid-Cretaceous felsic intru- deposit types (Werner et al., 1998). An example is sions hosting four vein systems with different the Washington copper–molybdenum–tungsten tungsten/molybdenum ratios (Noble et al., 1984). breccia pipe in Sonora, Mexico (Simmons and Mineralisation is centred on a felsic porphyry Sawkins, 1983). dyke complex and includes stockworks, sheeted vein systems, disseminations and skarns but Stratabound deposits mostly comprises typical porphyry-style crackle breccias showing many similarities with porphyry Tungsten mineralisation in stratabound deposits molybdenum deposits. Resources, which are is confined to a single stratigraphic unit, although NI43-101 compliant, are estimated to include 30.8 they may not be strictly conformable to bedding, million tonnes @ 0.114% WO3 in the measured i.e. mineralisation may cross bedding planes. category, 192.6 million tonnes @ 0.1% WO3 in the Stratabound deposits occur in volcano-sedimen- indicated category and a further 201.2 million tary sequences and are considered to be syngenetic tonnes @ 0.089% WO3 in the inferred resource in origin. They can be distinguished from skarn category (Molavi et al., 2011). deposits which are largely controlled by the com- The Xingluokeng tungsten–molybdenum deposit position of the host rock lithology and are assumed of the Fujian province is hosted in a late Jurassic to be epigenetic. Stratabound tungsten mineralisa- Yanshanian granite porphyry stock in the Wuyishan tion occurs within iron–magnesite and dolomitic metallogenic belt. The central zone underwent marbles in the Eastern Alps (Neinavale et al., strong silicic and potassic alteration and is enriched 1989). Many stratabound tungsten occurrences in rare earth elements (Zhang et al., 2008). appear to have been affected by later mobilisation The Shizhuyuan deposit in the Dongpo ore- and reconcentration and therefore their syngenetic field of Hunan Province is a world-class poly- origin is questionable (Werner et al., 1998). metallic tungsten deposit. The mineralisation is Examples of this type of tungsten deposit diverse in character and has a complex origin include the Mittersill-Ferbertal deposits in the related to multiple phases of granite intrusion Salzburg province of Austria and Damingshan in (Lu et al., 2003). It comprises dominantly W–Mo– China (Ma, 1982). The Cambrian-age Mittersill- Bi–Sn–F calcic skarn-greisen zones developed Ferbertal orefield comprises several lenses of around the late Jurassic Qianlishan granite com- scheelite-rich quartzite, an underlying vein-stock- plex of the Yanshanian granitoid province. The work zone, an eruption breccia and quartz-rich highest tungsten grades occur in vein/stockwork aureole to a granitoid intrusion. Further scheelite 392 teresa brown and peter pitfield enrichments occur along shear zones. Geochro- been worked in the past, include Golconda in nological and geochemical data on the Ferbertal Nevada, USA (Kerr, 1940; Marsh and Erickson, deposit indicate a genetic link with mantle- 1975) and Uncia in Bolivia (Werner et al., 1998). dominated granitic melts (Eichhorn et al., 1999). Subsequent metamorphic events and granitic Placer deposits intrusions have remobilised the scheelite (Höll Placer deposits are concentrations of heavy and and Eichhorn, 2000). The Cambrian-Ordovician chemically resistant minerals that occur in Damingshan tungsten deposit in the Danchi sediments. Wolframite and scheelite, although metallogenic belt in Guangxi Province includes heavy, will eventually decompose during weath- Late Cretaceous vein and stockwork wolframite ering and therefore tend not to be preserved mineralisation as well as the massive stratiform long enough to form widespread placer deposits. types (Li et al., 2008). However, they do occur, in both alluvial and coastal sediments, albeit they are usually small Pegmatite deposits in size. Typically, they are located very close to Pegmatites are coarse-grained igneous rocks, gen- the bedrock deposit from which they were origi- erally of granitic composition, formed by the late- nally derived (Werner et al., 1998). A few tungsten- stage crystallisation of magma and containing bearing placer deposits have been worked on an many incompatible elements such as lithium, industrial scale, for example in the Heinze Basin beryllium, niobium, tantalum, tin and uranium. in Burma (Myanmar) (Goosens, 1978) and in the Tungsten is not a common constituent of pegma- Dzhida district of eastern Siberia. tites and tungsten-bearing pegmatite deposits are therefore rare. Grades for tungsten tend to be low Brine and evaporite deposits but it can be extracted as a by-product. For Tungsten-bearing brines and evaporite deposits example, tungsten occurs in the pegmatite which occur in recent lakes and/or palaeolake settings in is worked primarily for tantalum at the Wodgina arid regions of Asia and North America. The tung- mine in Western Australia, and it is also found in sten is thought to have been leached from bed- the Okbang deposit in South Korea (Chung, 1975). rock deposits by hot uids. The most significant example of this type of deposit is the Searles Pipe deposits Lake deposit in California, USA (Guerenko and Pipe deposits can be cylindrical or irregular, elon- Schmincke, 2002; Altringer, 1985). Stratabound gated or bulbous masses or quartz that occur at the scheelite at Halls Creek in Western Australia is of margins of granitic intrusions. Mineralisation, evaporitic origin analogous to continental-sabkha most frequently wolframite, is often erratically dis- playa basins of the Mojave desert (Todd, 1989). tributed in high-grade shoots or pockets containing up to 20 per cent wolframite, but deposits tend to be small. Examples include the Wolfram Camp Extraction methods, processing deposit in Queensland, Australia (Plimer, 1975). and beneficiation

Hot-spring deposits Extraction These deposits are probably derived from bed- Tungsten mining techniques are similar to many rock tungsten-bearing deposits and are formed other metals of similar occurrence. Most tung- by circulating hot ground water. Deposits of sten is mined from sub-surface (or underground) calcareous tuffs or travertine are formed by pre- mines. A few tungsten mines have used surface cipitation as this hot ground water cools and tung- (open-pit) methods, but in most cases, these sten mineralisation has been found in selected mines were later converted to underground mines locations. Examples, where these deposits have to access deeper ores. Tungsten 393

Surface mining roughly horizontal levels at various depths below the surface. The particular method used for a mine Open-pit mining is generally cheaper and safer will depend on the size, shape and grade of the ore to operate than underground mining and is body, its depth below the surface and the compe- normally used if a near-surface ore body occurs. tence of the ore zone and wall rocks (Shedd, 2011). Open-pit mining can reach depths of several hundred metres but rarely exceeds 100 metres in tungsten mine operations before going under- Processing ground. Some placer deposits are amenable to The first phase of processing the ore is beneficia- strip mining or dredging operations. tion at the mine site to increase the tungsten content. The resulting concentrate containing Underground mining more than 65 per cent tungsten trioxide can Underground mining is preferred when surface either be used directly for production of ferro- mining is, or becomes, prohibitively expensive. A tungsten and steel manufacture, converted to a major factor in the decision to mine underground number of intermediate tungsten compounds by is the ratio of waste to ore (strip ratio). Once this hydrometallurgical processes or further refined ratio becomes large, surface mining is no longer to pure tungsten using pyrometallurgical tech- economic. Ore is mined in stopes on a number of niques (Figure 16.3).

Ore from mine Crushing & milling

Pre-concentration stage (photometric or UV fluorescence and air blast separation)

Ore beneficiation (gravity, froth flotation, magnetic and/or electrostatic separation

Direct use in Tungsten concentrate Scrap and residues steel manufacture (65–75% WO3) (40–95% WO3)

Hydrometallurgy (pressure leaching, filtration, solvent extraction and precipitation)

Various Intermediate tungsten uses compounds, e.g. APT

Pyrometallurgy

Figure 16.3 Simplified ow diagram illustrating the generic steps in processing Various Tungsten metal, trioxide, carbide or alloys tungsten. uses 394 teresa brown and peter pitfield

Ore benefication that can be readily purified and then reduced. In modern plants most tungsten concentrates are The tungsten ore is first crushed and milled to processed chemically to an ammonium paratung- liberate the tungsten mineral grains. The slurry state (APT) intermediate. containing tungsten minerals and waste rock is In the hydrometallurgical processes, scheelite then concentrated using several methods (e.g. is first digested by sodium carbonate pressure gravity, froth otation, magnetic and electrostatic leaching and wolframite by sodium hydroxide. separation) depending on the characteristics and Wolframite–scheelite mixtures can be successfully composition of the ore. Some plants incorporate a digested by a pressure leach with sodium hydroxide pre-concentration stage before other conven- under simultaneous mechanical activation. tional processing methods. This usually involves Tungsten scrap with only a few exceptions is an ore-sorting process whereby a continuous easier to convert to APT than ore concentrates stream of rock particles is subjected to a ‘sensing’ because it does not contain detrimental elements stage and are subsequently separated using an like phosphorus, arsenic and silicon. The tungsten air blast. The sensing stage can consist of a scrap is first oxidised to facilitate dissolution in photometric method if the dark tungsten ore the alkaline leach process. Dissolution is achieved minerals are contained within white quartz or a either by sodium hydroxide pressure digestion or UV-uorescence process if the ore is scheelite an oxidising alkaline melt or by electrolysis. (Lassner and Schubert, 1999). Following the dissolution process the undis- After that, scheelite ore can be concentrated solved digestion residues are removed by filtra- by gravity methods, often combined with froth tion. However, the sodium tungstate solution otation, whilst wolframite ore can be concen- after filtration may still contain contaminants trated by gravity, sometimes in combination with such as molybdenum. These can be removed by magnetic separation. Gravity methods usually precipitation under carefully controlled condi- involve the use of spirals, cones and/or vibrating tions or by ion exchange and adsorption. tables. Magnetic separation is sometimes also The sodium tungstate solution is converted into used to clean scheelite (i.e. to remove magnetic ammonium isopolytungstate by either solvent gangue minerals from the ore). Electrodynamic or extraction (liquid ion exchange using an organic electrostatic separators are only used to separate phase) or ion exchange by resins. The resultant scheelite–cassiterite mixtures (Lassner and ammonium isopolytungstate solution from Schubert, 1999). either of these processes is evaporated whereby ammonia and water are volatilised. The ammonia Direct use in steel manufacture concentration decreases relative to tungsten tri- Wolframite concentrates can be smelted directly oxide and paratungstate, as the ammonium salt has with charcoal or coke in an electric arc furnace a low solubility, and crystallises out. The degree of to produce ferrotungsten (FeW) which is used as evaporation depends on the purity of the feed solu- alloying material in steel production. Pure tion and the required purity of the APT, which may scheelite concentrate may also be added directly range from 90 to 99 per cent. The tungsten-bearing to molten steel. crystal slurry is filtered from the mother liquor, washed with deionised water and dried (Lassner and Schubert, 1999). Hydrometallurgy The production of tungsten metal from wol- Pyrometallurgy framite and scheelite concentrates and tungsten scrap (which can contain between 40 and 95 The calcination under oxidising condition con- per cent tungsten) requires the conversion of the verts APT to tungsten trioxide. This is usually feed material into an intermediate compound carried out in a rotary furnace at 500–700 °C. Tungsten 395

Calcination under slightly reducing conditions of alkali and other metallic impurities. It is results in tungsten blue oxide. The latter can be produced from APT in an electrolytic cell. carried out in either a push-type or rotary furnace Tungsten trioxide (WO3) or tungstic oxide, is a at temperatures between 400 °C and 900 °C. light yellow powder synthesised by oxidative The conversion of either oxide into tungsten calcination of ammonium paratungstate. It can metal powder requires further treatment in also be produced by calcination of tungstic acid. a furnace, almost exclusively under hydrogen Tungsten blue oxide (TBO) is a dark blue or black reducing conditions. This is carried out either in powder with a nominal chemical formula of WO2.97. push-type or rotary furnaces with carefully con- It is not a well-defined chemicalcompound but a trolled temperature zones varying between 600 °C mixture of constituents. TBO is formed by calcina- and 1100 °C. The hydrogen not only provides a tion of APT under slightly reducing conditions. reducing environment, it also acts to remove any Tungsten metal powder (W) is produced water vapour which forms during the process from yellow or blue tungsten oxide in hydrogen (Lassner and Schubert, 1999). reduction furnaces. Tungsten is sometimes used in Most uses of tungsten powder, including in powder form but commonly must be consolidated the production of tungsten carbide or the various into solid form. The metal powder is not smelted tungsten alloys, involves further heat treatment. directly because tungsten has the highest melting Unlike other metals, compaction and sintering point of any metal. The pure tungsten powder is processes are normally used, rather than smelting pressed into bars, and then sintered at high tem- technology, because tungsten has an exceptionally perature (1800–2500 °C). These are worked and high melting temperature. Compaction involves rolled at lower temperatures to increase density pressing tungsten powder in rigid dies or exible and ductility, usually under a hydrogen atmosphere moulds and these are then sintered at high temper- to prevent metal oxidation. atures to increase their strength and density, often Tungsten–heavy-metal alloy (WHAs) are a in a reducing environment created by the presence category of tungsten alloy that typically contain of hydrogen (Lassner and Schubert, 1999). 90–98 per cent by weight of tungsten in combi- nation with a mix of nickel, iron, copper, and/or cobalt. These metals serve as a binder to hold the tungsten particles in place after sintering. These Specifications and uses heavy metal alloys do not require millwork to enhance their properties and are machined much Specifications more easily than pure tungsten. The most common Tungsten is used and traded in a variety of forms are tungsten–nickel–iron alloys, although forms (Lassner and Schubert, 1999; International the addition of cobalt enhances both strength and Tungsten Industry Association (ITIA), 2012), the ductility. The tungsten–molybdenum-nickel–iron most important of which are: alloy has greater strength but reduced ductility.

Ammonium paratungstate (APT) is a white Tungsten carbide (WC) and semi-carbide (W2C) free-owing crystalline salt with the chemical are produced by the reaction of tungsten metal ⋅ formula (NH4)10(W12O41) 5H2O. It is the main powder with pure carbon powder at 900–2200 °C intermediate and also the main tungsten raw in a furnace, a process called carburisation. material traded in the market. APT is usually Other methods include a uid-bed process that

calcined to yellow trioxide (WO3) or blue oxide react either tungsten metal or TBO with a

(W20O58). carbon monoxide/dioxide mixture and hydrogen Ammonium metatungstate (AMT) with the at 900–1200 °C. By melting tungsten metal and ⋅ formula (NH4)6H2W12O40 xH2O, is a highly soluble WC together, a eutectic composition of WC and white crystalline powder and a source of high-purity, W2C is formed. This melt is cast and rapidly water-soluble tungsten, which is essentially free quenched to form extremely hard solid particles. 396 teresa brown and peter pitfield

The solids are crushed and classified to various 9% mesh sizes. Tunsgten carbide is, quantitatively, the most 8% important tungsten compound. In its most basic 58% form, it is a fine grey powder, but it can be pressed and formed into shapes for a multitude of uses. Tungsten carbide is approximately three times stiffer than steel and is much denser than steel or titanium. It has a high melting point (2870 °C), is extremely hard (8.5–9.0 on Mohs scale, comparable with corundum or sapphire) and can only be polished and finished with abrasives of superior hardness such as silicon carbide, cubic boron nitride and diamond. It also has a low electrical resistivity, 25% comparable with some metals (e.g. vanadium).

Ferrotungsten (FeW and Fe2W) is an alloy with a steel-grey appearance formed by combining iron and tungsten. There are two grades with tungsten contents in the range 75–82 per cent and 70–75 per cent. Ferrotungsten is a remarkably robust Hard metals Mill products alloy because of its high melting point. By Steel and Others combining iron with tungsten, the brittleness of other alloys metallic tungsten is much reduced. Non-ferrous tungsten alloys in current use Figure 16.4 The main global uses of tungsten. (Data include composites with copper, silver, nickel and from Wolf Minerals, 2012.) rare metals. The combination of tungsten (or tungsten carbide) with these metals produces relatively hard, heat-resistant materials with products for many applications. These properties superior wear resistance, robust physical prop- include: erties at elevated temperatures, and good electrical ● very high melting point; and thermal conductivity. The tungsten compo- ● very high density; nent imparts a high density, high vibration and ● extreme strength; damping capacity, excellent radiation shielding, ● high wear resistance; higher temperature strength and thermal shock ● high tensile strength; resistance and improved corrosion resistance. ● low coefficient of expansion; Superalloys containing tungsten are high- ● high thermal and electrical conductivity. performance alloys that exhibit excellent The most important uses are shown in Figure 16.4. mechanical strength and creep resistance at high temperatures, good surface stability, and corro- Hard metals sion and oxidation resistance. A superalloy’s base alloying element is usually nickel, cobalt, or Hard metals consist of tungsten carbide and nickel–iron. cemented carbides, which are formed from tung- sten carbides and cobalt, or occasionally other metals such as titanium, tantalum and niobium. Uses They are very hard materials that are used for The unique range of properties of tungsten makes cutting, drilling and wear-resistant parts or it an essential component in a wide range of coatings. Tungsten 397

Tungsten carbide (WC) is important in the gas turbine and marine turbine industries due to metal-working (40 per cent of the total WC their very high corrosion and wear resistance in market), mining (30 per cent), and petroleum a variety of challenging environments. Examples (20 per cent) industries. Its main applications are in of applications include ue-gas desulfurisation machine tools especially where they need to with- systems, chemical processing components such stand higher temperatures (e.g. punches, drill tips, as heat exchangers, industrial furnaces, erosion tile and glass cutters), in armour-piercing ammu- shields and jet-engine combustion chambers. nition, sports equipment and domestic items such The unique properties of conventional WHAs – as scratch-resistant jewellery and the rotating ball high density, high strength, high ductility, good in the tips of ballpoint pens. Fine- and ultra-fine- corrosion resistance, high radiation adsorption grained WC hard metals are becoming increas- capability, and reasonably high toughness – make ingly important because they last even longer than them well suited for a variety of defence and conventional WC and can be used for specialist civilian applications including X-ray and radia- applications where very fine tools are required, e.g. tion shields, counter weights, kinetic-energy the drilling of holes in electronic circuit boards. penetrators, vibration dampening devices, med- ical devices for radioactive isotope containment, heavy-duty electrical contact materials and gyro- Steel and other alloys scopes. Due to eco-toxicity concerns WHAs have Tungsten was one of the first elements to be been used as weights in fishing to replace lead. systematically alloyed with steel to improve its A tungsten alloy including a proportion of tin, properties. Special steel alloys, such as high-speed copper or nylon is also used as a substitute for steels, heat-resistant steel and tool steels, are lead shot in shotgun pellets. largely utilised in metal cutting and specialist Tungsten-copper (WCu) alloys are dense, high- engineering applications where hardness and performance, easily machinable materials with strength are required, particularly over a wide excellent thermal and electrical conductivity. temperature range. WCu alloys are widely used in the electrical con- High-speed steel (HSS) is a variety of tool steel tacts, high-voltages switches, electrodes, circuit in which tungsten is a common heat-resisting breakers, refractory parts, self-cooling heat sinks, alloying element. It is superior to the older high- etc. As electrodes they exhibit low wear and carbon steel in that it can withstand higher tem- maintain high contour sharpness. peratures without losing its hardness. This property Nickel–tungsten (WNi) alloys remain stable allows HSS to cut faster than high-carbon steel, indefinitely at room temperature and are highly which gave rise to the name. There are many differ- resistant to decomposition when heated. WNi ent types and designations of high-speed steel, each electroplating is harder and longer lasting than with its own combination of added metals. chrome. Furthermore, the electroplating process Superhigh-speed steels have very high wear is more efficient than that for chrome, because resistance, improved temperature resistance or multiple layers can be applied in one step, and it both. These steels are used for high cutting speeds, is potentially safer and less environmentally dry or semi-dry cutting or simply to maximise tool polluting to produce. This is a relatively new life. Because of the high hardness and wear resis- technology with potential application in coating tance some grades are considered to provide a shock absorbers, print rolls and connectors for bridge between HSS and hard metals (Tarney, portable electronics (Nanomaterials, 2009). 2004). Heat treatment (usually with a laser or elec- Several other tungsten alloys exist includ ing tron beam), surface finish and coatings can provide molybdenum–tungsten (MoW), tungsten– significantly enhanced HSS tool performance. rhenium (WRe), tantalum–tungsten (TaW) and Superalloys containing tungsten have niobium–tungsten (NbW). These have superior particular applications in the aerospace, industrial properties in terms of strength, corrosion resistance, 398 teresa brown and peter pitfield temperature stability, etc. and are used for circuits and circuit boards through the process of specialist applications in the glass, aerospace and chemical vapour deposition. The semiconductor electrical industries. industry consumes around 200 tonnes of WF6 per year worldwide. It is also used to deposit films of Mill products tungsten on ceramics or metal products such as bearings. The term ‘mill products’ is used in the industry to refer to tungsten wire, sheets or rods. The combination of extremely high melting point, Recycling, re-use and resource efficiency conductivity and ductility of pure metallic tung- sten makes it ideal for electrical and electronic As the price of tungsten has increased substan- applications, most notably in incandescent light- tially in the last decade there are economic and bulb filaments, discharge lighting electrodes and environmental imperatives to use tungsten more vacuum tubes, and heating elements. In medical efficiently, to seek substitutes for less-demanding X-ray tubes both the filament and target is usually applications and to recycle it more effectively. tungsten or tungsten alloy. Unalloyed tungsten is Recycling is an important element of the world’s also used in electronic circuit interconnectors, tungsten supply. It is estimated that more than filaments in vacuum-metallising furnaces, and for 30 per cent of total world supply is from recycled electrical contacts such as the distributor points sources (EC, 2010). In some countries this figure is in automotive ignition systems. Its high melting much higher, for example in the USA 46 per cent of point is the basis for the use of tungsten welding tungsten supply is from recycled material (Shedd, electrodes in tungsten inert (TIG) welding, where 2011). The tungsten processing industry is able to it is often doped (0.5–2% weight/weight) with treat almost every kind of tungsten-containing lanthanum, cerium, thorium, or zirconium. Water- scrap and waste to recover tungsten and, if present, cooled tungsten tips are used for non-consumable other valuable constituents. electrode vacuum-arc melting of alloys. Old scrap Chemical and other specialist applications Old scrap, also referred to as end-of-life scrap, consists of tungsten-bearing products that are no Many of the intermediate tungsten products have longer being used. These include: uses in their own right. For example, ammonium ● cemented carbide parts; paratungstate (APT) can be used as a colouring ● tungsten metal and tungsten alloy parts; agent in the porcelain industry or in catalysts, ● old superalloy scrap; phosphors and absorbent gels; ammonium meta- ● tungsten-bearing catalysts. tungstate (AMT) is used as a reagent for chemical It is estimated that on average 10–25 per cent analysis during medical diagnosis or as a corro- of old scrap is recycled globally (UNEP, 2011). sion inhibitor; and tungsten trioxide (WO ) can 3 However, this figure will vary greatly between be used as a catalyst or as a pigment in ceramics individual countries, for example in the USA it is and paints. as high as 66 per cent (Shedd, 2011). Other substances containing tungsten have a wide range of applications. For example, sodium New scrap tungstate is a widely used tungsten chemical with applications in the manufacture of organic This is generated during the processing of tung- dyes and pigments, catalysts, fireproofing of tex- sten concentrates, scrap, and chemicals to make tiles and hard surfacing for graphite crucibles. metal powder, carbide powder, or chemicals, and

Tungsten hexauoride (WF6) gas is most com- during the fabrication of tungsten products from monly used in the production of semiconductor these materials. Recovery rates of new scrap are Tungsten 399 high, and metal losses are kept to a minimum by where optimum performance is required at high internal recycling of the scrap generated. temperatures. Unlike many other metals, there Alloy industries that produce parts by cast or are no substitutes for tungsten that do not involve wrought processes typically generate large quan- a considerable cost increase and compromise in tities of new scrap. During the casting of corrosion- product performance. In the aerospace and resistant alloys 40 to 60 per cent of the melt defence industries where product performance is becomes new scrap (Shedd, 2011), which is either paramount, substitution is generally avoided recycled in-house or sold to be recycled elsewhere. (Pitfield and Brown, 2011). Potential substitutes for cemented tungsten Unrecovered scrap carbides or hard metals include cemented carbides based on molybdenum carbide and In various applications scrap tungsten is not titanium carbide, ceramics, ceramic–metallic recovered, either due to the lack of an efficient composites (cermets), diamond tools, and tool collection system, or because it is not economic steels (EC, 2010). Potential substitutes for other to recycle owing to contamination. Most uses of applications are: tungsten in chemical applications, except for cat- ● molybdenum can replace tungsten in certain alysts, are dissipative. Tungsten is also lost in use mill products; due to wear of cemented carbide parts and hard- ● molybdenum steels can substitute for tungsten faced products, arc erosion of electrical contacts steels for some applications; and electrodes, and the oxidation of alloys under ● lighting based on compact uorescent lamps, high-temperature conditions. low-energy halogen light bulbs and light-emitting diodes (LEDs) are gradually replacing the tradi- Recycling methods tional use of tungsten in light bulb filaments as The tungsten industry uses a variety of recycling inefficient incandescent light bulbs are being methods to optimise recovery and reduce costs. phased out; Metallurgical and hydrometallurgical techniques ● depleted uranium can be used in weights and have the ability to treat all scrap types and remove counter-weights instead of tungsten alloys or impurities. unalloyed tungsten, but generally it has been Contaminated cement carbide scrap, tungsten tungsten that has substituted for depleted grindings and powder scrap can be oxidised and uranium for health and environmental reasons; chemically processed to APT in a similar way to ● depleted uranium alloys can also be used in that used for the processing of tungsten ores. If armour-piercing projectiles instead of cemented present, cobalt, tantalum and niobium are also tungsten carbides or tungsten alloys. recovered separately. Other tungsten-containing Because of its perceived benign impacts on scrap might require a modified process (EC, 2010; human and environmental health, tungsten Roskill, 2010). alloys have been the preferred substitute for lead- Tungsten in high-speed steel is frequently based munitions since the mid-1930s. recycled, and a typical melt contains 60 to 70 per cent scrap, including internally generated scrap (EC, 2010; Roskill, 2010). Environmental aspects of the life cycle of the metal and its products

Substitution Tungsten is much less toxic than other metals, such as lead or mercury, for which it is some- Because of the unique combination of properties times used as a substitute, e.g. in ammunition. of tungsten, there are limited options for However, there is growing evidence that tungsten substitution in many applications, especially should not be considered as completely ‘non-toxic’ 400 teresa brown and peter pitfield and that it is not as environmentally inert as pre- World resources and production viously thought (Petkewich, 2009). Concern regarding the health impacts of tung- Resources and reserves sten in the environment has been expressed fol- lowing the identification of three clusters of World tungsten resources have been estimated leukaemia in the USA, which are associated with at seven million tonnes (contained tungsten elevated tungsten levels. It has not yet been metal) including deposits that have so far not proven that tungsten is the cause of these clusters been proven to be economically workable and research is continuing (Koutsospyros et al., (Hinde, 2008). It is believed that 30 per cent 2006; and several others as listed in Strigul, 2010). of the resources are wolframite (76.5 per cent Tungsten levels in water have been subject to reg- tungsten trioxide) and 70 per cent are scheelite ulation in Russia for some years, but similar reg- (80.5 per cent tungsten trioxide) ores (Hinde, ulations do not yet exist in the USA nor in the EU 2008). (Strigul, 2010). Werner et al., (1998) noted that the ten largest A recent study by Clauson and Korte (2009) known deposits at that time were located in found that particles of tungsten metal could oxi- Kazakhstan, Canada, China and Russia. However, dise in air and then dissolve into soil through the since then new deposits have been discovered, for action of rainwater. The tungsten, in this more example O’Callaghans in Australia, and previ- soluble form, became mobile in the soil and ele- ously identified occurrences have been found to vated levels were detected in local groundwater. contain more tungsten than previously thought, Elevated levels of tungsten have been found in for example Hemerdon in the United Kingdom. It soils, trees and plant tissue, aquatic systems, the is clear that China, Kazakhstan and Russia have atmosphere, animals and humans in various considerable resources but detailed informa tion locations (see Koutsospyros et al., 2006 for a is difficult to obtain. Resources in Canada are discussion of the locations and values reported). known to exceed one million tonnes of contained Often these appear to be associated with natural tungsten, whereas resources in Australia are tungsten deposits but also occur in the vicinity of believed to be nearly 0.5 million tonnes of anthropogenic activities such as the processing of contained tungsten. tungsten ore or metal, military firing ranges, the The USGS (Shedd, 2012a) estimated in January use of agrichemicals and the disposal of waste in 2012 that reserves stood at 3.1 million tonnes various forms. (contained tungsten metal) with more than 60 The dissolution of tungsten powder appears to per cent of these located in China (Figure 16.5). cause soil acidification with associated adverse The National Bureau of Statistics of China consequences on the micro-organisms, inver- reported that their reserves of tungsten in 2010 tebrates and plant communities (Strigul et al., were approximately 1.75 million tonnes of 2005). However, the geochemistry of tungsten is contained metal (NBSC, 2012) and it is believed complicated and its mobility, bioavailability and this is concentrated in the provinces of Hunan toxicity in the environment depends on the exact and Jiangxi (Pitfield et al., 2010). These two form of the tungsten compounds (Strigul et al., provinces also receive the highest proportion of 2009), the environmental conditions (e.g. compo- China’s production quota (Shedd, 2012b). sition of chemical substrates, pH, oxidation Deposits of tungsten are known to exist in many states) and the nature of the exposure pathway other Chinese provinces but it is not known (Bednar et al., 2008; Strigul et al., 2005). whether they are categorised formally as Many studies have been conducted in recent resources or reserves. years (Strigul, 2010) but the environmental and In Russia, reserves of tungsten are believed to health impacts of tungsten are not yet fully be mainly located in the North Caucases area understood. and the Far East region. Canadian reserves are Tungsten 401 dominated by the huge deposit at Mactung and Production the nearby operating mine at Cantung, which are In 2011, total world production was nearly 73,000 located in the Yukon and North West Territories, tonnes of tungsten (metal content of concen- respectively. trates). This was a six per cent increase compared to 2010 and 173 per cent higher than the low of 20% 1999 shown in Figure 16.6 (BGS, 2013). Tungsten is currently produced in approxi- 62% mately 20 countries. China has been the world’s leading tungsten producer for many years. In 1989 it accounted for 58 per cent of the world 2% total but this rose to reach a peak of 89 per cent in 2004. In recent years this proportion has fallen 4% slightly but it was still 82 per cent in 2011 (Figure 16.7). China’s output was produced mainly in Jiangxi 4% and Hunan provinces, accounting for 44 per cent and 24 per cent of the total respectively (Research in China, 2011). Its major operating mines are at 8% Shizhuyuan in Hunan Province and Yangchulin, Xingluokeng and Xianglushan in Jiangxi Province. There are further mines, in these and other prov- inces, and new mines are believed to have opened but accurate information is difficult to obtain. China Russia USA Russia’s output is believed to come from Tyrnyauz Canada Bolivia Other countries in the North Caucases and the Vostok-2 area of the Far East region, but Russian mines were Figure 16.5 Location of the estimated world reserves of reported to be struggling in the economic crisis tungsten, as at December 2011 (Shedd, 2012a). (Levine, 2011). Historically, Bolivia has had

80

70

60

50

40

30

Thousand tonnes 20

10

0 Figure 16.6 Mine production of tungsten, 1992–2011. (Data from British 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Geological Survey World Mineral World total Statistics database.) 402 teresa brown and peter pitfield

1% (mainly from the Kara mine on Tasmania) and 1% Mongolia (BGS, 2013). 1% 1% 1% 2% The concentration of production in China is 2% a relatively recent occurrence, as illustrated by 3% Figure 16.8. Prior to the 1990s, China’s share of the total world production was less than 35 6% per cent. Between 1980 and 1990, although the total world production remained approximately the same, China’s share of that total more than doubled from 29 per cent to 62 per cent. Total world production had dropped significantly by 2000 and in tonnage terms China’s output fell too, but other countries’ production levels reduced by proportionally greater extents. The result was that China’s share of the world total increased to 77 per cent despite the decrease in output. Between 2000 and 2010 total world pro- duction recovered, and in 2010 was higher than 1990, but the majority of this increase has been 82% in China. It is likely that this pattern of supply is related to tungsten prices and market availability. In the China Russia Canada Bolivia late 1970s and the 1980s there was a significant Rwanda Austria Portugal Peru increase in the availability of tungsten concen- trates and intermediate products from China at Spain Others countries (x11) cheaper prices than the rest of the world. This led to oversupply and a significant fall in tungsten Figure 16.7 Mine production of tungsten in 2011. prices, with the result that many mines could not (Data from British Geological Survey, 2013.) sustain economic production and world output decreased. The cost of production was probably numerous small-scale mining operations extract- lower in China than other countries and conse- ing tungsten and this is thought to still be the quently output there fell less. As demand and case (Anderson, 2011). However, output from prices recovered, Chinese producers were able to Bolivia has increased significantly in recent years react more quickly and therefore China’s output from approximately 400 tonnes of contained has grown much more rapidly than that in the tungsten in 2004 to 1400 tonnes in 2011 (BGS, rest of the world. The sub sequent imposition of 2013). Austria’s output comes from the Mittersill production and export quotas in China has mine, operated by Wolfram Bergbau and Hütten pushed prices up further and encouraged both the AG, while Canada’s is from the Cantung mine in exploration for new deposits and the development the Yukon Territory, which is operated by North of new mines outside of China. American Tungsten Corp. Other countries producing tungsten in 2011 included: Portugal (mainly from the Panasqueira Future supplies mine), Peru (mainly from the Pasto Bueno mine), Rwanda, Spain, Uzbekistan, Brazil, Thailand, Burma A recent increase in prices for most forms of (Myanmar), North Korea, Burundi, Kyrgyzstan, tungsten, and concerns over security of supply, Democratic Republic of Congo, Uganda, Australia particularly from China, have caused a revival in Tungsten 403

70 Other Countries

Canada 60 Austria

50 Russia

Australia 40 Portugal

30 Korea, Republic of

Korea, Dem. P.R. Of 20 Thousand tonnes (W content) Bolivia

USA 10 Soviet Union

0 China 1950 1960 1970 1980 1990 2000 2010

Figure 16.8 Evolution of tungsten production concentration, output by country for 1950, 1960, 1970, 1980, 1990, 2000 and 2010. (Data from British Geological Survey World Mineral Statistics database.)

the development of new mines and the reopening reported for the Northern Dancer deposit in of inactive mines in other parts of the world. Canada’s Yukon Territory. Its owner, Largo New production has been recorded in Australia Resources, completed a scoping study in 2008 (2005), Peru (2006) and Spain (2008). and a Preliminary Economic Assessment in In North America, the Andrew mine in 2011 (Largo, 2011a). California, USA was reopened in 2007. The In New Brunswick, eastern Canada, Northcliff Cantung mine in Canada was reopened, again, in Resources completed a full feasibility study in October 2010 following a short suspension in 2013 and is conducting an environmental impact 2009. Resources at Cantung are reported to con- assessment at its Sisson Brook project. The project tain more than 40,000 tonnes of tungsten. Its has a resource of more than 270,000 tonnes of owners, North American Tungsten Corp., also contained tungsten (Northcliff, 2012). own the nearby Mactung tungsten deposit and In South America, Largo Resources announced completed a feasibility study there in 2009, which in December 2011 that it had shipped the first reported a contained tungsten resource of more tungsten concentrates from its Currais Novas than 370,000 tonnes. The Company is currently project in Brazil, which is reworking tailings from evaluating the options for the deposit and seeking past mining operations at Barra Verde and Boca de the necessary environmental permits (North Lage (Largo, 2011b). American Tungsten Corp., 2011 and 2012). In Europe, Wolf Minerals have completed the An even larger resource of more than construction of a new link road at the Hemerdon 390,000 tonnes of contained tungsten has been project in the United Kingdom. A favourable 404 teresa brown and peter pitfield

Definitive Feasibility Study was completed in Dolphin mine on King Island (between the main- 2011 and finances have been put in place. The land Australia and Tasmania). The underground resource contains more than 460,000 tonnes of mine on King Island was operated for tungsten tungsten and mine construction commenced in up until 1990 when it closed due to low prices. 2013 (Wolf, 2012). The resource is reported to still contain over In Spain, Ormonde Mining is aiming to start 190,000 tonnes of tungsten. A DefinitiveFeasibility production in 2014 or 2015 at their Barruecopardo Study was completed early in 2012 and the project. This location, which has been worked required permits and approvals are all in place in the past to shallow depths for tungsten, has a (King Island, 2012). reported resource of more than 50,000 tonnes A summary of selected major deposits under of contained tungsten. A Definitive Feasibility development and projects where new production Study was completed in February 2012 and the is expected in the near future is shown in project is moving through the permitting pro- Table 16.4. cess (Ormonde, 2012). In Asia, Woulfe Mining Corp. are redeveloping the former mine at Sangdong in South Korea. The mine was forced to close in 1992 due to low World trade tungsten prices, but it is still reported to contain resources of more than 280,000 tonnes of con- Tungsten is traded in a variety of forms, chiey tained tungsten. A feasibility study was completed as ores and concentrates, as intermediate prod- in April 2012 with production forecast for 2014 ucts such as APT, and as tungsten oxide (or or 2015 (Woulfe, 2012). trioxide). Although China is by far the largest In Australia, Newcrest Mining Ltd discovered producer of tungsten ores and concentrates, it the O’Callaghan’s deposit in 2008 near to the does not export much material in these forms, Telfer gold mine in the Pilbara region of Western preferring to process the raw material into Australia. The deposit is reported to contain either APT or other products. The main export- more than 200,000 tonnes of tungsten (Newcrest, ing countries in 2011 are shown in Figure 16.9a. 2012) and work towards a prefeasibility study China is the largest importer of tungsten ores continued in 2012. Also in Western Australia, and concentrates, followed by the USA and Hazelwood Resources is conducting a Definitive Austria (Figure 16.9b). Feasibility Study on its Cookes Creek tungsten Due to its high economic importance, together deposit, which has a reported resource of more with concerns over access to raw tungsten mate- than 40,000 tonnes of contained tungsten rials and uncertainty over the ability of Western (Hazelwood, 2012). producers to compete with China in the sale of Developments are also taking place at the tungsten products, the European Commission historical Mount Carbine mine in Queensland, (EC) has identified tungsten as one of its 14 Australia, where Carbine Tungsten Ltd com- ‘critical’ raw materials (EC, 2010). This means menced reprocessing of tailings in 2012. The that they consider there to be a risk of supply company proposes to follow this by processing shortage with an associated potential impact low-grade stockpiles on the site before progress- on the European Union economy. Strategies to ing on to mining of the hard rock resource. reduce this risk are being considered and these A feasibility study for the latter was completed include the diversification of supply sources, in 2012. It is reported that there is more than resource efficiency and recycling. Various coun- 40,000 tonnes of contained tungsten remaining tries, including China, the USA, Japan and Russia, at the mine (Carbine, 2012). King Island Scheelite maintain stockpiles of critical raw materials, Ltd is also planning to reopen the former including tungsten. Table 16.4 Selected major tungsten deposits under development and those where production is expected in the near future.

Current Status Resources (tonnes Name Country (as at January 2013) Possible Production contained tungsten)

Hemerdon United Kingdom Feasibility study completed May 2011, mine 2014 >460,000 construction expected 2013 Barruecopardo Spain Feasibility study completed Feb 2012, 2014 or 2015 >50,000 commissioning expected 2014 Sangdong South Korea Feasibility study completed April 2012, 2014 or 2015 >280,000 commissioning expected 2014 King Island Australia Feasibility study completed Feb 2012, all permits 2015 ? >190,000 in place, raising funding Sisson Brook Canada Feasibility study completed 2013. Environmental 2015 ? >270,000 impact assessment ongoing Cookes Creek Australia Feasibility study ongoing Unknown >40,000 Mactung Canada Feasibility study completed 2009, some Unknown >370,000 environmental permitting in 2012 Northern Dancer Canada Preliminary economic assessment completed in 2011 Unknown >390,000 O’Callaghans Australia Prefeasibility study ongoing Unknown >200,000

Note: Resources are from all categories and in some cases include reserves.

(a) 10,000

9000

8000

7000

6000

5000 Tonnes 4000

3000

2000

1000

0

USA Italy India Bolivia Spain Japan Canada Germany Portugal Australia Netherlands Rep. of Korea United Kingdom Other countries Russian federation China (inc Hong Kong)

Ores & concentrates Ferro-tungsten & ferro-silico-tungsten Metal (incl waste & scrap) Tungstates

Figure 16.9a Main tungsten exporting countries, 2011. Note: ‘tungstates’ is predominantly APT. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.) 406 teresa brown and peter pitfield

(b) 10,000

9000

8000

7000

6000

5000 Tonnes 4000

3000

2000

1000

0

Italy USA Brazil Austria Japan France Germany Slovakia Malaysia NetherlandsSwitzerland Kazakhstan Rep. of Korea United Kingdom Other countries

China (inc Hong Kong)

Ores & concentrates Ferro-tungsten & ferro-silico-tungsten Metal (incl waste & scrap) Tungstates

Figure 16.9b Main tungsten importing countries, 2011. Note: ‘tungstates’ is predominantly APT. (Data from British Geological Survey World Mineral Statistics database and UN Comtrade, 2013.)

Prices despite a decrease at the end of 2008. Prices for APT have been more volatile, with a notable Prices for tungsten concentrates produced by decrease towards the end of 2008, in common mines, and the intermediate tungsten powders pro- with many other metal prices, caused by the duced by the secondary processors, are quoted in global recession, followed by a significant increase terms of metric tonne units (mtu). A metric tonne in the second half of 2009 which continued unit is the equivalent of 10 kilograms of contained through most of 2011. Overall, between the first tungsten trioxide. Ten kilograms of tungsten quarter of 2004 and the third quarter of 2011, trioxide contains 7.93 kilograms of tungsten metal. prices for ore have risen by more than 200 per Tungsten is traded either on undisclosed cent, while those for APT over the same period supply contracts between the primary producers, have increased by more than 500 per cent. secondary processors and tertiary manufacturers, or via traders. The Chinese price for APT is the benchmark in the tungsten market. Outlook Tungsten prices for the main traded forms show a similar trend with a very sharp increase The worldwide supply of tungsten is dominated during the later part of 2004 through to the early by China, with more than 70 per cent of total part of 2006 (Figure 16.10). Since then prices for world production in every year since 1994 and ores and concentrates have generally stabilised, more than 60 per cent of estimated world Tungsten 407

500 450 400 350 300 250 200 150

US$ per metric tonne unit 100 50 0 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 Q1 Q3 2004 2005 20062007 2008 2009 2010 2011 Tungsten Ore European, Min. content WO3 65% CIF, quarterly average Tungsten APT Chinese, No1 Hong Kong FOB, quarterly average

Figure 16.10 Tungsten ore and APT price trend 2004–2011 (Data from Metal Bulletin.)

reserves in 2011. China imposes strict controls development of this domestic industry has on its tungsten industry by limiting the number resulted in China becoming the world’s largest of exploration and mining licences issued within consumer of tungsten, as well as the largest its borders, applying production quotas and producer, and this is reected in the sharp imposing export taxes and quotas (Shedd, 2012b). increase in China’s imports of tungsten ores Production quotas have been increased in recent (Figure 16.12). It was reported in 2010 that China years. However, export quotas have been reduced accounted for 37 per cent of the world’s consump- from 18,075 tonnes in 2003 to 15,400 tonnes in tion of tungsten (Toovey, 2010) and the main con- 2012 (Hayes, 2011; Smith, 2011). It is believed suming industries within the country have that the higher prices for tungsten ore and APT continued to grow. It is likely that China’s tung- shown in Figure 16.10 are principally due to sten industry will remain dominant and that these export restrictions. Further reductions in these trends will continue. China’s exports of the export quota are likely, although China’s ‘tungstates’ (which includes APT) have fallen production output is expected to continue to steadily since their peak in 1995 and it is possible increase. that supplies of Chinese APT to the rest of the Whereas in the past China has exported con- world may be further restricted in future if the siderable quantities of tungsten ores, in recent downstream tungsten metal industries in China years its exports have been concentrated on the continue to develop. higher-value intermediates, such as APT, and to a Concerns have been raised that China’s lesser extent on tungsten metal products export controls disproportionately disadvan- (Figure 16.11). This reects the development of tages non-Chinese consumers (Smith, 2011), and the domestic tungsten industry in China, and in March 2012 the United States of America, may also be the underlying reason for the export European Union and Japan all requested formal quotas, i.e. that China is seeking to preserve its ‘consultations’ with China though the World resources and support its domestic industry. The Trade Organisation (WTO, 2012a). These 408 teresa brown and peter pitfield

35000

30000

25000

20000 Tonnes 15000

10000

5000 Figure 16.11 China exports of tungsten ores and concentrates compared to 0 exports of tungstates (mainly APT) and tungsten metal, 1980–2010. (Data from

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 British Geological Survey World Mineral Statistics database and UN Ores & concentrates Metal Tungstates Comtrade, 2013.)

14,000

12,000

10,000

8000 Tonnes 6000

4000

2000 Figure 16.12 China imports of tungsten ores and concentrates, 0 tungstates (mainly APT) and tungsten metal, 1992–2010. (Data from British

1992 1993 1994 1995 1996 1997 1998 1999 2000 2002 2001 2003 2004 2005 2006 2007 2008 2009 2010 Geological Survey World Mineral Statistics database and UN Ores Metal Tungstates Comtrade, 2013.)

requests are the first stage in disputes through Higher prices, and the recognition of the risks the WTO and enable discussions between parties associated with production being concentrated in to take place in an attempt to find solutions. a few producing countries (BGS, 2012), has encour- However, these discussions did not resolve the aged the development of tungsten mines and pro- concerns and in July 2012 the Dispute Settlement jects outside of China. Whilst it is inevitable that Body within the WTO formally agreed to estab- not all proposals will come to fruition, several lish a panel to examine the issues (WTO, 2012b). new non-Chinese mines are likely to start pro- A report and WTO ruling is expected in 2013. ducing in the near future, as shown in Table 16.4. Tungsten 409

Tungsten use is largely inuenced by condi- 9 pp. http://www.bgs.ac.uk/mineralsuk/statistics/ tions in the global economy. As a consequence, riskList.html the consumption of tungsten in most Western British Geological Survey (BGS). (2013) World Mineral countries declined significantly in 2009 as a Production 2007–2011, British Geological Survey, Keyworth, Nottingham. result of the global recession but subsequently Carbine Tungsten Ltd. (2012) Quarterly Activities recovered. Future consumption of tungsten will Report for quarter ended 30 September 2012. http:// mainly depend on the performance of sectors that www.carbinetungsten.com.au use hard metals (tungsten carbide and cemented Chang, Z. (2005) World Skarn Deposits - China. carbide), particularly in machine tools, which Appendix 2 of Mernert et al. (2005) World Skarn include the manufacturing, metal working, min- Deposits. In: Hedenquist, J.W., Thompson, J.F.H., ing, construction and petroleum industries. Goldfarb R.J. and Richards J.R. (eds.) Society of Further development of specialist steel alloys and Economic Geologists One hundredth Anniversary ‘superalloys’ for a variety of applications requiring Volume (1905–2005). high degrees of wear and temperature resistance Christie, T and Brathwaite, B. (1996) Mineral com- are also likely to require significant quantities of modity report 12 - Tungsten. Ministry of Economic Development, New Zealand. http://www.nzpam. tungsten. Consumption in this area will depend govt.nz/cms/pdf-library/minerals/publications/ on developments in the aerospace, industrial tur- Commodity%20Reports bines and similar industries. Future develop- Chung, C.K. (1975) Tungsten and molybdenum deposits ments of new technologies, such as the current in South Korea. Korean Institute of Mineral and designs for fusion reactors, may also require Mining Engineers Journal 12, 192–199. significant quantities of tungsten. Clauson, J.L. and Korte, N. (2009) Environmental fate of tungsten from military use. Science of the Total Environment 407, 2887–2893. Acknowledgements Cox, D.P. and Bagby, W.C. (1986) Descriptive model of W veins. In: Cox, D.P. and Singer, D.A. (eds.) Mineral deposit models: U.S. Geological Survey Bulletin The authors publish with the permission of the 1693, p. 64. Executive Director of the British Geological Survey. De Roo, J.A. (1988) Structural controls on the emplace- ment of the vein-type tungsten-tin ore at Mount Carbine, Queensland, Australia. Economic Geology References 83, 1170–1180. Dick, L.A. and Hodgson, C.J. (1982) The Mactung tung- Altringer, P.B. (1985) Recovery of Tungsten from Searles sten-copper-zinc contact metasomatic and related Lake Brines by an Ion-exchange Process. Bull. US deposits of the northeastern Canadian cordillera. Dept of the Interior, Bureau of Mines, 682, pp. 47. Economic Geology 77, 845–867. Anderson, S.T. (2011) Mineral Industry of Bolivia. 2009 Eichhorn, R., Höll, R., Loth, G. and Kennedy, A. (1999) Minerals Yearbook, United States Geological Survey, Implications of U-Pb SHRIMP zircon data on the age pp. 17. and evolution of the Ferbertal tungsten deposit Bednar, A.J., Jones, W.T., Boyd, R.E., Rindelberg, D.B. (Tauern Window, Austria). Geologische Rundschau and Larson, S.L. (2008) Geochemical parameters 88, 496–512. inuencing tungsten mobility in soils. Technical Einaudi, M.T., Meinert, L.D. and Newberry, R.J. (1981) reports: Heath metals in the environment. American Skarn deposits. Economic Geology 75th Anniversary Society of Agronomy, Crop Science Society of Volume, 317–391. America, and Soil Science Society of America. Elliott, J.E. (1992) Tungsten - geology and resources of Brand, A.A. (2008) Mineralogy, geochemistry and geochro- deposits in southeastern China. In: DeYoung Jr, J.J. nology of the Northern Dancer tungsten-molybdenum and Hammarstrom J.M. (eds.) Contributions to com- deposit, Yukon and British Columbia. MSc thesis, modity geology research. U.S. Geological Survey University of British Columbia. Bulletin 1877, I1–I10. British Geological Survey (BGS). (2012) The risk list. Elliott, J.E., Kamilli, R.J., Miller, W.R. and Livo, K.E. British Geological Survey, Keyworth, Nottingham. (1995) Vein and greisen Sn and W deposits. In: Du 410 teresa brown and peter pitfield

Bray, E.A. (ed.) Preliminary compilation of descrip- International Tungsten Industry Association (ITIA) tive geoenvironmental mineral deposit models. (2012). Tungsten Products. http://www.itia.info/ U.S. Geological Survey Open-File Report 95-831, tungsten-products.html 62–69. Jarchovsky, T. (2006) The nature and genesis of greisens European Commission (EC). (2010) Critical raw mate- stocks at Krásno, Slavkovský les area – western rials for the EU. Report of the Ad-hoc Working Bohemia, Czech Republic. Journal of the Czech Group on defining critical raw materials. European Geological Society 51, 201–215. Commission Enterprise and Industry. http://ec. Keats, W. (1981) Alteration and replacement of wol- europe.eu/enterprise/policies/raw-materials/critical/ framite in the Hemerdon tungsten deposit, Devon. index_en.htm PhD thesis, Open University. Goossens, P.J. (1978) The metallogenic provinces of Kelly, W.C. and Rye, R.O. (1979) Geologic, uid Burma—their definitions, geologic relationships, inclusion, and stable isotope studies of the tin-tungsten and extension into China, India, and Thailand. In: deposits of Panasqueira, Portugal. Economic Geology Nutalaya, P. (ed.) Proceedings of the Third Regional 74, 1721–1822. Conference on Geology and Mineral Resources of Kerr, P.F. (1940) Tungsten-bearing manganese deposits Southeast Asia, Bangkok, Thailand, 14–18 November at Galconda, Nevada. Geological Society of America 1978, 431–492. Bulletin 51, 1359–1389. Gu, J Y. (1982) Morphological zoning of the vein-type King Island Scheelite Ltd. (2012) Dolphin Project – tungsten deposits in southern China. In: Hepworth, overview and press releases. http://www.kingisland- J.V. and Zhang, Y.H. (eds.) Proceedings of Symposium scheelite.com.au/index.php?id=18 on Tungsten Ore deposits, Geological Publishing Kooiman, G.J.A., McLeod, M.J. and Sinclair, W.D. House, Beijing, 269–278. (1986) Porphyry tungsten-molybdenum orebodies, Guerenko, A. and Schmincke, H.U. (2002) Tungsten in polymetallic veins and replacement bodies and Searles Lake. Mining Engineering 11, 301–303. tin-bearing greisens zones in the Five Tower zone, Guiliani, G., Li, Y.D. and Sheng, T.F. (1988) Fluid Mount Pleasant, New Brunswick. Economic Geology inclusion study of Zihuashan tungsten deposit in 81, 1356–1373. southern Jiangxi province, China. Mineralium Koutsospyros, A., Braida, W., Christodoulatos, C., Deposita 23, 24–33. Dermatas, D. and Strigul, N. (2006) A review of tung- Hayes, M. (2011) China’s dominance in tungsten sten: From environmental obscurity to scrutiny. to continue as exports decrease. Fast Markets Ltd, Journal of Hazardous Materials 136, 1–19. 15 November 2011. Kwak, T.A.P. (1987) W-Sn skarn deposits and related Hazelwood Resources Ltd. (2012) Cookes Creek metamorphic skarns and granitoids. Developments Tungsten - overview. http://www.hazelwood.com.au/ in Economic Geology, 24, Elsevier Publ. Co., pp. 451. Projects/Cookes_Creek_Tungsten/ Kwak, T.A.P., and Tan, T.H. (1981) The geochemistry of Hinde, C. (ed.). (2008) Tungsten. Mining Journal Special zoning in the skarn minerals at King Island (Dolphin) Publication (June 2008). p16 http://www.mining- mine. Economic Geology 76, 468–497. journal.com/_data/assets/supplement_file_attach- Landis, Gary P. and Rye, R.O. (1974), Geologic uid ment/0014/ 123314/tungsten-scr.pdf inclusion and stable isotope studies of the Pasto Höll, R. and Eichhorn, R. (2000) Tungsten minera- Bueno tungsten-base metal ore deposit, northern lisation and metamorphic remobilisation in the Peru. Economic Geology 69, 1025–1059. Ferbertal scheelite deposit (Central Alps, Austria). Largo Resources Ltd. (2011a) Largo begins drill program In: Spry, P.G., Longstaff, F.J. and Vokes F.M. (eds.) for current prefeasibility study at Northern Dancer Metamorphosed and metamorphogenic Ore Deposits: tungsten-molybdenum project, Yukon. Press release 6 Rev. Economic Geology 11, 233–264. July 2011. http://www.largoresources.com/Investors/ Huang, C. and Xiao, Z. (1986) A general review of the Press-Releases/ tungsten deposits in China. In Beus, A.A. (ed.) Largo Resources Ltd. (2011b) Largo Resources announces 1986 Geology of Tungsten. International Geological first commercial shipment from Currais Novos tung- Correlation Programme Project 26 ‘MAWAM’ sten project. Press release 14 December 2011. http:// (Mineralisation Associated with Acid Magmatism). www.largoresources.com/Investors/Press-Releases/ United Nations Educational, Scientific and Cultural Lassner, E. and Schubert, W-D. (1999) Tungsten: Organisation, Paris, France. Properties, Chemistry, Technology of the Element, Tungsten 411

Alloys and Chemical Compounds. Kluwer Academic/ Mining Magazine (1979) Hemerdon – Britain’s largest tung- Plenum Publishers, New York, USA pp. 422. sten deposit. Mining Magazine October 1979, 342–351. Levine, R.M. (2011) The Mineral Industry of Russia. Molavi, M., Hollett, G., Board, W.S., Bolu, M., Smith, P., 2009 Minerals Yearbook. United States Geological Lemieux, J., Netherton, D.E., Nyland, E. and Weston, Survey, pp. 26. S. (2011) Preliminary Economic Assessment for the Li, G.L. Hua, R.M. Zhang, W.L. et al. (2011) He-Ar iso- Northern Dancer Project, Yukon, Canada. Prepared tope composition of pyrite and wolframite in the by AMC for Largo Resources Ltd. Filed on www. Tieshanlong tungsten deposit, Jiangxi, China: sedar.com on 7 April 2011. Implications for uid evolution. Resource Geology Nanomaterials (2009) New nickel-tungsten alloy that is 61, 356–366. not only safer chrome but also more durable. News Li, S.R., Wang, D.H., Liang, T., Qu, W.J. and Ying, L.J. article posted on the A-Z of Nanotechnology web- (2008) Metallogenic epochs of the Damingshan tung- page. 19 May 2009. http://www.azonano.com/news. sten deposit in Guangxi and its prospecting potential. asp?newsID=11591. Acta Geologica Sinica 82 873–879 (in Chinese with Narciso, H., Iakovlev, I., de Ruijter, M.A. et al.(2009) English abstract). Amended Technical Report on the Mactung Property, Li, X-H., Liu, D., Sun, M., Li, W-X., Liang, X-R. and Liu prepared by Wardrop et al. for North American Y. (2004) Precise Sm-Nd and U-Pb isotopic dating of Tungsten Corp. Ltd. www.sedar.com on 18 May 2010. the supergiant Shizhuyuan polymetallic deposit and National Bureau of Statistics of China (NBSC). (2012) its host granite, SE China. Geological Magazine 141 Statistical Yearbook of China 2011. http://www.stats. (2), 225–231. gov.cn/english/statisticaldata/yearlydata/. Liu, W. (1980) Geological Features of Mineralisation of Neinavale, H., Thalmann, F., Ataii, B. and Beran, A. the Xingluokeng Tungsten (Molybdenum) Deposit, (1989) Wolframite and scheelite-bearing carbonate Fujian Province. In: Hepworth, J.V. and Lu, H.Z. (eds.) rocks of the Nock Mountains, Austria: A new type Tungsten Geology, China, ESCAP/RMRDC, of tungsten mineralisation in the Eastern Alps. Bandung, Indonesia, 338–348. Mineralium Deposita 24, 14–18. Liu, Y.J and Ma, D.S. (1993) Vein-type tungsten deposits Newberry, R.J. and Einaudi, M.T. (1981) Tectonic and of China and adjoining regions. Ore Geology Reviews geochemical setting of tungsten skarn mineralization 8, 233–246. in the Cordillera. Arizona Geological Society Digest Lu, H-Z., Liu, Y., Wang, C., Xu, Y. and Li, H. (2003) 14, 99–111. Mineralisation and uid inclusion study of the Newcrest Mining Ltd. (2012) December 2011 Shizhuyuan W-Sn-Bi-Mo-F skarn deposit, Hunan Resources and Reserves Statement. Market Release Province, China. Economic Geology 98, 955–974. 10 February 2012. http://www.newcrest.com.au// Ma, L. (1982) The geological characteristics of resources.asp. Damingshan sedimentary-magmatic- hydrothermal Noble, S.R., Spooner, E.T.C. and Harris, F.R. (1984) The wolframite deposit, Guangxi Zhuang Autonomous Logtung large tonnage, low-grade W (scheelite)-Mo Region, China. In: Hepworth, J.V., and Yu, H.Z., (eds.) porphyry deposit, south-central Yukon Territory. Symposium on Tungsten Geology [Jiangxi, China, Economic Geology 79, 848–868. October 1981]: Bandung [Indonesia], ESCAP/RMRDC North American Tungsten Corp. (2011) North American and Beijing, Geological Publishing House, 374–384. Tungsten establishes special committee for Mactung, Marsh, S.P. and Erickson, R.L. (1975) Integrated geologic releases second quarter financial results and provides and geochemical studies, Edna Mountain, Nevada. an update on Cantung operations. Press release 30 May In: Elliot, I.L. and Fletcher, W.K. (eds.) Geochemical 2011. http://www.natungsten.com/. Exploration 1974. Elsevier Publ. Co., Amsterdam, North American Tungsten Corp. (2012) North American Netherlands, 237–250. Tungsten Achieves Major Permitting Milestone for Meinert, L.D., Dipple, G.M. and Nicolescu, S. its Mactung Project, Yukon. Press release 16 October (2005) World Skarn Deposits. In: Hedenquist, J.W., 2012. http://www.natungsten.com/. Thompson, J.F.H., Goldfarb, R.J. and Richards, J.R. Northcliff Resources Ltd. (2012) Management Discussion (eds.) Society of Economic Geologists One hundredth and Analysis report for period ended 31 October 2012 Anniversary Volume (1905–2005), 299–336. released on http://www.sedar.com Metal Bulletin Online (2012) Price information. www. Ormonde Mining Plc. (2012) Barruecopardo tungsten metalbulletin.com project summary. http://www.ormondemining.com/ 412 teresa brown and peter pitfield

en/projects/barruecopardo_tungsten_project/sum- http://pubs.usgs.gov/circ/circ1196-R/. (Supersedes U.S. mary_btp#Summary Geological Survey Open-File Report 2005–1028.) Petkewich, R A. (2009) Unease over tungsten. Science Shedd, K B. (2012a) Tungsten. Mineral Commodity and Technology, 19 January 2009, 87, 3, 63–65. http:// Summaries, United States Geological Survey. http:// pubs.acs.org/cen/science/87/ 8703sci2.html minerals.usgs.gov/minerals/pubs/commodity/tung- Pitfield, P.E.J. and Brown, T.J. (2011) Tungsten. Mineral sten/680400.pdf Profile series, British Geological Survey. http://www. Shedd, K.B. (2012b) Tungsten. 2010 Minerals Yearbook. bgs.ac.uk/mineralsuk United States Geological Survey. http://minerals. Pitfield, P.E.J. Brown, T J and Idoine, N E. (2010) Mineral usgs.gov/minerals/pubs/commodity/tungsten/myb1- Information and Statistics for the BRIC countries 2010-tungs.pdf 1999–2008. British Geological Survey. http://www. Simmons, S.F. and Sawkins, F.J. (1983) Mineralogic and bgs.uk/mineralsuk uid inclusion studies of the Washington Cu-Mo-W- Plimer, I. 1975. Wolfram Camp wolframite-molybdenite- bearing breccia pipe, Sonora, Mexico. Economic bismuth quartz pipes, North Queensland. In: Knight, Geology 78, 521–526. C.L. (ed.) Economic Geology of Australasia and Papua Smith, M. (2011) China cuts 2012 export quotas for tung- New Guinea v.1 – Metals. Australasian Institute of sten. Tungsten Investing News, 9 November 2011. Mining and Metallurgy Monograph Series 5, 760–762. Snow, R.J. and Coker, W.B. (1986) Overburden geo- Ray, G.E. (1995) Tungsten Skarns, in Selected British chemistry related to W-Cu-Mo mineralisation at Columbia Mineral Deposit Profiles, Volume 1 - Sisson Brook, New Brunswick, Canada: an example Metallics and Coal, Lefebure, D.V. and Ray, G.E., of short- and long-distance glacial dispersal. Journal Editors, British Columbia Ministry of Employment of Geochemical Exploration 28 (1987), 353–368. and Investment, Open File 1995-20, 71–74. http:// Soloviev, S.G. and Krivoschchekov, N.N. (2011) www.empr.gov.bc.ca/Mining/Geoscience/ Vostok-2 gold-base-metal-tungsten skarn deposit, MineralDepositProfiles/ListbyDepositGroup/Pages/ Central Sikhote-Alin, Russia. Geology of Ore Deposits KSkarn.aspx#K05 53, 478–500. Research in China (2011) China Tungsten Industry Strigul, N., Koutsospyros, A., Arienti, P., Christodoulatos, Report 2010-11, Abstract. http://www.researchinchina. C., Dermatas, D. and Braida, W. (2005) Effects of tungsten com/Htmls/Report/2011/6076.html on environmental systems. Chemosphere, 61, 248–258. Roskill Consulting Group. (2010) A custom study of the Strigul, N., Galdun, C., Vaccari, L., Ryan, T., Braida, W. market for tungsten. Commissioned by Wolf Minerals and Christodoulatos, C. (2009) Inuence of specia- Ltd. Roskill Consulting Group, London, UK. tion on tungsten toxicity. Desalination 248, 869–879. Rubinstein, J.P. and Barsky, L.A. (2002) Mineral Deposit Strigul, N. (2010) Does speciation matter for tungsten and Resources of the Former Soviet Union, Vol 1: ecotoxicology? Ecotoxicology and Environmental Non-Ferrous Metal Ores: Deposits, Minerals and Safety 73, 1099–1113. Plants. Taylor and Francis, pp. 411. Tarney, E. (2004) High speed steel properties: different Rudnick, R L and Gao, S. (2003) The Composition of the grades for different requirements. Article for Gear Continental Crust. In: The Crust Vol. 3, Treatise on Technology Journal, Randall Publishing, Elk Grove Geochemistry (eds. Holland, H.D. and Turekian, K.K.). Village, IL60009, USA. http://www.technology.com/ Elsevier-Pergamon, Oxford, United Kingdom. 1–64. pa/members/sepoct04/section9. pdf Schmidt, S. (2012) From deposit to concentrate: The Todd, R.G. (1989) Evaporites and strata-bound tungsten basics of tungsten mining. Part 1: Project generation mineralisation. Geology 17, 139–143. and project development. International Tungsten Toovey, L.M. (2010) China’s tungsten demand to grow Industry Association newsletter, June 2012. 8 per cent this year. Tungsten Investing news, 22 July Schubert, W.D. and Lassner, E. (2009) Tungsten. 2010. International Tungsten Industry Association. http:// Tornos, F., Galindo, C., Crespo, J. L., and Spiro, B. F. (2007) www.itia.info/news-2-title.html Geochemistry and origin of calcic tungsten-bearing Shedd, K.B. (2011) Tungsten recycling in the United skarns, Los Santos, Central Iberian Zone, Spain. States in 2000, chap. R of Sibley, S.F., Flow studies for Canadian Mineralogist 46, 87–109. recycling metal commodities in the United States: United Nations (UN) Comtrade (2013) Commodity U.S. Geological Survey. Circular 1196–R, p. R1–R19, Trade Statistics Database. http://comtrade.un.org/db. Tungsten 413

United Nations Environment Programme (UNEP). Zaraisky, G.I. and Dubinina, E.O. (2001) On duration (2011) Recycling Rates of Metals, A Status Report. of the magmatic and hydrothermal processes at International Resource Panel. Working Group on the Akchatau greisen W-Mo deposit in central Kazakhstan. Global Metal Flows. In: Piestrzynski et al. (eds.) Mineral Deposits at the Wang, F.Y., Li, C.Y., Ling, M.X., Zhang, H., Sun, Y.L. Beginning of the 21st Century, Swets and Zeitlinger and Sun, W. (2011) Geochronology of the Xihuashan B.V., Lisse, Netherlands, 505–510. tungsten deposit in southeastern China: constraints Zhang, J., Chen, Z., Wang, D., Chen, Z., Liu, S. and from RE-Os and U-Pb dating. Resource Geology 61, Wang, C. (2008) Geological characteristics and metal- 414–423. logenic epoch of the Xungluokeng tungsten deposit, Werner, A B T., Sinclair, W.D. and Amey, E.B. (1998) Fujian Province. Geotectonica et Metallogenia 32, International Strategic Mineral Issues Summary 92–97. Report – Tungsten. U.S. Geological Survey Circu- Zhang, R., Lu, J., Zhu, J., Wang, R., Chen, J. and Gao, lar 930-0. http://pubs.usgs.gov/pdf/circular/c930-0. J. (2011) Re-Os and U-Pb geochronology of large html Xintianling skarn-type scheelite deposit, Nanling Wolf Minerals Ltd. (2012) Hemerdon tungsten and tin Range, China. 11th Biennial Meeting SGA 2011. project. http://www.wolfminerals.com.au/investor- https://sga2011conference services.net/ resources/ relations/press-releases Zhang, Z., Mao, J., Yang, J., Wang, Z. and Zhang, Z. (2003) World Trade Organisation (WTO) (2012a) US, EU and Geology and genesis of the Ta’ergou skarn-quartz vein Japan file disputes against China. WTO 2012 News tungsten deposit in north Qilian Caledonian orogenic Items, 13 March 2012. http://www.wto.org/english/ belt, northwest China. Resource Geology 53, 101–114. news_e/news12_e/dsrfc_13mar12_e.htm Zhao, Y., Wenwei, L., Chengsi, B., Daxin, L. and World Trade Organisation (WTO) (2012b) Panel Chongjun, J. (1990) Skarn Ore Deposits in China. established on China’s rare earth exports. WTO News Beijing, Geological Press, pp. 354. Items, 23 July 2012. http://www.wto.org/english/ Zhaolin, L. and Zhongfang, Y. (1996) Rock forming and news_e/news12_e/dsb_23jul12_e.htm ore-forming temperatures of Lianhuashan tungsten Woulfe Mining Corp. (2012) Sangdong – Overview. deposit, Guangdong Province. Chinese Journal of http://www.woulfemining.com/s/Sangdong.asp Geochemistry 15, 239–248. Appendix 1 Units and symbols used in this volume

Throughout this publication units are, in general, has the unit of ‘metres per second’ can be shown expressed in SI units (i.e. according to the as ‘m/s’ or ‘m s–1’. In this volume, the symbol ‘ / ‘ International System of Units or Système interna- has been used to represent the word ‘per’, which tional d’unités). The main exception to this is the represents the action of dividing one unit by unit for temperature where the more commonly another. Symbols for compound units involving recognised degrees Celsius (°C) is used in preference multiplication (rather than division) can be either to Kelvin (K). The intervals between degree Celsius joined using a raised dot or a non-breaking space, and Kelvin are identical; the only difference is the for example ‘N · m’ or ‘N m’. In this volume the position of zero. The abbreviations or symbols used non-breaking space has been used. in this book are shown in Table A.1.1 together with For some properties the commonly used units the unit name in full and the property for which are not included within the SI system. In these they are used. In some instances the internationally instances the values for the property have been con- recognised Standard Prexes are used alongside the verted into SI units. Similarly, within the SI system SI units and these are also included in Table A1.1. certain properties can be described using more than There are several acceptable methods of dis- one compound unit. The conversion factors bet- playing compound units, for example speed which ween certain units are shown in Table A1.2.

Table A1.1 Units used in this volume

Symbol or abbreviation Unit name Property

/°C per degree Celsius coefficient of thermal linear expansion °C degree Celsius temperature µΩ m micro-ohm metre (10−6 ohm metre) electrical resistivity GPa gigapascal (109 pascal) pressure or stress (including Young’s modulus) J/(g °C) joule per gram degree Celsius specific heat capacity J/g joule per gram latent heat of fusion or vaporisation kg kilogram mass

(continued )

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Appendix 1 415

Table A1.1 (cont’d )

Symbol or abbreviation Unit name Property kg/m3 kilogram per cubic metre density kJ/kg kilojoule per kilogram latent heat of fusion or vaporisation km kilometre length m metre length m3/kg cubic metre per kilogram mass magnetic susceptibility MN/m2 meganewton per square metre Brinell hardness MPa megapascal (106 pascal) pressure or stress (including tensile strength) nΩ m nano-ohm metre (10−9 ohm metre) electrical resistivity pm picometre (10−12 metre) length S/m siemen per metre electrical conductivity V volt electric potential difference or electromotive force W/(m °C) watt per metre degree Celsius thermal conductivity Ω m ohm metre electrical resistivity

Table A1.2 Conversion factors for units

Property name Convert from unit Convert to unit Conversion required

Brinell hardness meganewton per square megapascal MPa no conversion necessary metre MN/m2 because 1 pascal is equal to 1 newton per square metre Density gram per cubic centimetre kilogram per cubic metre multiply by 1000 g/cm3 kg/m3 Latent heat of fusion or joule per gram J/g kilojoule per kilogram no conversion necessary vaporisation kJ/kg Radius of atoms or cations Ångstrom Å picometre pm multiply by 100 Specific heat capacity joule per kilogram degree joule per gram degree divide by 1000 Celsius J/(kg °C) Celsius J/(g °C) Specific heat capacity joule per gram degree joule per gram kelvin J/(g K) no conversion necessary Celsius J/(g °C) Temperature degree Celsius °C kelvin K add 273 Thermal conductivity watt per metre degree watt per centimetre degree divide by 100 Celsius W/(m °C) Celsius W/(cm °C) Thermal conductivity watt per metre degree watt per metre kelvin no conversion necessary Celsius W/(m °C) W/(m K) Appendix 2 Geological time periods (simplified)

Eon Era Period Age (Ma, million years before present) Quaternary 2.6 Neogene 23 also known as Tertiary Cenozoic Palaeogene 65 Cretaceous 145 Jurassic 199 Mesozoic Triassic 251 Permian 299 Phanerozoic Carboniferous 359 Devonian 416 Silurian

Palaeozoic 443 Ordovician 488 Cambrian 542

Proterozoic 2500 Precambrian Archaean

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. Glossary of technical terms

A alpha particle a positively charged particle con- taining two protons and two neutrons, identical absorption spectrum when electromagnetic radia- to the nucleus of a helium atom, which may be tion is passed through a substance, certain wave- emitted during radioactive decay. lengths are absorbed by that substance; the variation in absorption intensity is the absorption spectrum. aluminosilicate mineral a mineral composed primarily of aluminium, silicon and oxygen, accessory mineral a mineral that is present in together with other cations. small amounts in a rock but is not considered to be characteristic of the rock. aluminothermic reduction an exothermic reac- accretionary terrane an assemblage of frag- tion involving aluminium oxide and iron oxide as ments of different tectonic plates joined together reducing agents. along a convergent plate boundary. See also: plate, amorphous lacking a clear shape; non-crystal- convergent plate boundary. line. actively spreading ridge a ridge created by the amphoteric a molecule that can react as an acid forces of plate tectonics at divergent plate mar- as well as a base (alkali). gins. Typically occur under the oceans, where it is known as ‘oceanic spreading,’ or ‘mid-ocean anhydrous containing no water. spreading’, but also on land at ‘rift zones’. anode slimes material collected at the bottom albitite a rock composed almost totally of the of electrolytic cells during the refining of copper feldspar mineral albite. metal. alkaline igneous rock rock formed from magmas anorthosite a type of igneous rock largely com- enriched in the alkalis, sodium and potassium; posed of plagioclase feldspar. they are characterised by the presence of alkali anthroposphere that part of the biosphere feldspar and/or feldspathoids. that is made or modified by humans for use in alluvial placer deposit a type of mineral deposit human activities; sometimes referred to as laid down in streams or rivers. technosphere.

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 420 Glossary of technical terms argillic alteration hydrothermal alteration of blister an intermediate stage in copper metal rocks, generally at low temperature, that results production, with a copper content of 97–99%, in the formation of clay minerals such as kaolin- before refining into pure metal. ite, smectite or illite. breakdown voltage the minimum voltage A-type granite (or granitoid) derived from man- required to cause part of an insulator to begin tle depths and emplaced in zones of large-scale conducting electricity. extension of the Earth’s crust. breccia a rock composed of angular fragments of one or more rock types cemented by a fine- B grained groundmass. back-arc rift zone a rift zone that sometimes occurs near to convergent plate boundaries as a C result of the subduction processes, particularly if the converging plates are both oceanic. They are calc-alkaline a suite of igneous rocks which often associated with island arcs. evolves by fractional crystallisation from basic rock types, such as basalt or gabbro, to acid Basel Convention on the Control of Transbound- rock types, such as rhyolite or granite, with ary Movements of Hazardous Wastes and their little or no associated iron enrichment. See Disposal a convention adopted in Basel, Swit- also: tholeiitic. zerland in 1989 with the objectives of protecting human health and the environment from the calcine to heat a substance to a temperature at adverse effects of hazardous wastes. which it is oxidised or reduced, or loses water or at which carbonate compounds are decomposed. basement terranes underlying, generally older The term is also used to describe the product of and deformed, rock units. this process. Bayer liquor the uid produced by digesting capacitor an electrical component used to store bauxite in the Bayer process from which the alu- energy in an electric field. mina is precipitated. carbonaceous a substance rich in carbon. Bayer process the main industrial process used to refine bauxite into alumina (aluminium oxide). carbonate a compound containing the carbon- bearing anion CO 2–. beach placer deposit mineral deposit formed 3 along coastlines by the action of the sea. carbonatisation the entire or partial replacement of rocks by carbonate minerals, such as calcite or Beijing consensus a term used to describe dolomite. alternative plans for economic development com- pared to the Washington consensus; as the name carbonatite igneous rock composed of more suggests the model for these alternatives is China. than 50% carbonate minerals. beneficiation the treatment of ore to separate catalyst a substance that changes the rate of a the desired mineral from the gangue. May involve chemical reaction whilst not being consumed by a range of processes such as crushing, washing, the reaction itself. sizing of particles and concentration of the cathode depolariser a substance that removes desired mineral by various methods. hydrogen build up on the cathode of a battery. bioleaching a type of leaching that utilises cathodic protection a technique used to control bacteria to extract the metals. See also: leaching. corrosion of a metal surface by making it the biosphere all life on Earth. cathode of an electrochemical cell. Glossary of technical terms 421 chalcogenide glass a glass containing one of the and which are sold to provide funds for armed elements from Group 16 of the Periodic Table. groups. chalcophile a group of elements that have a continental brines uids containing a high level strong affinity with sulfur. of dissolved solids that occur close to the surface of continental crust. chloritisation the alteration of ferromagnesian minerals to chlorite. continental plate plate of the Earth’s crust form- ing the land areas; typically 35 to 50 km thick and chromatic dispersion an effect that results from more variable in composition, but higher in a variation in propagation delay with wavelength; silica, than oceanic crust. the most familiar example is the rainbow, where convergent plate boundary the boundary bet- white light is split into its component wave- ween two plates which are moving towards lengths (as evidenced by different colours) as a each other. The crust is destroyed at these result of dispersion. boundaries. Also known as ‘destructive plate chromitite a rock containing a high concentration boundary’. of the mineral chromite (an iron chromium oxide, cosmogenic produced by cosmic rays. FeCr2O4). Cottrell ash (or dust) ash or dust collected in a circular economy a circular economy is a model type of electrostatic precipitator which was in- which aims to reduce the generation of waste, re- vented by Frederick Cottrell and designed to use materials for a new purpose or treat waste remove dust from industrial exhaust fumes. materials as a resource and recycle them into new products. covalent bond a form of chemical bond which is characterised by the sharing of pairs of electrons cleavage cleavage planes are planes of relative between two atoms. weakness as a result of the positioning of atoms the arrangement of atoms within a crystal and allow the crystal to be more crystalline structure in a crystal; there are seven basic 3-dimensional easily split. patterns that can be formed: triclinic, monoclinic, coefficient of expansion the degree of expansion orthorhombic, rhombohedral, tetragonal, hexago- of a substance as temperature increases; a low nal and cubic. coefficient means that an element expands little as temperature rises. D collision tectonics convergent plate margins. See also: orogeny, orogenic belts, subduction and decrepitation the breaking apart of a substance obduction. on heating. conchoidal the description of fractures in diadochic replacement the replacement of one brittle materials that do not follow any natural element in a crystal structure with another. planes of weakness. Often results in curved die casting a metal casting process that involves breakage surfaces. forcing metal into a mould cavity under high pressure. concordant intrusive igneous rock bodies which are parallel to bedding planes. The opposite is dielectric capacity the ability to hold an electric described as ‘discordant’. charge. conflict metals or minerals metals or minerals differentiates different types of igneous rocks which are extracted in areas experiencing conict formed by the process of magmatic differentiation 422 Glossary of technical terms in which a single magma body evolves over time electrodeposition an industrial process whereby to produce rocks of varying composition. particles are deposited from a liquid on an elec- trode using the inuence of an electrical field. digestion the dissolving of a substance in a hot uid. electroluminescent display material that emits light in response to the passage of an electrical diode semiconductor device which allows current. current to ow in one direction but not the other. electrolysis a method of using a direct electrical disseminated when a substance is evenly dis- current to drive a chemical reaction. tributed throughout another substance. Com- monly used in describing the distribution of ore electron beam refining a process that uses a minerals. high-energy electron beam as a heat source. divergent plate boundary the boundary between electron shell a grouping of electrons around an two plates which are moving away from each atom’s nucleus. The configuration of electrons in other. New crust is generated at these boundaries. the outermost shells determines the chemical Also known as a ‘constructive plate boundary’. properties of the atom. dopant a trace impurity element that is inserted electronegative elements elements that have into a substance to confer particular, usually the potential to attract electrons to themselves. electrical or optical, properties. electronegativity the tendency of an atom to ductility the ability of a solid material to deform attract electrons towards itself. under tensile stress. electrophoretic display an electronic display that is designed to mimic the appearance of ink on paper. E electrostatic concentrator a process that uses an elasticity the ability of a material to return to electrical charge to separate minerals. its original shape after deformation. electrostatic discharge the sudden ow of electrical conductivity a measure of how easily electricity between two objects; it can be caused an electrical current will pass through a material; by a build-up of static electricity or by electrostatic the opposite of electrical resistivity. induction. electrical resistance furnace a furnace that uses electrowinning an electrodeposition process the electrical resistance of a substance as the whereby an electrical current is passed from an means for creating heat. inert anode through the liquid leach solution con- taining the metal to a cathode; the metal is extract- electrical resistivity a measure of how strongly ed from the liquid and deposited on the cathode. a material opposes the ow of electricity; the opposite of electrical conductivity. eluvial placer deposit a residual type of placer deposit formed by the removal of lighter material. electrochemical potential see electrochemical series. emission spectrum the spectrum of frequencies of electromagnetic radiation (including light) electrochemical series a list of chemical ele- emitted by an element’s atoms or a compound’s ments in order of their standard potentials to a molecules. hydrogen electrode, (also known as their electro- chemical potential). The more negative the value end-of-life recycling rate the proportion of a the stronger reducing agent the metal is, and the metal that is discarded at the end of its life which more positive the stronger the oxidising agent. is actually recycled. Glossary of technical terms 423 epigenetic deposition or alteration by external fractional distillation the separation of a mix- agents after the rocks have formed. ture into its component parts by heating to a temperature at which one or more fractions will epithermal formed by warm waters at shallow vaporise. depths. fumarole precipitates fumaroles are openings in evaporite sedimentary rocks produced by the the earth’s crust, usually in the vicinity of vol- crystallisation of minerals from evaporating wa- canoes, which emit steam and other gases; on ter bodies, either the sea or a lake. cooling, contained elements, such as sulfur, may exhalative processes the venting of hydrother- be deposited to form new rock. mal uids on the sea oor. exothermic reaction a reaction that releases G energy; the opposite of endothermic reaction. gabbroic relating to a gabbro, a coarse-grained igneous rock composed of the minerals plagio- F clase feldspar, pyroxene and, commonly, olivine. felsic light-coloured minerals, such as quartz, geosphere the solid parts of the Earth. feldspar and feldspathoids, found in igneous geothermal field an area of the Earth’s crust that rocks. Also used to describe rocks which are rich is unusually hot. in these minerals. geothermal gradient the rate at which tempera- fenite high-temperature metasomatic rocks ture increases with depth in the Earth’s crust. composed mainly of potassium and sodium feld- spars. grade the concentration of a mineral or metal in its ore; expressed as a percentage content, or as ferromagnetic having a high susceptibility to ‘parts per million’ (ppm) or ‘grams per tonne’ (g/t). magnetisation, the strength of which depends on that of the applied magnetising field, and that gravity concentration the process of separating may persist after removal of the applied field. two mineral types based on their specific gravity. This is the kind of magnetism displayed by iron, It can be performed wet or dry, depending on the nickel and cobalt and many of their alloys. particular method used. flood basalt suite areas of basalt resulting greisen deposit mineral deposit comprising from extremely large volcanic eruptions which quartz and mica formed by post-magmatic meta- have covered the land or sea oor on a regional somatic alteration of granitic rocks. scale. greisenisation the formation of greisen, a highly flotation, also known as ‘froth flotation’ a process altered granitic rock composed of quartz and for separating hydrophobic and hydrophilic mate- muscovite mica, often rich in uorine. rials in a mineral concentrate. The concentrate and reagents are mixed in water to form a slurry H which is aerated to create bubbles; the hydropho- Hall effect the production of a voltage difference bic materials attach to the bubbles and rise to the across a conductor, transverse to an electric surface in a froth which can be skimmed off. current in the conductor and a magnetic field per- footwall rocks located immediately below a pendicular to the current. mineral deposit. Hall–Héroult process the main industrial pro- fractional crystallisation the precipitation and cess used for smelting alumina (aluminium oxide) segregation of minerals from magma. to produce aluminium metal. 424 Glossary of technical terms halogens the elements uorine, chlorine, bro- in situ leaching the extraction of metal from an mine, iodine and astatine. ore without the need for conventional mining. A leach solution is pumped via boreholes into the hard metal a mixture of cemented carbides con- ore and the solution containing dissolved metal is taining one or more of tungsten, titanium, tanta- pumped back to surface. See also: leaching. lum or vanadium embedded in a matrix of cobalt or nickel by sintering; widely used for cutting tools. interstitial lung disease a group of diseases affecting the tissue and space around the air sacs heap leaching metal extraction in which metals of the lungs. are leached from a heap of crushed ore by a perco- lating reagent. See also: leaching. intracontinental extensional basin depression in a continental plate where the forces of plate humic acid one of the principal components of tectonics are divergent; may lead to formation of soil, peat and coal produced by the biodegrada- a new rift system. tion of dead organic matter. ion an atom with either a positive electrical charge humin chemical compounds in soil, peat and (a ‘cation’) or negative electrical charge (an ‘anion’). coal which do not breakdown when treated with diluted alkali solutions. ion adsorption deposit a type of residual deposit in which elements released during the weathering hydrogen reduction a process that involves a of granites are adsorbed by kaolin and other clay molecule gaining hydrogen atoms. minerals. hydrolysis the decomposition of compounds by ion exchange reversible exchange of ions interaction with water; can also be carried out in contained in a crystal for different ions in solution the presence of acids or alkalis. without destruction of the crystal structure. It is hydrometallurgy processes that use aqueous sometimes carried out using resins which are small chemistry to recover metals. beads fabricated from an inorganic polymer. hydrosphere the combined mass of water found ionic radius the radius of an ion. on Earth. island arc typically a curved line of volcanoes hydrothermal related to processes involving hot forming islands near a convergent plate boundary uids. where both plates are oceanic. hydroxide a chemical compound containing an isotope forms of a particular chemical element oxygen atom and a hydrogen atom bound together. differing in the number of neutrons contained in their nucleus. hypabyssal, also known as ‘sub-volcanic’ igne- ous rocks that originate from shallow or medium I-type granite (or granitoid) derived from the depths within the crust. melting of mainly igneous rocks.

I K igneous rock rock types formed from the cooling komatiitic basalt a type of ultramafic mantle- and solidification of magma. derived volcanic rock containing high levels of magnesium but low levels of silicon, potassium immiscible liquids that cannot mix with each and aluminium. other. L incompatible element an element that because of its size or charge does not fit into the crystalline late magmatic fluids uids remaining in final structure of minerals. stages of crystallisation of a magma. Glossary of technical terms 425 laterite deposits formed by intensive weathering, magnetic concentration a process of separating usually in tropical settings, resulting in the two mineral types based on their magnetic prop- enrichment of some elements and the depletion of erties. others. magnetoencephalography a technique for map- leaching a widely used technique for extracting ping brain activity by recording magnetic fields metals from ores by converting the metal into produced by electrical currents occurring natu- salts soluble in an aqueous liquid; it can be car- rally in the brain. ried out in a number of ways, using either acid or malleability the ability of a solid material to alkaline liquids. See also: heap leaching; biole- deform under compressive stress. aching; in situ leaching; pressure leaching. matte molten metal formed during smelting leukocyte scintigraphy leukocytes are white processes. blood cells which form the body’s immune sys- tem; scintigraphy is a form of diagnostic test in mesothermal hydrothermal deposits formed at which radioisotopes are used to ‘tag’ cells so they intermediate temperature and pressure. can be detected using scanning detectors. metalloids also known as semi-metals; ele- ligand molecular groups bonded around a central ments which are in between metals and non- metal atom. metals in the Periodic Table and display prop- erties similar to both. Generally these include lignite a soft brown coal with characteristics boron, silicon, germanium, arsenic, antimony intermediate between those of peat and hard coal; and tellurium. it has a low calorific value and high content of water and volatile components. metaluminous a type of igneous rock where the aluminium oxide content is greater than the liquefaction the process of converting a solid content of sodium oxide and potassium oxide into a liquid by heating or lowering pressure, or of combined, but less than the content of calcium converting a gas to a liquid by cooling or raising oxide combined with sodium and potassium pressure. oxides. lithology a description of the physical charac- metamorphic rock rock types formed by the teristics of a rock unit. alteration of existing rock by heat and/or pressure. lithophile a group of elements that combine metamorphogenic deposits formed during meta- readily with oxygen. morphic processes as a result of high pressure lixiviant the liquid medium used in hydromet- and/or temperature. allurgy to selectively extract the desired metal metasedimentary sedimentary rock that shows from an ore or mineral. evidence of metamorphism. M metasomatic the process of chemical alteration of a rock by interaction with hydrothermal or mafic dark-coloured minerals such as olivine, other uids and replacement of one mineral by pyroxene and amphibole found in igneous rocks, another without melting. which are rich in iron and magnesium. The term can also be used to describe rocks which are rich metastable a state which is not the most stable in these minerals. (full equilibrium) and yet might be long lived. magmatic related to magma, molten rock and meteoric water water derived from rain or sur- uid originating deep within or below the Earth’s face water courses that circulates through the crust. rocks or is stored in pore spaces. 426 Glossary of technical terms meteoritic abundance the abundance of ele- ophiolite a section of the Earth’s oceanic crust ments in meteorites; for many elements this is and underlying upper mantle that has been em- commonly used as a best estimate of the abun- placed above sea level or onto continental crust dance of the elements in the entire solar system. by plate tectonic processes. mid-ocean spreading see actively spreading ridges. ore the part of mineral deposit comprising the metallic minerals that can be economically ex- mineral deposit an accumulation of a mineral, tracted. or group of minerals, of sufficient size and quality that may allow extraction under favourable organolithium molecules with a direct bond economic conditions. between carbon and lithium. mineral deposit model systematically arranged organophile elements that have an affinity for information describing the essential attributes, organic compounds. both descriptive and genetic, of a class of mineral orogenic belt a structurally complex tract of deposits. land comprising a mountain chain, or former mineral occurrence the presence in the Earth’s mountain chain, created during an orogeny. crust of a mineral, or group of minerals, which orogeny a period of mountain building; they may indicate a concentration of economic, occur at convergent plate boundaries where one scientific or technical interest. or both plates are continental. monochromator an optical device that trans- oxidation see oxidation state mits only a narrow band of wavelengths of light or other radiation. oxidation state the degree of oxidation of an atom in a chemical compound. An increase in N oxidation state through a chemical reaction is known as ‘oxidation’; a decrease is known as a non-stoichiometric chemical compounds with ‘reduction’. Oxidation state is recorded as an inte- an elemental composition that cannot be repre- ger and can be either positive, negative or zero. See sented by a ratio of whole numbers. also valency. noritic relating to a norite, an igneous rock oxide a chemical compound containing at least in which orthopyroxene is the dominant mafic one atom of oxygen and one other element. mineral. oxyhydroxide a mixed oxide and hydroxide. O P obduction the overthrusting of continental crust by oceanic crust or mantle rocks at a con- para-autochthonous rocks found close to the vergent plate boundary. See also: plate, plate mar- location of their formation. gins, plate tectonics. paramagnetic a form of magnetism that only oceanic plate plate of the Earth’s crust forming occurs in the presence of an externally applied the ocean oor; generally less than 10 km thick magnetic field. and composed largely of igneous rocks in layered pegmatite very coarse grained igneous rock, form. commonly of granitic composition. oceanic spreading see actively spreading ridges. peralkaline igneous rocks which have a higher offgas stream / offgas exhaust gases emitted dur- molecular proportion of combined sodium and ing industrial processes. potassium than aluminium. Glossary of technical terms 427 peraluminous a type of igneous rock where the power density the rate of energy transfer over a aluminium oxide content is greater than the given time per unit volume of material. content of calcium oxide, sodium oxide and power-added efficiency a metric for rating the potassium oxide combined. efficiency of a power amplifier. permanent magnet an object that is magnetised and creates its own persistent magnetic field. precipitation the formation of a solid from a uid because of a change in solubility, either due photometric sorting a method of concentrating to temperature variation or the introduction of minerals using optical properties to distinguish another substance. between the required mineral and gangue. pressure leaching a form of leaching carried out photovoltaic the method of generating electrical under increased pressure. power by converting solar radiation into direct electrical current, as used in photovoltaic cells. primary raw material material extracted from the geosphere. piezoelectricity electricity that builds up in solid substances, such as certain crystals and primitive magma primary melt from the man- ceramics, as a result of pressure. tle, unaffected by later processes such fractional crystallisation, magma mixing or contamination placer deposit an accumulation of valuable by crustal rocks. minerals formed as a result of differences in specific gravity. They can be alluvial, eluvial or protolith the original unmetamorphosed rock beach types. from which a given metamorphic rock is formed. plate the Earth’s crust is divided into large slabs pyroclastic deposits deposits of rock material of solid rock, known as plates, which are con- that has been expelled aerially from a volcanic stantly moving in relation to each other. See also: vent, such as agglomerate, tuff and ash. convergent plate boundary, divergent plate boundary, transform plate boundary. pyrometallurgy processes that use heat to bring about physical or chemical transformations that plate margin the edge of a plate, also known as enable the recovery of metals. a plate boundary. pyrophoric capable of igniting spontaneously plate tectonics collective term for the science in air. related to the moving of Earth’s plates. pyroxenite a rock containing a high concentration pneumoconiosis a restrictive lung disease often of the mineral pyroxene. caused by the inhalation of dust during mining. porphyritic an igneous rock texture comprising some large crystals surrounded by a finer-grained R groundmass. redox reduction-oxidation. See: oxidation states. porphyry deposit orebody associated with por- reducing agent the element or compound in a phyritic intrusive rocks and the uids associated reduction-oxidation reaction that donates an with them. See also: porphyritic. electron. powder metallurgy processes involving the reduction see oxidation state blending of finely ground materials, pressing them into a desired form and then heating them refining the purification of an impure metal, in a controlled atmosphere to bond the material commonly by hydrometallurgy or pyrometal- together. lurgy. 428 Glossary of technical terms refractive index a number that describes how semiconductor a material with electrical con- light, or other radiation, will pass through a ductivity intermediate between that of a conduc- particular substance. tor and an insulator. The conductivity may be affected by temperature, light, electric and refractory metals a class of metals which are magnetic fields or by the addition of small highly resistant to heat and wear. amounts of other materials (‘dopants’). replacement deposit mineral deposit formed by sericitisation a type of alteration which leads to chemical processes that remove certain elements the formation of sericite, a form of white mica. and leave others in their place. serpentinisation the conversion of ultramafic reserves the part of a resource which has been rock into serpentinite by low-temperature fully geologically evaluated and is commercially metamorphism. and legally mineable using current technology. resource nationalism the tendency of people shale a fine-grained, finely laminated, clastic and governments to assert control over natural sedimentary rock composed of clay minerals and resources located in their territory. The control tiny grains of other minerals, most commonly may be exerted in many ways, including nation- quartz. alisation or partial state ownership, imposition of siderophile a group of elements that have a trade restrictions through taxation, quotas, lev- strong affinity with iron. ies, etc. silicate a compound with structure based on resources a concentration of minerals or a body 4– anion of SiO4 . of rock that is, or may become, of potential economic interest for the extraction of a mineral siliciclastic rock silica-rich fragmentary rock. commodity. silicification the entire or partial replacement reverberatory furnace metallurgical furnace in of rocks with silica, in the form of quartz, chalce- which the material being processed is kept sepa- dony or opal. rate from the fuel but is exposed to the combustion sill a tabular sheet of igneous rock that has been gases and radiant heat. emplaced between layers of older rock. rift zone a region of the crust at a divergent skarn deposit mineral deposit consisting of plate boundary associated with a linear zone of metamorphic rocks formed as a result of chemical faulting and volcanic activity. alteration by hydrothermal or other uids. smelting the production of metal from its ore, S normally by heating with a chemical reducing secondary material material sourced from recy- agent, commonly carbon. cling. solid solution series a continuum of solid crys- sedimentary-exhalative (SEDEX) mineralisa- tals containing two or more chemical elements tion formed at or below the palaeo-seaoor by where each element can substitute for the other uids exhaling from underneath. within the crystal structure. sedimentary basin a geologically depressed area solvent extraction a method used to separate containing sedimentary rocks. compounds based on their relative solubilities in two immiscible liquids. sedimentary rock rock types formed from the settling and accumulation of particles or the pre- sputtering a process used in thin-film deposi- cipitation of material from solution. tion whereby atoms are ejected from a target Glossary of technical terms 429 material by bombarding them with particles; quartz is either absent or present in only small these atoms are deposited in a thin layer across a quantities. required surface. synform a downward bending fold; generally stockwork an irregular network of many cross- the term ‘syncline’ is only used when it can be cutting veins. demonstrated that the strata in the centre of the fold are younger. stratabound deposits where the mineralisation is confined to a single stratigraphic unit. syngenetic formed contemporaneously with rock formation. stratiform bedded or layered. synorogenic contemporaneous with orogenic stripping ratio the ratio of the amount of waste processes. rock removed to ore mined. synsedimentary deposit mineral deposit formed structurally controlled the localisation of a at the same time as the host sedimentary rock. mineral deposit as a result of geological structure, such as a fold or fault. syntectonic metamorphism metamorphism syn- chronous with, and resulting from, tectonic activity. S-type granite (or granitoid) derived from the melting of mainly sedimentary rocks. T subduction processes that occur at convergent plate margins where one plate (usually an oceanic technosphere see anthroposphere. plate) is forced below another. May result in tensile strength the maximum load a material formation of an oceanic trench and either an is- can withstand when being stretched. land arc (if both plates are oceanic) or orogenic belt (if one plate is continental). teratogenic a substance that may cause mal- functions of an embryo or foetus. sulfate a compound containing the sulfur-bear- 2- ing anion SO4 . thermal aureole a zone in the country rock surrounding an igneous intrusion within which sulfosalts a complex set of minerals which gen- metamorphic changes, mainly thermal, have erally contain a metal, semi-metal and sulfur. occurred. superalloy high-performance alloy with particular thermal conductivity a measure of how easily applications in aerospace and gas turbines due to heat is transferred through a material. their strength, resistance to creep, and resistance to wear and corrosion. thermoelectric device device that converts tem- perature differences to electrical voltage, and vice superconductor a material with zero electrical versa. resistivity; this happens to most metals when cooled below a certain, very low, temperature. tholeiitic a suite of igneous rocks which evolves by fractional crystallisation from basic rock supply chain the system of organisations, peo- types, such as basalt or gabbro, to acid rock types, ple, resources and activities that are required to such as rhyolite or granite, with marked enrich- move a product from the raw materials stage to a ment in iron. See also: calc-alkaline. finished product delivered to a customer. toll refining the refining of metal-bearing supracrustal deposited on existing basement concentrates, or scrap, for which the refinery is rocks. paid a fee but the ownership of the metal recov- syenite a coarse-grained alkaline igneous rock ered remains with the owner of the feed with the same composition as granite except that material. 430 Glossary of technical terms transform plate boundary the boundary bet- vapour pressure the pressure exerted by a ween two plates which are moving sideways vapour when it is in equilibrium with its solid or past each other along ‘transform’ faults. Crust is liquid condensed phases at a given temperature neither destroyed not created at these bound- in a closed container. aries. Also known as a ‘conservative plate vein deposit mineral deposit in narrow, elon- boundary’. gate cavities of the parent rock such as cracks or transition metal elements in Groups 3 to 12 of fissures. the Periodic Table. volatile in a chemical context, the tendency of a substance to vaporise. U volcano-sedimentary sequence a mixture of ultrabasic an igneous rock composed of less rocks of volcanic and sedimentary origin laid than 45% silica (SiO ). 2 down in layers. ultramafic igneous rocks containing more than 90% mafic minerals. W wafer also known as ‘substrate’; a thin slice of V semiconductor material used in the fabrication of vacuum arc melting a form of pyrometallurgy integrated circuits or other electronic devices. carried out in a vacuum arc furnace. Washington consensus a term used since 1989 valency the number of chemical bonds formed to describe 10 relatively specific economic policy by the atoms of a given element; monovalent prescriptions considered as ‘standard’ by Washing- (single bond), divalent (two bonds), trivalent ton D.C.-based institutions such as the (three bonds), etc. International Monetary Fund and the World Bank. value chain the series of processes through X which a product passes on the way to a customer; at each stage additional value is added to the xenolith a fragment of one type of rock enveloped product. by a different rock type during the latter’s formation. Index

Alaskan/Alaskan-Ural type global production 87–90 autocatalysts 297–9, 307, 309 deposits 289 gold–antimony ore 72–7 AZO see aluminium-doped zinc oxide alkaline granites and syenites 363, greenstone-hosted quartz-carbonate 367–8 vein deposits 76–8 basalt deposits 288, 292–3 alkaline igneous rock deposits 319, historical uses 70 Basel Convention 60 323–4 hot spring deposits 74, 76–7 bastnäsite 315, 327 Alpine-type (APT) deposits 181–2, 184 major deposit classes 72–8 batteries aluminium mineralogy 70–72 antimony 85–6 beryllium alloys 102, 103, 113, 116 mining methods 78–9 cobalt 140, 142–3, 145–6 gallium recovery 151, 153 polymetallic base metal vein gallium 155 magnesium alloys 269, 280 deposits 74, 76 lithium 239–41, 244, 255–8 mining industry 21 prices 92–4 rare earth elements 330 recycling and reuse 52 processing, beneficiation and bauxite 151, 153 resource criticality 6, 8 conversion to metal 79–82 Bayer process 151, 153, 155 aluminium-doped zinc oxide (AZO) 224 projects under development 90–91 BEE see black economic empowerment ammonium metatungstate (AMT) 395, recycling and reuse 57, 85, 88 Beja process 153 398 reduced magmatic deposits 74, 76, 78 bertrandite 106, 109–111 ammonium paratungstate (APT) 394, resources and reserves 86–7, 90–91 beryl 106, 107, 110 395, 398–9, 406–7 size and grade of deposits 74 beryllia ceramics 100, 103 ammonium perrhenate (APR) 352 specifications 82 beryllium 99–121 ammonothermal 160 substitution 86 beryllia ceramics 100, 103 AMT see ammonium metatungstate world trade 91–3 beryllium oxide production from anode slime residues 212 antimony trioxide (ATO) 82, 84, 95 beryllium hydroxide 113–15 antimony 70–98 apartheid 34 contemporary uses 100–103, 118–19 abundance in the Earth 71 apatites 317, 324 deposit characteristics 107–10 carbonate replacement deposits 72, Apex Mine, Utah 152 discovery and isolation 99 76–8 APR see ammonium perrhenate distribution and abundance in Earth’s contemporary uses 70, 82–5 APT see Alpine-type (APT) deposits; crust 100 definitions and characteristics 70 ammonium paratungstate environmental factors 116–18 deposit characteristics 76 Araxá deposit 363, 365–6, 379 future developments and environmental factors 94–5 arsenide ores needs 118–19 epithermal deposits 72–7 antimony deposits 77 global distribution 107, 108 future developments and needs 96 cobalt 122–3, 125, 138 global production 103–5 global consumption 84 artisanal mining 22, 371–2 high-beryllium alloys 102–3 global distribution 73, 75, 86 ATO see antimony trioxide hydrothermal deposits 110

Critical Metals Handbook, First Edition. Edited by Gus Gunn. © 2014 John Wiley & Sons, Ltd. Published 2014 by John Wiley & Sons, Ltd. 432 Index

beryllium (cont’d ) carbonatite deposits magmatic deposits 129–30, 134 low-beryllium alloys 100, 102 niobium and tantalum 363–6 manganese nodules 132, 140 metal and alloy production from platinum-group metals 289 mineralogy 122–3 beryllium hydroxide 113–15 rare-earth elements 319–23 mining industry 27, 28, 34 metal preparation 269 carbo-thermic reduction 266–7 physical and chemical mineralogy 106–7 carcinogenicity 117 properties 122–3 mining methods 110–11 cars 43–4, 58–60 prices 144–5 pegmatite deposits 107–10 casting cover-gases 273–4 recycling and reuse 43, 57, 138 prices 118 catalysts resources and reserves 139–40 processing of ores to beryllium cobalt 142 specification 140–142 hydroxide 111–12 germanium 189–91 substitution 142–3, 146 properties 99–100, 102–3 platinum-group metals 297–9, 307, world trade 139, 141 pure beryllium metal 102–3 309 cobalt–chromium alloys 122 rare element zoning 107, 109 rare earth elements 330 collection schemes 47–8, 52–4 recycling and reuse 57, 113–16 recycling and reuse 43 columbotantalite 362–3, 371 resources and reserves 106–7 rhenium 340–341, 352–6, 358 combustion 81 substitution 116 CBD see chronic beryllium disease comminution 294–5 world trade 105 CCZ see Clarion–Clipperton Zone compact fluorescent lightbulbs BGS see British Geological Survey Central African Copperbelt 126–8, (CFL) 161 Bielfeld and Laspeyres process 153 135–40, 142, 198 computers see Waste Electrical and bismuth alloys 215–16 CFL see compact fluorescent lightbulbs Electronic Equipment black economic empowerment chalcopyrite deposits 206 concentrator photovoltaics (CPV) 162, (BEE) 28 Chromitite reef type deposits 286, 288, 170 blast furnaces 85–6 291 ConRoast process 296 Bou Azer, Morocco 125 chromium 34 construction minerals 1 breccia deposits 391 chronic beryllium disease contact type deposits 288, 291 British Geological Survey (BGS) (CBD) 116–18 continental brines 232–4, 236–8, 241–2, beryllium 103–5 CIGS see copper indium gallium 249–51, 392 cobalt 138, 140 selenium Copaux–Kawecki process 111 lithium 247 CIM see Canadian Institute of Mining, copper and copper ores rare earth elements 333–4 Metallurgy and Petroleum beryllium alloys 100, 101, 113, 115, Bushveld Complex 286, 290–292 CIS see copper indium selenide 118–19 by-product metals civil war see geopolitical risk cobalt recovery 125–7, 129–30, 135–6 cobalt 123, 125 Clarion–Clipperton Zone (CCZ) 132 germanium recovery 178–9, 181–3 germanium 193–4 closed-loop recycling systems 58–60, indium recovery 206, 208–14, 217 indium 226–7 62, 66, 300–301, 309 mining industry 26–31, 35 mining industry 28–9 coal deposits 185–6, 188, 194, 197, 199 recycling and reuse 41–2, 52 rare earth elements 326 cobalamin 143 resource criticality 6–7 recycling and reuse 46 cobalt 122–49 rhenium recovery 342–50 resource criticality 6–8 contemporary uses 122, 140–142 tungsten alloys 397 deposit characteristics 123–34 copper–indium–gallium–selenide cadmium distribution and abundance in the (CIGS) 155, 163–4, 166–7, 171, 216 alloys 215–16 Earth 122 copper–indium–selenide (CIS) 163, 216 cadmium telluride 163 environmental factors 143–4 CPV see concentrator photovoltaics mining industry 28 extraction, processing and criticality see resource criticality Canadian Institute of Mining, refining 134–8 cyanidation 81 Metallurgy and Petroleum ferromanganese oxyhydroxide Czochralski (CZ) method 155–6, 188–9 (CIM) 4 crusts 123, 132–4, 138 capacitors 375, 382 future developments and needs 144–6 de la Bretèque process 153 capital investment 24, 30–31 global distribution 123–4 Defence Advanced Research Projects carbon nanotubes 224 global production 138–40 Agency (DARPA) 170 carbonate-hosted zinc-lead historical uses 122 demand rationing 26 deposits 182–4 hydrothermal deposits 123–9, 135–6 depleted uranium 399 carbonate ores 316 iron-oxide–copper–gold deposits 125, deuterium-tritium (DT) reaction 257–8 carbonate replacement deposits 74, 129 didymium 312 76–8 laterite deposits 130–132, 134–5 dietary ingestion 143–4 Index 433

dismantling processes 47–8, 51–4 fibre optics 189–91, 199 sphalerite deposits 151–2 disseminated deposits 390 Financial Stability Act (2010) 377 substitution 163 dolomite deposits 366 Fischer–Tropsch process 354, 358 world trade 167–9 DT see deuterium-tritium (DT) reaction flame retardants 83, 85 gallium-doped zinc oxide ductile-to-brittle transition flash smelting 134–5 (GZO) 170–171 temperature 341 flotation processes gallium–indium–tin alloys 216–17 dunite pipes 288, 291–2 germanium 186, 188 galvanic corrosion 271–2 platinum-group metals 294–5 gas turbines 353–4, 356–8 economic forecasting 24, 25, 39 rare earth elements 326 geopolitical risk economic viability 3–5 fluorcarbonate ores 316 cobalt 142, 144 electric vehicles (EV) 240, 244, 255–7 fluoride ores 316 magnesium 261 electrical insulation 103 fluorite 317 mining industry 26, 28, 32–4, 37 electrodeposition 81–2 fluorspar 37 niobium and tantalum 376–7, electrolytic processes fluxes 239, 271 380–381 indium 212, 214 fly ash 152 platinum-group metals 308–9 lithium 239–40 forensic fingerprinting 377 policy attractiveness 34–6 magnesium 263–4, 267, 272–3, Frary and Pechiney process 153 recycling and reuse 43 279–80 fuel cells 307–8, 328 geothermal brines 234, 251 End-of-life Vehicles (ELV) fusion reactors 103 germanium 177–204 Directive 47, 60 accumulation in sulfide energy minerals 2 gallium 150–176 deposits 181–5 environmental factors bauxite deposits 151 contemporary uses 189–91, 199–200 antimony 94–5 contemporary uses 153–63 deposit characteristics 179–86 beryllium 116–18 deposit characteristics 151–2 discovery 177 cobalt 143–4 discovery 150 distribution and abundance in the gallium 163–4 environmental factors 163–4 Earth 177–8 germanium 192 future developments and enrichment in lignite and coal 185–6, indium 225–6 needs 166–7, 170–172 188, 194, 197–9 lithium 241 gallium antimonide 157 environmental factors 192 magnesium 272 gallium arsenide 153–9, 163, 166, extraction methods 186 mining industry 25 224–5 first reduction metal 188 niobium and tantalum 376 gallium chemicals 159–60 future developments and platinum-group metals 301–2 gallium metal 157 needs 196–7, 199–200 rare earth elements 330–331, 336–7 gallium nitride 153, 160–163, 170 germanium dioxide 188, 197 recycling and reuse 42, 43 gallium phosphide 162 germanium tetrachloride 188 resource criticality 14 global distribution 150–151 global distribution 179–81 rhenium 356 global resources and global production 194–6 tungsten 399–400 production 164–7 mineralogy 178–9 epitaxial layers 156–7, 160 IREDs, LDs and LEDs 158–64, mining industry 28, 37–8 epithermal deposits 72–7, 210–211 170–172 physical and chemical equity financing 22–3 mineralogy 150–151 properties 177–8 essential nutrients 143–4 mining industry 29–30, 37 prices 197–8 ETFs see exchange-traded funds photovoltaics 162–4, 170–171 processing and beneficiation 186–8 European Commission (EC) 12–13, 35 physical and chemical recycling and reuse 47, 57, 189–91, eutectic alloys 157 properties 150–151 199 EV see electric vehicles prices 167–9 resources and reserves 192–4 evaporite deposits 392 primary production 164–5 single crystals 188–9 exchange-traded funds (ETFs) 298, 307 primary recovery 152–3 specifications 188–9 exploration companies 22–3 recycling and reuse 41, 57, 153–5, substitution 191–2 165–7 supply and demand 198–200 ferberite 386 refining and purification 155 world trade 197 ferromanganese oxyhydroxide resource criticality 6 zone-refined metal 188 crusts 123, 132–4, 138 secondary production 165–6 GHG see greenhouse gases ferro-niobium 376, 380 secondary recovery 153–5 global diversified miners 21 ferrotungsten 396, 406 semiconductors 153–60, 163 Global Mineral Resource Assessment fibre glass 308 specifications 157–63 Project 6 434 Index

gold ICGLR see International Conference on industrial minerals 1–2 antimony deposits 73–80, 90 the Great Lakes Region infrared emitting diodes (IREDs) 158–9 antimony ores 73–7 Idaho Cobalt Belt (ICB) 125, 138 infrared optics 189–91, 199 cobalt ores 125, 129 IMA see International Mineralogical inhalation exposure 117 mining industry 21, 24 Association intermetallic compounds 101 recycling and reuse 41–2, 47–8 IMDG see International Maritime International Conference on the Great granite deposits 363, 367–8, Dangerous Goods Lakes Region (ICGLR) 377 370–371 imperial smelting process (ISP) 212 International Maritime Dangerous granitic pegmatites 107–10 indigenisation 32 Goods (IMDG) Code 376 gravity methods 394 indium 204–30 International Mineralogical Association greenhouse gases (GHG) 272–5, 302 abandoned production 221–2 (IMA) 178 greenstone-hosted quartz-carbonate abundance in Earth’s crust 205 international reporting codes 4 vein deposits 76–8 alloys and solders 215–16 IOCG see iron-oxide–copper–gold greisen deposits 390 base-metal-rich epithermal ion adsorption deposits 322–4 Grignard reagents 269 deposits 210–211 ion-exchange processes 238 GZO see gallium-doped zinc oxide base-metal-rich tin–tungsten and IREDs see infrared emitting diodes skarn deposits 210 iridium see platinum-group metals Harris Process 213 base-metal sulfide deposits 209 Irish-type (IRT) deposits 181–2, 184 HB see horizontal Bridgman (HB) contemporary uses 204, 214–17 iron and iron ores method deposit characteristics 206–10 germanium recovery 179, 182–3 hectorite clays 234–5, 248–9, discovery 204 indium recovery 206 252–3 environmental factors 225–6 mining industry 36 high nitrogen pressure solution extraction methods and rare earth elements 325 (HNPS) 160 processing 210–214 recycling and reuse 52 high-pressure acid leach (HPAL) future developments and needs 226–7 resource criticality 6, 8 technology 135, 138 global distribution 207 iron-oxide–copper–gold (IOCG) high-speed steel (HSS) 397 global production 218–20 deposits 125, 129, 324 HNPS see high nitrogen pressure historical uses 204 IRT see Irish-type (IRT) deposits solution mineralogy 205–6 isotropy of beryllium 102 Hoboken universal process 54–5 mining industry 28, 37 ISP see imperial smelting process horizontal Bridgman (HB) mining methods 210–212 ITO see indium–tin oxide method 155–6 physical and chemical hot spring deposits properties 204–5 jadarite 235–6, 244, 253 antimony 72, 76–7 polymetallic vein-type deposits 209, jarosite 152, 212 indium 211 211 jewellery 298 tungsten 392 prices 223–4 Joint Ore Reserves Committee HPAL see high-pressure acid leach processing, beneficiation and (JORC) 4 HSS see high-speed steel conversion to metal 212–14 Joint Research Council (JRC) 13 hübnerite 386 production from copper ores 213–14 hydride vapour phase epitaxy production from tin ores 214 Kambalda Dome 129 (HVPE) 160 projects under development 221 Kipushi-type (KPT) deposits 181–2, hydrometallurgical processes recovery from secondary sources 214 184, 196–7 antimony 80, 81 recycling and reuse 41, 42, 46, 51, 57, Kjellgren–Sawyer process 111 cobalt 134 220–221, 224 Komatiite related deposits 288, 293 germanium 187 resource criticality 5, 8 KPT see Kipushi-type (KPT) deposits lithium 240 resources and reserves 217–18 Kroll process 269 platinum-group metals 296 semiconductors 215, 216 recycling and reuse 47–8, 54, 56 specifications 214–15 Landsat images 235 tungsten 394 substitution 224–5 lanthanides see rare earth elements hydrothermal deposits world trade 222–3 laser diodes (LDs) 158–61 antimony 72, 74, 76, 77 indium–boron oxide 216 laterite deposits 130–132, 134–5, 289 beryllium 110 indium gallium aluminium LCA see life-cycle analysis cobalt 123–9, 135–6 phosphide 162 LCD see liquid crystal displays platinum-group metals 289 indium gallium phosphide 162 LCEs see lithium carbonate equivalents rare earth elements 317, 319–24 indium–tin oxide (ITO) 46, 51, 54, 171, LCI see life-cycle inventories hypothetical resources 4 204, 215, 223–6 LCT see Lithium-Caesium-Tantalum Index 435

LDs see laser diodes pegmatites 232, 236–8, 241–9, 255 twentieth century production leaching prices 254–5 processes 266–7 antimony 81 production costs 248–9 world trade 277–8 beryllium 111, 112 properties 230, 232 magnet alloys 140–142, 157 cobalt 134–8 recycling and reuse 240–241 magnet ores 328–9, 334 germanium 186 resources and reserves 241–5 magnetic separation 394 lead and lead ores salars 231–8, 241–4, 249–51 magnetite 179 germanium 181–4 specifications and uses 238–40, manganese indium alloys 215–16 254–8 ferromanganese oxyhydroxide recycling and reuse 52 substitution 240–241 crusts 123, 132–3, 138 resource criticality 8 terminology 230 steel alloys 34 LEC see liquid encapsulated world trade 253–4 manganese nodules Czochralski lithium carbonate equivalents cobalt 133, 140 LEDs see light emitting diodes (LCEs) 230 germanium 179 lepidolite 232, 247 lithium niobate 374 rare earth elements 325 life-cycle analysis (LCA) 274–5, lithium tantalate 374 market capitalisation 21, 22 399–400 Lithium-Caesium-Tantalum (LCT) matte-smelting process 296 life-cycle inventories (LCI) 273 association 107 MBE see molecular beam epitaxy light emitting diodes (LEDs) LIX see liquid ion exchange medical applications gallium 155, 158–9, 161–4, London Metal Exchange (LME) 144 indium 217 170–172 LPE see liquid phase epitaxy platinum-group metals 299–300, 308 germanium 199 lubricants 239 MEG see Metals Economics Group indium 216 memory and storage applications 200, lignite deposits 185–6, 194 magmatic deposits 299–300 limonite 152 cobalt 129–30, 134 mercury 78 liquid crystal displays (LCD) platinum-group metals 292 Merensky Reef type deposits 286, 288, gallium 161–2, 170 rhenium 346, 349–50 290–291, 295 indium 215, 221, 223–7 magnesium 261–83 metal organic vapour phase epitaxy recycling and reuse 46, 51 casting cover-gas issues 273–4 (MOVPE) 157, 160 liquid encapsulated Czochralski (LEC) contemporary uses 261, 267–9, metallothermic reduction 263, 265–7, method 155–6 280–281 273–5, 279 liquid ion exchange (LIX) 350–351 deposit characteristics 263 metallurgical metals recovery see liquid phase epitaxy (LPE) 157, 166 distribution and abundance in the hydrometallurgical processes; lithium 230–258 Earth 262 pyrometallurgical processes abundance in the Earth 230 electrolytic processes 263–4, 267, Metallurgical Miners’ Association of chemical and non-chemical 272–3, 279–80 China (MMAC) 36 demand 238 environmental factors 272 Metals Economics Group (MEG) 23 continental brines 232–8, 241–2, extraction methods, processing and metals life cycle 45–6, 62, 63 249–51 beneficiation 263–7 meteorite impacts 292, 293 current producers 245–8 future developments and needs 277, microlite 363 discovery 230 279–81 military applications 158–9 environmental factors 241 global production 275–7 mineral commodities, definition 2 extraction methods and life-cycle analysis 274–5 mineral occurrences, definition 2 processing 236–8 metallothermic reduction 263, mineral sands 324, 326 future developments and 265–7, 273–5, 279 mining industry 20–39 needs 249–53, 255–8 mineralogy 262–3 artisanal mining 22 geothermal brines 234, 251 nineteenth century production capital and operating costs 24–5, 30 global distribution 231 processes 266 constraints on mineral supply global production 244–5 physical and chemical response 27–36 hectorite clays 234–5, 248–9, 252–3 properties 261–2 demand rationing 26 jadarite 235–6, 244, 253 prices 277, 279 economic constraints 29–31 lithium carbonate 239–40, 253–4 recycling and reuse 269–71 economic forecasting 24, 25, 39 mineralogy and deposit resources and reserves 275–7 environmental factors 25 characteristics 230–232 specifications 267–8 geopolitical risk and policy mining industry 24 substitution 271–2 attractiveness 33–5, 38 oilfield brines 234, 251–2 supply security 261 industry dynamics 23–7 436 Index

mining industry (cont’d ) neodymium–iron–boron magnets 328, indium 226 institutional constraints 31–4 334 platinum-group metals 301 market capitalisation 21–2 neutron moderation 102 Oddo Harkins effect 314 market size and market share 36 nickel OECD see Organisation for Economic miners and explorers 21–3 beryllium alloys 113, 115 Cooperation and Development nationalisation 32 cobalt ores 129–35 oilfield brines 234, 251–2 natural constraints 27–8 mining industry 26, 27, 36 ongonites 110 policy issues 38–9 platinum-group metal alloys 308 open-loop recycling systems 58–9, prospect development 24 recycling and reuse 52 61–2, 300, 309 resource criticality 20, 24, 34–8 resource criticality 7, 8 operating costs 25–6, 30–31 role of China 34–8 tungsten alloys 397 ophiolites 289 securitisation 20, 32 nickel–copper-dominant deposits 288, Organisation for Economic Cooperation supply and demand 20–21, 26–7 292–7, 309 and Development (OECD) 377 Minor Metals Trade Association niobium 361–84 osmium see platinum-group metals (MMTA) 82, 352, 357 alkaline to peralkaline granites and oxide ores Mississippi Valley-type (MVT) syenites 363, 367–8 antimony 79, 81 deposits 181–4, 196 carbonatite deposits 363–6 germanium 179, 181 MLCC see multi-layer ceramic deposit characteristics 363–71 rare earth elements 316 capacitors discovery 361 MMAC see Metallurgical Miners’ distribution and abundance in the palladium see platinum-group metals Association of China Earth 361–2 peak production concept 9 MMIC see monolithic microwave environmental factors 376 pegmatite deposits integrated circuits extraction methods and beryllium 107–10 MMTA see Minor Metals Trade processing 371–4 lithium 232, 236, 241–9, 255 Association future developments and niobium and tantalum 363 mobile phones 158, 239 see also Waste needs 379–82 tungsten 392 Electrical and Electronic geopolitical risk 376–7, 380–381 peralkaline granites and syenites 363, Equipment global distribution 364 367–8 modifying factors 5 global production 377–80 peraluminous granites 363, 370–371 molecular beam epitaxy (MBE) 157, 160 mineralogy 362–3 peraluminous pegmatites 363, 368–70 molybdenites 342–5, 349–50 mining industry 28 PET see polyethylene terephthalate molybdenum 28–29, 342–5, 347–50 peraluminous granites 363, 370–371 petalite 232 monazite 315–17, 323, 326–7 peraluminous pegmatites 363, PGM see platinum-group metals monolithic microwave integrated 368–70 phosphate ores 316 circuits (MMIC) 158–9 physical and chemical photovoltaics MOVPE see metal organic vapour phase properties 361–2 gallium 162–4, 170–171 epitaxy prices 380–381 germanium 190–192 multi-layer ceramic capacitors recycling and reuse 43, 57, 375 indium 221, 223 (MLCC) 299, 301 resources and reserves 377–80 platinum-group metals 308 multi-stage solvent extraction 327–8 specifications and uses 374–5 Pidgeon process 263, 265–7, 273–5, 279 MVT see Mississippi Valley-type (MVT) substitution 375 pipe deposits 392 deposits niobium-yttrium-fluorine (NYF) placer deposits 289, 324, 325, 392 association 109 platinum-group metals (PGM) 284–309 National Bureau of Statistics of China NIOSH see National Institute of cobalt 129–30 (NBSC) 400 Occupational Safety and Health contemporary uses 284, 297–300, National Institute of Occupational non-ferrous tungsten alloys 396 307–9 Safety and Health (NIOSH) 95 Norilsk region 130 deposit characteristics 285–94 National Research Council (NRC) NORMs see naturally occurring environmental factors 301–2 lithium 240–241 radioactive materials extraction methods 294 mining industry 35 nuclear weapon pits 157 future developments and needs 306–9 resource criticality 12, 14 NYF see niobium-yttrium-fluorine global distribution 287 nationalisation 32 (NYF) association global production 302–4 naturally occurring radioactive mineralogy 285–6 materials (NORMs) 376 occupational health mining industry 27–8, 35 NBSC see National Bureau of Statistics antimony 95 nickel–copper-dominant of China beryllium 117 deposits 288, 292–7, 309 Index 437

nomenclature 284 future developments and needs 329, lifecycle structures 58–62 PGM-dominant deposits 286–92, 332–3, 336–7 lithium 239–40 294–6, 309 global distribution 318 magnesium 269–71 prices 306–7 global production 331–3 metallurgical metals recovery 47–8, processing methods 294–7 inclusion of yttrium and 54–7 properties and abundance in the scandium 312 metals life cycle 45–6, 62 Earth 284–5 ion adsorption deposits 319, 324 mining industry 26 recycling and reuse 41–4, 46, 49, 51, iron-oxide-apatite deposits 316–17, mining and recycling as 54, 57–62, 300–301, 309 324 complementary systems 64–6 resource criticality 5, 8, 13, 14, 16 IUPAC definition 312 mobile phones 42–3 resources and reserves 302 lanthanide contraction 313 niobium and tantalum 375 specifications 297 mineralogy 315–17 platinum-group metals 300–301, 309 substitution 301 mines and advanced projects 317, rare earth elements 328–31 supply security 308–9 319–22 rationale and benefits 41–3 world trade 304–6 mining industry 24, 25 recyclability factors 48–50, 62, 63 policy attractiveness 33–5 mining methods 325 recycling challenges 48–50 polyethylene terephthalate (PET) physical and chemical recycling rates and improvement antimony 83–5, 95 properties 312–5 pathways 61–2 germanium 189, 191, 199 placer deposits 324, 326 recycling technologies 51–8 polymetallic vein deposits prices 334–6 recycling value chain 47–8, 51–7 antimony 72, 76 processing and beneficiation 325–8 resource criticality 9–10, 16 indium 209, 211 recycling and reuse 41, 54, 58, resource security as societal driver 64 porphyry deposits 328–31 rhenium 354–5 germanium 182–3 resource criticality 5–8, 14, 16 seven conditions of effective indium 208, 211 resources and reserves 331–3 recycling 50–51 platinum-group metals 289 seafloor deposits 325 status and challenges 45–51 rhenium 342–8 separation from each other 312 tungsten 398–9 tungsten 390–391 size and grade of ore deposits 317, urban mine concept 41–3, 66 pre-processing 47–8, 51–4 319 waste and resource legislation 47 primary resources 9–10 specifications and uses 328–9 reduced magmatic deposits 74, 76, 78 prospect development 24 substitution 330 REE see rare earth elements pyrochlore 362–3, 366–7, 371–2 world trade 333–4 reserve base, definition 5 pyrometallurgical processes rare element zoning 107, 109 reserves, definition 5 antimony 79, 81 Raw Materials Group 36 resource classification 4 cobalt 134 Raw Materials Initiative 62 resource criticality 1–18 germanium 187–8 REACH guidelines 376 assessments of criticality 11–15 lithium 239 recycling and reuse 41–67 construction minerals 2 recycling and reuse 47–8, 54–7 antimony 58, 85, 88 criticality concept 10–18 tungsten 394–5 automotive PGM applications 60 crustal abundances 2 beryllium 57, 113–16 definitions and terminology 3–4 quantum dots 224 cars 43 demand and usage patterns 6 cobalt 142 economic viability 4–6 rare earth elements (REE) 312–39 dedicated processes for metal energy minerals 2 alkaline igneous rock deposits 319, combinations 56–7 geological assessment 6 323–4 electronic PGM applications 60 geology and technology of by-products, co-products and waste EU critical metal recycling metals 1–5 products 325 status 57–8 improving criticality carbonatite-related deposits future developments and needs 62–6 assessment 14, 15 317–23 gallium 153–5, 165–7 industrial minerals 1–2 deposit characteristics 317–25 germanium 189–91, 198–9 key concepts 1–4 didymium 312 global end-of-life recycling rates 49 metals 2 distribution and abundance in the improvements in primary production mining industry 20, 24, 34–8 Earth’s crust 313–15 and wastes 45 peak production concept 9 environmental factors 330–331, indium 220–221, 224 policy implications 16 336–7 industrial PGM applications 59–60 recycling and reuse of extraction and separation 327–8 innovation needs 62–4 metals 9–10, 16 438 Index

resource criticality (cont’d ) sediment-hosted massive sulfides (SMS) Sudbury Igneous Complex resources and reserves 4–5 germanium 181–2, 184, 196 (SIC) 129–30, 292, 293 supply and demand 6–9 indium 208–9, 211, 217–18, 222 sulfide ores Resource Super Profits Tax 32 selenium 28 antimony 79–81 reverberatory furnaces 85 semiconductors cobalt 122–3, 125, 134–6 rheniite 341–2 gallium 153–60, 163 gallium 151–2 rhenium 340–360 germanium 189, 199 germanium 178–9, 181–5, 193, 197 contemporary uses 340, 352–4, indium 215, 216 indium 205–6, 209 356–9 tungsten 398 platinum-group metals 285–6 deposit characteristics 342–6 shales 289 rhenium 341–2 discovery 340 SIC see Sudbury Igneous Complex sulfur hexafluoride 273–4 distribution and abundance in the silicates 178–9, 316 superalloys Earth’s crust 341 silico-thermic reduction 267 cobalt 140–142, 145–6 environmental factors 356 silver 42–3, 49 platinum-group metals 308 extraction methods, processing and skarn deposits 210, 389–90 rhenium 352–5, 359 beneficiation 350–351 skutterudite 123 tungsten 396, 397, 409 future developments and slag/tailings recovery supply risk 12–14 needs 348–9, 358–9 antimony 79–81, 85 surface-coating applications 216 global distribution 347 cobalt 136–7 syelite deposits 363, 367–8 global production 340, 346–50 germanium 185, 192 synorogenic hydrothermal magmatic nickel–copper–PGE indium 213 deposits 77–8 deposits 346, 349–50 niobium and tantalum 372–3, 379 mineralogy 341–2 recycling and reuse 45, 57 tailings see slag/tailings recovery mining industry 28, 35 SMS see sediment-hosted massive Tanco deposit 368–70 physical and chemical sulfides tantalum 361–84 properties 340–341 sodium antimonate 84–5 alkaline to peralkaline granites and porphyry deposits 342–8 solid-bed ion exchange 350–351 syenites 363, 367–8 prices 352–3, 355, 357–8 solid-state lighting (SSL) 161–2 carbonatite deposits 363–6 recycling and reuse 354–5 solution-annealing 113 deposit characteristics 363–71 resources and reserves 346–50 solvent extraction 111–12, 327–8 discovery 361 sediment-hosted copper speculative resources 4 distribution and abundance in the deposits 345, 348 sphalerite deposits 151–2, 183, 205–6 Earth 361–2 specifications 352–4 spodumene 232, 236, 243–5, 254 environmental factors 376 substitution 355–6 SSL see solid-state lighting extraction methods and uranium deposits 346, 349 State Reserves Bureau (SRB) 35–6 processing 371–4 vein molybdenum deposits 343, static life index 27 future developments and 345 stibnite 70–73, 78–81 needs 379–82 world trade 356–7 stock market financing 23 geopolitical risk 376–7 rhodium see platinum-group metals stockpiling global distribution 364 rift-and continental-flood basalts 288, antimony 89–90 global production 377–80 292–3 beryllium 99, 101–3, 111 mineralogy 362–3 ruthenium see platinum-group metals China and criticality 34–6 mining industry 28 cobalt 137 peraluminous granites 363, 370–371 salars 232–5, 237, 238, 241–4, 249–51 germanium 185 peraluminous pegmatites 363, samarium–cobalt magnets 328 indium 223, 227 368–70 scandium see rare earth elements lithium 246 physical and chemical scheelite 386, 390–392 niobium and tantalum 372, 379 properties 361–2 Schwenzfeir process 113, 114 policy issues 39 prices 380–381 SEC see Securities and Exchange rare earth elements 331–2 recycling and reuse 57, 375 Commission recycling and reuse 62 resources and reserves 377–80 secondary resources 9 resource criticality 11, 14, 34–7, 39 specifications and uses 374–5 Securities and Exchange Commission rhenium 354, 357 substitution 375 (SEC) 377 tungsten 404 taxation 34 securitisation 20, 32 stockwork deposits 387, 389 TCOs see transparent conducting sediment-hosted copper stratabound deposits 391–2 oxides deposits 345, 348 structural scarcity 7 technical scarcity 7 Index 439

tellurium 7–8, 28, 35 specifications 395–6 rhenium 343, 345 TFSC see thin-film solar cells steel and other alloys 397–8, 409 tungsten 387, 389 TFT see thin-film transistor stratabound deposits 391–2 vertical Bridgman (VB) method 155–6 thermometers 157 substitution 399 vertical gradient freeze (VGF) thin-film deposition 157 vein and stockwork deposits 387, 389 method 155–6 thin-film solar cells (TFSC) 163, 167, world trade 404–8 vitamins 143–4 223 tungsten carbides 375, 395–7 VMS see volcanogenic massive sulfide thin-film transistor (TFT) LCDs 161–2 tungsten–heavy-metal alloys Voisey’s Bay 130 tin and tin ores. see also indium–tin (WHAs) 395, 397 volatilisation, antimony 79–82 oxide volcanogenic massive sulfide (VMS) indium alloys 215–16 UBC see Used Beverage Containers deposits indium recovery 210, 214 unconformity-related deposits 289 germanium 181, 184, 197 niobium and tantalum 363, 372–3, undiscovered resources 4 indium 206, 208–9, 211, 217–18, 222 379 United Nations Environment vulnerability to supply restriction resource criticality 8 Programme (UNEP), beryllium 115 (VSR) 14 titanium 269 United States Bureau of Mines TMI see trimethylindium (USBM) 5 Waste Electrical and Electronic total leaching 296 United States Environmental Protection Equipment (WEEE) transparent conducting oxides Agency (USEPA) germanium 191 (TCOs) 170 beryllium 117 lithium 239–40 trimethylindium (TMI) 216 magnesium 273–4 platinum-group metals 300 tungsten 385–413 rare earth elements 331 recycling and reuse 42–3, 47, 50, antimony deposits 78 United States Geological Survey 52–3, 59–60, 64 breccia deposits 391 (USGS) 5, 7 WHAs see tungsten–heavy-metal alloys brine/evaporite deposits 392 antimony 87 whole ore leach (WOL) method 136–7 contemporary uses 385, 396–8 beryllium 118 wireless communications 159 deposit characteristics 386–92 cobalt 139–40 wodginite 363 discovery 385 indium 217–18, 222 WOL see whole ore leach disseminated or greisen deposits 390 lithium 241 wolframite 386, 392, 394 distribution and abundance in the mining industry 27–8 World Trade Organisation (WTO) Earth’s crust 385 niobium and tantalum 380 magnesium 279 environmental factors 399–400 rare earth elements 312 mining industry 32 extraction methods 392–3 tungsten 400 tungsten 407–8 future developments and needs 402–9 uranium and uranium ores global distribution 388 metal preparation 269 X-ray transparency 102 global production 401–5 mining industry 33 xenotime 315–16, 326–8 hot spring deposits 392 rhenium 346, 349 indium recovery 210 tungsten alloys 399 yttrium see rare earth elements life-cycle analysis 399–400 urban mine concept 41–3, 66 yttrium aluminium garnet (YAG) mill products 398 USBM see United States Bureau of lasers 328 mineralogy 386 Mines mining industry 36 Used Beverage Containers (UBC) 270 zinc and zinc ores pegmatite deposits 392 USEPA see United States gallium recovery 151–3 physical and chemical Environmental Protection Agency germanium recovery 178–9, 181–4, properties 385–6 USGS see United States Geological 195 pipe deposits 392 Survey indium recovery 206, 212, 217–20, placer deposits 392 226 porphyry deposits 390–391 vacuum melting 113, 114 mining industry 28 prices 406–9 vanadium 34 resource criticality 8 processing and beneficiation 393–5 VB see vertical Bridgman (VB) method zircon 317 recycling and reuse 57, 398–9 vein deposits zirconium 269 resources and reserves 400–405 antimony 74, 76–8 zirconium–indium-based fluoride skarn deposits 389–90 indium 209, 211 glass 191