The Serra Pelada Au - Pd - Pt Deposit (Carajás, Amazon Craton, Brazil) Geology, Minralogy, and Hydrothermal Geochemistry of Ore Formation

Research Collection

Doctoral Thesis

Author(s): Berni, Gabriel V.

Publication Date: 2014

Permanent Link: https://doi.org/10.3929/ethz-a-010278913

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For more information please consult the Terms of use.

ETH Library DISS. ETH NO. 21929

THE SERRA PELADA Au - Pd - Pt DEPOSIT (CARAJÁS, AMAZON CRATON, BRAZIL): GEOLOGY, MINERALOGY, AND HYDROTHERMAL GEOCHEMISTRY OF ORE FORMATION

A thesis submitted to attain the degree of

DOCTOR OF SCIENCES of ETH ZURICH

(Dr.sc. ETH Zurich)

presented by

GABRIEL VALENTIM BERNI

Master in Geology

Universidade Federal de Minas Gerais

Born on November 26.11.1982

Citizen of Brazil

accepted on the recommendation of

Prof. Dr. Christoph A. Heinrich

Prof. Dr. Bernd Lehmann

Dr. Victor J. Wall

2014

Declaration of originality

I hereby confirm that I am the sole author of the written work here enclosed and that I have compiled it in my own words. Parts excepted are corrections of form and content by the supervisor.

Title of Work:

THE SERRA PELADA Au - Pd - Pt DEPOSIT (CARAJÁS, AMAZON CRATON, BRAZIL): GEOLOGY, MINERALOGY, AND HYDROTHERMAL GEOCHEMISTRY OF ORE FORMATION

Authored by:

Name: First name: BERNI GABRIEL VALENTIM

With my signature I confirm that have committed none of the forms of plagiarism described in the ‘Citation etiquette’ information sheet.

*I have documented all methods, data and processes truthfully. *I have not manipulated any data. *I have mentioned all persons who were significant facilitators of the work. *I am aware that the work may be screened electronically for plagiarism.

Copyright PhD Thesis, Department of Earth Sciences, ETH Zürich The author hereby agrees that after 31. July 2014 the PhD thesis may be copied and used for private and personal scholarly use. However, this is to emphasize that the making of multiple copies of the thesis or the use of the thesis for commercial purposes is strictly prohibited. When making use of the results found in this thesis, the normal scholarly methods of citation are to be followed.

Confidentiality moratorium This PhD thesis is a collaboration work between Colossus Minerals Inc. and ETH Zurich. The thesis is presently available for reading to teaching staff of the Department Earth Sciences but has to be treated confidentially. After the confidentiality period of 31. July 2014, this PhD thesis may be published.

GABRIEL V. BERNI PROF. DR. C.A HEINRICH

TABLE OF CONTENTS

Table of Contents SUMMARY ...... vii

ZUSAMMENFASSUNG ...... ix

CHAPTER I - INTRODUCTION ...... 1

1.1 – PREFACE ...... 1

1.2 – AIMS OF RESEARCH ...... 2

1.3 – THESIS ORGANIZATION ...... 2

1.4 – NOBLE METALS IN BLACK ROCKS: HYDROTHERMAL CONSTRAINTS AND GENETIC MODELS FOR UNUSUAL SEDIMENTARY PGE DEPOSITS ...... 3 1.4.1 – Transport of Pd, Pt and Au in hydrothermal fluids...... 3 1.4.2 – Genetic Models for some unconventional Pt and Pd deposits ...... 4

CHAPTER II - THE SERRA PELADA Au - Pd - Pt DEPOSIT, CARAJÁS, BRAZIL: GEOCHEMISTRY, MINERALOGY AND ZONING OF HYDROTHERMAL ALTERATION ...... 9

2.1 – ABSTRACT ...... 9

2.2 – INTRODUCTION ...... 10

2.3 – REGIONAL SETTING ...... 12

2.4 – GEOLOGY OF THE SERRA PELADA Au - Pd - Pt DEPOSIT ...... 13 2.4.1 – Stratigraphy ...... 13 2.4.2 – Structural setting ...... 18

2.5 – ANALYTICAL METHODS AND SAMPLE MATERIAL ...... 19

2.6 – HYDROTHERMAL ALTERATION AND MINERALIZATION ...... 20 2.6.1 – Distal alteration types ...... 20

2.6.2 – Hydrothermal alteration of the mineralized zones...... 23

2.7 – GEOCHEMISTRY OF HYDROTHERMAL ALTERATION ...... 26 2.7.1 – Bulk-rock geochemistry ...... 26 2.7.2 – Au-Pt-Pd distribution within major alteration types ...... 28

2.8 – DISCUSSION ...... 32 2.8.1 – Late Cretaceous bonanza style upgrading at Serra Pelada ...... 32 2.8.2 – Timing of hydrothermal Au - Pd - Pt mineralization with respect to regional deformation history ...... 33 2.8.3 – Hydrothermal alteration and mineralization at Serra Pelada: variant of an unconformity-related U deposits ...... 34

2.9 – ACKNOWLEDGMENTS FOR THE CHAPTER ...... 35

CHAPTER III - ORE MINERALOGY OF THE SERRA PELADA Au - Pd - Pt DEPOSIT AND IMPLICATIONS FOR ORE FORMING PROCESSES ...... 37

3.1 – ABSTRACT ...... 37

3.2 – INTRODUCTION ...... 38

3.3 – GEOLOGICAL SETTING AND MINERALIZATION OF THE SERRA PELADA Au - Pd - Pt DEPOSIT ...... 38

3.4 – ANALYTICAL TECHNIQUES AND SAMPLE MATERIAL ...... 40

3.5 – ORE PETROGRAPHY ...... 41 3.5.1 – Carbon-rich argillic ore ...... 41 3.5.2 – Kaolinite-rich argillic ore ...... 43

TABLE OF CONTENTS

3.6 – MINERAL ASSEMBLAGES ...... 45

3.7 – COMPOSITION OF ORE AND ACCESSORY MINERALS...... 46 3.7.1 – Copper and silver sulfide and selenide minerals ...... 46 3.7.2 – Au-rich phases ...... 48 3.7.3 – Pd-Pt-rich mineral phases ...... 53 3.7.4 – Secondary PGE alloys ...... 53

3.8 – DISCUSSION ...... 56 3.8.1 – Temperature constraints from the composition of selenide minerals...... 56 3.8.2 – Serra Pelada, a sulfide-deficient hydrothermal system: evidences from selenide and sulfide phase relations ...... 58

3.9 – CONCLUSIONS ...... 60

3.10 – ACKNOWLEDGMENTS FOR THE CHAPTER ...... 60

CHAPTER IV - FLUID MIXING AND REDOX GRADIENTS AT THE SERRA PELADA Au - Pd - Pt DEPOSIT ...... 61

4.1 – ABSTRACT ...... 61

4.2 – INTRODUCTION...... 62

4.3 – GEOLOGICAL SETTING AND SAMPLING MATERIAL ...... 63

4.4 – ANALYTICAL METHODS ...... 68

4.4 – RESULTS ...... 69 4.4.1 – Microthermometry ...... 69 4.4.3 – Compositional trends within FIA’s ...... 76

4.5 – THERMODYNAMIC CONSTRAINTS FROM FLUID INCLUSION LA-SF-ICPMS ANALYSIS AND MINERAL EQUILIBRIA...... 82 4.5.1 – Pressure and temperature ...... 82 4.5.2 – pH...... 83

4.5.3 – pH / ƒO2 (g) relationships from mineral assemblages and fluid inclusion data ...... 84

4.6 - GENETIC MODEL ...... 86

CHAPTER V - CONCLUSIONS AND OUTLOOK ...... 91

REFERENCES ...... 95

ACKNOWLEDGMENTS ...... 101

APPENDIX I...... 102

APPENDIX II ...... 110

CURRICULUM VITAE ...... 119

SUMMARY SUMMARY

The Serra Pelada Au - Pd - Pt deposit was discovered in the late 1980’s and was the site of the last major gold rush in South America’s history. The deposit is located within the Carajás Mineral Province at the eastern border of the Amazon craton, which also hosts world class iron ore and iron-oxide hosted copper-gold deposits. Even though the deposit became world-known when the mine was active, scientific research into the genesis of the deposit was limited by access to relatively un-weathered primary ore samples permitting detailed mineralogical and fluid inclusion analysis. In this research project, we used fresh drill core samples from a recent drilling campaign to characterize the hydrothermal alteration at the Serra Pelada deposit, by combining bulk-rock geochemistry and LA-ICPMS analysis of minerals and alteration halos, detailed mineralogy, and fluid inclusion studies on petrographically constrained samples from distal and local alteration zones.

The present thesis is structured by starting with a short introduction, followed by three main chapters and a summary of conclusions and outlook to future research and exploration for similar deposits (Chapter V). The main chapters are co-authored papers, of which I am always the leading author and have carried out most of the research and interpretation. One paper has already been accepted for publication in an international journal, and the other two manuscripts will be submitted upon committee’s approval.

The Serra Pelada Au - Pd - Pt deposit Carajás, Brazil: geochemistry, mineralogy and zoning of hydrothermal alteration (chapter II): Bulk-rock, detailed microscopy and LA-ICPMS analysis from well-preserved samples were used to characterize the primary features of the hydrothermal alteration at the deposit. Three major alteration zones comprise distal reducing and oxidizing alteration fronts plus the ore zone, which contains silicified portions, carbon-, kaolinite-, and hematite-rich argillic alteration types. The mineral paragenesis includes kaolinite, quartz, sericite, amesite (Mg-rich aluminum-silicate), monazite, rutile and either amorphous carbon or hematite, together with a complex ore mineralogy which is detailed in Chapter III. Major element changes during hydrothermal alteration include C and Mg addition, K depletion, localized silica loss and silicification with notable introduction of trace elements including LREE, Bi, Pb, U, V, Cu, Co, Ni and As. The results of this chapter were used to set the basis for the detailed mineralogy and fluid inclusion studies.

Ore mineralogy of the Serra Pelada Au - Pd - Pt deposit and implications for ore forming processes (chapter III): Fresh samples of argillic breccias and quartz-kaolinite veins were selected to describe the complex primary mineralogy within high and low grade samples which locally record extreme high-grades ore that reach >1000 g/t gold, palladium and platinum each. A complex

vii

SUMMARY paragenesis including sulfide, selenide, sulfate and oxide minerals was identified. The results of this work established thermodynamic constraints on the deposit genesis, suggesting that hydrothermal Au-Ag-Se mineralization occurred around 260 °C.

Fluid mixing and redox gradients at the Serra Pelada Au - Pd - Pt deposit (chapter IV): Fluid inclusions from distal alteration zones and silicified portions within mineralized zone record a wide range of salinity (5-22 wt. % NaCleqv.) but relatively constant homogenization temperatures (160 ± 25 ºC). By combining calculated fluid inclusion isochores with the independent mineralogical geothermometers (Chapter III), the formation of the deposit was constrained at about 260 ºC and 2 kbar. Individual fluid inclusions were analyzed by LA-SF-ICPMS analysis for the concentrations of K, Mg, Ni, Cu, As, Sb, Cs, Ba, Ce, Au, Pb, Bi and U. Using the potassium concentration of fluid inclusions, the pH of the system was estimated based on the kaolinite-muscovite transformation. A range of oxygen fugacities is indicated for the precipitation of the ore mineral assemblage. The combination of fluid-chemical, mineralogical and experimental-thermodynamic constraints was used to propose a genetic model for the deposit, involving deep infiltration of a highly oxidized surface brine that first interacted with fully oxidized greenstones, before mixing with a reduced and methane-bearing fluid derived from reduced metasedimentary rocks, to precipitate the high-grade carbonaceous Au - Pd - Pt selenide and arsenide ore.

viii

ZUSAMMENFASSUNG

Die Serra Pelada Au - Pd - Pt Lagerstätte wurde in den späten 1980er Jahren entdeckt und war der Ort des letzten grossen Goldrausches in der Geschichte Südamerikas. Die Lagerstätte liegt in der Carajás Mineral Province am Ostrand des Amazonas Kraton in Brasilien, das auch weltklassige Eisenerz- und Eisenoxid – Kupfer – Gold Lagerstätten enthält. Obwohl die Mine als sie aktiv betrieben wurde Weltbekanntheit erlangte, war der Zugang zu unverwitterten Erzproben eingeschränkt. Das wiederum erschwerte wissenschaftliche Arbeiten über die Mineralogie, Fluideinschlüsse und der Genese von Sera Pelada. Für dieses Forschungsprojekt verwendeten wir unverwitterte Bohrkernproben der jüngsten Explorationskampagne um hydrothermale Gesteinsveränderungen in der Sera Pelada Lagerstätte zu beschreiben. Für diesen Zweck machten wir Gesamtgesteinsgeochemie, LA-ICPMS Analysen von Mineralien, Alterationshöfen und untersuchten die Mineralogie der unverwitterten Erze. Gesteinsproben aus distalen und lokalen Alterationszonen wurden eingehend petrografisch dokumentiert und danach für Fluideinschlussstudien verwendet.

Die vorliegende Doktorarbeit beginnt mit einer kurzen Einleitung. Danach folgen drei Haupteile, eine Zusammenfassung der Schlussfolgerungen und zuletzt blicken wir auf mögliche zukünftige Forschungs- und Explorationsprojekte in ähnlichen Lagerstätten. Die Hauptkapitel sind als wissenschaftliche Publikation geschrieben bei denen ich Ersttautor bin und auch die meiste Forschungs- und Interpretationsarbeit geleistet habe. Eine Arbeit wurde von einer internationalen Fachzeitschrift zur Publikation angenommen und die beiden anderen Manuskripte werden zur Veröffentlichung eingereicht, vorausgesetzt das Gutachter zustimmen.

Geochemie, Mineralogie und hydrothermale Zonierung (2. Kapitel): Gesamtgesteinschemie, detaillierte Mikroskopie und LA-ICPMS Analytik wurden an unverwitterten Proben durchgeführt mit dem Ziel die primären Merkmale der hydrothermalen Alteration in der Lagerstätte zu bestimmen. Die drei grossen Mineralalterationszonen umfassen eine distale reduzierende- und eine oxidierende Alterationsfront sowie die Vererzung, die teilweise silizifiziert, Kohlenstoff-, Kaolinit- und hämatitreich alteriert ist. Die Mineralparagenese mit Kaolinit, Quarz, Serizit, Amesit (Mg reiches Aluminiumsilikat), Monazit, Rutil, amorpher Kohlenstoff oder Hämatit und eine komplexe Erzmineralogie ist Inhalt des 3. Kapitels. Hauptelementveränderungen durch die hydrothermale Alteration zeigen eine C und Mg Aufnahme, K Verringerung, lokaler Silikaverlust aber auch Silizifikation mit beträchtlicher Anreicherung von Spurenelementen einschliesslich LREE, Bi, Pb, U, V, Cu, Co, Ni und As. Die Resultate aus diesem Kapitel sind die Grundalge für die folgenden Studien zur Mineralogie und zur Fluidzusammensetzung.

ZUSAMMENFASSUNG Die Erzmineralogie der Serra Pelada Au - Pd - Pt Lagerstätte und Rückschlüsse auf den Vererzungsprozess (3. Kapitel): An unverwitterten und phyllitisch alterierte Brekzien und Quarz und kaolinitführende Adern wurde die primäre Mineralogie der Erze, die jeweils bis zu >100g/t Gold, Palladium und Platin enthalten, untersucht. Die Mineralvergesellschaftung zeigt verschiede Sulfide, Selenide, Sulfate und Oxide. Die Paragenese und die darauf aufbauenden thermodynamisches Modelrechnungen deuten darauf hin, dass die hydrothermale Au-Ag-Se Mineralisation bei ca. 260°C stattgefunden hat.

Fluidmischung und Redox Gradienten in der Serra Pelada Au - Pd - Pt Lagerstätte (4. Kapitel): Flüssigkeitseinschlüsse aus distalen Alterationszonen und aus silizifizierten Bereichen der

Vererzung zeigen eine grosse Salinitätsvariation (5-22 wt. % NaCleqv.) aber ziemlich konstante Homogenisationstemperaturen (160 ± 25°C). Aufgrund berechneter Fluideinschlussisochoren und einem unabhängigen Mineralthermometer (3. Kapitel) wurde die Lagerstätte bei etwa 270 °C und 2 kbar gebildet. In einzelnen Fluideinschlüssen wurde mittels LA-SF-ICPMS die Konzentrationen von K, Mg, Ni, Cu, As, Sb, Cs, Ba, Ce, Au, Pb, Bi und U gemessen. Anhand der Kaliumkonzentrationen in Fluideinschlüssen und der Kaolin – Muskovit Mineralumwandlung wurde der pH des Systems abgeschätzt. Zusätzlich wird ein Bereich von Sauerstofffugazitäten für die Entstehung der Erzminerale angegeben. Eine Kombination aus Fluidchemie, Mineralogie und experimentell- thermodynamische Rahmenbedingungen wurde verwendet um ein generisches Lagerstättenmodel vorzuschlagen. Eine hoch oxidierte Oberflächensole infiltrierte in der Tiefe vollständig oxidierte, (ultra-) mafische Metavulkanite. Danach mischte sich die Sole sich mit einem aus reduzierten Metasedimenten stammenden, methanhaltigen Fluid und fällte kohlenstoffführende und sehr Au - Pd - Pt reiche Selenid- und Arseniterze aus.

CHAPTER I

CHAPTER I INTRODUCTION

1.1 – PREFACE

The Serra Pelada Au - Pd - Pt deposit was discovered in the begging of the 1980’s, within the dense, rain-forest covered and relatively unexplored Carajás region in north Brazil. A few months after its discovery, artisanal miners started to come from all over the country, totalizing more than 50,000 people working on site during peak of activities, making Serra Pelada the site of the biggest gold rush ever seen in South America’s modern history (Fig. 1.1).

Figure 1.1 - The Serra Pelada open pit in September 1983 (photo by Arthur Bernadelli).

Although the Serra Pelada deposit has been mined between 1980 and 1986, a few exploration and even less scientific work has been done in the area. The deposit hosts extremely high-grade Au and platinum group elements (PGE) mineralization within hydrothermally altered metasedimentary rocks, a mineralization style which is poorly described in the literature, hence, lacking of scientific research. The extreme tropical weathering and the state of preservation of exploration drill cores (performed by CVRD in 1980-90’s) have always hindered the 1

CHAPTER I identification of the primary mineralogy and hydrothermal alteration features. After 2003, a joint-venture between Colossus Minerals Inc. and the mining cooperative of the artisanal miners of Serra Pelada (COOMIGASP) started a new evaluation of the deposit, providing access to new, fresh and well preserved samples used in this study.

1.2 – AIMS OF RESEARCH

The aim of this scientific research is to understand the key features necessary for the genesis of such an exceptional ore deposit as the Serra Pelada Au - Pd - Pt deposit, by detailed field documentation and sampling, mineralogy and fluid inclusion studies with fluid-mineral equilibria modeling. Detailed research objectives are:

 Determine the large and small scale hydrothermal features and zoning of the Serra Pelada deposit.  Identify the primary mineral paragenesis and deposit-scale geochemical controls on high- grade ore deposition based on detailed mineralogy and fluid inclusion studies.  Investigate the geochemical processes related to the Au - Pd - Pt mineralization and compare with similar deposits for developing targeting criteria to explore for similar deposits in the same region and beyond.

1.3 – THESIS ORGANIZATION

Chapter 1 defines the aims of this research and introduces the topic of sediment-hosted Au- PGE deposits with focus on deposits with similar geological and hydrothermal features.

Chapter 2 describes the regional and local geological features of the deposit and the zoning of the hydrothermal alteration. It is a descriptive chapter, based on detailed field documentation, drill core logging, sampling, petrology, bulk-rock and LA-ICPMS geochemistry, setting up background for the following chapters on detailed mineralogy and fluid chemistry.

Chapter 3 presents the results of the detailed investigation (scanning electron microscope imaging and electron microprobe analysis) of the ore and gangue mineralogy associated with the both primary (epigenetic) and secondary (supergene) Au - Pd - Pt mineralization at the Serra Pelada deposit.

Chapter 4 comprises the results of the fluid inclusion studies associated with main alteration zones described in Chapter 2. It includes detailed fluid inclusion petrography,

CHAPTER I microthermometry and LA-SF-ICPMS analysis and a fluid-mineral equilibria modeling based on the results.

Chapter 5 is a summary of the thesis conclusions and an outlook for future research and exploration on sediment-hosted epigenetic PGE deposits.

Appendixes include the thermodynamic calculations used to re-construct the diagrams in Chapter IV (Appendix I) and a list of the samples collected from drill cores of the Serra Pelada project (Appendix II). Digital appendix are included in a DVD within this thesis or could be also provide by request. They include the bulk-rock and LA-ICPMS geochemical dataset used at the second chapter, individual fluid inclusion analysis and the sample collection list.

1.4 – NOBLE METALS IN BLACK ROCKS: HYDROTHERMAL CONSTRAINTS AND GENETIC MODELS FOR UNUSUAL SEDIMENTARY PGE DEPOSITS

A great number of ore deposits of different commodities are hosted or directly associated with carbon-rich sedimentary or metasedimentary rocks. Some of these host low-grade PGE mineralization, normally as a by-product of Au, Cu and/or U, such as the Kupferschiefer Cu ± PGE deposits of northeast Germany/southwest Poland, the Russian Cu ± PGE and Au ± PGE deposits of Udokan and Sukhoi log Siberia, the Coronation hill U ± Au ± PGE deposit in Australia and the Serra Pelada Au - Pd - Pt deposit in Brazil, focus of this study.

1.4.1 – Transport of Pd, Pt and Au in hydrothermal fluids

Studies that consider the Pt and Pd solubility (Mountain and Wood, 1988; Gammons et al., 1992; Gammons, 1996; Wilde et al., 2003) have shown that Pt and Pd can be transported by hydrothermal fluids mainly as chloride, hydroxide or bisulfide complexes over nearly all geological conditions.

Chloride complexing is important only under oxidizing and acidic conditions. Concentrations of 10 ppb of Pt and/or Pd could be transported as chloride complexes in solutions in equilibrium with hematite at low pH (< 3) at temperatures of around 300° C - Mountain and Wood, 1988).

At highly oxidizing conditions (Mn2O3/Mn3O4 boundary), concentrations of Pt and Pd can exceed 1 ppm (Gammons et al., 1992). PGE solubilities as chloride complexes are much lower (< 1 ppb) at oxidation states more typical of natural hydrothermal fluids (e.g., pyrite/hematite or aqueous sulfide/sulfate boundaries), and will probably be less than the total contribution from other species (e.g., bisulfide, ammonia, or hydroxide complexes). This fact implies severe

CHAPTER I constraints on any conceptual model involving the transport and deposition of PGE as chloride complexes at low to moderate temperature (Gammons et al., 1992).

Hydroxide complexes: because hydroxyl is always present in a water solvent and hydrolysis becomes more significant with increasing temperature, the importance of Pt and Pd hydroxide complexes should also be considered (Mountain and Wood, 1988). The stability of M(OH)4-2 complexes predominate over chloride complexes over most of the geologically reasonable conditions, and Pt and Pd becomes appreciable soluble (up to 10 ppb of Pt or Pd) in near-neutral to basic solutions under reducing conditions, at temperatures from 25° to 300 °C (Mountain and Wood, 1988).

Bisulfide complexes: The soft nature of the Pd+2 and Pt+2 ions suggests that bisulfide complexing should be important, in analogy with Au. Under conditions where the solubility of HS- complexes reaches a maximum at 300° C, concentrations of 10 ppt to 1 ppb could be attained close to the pyrite + pyrrhotite + magnetite triple point (Mountain and Wood, 1988). Bisulfide complexing may be an important mechanism for transporting PGE metals under moderately reducing, near-neutral pH conditions, with high total sulfur, at intermediate temperatures near 250° C (Mountain and Wood, 1988).

There is also the chance that ligands involving selenium, tellurium, arsenic, antimony, bismuth or some combination of these with or without sulfur may also play an important role in platinum-group element transport although it is not possible to evaluate them with the available thermodynamic data (Mountain and Wood, 1988).

1.4.2 – Genetic Models for some unconventional Pt and Pd deposits

Unconformity-related U ± Au ± PGE deposits

Unconformity-related U deposits contain the largest and richest resources of this metal on earth. Although U is usually the principal commodity, many Australian deposits also contain significant quantities of Au and PGE, which are paragenetically associated with the U mineralization. For example, Pt and Pd occur at the Coronation Hill unconformity-related (Au-U) deposit (Fig. 1.2 A – Wilde et al., 2003). Significant but sub economic amounts of Pd are associated with Au near the Jabiluka deposit (Fig. 1.2 B), with grade exceeding 0.1 ppm Pd. Both deposits are spatially and genetically associated with the unconformity between a thin (< 2km), middle Proterozoic, coarse grained clastic rift-fill sequence (Ojakangas, 1986 – in Wilde et al., 2003) and lower Proterozoic amphibolite facies metamorphic basement which is locally rich in carbon and sulfide

CHAPTER I

(Wilde at al., 2003). Two styles of mineralization are present at the Coronation Hill area. The mixed U-Au-PGE mineralization is restricted to carbonaceous units and Au-PGE mineralization occurs in a number of different lithological units containing feldspar. Fluid inclusion data indicated that U, Au and PGE were transported in the same highly oxidized, low pH and very calcium-rich brine of intermediate temperature (200° C), probably as oxy-chloride and chloride complexes (Fig. 1.2 C - Mernagh et al., 1994, 1998).

Figure 1.2 - Sections through the Coronation Hill (A) and Jabiluka (B) unconformity-related U ± Au ± PGE deposits (Wilde, 1988; Carville et al., 1990 – in Wilde et al., 2003). (C) Simplified diagram showing the fluid flow at the Coronation Hill deposit (Mernagh et al., 1994).

Sediment-hosted Cu deposits

Stratiform Cu deposits share some characteristics with unconformity-related deposits, including the occurrence in an intracratonic rift environment and dominance of “oxidized” (hematite and sulfate rich) sedimentary rocks (Wilde et al., 2003). High levels of Pt and Pd have been documented in the Kupferschiefer deposits of Poland, among others deposits in Russia and

CHAPTER I

Kazakhstan (Kucha and Przybylowicz, 1999; Piestrzynski and Sawlowicz, 1999; Piestrzynski et al., 2002 – in Wilde et al., 2003).

The Kupferschiefer Cu and associated Au - PGE deposits occur within a dominantly terrestrial, red-bed sequence of evaporitic sediments and bimodal volcanic rocks unconformably overlying gray-colored, deformed Carboniferous clastic and Proterozoic metamorphic rocks (Kucha and Przybylowicz, 1999; Piestrzynski and Sawlowicz, 1999; Piestrzynski et al., 2002 – in Wilde et al., 2003). Economic Cu is spatially associated with the carbonaceous shale, although only 20% of the Cu is actually hosted by it. Au - Pd - Pt mineralization generally occurs below the economic Cu mineralization at the Polkowice-Sieroszowice mine, hosted by sandstones and the lower most shale unit, both relatively poor in organic carbon (Piestrzynski et al., 2002 – in Wilde et al., 2003).

Orogenic Au-Pt (Pd) deposits in carbonaceous metasedimentary rocks

Many occurrences of “black-shale” hosted Au deposits with potential economic PGE have been reported in Russian deposits, included in the orogenic gold classification as proposed by Groves et al., (1998). The Sukhoi log deposit in the Lena gold fields is the deposit for which data on PGE grades and distribution are most available (Wilde et al., 2003). Gold resources exceed 75 million onces at an average of 2.7 ppm of Au, with total PGE grades at the same order. Pt grades of about 1 ppm are common and generally exceed Pd, with the highest grades reported of 1.45 g/t Pt over 102.3 m and 2.42 g/t over 40.5 m (Distler et al., 1996). The deposit host rocks are clastic metasedimentary rocks with abundant carbonaceous matter and sulfide. At the majority of the Russian deposits, Au-PGE deposition can be related to the emplacement of quartz veins that postdate deformation (Wilde et al., 2001). This temporal relationship is, however, disputed at Sukhoi log where much of the gold is interpreted to be related to early pyrite, containing Pt minerals, rather than quartz veins (Buryak, 1982; Buryak & Khmelevskaia, 1997 – in Wilde et al., 2003). Fluid inclusion data indicate fluids of low salinity and unusually high CO2 concentrations (Wilde et al., 2003).

The Serra Pelada Au - Pd - Pt deposit: Exceptional among unconventional Pt and Pd deposits

The Serra Pelada Au - Pt - Pd mineralization has exceptional characteristics comparing to the unconventional deposits mentioned above and others. One of the main features is the PGE ratio compared to the main commodity. At Serra Pelada, the Pt and Pd contents play a big role on deposit’s viability and represents a considerable part of the deposit’s reserves (110 t of Au, 35 t Pd and 18 t Pt – Grainger et al., 2002) whereas in all other sediment-hosted epigenetic deposits

CHAPTER I the PGE are by-product of another major commodity. The degree of metal enrichment is another unusual feature of the Serra Pelada deposit. While in all other PGE deposits the Pt and Pd contents are mostly at the ppb range, locally with high grade intersections up to a few ppm, at Serra Pelada grades above 1 ppm of Pd+Pt are common and high grade intersections can host up to 4631 ppm Au, 1600 ppm Pt and 1730 ppm Pd – Colossus Minerals Inc., 2011)

This research project is a study of the Serra Pelada Au - Pd - Pt deposit to determine the key parameters necessary to form such a usual ore deposit, focused on regional and local description of the hydrothermal alteration features, mineralogy and fluid chemistry to be able to develop criteria for exploration of similar deposits.

CHAPTER II

THE SERRA PELADA Au - Pd- Pt DEPOSIT, CARAJÁS, BRAZIL: GEOCHEMISTRY, MINERALOGY AND ZONING OF HYDROTHERMAL ALTERATION

(Gabriel V. Berni, Christoph A. Heinrich, Lydia M. Lobato, Victor J. Wall, Carlos A. Rosière, Marcelo A. Freitas)

Published at Economic Geology: econgeol.geoscienceworld.org/content/109/7/1883.full.pdf

2.1 – ABSTRACT

The Serra Pelada Au - Pd - Pt deposit is located within the Carajás Mineral Province, which also hosts the world class iron ore and iron-oxide hosted copper - gold deposits of the Amazon craton in Brazil. The unusual low-temperature hydrothermal mineralization at Serra Pelada is epigenetic, hosted by metasedimentary rocks of the Águas Claras Formation and structurally localized in the Serra Pelada overturned syncline. The ore bodies are controlled by the intersection of subvertical NE-trending fault zones with metasiltstones, mainly at the syncline’s hinge, with minor ore occurrences at the upper and lower limb. Intense tropical weathering over the last 70Ma has completely overprinted the shallow ore in and near the flooded open pit, but primary hydrothermal features are preserved in deeper drill core delineating the remaining resource. Gold, platinum and palladium mineralization is associated with intense argillic alteration, hematite breccias and silicification, with the highest grade ore hosted by brecciated metasiltstones that are highly enriched in amorphous carbon. Distal alteration zones comprise a reducing and an oxidizing alteration front. The hydrothermal mineral paragenesis comprises kaolinite, quartz, sericite, amesite (Mg-rich Al-silicate), amorphous carbon, hematite, monazite, rutile, pyrite and a complex assemblage of Bi, Ag, Pb, Cu, Co, Ni, Pt, Pd and Au bearing, chalcogenide (S, Se) and arsenide (As, Sb) minerals. Major element changes during hydrothermal alteration include C and Mg addition, K depletion, localized silica loss and silicification with notable introduction of trace elements including LREE, Bi, Pb, U, V, Cu, Co, Ni and As. The hydrothermal alteration and the element association of the Serra Pelada deposit show geochemical similarities with unconformity-related uranium deposits, which may also be enriched in Au, Pd and Pt and were formed by mixing of fluids that interacted with a highly oxidized cover sequence and highly reduced rock packages in structures of brittle-ductile strain.

CHAPTER II

2.2 – INTRODUCTION

The Serra Pelada Au - Pd - Pt deposit is located within the Itacaiúnas Belt, in the northern part of the Carajás Mineral Province and near the eastern edge of the Amazon craton in Brazil. The Itacaiúnas Belt comprises an Archean granite – gneiss basement and two metavolcanosedimentary sequences, the Itacaiúnas Supergroup and the Andorinhas Supergroup, which are overlain by two metasedimentary sequences. A set of granitic and mafic/ultramafic intrusions from Archean to Proterozoic ages also occurs in the province. Well known for the giant iron mines, the Carajás Mineral Province also hosts big (<300 Mt) iron-oxide hosted copper gold deposits (IOCG; e.g., Cristalino, Salobo, Igarapé-Bahia, Sossego), as well as some Au, Ni and Mn deposits (Grainger et al., 2008).

The Serra Pelada deposit was discovered in the early 1980´s, with over 80,000 artisanal miners working in a 400 by 300 meters open pit during the peak of activities in the following years. Site of the largest gold rush in South America, the deposit has become world famous by its spectacular metal concentration at near surface (e.g., drill hole FD-32 at 54.5-55.0 m containing 132,000 ppm Au, 11,000 ppm Pd and 359 ppm Pt; Cabral et al., 2002), containing gold as coarse-grained aggregates and nuggets up to 50 kg. An estimated production of 32.6 tons of Au was manually extracted from the open pit, which collapsed and flooded in 1984 (Meireles and Silva, 1988).

Two styles of ore are present in the Serra Pelada deposit: (1) The near-surface, bonanza style mineralization which formed the nuggets and wire-shaped particles of black palladian gold (“ouro preto”; Cabral et al., 2002, 2011); (2) The deep-seated mineralization, as the focus of this study which also hosts high metal concentrations (e.g., drill hole SPD099 at 239.95-242.30 m containing 4630 ppm Au, 1730 ppm Pd, 1600 ppm Pt; Colossus Minerals Inc., 2011) but is characterized by a sulfide-selenide-arsenide mineralogy.

First sub-surface exploration took place during the late 1980´s and early 1990´s, managed by VALE (formerly Companhia Vale do Rio Doce – CVRD). In 2007, the mining license was transferred to the mining cooperative of artisanal miners of Serra Pelada (COOMIGASP). A joint venture with the Canadian company Colossus Minerals Inc. started a new evaluation of the deposit in 2009, allowing access to fresh drill core samples (Wall et al., 2009).

Published work on the Serra Pelada area includes geological descriptions of the mine sequence (Jorge João et al., 1982; Meireles et al., 1982; Meireles and Silva, 1988; Tallarico et al., 2000), the

CHAPTER II ore geochemistry (Moroni et al., 1999, 2001; Grainger et al., 2002), and the mineralogy and age dating with a main focus on the near-surface bonanza ore (Cabral et al., 2002, 2011).

Genetic models include an intrusion-related model (Tallarico et al., 2000) and Grainger et al. (2002) described the hydrothermal alteration within the ore zone and suggested a distal, low- temperature variant in the broad genetic group of iron-oxide copper-gold deposits of the Carajás Mineral Province. The hydrothermal process is thought to have occurred during Paleoproterozoic times, based on U-Pb ages of monazite intergrown with primary gold and PGE minerals (1861 ± 45 Ma – SHRIMP 208Pb/232Th, Grainger et al., 2002). Cabral et al. (2011) dated a Mn-Ba oxide aggregate intergrown with palladiferous gold and Pd-Pt minerals within the bonanza ore and proposed a Late Cretaceous (75 ± 6 Ma; 40Ar/39Ar) overprint of the Paleoproterozoic mineralization.

In previous studies, the deep weathering has limited the interpretation of primary features of hydrothermal alteration and mineralization, which is now more accessible by the recent drilling campaign. The present paper aims at characterizing the ductile to brittle deformation history of fresh ores and host rocks and their hydrothermal alteration in relation to primary variations in sedimentary rock composition, in preparation for more detailed studies of the primary ore mineralogy and the fluid-compositional evolution of the deposit based on fluid inclusions.

CHAPTER II

2.3 – REGIONAL SETTING

The basement assemblage of the Itacaiúnas Belt (Fig. 2.1) consists of granite – gneiss terrains known as the Pium Complex (3002 ± 14 Ma; Pidgeon et al., 2000) and Xingu Complex (2,859 ± 2 Ma; Machado et al., 1991). Overlying supracrustal rocks are related to two different metavolcanosedimentary sequences. The older Andorinhas Supergroup is represented in the area of Figure 1 by the Rio Novo Group (ca. 2.9 Ga; Avelar et al., 1999; Villas and Santos, 2001) and consists of mafic to ultramafic schists with minor felsic rocks and banded iron formations (BIF). The younger Itacaiúnas Supergroup (ca. 2.76 Ga; Machado et al., 1991) is dominated by mafic metavolcanic rocks and BIF.

Two metasedimentary units unconformably overlie the metavolcanosedimentary sequences: (1) The Águas Claras Formation (hosting the Serra Pelada Au - Pd - Pt deposit) is composed of metasandstones and metasiltstones deposited in fluvial to marine environments during late Archean to lower Paleoproterozoic times (Nogueira et al., 1995). It unconformably overlies the Itacaiúnas Supergroup in the western and the Andorinhas Supergroup in the eastern portion of the area shown in Figure 1; (2) The Caninana unit (ca. 2.0 Ga; Prado-Pereira, 2009) is also unconformably overlying the Itacaiúnas Supergroup and is composed of metasandstones and metaconglomerates deposited in a fluvial environment but its relationship with the Águas Claras Formation is yet to be established.

Intrusive rocks include the Luanga mafic/ultramafic complex (2,763 ± 6 Ma; Machado et al., 1991), syntectonic calc-alkaline granitoids and diorite bodies of the Plaquê Suite (ca. 2.74 Ga; Barros et al., 2000) and granites of the Estrela Complex (ca. 2.57 Ga; Machado et al., 1991). These older intrusions only cut the metavolcanosedimentary sequences and their basement, and are pervasively deformed. Anorogenic granitoids, mostly large alkaline A-type granite intrusions such as the Cigano and Central Carajás plutons (ca. 1.88 Ga; Machado et al., 1991), intrude all former rock units including the Águas Claras Formation. The metavolcanosedimentary and metasedimentary rocks of the Itacaiúnas Belt forms a set of WNW- and WSW-trending folds that locally progress to axial-plane thrust faults (Rosière et al., 2005), notably the Carajás and Cinzento strike-slip systems (Fig. 2.1). Pinheiro and Holdsworth (2000) interpret these faults as late Archean in age, reactivated in the Paleoproterozoic before being truncated by the anorogenic granites. The last major tectonic structure shown in Figure 1 is the frontal thrust of the Araguaia Fold Belt over the eastern border of the Amazon craton (Braziliano or Pan-African event, ca. 600 Ma; Almeida et al., 1981).

CHAPTER II

Figure 2.1 – Regional map of the Itacaiúnas Belt (IB - Continent-scale location on upper left; modified from Rosière et al., 2005). Ages referenced in the text. Inset shows location of Serra Pelada detailed in Figure 2.

2.4 – GEOLOGY OF THE SERRA PELADA Au - Pd - Pt DEPOSIT

2.4.1 – Stratigraphy

The Serra Pelada Au - Pd - Pt deposit is hosted by the Águas Claras Formation, which unconformably overlies the Rio Novo Group (Tallarico et al., 2000; Grainger et al., 2002) in the northeastern portion of the Itacaiúnas Belt (Fig. 2.2). Here, the Rio Novo Group consists of mafic and ultramafic schists, BIF and minor felsic rocks (Tallarico et al., 2000). It is intruded by the Luanga mafic/ultramafic complex (U-Pb 2,763 ± 6 Ma; Machado et al., 1991) and by the Formiga Granite (uncertain age), which hosts some Au mineralization (Grainger et al, 2002).

In the Serra Pelada area, the Águas Claras Formation is formed by alternating metasandstone and metasiltstone units. The lower portions of the sequence outcrop at the Elefanto syncline and consist of basal metasandstones with polymictic conglomerate layers and laminated

CHAPTER II carbonaceous and red metasiltstones. The upper part of the sequence outcrops at the Serra Pelada syncline and Cedro anticline and is composed by metasandstones, dolomitic metasandstones with layers of metasiltstone and metaconglomerates, undifferentiated gray metasiltstones and red metasiltstone and an upper metasandstone sequence (Fig. 2.2).

The mine sequence at the Serra Pelada deposit (Fig. 2.3) overlies the mafic and ultramafic schists of the Rio Novo group (Fig. 2.4 A) and consists of the following rock types, stratigraphically from bottom to top in the lower limb of the Serra Pelada syncline, as shown in Figures 4B to 4F:

Dolomitic sandstone: Comprises impure fine- to coarse-grained sandstones cemented by fine- grained dolomite (Fig. 2.4 B), and minor interlayers of mono- and polymictic conglomerates (Fig. 2.4 C) with clasts of rounded quartz and minor fragments of BIF within a sandy and dolomitic matrix. This unit is weathered to pure friable sand near the surface and locally down to 300 m depth.

Gray metasiltstone: Made of a thin, laminated metasiltstone composed of quartz, fine-grained muscovite (sericite), organic matter <0.1 wt. % Ctotal) and rare pyrite (Fig. 2.4 D, E). In the Serra Pelada syncline this rock forms a continuous 15-20 m thick layer at the base of the red metasiltstone. This lithology hosts the bulk of the precious-metal resource of the Serra Pelada deposit (Grainger et al., 2002).

Red metasiltstone: Composed of centimeter-scale pink layers of fine sandstone and millimeter-thin laminar darker red pelitic layers (Fig. 2.4 F). The mineralogy comprises muscovite, kaolinite, quartz, hematite and goethite (in decreasing abundance, Moroni et al., 2001). The rock shows well preserved sedimentary structures such as ripple marks and graded beds with coarse-grained bottom foresets (e.g. Fig 4F, steep structure). A weakly-schistose texture defined by a weak grano- to nematoblastic alignment of phyllosilicate minerals, hydrated Fe-oxide minerals and quartz within discrete zones, indicates the effects of low-grade metamorphism. Mineral assemblages within these minor zones include only locally developed muscovite (Grainger, 2003).

CHAPTER II

Figure 2.2 – District scale geological map of the Serra Pelada area (modified from Berni, 2009). Inset shows location of Serra Pelada Au - Pd - Pt deposit shown in Figure 3.

CHAPTER II

Figure 2.3 – Geological Map of the Serra Pelada Au - Pd - Pt deposit (modified from Berni, 2009).

CHAPTER II

Figure 2.4 – Least altered rocks in and around the Serra Pelada Au - Pd - Pt deposit. (A) Chlorite-talc schist from the Rio Novo Group. (B) Fine grained dolomitic sandstone. (C) Dolomitic conglomerate with red fragments of banded iron formation. (D) Least altered gray siltstone (<0.1 wt. % Ctotal). (E) Photomicrograph of slightly altered gray siltstone showing carbon mobilization and re-precipitation along S2 crenulation plans (subvertical structure). (F) Red siltstone with bright sandy and darker red clay-rich layers. Scale bars are 1cm.

Intrusive rocks: Diorite bodies of uncertain age intruded the Rio Novo Group as well as the metasedimentary rocks of the Águas Claras Formation. Tallarico et al. (2000) described the contact metamorphism associated with the diorite intrusion SW of the deposit (Fig. 2.3 A), where the occurrence of actinolite + calcite and the local association of diopside indicate a minimum temperature of 550°C for the peak metamorphic conditions (dolomite + quartz + H2O = actinolite + calcite + CO2). Diorites, all sedimentary and earlier magmatic rocks were

CHAPTER II subsequently deformed and metamorphosed under lower greenschist facies (Grainger et al., 2002).

2.4.2 – Structural setting

Textures in the metasedimentary rocks of the Águas Claras Formation near Serra Pelada record two shortening deformation events (D1 and D2). The north-northwest verging D1 event was related to regional-scale, tight and asymmetrical NE-trending F1 folds, mapped as the Serra Pelada and Elefanto synclines (Grainger et al., 2002) and the Cedro anticline (Berni, 2009). All three regional-scale folds plunge to SW. The Elefanto syncline and the Cedro anticline have moderately-dipping limbs (45-60º) and axis plunge (30-40º), whereas the Serra Pelada Syncline has shallow-dipping limbs (15-30º) and a gently plunging axis (15-20º). Progressive folding led to axial-plane thrust faults, possibly with associated oblique movement, which resulted in a stacking of the regional-scale F1 folds such as the Serra Pelada and the Elefanto synclines (Fig. 2.2). The

D2 event is related to a low-strain deformation phase, with associated NE-trending subvertical faults and the development of a local crenulation cleavage and centimeter-scale folds. Later NW- trending faults offset all former structures and control the emplacement of Jurassic gabbro dikes (Tallarico et al., 2000).

The diorite intrusion pre-dates S1 foliation and is the oldest intrusion in the area. The intrusion of the Cigano A-type granite (U-Pb age of 1.88 Ga; Machado et al., 1991) and a north- northwest-trending gabbro dike (Jurassic Rb-Sr age of 198 Ma; Meireles et al., 1982) postdate the S1 metamorphic foliation. The Serra Pelada deposit is localized in the homonymous syncline, but ore formation clearly postdates the D1 event that formed this reclined fold (Grainger et al., 2002). The Serra Pelada syncline has a fold closure of semi-cylindrical geometry, showing little or no apical thickening.

Later structures related with the D2 event control the location of mineralization (Fig. 2.5 A). These are subvertical NE-trending fault zones associated with a non-penetrative crenulation cleavage and small-scale open folds, which overprint the gently dipping axial-plane foliation (Fig. 2.4 E). Ore grade distribution based on 3D modeling of drilling assays from the Serra Pelada deposit reflects this structural control on the large scale, defining two ore zones (Fig. 2.5 B): (1) The Main Zone, located at the syncline’s hinge and hosted by hydrothermally altered gray and red metasiltstones, with minor occurrences hosted by the dolomitic sandstones. It is subvertical and grossly tabular in shape, with approximately 100 m in height, a horizontal width up to 40m, at least 900m length extending from the open pit down along the gently plunging axis of the

CHAPTER II syncline towards SW. The main high-grade ore shoots (> 50 ppm Au) are narrow, rod-shaped, NE-trending and plunge gently to SW. (2) The Limb zones are located about 150-200 m west of the fold hinge mainly at the lower limb and locally at the upper limb of the syncline. Mineralization is hosted mainly by hydrothermally altered gray metasiltstone and is controlled by the intersection of another fault zone parallel to the main zone at the lower limb of the syncline which resulted in a flat-lying ore shoot, having 10-15 m in vertical thickness, and approximately 50 m in width.

Figure 2.5 – Structural features of the Serra Pelada Au - Pd - Pt deposit. (A) Outcrop scale structural controls of mineralization with the main ore zone (horizontal view, after Gaál, 1997). (B) Schematic 3D view of the Serra Pelada syncline and the two ore zones of the deposit.

2.5 – ANALYTICAL METHODS AND SAMPLE MATERIAL

Samples of the least-altered siltstones and of altered and mineralized rocks were selected from 28 different drill cores within the deposit plus samples from two high-grade intervals from the exploration program of Colossus Minerals Inc. Description of the drill hole depth, individual methods, detection limits, and quality control data are listed in Table 1 of the digital repository. Whole-rock analyses of major, trace and noble metals (Au, Pt and Pd) were carried out by Genalysis Intertec Laboratory Services (Perth, Australia) on 349 samples. Drill core samples were pulverized, homogenized and aliquots of 50 g were dissolved in sodium peroxide fusion and hydrochloric acid. Major elements, plus B, Cu, Sc, V, Zn, S, Co and Cr were analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES) and other trace elements and REE by inductively coupled plasma mass spectrometry (ICP-MS). Precious metal (Au-Pt- Pd) analyses where made by fire assay (25 g) prior to LA-ICPMS.

CHAPTER II

In order to have a better understanding of the bulk rock data obtained from core intersections without close spatial control, we quantified bulk-rock ratios of major element from small-scale variations of sedimentary layers, mm-scale barren argillic halos, quartz-kaolinite veinlets and some massive carbon- and kaolinite-rich argillic rocks by LA-ICPMS analysis of polished slabs of these very fine-grained rocks. We used the Excimer LA-ICPMS at ETH Zürich (Günther et al., 1997) and large shallow ablation pits of 110µm diameter, which yielded quite homogeneous signals allowing evaluation of averaged major element ratios in the rock. Analytical and standardization procedures were performed using the NIST SRM 610 as an external reference material according to Heinrich et al. (2003). Element ratios were calculated using the SILLS software (Guillong et al., 2008). Table 2 (see digital repository item) reports the core-scale bulk- rock analyses and small-scale LA-ICPMS ratios for the selected samples.

For detailed mineralogy studies we used a JEOL JSM - 6390LA scanning electron microscope (SEM) attached to a Thermo Scientific UltraDry EDS detector at ETH Zürich.

2.6 – HYDROTHERMAL ALTERATION AND MINERALIZATION

The mineralization at Serra Pelada is a fracture-controlled system on all scales. Two contrasting distal hydrothermal alteration comprises chlorite / carbon alteration and wide spread hematite alteration in metasedimentary rocks and metavolcanic rocks at different location within of the deposit vicinity (Fig. 3 B). The main mineralized zone is located at the hinge of the fold and comprises carbon-rich argillic alteration, silicification and hematite breccias (Fig. 2.6).

2.6.1 – Distal alteration types

Chlorite-carbon alteration: the chlorite-carbon alteration is the most widespread alteration feature within the metasiltstone sequence. It transgressively overprints the red metasiltstone sequence along the contact between the dolomitic sandstone and the gray metasiltstone, up to 350 m away from the ore zone (Fig 3B). The alteration features are readily recognized by the precipitation of chlorite and carbon along fractures and the compositional layering of the red metasiltstone (Fig. 2.7 A). Carbon precipitation is visible within veins that cross-cut the already chlorite altered metasiltstone (Fig. 2.7 B), or along S2 crenulation planes (Fig. 2.7 C and 4 E). The alteration replaces low-grade metamorphic muscovite by chlorite along fractures and their halos with associated carbon addition, which turns the red metasiltstone into a green-grayish colored rock. The alteration assemblage comprises chlorite + amorphous carbon + muscovite + quartz with minor pyrite, chalcopyrite and goethite (pseudomorphs after pyrite or magnetite).

CHAPTER II

Figure 2.6 – (A) The possibly best photographic record of the ore zone, exposed during open-pit mining activity (note miner for scale). (B) Interpretation of the alteration zones within the photograph in A. (C) Geological cross section (see Fig. 3 for location) with alteration zones and Au + Pt + Pd grades along exploration drill cores, constructed from recent exploration data.

Hematite alteration: this is a patchy alteration feature which is observed up to 250 m away from the ore zone. In restricted areas of subvertical fracturing, dolomitic sandstone is changed to a hematite dusted recrystallized marble with numerous dissolution vugs (Fig. 2.3 B, 2.7 D). These vugs are lined by cm-sized euhedral crystals of co-precipitating quartz, hematite - dusted dolomite (Fig. 2.7 E) and late calcite. Growth zones within the quartz and dolomite crystals include abundant hematite, with minor pyrite, chalcopyrite and carrollite-fletcherite (Cu(Ni,Co)2S4). Bulk (110 µm ablation pits). LA-ICPMS analyses of these growth zones yield concentrations of 100’s ppb of Pt, evidencing the distal relationship of this alteration with the ore mineralization. Within the red metasiltstone sequence, comparable hematite alteration is restricted to a few quartz-hematite veins crosscutting the earlier S1 foliation. Regionally, quartz- hematite breccias are observed near the main fault zones hosted by metasedimentary rocks or by the metavolcanic basement up to 6 km from the deposit.

CHAPTER II

Figure 2.7 – Distal alteration zones around the Serra Pelada Au - Pd - Pt deposit. (A) Chlorite + carbon alteration at subvertical crack, extending along compositional layering of the red siltstone. (B) Carbon-quartz veins within chlorite altered red siltstone. (C) S2 oriented chlorite totally overprinting red siltstone. (D) Coarse-grained, hematite-altered dolomitic sandstone with dissolution vugs lined by hematite-dusted dolomite. (E) Detail of vug with co-precipitated crystals of quartz, dolomite and hematite, from large fracture zone cutting through the rocks shown in D; growth zones in crystals are lines with fine-grained sulfides enriched in Ni, Pt and other ore-related elements. Scale bars are 1cm.

CHAPTER II

2.6.2 – Hydrothermal alteration of the mineralized zones

Argillic alteration overprints carbon-enriched metasiltstones (> 10 wt. % Ctotal) and is most closely associated with the bulk of the Au-Pt-Pd resource within the deposit. It can be subdivided in three subtypes, the carbon-, kaolinite- and hematite-rich argillic zones (Fig 6). Most of the ore occurs with carbon-rich argillic alteration, where the ore texture varies from a coherent carbon- enriched metasiltstone with stylolitic quartz – kaolinite veins (Fig. 2.8 A) to high-grade friable breccias with metasiltstones fragments within a carbon – quartz – kaolinite matrix (Fig. 2.8 B). The stylolitic veins are mostly associated with flat fractures, consistent with subhorizontal shortening associated to the D2 deformation event. Carbon-poor zones of intense kaolinite-rich (Fig. 2.8 C) and hematite-rich argillic alteration (Fig. 2.8 D) occur as thin discontinuous bodies between the main carbon-rich argillic ore (Fig. 2.6) and often present breccia textures. It replaces the metamorphic muscovite by kaolinite, locally associated with coarse-grained hydrothermal muscovite and locally amesite. Gold occurs as aggregates of fine pureness (< 2 wt. % Ag) and fischesserite (Ag3AuSe2), as isolated grains within the matrix, or included in quartz, monazite, hematite sulfides and selenides. Palladium minerals are arsenopalladinite (Pd8(As,Sb)3), isomertieite (Pd11Sb2As2) and mertieite (Pd11(Sb,As)4). Palladseite (Pd17Se15) was also identified by Grainger (2003). Platinum phases are fine grained braggite (Pt,Pd,Ni)S, Sperrylite (PtAs2), sudovikovite (PtSe2) and isoferroplatinum (Pt3Fe). Palladian gold was reported by Cabral et al. (2002), containing inclusions of Pd-arsenides, isomertieite, palladseite, sudovikovite and a Pt-Pd- Se phase, but this texturally overgrowing alloy was found only within the oxidized ore within and just below the former open pit mine. Other accessory minerals within the primary hydrothermal ore comprise pyrite, chalcopyrite, covellite, selenian covellite, Ag-Cu-Fe sulfide, carrollite-fletcherite (Cu(Co,Ni)2S4), galena, barite, anhydrite, rutile, clausthalite (PbSe), naumannite (Ag2Se) and litochlebite (Ag3PbBi4Se8). Detailed investigation of this complex mineralogy is in progress. Barren argillic alteration extends to the inner hinge of the Serra Pelada syncline and consists of zones of carbon removal (bleaching) along S2 crenulation planes, with associated sericite, kaolinite, quartz, hematite and/or rare pyrite (Fig. 2.8 E).

CHAPTER II

Figure 2.8 – Alteration types from deep near-vertical drill holes intersecting the ore zone. (A) Coherent carbon-enriched siltstone with stylolitic quartz - kaolinite vein containing gold, fischesserite with minor primary hydrothermal sulfides including Se-covellite and chalcopyrite. (B) High grade carbonaceous breccia (~ 640 ppm Au, 360 ppm Pd, 255 ppm Pt). (C) Kaolinite-rich argillic alteration. (D) Hematite-rich argillic alteration, weakly gold-mineralized. (E) Barren sericite alteration extending vertical along S2 crenulation. (F)

Deformed pre-ore quartz vein showing that at least some silicification predated or was coeval with the S1 axial plane foliation. (G) Fine-grained later siliceous alteration showing near-coeval carbon and hematite veinlets. (H) Fine-grained siliceous alteration with vugs partly filled by coarse quartz, with globular and platy hematite at the base and along growth zones of the euhedral crystals. All scale bars are 1cm.

CHAPTER II

Siliceous alteration envelops the argillic altered rocks mainly at the upper and lower limb of the Serra Pelada syncline and locally at the inner hinge (Fig. 2.6). The dolomitic sandstones in the outer part of the fold hinge and the metasiltstones at the inner hinge are changed to fine- to coarse-grained quartz with minor hematite and pyrite. Mineralization is sub-economic (mostly sub-ppm Au grades) and locally higher grades are normally restricted to silicified gray metasiltstone (Grainger et al., 2002). Silicification includes early crack-seal veins and even recrystallized veins that were folded by the metamorphic S1 deformation (Fig. 2.8 F). The greater part of silicification around the ore body does not show any S1-related structure, indicating that it occurred during ore formation. It is formed by fine grained massive to coarse and locally vuggy quartz, with contemporaneous hematite and carbon veinlets (Fig. 2.8 G), suggesting that reduced and oxidizing alteration fronts formed at overlapping times later in the deformation history. The vugs are patchy dissolution features, rather than fracture openings. They are typically lined by fine-grained or globular hematite that is overgrown by clear euhedral quartz crystals (Fig. 2.8 H), indicating local silica dissolution and re-precipitation. The mineral assemblage is quartz with minor fine-grained muscovite, kaolinite, hematite, pyrite and rare chalcopyrite. The siliceous alteration corresponds to the jasperoid alteration described by Tallarico et al. (2000) and Grainger et al. (2002), terminology defined as fine-grained to cryptocrystalline silica replacement of carbonate rocks by Neuendorf et al. (2005). At Serra Pelada, silicification replaces both metasiltstone and dolomitic sandstone, and therefore the more general term is used here.

Hematite breccias locally with economic ore grades occur in the Limb Zone, with a few low grade occurrences also in the Main Zone of the deposit (Fig. 2.6 C). This breccia is made up of angular fragments of metasiltstone and fine grained silica within a goethite matrix (weathered hematite/pyrite). In the Limb Zone, the breccias are matrix supported and contain a greater proportion of metasiltstone fragments, whereas similar breccias in the main zone are clast supported, with predominantly siliceous fragments. These breccias host high-grade Pd-Pt mineralization up to hundreds ppm Pd+Pt, with particularly high Pd+Pt/Au ratios.

CHAPTER II

2.7 – GEOCHEMISTRY OF HYDROTHERMAL ALTERATION

2.7.1 – Bulk-rock geochemistry

Figures 9A and B show the bulk concentration ranges of typically immobile elements such as Al, Ti and Zr for least altered host rocks and variably mineralized samples. Within the range of the primary sedimentary variations, those elements are kept in the same proportions in the protolith and altered rocks, and therefore can be used to assess significant element changes during the hydrothermal alteration, at least in a qualitative way. Major element variations observed between the least altered metasedimentary rocks and the hydrothermally altered rocks within the ore zone can be explained by variable additions of C and Mg and loss of K with localized silica addition and loss (Fig. 2.9).

Some of the kaolinite-rich argillic samples show elevated concentrations of residual immobile elements (yellow triangles in Fig. 2.9 A and B), indicating net mass removal of K and also some Si. This is shown in Figure 9C, where the K/Al values within samples of the kaolinite-rich argillic reach much lower values and decrease faster (in respect to K/Si) than those in carbon- rich argillic samples. The Si/Al ratio slightly increases or the ratio is kept at a nearly constant value, in respect to Si/K values, implying smaller silica addition associated with this type (Fig. 2.9 D).

In the carbon-rich argillic alteration and many other ore-related rocks, the concentrations of Al, Ti and Zr are lowered in near-constant proportions compared to the least altered siltstones (Figs. 2.9 A, 2.9 B), requiring massive dilution by addition of C, Si and Mg with localized silica loss (Fig. 2.9 D). Most of the carbon-rich argillic samples show a similar trend as observed within the kaolinite-rich argillic samples, with a Si/K ratios increase and nearly constant Si/Al implying K loss. The same trend is also visible within small-scale barren argillic alteration haloes (orange crosses in Figs. 2.9 D). For some of the carbon-rich argillic rocks, the values of Si/K increase together with Si/Al within the carbon-rich argillic rocks, implying a Si input, in total exceeding the mass loss by K leaching in this alteration type. A few samples show smaller Si/Al values than the least altered metasedimentary rocks, suggesting localized Si-loss. (Fig. 2.9 D). Mg enrichment is also associated with many kaolinite-rich and some of the carbon-rich argillic samples, as shown in Figure 9E, where the higher increase of Mg/Al values compared to Mg/K cannot be explained by K loss only (cf. two dashed arrows in Fig. 2.9 E).

CHAPTER II

Figure 2.9 – Bulk-rock concentrations of Ti, Al and Zr, as a selection of relatively immobile elements maintaining approximately constant ratios during hydrothermal silicification, hematite and/or carbon addition (A, B). Ratio plots (C, D and E) comparing selected major elements proportions for least altered rocks and main alteration types within the Serra Pelada Au - Pd - Pt deposit, showing K-loss during ore-related kaolinitization, silicification and Mg-alteration. Big symbols are whole rock, small symbols are LA-ICPMS analyses. Red circles represent the stoichiometric compositions of minerals: M = muscovite; K = kaolinite; A = Amesite; Q = Quartz. 27

CHAPTER II

Structural relations indicate that this intense carbon addition occurred at the same time as the distal chlorite-carbon alteration (Figs. 4E and 7B), raising the carbon content from 0.1 wt. %

Ctotal in the least altered gray siltstone up to 10% wt. % Ctotal within the carbon-rich argillic alteration. The association of the most carbon-enriched zone with S2 crenulations, fracturing and brecciation indicates that carbon addition was a hydrothermal process associated with ore formation. There is no direct correlation of Au+Pd+Pt grades with carbon content, but the highest grades in primary ore consistently occur in samples with carbon content higher than 3 wt. % (Fig. 2.10 A).

The siliceous alteration zone is dominated by dilution of Al, Ti and Zr by quartz, requiring strong Si addition in most samples of massive granular and vuggy quartz. The low Zr and Ti concentrations in these rocks argue against a primary origin of silica-rich rocks as sandstone lenses, in which detrital rutile and zircon should be enriched compared to clay minerals (i.e., Al) in shales. Note that the high K/Al > 0.4 (at very low K, Al) reported in Figure 9C for some of the most siliceous samples must be analytical artifacts due to incomplete dissolution of minor clay minerals, rather than indicating the presence of K-feldspar.

Precious metals and related trace-elements including Bi, Pb, U, Se, Te, As, Sb and V are highly enriched compared to the least altered metasedimentary rocks and broadly correlate with Au + Pt + Pd ore grades (Fig. 2.10 B-F), in contrast to the dilution trends exhibited by the immobile elements. This is consistent with the textural evidence that monazite, clausthalite and Bi-V oxides are commonly being intergrown with Au, Pt and Pd selenides, Sb and As bearing minerals. Significant enrichment of all these ore-related trace elements is also largely associated with elevated carbon contents. Geochemical analyses for Rh and Ir carried out in a few mineralized samples (carbon-rich argillic rocks only – Figs. 10G and H) show good correlation for Rh with Au+Pd+Pt grades and a more broad correlation for Ir. All altered rocks also display highly enriched values of LREE (in monazite) and elevated Cu, Co and Ni.

2.7.2 – Au-Pt-Pd distribution within major alteration types

From the hydrothermal alteration we can therefore distinguish two main ore types; the carbon- rich argillic ore (80-85% of the ore tonnage) and the kaolinite-rich argillic ore (5-10%). Other ore types are minor and have a more variable grade distribution, including the hematite breccia restricted to the lower Limb Zone (~5% of the resource) and a few occurrences within siliceous alteration and hematite-rich argillic rocks (< 5 %).

CHAPTER II

The Au - Pd - Pt distribution is broadly correlated, with Au >Pd ≥ Pt (Moroni et al., 2001) but with variable ore-metal ratios associated with the different alteration types. Figure 11 shows plots of Au vs. Pd+Pt and Pt vs. Pd for the different alteration types within the deposit. Samples of the high-grade bonanza ore that was mined are represented by samples from the old pit including data published by Cabral et al. (2002). Lines of constant element ratio are displayed for reference.

Figure 2.10 – Variation diagrams for carbon and selected ore-related trace elements as a function of total Au + Pd + Pt ore grade, comparing the least altered rocks with main alteration types within the Serra Pelada deposit. Note that that high-grade ores contain orders of magnitude more Ctotal than reduced country rocks, with a correlation between precious metal grades and concentrations of Pb, U, Bi, Ag, As + Sb and (at very low levels) Ir and Rh.

CHAPTER II

For the highest grades, Au and Pd + Pt are clearly correlated in all argillic altered rocks, showing that the exceptional element association and their extreme degree of enrichment are related with each other. However, very high Au or Pt + Pd grades also occur with low to modest grades of the other metal, in all types of alteration.

The ore grade within the carbon- and the kaolinite-rich argillic rocks is extremely variable, ranging from weight percent’s down to less than a ppm in the low-grade samples. Generally Au grade is higher than Pt + Pd (Fig. 2.11 A, B), except for some samples with high Pd+Pt to Au ratios. The majority of samples of the deep seated mineralization is somewhat enriched in Au relative to Pd+Pt, but samples from the bonanza ore show the highest Au to Pt + Pd ratios. Likewise, the Pt to Pd ratios within the carbon- and argillic-rich ores are relatively constant around 1 for the deep sited mineralization, while Pd can be enriched relative to Pt up 12x in the bonanza ore samples (Fig. 2.11 C, D).

The hematite-argillic and the siliceous alteration are generally poor in metal contents, with a few high-grade samples. The Au to Pd+Pt ratios are extremely variable within both types of alteration (Fig. 2.11 E, F). The same differences observed between the bonanza ore and the deep mineralization are also evident with the hematite-rich argillic ore (Fig. 2.11 G). As expected since siliceous-rich rocks are less susceptible to weathering, the same is not observed in the siliceous samples within the pit (Fig. 2.11 H).

CHAPTER II

Figure 2.11 – Metal concentrations grouped according to different ore types in the Serra Pelada deposit, linked with logging information from more than 60,000 m of diamond drill core (data obtained by Colossus Minerals Inc. from all drilling campaigns between 1980’s to 2008). Note that near-surface samples that are largely oxidized to Fe- and Mn-oxy-hydroxides (from Cabral et al., 2002) are most extremely enriched in precious metals, with depletion of Pt relative to Au and Pd.

CHAPTER II

2.8 – DISCUSSION

2.8.1 – Late Cretaceous bonanza style upgrading at Serra Pelada

In a recent paper, Cabral et al. (2011) interpret the genesis of the bonanza-style ore, with coarse wires of palladian gold (e.g. FD 0032), to be a product of a Late Cretaceous hydrothermal event (75 ± 6 Ma) and implied that such a late hydrothermal event was the main process that generated the spectacular precious metal endowment of this unusual deposit. Their conclusion is based on 40Ar-39Ar dating of a Mn-Ba oxide aggregate from a thoroughly oxidized high-grade ore sample, which the authors consider to be coeval with hydrothermal quartz and hematite of the type we describe in the present paper, rather than being a product of the long lasting lateritization of the Carajás region that started around the same time (~72 Ma; Vasconcelos et al., 1994). Additional geochemical arguments for a Cretaceous hydrothermal process are (1) iridium concentrations were supposedly depleted in the bonanza-ore style mineralization compared to the primary ores, requiring a hydrothermal overprint because Ir is extremely immobile in surface waters (Anbar et al., 1996 in Cabral et al., 2011); and (2) the supposed Bi- rich affinity of the deep-seated mineralization compared to As, Sb, Se and Hg enrichment of the bonanza style.

Our results show that none of these hydrothermal features described by Cabral et al. (2011) are restricted to the fully-oxidized bonanza-grade ore, and that they were generated by the hydrothermal processes that formed the primary high-grade ore at depth. Cavity-filling tabular hematite is widely and intimately associated with euhedral quartz in the periphery of the primary carbonaceous ore (e.g. Figs. 7D and 7E). Both are indeed hydrothermal, but unlikely to be coeval with the Cretaceous Mn-Ba oxides dated by Cabral et al. (2011). The Au-Pt-Pd concentrations show a good correlation with Bi, but the association with As, Sb, Te and Se is equally pronounced in the primary ore and not a geochemical feature restricted to the oxidized bonanza ore (Fig. 2.10 F). Clearly hydrothermal quartz - kaolinite veins (Fig. 2.8 A) and other argillic-altered rocks contain selenides, Sb-arsenides and sulfides as primary minerals, and in the bonanza ore mineralization these minerals are found as relict inclusions in palladian gold. Selective Ir depletion would be a valid argument against an origin of the bonanza ore by weathering, but the argumentation that the deep-seated mineralization is strongly enriched in Ir is not clear. Ir analyses presented in this contribution and also by Grainger, 2003 show that the concentration of Ir is quite variable within high-grade Au - Pd - Pt samples of the deep-seated mineralization and samples of high-grade Au+Pd+Pt (< 100 ppm) also show low Ir concentrations (Fig. 2.10G). 32

CHAPTER II

Even though the bonanza-style ore contains the spectacular gold nuggets and extremely high- grade ore that triggered the discovery of the deposit and the gold rush, we conclude that this overprint was not the main process that generated this unusual ore deposit. Deep seated high- grade ore dominated by Bi-Se-Te-As-Sb-bearing Pt and Pd minerals already produced spectacular ore grades up to 5000 ppm Au+Pd+Pt, with a metal enrichment of 7 orders of magnitude compared with ordinary crustal rocks. Mn – Fe – oxide rich bonanza ore formation involved a further enrichment by at most 1 order of magnitude compared to the primary, dominantly carbonaceous ore (Fig. 2.11). This emphasizes that exploration criteria should be focused on the primary mineralization features at regional to local scale, which we will further discuss in the following sections.

2.8.2 – Timing of hydrothermal Au - Pd - Pt mineralization with respect to regional deformation history

The alteration zoning around the Serra Pelada syncline's hinge implies that this pre- mineralization structure, caused by regional deformation and metamorphism (D1) in the late Archean, was an important factor in the ground preparation for localizing the ore body.

However, mineralization occurred later during tectonic reactivation (D2) in Paleoproterozoic times. Grainger et al. (2008) dated hydrothermal monazite intergrown with Au and Pt-Pd minerals (U-Pb SHRIMP age of 1861 ± 45 Ma) within the Serra Pelada deposit, indicating that this is the most likely age of primary low-temperature hydrothermal Au - Pd - Pt enrichment to grades of thousands of ppm of each of the precious metals. We suggest that the D2 event was associated with minor regional deformation in the Paleoproterozoic, well after major pervasive deformation of the Carajás block in the late Archean. Low-strain, mostly brittle deformation locally reactivated the main E-W fault systems of Archean age, and developed the main NE fault zones controlling the Serra Pelada mineralization. Temporal overlap of the interpreted age of mineralization with the emplacement of Paleoproterozoic alkaline granitoids (1883 ± 2 Ma - U- Pb on zircon by SHRIMP, Machado et al. 1991) indicates a common, anorogenic thermotectonic event in the region, but does not necessarily prove a genetic link of hydrothermal ore formation to the distal granitoid plutons and any magmatic fluids they may have generated, as tentatively implied by Grainger et al. (2008).

CHAPTER II

2.8.3 – Hydrothermal alteration and mineralization at Serra Pelada: variant of an unconformity-related U deposits

Textural and geochemical data presented in this paper demonstrate a hydrothermal, epigenetic origin of the primary high-grade Au, Pd and Pt mineralization at Serra Pelada. Deep-reaching carbonate dissolution in the dolomite-cemented sandstones and supergene modification of the primary ore mineralogy to palladian gold alloys and Mn-Fe oxy-hydroxides has largely overprinted the primary hydrothermal alteration and high-grade mineralization, which therefore has been missed or rejected in earlier studies. We have shown here that quartz, hematite and kaolinite associated with the primary sulfide and selenide mineralization at the Serra Pelada Au - Pd - Pt deposit (e.g. Fig. 2.8A) are not a product of weathering. The ore body geometry and associated argillic alteration of the outer portions of the fold hinge are structurally controlled by the D2 brittle-ductile deformation event. Distinct stylolitic quartz-kaolinite veins, which are limited to the ore zone and clearly associated with primary mineralization, are inconsistent with an entirely supergene origin of the argillic alteration.

Textural features including small-scale alteration halos along cracks (Fig. 2.6A), crystal-lined vugs within the dolomitic sandstone (Fig. 2.6D) and the siliceous halo around the deposit (Fig. 2.8H), stylolitic veins (Fig 2.8A) and small-scale argillic alteration halos (Fig. 2.8E) show unequivocally that these are hydrothermal alteration products. Geochemical changes such as K loss associated with Si, C and Mg addition fit the observed mineralogical changes superimposed on primary sedimentary variations. The observed variation within the least altered rocks are much smaller than the observed loss and gains of K, Si, C and Mg (Fig. 2.9), which are most likely related to the replacement of muscovite by kaolinite ± amesite and the precipitation of quartz, as observed in the carbon-rich and kaolinite-rich argillic alteration.

Reducing (amorphous carbon) and oxidizing (platy and globular hematite) alteration fronts occurred at overlapping times in the stability field of kaolinite, as recorded in fractures and primary growth zones of massive and euhedral quartz and dolomite crystals. The close concentration of high-grade Au - Pd - Pt mineralization in rocks with high carbon content indicates that preexisting carbon, as well as additional precipitated carbon that was probably advected as CH4, acted as an essential trap for the precipitation of the precious metals as well as the suite of other redox-sensitive elements. This element suite includes Au, Pd, Pt, Bi, U, As, Se, Te and V, which are variably enriched in the primary ore at Serra Pelada. The same element suite also characterizes unconformity-related uranium deposits (Wilde et al., 1989), and particularly the low-temperature hydrothermal Au – Pd ± U deposit of Coronation Hill located in the same

CHAPTER II region as the large unconformity- related uranium deposits of Australia (Mernagh et al., 1994). These deposits are interpreted to be the product of fluid mixing between a surface-derived, deeply circulating basin brine with a reducing fluid originating from organic carbon-bearing metasedimentary rocks in the basement underlying the unconformity (Wilde et al., 1989; Mernagh et al., 1994, 1998; Alexandre et al., 2005; Dérome et al., 2005; Cuney et al., 2010). We propose that the Serra Pelada deposit was localized at the site of reduction of a suite of elements that were transported in a highly oxidized fluid that had previously contacted mafic or ultramafic rocks while traversing the tectonically stacked Archean basement. Similar low-temperature hydrothermal redox processes precipitated high-grade Au, Pd and Pt ore together with these elements by formation of minerals that are largely insoluble at reducing conditions. Such a variant of an unconformity-related redox model is a more likely explanation for the origin of Serra Pelada than a magmatic-hydrothermal process of iron-oxide-copper-gold affinity as previously proposed. An ongoing study of fluid inclusions and hydrothermal geochemistry aims at refining the proposed redox process that generated the element association and exceptional ore grades at Serra Pelada.

2.9 – ACKNOWLEDGMENTS FOR THE CHAPTER

We gratefully acknowledge Colossus Minerals Inc. for providing data, permitting publication and financially supporting the ongoing scientific research at the deposit. GVB is thankful to all Colossus employees and the garimpeiros from Serra Pelada, for all their assistance during the time on site. We also thank Markus Wälle for help with the LA-ICPMS analysis.

CHAPTER III

ORE MINERALOGY OF THE SERRA PELADA Au - Pd - Pt DEPOSIT AND IMPLICATIONS FOR ORE FORMING PROCESSES

(Gabriel V. Berni, Christoph A. Heinrich, Lydia M. Lobato, Victor J. Wall)

*To be submitted to Mineralium Deposita

3.1 – ABSTRACT

The Serra Pelada Au - Pd - Pt deposit was discovered in the late 1980’s and was the site of the last major gold rush in South America’s history. Located at the eastern border of the Amazon Craton, the rocks at Serra Pelada have experienced intense tropical weathering for about 70 Ma. Nevertheless, deep drill core samples of argillic breccias and quartz-kaolinite veins preserve the primary mineralogy and hydrothermal alteration features, with locally extreme ore grades reaching hundreds of ppm Au, Pd and Pt each. The mineralization at Serra Pelada is hosted by hydrothermally altered metasiltstones and dolomitic metasandstones at the hinge zone of a recumbent syncline, comprising zones of hematite, carbon-chlorite, argillic, and siliceous alteration. The main hydrothermal gangue minerals are kaolinite, sericite, amesite, quartz, hematite, monazite, florencite and amorphous carbon. Accessory minerals are pyrite, chalcopyrite, Se-covellite, klockmannite, krutaite, athabascaite, naumannite, clausthalite, Bi-V oxides, barite, and anhydrite. Ore minerals consist of electrum with variable metal contents (Ag,

Pd, Pt, Cu, Fe), palladian-gold, palladseite (Pd17Se15), Sb-rich isomertieite (Pd11Sb2As2), mertieite-

II (Pd8(Sb, As)3), sperrylite (PtAs2), sudovikovite (PtSe2) and fischesserite (Ag3AuSe2). The composition of fischesserite varies from the ideal formula towards more Ag as well as more Au rich compositions. This indicates that fischesserite formed in the disordered solid-solution form, which is stable above 260ºC and sets a minimum temperature for the formation of the Serra

Pelada Au - Pd - Pt deposit. The precipitation of selenium-rich copper minerals requires ƒSe2 (g) to ƒS2 (g) ratios higher than unity, which can occur in the sulfate predominance field at ƒO2 (g) conditions well above the hematite-magnetite equilibrium. Primary ore and gangue minerals at Serra Pelada comprise a suite of elements including Bi, Se, As, U, Au, Pt and Pd that are best transported in oxidizing conditions and will be precipitated upon reduction, consistent with the carbonaceous mineralogy of most of the ore. Secondary processes have leached out most of the As, Sb, Bi, Se and recrystallized some of the Au (visible from the gold nuggets found within laterite), leaving behind a series of metallic alloys of variably complex composition. 37

CHAPTER III

3.2 – INTRODUCTION

Discovered by artisanal miners in the beginning of 1980, the Serra Pelada Au - Pd - Pt deposit was the site of the biggest gold rush in South American recent history, with a total of 32.6 t of gold manually extracted from a 300 x 400 m open pit (Meireles and Silva, 1988). Exploration work at Serra Pelada was performed by VALE (formerly CVRD) between 1980 to late 1990’s and a re-evaluation of the deposit started in 2007 by Colossus Minerals Inc. The deposit is exceptional in terms of its unusual Au - Pd - Pt metal association, ore grade and size, with total reserves estimated at 110 t Au, 35 t Pd and 18 t Pt (Grainger et al., 2002).

The extensive tropical weathering in the Serra Pelada region and the poor preservation of old drill core held back the collection of mineralogical data relevant to the primary mineralization of Serra Pelada. Detailed studies of the bonanza-style ore were published by Cabral et al. (2002a, b), but no detailed study of the primary mineralogy and associated gangue minerals was ever published. Grainger (2003) describes the primary hydrothermal alteration within the ore zone, but it remained uncertain whether the identified platinum-group minerals (PGM) could account for the reported Pd + Pt grades and little was known about accompanying sulfide and selenide minerals. Identifying the primary mineral assemblage and related gangue minerals is essential for metallurgical processing of the remaining subsurface ore and constrains the thermodynamic conditions during the ore-forming process. In this study, we used least weathered samples from the 2007-2009 drilling campaign in order to characterize the primary mineralogy at the Serra Pelada Au - Pd - Pt deposit.

3.3 – GEOLOGICAL SETTING AND MINERALIZATION OF THE SERRA PELADA Au - Pd - Pt DEPOSIT

The Serra Pelada Au - Pd - Pt deposit is located in the northern part of the Carajás Mineral Province and near the eastern edge of the Amazon craton in Brazil. The deposit is hosted by a late-Archean to Paleoproterozoic folded metasedimentary sequence named Águas Claras Formation (Nogueira et al., 1995), which consists of low-grade metamorphosed dolomitic sandstone, conglomerates, red and gray siltstones (Meireles and Silva, 1988; Tallarico et al., 2000; Grainger et al., 2002). The Águas Claras Formation overlies the metavolcanosedimentary rocks of the Rio Novo Group and the mafic/ultramafic intrusion of the Luanga complex, which contains sub-economic, orthomagmatic enrichments of Cr and platinum group elements (PGE - Tallarico et al., 2000; Grainger et al., 2002). Diorite rocks of unknown age intrude the Águas Claras Formation and are folded and trusted within the metasedimentary sequence. Large alkaline A-type anorogenic granitoids (ca. 1.88 Ga; Machado et al., 1991) intrude all former rock

CHAPTER III units including the Águas Claras Formation. The geological setting of the Serra Pelada deposit is described in more detail by Tallarico et al. (2000), Grainger et al. (2002) and Berni et al. (2014).

The mineralization at Serra Pelada is hosted by hydrothermally altered metasiltstones and metasandstones at the hinge zone of a reclined syncline (Moroni et al., 2000; Tallarico et al., 2000; Grainger et al., 2002; Berni et al., 2014). Hydrothermal alteration comprises distal and spatially separated chlorite - carbon and hematite alteration, and the mineralized zone is located at the syncline’s hinge (Berni et al., 2014 - Fig. 3.1 A). The mineralized zone comprises carbon-, kaolinite- and hematite-rich argillic alteration of the metasiltstones and a silicification halo in the dolomitic metasandstones at the outer syncline’s hinge (Berni et al., 2014 - Fig. 3.1 B).

Figure 3.1– Hydrothermal alteration zoning at Serra Pelada Au - Pd - Pt deposit. (A) Distal alteration features showing the location of carbon-chlorite, hematite and mineralized zone. (B) Hydrothermal alteration zone within the main mineralized zone at the hinge zone of the Serra Pelada syncline (modified from Berni et al., 2014).

CHAPTER III

3.4 – ANALYTICAL TECHNIQUES AND SAMPLE MATERIAL

High- and low-grade samples from the main ore zone at the Serra Pelada deposit were selected from drill cores for detailed transmitted and reflected light microscopy. Out of a greater number, five samples, including low- and high-grade carbon- and kaolinite-rich argillic ores (Table 3.1) were selected for detailed scanning electron microscopy (SEM) and electron microprobe (EMP) analysis at ETH Zurich. The presence and abundance of sulfide and selenide minerals were also considered as a criteria for selecting the least weathered samples. Polished sections were investigated using a JEOL JSM - 6390LA SEM attached to a Thermo Scientific UltraDry EDS detector. EMP analyses were performed using a JEOL JXA8200 instrument with 5 wavelength- dispersive spectrometers. Analyses were performed with 15 kV acceleration voltage and a beam current of 2 x 10-8 A. Analyzed elements, X-ray lines and standards as follow: Au, Mα, pure metal; Pt, Mα, pure element; Pd, Lα, pure metal; Ag, Lα, pure element; Se, Lα, pure element; As,

Lα, Skutterudite (CoAs3); Sb, Lα, pure element; S, Kα, pyrite (FeS2); Cu, Kα, pure element; Ni, Kα, Millerite (NiS); Fe, Kα, pure element. For fischesserite and naumannite EMP analysis, a naumannite standard was used for Ag and Se for a better matrix-fit. For copper selenides measurements, pure element standards were used, since their selenium content can be up to two times higher than in naumannite.

Table 3.1 – Drill core, depth and metal contents of ore samples investigated in detail.

Au Pt Pd Drill core Depth Ore Type (ppm) (ppm) (ppm) SPD001 233.7 m Kaolinite-rich argillic breccia. 1.93 77.6 74.6

Coherent carbon-rich sitstone cut SPD001 214.5 m 7.34 0.13 0.20 by quartz-kaolinite veins.

Kaolinite-rich argillic breccia with SPD055 223.1 m 28.4 83.5 91.3 carbonaceous fragments.

SPD095 238.5 m Kaolinite-rich argillic breccia. 138 429 231

SPD002 229.0 m Carbonaceous breccia. 446 181 245

CHAPTER III

3.5 – ORE PETROGRAPHY

3.5.1 – Carbon-rich argillic ore

This ore type hosts the bulk of the remaining subsurface metal resources of the Serra Pelada deposit. The gold, platinum and palladium content varies from less than a ppm and can reach thousands of ppm’s in the highest-grade samples (Berni et al., 2014). The ore grade shows a direct relationship with the ore texture, from the low grade, carbon-enriched and coherent metasiltstone to the high grade carbonaceous breccia, as shown in figures 3.2 A to D including the least altered metasiltstone for comparison.

The coherent carbonaceous metasiltstone shows a lepidoblastic texture defined by the S1 foliation, made of thin layers mainly composed by muscovite, minor kaolinite, quartz and variables amounts of amorphous carbon (up to 16 wt. % - Berni et al., 2014). Carbon is precipitated as thin lamellae and veinlets along the sedimentary layering, the S1 metamorphic foliation or along the steep crenulation cleavage (S2 – Fig. 3.2 B). At places of stronger post-S1 deformation, the pre-mineralization texture is partially or completely obliterated to breccias made of carbonaceous metasiltstone fragments within a carbon- and kaolinite-rich matrix, which hosts the high-grade mineralization (Figs. 3.2 C, D). The breccias display a chaotic texture, defined by quartz – kaolinite aggregates and metasiltstone fragments surrounded by a matrix composed of amorphous carbon and kaolinite with minor sericite and quartz.

The low-grade mineralization hosted by the coherent metasiltstone is associated with sparse, centimeter-scale, stylolitic quartz-kaolinite veins of shallow dipping or sub-horizontal orientation such as in figure 3.2 C. The veins are formed by coarse-grained (500 µm to 1 mm) subhedral to anhedral quartz at vein core, with fibrous quartz crystals at the vein walls within a matrix made of coarse-grained kaolinite and minor sericite (Fig. 3.2 E). Locally, quartz crystals also contain abundant inclusions of carbonaceous matter. Main accessory minerals are monazite, rutile, hematite and florencite, (La,Ce)Al3(PO4)2(OH)6, barite, anhydrite and Ag-Pb-Cu bearing sulfides and selenides. Gold, platinum, palladium and associated minerals are included in quartz, located at grain boundaries or adjacent to carbonaceous matter. In the high grade carbonaceous breccias, the ore minerals and related accessory minerals are precipitated within the matrix enveloping metasiltstone fragments (Fig. 3.2 F) or within small kaolinite-quartz veinlets.

CHAPTER III

Figure 3.2 – (A) Least altered metasiltstone with carbonaceous layer (< 0.1 wt. % Ctotal). (B) Coherent carbon- enriched metasiltstone with precipitated carbon along subhorizontal S1 and steeply-oriented S2 crenulation planes. (C) Deformed carbon-rich argillic ore with gold and selenide bearing quartz-kaolinite vein. (D) 42

CHAPTER III

Carbon-rich argillic breccia with coarsen kaolinite - quartz aggregates and several generations of carbon veinlets. (E) Detail on mineralized quartz - kaolinite veinlet of (C) with abundant opaque mineral inclusions which include gold, sulfide and selenide minerals. (F) High-grade carbon-rich argillic breccia with partially obliterated metasiltstone fragment (top center) and abundant coarse-grained gold (shiny grains, ca. 1% Au of the photo area). All photos from A to E were taken using transmitted light and crossed-polars, except for F, which was taken using reflected light. Mineral abbreviations: Kao = kaolinite; Ser = sericite; Anh = anhydrite; Qz = quartz; Bi-V-O = bismuth-vanadium oxides.

3.5.2 – Kaolinite-rich argillic ore

The kaolinite-rich argillic ore hosts extremely high grade mineralization at elevated (Pd+Pt)/Au ratios and represent a subordinate primary ore type at Serra Pelada. This ore type varies between massive kaolinite with variable amounts of hematite inclusions to a brecciated ore made of quartz and metasiltstone fragments within a kaolinite - sericite - amesite - hematite matrix (Berni et al., 2014). The fragments of metasiltstone commonly show altered rims (Fig. 3.3 A) and/or are cut by several generations of kaolinite - quartz - hematite veinlets (Fig. 3.3 B). Quartz occurs as anhedral grains with undulose extinction or as free-standing euhedral centimeter-scale quartz crystals in irregular vugs (Fig. 3.3 A).

The mineralogy comprises coarse-grained kaolinite, sericite and amesite (60-80% in total), with variable amounts of hematite (2-15%), amorphous carbon (up to 1%) and quartz (1-10 %). Kaolinite and lesser sericite and amesite occur as coarse- to fine-grained fibro-radial crystals within the matrix, but also as coarse-grained aggregates with quartz and sericite in veinlets. Massive kaolinite aggregates contain abundant inclusions of monazite, hematite and rutile (Fig. 3.3 C), which are also found in the breccia matrix and constitute the main accessory minerals. Quartz normally occurs as breccia fragments and is rare within the matrix. Hematite is found as tabular crystals within the matrix or as euhedral inclusions in quartz and rutile (Fig. 3.3 D).

The ore minerals occur within grain boundaries among the kaolinite - hematite matrix (Fig. 3.3 E) and as inclusions in quartz, rutile, hematite and monazite. (Fig. 3.3 F). They comprise the same mineralogy of selenide, sulfide, gold and PGE minerals as observed within the carbon-rich argillic ore, which are detailed in the following section.

CHAPTER III

Figure 3.3 – Kaolinite-rich argillic ore. (A) Argillic breccia with fragments of altered metasiltstone and euhedral quartz crystals. (B) Photomicrograph (transmitted light) of an argillic breccia showing a fragment cut by a mineralized quartz - kaolinite - hematite veinlet. (C) Photomicrograph (transmitted light) showing the kaolinite-rich ore with abundant inclusions of hematite, monazite and rutile. (D) Back-scattered image showing a large rutile crystal with inclusions of hematite and barite. (E) Photomicrograph (transmitted light, cross polar) of the kaolinite-rich argillic ore, with kaolinite-sericite aggregates within a carbon-rich matrix. (F) Reflected light photomicrograph of the same area shown in (E), with coarse and fine grained PGM within the carbon-rich matrix. Mineral abbreviations: Rut = rutile; Mnz = monazite; Hem = hematite; Bar: barite; PGM = Platinum group minerals; Kao = kaolinite; Ser = sericite.

CHAPTER III

3.6 – MINERAL ASSEMBLAGES

Textural relations and mineral intergrowths are well-preserved within kaolinite-quartz veins from the least weathered samples, were two hypogene ore assemblages were defined based on microscopic overgrowth textures (Table 3.2, Fig. 3.4):

(A) An early assemblage made of coarse grained pyrite and chalcopyrite with associated millerite (NiS), cobaltite (CoAsS), Pb and Ag sulfides and selenides. The silver minerals commonly occur as rims of acanthite (Ag2S), naumannite (Ag2Se) or mckinstryite (Ag,Cu)2S growing on pyrite;

(B) A later sulfur deficient assemblage with Se-covellite, chalcocite, a variety of Cu, Pb sulfides and selenides, barite, anhydrite and Bi-V oxides coexisting with gold, platinum and palladium minerals. The ore minerals are fischesserite (Ag3AuSe2), electrum (Au,Ag), palladian-gold

(Au,Pd), sperrylite (PtAs2), Se-bearing braggite (Pd,Pt,Ni)S, sudovikovite (PtSe2), isomertieite

(Pd11Sb2As2) and mertieite-II (Pd8(Sb,As)3). One crystal of litochlebite (Ag2PbBi4Se8) was identified at the edge of a pyrite crystal. Primary ore minerals within the high-grade carbon-rich and kaolinite-rich breccias are rare and mostly preserved as inclusions within refractory accessory minerals (e.g., monazite, rutile) or as relicts within the breccia matrix, normally showing rims altered to pure metal alloys.

Table 3.2 – Summary of the main identified primary and secondary mineral assemblages within the Serra Pelada deposit

Gangue Minerals Ore and Acessory Minerals

Assemblage A Assemblage B

Pyrite, chalcopyrite, Se-covellite, krutaite (CuSe2), Amorphous carbon, chalcocite, millerite (NiS), klockmannite (CuSe), athabascaite kaolinite, sericite, amesite, cobaltite (CoAsS), galena, (Cu5Se4), anhydrite, barite, Bi-V oxides,

quartz, hematite, monazite, acanthite (Ag2S), electrum, fischesserite (Ag3AuSe2),

rutile apatite. naumannite (Ag2Se) and sudovikovite (PtSe2), palladseite

argillic ore Carbon-rich clausthalite (PbSe). (Pd17Se15) and sperrylite (PtAs 2).

barite, electrum, palladian gold, Kaolinite, sericite, amesite, Pyrite, chalcopyrite, sudovikovite (PtSe2), Se-Braggite quartz, amorphous carbon, naumannite (Ag Se), ((Pt,Pd,Ni)S) , isomertieite (Pd Sb As ), hematite, monazite, rutile, 2 11 2 2 galena, clausthalite (PbSe). mertieite-II (Pd (Sb,As) ), sperrylite apatite, florencite. 8 3 argillic ore (PtAs ). Kaolinite-rich Kaolinite-rich 2

The secondary, supergene mineralogy is mainly composed of goethite, cryptomelane and quartz relicts with associated phosphates such as waylandite (BiAl2(PO4)(OH)6. Crandallite and woodhouseite were also reported by Moroni et al., 2001. Gold, platinum, palladium and

CHAPTER III associated ore minerals are converted into metal alloys within the most weathered samples. These include Pd-Cu alloys, Pt-Pd-Cu-Fe alloys of variable composition, and native gold containing less than 2 wt. % of the other metals.

3.7 – COMPOSITION OF ORE AND ACCESSORY MINERALS

3.7.1 – Copper and silver sulfide and selenide minerals

A variety of copper selenide and S-poor copper sulfide minerals are associated and commonly intergrown with Au and PGM in the best-preserved samples. These minerals formed late relative to pyrite and chalcopyrite, growing at the edges and as rims of pyrite crystals, but also as individual crystals in the matrix or included in quartz. Compositions of analyzed sulfide and selenide minerals are listed in Table 3.3.

Naumannite (Ag2Se) and clausthalite (PbSe) are the most abundant selenide minerals, normally associated with pyrite and chalcopyrite (assemblage A) and are more widespread than their sulfide analogs (acanthite and galena). Crystals of mckinstryite (Ag,Cu)2S and an unknown phase

(approximately (Ag,Cu)8Se3,S2) were also identified. The minerals are present as individual crystals, as inclusions in late sulfides such as shown in figure 3.4 A, or at the edges of pyrite crystals showing complex intergrowth between sulfides and selenides (Fig. 3.4 B).

Table 3.3 – Electron Microprobe Analyses of Ag and Cu selenides and sulfides minerals.

Grain Se S Cu Ni Fe Ag Au Pt Pd Total Mineral Formula No.

1 45.4 0.1 51.9 0.0 3.4 0.1 0.0 0.0 0.0 101.0 Athabascaite Cu5Se4

2 48.3 0.2 51.1 0.0 2.3 0.2 0.0 0.0 0.0 102.0 Athabascaite Cu5Se4 1 55.3 0.0 42.2 0.0 3.8 0.4 0.0 0.0 0.0 101.8 Klockmannite CuSe 2 57.1 0.0 44.0 0.0 0.2 0.3 0.0 0.0 0.0 101.7 Klockmannite CuSe 2 57.2 0.0 42.1 0.1 0.1 0.3 1.4 0.0 0.0 101.2 Klockmannite CuSe 1 73.2 0.0 29.2 0.0 0.0 0.0 0.0 0.0 0.0 102.4 Krutaite CuSe2 1 8.4 26.8 61.2 0.0 0.2 0.2 0.1 0.0 0.0 96.9 Se-covelite CuS 1 8.9 26.7 61.4 n.a. 0.1 0.2 0.1 0.0 0.0 97.4 Se-covelite CuS 2 8.1 27.3 60.6 0.0 3.3 0.3 0.0 0.0 0.0 99.6 Se-covelite CuS

1 0.1 22.0 77.1 0.0 0.2 1.6 0.0 0.0 0.0 101.1 Chalcocite Cu2S

1 26.7 na. na. na. na. 73.3 0.0 0.0 na. 100.0 Naumannite Ag2Se 2 25.2 na. na. na. na. 73.5 0.0 0.0 na. 98.7 Naumannite Ag2Se 3 27.1 na. na. na. na. 72.3 0.0 0.1 na. 99.4 Naumannite Ag2Se 3 27.6 na. na. na. na. 73.8 0.0 0.0 na. 101.4 Naumannite Ag2Se

1 3.3 11.9 1.0 0.0 0.5 85.8 0.0 0.0 0.0 102.4 Acanthite Ag2S 2 1.9 11.7 0.9 0.0 0.0 84.2 0.0 0.0 0.0 98.6 Acanthite Ag2S

1 0.1 13.7 24.7 0.0 0.2 64.1 0.0 0.0 0.0 102.8 Mckinstryite (Ag,Cu)2S 1 21.4 5.6 16.4 0.0 0.1 57.0 0.0 0.0 0.0 100.4 Un-named Ag- (Ag,Cu)8Se3S2 Cu selenide

CHAPTER III

Covellite and chalcocite are the main late-stage copper sulfides, (Assemblage B) normally found in complex intergrowths with each other and other ore minerals (Fig. 3.4 A). A variety of Se-rich covellite with up to 8.85 wt. % Se was also identified (Table 3.2). Copper selenides occur as individual grains, along the edges of pyrite crystals (Fig. 3.4 C), included in quartz or disseminated within the kaolinite matrix. They consist mainly of klockmannite (CuSe) and more rare krutaite (CuSe2) and athabascaite (Cu5Se2).

Figure 3.4 – Main sulfide and selenide minerals in primary Serra Pelada Au - Pd - Pt ore. (A) Crystal of chalcocite and selenian covellite with inclusions of an Ag-Cu-Se-S phase and naumannite. (B) Pyrite crystal rimed by naumannite, acanthite, mckinstryite and chalcopyrite. (C) Pyrite crystal with klockmannite at its edge. Mineral abbreviations: Kl = klockmannite; Se-Cov = selenian covellite; Nau = naumannite; Aca = acanthite; Mck = mckinstryite; Cla = clausthalite; Py = pyrite; Cha = chalcocite; Cpy = chalcopyrite; Bar = barite.

CHAPTER III

3.7.2 – Au-rich phases

Three Au-bearing phases were identified and include electrum, fischesserite (Ag3AuSe2) and palladian gold. Fischesserite was found in a well-preserved quartz-kaolinite vein (Fig. 3.2A), as anhedral crystals up to 30 µm in size. They occur within the kaolinite-sericite-quartz matrix or included in quartz, usually intergrown with electrum, selenian covellite (Fig. 3.5 A) and bismuth- vanadium oxides (Fig. 3.5 B), as individual crystals or as a thin rim at the edge of pyrite (Fig. 3.5 C) and electrum grains (Fig. 3.5 D). The composition of most of the analyzed grains deviates significantly from the stoichiometric fischesserite (Ag3AuSe2) and naumannite (Ag2Se) compositions, and the fischesserite varies over a consistent solid solution trend from more Ag- rich to more Au-rich compositions that are also elevated in Se (Fig. 3.5 E – Table 3.4). This is clearly is not an effect of mixed signals during EMP analysis since there are no naumannite crystals co-existing with fischesserite.

Table 3.4 – Electron Microprobe Analyses of Fischesserite.

wt % 100 % Normalized Grain Se Ag Au Total Se Ag Au No. 1 24.5 48.1 28.8 101.5 24.1 47.5 28.4 2 27.5 44.0 27.4 98.9 27.8 44.5 27.7 2 27.0 42.0 30.2 99.3 27.2 42.4 30.5 3 26.1 43.9 30.7 100.7 25.9 43.6 30.5 3 25.2 46.3 27.1 98.5 25.6 47.0 27.5 4 26.8 46.9 27.7 101.3 26.4 46.3 27.3 4 24.3 47.7 29.4 101.4 23.9 47.1 29.0 5 25.0 54.7 22.3 101.9 24.5 53.6 21.9

6 23.5 52.5 26.6 102.6 22.9 51.2 25.9 6 24.1 53.6 24.2 101.9 23.6 52.6 23.8 6 23.3 53.9 25.3 102.5 22.7 52.6 24.7 7 23.6 47.2 30.1 101.0 23.4 46.8 29.9

7 24.3 52.7 23.8 100.8 24.1 52.3 23.7 8 23.4 49.1 26.0 98.4 23.7 49.9 26.4 8 25.2 49.6 25.0 99.7 25.3 49.7 25.0 9 20.5 52.3 28.7 101.6 20.2 51.5 28.3

9 22.9 55.8 23.2 101.9 22.5 54.8 22.7 9 22.0 57.3 21.4 100.7 21.8 56.9 21.3 10 20.9 52.3 28.5 101.7 20.5 51.5 28.0 10 22.3 56.3 19.1 97.7 22.9 57.6 19.5

11 24.3 49.4 27.1 100.7 24.1 49.0 26.9 11 24.3 52.4 21.6 98.3 24.7 53.3 22.0 12 23.1 54.7 21.8 99.6 23.2 54.9 21.9

CHAPTER III

Figure 3.5 – Back-scattered images and composition of fischesserite from the carbon-rich argillic ore. (A) Fine-grained intergrowth between electrum, fischesserite and Se-covellite. (B) Aggregate showing co-existing electrum, Bi-V oxides and fischesserite invaded by needle-shaped crystals of supergene waylandite (BiAl3

(PO4)2(OH)6. (C) Pyrite crystal overgrown by klockmannite and fischesserite. (D) Early pyrite overgrown sequentially by Se-covellite, electrum, fischesserite and klockmannite, suggesting a sulfur-decrease (selenium increase) pathway. (E) Triangular diagram showing compositional variations (wt. %) for fischesserite and non- co-existing naumannite grains; black open circles denote fischesserite and naumannite stoichiometric compositions. Mineral abbreviations: Kl = klockmannite; Au = electrum; Bi-V-ox = bismuth-vanadium oxide; Py = pyrite; Fs = fischesserite; Se-cov = selenian covellite; Way = waylandite.

CHAPTER III

Electrum is the most abundant Au-bearing mineral at the Serra Pelada deposit, and is present in all studied samples. In the best preserved, un-weathered samples, primary electrum is mostly fine grained and only locally reaches >50 µm (Fig. 3.6 A). It forms anhedral crystals scattered in the kaolinite matrix, included in quartz, intergrown with or included in rutile, monazite, selenian covellite, hematite and fischesserite. In weathered samples, supergene electrum has a larger grain size from >100 µm up to visible gold dendrites several mm in size, included or precipitated in contact with goethite and/or manganese oxides (Fig. 3.6 B) and associated secondary phosphates.

The composition of primary electrum shows a variable content of Pd and Ag (from 3 to 9 wt. % - Fig. 3.6 C, D) and also Pt, Fe and Cu (Fig. 3.6 E). All compositional plots shown in Figure 3.6 (table 3.4) exhibit a clear trend of primary electrum and palladian gold towards supergene gold of high purity (>98% wt. Au). Palladian gold is restricted to the kaolinite-rich argillic ore, normally as isolated crystals in the kaolinite-rich matrix, and forms a distinct phase coexisting with silver- rich electrum (gray crystals in Fig. 3.6 A). This palladian gold is morphologically and chemically different from that described by Cabral et al. (2002a, b). These authors described palladian gold from the near-surface, bonanza-style ore, which is characterized by arborescent crystals coated by goethite and remarkably homogeneous composition (Au7Pd) free of silver. The palladian gold identified within the kaolinite-rich argillic ore of the present study is made of fine-grained (< 10 µm), anhedral crystals (Fig. 3.6 A) with a different composition (Table 3.5) from the bonanza- related ore, showing a higher content of palladium (6-11 wt. % Pd – Fig. 3.6 F) and also contains considerable amounts of silver (up to 2.3 wt. % - Fig. 3.6 C-E).

CHAPTER III

Figure 3.6 – Reflected light photomicrographs and compositional diagrams data of primary hydrothermal palladian gold, primary and secondary electrum grains within carbon- and kaolinite-argillic rich ore at Serra Pelada. (A) Fine-grained intergrown crystals of palladian gold (grayish grains) and electrum. (B) Typical secondary ore showing coarse-grained fine pureness Au (> 98 wt. % Au), normally included or associated with goethite and Pd-Pt alloys. (C) Triangular diagram showing compositional variations of Ag, Pd and Au in electrum and palladian gold. (D-F) Variation diagrams between total Au (wt. %) against ∑ of other metals (Ag, Pt, Pd, Fe and Cu), Ag and Pd showing a trend of increasing gold pureness within the clearly weathered samples.

CHAPTER III

Table 3.5 – Electron Microprobe Analyses of Electrum and Palladian-gold.

Grain No. As Se Cu S Pd Fe Ag Au Pt Total Mineral

1 0.0 0.1 0.0 0.0 0.0 0.0 8.9 92.8 0.0 101.8 Electrum - primary

2 0.0 0.0 0.3 0.0 2.8 2.2 8.7 81.9 2.9 98.79 Electrum - primary

3 0.0 0.1 0.0 0.0 0.0 0.0 7.9 93.7 0.0 101.6 Electrum - primary 4 0.0 0.1 0.0 0.0 0.0 0.0 6.4 89.6 0.0 96.1 Electrum - primary 5 0.0 0.0 0.0 0.0 0.0 0.0 5.1 96.9 0.0 102.0 Electrum - primary 6 0.0 0.1 0.0 0.0 0.0 0.0 4.9 96.0 0.0 101.0 Electrum - primary 7 0.0 0.0 0.0 0.0 0.1 0.0 4.8 96.6 0.7 102.2 Electrum - primary 8 0.0 0.0 0.0 0.1 0.0 0.0 4.8 97.0 0.0 101.9 Electrum - primary 9 0.0 0.0 0.0 0.0 1.1 0.0 4.7 95.9 0.1 101.8 Electrum - primary 10 0.0 0.0 0.0 0.0 0.2 0.0 4.6 98.8 0.1 103.8 Electrum - primary 11 0.0 0.0 0.0 0.1 0.0 0.0 4.5 97.3 0.0 101.9 Electrum - primary 12 0.0 0.1 0.0 0.0 0.0 0.0 4.5 95.9 0.0 100.5 Electrum - primary 12 0.0 0.1 0.0 0.0 0.0 0.0 4.4 96.4 0.0 101.0 Electrum - primary 13 0.0 0.2 0.0 0.0 0.0 0.0 4.4 95.1 0.0 99.7 Electrum - primary 13 0.0 0.2 0.3 0.1 0.0 0.0 4.4 93.5 0.0 98.4 Electrum - primary 13 0.0 0.1 0.0 0.0 0.0 0.0 4.4 96.6 0.0 101.1 Electrum - primary 14 0.0 0.1 0.0 0.0 0.0 0.0 4.2 94.7 0.0 99.1 Electrum - primary 15 0.0 0.1 0.0 0.0 0.0 0.3 4.0 94.7 0.0 99.1 Electrum - primary 16 0.0 0.0 0.1 0.1 0.0 0.0 3.6 97.9 0.0 101.7 Electrum - primary 16 0.0 0.1 0.0 0.0 0.0 0.0 3.6 98.0 0.0 101.7 Electrum - primary 17 0.0 0.0 0.0 0.0 0.0 0.0 3.5 97.2 0.0 100.7 Electrum - primary 17 0.0 0.1 0.1 0.1 0.0 0.0 3.5 98.2 0.0 102.0 Electrum - primary 18 0.0 0.0 0.1 0.1 0.0 0.0 3.4 95.0 0.0 98.6 Electrum - primary 19 0.0 0.1 0.0 0.0 0.0 0.0 3.3 97.1 0.0 100.5 Electrum - primary 19 0.0 0.1 0.0 0.0 0.0 0.0 3.3 98.2 0.0 101.6 Electrum - primary 20 0.0 0.0 0.1 0.1 0.0 0.0 3.2 95.0 0.0 98.3 Electrum - primary

21 0.0 0.1 0.0 0.0 0.0 0.0 3.2 97.0 0.0 100.3 Electrum - primary

1 0.0 0.0 0.2 0.1 0.0 0.1 2.7 98.4 0.0 101.6 Electrum-supergene 2 0.0 0.1 0.5 0.1 0.0 0.8 2.7 96.1 0.0 100.2 Electrum-supergene 3 0.0 0.0 0.3 0.1 0.0 0.2 2.1 96.9 0.0 99.6 Electrum-supergene 4 0.0 0.0 0.1 0.0 0.1 0.0 1.9 97.9 0.1 100.1 Electrum-supergene 6 0.0 0.0 0.0 0.0 0.1 0.0 1.7 99.9 0.0 101.7 Electrum-supergene 7 0.0 0.1 0.0 0.0 0.0 0.0 1.5 99.2 0.0 100.7 Electrum-supergene 8 0.0 0.1 0.0 0.0 0.3 0.8 1.5 96.9 0.0 99.5 Electrum-supergene

9 0.0 0.0 0.0 0.0 1.8 0.0 1.3 96.6 0.0 99.8 Electrum-supergene

10 0.0 0.0 0.0 0.0 2.6 0.3 0.2 97.5 0.2 100.9 Electrum-supergene

1 0.0 0.0 0.0 0.0 6.2 0.0 1.4 93.8 0.0 101.4 Palladian-gold 2 0.0 0.0 0.0 0.1 5.8 0.0 1.5 89.7 0.0 97.0 Palladian-gold 3 0.0 0.1 0.0 0.0 8.3 0.0 1.9 87.4 0.0 97.7 Palladian-Gold 4 0.0 0.1 0.0 0.0 8.8 0.0 1.9 91.8 0.0 102.6 Palladian-gold 5 0.0 0.1 0.0 0.0 11.4 0.0 2.3 88.0 0.0 101.8 Palladian-gold

CHAPTER III

3.7.3 – Pd-Pt-rich mineral phases

Primary platinum and palladium minerals at the Serra Pelada Au - Pd - Pt deposit are sulfide, selenide and arsenic-antimony minerals. They are mostly found as inclusions in quartz (Fig. 3.7 A, B), hematite (Fig. 3.7 C), or other gangue mineral and less commonly as individual grains in the kaolinite matrix, normally showing rims which were later partly weathered to pure metal alloys (Fig. 3.7 D). The main palladium arsenide-antimony mineral are mertieite-II (Pd8(As,Sb)3) and isomertieite (Pd11As2Sb2) showing low contents of Pd. It occurs as individual anhedral fine- grained (<30 µm) crystals within the kaolinite matrix, in quartz – kaolinite – hematite, veinlets or included in quartz, rutile, hematite or monazite. Isomertieite (Pd11As2Sb2) and palladseite

(Pd17Se15) were reported by Cabral et al. (2002a) from bonanza-style ore. Platinum minerals consist of sudovikovite (PtSe2), sperrylite (PtAs2) and ferroan-platinum (Pt,Fe). Palladium- platinum sulfides are normally very fine-grained, at the limit of EMPA analysis, but one crystal of Se-braggite ((Pt,Pd,Ni)S-Se yielded 48 wt. % Pd and 15% Pt (Table 3.6).

3.7.4 – Secondary PGE alloys

During deep tropical weathering, calcareous rocks have been largely decalcified and manganese oxides formed by weathering of the dolomitic sandstone unit cemented by Mn-rich dolomite. Grainger (2003) interprets the formation of the collapse breccias to be related to volume reduction due to decalcification of the sequence of impure dolomite and dolomitic sandstone. Primary hydrothermal Fe- oxides and sulfides of the ore zone were transformed to hydrated Fe- oxides including goethite.

The primary PGM react to purer metals of variable minor-element composition, normally included or surrounded by goethite or manganese oxides. Some of these occur as extremely fine- grained crystals, mainly among the matrix grain boundaries such as in figure 3.7 D or may show colloform textures (Fig. 3.7 E), and their different morphology correlates with distinct compositional trends (Fig. 3.7 F). Compositions vary between essentially palladium-rich and platinum-rich alloys with minor contents of Au, Cu and Fe, coexisting with essentially pure supergene gold. The supergene PGE alloys can be divided into (1) an essentially Pd-rich alloy (up to 68 wt. % Pd) with considerable amounts of Cu (up to 16 wt. %), Ag (up to 15 wt. %) and traces of iron (> 1 wt. %); and (2) a Pt-rich needle-shaped alloy showing an inverse-proportional relationship between Pt and Pd (up to 46 wt. % Pd), small amounts of gold (5-8 wt. % Au) and silver (1-3 wt. %). Table 3.6 summarizes the composition of the analyzed Pt- and Pd-rich alloys.

CHAPTER III

Figure 3.7 – Back-scattered images of relicts of primary Pt and Pd minerals in the process of partial decomposition to secondary, likely supergene minerals. (A) Kaolinite-rich argillic ore with secondary fine- grained Pd-Pt alloys, coarse-grained gold and primary relict minerals included in quartz. (B) Detail of a quartz crystal shown in A, with abundant inclusions of sperrylite(Spy) and sudovikovite (Sud). Note that a coarser secondary Pt-Pd alloy precipitated within a later crack. (C) Detail of hematite flakes with fine-grained associated (Pt,Pd,Ni)(S,Se). (D) Mertieite-II showing rims weathered to Pd-Cu alloy. (E) Colloform texture of some of the fine-grained Pt- and Pd-rich alloys. (F) Pt-Pd alloy aggregate showing needle-shaped Pt-rich alloys and fine grained anhedral Pd-rich alloys.

CHAPTER III

Table 3.6 – Electron Microprobe Analyses of Pt-Pd Minerals and alloys.

No. As Se Cu S Sb Pd Ni Fe Ag Au Pt Total Mineral

1 9.6 n.a. n.a. n.a. 17.7 70.5 n.a. n.a. n.a. n.a. n.a. 97.8 Pd poor Isomertieite 1 9.9 n.a. n.a. n.a. 17.9 70.1 n.a. n.a. n.a. n.a. n.a. 97.9 Pd poor Isomertieite 2 9.9 n.a. n.a. n.a. 17.9 70.2 n.a. n.a. n.a. n.a. n.a. 97.9 Pd poor Isomertieite 2 9.9 n.a. n.a. n.a. 17.9 70.2 n.a. n.a. n.a. n.a. n.a. 98.0 Pd poor Isomertieite 3 9.6 n.a. n.a. n.a. 18.0 70.8 n.a. n.a. n.a. n.a. n.a. 98.4 Pd poor Isomertieite 3 9.7 n.a. n.a. n.a. 17.9 70.9 n.a. n.a. n.a. n.a. n.a. 98.5 Pd poor Isomertieite 3 10.0 n.a. n.a. n.a. 18.1 71.2 n.a. n.a. n.a. n.a. n.a. 99.3 Pd poor Isomertieite 4 10.7 n.a. n.a. n.a. 18.2 71.6 n.a. n.a. n.a. n.a. n.a. 100.4 Pd poor Isomertieite 4 10.7 n.a. n.a. n.a. 18.2 71.7 n.a. n.a. n.a. n.a. n.a. 100.6 Pd poor Isomertieite

1 3.6 n.a. n.a. n.a. 28.2 67.4 n.a. n.a. n.a. n.a. n.a. 99.2 Pd poor Mertieite-II 1 3.9 n.a. n.a. n.a. 28.5 67.2 n.a. n.a. n.a. n.a. n.a. 99.6 Pd poor Mertieite-II 2 3.8 n.a. n.a. n.a. 28.6 67.5 n.a. n.a. n.a. n.a. n.a. 99.9 Pd poor Mertieite-II 2 3.9 n.a. n.a. n.a. 28.7 67.4 n.a. n.a. n.a. n.a. n.a. 99.9 Pd poor Mertieite-II 3 3.7 n.a. n.a. n.a. 28.7 67.7 n.a. n.a. n.a. n.a. n.a. 100.0 Pd poor Mertieite-II 3 3.7 n.a. n.a. n.a. 28.7 67.8 n.a. n.a. n.a. n.a. n.a. 100.2 Pd poor Mertieite-II

1 n.a. 0.0 0.1 0.1 n.a. 3.4 n.a. 15.3 0.6 7.5 76.3 103.3 Ferroan platinum

1 0.2 7.3 0.2 13.8 0.0 48.6 1.6 0.2 8.0 1.2 15.2 96.3 Se-Braggite 1 n.a. 0.1 15.0 0.0 n.a. 67.9 n.a. 0.9 14.3 0.2 1.0 99.4 Pd-rich alloy 2 n.a. 0.1 14.3 0.0 n.a. 67.0 n.a. 1.7 15.1 0.2 1.0 99.4 Pd-rich alloy 3 n.a. 0.1 14.6 0.0 n.a. 69.0 n.a. 1.3 13.9 0.1 1.2 100.2 Pd-rich alloy 4 n.a. 0.0 16.3 0.0 n.a. 67.8 n.a. 1.0 14.6 0.2 1.8 101.7 Pd-rich alloy 5 n.a. 0.1 14.5 0.0 n.a. 55.9 n.a. 4.0 12.1 0.3 2.7 89.5 Pd-rich alloy 6 n.a. n.a. n.a. n.a. n.a. 46.4 n.a. n.a. 3.6 5.1 46.4 101.5 Pt-rich alloy

7 n.a. n.a. n.a. n.a. n.a. 37.8 n.a. n.a. 2.8 5.6 51.9 98.0 Pt-rich alloy

8 n.a. n.a. n.a. n.a. n.a. 28.8 n.a. n.a. 2.4 6.0 63.0 100.2 Pt-rich alloy

9 n.a. n.a. n.a. n.a. n.a. 24.8 n.a. n.a. 1.9 6.2 67.5 100.5 Pt-rich alloy

10 n.a. n.a. n.a. n.a. n.a. 22.9 n.a. n.a. 1.8 6.8 70.6 102.0 Pt-rich alloy

n.a. n.a. n.a. n.a. n.a. 20.3 n.a. n.a. 1.4 7.2 73.6 102.5 Pt-rich alloy 11 12 n.a. n.a. n.a. n.a. n.a. 15.5 n.a. n.a. 1.1 7.4 79.7 103.7 Pt-rich alloy 13 n.a. n.a. n.a. n.a. n.a. 13.8 n.a. n.a. 1.0 7.4 77.3 99.5 Pt-rich alloy 14 n.a. n.a. n.a. n.a. n.a. 13.4 n.a. n.a. 1.0 7.8 78.8 101.0 Pt-rich alloy 15 n.a. n.a. n.a. n.a. n.a. 17.2 n.a. n.a. 1.4 7.8 74.7 101.2 Pt-rich alloy 15 n.a. n.a. n.a. n.a. n.a. 21.4 n.a. n.a. 1.8 7.9 69.6 100.7 Pt-rich alloy 17 n.a. n.a. n.a. n.a. n.a. 16.3 n.a. n.a. 1.2 8.0 76.1 101.6 Pt-rich alloy 18 n.a. n.a. n.a. n.a. n.a. 16.6 n.a. n.a. 1.6 8.2 75.1 101.4 Pt-rich alloy 19 n.a. n.a. n.a. n.a. n.a. 12.4 n.a. n.a. 1.0 8.3 80.6 102.2 Pt-rich alloy 20 n.a. n.a. n.a. n.a. n.a. 19.8 n.a. n.a. 1.4 8.3 73.0 102.6 Pt-rich alloy

CHAPTER III

3.8 – DISCUSSION

The hydrothermal mineralogy of Serra Pelada provides indications for the physical conditions of ore formation at elevated temperatures and indicates a distinct chemical environment of low sulfur but high selenium activity involving strong redox gradients.

3.8.1 – Temperature constraints from the composition of selenide minerals

The Cu-Se system is a useful system for geothermometry (Simon and Essene, 1996). Published phase diagrams in the Cu-Se system are based on experimental work by Heyding (1966) and

Chakrabarti and Laughlin (1981) and include Cu2Se, Cu3Se2, CuSe, CuSe2 and a Cu2-xSe phase of variable composition. All experimentally synthesized phases correspond to minerals that are found in nature, and three of them are present in association with Au-bearing minerals at the

Serra Pelada deposit, namely CuSe (klockmannite), athabascaite (Cu5Se3) and CuSe2 (krutaite). Klockmannite is stable in three structural modifications up to its melting point at 377°C, while krutaite has a melting point of 332°C (Chakrabarti and Laughlin, 1981), setting an upper temperature limit for the formation of the Serra Pelada deposit.

Experimental data on the ternary Au-Ag-Se phase diagram was published by Wiegers (1976) and more recently by Echmaeva and Osadchii (2009) is represented by the pseudo-binary temperature - composition section Ag2Se – Ag3AuSe2. The phase diagram (Echmaeva and Osadchii, 2009) in figure 3.8 shows three fields: (1) a two-phase field below 130 ºC, where the ordered forms of β-Ag3AuSe2 (fischesserite) and β-Ag2Se (naumannite) coexist. The mutual solubility between these two phases is very small and compositions are close to stoichiometric

(Wiegers, 1973); (2) a two phase field above 130.85 ºC, where β-Ag3AuSe2 co-exists with the disordered body-centered cubic <(BCC) solid solution, taking up increasing proportions of

Ag3AuSe2, into the BCC solid solution up to 267 ºC, the highest point of the transition curve – Fig. 3.8) where the fischesserite composition attains the BCC structure; (3) a one phase field above the α-phase transition curve, where BCC phase is stable across the entire composition range up to the melting point of Ag3AuSe2 at 742 ºC (Wiegers, 1973). The disordered modification of the BCC solid solution can also be enriched in Au beyond the composition of

Ag3AuSe2, where specific sites in the crystal structure are gradually occupied by excess Au atoms, thereby leading to higher gold contents than stoichiometric fischesserite (Wiegers, 1976; Simon and Essene, 1996).

The composition of “fischesserite” grains at Serra Pelada show both higher Ag to (Au + Se) ratios than stoichiometric fischesserite (Table 3.3, Fig 3.5). It would be expected from the phase

CHAPTER III diagram that any Ag-rich “fischesserite” could exist only within the high-temperature phase transition, since no solid solution exists between the low temperature phases. The composition of the Ag-rich “fischesserite from Serra Pelada corresponds to a solid solution containing 0.1 to

0.3 weight fraction of Ag2Se and 0.7 to 0.9 of Ag3AuSe2 (Fig. 3.8). Hence, the presence of fischesserite grains with composition close to stoichiometric, co-existing with other grains that show higher Ag and locally Au and Se, suggest that the “fischesserite” from Serra Pelada was originally precipitated as the BCC solid solution at about 260 and 267 ºC since at higher temperatures a single phase of fixed composition would the stable. This does not preclude that the original high temperature fischesserite (α-phase) partly ex-solved upon cooling on a submicroscopic scale, but it may also have been preserved meta-stably since experimental data records large hystereses during slow cooling (Wiegers, 1976).

The relatively high temperature compared to prior estimates (175 ± 25 °C – Grainger, 2003) is consistent with the crenulation structures (Berni et al., 2014; Fig. 3.2 A) showing that mineralization occurred at conditions where rocks were deformed partially in brittle, partially in ductile regime in association with stylolitic veins and crenulation cleavage.

Figure 3.8 – Phase diagram for the pseudobinary Au2Se-Ag3AuSe2 system (from Echmaeva and Osadchii, 2009) showing compositional range of “fischesserite” grains from the Serra Pelada Au - Pd - Pt deposit. The refined position of the eutectoid composition is shown in the inset. BCC is a solid solution with a body- centered cubic structure.

CHAPTER III

3.8.2 – Serra Pelada, a sulfide-deficient hydrothermal system: evidences from selenide and sulfide phase relations

Selenide minerals generally display more restricted stability fields within the ƒSe2 (g) vs. ƒS2 (g) and

ƒSe2 (g) vs. ƒO2 (g) diagrams than their sulfide counterparts, and therefore can be used to constrain those variables. In addition, the presence of some selenide minerals can also be used to assess whether a mineral association may represents an equilibrium assemblage or not (Simon et al., 1997).

Figure 3.9 depicts the stability fields of Cu, Fe, Au, Ag and Bi selenide, sulfide and oxide minerals with respect to the fugacity of S2, Se2 and O2. Diagrams of log ƒSe2(g) vs. log ƒS2(g) (Fig.

3.9 A) and log ƒSe2(g) vs. log ƒO2(g) (Fig. 3.9 B) follow Simon and Essene (1996) for a temperature of 300°C, which is close to the formation of the Serra Pelada deposit, even though phase relations would not change significantly at a lower temperature of 150 ºC. The two diagrams show that the paragenetically early assemblage A (yellow circle A in Fig. 3.9) comprising pyrite + chalcopyrite + acanthite + naumannite cannot coexist with the copper selenides klockmannite (CuSe) + krutaite (CuSe2 ; point B in Fig. 3.9). Assemblage A is supported by textural evidence shown in Figure 3.4C, where pyrite is totally rimmed by naumannite (Ag2Se) and acanthite (Ag2S), indicating at least local equilibrium. A later assemblage B with Se-covellite + chalcocite + klockmannite + krutaite is consistent with the presence of Bi- V oxides, fischesserite and electrum. The co-existence of Bi-V oxides and fischesserite requires high ƒO2 (g) and high ƒSe2 (g) values (point B in Fig. 3.9 A and B). In figure 3.9, the stability fields

AuSe and Bi2O3 are shown as an approximation for lack of thermodynamic data for the more complex solid solution phases. At least qualitatively, however, the mineral associations at Serra Pelada together with published phase stability relations indicate a gradient from moderately sulfidic conditions with pyrite + chalcopyrite + silver sulfide ± hematite (assemblage A) towards an increase in ƒSe2(g) and ƒO2(g) stabilizing Cu and Au-Ag selenides + Bi-V-oxides + hematite (assemblage B; Fig. 3.9).

This chemical gradient can be explained in the larger context of alteration zoning and inferred fluid processes at Serra Pelada. Berni et al. (2014) describes distal alteration zones including spatially separated carbon-chlorite and hematite alteration at Serra Pelada, suggesting fluid mixing between highly oxidizing (well in the hematite stability field) and highly reduced fluid end-members (carbon precipitation field). Assemblage A can be explained by reduction of

HSO4-2/ SO4-2 to H2S, leading to the precipitation of early sulfide minerals.

The sulfide-bearing assemblage was probably dominated by the reducing influence of organic carbon, either by fluid-rock reaction with solid carbon or by mixing or an oxidized fluid with a 58

CHAPTER III greater proportion of reduced (e.g. methane bearing) fluid. The presence of silver selenide in sulfide-bearing assemblage shows that the fluid in contact with assemblage A was Se-bearing, but perhaps less so than in the high ƒSe2 (g) assemblage B due to depletion by incipient selenide mineral precipitation. The sulfide-free and more oxidized assemblage B was dominated by a Se- rich fluid, which contained sulfate and probably carried the main economic ore metals Au, Pd and Pt as well as Cu. We infer that the primary ore minerals were precipitated as selenides, arsenides and minor sulfide minerals as a result of reduction by overwhelming reaction of this oxidized fluid with carbonaceous matter or by a mixture of the reducing (e.g. CH4-bearing) fluid.

Figure 3.9 – Composition diagram showing the relative stability fields of selenide, sulfide and oxide from Simon and Essene. (1996). Yellow symbols labelled A and B represent the position of identified mineral assemblages. Abbreviations: bn = bornite; py = pyrite; cpy = chalcopyrite; po = pyrrhotite; fa = fayalite; hem = hematite; mt = magnetite;

The mineralogy and suggested process of ore formation of Serra Pelada shows similarities to other selenide bearing deposits, including unconformity-related uranium deposits and low- temperature selenide vein and epithermal Au-Ag deposits. Low-temperature selenide veins occur in several locations in Germany (Stanley et al., 1990) and Bolivia (Grundmann et al., 1990), in which carbonate minerals, hematite, rare sulfides, gold, and platinum and palladium selenides were formed at temperatures around 150 ºC (Shepherd et al., 2005).

Unconformity-related uranium deposits contain Cu, Co and Ni selenides associated with uraninite, hematite, Cu, Co and Ni sulfides. Such deposits occur in the Bohemian massive in Czech Republic, in the Massif Central of France, in the Athabasca basin of Saskatchewan (Simon et al., 1997). Deposition temperature in these deposits is typically low, as indicated by the presence of umangite (Cu3Se2), which is stable below 112 ºC only (Simon et al., 1997). Unconformity-related uranium deposits in Northern Australia were probably formed at higher

CHAPTER III temperature (~200 ºC at ƒO2 (g) above the magnetite-hematite buffer (Wilde and Wall, 1987). The Au-Pd-Pt-U deposit of Coronation Hill in the Northern Australian uranium province also contains abundant Pt and Pd selenides, precipitated at a minimum temperature of 160°C based on fluid inclusion homogenization (Mernagh et al, 1994). The Serra Pelada deposit was probably formed at even higher temperature and pressure than the Australian unconformity-related deposits, yet shows mineralogical and geochemical similarities with low temperature vein deposits, with the notable presence of Pd and Pt selenides.

3.9 – CONCLUSIONS

1. The primary ore mineralogy of the Serra Pelada Au - Pd - Pt deposit is characterized by sulfide, selenide, arsenide and antimony-arsenide Au - Pd - Pt minerals. The main primary ore assemblages range from a more reduced assemblage including sulfides and minor selenides (pyrite + chalcopyrite + naumannite + acanthite + clausthalite + galena) to a more oxidized assemblage of Cu selenides, sulfoselenides and oxides (Se-covellite, klockmannite + krutaite + fischesserite + electrum + Bi-V oxides anhydrite + barite + hematite).

2. The deposit was probably formed at a minimum temperature of ca. 260-267 °C and maximum of 332 °C, as indicated by the composition of fischesserite solid solution, mineral which is part of the more oxidized assemblage overprinting the early sulfide-bearing assemblage.

3. The range of hydrothermal mineral assemblages in the Serra Pelada deposit indicates a fluid- compositional gradient between low ƒO2 (g) and moderately high ƒS2 (g) towards conditions where ƒSe2 (g) to ƒS2 (g) ratios were higher than unity and ƒO2 (g) values high up in the hematite stability field.

4. The secondary, weathering related mineralogy is associated mainly with goethite, manganese oxides and phosphates. The majority of primary Pt and Pd minerals are weathered into near pure metal alloys and are only preserved within mineral inclusions or slightly weathered samples.

3.10 – ACKNOWLEDGMENTS FOR THE CHAPTER

We thank Colossus Minerals Inc. for the funding of this research and permitting the publication of the data. GVB is thankful to all Colossus employees and the garimpeiros from Serra Pelada, for all their assistance during the time on site.

CHAPTER IV

FLUID MIXING AND REDOX GRADIENTS AT THE SERRA PELADA Au - Pd - Pt DEPOSIT

(Gabriel V. Berni, Christoph A. Heinrich, Markus Wälle, Vic Wall)

*To be submitted to Economic Geology

4.1 – ABSTRACT

Serra Pelada is a world class Au - Pd - Pt deposit located near the eastern border of the Amazon Craton. The mineralization is hosted by metasedimentary rocks of the Águas Claras Formation and is structurally hosted by the recumbent Serra Pelada syncline, which is tectonically overlain by the mafic/ultramafic rocks of the Rio Novo Group to lie above the present exposure of the deposit. The fold structure consists of dolomitic metasandstones at the outer part of the fold hinge, a layer of carbonaceous siltstone in which the orebodies are located, and a red siltstone in the core of the fold. Distal hydrothermal alteration comprises spatially separated chlorite/carbon and hematite alteration zones. The alteration within the mineralized zones include silicification, carbon-, kaolinite- and hematite-rich argillic alteration. Fluid inclusion hosted by quartz and dolomite crystals from distal alteration zones and from silicified portions within the mineralized zone record a wide range of salinity (5-22 wt. % NaCleqv.) but relatively constant homogenization temperatures (160 ± 25 ºC). Fluid inclusion isochores were combined with mineralogical geothermometers to constrain the formation of the deposit at about 270 ºC and 1.7 kbar. Element concentration in individual fluid inclusions were determined by LA-SF- ICPMS analysis for K, Mg, Ni, Cu, As, Sb, Cs, Ba, Ce, Au, Pb, Bi and U, using a highly sensitive sector-field ICP mass spectrometer. Element concentrations of alkali metals, Pb, Cs and to a lesser extent Ba and Sb correlate with salinity. The maximum pH of the oxidizing/acidic fluids was below 3±0.5, estimated by using potassium concentrations of fluid inclusion and assuming muscovite-kaolinite equilibrium as an upper pH limit. The arsenic concentration of the fluid inclusions (ca. 100 ppm) was used to estimate the stabilization of sperrylite (PtAs2) relative to native platinum at about 5 log ƒO2(g) units higher than the hematite-magnetite buffer. Experimental thermodynamic data show that such an extremely oxidizing, hematite-saturated fluid is capable of significant (>10-100ppb) hydrothermal transport of Au, Pd and Pt in the form of chloride complexes. The occurrence of hydrothermally precipitated carbon implies that a reducing fluid was also present, and the structural setting in conjunction with vein

CHAPTER IV relationships suggests that mixing between a reducing and an acidic, oxidizing and metal- carrying fluids was the main process driving ore metal precipitation. Genesis of the Serra Pelada Au - Pd - Pt deposit shows similarities with unconformity-related U ± Au ± PGE deposits. Major differences in metal proportions of this broad class of ore deposits in different ore provinces are primarily caused by differences in source rocks available for oxidative leaching. The Au-PGE-rich but U-poor nature of Serra Pelada is probably related to the abundance of overthrust mafic and ultramafic source rocks and immature sediments derived from them.

4.2 – INTRODUCTION

Serra Pelada is an Au - Pd - Pt deposit located at the Carajás Mineral Province, eastern border of the Amazon craton, north Brazil. The deposit was discovered by artisanal miners in the early 1980’s and was the site of the biggest gold rush in South America’s late history. As a result of the tough mining conditions in an overcrowded pit, only few exploration drill core data and little descriptive work were published during the time the mine was operating (Jorge João et al., 1982; Meireles and Silva, 1982 and 1988). Recently, primary hydrothermal features were recognized by Berni et al. (2014), and the geology and materials of that recent study is the basis of the fluid inclusion study presented here.

An epigenetic mineralization has been suggested by previous authors (Moroni et al., 1999, 2001; Tallarico et al., 2000; Grainger et al., 2002), but a reliable genetic model has never been suggested due to the lack of proper characterization of the primary mineralogy and the hydrothermal fluid chemistry in this intensely weathered terrain. An intrusion related hydrothermal model has been suggested by Moroni et al. (2001) and Tallarico et al. (2000), while Grainger et al. (2002) proposed an distal association with the IOCG deposits of the Carajás Mineral province. Cabral et al. (2011) suggested a Late-Cretaceous hydrothermal overprint over the primary Paleoproterozoic (Grainger et al., 2002) mineralization.

Based on the preceding geological (Chapter II) and mineralogical studies (Chapter III) we performed detailed petrography, microthermometry and LA-SF-ICPMS analysis of fluid inclusions to constrain the hydrothermal process of ore formation. In conjunction with thermodynamic data, we suggest that redox gradients between highly oxidized, metal carrier fluids and reduced fluids of variable salinity controlled the metal precipitation at the Serra Pelada deposit.

CHAPTER IV

4.3 – GEOLOGICAL SETTING AND SAMPLING MATERIAL

The Serra Pelada Au - Pd - Pt deposit is located in the northern part of the Carajás Mineral Province, known as the Itacaiúnas Belt (Grainger et al., 2002). This major mining district is the host of world-class iron ore and IOCG deposits within two metavolcanosedimentary sequences and a granite-gneiss basement, which are overlaid by the Águas Claras Formation, host the Serra Pelada deposit. From stratigraphic base to top, the mine sequence at the Serra Pelada deposit comprises dolomitic metasandstones and metaconglomerates, a 20-30 m thick layer of gray metasiltstone and a red metasiltstone.

The mineralization at the Serra Pelada deposit is hosted by hydrothermally altered, metasedimentary rocks at the hinge of a recumbent syncline (Moroni et al., 1999; Tallarico et al., 2000; Grainger at al., 2002; Berni et al. 2014). Alteration fronts of contrasting redox potential are present within the deposit, comprising spatially separated distal chlorite/carbon and hematite alteration zones with a mineralized zone at the hinge of the syncline and in minor orebodies at lower and upper limbs (Fig. 4.1 A). Hydrothermal alteration in the ore zone comprises carbon-, kaolinite- and hematite-rich argillic alteration, which are enveloped by siliceous alteration around the fold hinge (Berni et al., 2014). The main hydrothermal minerals are quartz, kaolinite, sericite and monazite with a complex ore mineralogy made of sulfide, selenide, and antimony-arsenide minerals. A more detailed description of the geological setting and the hydrothermal alteration of the Serra Pelada Au-PGE deposit can be found in the second chapter of this thesis.

Out of a greater number of samples cut for doubly-polished plates, 9 samples (Table 4.1) were studied in detail, providing textural information and quantitative data from 64 fluid inclusion assemblages including 3 to 18 individual inclusions each. Table 4.2 depicts alteration zones, mineral host, and range of salinity and homogenization temperatures of each sample, average composition of fluid inclusion assemblages are listed in table 4.3 and individual fluid inclusion analyses are provided within the digital repository.

Selected samples consist of quartz and dolomite crystals from veins and partly open cavities including: (1) Distal hematite alteration, which converts the dolomitic sandstone into a red- dusted marble containing vugs lined by coarse-grained quartz and dolomite crystals (Fig. 4.1 B). (2) Distal carbon/chlorite alteration, which overprints the red siltstone at the lower limb of the Serra Pelada syncline, and changes the red, hematite bearing metasiltstone into a green, chlorite- bearing and locally carbon-rich metasiltstone. Chlorite replaces the metamorphic sericite of the red metasiltstone and carbon is precipitated, along the metamorphic banding or together with

CHAPTER IV quartz in veins (Fig. 4.1 C). (3) Euhedral quartz from breccia zones and vugs within the silicified halo from the ore zone (Fig. 4.1 D).

Figure 4.1 – Main alteration zones of the Serra Pelada Au - Pd - Pt deposit (Berni et al., 2014), sample material for fluid inclusion studies and their location. (A) Cross-section along the deposit showing the main lithological units, distribution of alteration zones and sample area location. (B) Red-dusted recrystallized dolomitic sandstone with co-existing coarse-grained dolomite with plentiful hematite inclusions and quartz separated by hematite-sulfide growth zones. (C) Chlorite altered red-siltstone with precipitated amorphous carbon and coarse grained quartz (red portions corresponds to goethite related to later weathering. (D) Fine- grained silicified dolomitic sandstone with coarse-grained quartz and hematite.

At the periphery of the Serra Pelada ore zone, the hematite alteration is made of a red-dusted recrystallized marble containing centimeter-scale quartz and dolomite crystals (Fig. 4.2 A). These breccia zones are preserved only in a few deep drill holes that extend below the level of

CHAPTER IV supergene leaching of the dolomitic sandstones, at locations where the shallow-dipping contact with metasiltstone was cut by brittle and probably steep fracture zones. Distinct growth zones within the vug crystals are lined with numerous micron-sized inclusions of hematite, sulfides (pyrite, chalcopyrite and Ni-sulfides) with a small-scale enrichment in Au, Pd and Pt, detected by LA-ICPMS analysis (Fig. 4.2 B). The dolomite (Fig. 4.2 C, D) and quartz (Fig. 4.2 E, F) crystals also contain abundant primary, pseudosecondary and secondary fluid inclusion assemblages consisting of two-phase, aqueous fluid inclusions.

Table 4.1 – Drill hole, depth, and alteration zone of the samples used for fluid inclusion analysis.

Drill hole Sample Alteration Description and depth

SPD-141 Chlorite altered red siltstone with quartz-carbon vein and S.196 Chlorite/carbon 68.2 m euhedral crystals at vein core. FD-0126 Coarse-grained silicified dolomitic sandsone with abundant S.064 Siliceous 223.1m hematite inclusions. SPD-007 Coarse-grained hematite dusted (red) silicified dolomitic S.136 Siliceous 198.6 m sandstone with vuggs filled by well formed quartz and hematite. SPD-051A Coarse-grained silicified dolomitic sandstone with vugs filled by S.140 Siliceous 214.5 m well formed quartz and hematite. SPD-007 Brown, clay-rich breccia with free standing quartz crystals within S.166 Siliceous 195.3 m the outer silicification zone. SPC-022 Recristalized dolomitic sandstone with vugs containing S.145 Hematite 320.0 m centimenter scale quartz crystals.

Recristalized dolomitic sandstone with vugs filled by centimenter SPC-022 S.147 Hematite scale coexisting quartz and dolomite crystals and abuntant 351.6 m hematite. SPC-022 Recristalized dolomitic sandstone with vugs containing S.141 Hematite 351.7 m centimenter scale dolomite and quartz crystals. S.142 Hematite SPC-022 Recristalized dolomitic sandstone with vugs containing 268.3 m centimenter scale dolomite cystals.

The chlorite-carbon alteration is a reduction front where precipitation of quartz is rare, but clearly associated with hydrothermal carbon precipitation in some samples (e.g. Fig. 4.1 C). Fluid inclusion assemblages with measurable size for microthermometry and LA-ICPMS analysis where identified only in one sample from this alteration type. The sample contains a quartz vein with abundant inclusions of amorphous carbon in the silicified wall rock, sub-euhedral fine- grained quartz at the vein walls and euhedral quartz towards the vein and goethite pseudomorphs at the vein core (after hematite/pyrite - Fig. 4.3 A). Fluid inclusions assemblages are made of aqueous fluid inclusions (Fig. 4.3 B) of primary and secondary origin. Locally, vapor or empty looking inclusions are found as isolated inclusions, or irregularly distributed within quartz crystals of this alteration zone.

CHAPTER IV

Figure 4.2 – Quartz-dolomite veins, vug textures and fluid inclusion assemblage relationships within some of the studied samples. (A) Vuggy dolomite with dark growth zones defined by numerous solid inclusions in relation to successive fluid inclusion assemblages. (B) Transient LA-SF-ICPMS signal of a 130µm ablation pit covering the dark growth zones shown in A, identifying micron-sized mineral inclusions including hematite and elements associated with the Serra Pelada Au - Pd - Pt mineralization, including U, and Bi as well as Au, Pt and Pd. (C) Pseudosecondary fluid inclusion assemblage hosted by dolomite. (D) Detail on an aqueous fluid inclusions hosted by dolomite. (E) Pseudosecondary fluid inclusion assemblage hosted by quartz within hematite distal alteration zone. (F) Detail on an aqueous fluid inclusions hosted by quartz.

CHAPTER IV

Figure 4.3 – (A) Fine-grained quartz with amorphous carbon inclusions and associated euhedral quartz. (B) Detail on an aqueous secondary fluid inclusions hosted by quartz within the chlorite/carbon alteration. (C) Vuggy quartz within silicified dolomitic sandstone with well-defined growth zones marked by globular hematite inclusions, secondary and pseudosecondary inclusion trails. (D) Detail on an aqueous fluid inclusion assemblage hosted by quartz within the siliceous alteration (E) Transient LA-SF-ICPMS signal of a fluid inclusion with a kaolinite accidentally trapped crystal (detail upper right) hosted by dolomite. Note the shift between the Na and K peaks (in solution) in comparison to Al peak.

Samples for characterizing the fluids within the ore zone were taken from the siliceous alteration, which replaces both metasandstone and metasiltstones near and in the mineralized zone at the Serra Pelada deposit. The siliceous alteration generally host low-grade Au - Pd - Pt mineralization at the periphery of high-grades ore shoots within the deposit, and is temporally related to the argillic alteration and ore metal introduction in the main orebody (Grainger et al., 2002;Berni et al., 2014). The alteration is formed by coarse-grained to microcrystalline quartz

CHAPTER IV with inclusions of sericite and hematite. Vugs with euhedral quartz lined by hematite (Fig 4.3 C) host primary, pseudosecondary and secondary containing two-phase, aqueous fluid inclusions (Fig 4.3 D) which were selected for detailed studies. Some inclusions host mineral daughter crystals which were identified as kaolinite (Fig 4.3 E) and sericite.

4.4 – ANALYTICAL METHODS

Samples with numerous, adequately sized (>15 µm) and clearly-defined assemblages of fluid inclusions were selected for detailed petrography studies and to establish their relative chronology. Groups of fluid inclusions that are coevally formed (according to the concept of fluid inclusion assemblage – FIA; Goldstein and Reynolds, 1994) were selected and analyzed by microthermometry in order to determine their apparent salinity (from ice melting temperatures) and homogenization temperatures. We used a Linkam THMSG-600 heating-freezing stage connected to a Leitz optical microscope, calibrated with synthetic fluid inclusions of pure CO2,

NaCl-H2O mixture of eutectic composition and pure water of known homogenization temperature for the microthermometry measurements.

In-situ analyses of individual fluid inclusion were performed by laser ablation inductively coupled sector field mass spectrometry (LA-SF-ICPMS) with a beam-homogenized 193 nm ArF Eximer laser ablation system (Günther et al., 1997) connected to a high-sensitivity sector field mass spectrometer (ThermoFisher, Element XR; Wälle and Heinrich, 2014). Preliminary runs were analyzed by LA-ICPMS, connected to a quadrupole ICPMS (Perkin Elmer Elan 6100 DRC; see result section for details). After some test-runs, we selected a reduced set of detectable isotopes related to the mineralization at Serra Pelada (Na23, Mg24, Al27, K39, Ni60, Cu65, As75, Sb121, Cs133, Ba137, Ce140, Au197, Pb208, and Bi209) and applied an increased dwell time of 30 ms on Au197 and U238 to improve sensitivity and 10 ms for all other elements. We ablated with a pulse rate of 10 Hz and energy densities per pulse of ca. 30 J/cm2 for quartz and 20 J/cm2 for dolomite hosted inclusions. Analytical and standardization procedures were performed according to Heinrich et al. (2003), using the NIST SRM 610 as an external reference material. Element concentrations of individual fluid inclusion were calculated using the SILLS software (Guillong et al., 2008).

CHAPTER IV

4.4 – RESULTS

4.4.1 – Microthermometry

Salinities were estimated from ice melting temperature and calculated as wt. % NaCl equivalent

(NaCleqv.) using data from Bodnar and Vityk (1994). During microthermometry measurements, ice was clearly the last phase to melt. Cyclic cooling and heating experiments were performed to induce hydrohalite nucleation, which was observable only within highly saline FIA’s (> 20 wt. %), where the calcium estimated calcium content was relatively small, generally between 1-2 wt. % CaCl). Within the majority of the low salinity FIA’s, the melting temperature of hydrohalite would lead to senseless results considering the NaCl-CaCl-H2O system either due to metastability of hydrohalite or the presence of magnesium component. All inclusions homogenized by bubble disappearance, providing a value for a minimum temperature.

Most fluid inclusion assemblages show consistent homogenization temperatures within ± 15 °C, with averages ranging from 140 ºC to 185 ºC (Fig. 4.4). This contrasts with a wide range of salinity from 2.6 to 23.5 wt. % NaCleqv. The average of final ice melting temperature (Tmice), calculated equivalent salinities (NaCleqv.), total homogenization temperatures (Thtot), time- relationships and the number of laser ablation analyses are summarized in table 4.2 and average composition of FIA’s are listed in table 4.3.

FIA’s hosted by dolomite and quartz within the hematite distal alteration zone display the biggest variation in salinities out of all analyzed samples. Salinity variations between primary fluid inclusion assemblages from 3.2 to 11.0 wt. % NaCleqv were observed within different growth zones such as shown in figure 4.2 A. Pseudosecondary FIA’s record smaller variation ranging from 4.4 to 6.2 wt. % NaCleqv., while secondary trails record variations from 3.9 to 23.5 wt. % NaCleqv.. Even though there is a large variation in salinity, all fluid inclusion assemblages have consistent total homogenization temperatures at about 165°C.

The quartz crystals filling vugs within the siliceous alteration also host fluid inclusion assemblages with consistent total homogenization temperatures around 160 ºC (except for a few

FIA of lower Thtot.). Salinity variations within these samples are smaller, from 3.3 to 8.8 wt. %

NaCleqv. although isolated fluid inclusions of higher salinity were observed during the microthermometry measurements.

Primary FIA’s in quartz from the chlorite/carbon alteration record a salinity variation from 2.6 to 11.9 wt. % NaCleqv. As observed for the fluid inclusion assemblages analyzed within the other

CHAPTER IV alteration zones, the total homogenization temperatures are consistent around 160 ± 25 °C (expect of one FIA).

Figure 4.4 - Plots of measured homogenization temperatures vs. total salinities. Each data point corresponds to an average value for one FIA, with 1σ standard deviation shown by gray error bars. Color codes at the legend denotes the alteration zone of the mineral host.

Table 4.2 – Summary of results from fluid inclusion microthermometry analysis.

No. No. No. LA-ICPMS Av. Std. Av. Std. Sample Alteration Host MT Time relationship NaCl Th FIA analysis eqv. NaCl tot Th meas. eqv. tot S.196 Chlorite/carbon Qz 5 36 17 Pseudosecondary 2.6 to 11.9 0.4 155 18

S.064 Siliceous Qz 4 34 20 Secondary/unknown 3.3 to 8.8 0.5 159 8

S.136 Siliceous Qz 2 42 20 Primary 3.4 to 4.0 0.4 159 8 S.136 Siliceous Qz 2 20 10 Pseudosecondary 3.8 to 4.4 0.4 161 8 S.136 Siliceous Qz 2 28 12 Secondary/unknown 3.3 to 5.0 0.2 163 12

S.140 Siliceous Qz 1 11 11 Primary 5.3 0.2 171 11 S.140 Siliceous Qz 1 12 10 Pseudosecondary 5.7 0.2 164 9 S.140 Siliceous Qz 1 4 4 Secondary/unknown 6.6 0.6 162 13

S.166 Siliceous Qz 2 23 9 Secondary/unknown 5.7 to 5.9 0.3 169 10

S.145 Hematite Dol 6 60 35 Primary 3.2 to 11.0 0.3 n.a n.a S.145 Hematite Dol 9 131 113 Pseudosecondary 4.5 to 6.3 0.1 n.a n.a S.145 Hematite Dol 1 8 6 Secondary/unknown 20.7 0.1 176 6

S.147 Hematite Qz/Dol 7 84 34 Pseudosecondary 4.2 to 5.4 0.3 168 9 S.147 Hematite Qz 7 58 33 Secondary/unknown 14.3 to 23.5 0.5 154 12

S.141 Hematite Qz 4 44 24 Secondary/unknown 3.9 to 20 0.8 165 10

S.142 Hematite Qz 5 40 39 Secondary/unknown 4.4 to 23.5 0.5 166 10

Abbreviations: No: number of; FIA: fluid inclusion assemblage; Qz: quartz; Dol: dolomite; No. MT meas. = number of microthermometry measurements. Std. Dev. = 1σ standard deviation.

CHAPTER IV

4.4.2 – Element concentration in fluid inclusions

FIA’s hosted by quartz and dolomite crystals from different alteration zones of the Serra Pelada deposit show consistent compositional variations (Table 4.3). Elemental plots shown in figures 4.5 to 4.7 are first grouped by each of the three major alteration types, and within each group sorted by increasing total salinity.

Major components

Samples from the chlorite/carbon distal alteration zone are Na dominated, and display the smallest concentrations of K (average ~1500 ppm). Within the siliceous alteration zone FIA’s consist of a higher content of potassium (average ~3000 ppm), reaching up to 10,000 ppm K within a consistent trend with increasing salinity. The FIA’s hosted by samples within the hematite alteration zone show a higher K average concentration (average ~7200 ppm). There is a consistent trend with increasing total salinity up to 7 wt. % NaCleqv., with concentrations up to 17,000 ppm (Fig. 4.5 A). Na/K ratios are also systematically variable (Fig. 4.5 B), with values above 20 within the carbon/chlorite alteration zone, and a decrease from 20 to 1 within FIA’s hosted by the siliceous alteration correlating with increasing salinity. The low to intermediate salinity FIA’s within the hematite alteration zone show a similar pattern in respect to Na to K ratios as the FIA’s hosted by siliceous, whereas the highest-salinity fluids have a Na/K ratio around 10.

A similar salinity correlation is observed for magnesium concentrations (Fig. 4.5 C). FIA’s hosted within the reduced alteration zone display lower Mg concentration, where FIA’s within the siliceous alteration show consistently higher magnesium concentrations (average ~750 ppm). The magnesium concentration within the hematite alteration correlates positively with total salinity, with average values reaching 10,000’s ppm in the most saline FIA’s.

CHAPTER IV

Figure 4.5 - Summary of average compositions for K, Na/K and Mg determined by LA-SF-ICPMS microanalysis, grouped by distal chlorite/carbon alteration – green; siliceous alteration in and around ore zone – blue; distal hematite alteration – red symbols for dolomite hosted and pink symbols for quartz hosted FIA’s. Each symbol represents the average value for a fluid inclusion assemblage (FIA), with the respective 1σ standard deviations of several detected values shown by the error bars (ignoring counting-statistical uncertainty). Within each of the three alteration groups, assemblages are sorted by increasing salinity

(NaCleqv.) indicated by colored values below the diagram. Thin dashes mark the lower detection limit when concentrations of all inclusions within the same assemblage were below detection.

CHAPTER IV

Minor and trace elements

Trace element concentrations for Cs, Ba, As and U were measureable for most of the analyzed fluid inclusion assemblages in many individual inclusions (Fig. 4.6; Table 4.3). Uranium concentrations (Fig. 4.6A) range from below detection in FIA’s from the chlorite/carbon alteration zone (average detection limit 0.15 ppm), to ~ 0.3 ppm in FIA’s of the siliceous zone, whereas in FIA’s from the hematite alteration zone some assemblages consistently host up to 2 ppm of U. Arsenic concentrations varies between 10 and 140 ppm within all analyzed FIA’s from all alteration zones, with consistent variations among assemblages but no clear petrographic trends or correlation with salinity (Fig. 4.6 B). Barium concentrations are quite variable within all analyzed FIA’s, ranging from 10-100 on average (Fig. 4.6 C). Cesium concentrations vary between 1 and 10 ppm (Fig. 4.6 D) within FIA’s from all alteration zones and show a consistent trend with increasing salinity. Copper concentrations are only occasionally above detection and averages display high variability within all alteration zones, with most assemblage averages on the order of 100 ppm or less (Fig. 4.6 E).

Nickel was detected in a few FIA’s and concentrations can reach up to a few hundreds ppm (Fig. 4.7 A). Cerium contents were measurable in about half of the analyzed FIA’s, with a distribution similar to U. The lowest Ce concentration are from FIA’s within the carbon / chlorite alteration (one FIA with 0.9 ppm Ce), reaching up to 2 ppm Ce within the siliceous alteration, and the highest Ce concentrations are associated with the hematite alteration (average 10 ppm Ce; Fig- 4.7 B). Lead concentrations do not show any correlation with alteration type, but correlate with increasing salinity, with values ranging from 2 to 100 ppm (Fig. 4.7 C). Detection limit for antimony was around 2 ppm for most assemblages and was detectable in a few FIA’s, ranging from 1-6 ppm where three assemblages yielded 20 to 35ppm of Sb (Fig. 4.7 D). Bi concentration was measured in a few assemblages and most values from 0.2 to 1 ppm Bi (Fig. 4.7 E). Sulfur and selenium were analyzed in the first test runs and values were always below the detection limit which ranged between 500 to 1500 ppm and 30-100 ppm respectively.

CHAPTER IV

Figure 4.6 – Average concentrations and detection limits of U, As, Ba, Cs and Cu determined by LA-SF- ICPMS microanalysis, following same data and legend structure as explained in caption of Figure 4.5.

CHAPTER IV

Figure 4.7 – Average concentrations and detection limits of Ni, Ce, Pb, Sb and Bi determined by LA-SF- ICPMS microanalysis, following same data and legend structure as explained in caption of Figure 4.5.

CHAPTER IV

Au and PGE concentrations in fluid inclusions

The concentration of gold in fluid inclusions was below detection limits within most analyzed FIA’s, as shown for all individual inclusion measurements in Figure 4.7. The detection limit was quite variable, ranging from 10 ppb to 10 ppm. Lowest detection values within one assemblage varied between 400 and 40 ppb. Unambiguous gold signals were detected in a few fluid inclusion assemblages of the oxidized hematitic alteration, including 2 to 3 individual inclusions in each assemblage indicating unambiguous detection at concentration levels of 40 to 400 ppb on average (Fig. 4.8). Measurement of Pt and Pd in individual fluid inclusions was abandoned after all initial tests yielded limits of detection above several ppm’s each, but a renewed attempt with a dedicated small element menu might be pursued as a future extension of this project.

Figure 4.8 – Au concentrations and detection limits of individual fluid inclusions microanalysis, following same data and legend structure as explained in caption of Figure 4.5. Colored dashes are detection limit values and red circles are measured values of individual inclusions with the counting-statistical uncertainty indicated by thin gray error bars.

4.4.3 – Compositional trends within FIA’s

The hydrothermal alteration at Serra Pelada is characterized by contrasting reducing and oxidizing ore assemblages and the composition of the fluid inclusion assemblages from the different alteration zones from the deposit suggest similar redox gradients.

Figure 4.8 shows selected element variation diagrams in which such compositional trends are observable. The sodium vs. potassium variation diagram (Fig. 4.9 A) points out that the more reducing chlorite/carbon alteration has consistently lower K concentration in comparison FIA’s hosted by the hematite alteration. The samples from the siliceous alteration have concentrations

CHAPTER IV plotting between the two distal alteration types, as previously indicated by the Na/K ratios plotted in Figure 4.5B. This could be explained as a mixing trend between a high- Na/K (chlorite/carbon alteration) and a low Na/K fluid end-member (hematite alteration). Diagrams with redox-sensitive elements such as uranium and cerium show similar mixing trends, with the lowest concentrations of U and Ce (mostly below detection) in the high Na/K fluids associated with chlorite – carbon alteration and the highest concentrations up to 1.8 ppm U (Fig. 4.9 B) and 10 ppm Ce (Fig. 4.9 C) in the hematite alteration. Concentrations from FIA’s within the siliceous alteration show intermediate values between the other two distal alteration types.

Figure 4.9 - Variation diagrams and mixing trends within FIA from different alteration zones from the Serra Pelada Au - Pd - Pt deposit. (A) Sodium in function of potassium. (B) Sodium to potassium ration in function of Ce. (C) Sodium to potassium ration in function of U.

CHAPTER IV

Table 4.3 – Average elemental composition of fluid inclusion assemblages (FIA’s) from LA-SF-ICPMS microanalysis. Values shown as

NaCl eqv. Th Na Mg K Ni Cu As Sb Cs Ba Ce Au Pb Bi U Sample Method Fia Id Type wt % °C ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm S.196 Fia-1 LA-SF-MS Ps 8.8 165 34440 1780 1420 50 159 39 <7.8 5.4 28 0.9 <1.1 13 < 0.4 < 0.09 1σ standard deviation 0.6 21 1720 420 1590 10 128 13 - 2.5 5 0.3 - 11 - - S.196 Fia-2 LA-SF-MS Ps 2.6 164 10250 160 530 < 17 < 15 140 <2.5 <0.9 4 <1.6 <0.85 <0.6 < 0.12 < 0.06 1σ standard deviation 0.3 14 1130 140 610 - - 30 ------S.196 Fia-3 LA-SF-MS Ps 8.7 160 33180 130 <375 < 12 145 60 3.2 4.0 53 <0.2 <0.58 36 < 0.16 < 0.03 1σ standard deviation 0.5 17 2540 35 - - 5 6 0.8 0.3 40 - - 13 - - S.196 Fia-5 LA-SF-MS Ps 11.8 133 47080 5750 940 < 100 1020 380 < 17 < 6.0 194 <2.3 <5.67 < 6 < 1.58 < 0.40 1σ standard deviation 0.3 26 1067 5330 200 - 1200 50 ------S.196 Fia-6 LA-SF-MS Ps 9.8 152 38570 550 712 < 16 160 < 14 < 2.3 6.4 14 <0.3 <0.08 10 < 0.25 < 0.07 1σ standard deviation 0.3 11 1200 240 390 - 120 - - 3.8 5 - - 7 - - S.064 Fia-1 LA-ICPMS Un 3.5 120 13920 n.a. 460 n.a. < 4 < 2 < 1.8 0.3 8 <0.3 <0.33 < 0.3 < 0.3 < 0.15 1σ standard deviation 1.0 17 3960 - 60 - - - - - 3 - - - - - S.064 Fia-2 LA-ICPMS Un 4.0 166 15500 n.a. 2920 n.a. < 15 < 7 < 3.4 < 1.0 85 <0.8 <1.06 < 2 < 1.36 < 0.75 1σ standard deviation 0.7 12 2960 - 2640 - - - - - 32 - - - - - S.064 Fia-3 LA-ICPMS Un 8.8 151 33340 n.a. 7375 n.a. 24 15 < 5.0 < 1.0 45 <2 <1.60 2.1 < 1.3 < 0.74 1σ standard deviation 2.0 19 8400 - 5600 - 10 10 - - 25 - - 0.2 - - S0.64-2 Fia-2 LA-SF-MS Sec 4.0 145 14410 130 1220 < 19 4 13 < 0.5 0.9 60 0.4 <0.22 1.6 < 0.07 < 0.01 1σ standard deviation 0.3 11 1390 90 1360 - 1 1 - 0.5 27 0.2 - 0.2 - - S.136 Fia-3 LA-SF-MS Ps 4.4 n.a. 16930 <85 1456 < 8 < 5 26 2.1 1.4 23 <0.1 <0.17 2.3 < 0.14 < 0.01 1σ standard deviation 0.3 - 1350 - 330 - - 3 1.3 0.3 3 - - 0.2 - - S136-2 Fia-1 LA-SF-MS Ps 3.8 161 14230 490 516 < 2 < 2 6 < 0.1 0.7 43 <0.1 <0.01 20 0.31 0.24 1σ standard deviation 0.5 7 1700 630 550 - - 3 - 0.3 8 - - 20 0.10 0.02 S136-2 Fia-2 LA-SF-MS Pr 3.4 159 13190 190 1550 99 2 8 0.8 0.4 14 <0.01 <0.01 5 0.04 0.05 1σ standard deviation 0.5 7 1870 125 1530 20 0.2 5 0.8 0.3 8 - - 4 0.01 0.004 S136-2 Fia-3 LA-SF-MS Pr 4.0 158 14790 1565 860 < 6 < 3 14 < 0.8 0.7 18 0.5 <0.12 9 0.35 0.20 1σ standard deviation 0.3 8 1990 2300 700 - - 7 - 0.5 9.0 0.5 - 7 0.06 0.13 S136-2 Fia-4 LA-SF-MS Un 3.3 n.a. 12560 120 1040 7 6 10 1.4 0.5 18 <0.2 <0.22 8 < 0.05 0.15 1σ standard deviation 0.2 - 850 70 1360 3 2 2 0.4 0.2 17 - - 6 - 0.140 S136-2 Fia-5 LA-SF-MS Un 5.0 n.a. 19253 670 1450 < 3 < 3 < 4 0.7 < 0.2 12 <0.1 <0.15 20 < 0.28 0.23 1σ standard deviation 0.1 - 280 70 460 - - - 0.2 - 9 - - 3 - 0.05

78 CHAPTER IV

Table 4.3 – Continued.

NaCl eqv. Th Na Mg K Ni Cu As Sb Cs Ba Ce Au Pb Bi U Sample Method Fia Id Type wt % °C ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm S.166 Fia-1 LA-ICPMS Un 5.7 n.a. 20820 303 4995 < 47 < 36 < 6 < 4.2 5.1 348 <1.2 <1.26 < 7 < 1.28 < 0.55 1σ standard deviation 0.3 - 1320 60 1400 - - - - 1.3 217 - - - - - S.166 Fia-2 LA-ICPMS Un 5.9 n.a. 21680 369 4504 < 36 < 13 < 4 < 2.8 5.9 724 <0.6 <0.96 < 0.9 < 0.63 < 0.37 1σ standard deviation 0.3 - 770 20 2050 - - - - 0.6 720 - - - - - S.140 Fia-1 LA-SF-MS Ps 6.6 n.a. 25920 410 3460 158 362 6.2 < 0.1 0.5 10 2.1 <0.11 131 0.37 0.41 1σ standard deviation 0.6 - 1860 260 3400 109 408 1.2 - 0.6 7 0.5 - 7 0.13 0.22 S.140 Fia-2 LA-SF-MS Pr 5.8 n.a. 22460 1970 11080 < 92 336 112 < 11 8.7 89 3.8 <1.8 93 < 0.96 < 0.52 1σ standard deviation 0.2 - 1060 870 3160 - 160 24 - 3.1 18 2.2 - 44 - - S.140 Fia-6 LA-SF-MS Pr 5.3 n.a. 20960 520 6980 133 169 41 2.5 2.1 41 1.5 <0.38 75 0.74 0.28 1σ standard deviation 0.2 - 760 220 3570 22 108 7 0.7 0.1 18 0.8 - 70 0.11 0.1 S.145 Fia-1 LA-SF-MS Ps 6.0 177 23390 matrix 12100 52 5 48 n.a. 3.0 13 12 < 0.07 matrix < 0.03 0.11 1σ standard deviation 0.1 7 370 - 2950 35 1 21 - 0.9 5 8 - - - 0.13 S.145 Fia-3 combined Ps 6.3 185 24820 matrix 13550 < 5 7 70 < 0.3 3.5 18 < 0.02 < 0.02 matrix < 0.07 0.10 1σ standard deviation 0.3 2 740 - 520 - 6 30 - 0.6 15 - - - - 0.05 S.145 Fia-3.1 LA-SF-MS Ps 6.1 178 23920 matrix 7200 < 3 5 75 3.7 2.7 17 < 0.06 0.09 matrix < 0.11 < 0.004 1σ standard deviation 0.2 5 800 - 2880 - 3 60 0.4 1.1 13 - 0.06 - - - S.145 Fia-4 LA-SF-MS Ps 5.1 169 17160 matrix 7840 35 < 1 27 1.2 2.4 162 3.7 < 0.02 matrix < 0.01 0.14 1σ standard deviation 0.1 10 640 - 1400 10 - 9 0.3 0.7 92 3.4 - - - 0.1 S.145 Fia-4.1 LA-SF-MS Ps 6.5 n.a. 25630 matrix 15700 < 5 9 71 n.a. 3.7 15 17 < 0.03 matrix < 0.05 0.06 1σ standard deviation 0.1 - 280 - 1950 - 8 19 - 0.6 10 3 - - - 0.06 S.145-2 Fia-1A LA-SF-MS Pr 3.3 n.a. 14310 matrix 1930 < 9 < 38 < 13 < 1.2 1.8 6 < 0.2 < 0.23 18 < 5.72 < 0.04 1σ standard deviation 0.6 - 2670 2030 - - - - 0.3 2.0 - - 3 - - S.145-2 Fia-2 LA-SF-MS Pr 11.0 n.a. 43990 matrix 185 < 30 50 < 17 < 2.3 < 0.8 93 < 0.3 < 0.45 < 1.2 < 0.26 < 6.7 1σ standard deviation 0.2 - 672 - 35 - 10 - - - 71 - - - - - S.145-2 Fia-0.1B LA-SF-MS Ps 5.8 n.a. 22860 matrix 12370 matrix 63 60 < 0.3 3 36 < 0.03 < 0.03 matrix < 0.02 < 0.02 1σ standard deviation 0.2 - 520 - 2670 - 68 21 - 1 56 - - - - - S.145-2 Fia-0.1-2B LA-SF-MS Ps 6.2 n.a. 24170 matrix 15880 < 3 12 67 1.8 3.7 78 3.2 < 0.04 45 < 0.03 < 0.01 1σ standard deviation 0.1 - 200 - 4790 - 7 29 0.5 3.4 134 5.3 - 11 - - S.145-2 Fia-0.1-3B LA-SF-MS Ps 4.5 n.a. 17530 matrix 7400 < 3 4 9 < 0.3 2.4 19 < 0.02 < 0.03 8 < 0.01 0.34 1σ standard deviation 0.1 - 500 - 310 - 2 6 - 0.3 16 - - 10 - 0.3 S.145-2 Fia-1B LA-SF-MS Pr 3.9 n.a. 15010 matrix 3370 < 5 < 3 51 < 0.5 1.8 26 2.2 < 0.06 < 0.3 0.82 < 0.2 1σ standard deviation 0.1 - 260 - 1570 - - 48 - 0.6 19 1.0 - - 0.24 - S.145-2 Fia-1C LA-SF-MS Pr 4.3 n.a. 16760 matrix 3160 < 14 < 22 < 12 < 1.8 1.0 <8.2 < 1.7 < 0.18 < 6.1 0.74 < 0.1 1σ standard deviation 0.1 - 360 - 4000 - - - - 0.3 - - - - 0.19 -

79 CHAPTER IV

Table 4.3 – Continued.

NaCl eqv. Th Na Mg K Ni Cu As Sb Cs Ba Ce Au Pb Bi U Sample Method Fia Id Type wt % °C ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm S.145-2 Fia-1-2C LA-SF-MS Pr 3.8 n.a. 14370 matrix 2450 < 6 < 5 < 4 < 0.75 1.2 <0.9 <0.1 < 0.07 < 0.2 < 0.06 < 0.03 1σ standard deviation 0.4 - 1010 - 1460 - - - - 0.3 ------S.145-2 Fia-3E LA-SF-MS Pr 20.7 n.a. 81740 matrix 2500 56 6 5 < 0.3 2.5 21 0.7 < 0.03 matrix < 0.04 < 0.1 1σ standard deviation 0.1 - 505 - 280 14 4 2 - 0.6 18 0.1 - - - - S.145-2 Fia-0.1D LA-SF-MS Ps 6.0 n.a. 23560 matrix 9130 < 1 < 1 48 < 0.2 2.4 18 < 0.1 < 0.08 matrix 0.14 0.08 1σ standard deviation 0.2 - 850 - 3660 - - 29 - 0.6 14 - - - 0.06 0.01 S.145-2 Fia-0.1-2D LA-SF-MS Pr 7.3 n.a. 28740 matrix 16400 < 15 <10.2 66 < 2.2 4.0 19 2.8 < 0.77 matrix 0.74 < 0.03 1σ standard deviation 0.2 - 1480 - 1910 - - 16 - 0.6 3 1.0 - - 0.28 - S.147 Fia-1 LA-ICPMS Ps 4.9 168 19700 n.a. 1920 294 734 7 < 1.4 2.7 37 4.3 < 0.24 40 < 0.09 < 1.025 1σ standard deviation 0.3 3 1080 - 450 195 696 1 - 0.2 1 2.7 - 43 - - S.147 Fia-2 LA-ICPMS Ps 4.6 166 18300 n.a. 2070 72 341 15 < 0.4 1.4 5 8 < 0.10 27 0.21 < 3 1σ standard deviation 0.3 12 1010 - 1370 94 467 8 - 0.8 1 8 - 24 0.70 - S.147 Fia-3 LA-ICPMS Ps 5.4 153 21280 n.a. 2280 33 87 24 < 0.7 1.4 9 1.6 < 0.27 6 < 0.17 < 0.08 1σ standard deviation 0.4 3 1820 - 140 27 110 15 - 0.3 8 0.2 - 2 - - S.147 Fia-4 LA-ICPMS Un 16.6 150 65440 n.a. 4450 531 4400 11 < 6.5 11.0 40 28 < 2.06 321 < 1.03 < 0.71 1σ standard deviation 1.1 12 170 - 2870 350 2400 1 - 1.0 16 12 - 203 - - S.141 Fia-1 LA-ICPMS Sec 4.0 174 15580 210 3770 < 37 32 32 3.0 5.0 27 < 1.2 < 1.22 9 < 0.60 < 165 1σ standard deviation 0.4 8 1650 190 2180 - 10 13 0.8 3.5 22 - - 4 - - S.141 Fia-2 LA-ICPMS Sec 4.6 165 18000 n.a. 2240 < 74 586 21 < 5.8 5.8 < 15 < 2.5 < 5.41 <5.6 < 2 < 3 1σ standard deviation 0.4 9 1050 - 800 - 128 4 - 4.6 ------S.141 Fia-5 LA-ICPMS Sec 4.3 162 16900 n.a. 1960 < 4 163 218 <1.5 10.6 9 8.9 < 1.42 1 < 0.66 < 0.67 1σ standard deviation 0.2 11 900 - 1990 - 66 74 - 3.3 2 4.6 - 0 - - S.141 Fia-6 LA-ICPMS Sec 20.0 160 83010 n.a. 17560 < 93 215 < 67 < 20 <4.9 241 < 6.6 < 14 26 < 10 < 7 1σ standard deviation 2.3 10 5650 - 5720 - 108 - - - 260 - - 16 - - S.142 Fia-1 LA-ICPMS Un 4.5 174 17690 matrix 8550 < 49 < 13 64 < 3.2 3.2 134 6.9 < 1.8 < 2.1 < 1.1 <0.60 1σ standard deviation 0.1 11 470 - 4850 - - 20 - 1.0 142 3.8 - - - - S.142 Fia-2 LA-ICPMS Un 23.1 162 87700 matrix 15070 < 45 < 12 < 5 < 2.3 2.2 59 < 1.2 < 1.4 39 < 0.72 < 0.22 1σ standard deviation 0.6 7 250 - 7900 - - - - 0.5 56 - - 15 - - S.142 Fia-3 LA-ICPMS Un 4.4 177 17420 matrix 9240 < 21 < 3 35 < 1.3 4.0 129 22.0 < 0.58 23 < 0.32 < 0.12 1σ standard deviation 0.2 6 810 - 2420 - - 18 - 3.3 80 12.8 - 4 - -

80 CHAPTER IV

Table 4.3 – Continued.

NaCl eqv. Th Na Mg K Ni Cu As Sb Cs Ba Ce Au Pb Bi U Sample Method Fia Id Type wt % °C ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm S.142 Fia-4 LA-ICPMS Un 23.5 162 92595 6180 14985 < 10 21 141 2.2 3.8 27 < 0.20 < 0.24 37 0.66 < 0.26 1σ standard deviation 0.3 7 1000 1830 5073 - 16 136 0.4 2.1 18 - - 9 0.20 - S.142 Fia-5 LA-ICPMS Un 21.7 155 86153 5465 10252 267 < 20 < 4 35.2 4.8 23 3.1 < 0.81 57 < 0.55 < 0.14 1σ standard deviation 1.5 23 4710 3060 3226 44 -- - 23.6 3.1 6 1.4 - 49 - - S.147-2 Fia-1 LA-SF-MS Ps 4.2 185 16690 matrix 3057 < 40 < 3 32 < 0.9 2.6 <10.7 < 0.5 < 0.15 matrix < 0.17 0.43 1σ standard deviation 0.2 5 900 - 2220 - - 13 - 1.2 - - - - - 0.16 S.147-2 Fia-2 LA-SF-MS Un 14.3 175 56275 3860 3220 < 77 < 20 < 9 < 2.1 <0.6 36 < 0.2 < 0.24 11 < 0.43 < 0.08 1σ standard deviation 0.1 6 300 1720 440 - - - - - 6 - - 2 - - S.147-2 Fia-3 LA-SF-MS Un 17.8 n.a. 70860 5590 3210 < 46 61 < 11 < 1.7 <0.4 30 < 0.05 < 0.17 40 < 0.3 < 0.04 1σ standard deviation 0.3 - 1410 3240 750 - 14 - - - 14 - - 30 - - S.147-2 Fia-4 LA-SF-MS Un 21.3 149 84090 4020 7220 < 60 < 9 < 10 < 2.0 7.4 129 <0.2 < 0.24 100 < 0.35 0.47 1 22 420 1300 300 - - - - 1.9 33 - - 19 - 0.5 σ standard deviation 0.4 S.147-2 Fia-5 LA-SF-MS Ps 4.0 169 17480 960 730 < 9 12 < 1.1 < 0.3 0.9 16 < 0.2 < 0.04 10 < 0.05 < 0.02 1σ standard deviation 0.5 5 1070 820 730 - 7 - - 0.4 18 - - 6 - - S.147-2 Fia-7 LA-SF-MS Un 15.7 166 61410 5190 7450 < 5 53 57 < 0.7 8.0 189 < 0.09 < 0.13 70 < 0.25 0.44 1σ standard deviation 0.7 16 1690 2060 1160 - 25 79 - 2.8 67 - - 22 - 0.19 S.147-2 Fia-8 LA-SF-MS Un 18.4 131 71978 12750 10420 < 22 < 28 < 20 < 3.4 6.9 <0.9 < 0.4 < 0.91 51 < 0.58 0.5 1σ standard deviation 0.2 11 1000 4100 1750 - - - - 0.7 - - - 16 - 0.13 S.147-2 Fia-9 LA-SF-MS Un 23.5 139 91230 6730 2840 <6 < 9 < 5 < 1.3 1.5 8 < 0.2 < 0.31 7 < 0.26 < 0.03 1σ standard deviation 0.4 7 3000 5100 680 - - - - 1.2 3 - - 5 - - S.147-2 Fia-13 LA-SF-MS Ps 5.0 171 19500 1300 2800 15 50 89 6.2 3.4 6 3.6 0.32 28 < 0.27 < 0.12 1σ standard deviation 0.2 9 1040 800 105 4 63 41 3.7 1.7 2 1.8 0.18 35 - - S.147-2 Fia-15 LA-SF-MS Ps 5.1 163 20320 2400 765 10 30 11 3.6 2.4 22 1.3 0.16 11 0.55 < 2.71 1σ standard deviation 0.2 12 800 2000 600 8 30 5 2.3 1.9 6 0.9 0.13 8 0.45 - S.147-2 Fia-16 LA-SF-MS Ps 16.7 141 65910 14670 11299 9 106 27 18.2 19.8 344 < 0.1 < 0.4 228 < 4 < 0.01 1σ standard deviation 0.2 3 610 10640 3030 2 91 5 25.0 11.9 165 - - 139 - - S.147-2 Fia-18 LA-SF-MS Ps 5.0 n.a. 19810 4690 11320 27 45 64 <0.5 4.1 11 0.7 < 0.05 26 0.36 0.35 1σ standard deviation 0.1 - 320 5370 19870 4 57 18 - 2.2 6 0.4 - 25 0.32 0.37 S.147-2 Fia-19 LA-SF-MS Ps 4.3 n.a. 16780 920 3430 36 41 29 4.1 1.8 19 0.9 0.36 8 0.88 1.81 1σ standard deviation 0.4 - 1530 420 1760 19 15 15 3.2 1.3 8 1.0 0.01 5 1.16 1.31

81 CHAPTER IV

4.5 – THERMODYNAMIC CONSTRAINTS FROM FLUID INCLUSION LA-SF-ICPMS ANALYSIS AND MINERAL EQUILIBRIA

4.5.1 – Pressure and temperature

The pressure and temperature conditions of Au - Pd - Pt mineralization at Serra Pelada can be estimated by combining microthermometry data and independent mineral geothermometers. Fluid inclusion isochores from microthermometry data measured in this study were calculated using the SOWAT software (Driesner and Heinrich, 2007). The conditions of formation of Serra Pelada deposit were estimated at about 270 ± 20 ºC and 1.5- 2 kbar (Fig. 4.10), by intersecting the isochores of fluid inclusions with 7-20 wt. % NaCleqv. homogenizing at about

165 °C with the temperature range defined by the disordered form of Ag3AuSe2 (> 260 ºC – Wiegers, 1976; Echmaeva and Osadchii, 2009). The stability limit of kaolinite relative to pyrophyllite constrains the temperature to be below 290°C at a pH2O of 2 kbar (Holdaway, 1971), consistent with the ore-mineral stabilities.

Figure 4.10 – Diagram illustrating pressure and temperature conditions for the formation of the Serra Pelada deposit, based on calculated fluid inclusion isochores, ore mineral geothermometers and the stability field of kaolinite and pyrophyllite according to Holdaway (1971).

CHAPTER IV

4.5.2 – pH

Kaolinite is a mineral which is stable within the primary mineral assemblage of the ore zone (Berni et al., 2014) and was also identified as accidentally trapped minerals within some fluid inclusion assemblages (Fig. 4.3E). This is an evidence that those fluids were on the stability field of kaolinite and can be used as an indicator of minimum acidity relative to the stability of muscovite according to the reaction:

K+ + 1.5 Al2Si2O5(OH)4 = K(Al2Si3O10)(OH)2 + 1.5 H20 + H+ (Kaolinite) (Muscovite)

Calculating log K for the reaction above using the Geochemist Workbench 9.0 Rxn software and the thermo.dat thermodynamic database therein (Bethke, 2008), we can calculate the pH at 270°C for an activity of K+ constrained by the measured potassium concentrations of fluid inclusions, which ranges from 1000 to 10’000 ppm within all alteration zones. Equating molal concentration with activity would result in K+ activities between 0.025 and 0.25, indicating a maximum pH of ca. 2.5 and 3.5 for a fluid in equilibrium with kaolinite (Fig. 4.11). Activity coefficients in the H+/K+ cation ratio partly cancel, but ion association to form KCl in high- temperature brines will reduce free potassium ion activity and permit somewhat higher pH values.

Figure 4.11 – pH vs. log [K+] diagram showing the pH variations assuming concentrations from natural fluid inclusions. The red line represents muscovite to kaolinite breakdown reaction. Dashed lines indicate potassium activities of 0.25 and 0.025 and the respective pH, [H2O]=1.

CHAPTER IV

4.5.3 – pH / ƒO2 (g) relationships from mineral assemblages and fluid inclusion data

The ore mineralogy at Serra Pelada comprises oxidizing and reducing mineral assemblages, suggesting that redox gradients were important in the genesis of the deposit (Chapter III of this thesis). The presence of copper, platinum selenides and Bi-V oxides suggest that the oxygen fugacity reached log values of ca. -20 at 300°C, also at high selenium activities (see Fig. 3.9). These highly oxidizing conditions can be cross-checked by calculating the stability fields of other ore minerals such as sudovikovite (PtSe2) and sperrylite (PtAs2), by considering activities of dissolved species based on fluid inclusion analyses.

Figure 4.12 is an activity diagram of log ƒO2(g) versus pH, showing the stability fields of some of the minerals occurring at Serra Pelada and the predominance fields of aqueous species for selenium (Fig. 4.12 A) and arsenic (Fig. 4.12 B) based on data from Mountain and Wood (1988). The temperature of 300 ºC was chosen since there is direct thermodynamic data available at this temperature, which is close enough to estimated temperature for the formation of the deposit (270 ºC). The published data and diagrams are for aqueous fluids at vapor saturation pressure, an d we ignore any shifts of equilibria caused by the higher pressure during ore fomation, for lack of volumetric data. The predominance fields of sulfur species, the stability fields of Fe minerals and graphite (Ohmoto, 1972) are shown for reference. Solubility contours of Pt as chloride complexes are shown, based on data of Mountain and Wood (1988) and estimated concentrations applicable to Serra Pelada. Solubility contours of gold as chloride complexes were calculated with Geochemists Workbench for 0.1 and 1 ppm. All calculations and estimations involved in the construction of the stability fields in Figure 4.12 are listed in Appendix I of this thesis.

Even though there is no information available on selenium concentration from fluid inclusion analyses, the detection limits for Se within the FIA’s measured in this study were around 30 ppm and thus set an upper limit for Se concentrations. The effect of selenium concentration on the stability field sudovikovite in respect to native platinum can be estimated by calculating how much it will expand with increasing selenium concentrations. Figure 4.12 A shows that saturation of selenide minerals will occur at highly oxidizing conditions, even at low Se concentrations. The boundary-line between the stability field of native platinum, luberoite and sudovikovite is represented by a thick area corresponding to a shift in oxygen fugacity due to an increase in the activity of dissolved selenium species from 10-6 to 10-5 (corresponding to about

0.45 to 4.5 ppm Se in solution). The activity diagram predicts that luberoite (Pt5Se4 – PtSe0.8 at Fig. 4.12 A) should precipitate as an intermediate phase between native platinum and

CHAPTER IV sudovikovite, but this mineral phase was not identified within the ore mineralogy. Sudovikovite was identified by semi-quantitative electron microscopy analysis, but grains are rare and too small for EMPA analysis, therefore the occurrence of luberoite (Pt5Se4) cannot be excluded. The presence of platinum arsenides can be used as another independent constraint indicating a highly oxidized ore assemblage. Figure 4.12 B depicts the stability field of sperrylite (PtAs2) relative to native platinum and luberoite in respect to pH and log ƒO2(g). The thickness of the lines separating the stability field of platinum minerals indicates the increase in oxygen fugacity on the saturation of sperrylite in respect to native platinum when arsenic activity as H3AsO3 increases from 10-4 to 10-3, which can be estimated from the high (ca. 100 ppm) concentrations measured in fluid inclusions from the Serra Pelada deposit.

The big circles plotted on both diagrams in Figure 4.12 are an attempt to interpret the range of pH and ƒO2 (g) conditions recorded by the deposit. The open circles refer to the conditions of inferred end-member compositions of fluids that converged from different source regions to generate the deposit; the full circles denote estimated conditions constrained by observed mineral assemblages within the deposit. The open circle (Ox) is a highly oxidized fluid, which is inferred to be the main carrier of dissolved ore metals. This end-member fluid is most acid, well in the kaolinite field at the minimum observed [Na+]/[K+], because it must be able to cause alteration of sericite to kaolinite in the ore zone. It is tentatively placed at a high ƒO2 (g) level near the aqueous As(V)/As(III) redox transition. H2As(V)O4– is the potentially dominant oxidant available above the sulfate/sulfide buffer (Heinrich and Eadington, 1986), based on the high As concentrations analyzed in the inclusions (~ 10–3m), which are the same order of magnitude than the free O2 concentration of air-saturated water (~ 10–5m, Millero et al., 2002). The reduced end member fluid (R) represents a fluid containing CH4 derived from organic carbon-rich lithologies, which must be able to precipitate hydrothermal carbon in the ore zone. Being equilibrated with the reduced metapelitic country rocks, it probably has a weakly acid pH consistent with muscovite stability and the highest Na/K ratios among the analyzed fluids (green points in Fig. 4.5 B).

The filled circles denote the conditions constrained by ore paragenesis (Chapter III of this thesis), which are all in the pH field of kaolinite alteration. Circle (A) denotes the pyrite and sulfide-bearing assemblage in the carbonaceous ore, but also locally includes hematite and is therefore placed near the sulfide/sulfate buffer. Circle (B) is the more oxidized, selenide rich assemblage which tends to overprint the sulfide-bearing assemblage and can only be stable in the hematite field with HSO4– >> H2S. It is placed at a fugacity range around to 25 log units,

CHAPTER IV where electrum would precipitate from the assumed gold concentrations in the fluid (ca. 100 ppb). The presence of Bi-V oxides and fischesserite indicates even higher values for oxygen fugacity (ca. 20 log units – Chapter III of this thesis). The closed circle in green (C) represents the conditions where graphite would precipitate according to Ohmoto (1972).

Figure 4.12 – pH vs. log ƒO2 diagram showing the stability fields of Pt metal, selenide (A) and arsenide (B) minerals. The pink lines ( ) separates the stability of native platinum metal and respective selenide (A) and arsenide minerals (B) for two different concentrations of Se (from 10-6 to 10-5 m) and As(from 10-4 to 10-3 m), indicated by the black arrows (towards higher concentration) within the pink shaded areas. The stability fields of Se and As species are shown in gray lines ( ) in figures A and B at concentrations of 10-6 and 10-4m respectively. The stability field of S aqueous species at a total S concentration of 0.005 m are shown in black dashed lines ( ). Solubility contours of Pt as PtCl42- and Pt(OH)4-2 complexes at a concentration of 10ppb Pt is shown in dot-dashed lines ( ). Solubility contours of gold as AuCl2- in the absence of Se and S are shown in orange lines ( ) for 1 and 0.1 ppm (Cl- activity of 1 m), which are only partially shown for clarity. Blue lines ( ) separate the stability fields of Fe minerals. Graphite stability field is separated by green lines ( ). Big circles represent the interpreted conditions of fluids and precipitated mineral assemblages, see text for explanation. Modified from Mountain and Wood (1988) with data from Geochemists Workbench (Bethke, 2008).

4.6 - GENETIC MODEL

The Serra Pelada Au - Pd - Pt mineralization is epigenetic and of hydrothermal origin, controlled by fluid-focusing step structures intersecting the fold hinge of a recumbent syncline (Berni et al., 2014). The identification of primary ore minerals (Chapter III of this thesis) and the fluid inclusion constraints presented here suggests that a redox process, driven by fluid mixing of a reducing (CH4-bearing) fluid with a highly oxidizing, acidic and metal carrying fluid, can explain the highly efficient metal precipitation. The presence of hydrothermally precipitated carbon and abundant hematite alteration which mutually cut and overprint each other (Fig. 2.8 G) is the 86

CHAPTER IV main evidence that such extreme redox gradients occurred at overlapping times during hydrothermal ore formation.

Under reducing conditions, carbon will be transported mainly as CH4 and will precipitate upon partial oxidation around 6 ƒO2 (g) units below the magnetite/hematite buffer (Ohmoto, 1972). Iron will also be transported mainly as Fe2+ (Heinrich and Seward, 1990) and will precipitate by reaction with oxidizing fluids. This could explain the localized net addition of hematite in vugs, veins (e.g. Fig. 4.2) and the red hematite-rich and the white kaolinite-rich argillic alteration (more oxidizing/acidic fluids) in close proximity with highly carbon-enriched argillic rocks hosting most of the highest-grade ore (Figure 2.6). The reducing fluids able to transport carbon can be generated by fluid interaction with weakly carbonaceous metasedimentary rocks, which are widespread in the Serra Pelada region. The limb-parallel pattern of the distal chlorite-carbon alteration (Fig. 4.1) suggests that fluid migration was preferably through S0/S1 structures (bedding-parallel foliation along the fold limbs).

By contrast, Au, Pt, Pd, arsenide or arsenate and selenide or selenate are transported together at highly oxidizing conditions and can be precipitated by initial reduction from conditions approximated by the point (Ox) to those represented by the assemblages A and B in figure 4.12.

The high log ƒO2(g) values estimated for the saturation of PtAs2 and PtSe2 in respect to native platinum implies that exceptionally oxidized fluids reached relatively deep portions of the crust

(ca. 5 km). Fluids buffered at such oxidized conditions, more than 10 log ƒO2 (g) units above the sulfate/sulfide buffer, can be generated by initial interaction of fluids with an oxygenated atmosphere, but elemental O2 is not very soluble. Dissolved O2 was present at ppm-levels in water saturated with the Proterozoic atmosphere, which probably contained a few mol % of O2

(g) (Kump, 2008), and even a small extent of reaction with reducing rocks (e.g. ~10–6 moles ferrous iron mineral per kg of fluid) will consume this free O2. Arsenic is present at higher concentrations than the maximum level of free dissolved O2, based on fluid inclusions from the Serra Pelada deposit, and is normally transported as As(V)-OH complexes at oxidizing conditions (Heinrich and Eadington, 1986). Arsenic (V) species therefore might partly buffer the oxidation state of the fluid at very high level, on its way to the ore deposit (Fig. 4.12 B).

Oxidized sulfur (SO4-2/HSO4-) may have been present in higher concentration than arsenate, but the slightest reduction to H2S (unless kinetically inhibited) will bring down the ƒO2 (g) towards the HSO4-/H2S line (Fig. 4.12). Even with 10–3 m arsenate as a weak but high-ƒO2(g) fluid buffer, the flow path of the oxidized precious-metal carrying fluid must have been dominated by thoroughly oxidized (effectively Fe2+-free) mineral assemblages, to reach the deposit for

CHAPTER IV effective precipitation of the high-grade ore. Field evidence of such flow paths is widespread even in mafic rocks of the Serra Pelada region, in the form of hematitic-altered crenulation cleavages and completely hematitized fault zones (Chapter II).

The presence of hydrothermal kaolinite and the high K contents within fluid inclusions from the hematite alteration, suggests that such oxidizing fluids were also quite acidic (kaolinite stability field, pH ca. 2.5-3.5). This is also supported by the siliceous replacement of the dolomitic sandstones around the fold-hinge and by dissolution vugs within the hematite altered sandy carbonates. Such acidic fluids cannot be readily explained by mineral buffering within the sedimentary rocks of the Serra Pelada region, because they contain widespread muscovite and because fluids buffered by muscovite + kaolinite should become less acid during heating along their down flow path (Montoya and Hemley, 1975). The presence of Mg-rich aluminum silicates at Serra Pelada and the relatively high concentration of magnesium at fluid inclusions (> 10.000 ppm) confirms that at least locally these fluids were also Mg-rich. The reaction of sericite with magnesium-rich fluids would produce amesite or Mg-Chlorite and could thereby generate acidity according to the reaction:

6H2O(l) + 1.33 KAl2(AlSi3O10)(OH)2 + 4Mg2+(aq) = Mg2Al((SiAl)O5(OH)4 + 2 SiO2+ 6.67H+(aq)+ 1.33 K+(aq) Muscovite Amesite Quartz A magmatic contribution to acidity, the predominant source of acidity by disproportionation of

SO2 to H2S is inconsistent with the high oxidation potential of the fluids, even though distal granitoid intrusions (Cigano Granite) overlap in age with the Serra Pelada mineralization and were previously suggested to be related to the ore formation (Grainger, 2003).

Fluids within this salinity range and at such highly oxidizing and acidic conditions would be able to transport about 1 ppm of Pt (Mountain and Wood, 1988) and even higher concentrations of gold (Wilde et al., 1989). Most of the measured gold concentration values in fluid inclusion are at about 100 ppb, suggesting that the fluids were metal under saturated. This implies that source rock availability was a major factor controlling the metal acquisition by the mineralizing fluid.

Gold and PGE were probably sourced from the hydrothermal leaching of mafic to ultramafic rocks equivalent to the volcanic and the Luanga mafic/ultramafic intrusive complex (Tallarico et al., 2000; Grainger et al., 2002), which originally overlaid the area of the deposit due to regional thrusting. The presence of Ni within the fluid inclusions analyzed in this study (Fig. 4.2) corroborates the interpretation that the highly-oxidized hematite – quartz – carbonate breccias represent pathways for fluids that had previously interacted with oxidized mafic/ultramafic source rocks. 88

CHAPTER IV

A schematic interpretation of possible fluid flow paths and processes that generated the metal concentration at the Serra Pelada deposit is presented in figure 4.13A, which is necessarily speculative due to the lack of detailed structural data for the tectono-stratigraphic package on a regional scale. The re-construction is based on a mapped, major pre-mineralization (S1) fault, which thrust the mafic and ultramafic metavolcanic rocks of the Rio Novo group against and over the Serra Pelada deposit, providing a plausible former source area from where the metals were leached. The mineralization is interpreted to be coeval with the mapped crenulation cleavage, evidenced by the subvertical attitude of the main orebody and local carbon precipitation along such cleavage surfaces (Berni et al., 2014). The mechanics driving fluid flow and ore precipitation is speculated to be a cyclic process by which fluids were repeatedly transferred from a source region for Au, Pt, and Pd in oxidized mafic metavolcanic and related ultramafic intrusive rocks of overlying the Rio Novo Group to the underlying hinge of the folded Serra Pelada metasedimentary rocks. In short periods of tectonic activation, oxidized fluids carrying metal would flow down to the deposit area, driven by the permeability increase (and pore pressure decrease) after brittle-ductile fracturing. In longer periods after a tectonic pulse, pore pressure in the mechanically weak metasedimentary rocks would increase towards lithostatic conditions, allowing time for reaction with reduced fluids derived more locally from the lower parts of the sequence, generating more reducing agent for the next batch of incoming oxidized fluid. This scenario is consistent with the more general model of fluctuating permeability in a critically-stressed crust (Ingebritsen and Manning, 2010), although transfer of surface fluids into the lithostatic domain of the crust remains a point of debate. The ultimate source of the fluids is likely to be an evaporitic surface environment providing adequate fluid salinity for metal complexation and a plausible source of Mg after precipitation of Ca-rich carbonates, to generate fluid acidity upon silicate alteration, as schematically indicated in Figure 4.13B.

The ore formation model for Serra Pelada derived here has several similarities with published interpretations for unconformity-related U ± Au ± PGE deposits, including with the Coronation Hill U ± Au ± PGE deposit. These include extreme redox-gradients evidenced by oxidizing/acidic and reducing mineral assemblages, but at rather lower temperatures (200 ºC). The mineralization in all of these deposits is interpreted to occur after diagenesis and deformation of the host rock, with a strong structural control. Reduction, driven by interaction of reducing and extremely oxidizing surface-derived fluids (Wilde et al., 1989; Mernagh et al., 1994) is suggested as the main mechanism for the ore mineral precipitation.

CHAPTER IV

Figure 4.13 – Schematic model idealizing the mineralizing process at Serra Pelada Au - Pd - Pt deposit. (A) Regional view showing the main lithologies and possible fluid paths. (B) Local inset indicating main alteration processes interpreted to be key features for the formation of the deposit. 90

CHAPTER V

CONCLUSIONS AND OUTLOOK

Serra Pelada is a unique deposit in terms of its exceptional Au - Pd - Pt ore metal grades, but it can be understood as an extreme variant of unconformity-related hydrothermal ore deposits, formed at a deep hydrological interface between an oxic and probably evaporitic Paleoproterozoic surface environment and reduced basement rocks containing carbon-bearing metasedimentary rocks. The main results of this scientific research can be listed as: (1) A qualitative analysis of major element exchange associated with the primary hydrothermal alteration and the recognition of redox gradients from the distal chlorite/carbon and hematite alterations, building on geological observation of large-scale structure and small-scale petrography. (2) Identification of the primary ore mineralogy, which is made by selenides, arsenides, antimony-arsenides, sulfide and sulfate minerals. The presence of AgSe2-Ag3AuSe2 solid solution suggests precipitation as the α-form of gold-silver selenide, which indicates a higher temperature for the formation of the deposit than previously considered. (3) Quantitative determination of major-and trace-element composition of the fluid inclusions associated the main ore zones within the Serra Pelada Au - Pd - Pt deposit, which shows that fluid mixing across an extreme redox gradient is the key process that generated this unusual deposit.

The improved understanding of the primary hydrothermal features of the deposit provided extra geological information for a better comparison between the Serra Pelada Au - Pd - Pt deposit with other selenide-bearing Au and PGE occurrences in the world. Redox gradients, driven by the mixing between highly oxidizing and reducing fluids, is a characteristic feature that the Serra Pelada deposit shares with many other deposit types, including sediment-hosted copper deposits such as the Kupferschiefer of Poland, the unconformity-related ore deposits of Coronation Hill and other U ± Au ± PGE deposits of Australia, and low temperature Se vein deposits of Germany.

Hydrothermal alteration and ore deposition at the Serra Pelada deposit shares the most similar temperature conditions, mineralogical and geochemical features with unconformity-related uranium deposits, even though Serra Pelada seems to have formed at the highest temperature of all selenide-bearing occurrences reported so far. None of the reports of other Se-rich Au - PGE deposits consider a magmatic contribution of heat or fluids as a necessary feature for ore formation, and the data reported in this theses also do not indicate any magmatic connection

CHAPTER V for Serra Pelada despite the approximate (± 45 Ma) temporal link to granite intrusions proposed by Grainger, 2003. Precise TIMS dating of the hydrothermal monazites, which are clearly associated with the mineralization at Serra Pelada, could help to test and possibly disprove the age overlap with the alkaline granitoid event of Carajás region.

Understanding how highly oxidizing surface fluids can be pushed to relatively deep portions of the crust (attaining lithostatic pressures at ~ 5km depth, as required by the thermobarometry reported in this study) remains a major genetic uncertainty. This hydromechanical challenge is common to Serra Pelada as well as unconformity-related uranium deposits. High As concentrations measured in the fluids suggest that the As(V)/As(III) buffer could contribute to keeping the oxidation state of a fluid many log ƒO2 (g) units above the sulfide/sulfate buffer during the fluid downward migration. Such extremely oxidized conditions are essential for Au, Pd and Pt transport as chloride complexes.

The genetic model proposed in this contribution is reasonably well constrained in terms of its chemistry, despite unknowns regarding the mechanism that drove the regional-scale flow of ore-forming hydrothermal fluids. Despite open questions, some regional geological features can be listed as potential exploration features that seem to be key ingredients for the formation of the Serra Pelada Au - Pt - Pd deposit:

(1) Regional presence of mafic to ultramafic rocks (ideally recording some orthomagmatic PGE ± Cr mineralization) is an essential source of metals, which at the time of ore formation should lie above the deposit (in thrust structures that lay above Serra Pelada before its exposure by erosion). The character of source rocks probably exerts a major control on final metal ratios in the deposit (e.g. presence or absence of U; selenide/arsenide/antimonide proportions, etc.). Sedimentary rocks derived from such mafic/ultramafic complexes would be equally important.

(2) Regional evidence of at least locally pervasive hematite alteration, indicating potential fluid channelways for extremely oxidized (sulfide and Fe2+ free) fluids descending from the paleosurface. Sedimentary rocks in the host rock package and any overlying cover sequence that were completely oxidized already at their stage of sedimentation (i.e. red siltstones; primary hematitic meta-sediments free of Fe2+) would be favorable, because they maintain surface- derived brines at their very high ƒO2(g) level during their descent to the site of ore deposition. Mg-alteration with high Mg/Fe chlorite or amesite may also be a regionally favorable indicator and may help generating fluid acidity that assists Au-PGE transport.

CHAPTER V

(3) Reduced host rocks are necessary for Au - Pt - Pd metal ore deposition. (Meta-) sedimentary rocks including at least mildly carbonaceous shales are probably the best source of reducing fluids driving the deposition of high-grade carbonaceous ores.

(4) An understanding of the tectonic deformation history of the terrain is essential, to consider possible scenarios of oxidized fluid pathways through essential source rocks, and to predict possible locations for fluid focusing and fluid mixing as sites of ore deposition (the intersection of recumbently folded reducing metasedimentary rocks with a later steep crenulation cleavage in the case of Serra Pelada).

The criteria mentioned above are somewhat generic and allow considerably different geological and geometric scenarios, but their application to new areas foremost requires a detailed geological knowledge of the prospecting ground. For the Carajás region, consistent regional mapping of primary lithologies including basin analysis, and of regional hydrothermal alteration including geochronological dating, could enhance the number of targets in the region. Better regional stratigraphy and structural geology could allow more reliable predictions of alternative fluid-flow scenarios and possible source regions. Extreme redox-gradients in a fracture- controlled fluid flow system remains the most critical requirement for creating such spectacular ore deposit, but the next Serra Pelada to be found might be hosted by a slightly different lithology and a totally different geometry.

REFERENCES

Alexandre, P., Kyser, K., Polito, P., and Thomas, D., 2005, Alteration mineralogy and stable isotope geochemistry of Paleoproterozoic basement-hosted unconformity-type uranium deposits in the Athabasca basin, Canada: Economic Geology, v. 100, p. 1547-1563.

Almeida, F.F.M., Hasui, Y., Brito-Neves, B.B., and Fuck, R.A., 1981, Brazilian structural provinces: an introduction. Earth Science Reviews, v. 17 p. 1-29.

Avelar, V. G., Lafon, J.M., Correia, Jr. F. C., Macambira, E.M.B., 1999, O magmatismo arqueano da região de Tucumã–província mineral de Carajás: novos dados geocronológicos: Revista Brasileira de Geociências, v. 29, p. 453-460.

Barros, C.E.M., and Barbey, P., 2000, Significance of garnet-bearing metamorphic rocks in the Archean supracrustal series of the Carajás mining province, northern Brazil: Revista Brasileira de Geociências, v. 30, p. 367-370.

Berni, G.V., 2009, Geologia e alteração hidrotermal do depósito de Au - Pd - Pt de Serra Pelada, Curionópolis, Pará, Brasil: Master thesis, Universidade Federal de Minas Gerais, 116p.

Berni, G.V., Heinrich, C.A., Lobato, L.M., Wall, V.J., Rosière, C.A., Freitas, M.A., 2014, The Serra Pelada Au – Pd– Pt Deposit, Carajás, Brazil: Geochemistry, Mineralogy and Zoning of Hydrothermal Alteration: Economic Geology, v.109, p.1883-1899.

Bethke, C.M., 2008, Geochemical and Biogeochemical Reaction Modeling, Cambridge University Press, 543 pp.

Bodnar, R.J., and Vityk, M.O., 1994, Interpretation of microthermometric data for H2O-NaCl fluid inclusions: Fluid inclusions in minerals: Methods and applications, p. 117-130.

Colossus Minerals Inc., 2011, Colossus Minerals drills 7.30 meters at 1494.7 g/t gold, 516.6 g/t platinum and 558.9 g/t palladium in extensions of Central Mineralized Zone at Serra Pelada, Brazil: News release. www.colossusminerals.com.

Cabral, A.R., Lehmann, B., Kwitko, R., and Cravo Costa, C.H., 2002a, The Serra Pelada Au - Pd - Pt deposit, Carajás mineral province, northern brazil: reconnaissance mineralogy and chemistry of very high grade palladian gold mineralization: Economic Geology, v. 97, p. 1127-1138.

Cabral, A.R., Lehmann, B., Kwitko, R., and Cravo Costa, C.H., 2002b, Palladium and Platinum Minerals from the Serra Pelada Au - Pd - Pt Deposit, Carajás Mineral Province, Northern Brazil: Canadian Mineralogist, v. 40, p. 1451-1463.

Cabral, A.R., Burgess, R., and Lehmann, B., 2011 , Late Cretaceous bonanza-style metal enrichment in the Serra Pelada Au - Pd - Pt deposit, Pará, Brazil: Economic Geology, v. 106, p. 119-125.

REFERENCES

Cuney, M., 2010, Evolution of uranium fractionation processes through time: Driving the secular variation of uranium deposit types: Economic Geology, v. 105, p. 553-569.

Chakrabarti, D. J., and Laughlin, D.E., 1981, The Cu-Se (copper-selenium) system: Bulletin of Alloy Phase Diagrams, v.2, p. 305-315.

Distler, V.V., Mitrafanov, G.L., Nemerov, V.K., Kovalenker, V.A., Mokhov, A.V., Semeikina, L.K., and Yudovskaya, M.A., 1996, Modes of occurrence of platinum group elements and their origin in the Sukhoi Log gold deposit (Russia): Geology of Ore Deposits, v. 38, p. 413– 428.

Derome, D., Cathelineau, M., Cuney, M., Fabre, C., Lhomme, T., 2005, Mixing of sodic and calcic brines and uranium deposition at McArthur river, Saskatchewan, Canada: A Raman and laser-induced breakdown spectroscopic study of fluid inclusions: Economic Geology, v. 100, p. 1529-1545.

Driesner T., and Heinrich C. A., 2007, The system H2O–NaCl. Part I. Correlation formulae for phase relations in pressure-temperature-composition space from 0 to 1000 ºC, 1 to 5000 bar, and 0 to 1XNaCl. Geochimica Cosmochimica Acta v. 71, p.4880-4901.

Echmaeva, E.A., and Osadchii, E.G., 2009, Determination of the Thermodynamic Properties of Compounds in the Ag-Au-Se and Ag-Au-Te Systems by the EMF method: Geology of Ore Deposits, v. 53, p. 247-258.

Gaál, 1997, Structural geology of the Serra Pelada gold deposit in Carajás, Pará state: Internal report.

Gammons, C.H., Yu, Y., & Bloom, M.S., 1992, Experimental investigation of the hydrothermal geochemistry of platinum and palladium: III. The solubility of Ag-Pd alloy + AgCl in NaCl/HCl solutions at 300°C. Geochimica et Cosmochimica Acta, v.57, p. 2469-2479. Gammons, C.H., 1996, Experimental investigations of the hydrothermal geochemistry of platinum and palladium: V. Equilibria between platinum metal, Pt (II), and Pt (IV) chloride complexes at 25 to 300°C: Geochimica et Cosmochimica Acta, v. 60, p. 1683– 1694.

Goldstein, R.H., Reynolds, T.J., 1994, Systematics of fluid inclusions in diagenetic minerals. Society of Sedimentary Geology Short Course Series 31.

Grainger, C.J., Groves, D.I., and Costa, C.H.C., 2002, The epigenetic sediment hosted Serra Pelada Au-PGE deposit and its potential genetic association with Fe-oxide Cu-Au mineralization within the Carajás mineral province, amazon craton, Brazil: Society of Economic Geologists Special Publication 9, p. 47-64.

Grainger, C.J., 2003., The Epigenetic sediment-hosted Serra Pelada Au-PGE deposit and its potential genetic association with varying styles of Cu-Au mineralization within the Carajás mineral province, amazon craton, Brazil. University of Western Australia, Doctoral Thesis, 259p. 96

REFERENCES

Groves, D.I., Goldfarb, R.J., Gebre-Mariam, M., Hagemann, S.G., and Robert, F., 1998, Orogenic gold deposits: A proposed classification in the context of their crustal distribution and relationship to other deposit types: Ore Geology Reviews, v. 13, p. 7–27.

Guillong, M., Meier, D., Allan, M., Heinrich, C., and Yardley, B., 2008, SILLS: A MATLAB- based program for the reduction of laser ablation ICP-MS data of homogeneous materials and inclusions. Mineralogical Association of Canada Short Course Series, v. 40, p. 328-333.

Günther D., Frischknecht R., Heinrich C. A., and Kahlert H. J., 1997, Capabilities of an argon fluoride 193 nm excimer laser for laser ablation inductively coupled plasma mass spectrometry microanalysis of geological materials. Journal of Analytical Spectrometry, v. 12, p. 939–944.

Grundmann, G., Lehrberger, G. and Schnorrer-Kohler, G., 1990, The El Dragon mine, Potosi, Bolivia: Mineralogical Record, v. 21, p. 133-146.

Heinrich, C.A., and Eadington, P.J., 1986, Thermodynamic predictions of the hydrothermal chemistry of arsenic, and their significance for the paragenetic sequence of cassiterite - arsenopyrite - base metal sulfide deposits: Economic Geology, v.81, p.511-529.

Heinrich, C.A., Seward, T.M., 1990, A spectrophotometric study of aqueous iron (II) chloride complexing from 25 to 200 ºC: Geochimica et Cosmochimica Acta, v.54, p. 2207-2221.

Heinrich, C.A., Pettke, T., Halter, W.E., Aigner-Torres, M., Audetat, A., Günther, D., Hattendorf, B., Bleiner, D., Guillong, M., and Horn, I., 2003, Quantitative multi-element analysis of minerals, fluid and melt inclusions by laser-ablation inductively-coupled-plasma mass-spectrometry. Geochimica et Cosmochimica Acta, v. 67, p. 3473-3497.

Heyding, R.D., 1966, The copper/selenium system: Canadian Mineralogist, v. 10, p. 207-215.

Holdaway, M.J., 1971, Stability of andalusite and the aluminum silicate phase diagram: American Journal of Science, v. 271, p. 97-131.

Ingebritsen, S.E. and Manning, C.E., 2010, Permeability of the continental crust: Dynamic variations inferred from seismicity and metamorphism: Geofluids, v. 10, p. 193-205.

Jorge João, X. da S., Neves, A.P., Leal, J.W.L., 1982, Ouro da Serra Pelada, aspecto da geologia e garimpagem: Sociedade Brasileira de Geologia, Simpósio de Geologia da Amazônia, 1, Belém, v.3, p.52-61.

Kump, L.R., 2008, The rise of atmospheric oxygen: Nature, v.451 p. 277-278.

Nogueira, A.C.R., Truckenbrodt, W., and Pinheiro R.V.L., 1995, Formação Águas Claras, Pré- Cambriano da Serra dos Carajás: redescrição e redefinição litoestratigráfica: Boletim do Museu Paraense Emílio Goeldi, n.7, p.177-197.

Neuendorf, K.K.E., Mehl, and J.P. Jackson, J.A., 2005, Glossary of Geology: American Geological Institute, Fifth edition, 779p.

REFERENCES

Machado, N., Lindenmayer, Z., Krogh, T.E., and Lindenmayer, D., 1991, U-Pb geochronology of Archean magmatism and basement reactivation in the Carajás area, amazon shield, Brazil: Precambrian Research, v. 49, p.329-354.

Meireles, E.M.; Teixeira, J.T.; Lourenço, R.S.; and Medeiros Filho, C.A., 1982, Geologia, estrutura e mineralização aurífera de Serra Pelada: Sociedade Brasileira de Geologia, Anais do 32º Simpósio Brasileiro de Geologia, Salvador, v. 3, p. 900-911.

Meireles, E.M. and Silva, A.R.B., 1988, Depósito de Ouro de Serra Pelada, Marabá, Pará: C. Schobbenhaus & C.E.S Coelho (eds), Principais Depósitos Minerais do Brasil, Brasília, DNPM/CVRD, v. 3., p. 547-557.

Mernagh, T.P., Heinrich, C.A., Lecky, J.F., Carville, D.P., Gilbert, D.J., Valenta, R.K., and Wyborn, L.A.I., 1994, Chemistry of low-temperature hydrothermal gold, platinum and palladium (±U) mineralization at Coronation Hill, Northern Territory, Australia: Economic Geology, v.89, p.1053-1073.

Mernagh, T.P., Wyborn, L.A.I., and Jagodzinski, E.A., 1998, Unconformity-related U ± Au ± platinum group-element deposits: Australian Geological Survey Organization Journal of Australian Geology and Geophysics, v. 17, p. 197-205.

Millero, F.J., Huang, F., and Laferiere, A.L., 2002b, The solubility of oxygen in the major sea salts and their mixtures at 25°C: Geochimica et Cosmochimica Acta, v. 66, n. 13, p. 2349- 2359.

Montoya, J.W. and Hemley, J.J, 1975, Activity relations and stabilities in alkali feldspar and mica alteration reactions: Economic Geology, v.70, p.557-583.

Moroni, M., Ferrario, A., and Girardi, V.A.V., 1999, Towards a model for the Serra Pelada Au- PGE deposit, Serra dos Carajás, Brazil: Stanley, C.J. (eds), Processes to processing: Balkema, Rotterdam, p. 759-785.

Moroni, M., Girard, V.A.V., and Ferrario, A., 2001, The Serra Pelada Au-PGE deposit, Serra dos Carajás (Pará State, Brazil): geological and geochemical indications for a composite mineralising process: Mineralium Deposita, v. 36, p. 768-785.

Mountain, B.W., Wood, S.A., 1988, Chemical controls on the solubility, transport and deposition of platinum and palladium in hydrothermal solutions: a thermodynamic approach: Economic Geology, v.83, p. 492-510.

Ohmoto, H., 1972, Systematics of Sulfur and Carbon Isotopes in Hydrothermal Ore Deposit: Economic Geology, v. 67, p. 551-578.

Pidgeon, R.T., Macambira, M.J.B., and Lafon, J.M., 2000, Th-U-Pb isotopic systems and internal structures of complex zircons from an enderbite from the Pium Complex, Carajás Province, Brazil: evidence for the ages of granulite facies metamorphism and the protolith of the enderbite: Chemical Geology, v. 166, p. 159-171.

REFERENCES

Pinheiro, R.V.L., and Holdsworth, R.E., 2000, Evolução tectonoestratigráfica dos sistemas transcorrentes Carajás e Cinzento, Cinturão Itacaiúnas, na borda leste do Craton Amazônico, Pará. Revista Brasileira de Geociências, v. 30, p. 597-606.

Prado-Pereira, R.M., 2009, Geologia da região sul da Serra Norte e características do minério de ferro do depósito de N8, Província Mineral de Carajás: Master Thesis, Universidade Federal de Minas Gerais.131p.

Rosière, C.A., Baars, F.J., Seoane, J.C.S., Lobato, L.M., Silva, L.L., Souza, S.R.C., and Mendes, G.E., 2005, Structure and iron mineralisation of the Carajás Province: Applied Earth Science, IMM Transactions section B, n.115, p.126-133.

Shepherd, T.J., Bouch, J.E., Gunn, A.G., McKervey, J.A., Naden, J., Scrivener, R.C., Styles, M.T., and Large, D.E., 2005, Permo-Triassic unconformity-related Au-Pd mineralisation, South Devon, UK: new insights and the European perspective: Mineralium Deposita v. 40, p. 24- 44.

Simon, G., Essene, E.J., 1996, Phase Relations among Selenides, Sulfides, Tellurides, and Oxides: I. Thermodynamic Properties and Calculated Equilibria: Economic Geology, v. 91, p. 1183- 1208.

Simon, G., Kesler, S.E., Essesne, E.J., 1997, Phase Relations among Selenides, Sulfides, Tellurides, and Oxides: II. Applications to Selenide-Bearing Ore Deposits: Economic Geology, v. 92, p. 468-484.

Stanley, C.J., Criddle, A.J. and Lloyd, D., 1990, Precious and base metals selenide mineralization at Hope’s Nose, Torquay, Devon: Mineralogical Magazine, v. 54, p. 485-493.

Tallarico, F.H.B., Coimbra, C.R., and Costa, C.H.C., 2000, The Serra Leste sediment hosted Au- (Pt-Pd) mineralization, Carajás province: Revista Brasileira de Geociências, v. 30, n.2, p. 226- 229.

Vasconcelos, P.M., Renne, P.R., Brimhall, G.H, and Becker, T.A., 1994, Direct dating of weathering phenomena by 40Ar/39Ar and K-Ar analysis of supergene K-Mn oxides: Geochimica et Cosmochimica Acta, v. 58, p. 1635-1665.

Villas, R.N., and Santos, M.D., 2001, Gold Deposits of Carajás Mineral Province: Deposits Types and Metallogenesis: Mineralum Deposita, v. 36, p. 300-331.

Wälle, M., and Heinrich, C.A., 2013, The Potential of Using a Sector Field ICP-MS for Analysis of Fluid Inclusions by Laser Ablation ICP-MS: Mineralogical Magazine, v. 77, p. 2441.

Wiegers, G.A., 1976, Electronic and ionic conduction of solid solutions Ag2-x-AuxSe ( 0 < X < 0.5): Journal of Less Common Metals, v. 48, p. 269-283.

Wilde, A.R., and Wall, V.J., 1987, Geology of the Nabarlek uranium deposit, Northern Australian Territory, Australia: Economic Geology, v. 82, p. 1152-1162.

REFERENCES

Wilde, AR., Bloom MS., and Wall VJ., 1989, Transport and deposition of gold, uranium and platinum-group elements in unconformity-related uranium deposits: The geology of gold deposits: the perspective in 1988, Economic Geology Monograph 6, p. 637-650.

Wilde, A.R., Layer, P., Mernagh, T., and Foster, J., 2001, The giant Muruntau gold deposit: Geologic, geochronologic and fluid inclusion constraints on ore genesis: Economic Geology, v. 96, p. 633–644.

Wilde, A., Edwards, A., Yakubchuk, A., 2003, Unconventional Deposits of Pt and Pd: A review with implications for exploration: Economic Geology Newsletter n°52.

Wall, V.J., Berni, G.V., and Grainger, C.J., 2009, Serra Pelada, Brazil: A high-grade, hydrothermal gold-platinum-palladium system. P.J. Williams et al (eds), Proceedings of the 10th. Biennial SGA Meeting Townsville Australia 2009, p. 74-76.

100

REFERENCES

ACKNOWLEDGMENTS

Many people should be acknowledged for contributing either directly or indirectly to this PhD project. And, as I am known for not having the best memory for such things, I first would like to thank everyone that contributed to it and might have been forgotten here.

A special thanks goes to Vic J. Wall. He was directly responsible for convincing the exploration company to finance this scientific research. Also, his judgment that an exploration geologist from Carajás should learn some geochemistry and thermodynamics is probably the only reason why I came to ETH Zürich. Since 2007, when I started my life in exploration geology, his guidance made of this period the greatest learning experience I ever had.

Another big thanks goes Christoph A. Heinrich, my supervisor. Even though it drove me crazy for most of the times, his unique character taught me to look at things from a different perspective. And to Lydia M. Lobato, my co-supervisor, that even from the distance was always available to help and discuss any kind of problems. Also a big thanks to my colleagues from the fluids and minerals research group, for the scientific help and many discussions. It was an amazing experience to have this opportunity of being between so many bright and nice people. A special thanks to Tobias Schlegel, which always made himself available for discussions at any time, and also for celebrations or complaining beers.

All the Colossus Minerals Inc. employees also deserve to be acknowledged. They provided an amazing assistance during the 3 years I was living and working at the Serra Pelada and when I came back as a PhD student. People like our driver like Paulinho, Linda or Vandinha, which no matter how bad was the situation going, could always cheer one up with their simplicity.

Also to Anastasia Papangelou, her company and affection during concentration and especially during relaxing moments was somehow my safe-harbor during this last year in Zürich.

Of course a big thanks to all my friends in Zürich, I have lived the best time of my life in this city (at least during summer :P) and it was great to meet you all! A special thanks to Alejandro Beltran- Trivino, parce, for his fellowship since my first days in Zürich. And, of course, to my dearest flat-mate Stefanie Luginbühl, for all the nice and a very few stressful moments we had lived together.

And least but not least, to my parents, which even though were suffering a bit from the distance, always supported every single decision I took I my life.

101

APPENDIX I APPENDIX I

Thermodynamic Calculations

This appendix provides the thermodynamic calculations used for constructing the pH vs. ƒO2(g) diagrams of Figure 4.12.

The stability fields and solubility contours for selenide and sulfur aqueous species, Fe minerals and Pt as chlorite complexes were taken from Mountain and Wood (1988). Graphite stability field is according to Ohmoto (1972). These are exactly as published and therefore calculations are not shown in this appendix.

The stability fields for sudovikovite (PtSe2), luberoite (PtSe0.8), sperrylite (PtAs2) and native platinum were re-calculated using thermodynamic constraints from fluid inclusion LA-SF- ICPMS data. A simplified version of the diagram, showing exclusively the minerals and aqueous species included in the reactions will be always precede the calculations. The lines defining the stability fields of minerals are labeled according the respective equation in the text.

1 - Stability fields of Platinum Selenides and aqueous species – Figure 4.12A

Figure 1 – pH – ƒO2 diagram showing the stability fields (Se activity of -6 m) of selenide aqueous species (thin dash-dotted lines), and platinum-selenide minerals (blue lines) which include: Native platinum (Pt), PtSe0.8 (luberoite) and PtSe2 (sudovikovite). The label on each line delineating the stability field of the minerals are a reference for the calculations. From Mountain and Wood, (1988).

102

APPENDIX I

1.1 - Native Platinum to PtSe0.8

H2SeO3 field

The boundary between the stability fields of native platinum relative to PtSe0.8 within the

H2SeO3 field is defined by the following reaction (A1 – Figure 1):

A1 1.25Pt(s) + H2SeO3(aq) = 1.25PtSe0.8(s) + O2(g) + H2O(l)

log K = log [H2O] + log ƒO2(g) + 0 - log [H2SeO3] - 0

log K = log ƒO2(g) - log [H2SeO3]

From the stoichiometry of the log K equation, the reaction will be represented by a horizontal line on the pH versus log ƒO2 (g) plots, since it depends only on oxygen fugacity and selenium concentration.

The effect of an increase of 1 order of magnitude in Se concentration will increase the oxygen fugacity of the saturation point in 1 log unit, which can be verified by calculating log K using values from calculated from Mountain and Wood (1988) diagram:

At activity of selenium = 10-6 m

log K = log ƒO2(g) - log [H2SeO3]

log K = -16.15 - (-6) = -10.15

Given Se activity.

Read from Mountain and Wood (1988).

At activity of selenium = 10-5 m

log ƒO2(g) = -10.15 + log [H2SeO3] log ƒO2 = -15.15

103

APPENDIX I

H2SeO3- field

The boundary between the stability fields of native platinum relative to PtSe0.8 within the

H2SeO3- field is defined by the following reaction (A2 – Figure 1):

A2 1.25Pt(s) + HSeO3-(aq) + H+(aq) = 1.25PtSe0.8(s) + O2(g) + H2O(l)

log K = log ƒO2(g) - log [H2SeO3] - pH

At the H3SeO3 /H2SeO3- boundary, an increase of one order of magnitude in selenium concentration will increase the oxygen fugacity of the saturation of PtSe0.8 in respect to native

Pt in 1 log unit. The reaction at H2SeO3- field will be represented by an inclined line with a negative slope defined by the pH and log ƒO2 (g) stoichiometry, with a decrease of 1 ƒO2 (g) log unit per unit of (increasing) pH.

H2Se field

The boundary between the stability fields of native platinum relative to PtSe0.8 within the H2Se field is defined by the following reaction (A3 – Figure 1):

A3 1.25Pt(s) + H2Se(aq) + 0.5O2(g) = 1.25PtSe0.8(s)+ H2O(l) log K = - log [H2Se] - 0.5log ƒO2(g)

From the stoichiometry of the log K equation, the reaction is represented by a horizontal line on the pH versus log ƒO2 (g) plots, since it depends only on oxygen fugacity and selenium concentration. The effect of an increase of 1 order of magnitude in Se concentration will decrease the oxygen fugacity of the saturation point in 2 log units, which can be verified by calculating log K using values from Mountain and Wood (1988) diagram:

At activity of selenium = 10-6 m

A3 – log K = -0.5log ƒO2 (g) -log [H2Se-] log K =-0.5*(-36.92)-(-6) =18.21+6=24.21

At activity of selenium = 10-5 m

24.21 = -0.5log ƒO2(g) + 5

-19.21 = -0.5log ƒO2(g) log ƒO2(g) = -38.92

104

APPENDIX I

H2Se- field

The boundary between the stability fields of native platinum relative to PtSe0.8 within the H2Se- field is defined by the following reaction (A4 – Figure 1):

A4 1.25Pt(s) + HSe-(aq) + H+(aq) + 0.5O2(g) = 1.25PtSe0.8(s) + H2O(l)

log K = -0.5log ƒO2(g) - log [H2Se-] - pH

At the H2Se/HSe- boundary, an increase of one order of magnitude in selenium concentration will decrease the oxygen fugacity of the saturation of PtSe0.8 in respect to native Pt in 2 log units. The reaction at HSe- field is represented by an inclined line with a positive slope defined by the pH and log ƒO2 (g) stoichiometry, with an increase of 2 ƒO2(g) log units per unit of pH.

1.2 - PtSe0.8 to PtSe2

H2SeO3 field

The boundary between the stability fields PtSe2 relative to PtSe0.8 within the H2SeO3 field is defined by the following reaction (B2 – Figure 1):

B1 1.25 PtSe0.8(s) + 1.5H2SeO3(aq) = 1.25PtSe2 (s)+ 1.5O2(g) + 1.5H2O(l)

log K = 1.5log ƒO2(g) – 1.5log [H2SeO3]

log K = -19*1.5 – 1.5*(-6) = -19.5

Given Se activity.

Read from Mountain and Wood, 1988.

At activity of selenium = 10-5 m

log K =1.5 log ƒO2(g) -1.5log [H2SeO3]

log ƒO2(g) = -18

The effect of an increase of 1 order of magnitude in Se concentration will increase the oxygen fugacity of the saturation point of PtSe2 in respect to PtSe0.8 in 1 log unit.

105

APPENDIX I

H2SeO3- field

The boundary between the stability fields PtSe2 relative to PtSe0.8 to within this selenide aqueous specie predominance field is defined by the following reaction (B2 – Figure 1):

B2 1.25PtSe0.8(s) + 1.5HSeO3-(aq) 1.5H+(aq) = 1.25PtSe2(s) + 1.5O2(g) + 1.5H20(l)

log K = 1.5log ƒO2(g) – 1.5log [H2SeO3] – 1.5log [H+]

log K = 1.5log ƒO2(g) – 1.5log [H2SeO3] – 1.5pH

At the H3SeO3 /H2SeO3- boundary, an increase of one order of magnitude in selenium concentration will increase the oxygen fugacity of the saturation of PtSe2 in respect to PtSe0.8 in

1 log unit. The reaction at H2SeO3- field will be represented by an inclined line with a negative slope defined by the pH and log ƒO2 (g) stoichiometry, with an decrease of 1 ƒO2 (g) log unit per unit of (increasing) pH.

H2Se field

The boundary between the stability fields PtSe2 relative to PtSe0.8 within this selenide aqueous specie predominance field is defined by the following reaction (B3 – Figure 1):

B3 1.25PtSe0.8(s) + H2Se (aq) + 0.5O2 (g) = PtSe2 (s) + H20 (l)

log K = -0.5log ƒO2(g) - log [H2Se]

From the stoichiometry of the log K equation, the reaction is represented by a horizontal line on the pH versus log ƒO2 (g) diagrams, since it depends only on oxygen fugacity and selenium concentration. The effect of an increase of 1 order of magnitude in Se concentration will decrease the oxygen fugacity of the saturation point in 2 log units, which can be verified by calculating log K using values from Mountain and Wood diagram:

log K = -0.5(-31.5) - (-6) = 21.75

Given Se activity.

Read from Mountain and Wood, 1988.

At activity of selenium = 10-5 m

0.5log ƒO2(g) = - log [H2Se]- log K

log ƒO2(g) = -(-5)-21.75 = -33.5

106

APPENDIX I

H2Se- field

The boundary between the stability fields PtSe2 relative to PtSe0.8 within this selenide aqueous specie predominance field is defined by the following reaction (B4 – Figure 1):

B4 1.25PtSe0.8(s) + HSe-(aq) + H+(aq) +0.5O2(g) = PtSe2 (s) + H20(l)

log K = -0.5log ƒO2(g) - log [H2Se] - log[H+]

At the H2Se/HSe- boundary, an increase of one order of magnitude in selenium concentration will decrease the oxygen fugacity of the saturation of PtSe2 in respect to native Pt in 2 log units. The reaction at HSe- field is represented by an inclined line with a positive slope defined by the pH and log ƒO2 (g) stoichiometry, with an increase of 2 ƒO2 (g) log units per unit of pH.

2 - Stability fields of Platinum Arsenide (sperrylite) – Figure 4.12B

Figure 2 – pH – ƒO2 (g) diagram with labels referring to the equations used for construction Fig.4.11 B

107

APPENDIX I

1.1 - Native Platinum to PtAs2

H3AsO3 field

The boundary between the stability fields of native platinum relative to PtAs2 is defined by the following reaction (A – Figure 2):

A1 Pt(s) + 2H3AsO3(aq) = PtAs2(s) +3H2O(l) + 1.5O2(g)

log K =1.5log[O2] - 2log[H3AsO3]

From the stoichiometry of the log K equation, the reaction is represented by a horizontal line on the pH versus log ƒO2 (g) diagram, since it depends only on oxygen fugacity and arsenic concentration.

The log K of reaction can be calculated, using values from Mountain and Wood (1988) diagram:

log K= 1.5logƒO2(g - 2log[H3AsO3]

log K =1.5logƒO2(g) -2log[H3AsO3] = 1.5(-27) -2(-4) = -32.5

Given As activity.

Read from Mountain and Wood, 1988

Therefore, we can apply the Arsenic concentration from Serra Pelada fluid inclusions LA-

ICPMS analysis. The stability field between native Pt and sperrylite (PtAs2) will shift 1.5 log units to higher values, as calculated below:

At As activity of = 10-2.9 m (~100ppm)

A2 log K = 1.5logƒO2(g) -2log[H3AsO3]

-32.5 = 1.5logƒO2(g) -2*(-2.875)

logƒO2(g) = -25.5

H2AsO3- field

B Pt(s) + H2AsO3- (aq)+ H+(aq)= 0.5PtAs2(s) +H2O (l)+ O2(g)

log K =logƒO2(g) -log[H3AsO3] - pH

108

APPENDIX I

At the H2AsO3/ H3AsO3- boundary, an increase in arsenic to the values measured in the fluid inclusions will shift the stability field between native PtSe0.8 over PtSe2 in 1.5 log units. The reaction at the H2AsO3- field is represented by an inclined line with a negative slope defined by the pH and log ƒO2 (g) stoichiometry, with a decrease of 1 ƒO2 (g) log units per unit of pH, going into the H2AsO4- field described below.

H2AsO4 - field

The stability field of native platinum relative in respect to will be defined by the following reaction:

C 0.5Pt(s) + H3AsO4- (aq)+ H+(aq)= 0.5PtAs2(s) +2H2O (l)+ O2(g)

log K =logƒO2(g) -log[H3AsO4-] - pH

At the H2AsO3/ H3AsO4- boundary, an increase in arsenic concentration to the values measured in the fluid inclusions at Serra Pelada will shift the stability field between native PtSe0.8 over

PtSe2 in 1.5 log units. At the H3AsO4- field, the line representing the reaction after the H2AsO3- boundary will be defined by a negative slope (defined by the pH and log ƒO2 (g) stoichiometry), with an increase of 1ƒO2 (g) log unit per unit of pH.

3 - Au solubility in the absence of Se, As or S

Au(s) + 2Cl-(aq) + H+(aq) + 0.25O2(g) = AuCl2-(aq) + 0.5H2O(l)

Log K = log[AuCl2-] + log[H2O] – log[Au(s)] – 2log[Cl-] – pH – 0.25logƒO2(g) log K at 300 °C = 4.1185 (Calculated using the Geochemist Workbench - Bethke, C.M., 2008

109

APPENDIX I I APPENDIX II Sample list This appendix is sample list including a brief description of the lithological unit, alteration features and Au + Pt + Pd grades (when not available sample were labeled with N.A.).

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

Fine grained to microcrystalline silica, overprinted by small and thin (3-4mm) S.001 SPD002 180.0 Sandstone Siliceous 0.04 quartz veins. Secondary hydrated iron oxides filling fractures. Weakly altered siltstone w ith disseminated hematite. Bedding parallel (10° to core S.002 SPD002 205.1 Siltstone - 1.35 axis) w ith parallel quartz-kaolinite vein. Secondary hydrated iron oxides filling fractures. Intense quartz-kaolinite veining (locally brecciated) hosted by carbon upgraded S.003 SPD002 215.7 Siltstone Carbon-rich argillic 1.40 siltstone. Coarse grained quartz grains associated w ith kaolinite. Secondary hydrated iron oxides also occur filling fractures. Carbon upgraded siltstone, bedding sub parallel to core axis, show ing S1 S.004 SPD002 225.8 Siltstone Carbon-rich argillic 11.5 foliation and small scale F1 folds. Carbonaceous micro breccia. All primary structures are obliterated. kaolinite S.005 SPD002 229.0 Siltstone Carbon-rich argillic 872 occurs as small aggregates, locally w ith quartz. Secondary hydrated iron oxides filling fractures. Intense kaolinite rich argillic zone in contact w ith carbon upgraded zone. S.006 SPD002 229.2 Siltstone Kaolinite-rich argillic 872 Essentially composed by kaolinite, locally w ith associated coarse grained quartz and carbon upgraded fragments. S.007 SPD002 238.2 Siltstone Carbon-rich argillic 54.9 Carbonaceous breccia w ith visible gold. S.008 SPD002 244.5 Siltstone Siliceous 8.69 Silicified mineralized siltstone, locally w ith quartz veining along bedding. hematite-rich argillic zone overprinting carbon upgraded zone. Locally some S.009 SPD002 252.2 Siltstone Hematite-rich argillic 1.71 preserved carbonaceous portions w ith intense kaolinite + iron oxide rich veining/brecciation. Strongly silicified siltstone(?). Coarse grained quartz w ith kaolinite veining along S.010 SPD002 267.1 Siltstone Siliceous 0.02 subvertical fracture. Carbonaceous micro breccia. All primary structures are obliterated. kaolinite S.011 SPD002 279.8 Siltstone Carbon-rich argillic 157 occurs as veins or small aggregates, locally w ith quartz. Extremely fine grained hematite-rich argillic zone w ith minor kaolinite and S.012 SPD002 289.6 Siltstone Hematite-rich argillic 0.35 fragments carbonaceous siltstone. Fine grained to microcrystalline silica, locally w ith iron oxides disseminated along S.013 SPD033 199.3 Sandstone Siliceous 0.04 specific silica bands. Fine grained to microcrystalline ailica w ith intense sericite-quartz veining, mostly S.014 SPD033 206.3 Siltstone Siliceous 0.16 w ith bedding parallel orientation (20° to core axis). 110

APPENDIX II

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

Carbon upgraded siltstone w ith preserved primary structures. Locally quartz- S.015 SPD033 213.2 Siltstone Carbon-rich argillic 0.35 kaolinite veining parallel to bedding orientation. Carbon upgraded siltstone w ith preserved primary structures. Spaced quartz- S.016 SPD033 220.3 Siltstone Carbon-rich argillic 0.30 kaolinite veining parallel to bedding orientation. Carbonaceous breccia, no primary structures preserved. Massive carbon, S.017 SPD033 225.4 Siltstone Carbon-rich argillic 2.14 locally w ith disseminated kaolinite. Carbonaceous micro breccia, no primary structures preserved. Massive carbon, S.018 SPD033 241.2 Siltstone Carbon-rich argillic 1.92 locally w ith disseminated kaolinite. Weakly silicified siltstone, locally carbon rich portions preserved. Hydrated iron S.019 SPD033 248.2 Siltstone Siliceous 7.80 oxides filling fractures. Silicified siltstone, w ith kaolinite and iron oxides filling fractures. Locally S.020 SPD033 258.4 Siltstone Siliceous 25.2 preserved sulfides along fractures. S.021 SPD033 274.5 Siltstone Siliceous 8.4 Silicified siltstone, locally w eak silicified layer w ith disseminated kaolinite. Fine grained to coarse grained quartz. Secondary hydrated iron oxides filling S.022 SPD013 221.1 Sandstone Siliceous 0.87 fractures. S.023 SPD013 241.8 Sandstone Distal hematite 0.39 Intense kaolinite veining w ith associated hematite and coarse quartz grains. S.024 SPD013 246.2 Siltstone Siliceous 0.18 Silicified siltstone w ith disseminated kaolinite. S.025 SPD013 253.8 Siltstone Carbon-rich argillic 63.8 Carbonaceous breccia. No primary structures preserved. Disseminated kaolinite.

S.026 SPD013 255.3 Siltstone Carbon-rich argillic 208 Carbonaceous breccia. No primary structures preserved. Disseminated kaolinite.

Kaolinite-rich argillic zone, w ith preserved compositional layers defined by Fe- S.027 SPD013 257.1 Siltstone Kaolinite-rich argillic 4.1 oxide rich and poor layers (parallel to core axis). Carbon upgraded siltstone. Kaolinite rich veins occurs parallel to bedding. Locally S.028 SPD013 261.0 Siltstone Carbon-rich argillic 1.0 quartz rich veins cutting the previous kaolinite rich vein. Carbonaceous micro breccia. No primary structures preserved. Disseminated S.029 SPD013 262.1 Siltstone Carbon-rich argillic 12.0 kaolinite. Kaolinite rich siltstone hosted argillic breccia. (fault breccia?). Carbonaceous S.030 SPD013 267.5 Siltstone Kaolinite-rich argillic 1.5 fragments w ithin a iron oxide rich clay matrix. Silicified siltstone, primary structures still preserved. Locally quartz veins w ith S.031 SPD013 271.9 Siltstone Siliceous 0.07 associated fresh sulfides. Hematite- rich argillic zone, mostly w eathered to goethite, locally w ith associated S.032 SPD013 276.3 Siltstone Hematite-rich argillic 0.9 coarse grained hematite crystals. Coarse grained silicified sandstone w ith hematite veining. Well-formed quartz crystals along S1 parallel fracture. Fine silica w ith iron oxides correspond the S.033 SPD008 156.1 Sandstone Siliceous 0.07 dark portions w hile the coarse grained quartz is related to the w hiter portions of the rock.

111

APPENDIX I I

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

Silicified sandstone, locally w ith less altered portions show ing coarse grained S.034 SPD008 166.8 Sandstone Siliceous 91.9 w ell-formed quartz crystals. Thin to microcrystalline quartz, locally w ith quartz-kaolinite veins. Compositional S.035 SPD008 167.9 Siltstone Siliceous 176 layering (S0) still preserved. Kaolinite and secondary iron oxides also occur filling fractures. Sandstone hosted fault breccia. Sandy (siliceous matrix) w ith kaolinite and S.036 SPD008 176.7 Sandstone Kaolinite-rich argillic 7.65 silicified fragments. Carbon upgraded siltstone. Compositional bedding (S0), S1 and S2 foliations S.037 SPD008 261.4 Siltstone Carbon-rich argillic 4.56 preserved. Locally small kaolinite veins along fractures. Carbonaceous breccia, no primary preserved structures. Massive carbon w ith S.038 SPD008 266.9 Siltstone Carbon-rich argillic 462 disseminated kaolinite. Locally quartz-kaolinite occurs as small veins. Fe-rich argillic zone (low er limb). Mostly composed by red argillic layers, locally S.039 SPD008 269.4 Siltstone Hematite-rich argillic 0.02 w ith some Fe-rich silicified portions w ith preserved hematite. S.040 SPD008 272.9 Sandstone Siliceous 0.10 Microcrystaline dark silica cut by a fine-grained w hiter silica vein. S.041 FD-0156 209.3 Siltstone Siliceous 2.41 Fine grained to microcrystaline dark silica, locally w ith secondary quartz veins. Carbon upgrade siltstone, locally w ith disseminated kaolinite and coarse quartz S.042 FD-0156 222.5 Siltstone Carbon-rich argillic 0.61 grains. Carbon upgrade siltstone, w ith preserved compositional layering (S0), S1 and S2 S.043 FD-0156 228.6 Siltstone Carbon-rich argillic 1.76 foliations. Small and thin kaolinite veins parallel to bedding//S1 foliation. Carbonaceous breccia. All primary structures are obliterated. kaolinite occurs as S.044 FD-0156 241.5 Siltstone Carbon-rich argillic 193 small aggregates, locally w ith quartz. S.045 FD-0156 262.7 Siltstone Carbon-rich argillic 4.40 Carbon upgraded siltstone, locally w ith thick quartz parallel to S1 foliation. S.046 FD-0156 287.5 Sandstone Siliceous <0.005 Coarse grained silicified sandstone. S.047 FD-0164 258.2 Siltstone Siliceous 5.59 Silicified siltstone, bedding preserved. Weakly altered siltstone w ith associated garnet (oxidized) w ithin specific S.048 FD-0119 233.1 Siltstone Chlorite-carbon <0.005 compositional layering S.049 FD-0119 259.6 Siltstone Chlorite-carbon <0.005 Weakly altered siltstone w ith cut by quartz-sulfide vein. S.050 SPD001 193.5 Siltstone Kaolinite-rich argillic 24.7 Silicified portion w ithin a argillic breccia zone. Locally disseminated kaolinite. Argillic breccia. Brow n argillic matrix w ith kaolinite-rich portions and S.051 SPD001 223.5 Siltstone Kaolinite-rich argillic 16.1 carbonaceous fragments. S.052 SPD001 227.2 Siltstone Siliceous 0.4 Fine grained silica zone w ith hematite dusting. Locally disseminated pyrite. S.053 SPD001 233.6 Siltstone Kaolinite-rich argillic 154 Kaolinite-rich argillic in contact w ith carbon-rich argillic alteration. Silicified portion w ithin kaolinite-rich argillic zone, w ith quartz-kaolinite-sulfide S.054 SPD001 234.0 Siltstone Kaolinite-rich argillic 154 vein. Carbon-rich argillic ore w ith associated kaolinite-quartz rich veins (bedding S.055 SPD001 241.5 Siltstone Carbon-rich argillic 7.63 parallel).

112

APPENDIX II

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

S.056 SPD007 147.3 Siltstone Siliceous 0.04 Fine to microcrystalline dark silica cross-cut by quartz-kaolinite vein. S.057 SPD007 161.7 Sandstone Kaolinite-rich argillic 252 Sandstone hosted breccia. Kaolinite-rich matrix w ith silicified fragments. Quartz-kaolinite vein cross-cutting carbon-rich argillic ore. Coarse grained quartz S.058 SPD007 169.2 Siltstone Carbon-rich argillic 32.3 crystals. Carbonaceous breccia. All primary structures are obliterated. Disseminated S.059 SPD007 181.7 Siltstone Carbon-rich argillic 123 kaolinite. Silicified siltstone w ith vuggy lined by w ell-formed quartz crystals. Great amount S.060 SPD007 191.2 Siltstone Siliceous 1.21 of disseminated hematite Hematite-rich argillic. Dense rock, low quartz w ith breccia texture w ith hematite- S.061 SPD007 203.2 Siltstone Hematite-rich argillic 0.03 kaolinite matrix involving more coherent fragments metasiltstone. Carbon-rich argillic ore, locally w ith a coarse grained, cross-cutting kaolinite- S.062 SPD007 223.1 Siltstone Carbon-rich argillic 4.60 quartz vein. S.063 SPD007 238.3 Siltstone Siliceous 6.79 Contact betw een internal silica zone and carbon-rich argillic ore. S.064 FD-0126 223.2 Siltstone Kaolinite-rich argillic 0.85 Brow n kaolinite-argillic matrix w ith siltstone and silicified fragments. Kaolinite-rich argillic breccia w ith fragments of metasiltstone and locally euhedral S.065 FD-0126 223.5 Siltstone Kaolinite-rich argillic 0.85 quartz crystals. Weakly altered siltstone (chlorite-carbon) w ith cross-cutting stylolitic quartz S.066 FD-0145 210.8 Siltstone - <0.005 sulfiide vein. Weakly altered siltstone w ith associated carbon rich layer (primary??) w ith S2 S.067 SPD054 179.9 Siltstone - <0.005 foliation cutting the carbon rich layer. Fine grained to microcrystalline dark silica locally involving coarse grained Fe-rich S.068 SPD041 153.8 Siltstone Siliceous 0.08 quartz. Preserved sulfiides. S.069 SPD041 154.2 Siltstone Siliceous 0.08 Fine grained silica w ith associated sulfides. S.070 SPD041 185.6 Siltstone Carbon-rich argillic 6.64 Carbonaceous breccia w ith cross-cutting quartz-kaolinite vein. Weakly altered sandstone w ith small hematite veinlets. Locally coarse grained S.071 SPD041 254.0 Sandstone Distal hematite 1.03 quartz crystals. Hematite-rich argillic zone w ith remaining carbon and associated coarse quartz S.072 SPD041 263.8 Siltstone Hematite-rich argillic 0.05 grains. S.073 SPGT002 172.2 Siltstone Kaolinite-rich argillic 1.33 Fine grained to microcrystalline silica w ith quartz-sericite veinlets. S.074 SPGT002 221.2 Siltstone - 0.85 Weakly altered siltstone w ith associated quartz-sulfide (chalcopyrite) veins. S.075 SPGT002 274.7 Siltstone - 27.5 Weakly altered siltstone, locally w ith quartz vein and visible gold (?) Carbonaceous breccia, all primary structures are obliterated. Locally kaolinite- S.076 SPGT002 287.4 Siltstone Carbon-rich argillic 107 quartz rich aggregates. S.077 SPGT002 288.8 Siltstone Hematite-rich argillic 1.23 Hematite-rich argillic alteration, locally w ith small kaolinite aggregates. Argillic breccia, siltstone hosted. Carbonaceous matrix w ith associated iron S.078 FD-0110 195.3 Siltstone Carbon-rich argillic 1301 oxides veinlets. Breccia w ith a hematite-rich matrix w ith carbonaceous fragments of siltstone S.079 FD-0110 200.2 Siltstone Hematite-rich argillic 61 w ith associated kaolinite aggregates.

113

APPENDIX I I

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

S.080 FD-0106 306.7 Sandstone Distal hematite 76.8 Sandstone hosted hematite-dusting alteration. Locally coarse grained hematite. Carbonaceous breccia, locally some fragments w ith preserved primary S.081 FD-0122 305.8 Siltstone Carbon-rich argillic 53.6 structures. Kaolinite occurs as small aggregates. S.082 FD-0113 293.8 Siltstone Carbon-rich argillic 3.26 Carbonaceous breccia. No primary structures preserved. Disseminated kaolinite. S.083 FD-0113 294.7 Siltstone Siliceous 86 Mineralized silicified siltstone, locally pyrite. S.084 FD-0117 220.7 Sandstone Kaolinite-rich argillic 423 Argillic breccia zone associated w ith siliceous zone w ithin the sandstone. S.085 SPD028 206.8 Sandstone Siliceous 218 Fine grained silica zone, locally w ith vugs lined by quartz and hematite. Kaolinite-rich argillic breccia, sandy hematite-rich (w eathered to goethite) matrix S.086 SPD028 210.0 Siltstone Kaolinite-rich argillic 97 w ith siltstone and kaolinite fragments. S.087 FD-0187 114.5 Siltstone - 0.48 Weakly altered (mineralized) red siltstone. Locally bedding parallel quartz-vein. S.088 FD-0187 116.3 Siltstone - 9.09 Weakly altered red siltstone w ith carbon along fracture. S.089 FD-0172 140.3 Siltstone Hematite breccia 1078 Hematite-goethite rich matrix w ith siltstone and silicified fragments. S.090 FD-0172 141.3 Siltstone Hematite breccia 897 Hematite-goethite rich matrix w ith siltstone and silicified fragments. S.091 SPC022 390.2 Dolomitic conglomerate Hematite - Iron oxides N.A Dolomitic sandstone w ith coarse grained (up to 2cm) dolomite crystals. Dolomitic sandstone w ith associated hematite dusting. Locally quartz and S.092 SPC022 431.0 Dolomitic conglomerate Distal hematite N.A hematite veins. Carbon upgraded siltstone w ith tw o generation of quartz-kaolinite veins. The first S.093 SPD055 207.4 Siltstone Carbon-rich argillic N.A one is oriented along bedding and deformed by S2(?) foliation. The second is filling subhorizontal cracks cutting S2. Carbon upgraded siltstone w ith good strucutral relationships. Visible folded S.094 SPD055 208.0 Siltstone Carbon-rich argillic N.A microlithons w ith associated S2 crenulation. Locally deformed (by S2) quartz vein. High grade sample (informal comunication - Grainger 07/2010). Carbon upgraded S.095 SPD055 212.6 Siltstone Carbon-rich argillic N.A siltstone w ith associated quartz-kaolinite vein along bedding orientation.

Kaolinite rich argillic zone in contact w ith carbon upgraded siltstone (High grade S.096 SPD055 213.5 Siltstone Kaolinite-rich argillic N.A sample - Grainger informal comunication 07/2010). Carbon upgraded siltstone (High grade - informal comunication - Grainger S.097 SPD055 217.2 Siltstone Carbon-rich argillic N.A 07/2010). Weak disseminated kaolinite, bedding subparallel to core axis. Carbon upgraded siltstone (High grade - informal comunication - Grainger S.098 SPD055 217.7 Siltstone Carbon-rich argillic N.A 07/2010). Weak disseminated kaolinite. S.099 SPD034 211.8 Siltstone Siliceous 7.14 Silicified metasilstone w ith preserved sulfides. Carbonaceous breccia, all primary structures are obliterated. Quartz-kaolinite S.100 SPD034 219.5 Siltstone Carbon-rich argillic 3859 aggregates and carbonaceous siltstone fragments w ithin a carbonaceous matrix. Kaolinite-rich alteration w ith associated talk (?) v.ein cutting fresh dolomic S.101 SPD060 83.5 Dolomitic conglomerate Kaolinite-rich argillic N.A sandstone. 114

APPENDIX II

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

S.102 SPGT001 272.6 Siltstone Kaolinite-rich argillic 133.59 Siliceous portion w ithin kaolinite-rich argillic breccia. S.103 SPGT001 274.0 Siltstone Kaolinite-rich argillic 189.29 Siliceous portion w ithin kaolinite-rich argillic breccia. S.104 SPD055 223.2 Siltstone Kaolinite-rich argillic N.A Kaolinite-rich argillic breccia w ith carbonaceous fragments. S.105 SPD031 208.7 Siltstone Carbon-rich argillic 2.02 Carbon-rich argillic ore w ith folded (S2?) quartz-kaolinte veins S.107 SPD058 258.0 Siltstone Carbon-rich argillic N.A Quartz-sulfide vein w ithin carbon-rich argillic ore. S.108 SPD056 267.3 Siltstone - N.A Weakly altered siltstone. Both S1/S2 foliations visible. S.109 FD-0140 340.0 Dolomitic conglomerate Distal hematite <0.005 Dolomitic sandstone - hematite dusting. S.110 FD-0140 341.0 Dolomitic conglomerate - <0.005 Coarse grained dolomitic sandstone. S.111 SPD010 268.4 Siltstone Kaolinite-rich argillic <0.01 Kaolinite-rich argillic zone w ith chlorite. S.112 FD-0120 244.7 Siltstone Chlorite-carbon <0.005 Chlorite rich siltstone. S.113 FD-0120 210.5 Siltstone Chlorite-carbon <0.005 Weakly altered siltstone. Locally chlorite rich layers. S.114 FD-0133 63.9 Dolomitic conglomerate Distal hematite <0.005 Hematite dusted dolomitic sandstone cut by hematite vein. S.115 Field Siltstone - N.A Red siltstone, w ith associated carbonaceous layers. S.116 FD-0145 233.0 Siltstone Chlorite-carbon N.A Chlorite rich siltstone locally w ith associated garnets. Quartz veins w ithin Rio Novo metavulcanic. Euhedral quartz crystals w ith S.117 Field Metavulcanic Quartz vein N.A alternating clear and milky zones parallel to crystal grow ing faces. Mamão Garimpo. Quartz veins w ith associated hematite/goethite w ithin Rio Novo metavulcanics. S.118 Field Metavulcanic Quartz-hematite breccia N.A Mamão Garimpo Hydrothermally altered metavulcanic rock. Disseminated hematite and goethite S.119 Field Metavulcanic Hematite breccia N.A replacing previous mineralogy. Mamão Garimpo. S.120 Field Metavulcanic - N.A Least altered metavulcanic, although rock is highly w eathered. Quartz vein w ithin metavulcanic. Euhedral quartz crystals grow ing tow ards S.121 Field Metavulcanic Quartz vein N.A fracture core.Cutia Garimpo. S.122 Field Metavulcanic Quartz vein N.A Least altered metavulcanic, although rock is highly w eathered. S.123 Field Metavulcanic - N.A Least altered metavulcanic, although rock is highly w eathered. Sandstone hosted hematite breccia close to the contact w ith the metavulcanic S.124 Field Sandstone Hematite breccia N.A sequence. Road from Caracol to Seu Helio's farm. S.125 Field Siltstone - N.A Grey silstone outcropping nearby Serra Pelada open pit. S.126 SPD095 240.5 Siltstone Barren-argillic 0.99 Black, silicified siltstone w ith sericite (?) alteration. Black, silicified siltstone w ith sericite alteration (?), cross cutting previous qz S.127 SPD095 245.4 Siltstone Barren-argillic 0.07 veins. S.128 SPD099 303.2 Siltstone Barren-argillic 0.01 Black, silicified siltstone w ith sericite(?) alteration. Black siltstone w ith sericite (?) alteration. Synchronous w ith step axial surface S.129 SPD099 235.7 Siltstone Barren-argillic 0.44 folding (D2). Sericite zones are preferentially dissolved and the vugs are filed by Qz-hem. S.130 SPD095 272.9 Siltstone Siliceous 0.49 Silicified zone w ith veinlets of remobilized (?) carbonaceous material.

115

APPENDIX I I

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

S.131 SPD051A 214.2 Siltstone Siliceous 1.31 White silicification zone in contact w ith black silicified carbonaceous siltstone. S.132 SPD099 251.6 Siltstone Siliceous 0.82 Previous silicified siltstone w ith cross-cutting clear qz-hem vein. Black siltstone w ith sericite alteration. Core of sericite halos locally host vugs S.133 SPD099 227.6 Siltstone Barren-argillic 0.08 w ith qz and hem. S.134 SPD099 265.7 Siltstone Siliceous 0.35 Silicified siltstone cutted by at least 2 qz vein generations. S.135 SPD099 265.9 Siltstone Siliceous 0.35 Silicified siltstone cutted by latter qz vein generatios. Coarse grained silica zone w ith vugs filled by w ell-formed Qz-hem. Hematite S.136 SPD007 198.6 Siltstone Siliceous 0.92 forms concentric crystals, sometimes w ith quartz in the edges. S.137 SPD079 249.2 Siltstone Siliceous 0.06 Fine grain silica zone cut by clear qz vein. S.138 SPD001 183.2 Undefined Siliceous 0.05 Fine grain silica zone cut by S1 folded qz vein. Fine grained silica zone w ith associated hematite dusting and vuggy filled w ith S.139 SPD001 227.2 Siltstone Siliceous 0.48 w ell-formed Qz crystals +hem , sometimes Qz + kaolinite + hem. Locally disseminated pyrite. Coarse grained silica zone w ith vugs filled by w ell-formed Qz-hem. The hematite S.140 SPD051A 214.5 Siltstone Siliceous 1.31 is mainly w eathered to goethite. Hematite also visible filling small cracks w ithin the massive silica portions. ematite-dusted dolomitic sandstone w ith coeval red - hem rich cm-quartz + large S.141 SPC022 351.7 Dolomitic conglomerate Distal hematite N.A dolomite crystals. S.142 SPC022 268.3 Dolomitic conglomerate Distal hematite N.A Hematite-dusted dolomitic sandstone w ith cm large euhedral dolomite crystals. S.143 SPC022 429.3 Dolomitic conglomerate Distal hematite N.A Coeval qz-kaolinite in a crack w ithin the red-hem dusted dolomite. S.144 SPC022 384.2 Dolomitic conglomerate Distal hematite N.A Red hematite dusted dolomitic sandstone cut by milcky quartz vein. Chloritized dolomitic sandstone w ith vugs filled by cm coarse-grained dolomite S.145 SPC022 320.0 Dolomitic conglomerate Distal hematite N.A crytals. S.146 SPC022 430.2 Dolomitic conglomerate Distal hematite Red altered dolomite w ith s1 folded qz vein (metamorphic?). Hematite-dusted dolomitic sandstone w ith coeval red - hem rich cm-large quartz S.147 SPC022 351.6 Dolomitic conglomerate Distal hematite and dolomite crystals. Red hem-bearing dolomite w ith vuggy filled by red dolomite + clear calcite S.148 SPC022 425.4 Dolomitic conglomerate Distal hematite crystals. S.149 SPC022 424.5 Dolomitic conglomerate Distal hematite Hematite-dusted dolomitic sandstone w ith vuggy filled by red dolomite crystals S.150 SPGT011 89.9 Chlorite-schist - <0.005 Chlorite-sericite schist w ith metamorphic qz veins. S.151 SPGT011 44.8 Chlorite-talc schist - 0.01 Chlorite-talc schist. S.152 SPGT012 86.8 Chlorite-schist - <0.005 Chlorite-talc schist. S.153 SPGT011 83.1 Chlorite-schist - <0.005 Metamorphic qz vein w ithin secite-talc schist. Coarse grained dolomitic sandstone w ith quartz and BIF fragments w ithin a S.154 SPC022 392.7 Dolomitic conglomerate - N.A dolomite matrix. S.155 SPD039 533.0 Dolomitic conglomerate - N.A Fine grained dolomitic sandstone. S.156 SPD099 262.4 Siltstone - 0.13 Fine grained carbonaceous siltstone w ith thick, sandy layering preserved. S.157 SPD127 156.8 Red Siltstone Chlorite-carbon N.A Red siltstone w ith chlorite replacing specific clay rich layers. 116

APPENDIX II

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

S.158 SPD127 156.0 Red Siltstone Chlorite-carbon N.A Red siltstone w ith chlorite replacing preferentially clay rich layers. S.159 SPD066 93.2 Red Siltstone - <0.005 Red siltstone w ith early metamorphic (?) vein. S.160 SPD066 79.5 Red Siltstone - <0.005 Red sitlstone cut by hem-qz vein. S.161 SPD001 200.9 Siltstone Kaolinite-rich argillic 17.95 Kaolinite-rich argillic breccias - Fe-rich matrix S.162 SPD001 220.1 Siltstone Kaolinite-rich argillic 10.97 Kaolinite-rich argillic breccias - Fe-rich matrix Brow n, clay-rich matrix w ith free standing Qz crystals. All the material got S.163 SPD095 260.6 Siltstone Kaolinite-rich argillic 2318 disaggregated after sampling. Brow n, clay-rich matrix w ith free standing Qz crystals. All the material got S.164 SPD095 261.4 Siltstone Kaolinite-rich argillic 3691 disaggregated after sampling. Brow n, clay-rich matrix w ith free standing Qz crystals. All the material got S.165 SPD095 261.4 Siltstone Kaolinite-rich argillic 3691 disaggregated after sampling. Brow n, clay-rich matrix w ith free standing Qz crystals. All the material got S.166 SPD007 195.3 Siltstone Kaolinite-rich argillic 0.61 disaggregated after sampling. S.167 SPD099 238.0 Siltstone Kaolinite-rich argillic 1.83 Argillic breccia zone w ith step foliation and some clear qz for FI studies. Argillic breccia zone. Brow n argillic matrix w ith kaolinite-rich portions and S.168 SPD099 242.1 Siltstone Kaolinite-rich argillic 7961.00 carbonaceous fragments. S.169 SPD099 226.9 Siltstone Kaolinite-rich argillic 0.00 High grade kaolinite rich argillic breccia zone. S.170 SPD095 238.5 Siltstone Kaolinite-rich argillic 797.91 High grade kaolinite rich argillic breccia zone. S.171 SPD095 238.7 Siltstone Kaolinite-rich argillic 145.60 Kaolinite rich argillic breccia w ith iron-oxide-rich matrix. S.172 SPD095 260.2 Siltstone Kaolinite-rich argillic 13.33 Brow n, clay-rich matrix (kaolinite +red spots), w ith free standing Qz crystals. S.173 SPD095 260.3 Siltstone Kaolinite-rich argillic 13.33 Brow n, clay-rich matrix (kaolinite +red spots), w ith free standing Qz crystals. S.174 SPD120 172.8 Siltstone Kaolinite-rich argillic 0.32 Brow n, clay-rich matrix (kaolinite +red spots), w ith free standing Qz crystals. S.175 SPD099 241.7 Siltstone Kaolinite-rich argillic 7961 Silicified fragments w ithin argillic breccia zone. S.176 SPD099 240.2 Siltstone Kaolinite-rich argillic 7961.00 Silicified fragments w ithin argillic breccia zone. Carbonaceous breccia w ith bleached siltstone fragments and quartz filling S.177 SPD099 250.0 Siltstone Carbon-rich argillic 0.00 dissolution (?) vug. S.178 SPD099 240.5 Siltstone Kaolinite-rich argillic 7961.00 Silicified fragments w ithin argillic breccia zone. S.179 SPD099 210.9 Siltstone Carbon-rich argillic 711.00 High grade carbonaceous breccia - PGE mineralogy. S.180 SPD099 249.3 Siltstone Carbon-rich argillic 111.00 High grade carbonaceous breccia - PGE mineralogy. S.181 SPD010 193.5 Siltstone Carbon-rich argillic 632.00 High grade carbonaceous breccia - PGE mineralogy. S.182 SPD095 262.8 Siltstone Siliceous 11.40 Silicified carbonaceous siltstone w ith vugs containing euhedral quartz. S.183 SPD120 186.4 Siltstone Carbon-rich argillic 0.75 Carbonaceous breccia/ kaolinite-qz veins. S.184 SPD120 213.5 Siltstone Carbon-rich argillic 13.13 Quartz-kaolinite vein w ithin carbonaceous ore. S.185 SPD120 213.5 Siltstone Carbon-rich argillic 13.142 Quartz-kaolinite vein w ithin carbonaceous ore. S.187 SPD137 252.5 Sandstone Kaolinite-rich argillic N.A Free-standing qz crystals in argillic zone w ithin sandstone layering. S.188 SPD063 125.1 Siltstone Chlorite-carbon N.A Chloritized siltstone cutted by qz-hem veins. Hematite dusted dolomitic sandstone w ith coeval red - hem rich quartz + large S.189 SPD067 332.7 Dolomitic conglomerate Distal hematite N.A dolomite crystals 117

APPENDIX I I

Sample Hole Id Depth Lithology Alteration Au+Pt+Pd Description

Hematite dusted dolomitic sandstone w ith coeval red - hem rich quartz + large S.190 SPD067 336.0 Dolomitic conglomerate Distal hematite N.A dolomite crystals Hematite dusted dolomitic sandstone w ith coeval red - hem rich quartz + large S.191 SPD056B 336.0 Dolomitic conglomerate Distal hematite N.A dolomite crystals S.193 SPD086A 212.8 Siltstone Hematite breccia 50.96 Hematite rich breccia w ith silicified and metasiltstone fragments. S.194 SPD141 38.2 Siltstone Carbon-rich argillic N.A Carbon-rich argillic ore but carbon veintlets w hich overprint S1 foliation. S.195 SPD141 65.2 Siltstone Carbon-rich argillic N.A Carbon-rich argillic ore but carbon veintlets w hich overprint S1 foliation. S.196 SPD141 68.2 Siltstone Carbon-rich argillic N.A Carbon-rich argillic ore but carbon veintlets w hich overprint S1 foliation. S.197 SPD056B 277.0 Siltstone Carbon-rich argillic 0.012 Carbon-rich argillic ore show ing both S1 and S2 foliations. S.198 SPD120 528.3 Siltstone Carbon-rich argillic 21.54 Carbon-rich argillic oree w ith chlorite. S.199 SPD139 278.8 Siltstone Chlorite-carbon N.A Chlorite altered metasiltstone w ith chlorite oriented along S2 crenulation. Fine grained amphibolite cut by quartz veins. Close to cutia garimpo S.205 Surface sample Amphibolite - N.A 655280/9332962. Fine grained quartz w ith associated pyrite. Cutia garimpo, 653155/9339970 S.206 Surface sample - Quartz vein N.A (approx.) Clear, milk quartz vein cut by fine grained darker silica veinlets, Cutia Garimpo, S.207 Surface sample - Quartz vein N.A 653155/9339970 (approx.) Fine grained early quartz vein cut by clear quartz vein w ith euhedral quartz S.208 Surface sample - Quartz vein N.A coated by manganese oxides.653155/9339970 (approx.) Mafic ultramafic rock from Luanga complex, highly w eathered. Road outcrop. S.209 Surface sample Ultramafic schist - N.A 657876/9340682 S.210 Surface sample Formiga Granite - N.A Gabbro intrusion nearby Cutia Garimpo. 655694/9344077. S.211 Surface sample Formiga Granite - N.A Granite samprolite. 653050/9344150 (appox.) S.212 Surface sample Formiga Granite - N.A Quartz vein w ith leached prysmatic crystals. 653050/9344150 (appox.) S.213 Surface sample Formiga Granite - N.A Quartz vein w ith K-feldspar. 653050/9344150 (appox.) Saprolite of porphyry dike intrunding metasedimentary sequence close to Cigano S.214 Surface sample Porphyry dike - N.A Granite. 653050/9344150 (appox.) Sandstone hosted hematite breccia close to the contact w ith the metavulcanic S.215 Surface sample Sandstone Hematite breccia N.A sequence. Road from Caracol to Seu Helio's farm. 644279/9339879. Sandstone hosted hematite breccia close to the contact w ith the metavulcanic S.216 Surface sample Sandstone Hematite breccia N.A sequence. Road from Caracol to Seu Helio's farm.644279/9339879. S.217 Surface sample Sandstone Talk alteration N.A Talk alteration close to dioritic intrusion. Nearby the Serra Pelada open pit. Red siltstone w ith chlorite-carbon alteration along vertical crack and spreading S.218 Surface sample Red Siltstone Chlorite-carbon N.A along bedding. Serra Pelada decline. S.219 Surface sample Grey siltstone Chlorite-carbon N.A Slightley chlorite alrered grey siltstone. Serra Pelada decline. S.220 Surface sample Red Siltstone - N.A Least altered red-siltstone. S.221 Surface sample Grey siltstone - N.A Carbonaceous siltstone from Elefanto sycline.

118

CURRICULUM VITAE

PERSONAL INFORMATION

Gabriel Valentim Berni Born on the 26th of November in Ipatinga (MG), Brazil E-mail: [email protected] Telephone: + 55 31 3223-3522 / + 55 31 9776-3522 (mobile)

ACADEMIC FORMATION

ETH Zürich, Switzerland 2014 Doctor of Sciences of ETH Zürich. Dissertation: The Serra Pelada Au - Pd - Pt Deposit (Carajás, Amazon Craton, Brazil): Geology, Mineralogy and Hydrothermal Geochemistry of Ore Formation. Supervisor: Prof. Dr. C. Heinrich.

UFMG - Federal University of Minas Gerais, Brazil 2009 Master degree in Geology - Applied Economic Geology. Thesis: Geology and Hydrothermal Alteration of the Serra Pelada AU-PGE deposit (in Portuguese).

UFMG - Federal University of Minas Gerais, Brazil 2007 Bachelor degree in Geology.

EDUCATION AND TRAINING

Colorado School of Mines, USA Nov2008 Sampling Theory, Sampling Practices and their Economic Impact. Instructor: Dr. Francis F. Pitard (ca. 40h).

Coffey Mining Brazil Sep2008 Surpac Basics. Instructor: Dr. Vandersoni Martins (ca. 30h).

Resource & Exploration Mapping, Australia Nov2007 Mapinfo and Discover. Instructor: David Splatt.

119

CURRICULUM VITAE WORK EXPERIENCE

ETH Zürich, ERDW, Switzerland 2010-Jul2014

Lab. Responsible of the fluid inclusion microthermometry and microscopy laboratory. Microprobe operator of a JEOL JXA8200 instrument including ca. 100 h independent experience. Assistant in BSc and Msc courses: Crystals and Minerals, Ore Deposits I and II. Professional reference: Prof. Dr. C. Heinrich.

Colossus Geologia e Participações LTDA. Nov2007 - Fev2010 Exploration Geologist - Serra Pelada Project, based in Pará State, Amazon Region, Brazil. Management of diamond and RC drilling campaign, drill-core logging, drill-core database management, 3D geological modeling and resource evaluation. Professional references: Dr. Vic Wall.; Dr. Augusto Kishida.

Geogeraes LTDA. Jun2007 - Out2007 Exploration Geologist. Geological mapping and natural gas prospection under PETROBRAS supervision (São Francisco Basin – MG – Brazil). Professional reference: Leandro Barros Reis.

Conselho Nacional de Desenvolvimento Científico e Tecnológico - CNPq. Jun2006 - Jun2007 Scientific Student Scholarship. Analysis of deformation in Pressure Fringes of the Nova Lima Group and relationship with orogenic gold mineralization. Academic Reference: Prof. Dr. Carlos Alberto Rosière.

Geogeraes LTDA. Jun2005 - Jun2006 Trainee Geologist. 3D Geological modeling and management of GIS geological data of different prospects and mines, including banded iron formation deposits of Caetité (Bahia) and Lithium- Tantalum pegmatites (Volta Grande mine – Nazareno – MG). Professional Reference: Leandro Barros Reis.

Geoaktivan Geologia e Meio Ambiente LTDA. Dec2004 - Jun2005 Trainee Geologist. Main activities involving managing diamond drilling campaign, drill-core logging and topographic surveying. Professional Reference: Ricardo Sholz.

Geogeraes LTDA. Apr2003 - Nov2004 Trainee Geologist. Digitalizing and managing geological raster and vector GIS data. Professional Reference: Leandro Barros Reis. ______

LANGUAGES Portuguese: mother tongue. English: Excellent written, understanding and oral skills. Spanish: Excellent understanding, intermediate written and oral skills.

120

CURRICULUM VITAE

COMPUTER SKILLS

Advanced knowledge ArcGIS (10) including topographic and geostatistical extensions. Advanced knowledge of MapInfo (9.0) and Discover. Advanced knowledge of Surpac (6.0.3) Advanced knowledge of GEMCOM (5.3)

PUBLICATIONS Berni, G.V., Heinrich, C.A., Lobato, L.M., Wall, V.J., Rosière, C.A., Freitas, M.A., 2014, The Serra Pelada Au – Pd– Pt Deposit, Carajás, Brazil: Geochemistry, Mineralogy and Zoning of Hydrothermal Alteration: Economic Geology, in press. Berni, G.V. 2009 - Geologia e Alteração Hidrotermal do Depósito de AU-PGE de Serra Pelada. Master thesis. 116p. Universidade Federal de Minas Gerais. Lobato, L.M.; Baltazar, O.F.; Reis, L.B.; Achtschin, A.B.; Baars, F.J.; Timbó, M.A.; Berni G.V.; Mendonça, B.R.V. de; Ferreira, D.V. 2005. Projeto Geologia do Quadrilátero Ferrífero - Integração e Correção Cartográfica em SIG com nota explicativa. Belo Horizonte: CODEMIG. 1 CD-ROM.

ORAL PRESENTATIONS

Berni, G.V., Heinrich, C.A., Lobato, L.M., Wall, V.J., 2013, The Sediment-Hosted Au - Pd - Pt deposit of Serra Pelada (Carajás Mineral Province Brazil): a variant of unconformity related U ± Au ± PGE deposits?: Oral presentation - Abstracts of the 26th International Applied Geochemistry Symposium, Rotorua, New Zealand. Berni, G.V., Heinrich, C.A., Lobato, L.M., Wall, V.J., 2013, The Sediment-Hosted Au - Pd - Pt deposit of Serra Pelada (Carajás Mineral Province Brazil): a variant of unconformity related U ± Au ± PGE deposits?: Oral presentation - Abstracts of the 26th International Applied Geochemistry Symposium, Rotorua, New Zealand. Berni, G.V.; Lobato, L.M.; Rosière, C.A.; Wall, V.; Grainger, C.J.; Freitas, M.A.; Carvalho, C.M.; Depósito de Au- PGE de Serra Pelada: Caracterização da Alteração Hidrotermal, 06/2009, II Simpósio Brasileiro de Metalogenia - Oral Presentation - Vol. 1, pp. 1-2, Gramado, RS, Brasil, 2009.

CONFERENCE ABSTRACTS & POSTERS

Berni, G.V., Heinrich, C.A., Lobato, L.M., Wall, V.J., 2013, Fluid Mixing Processes at the Serra Pelada Au-PGE deposit, Amazon Craton, Brazil: Evidences from Hydrothermal alteration, Gordon Conference, Andover, NH. Wall, V.J.; Berni, G.V. and Grainger, C.J. 2010. Serra Pelada, Brazil: A high-grade, hydrothermal gold-platinum- palladium system. Roundup 2010. Technical Abstracts, p98-100. Vancouver, B.C. Berni, G.V.; Ferreira, A.S.; Prado-Pereira, R.M. 2007. Mapeamento Geológico (Escala 1:25.000) da Região entre os Municipios de Barão de Cocais e São Gonçalo do Rio Abaixo. Trabalho de Graduação. Universidade Federal de Minas Gerais. Berni, G.V.; Ferreira, A.S.; Prado-Pereira, R.M. Rosière, A.C. 2007; Análise da Deformação no Grupo Nova Lima e Associação com a Mineralização Aurífera Baseada em Franjas de Pressão. Anais do XIV Simpósio de Geologia de Minas Gerais e X Simpósio de Geologia do Sudeste novembro de 2007. Rosière C.A.; Baars F.J.; Scoli, J.C.; Lobato, L.M. Costa, L.P.; Berni, G.V. 2006. A estrutura da Província Mineral de Carajás. In Simpósio Brasileiro de Exploração Mineral - SIMEXMIN,2006, Ouro Preto.

121