A Fundamental Study of the Reductive Leaching of Chalcopyrite Using Metallic Iron
A FUNDAMENTAL STUDY OF THE REDUCTIVE LEACHING OF CHALCOPYRITE USING METALLIC IRON
by
NED AM ABED
B. Sc., Chemical Engineering, Jordan University of Science and Technology, Jordan, 1989
A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE
in
THE FACULTY OF GRADUATE STUDIES
Department of Metals and Materials Engineering
We accept this thesis as conforming to the,required standard
THE UNIVERSITY OF BRITISH COLUMBIA
April 1999
©Nedam Abed, 1999 In presenting this thesis in partial fulfilment of the requirements for an advanced
degree at the University of British Columbia, I agree that the Library shall make it
freely available for reference and study. I further agree that permission for extensive
copying of this thesis for scholarly purposes may be granted by the head of my
department or by his or her representatives. It is understood that copying or
publication of this thesis for financial gain shall not be allowed without my written
permission.
Department of
The University of British Columbia Vancouver, Canada
DE-6 (2/88) ABSTRACT
A fundamental study of the reductive leaching (decomposition) of chalcopyrite was performed in both sulfate and chloride media. This was done to understand the physical chemistry of the leaching reactions and the possibility of developing a process flowsheet. The main objective of reductive leaching is to achieve the enrichment of chalcopyrite by rejection of iron and sulfur.
A chalcopyrite concentrate containing around 60% chalcopyrite and analyzing around
28% copper was leached under reducing conditions using metallic iron. Various parameters were studied to understand their effect on leaching kinetics, including : temperature, particle size, agitation, acid concentration, molar ratios, and others. The leaching data were analyzed to determine the leaching mechanism and develop a kinetic model.
The leaching reaction was found to be rapid on fresh surfaces of the concentrate, but slows markedly in one hour, as a film of products, mainly chalcocite, forms on particle surfaces.
Iron was also found to enter the leach solution as soluble ferrous ions and sulfur is released as hydrogen sulfide. The general leaching reaction may be written as :
+ 2+ 2CuFeS2(s) + Fe(s) + 6H (aq) ->' Cu2S(s) + 3Fe (aq) + 3H2S(g)
The proposed mechanism is a series of reactions. It is envisaged to be composed of two parts : a corrosion mechanism, which is iron dissolution, and galvanic mechanism, which is chalcopyrite reduction.
The kinetic analysis indicated that the leaching reaction, which is electrochemical in nature, follows the shrinking core model under product layer diffusion control, and the rate determining step is the transport of one or more of reaction species, through the product layer.
The reaction was dependent on the initial acid concentration, chalcopyrite particle size and molar ratio of iron to chalcopyrite. Moreover, the reaction was independent of the rate of agitation beyond that required to provide a well-mixed reaction mixture.
Under stoichiometric conditions, room temperature and atmospheric pressure, the conversion (decomposition of chalcopyrite to simpler copper sulfides) was always below 60%, unless the initial amount of the reductant was increased or very fine chalcopyrite particles were used.
ii For the studied experimental conditions, the developed parabolic leaching model for sulfate media takes the form :
-33,880" 2/3 l-3(l-Xb) + 2(l-Xb) 7 [FT] exp t R v RT J and for chloride media,
The parabolic leaching behavior was confirmed from the successful estimation of the related thermodynamic and kinetic properties of the leaching systems.
The analysis of temperature dependence indicated that leaching increases with increasing temperature up to 65 °C. The apparent activation energy for leaching in sulfate media was estimated to be ~33 kJ/mol, and for chloride media, ~26 kJ/mol under stoichiometric conditions, in the temperature range 25-65 °C.
Chemical analysis was extensively used based on wet chemistry methods, which were
capable of demonstrating the general reaction stoichiometry including the composition of the new solid phase. Further, qualitative analysis by SEM confirmed the findings of the kinetic and chemical analysis.
The findings from the fundamental study show that conversion is preferred under high
solid pulp density (SPD), and back reaction kinetics have essentially little or no adverse effect. In an attempt to improve the conversion and utilize the results for a possible flowsheet, the concentrate was leached in the presence of excess chloride content, at -35% SPD. Based on material balance calculations, at room temperature and under near stoichiometric conditions, greater than 80% of the added chalcopyrite can be decomposed to yield copper sulfides
(chalcocite) by rejection of iron and sulfur, using size fractions smaller than 74 jam. As a result of these findings, a process flowsheet was developed and further investigation is required to demonstrate the viability of the proposed process.
iii TABLE OF CONTENTS
Abstract ii
List of Tables vi
. List of Figures ix
Acknowledgment xiii
Section 1 Introduction 1
Section 2 Literature Survey : 8
2.1 Thermodynamics of Chalcopyrite Leaching 8
2.2 Oxidative Leaching of Chalcopyrite 30
2.3 Non-oxidative Leaching of Chalcopyrite 37
2.4 Reductive Leaching of Chalcopyrite 38
2.5 Halide Media Leaching of Chalcopyrite 50
Section 3 Objectives 52
Section 4 Proposed Leaching Mechanism with Metallic Iron 53
Section 5 Experimental Methods 57
5.1 Materials 57
5.2 Methods 59
5.2.1 The Kinetic Study 59
5.2.2 The Process Study 62
5.2.3 Analysis Techniques 64
5.2.4 Reaction Product Characterization 64
5.3. Calculations 65
Section 6 Results and Discussion 67
6.1 Analysis of Reaction Kinetics 67
6.1.1 Effect of Agitation 79
iv 6.1.2 Effect of Temperature 83
6.1.3 Effect of Particle Size 92 6.1.4 Effect of Initial Acid Concentration 100
6.1.5 Effect of Metallic Iron Addition 109
6.1.6 Effect of Chalcopyrite Addition 114
6.1.7 Effect of Solid Pulp Density (SPD) 119
6.1.8 Leaching Rates of Chalcopyrite Particles 121
6.2 Schematic Representation of the Leaching Process 126
6.3 Concluding Remarks 138
6.4 Process Development 139
6.4.1 Leaching 139
6.4.1.1 Acid Effect 139
6.4.1.2 Ferrous Chloride Effect 142
6.4.1.3 Metallic Iron Effect 142
6.4.1.4 Particle Size Effect 144
6.4.2 Solid / Liquid Separation 148
6.4.3 Iron Removal 148
6.4.4 Solid Washing Unit 148
6.4.5 Hydrogen Sulfide Collection Unit 149
6.4.6 Process Advantages 149
6.5 Hydrogen Sulfide Treatment 151
Section 7 Conclusions 156
Section 8 Recommendations for Future Research 158
Bibliography 162
Appendix I : Leaching Kinetics Models 172
Appendix II : Chemical Analysis 183
v LIST OF TABLES
Table 1.1 : Common copper minerals 3
Table 1.2 : Electronic and structural properties of selected sulfide and oxide minerals 6
Table 2.1 : Summary of physical and chemical properties of group 11 (IB) metals 9
Table 2.2 : Selected solubility data for copper and related species 10
Table 2.3 : Thermodynamic values for some common species and reactions in copper aqueous chemistry 13
Table 2.4 : Reactions and thermodynamic equations used in constructing the Eh-pH
diagrams 16
Table 2.5 : Selected heat capacity values for different species 21
Table 2.6 : Standard thermodynamic data for different species 22
Table 2.7 : Selected physical constants for copper and other species 23
Table 2.8 : The oxidative leaching of chalcopyrite in sulfate media 30
Table 2.9 : Reduction potentials of some metals and minerals at standard conditions 40
Table 2.10 : Summary of reviewed research on chalcopyrite reductive leaching 49
Table 5.1 : Detailed chemical analysis of the tested chalcopyrite concentrate 57
Table 5.2 : The mineralogical composition of the tested chalcopyrite concentrate 57
Table 5.3 : Particle size distribution of the concentrate 58
Table 6.1 : Sample experimental leaching data for selecting a leaching model by Wen's method (sulfate media, stoichiometric run, 25 °C) 69 Table 6.2 : Sample experimental leaching data for selecting a leaching model by Wen's method (chloride media, stoichiometric run, 65 °C) 70
Table 6.3 : Agitation speed effect on reaction kinetics (stoichiometric runs, sulfate media, 25 °C) 80
Table 6.4 : Agitation speed effect on reaction kinetics (stoichiometric runs, chloride media, 25 °C). 80 Table 6.5 : Temperature effect on reaction kinetics (0.1 M H2S04 solution, stoichiometric runs, 25-85 °C) 84
Table 6.6 : Temperature effect on reaction kinetics (0.1 M HC1 solution, stoichiometric runs, 25-85 °C) 85
Table 6.7 : Temperature dependence of reaction rates and related thermodynamic values for sulfate and chloride media 89
Table 6.8 : Particle size effect on reaction kinetics (sulfate media, stoichiometric runs, 25 °C) 93
Table 6.9 : Particle size effect on reaction kinetics (chloride media, stoichiometric runs,
25 °C) 94
Table 6.10 : Particle size dependence of reaction rates for sulfate and chloride media 94
Table 6.11 : Acid concentration effect on reaction kinetics (sulfate media, constant CuFeS2 and Fe additions, 25 °C) 101 Table 6.12 : Acid concentration effect on reaction kinetics (chloride media, constant
CuFeS2 and Fe additions, 25 °C) 101
Table 6.13 : Hydrogen ion dependence of reaction rates (sulfate media) 104
Table 6.14 : Hydrogen ion dependence of reaction rates (chloride media) 107
Table 6.15 : Metallic iron effect on reaction kinetics (sulfate media, constant CuFeS2 and
H2S04 additions, 25 °C) 110
Table 6.16 : Metallic iron effect on reaction kinetics (chloride media, constant CuFeS2 and HC1 additions, 25 °C) 110
Table 6.17 : Chalcopyrite effect on reaction kinetics (sulfate media, constant H2S04 and Fe additions, 25 °C) 115
Table 6.18 : Chalcopyrite effect on reaction kinetics (chloride media, constant HC1 and Fe additions, 25 °C) 115
Table 6.19 : Iron released in solution at various chalcopyrite additions (sulfate media,
constant H2S04 and Fe additions, 25 °C) 117
Table 6.20 : Iron released in solution at various chalcopyrite additions (chloride media, constant HC1 and Fe additions, 25 °C) 117 Table 6.21 : Recorded final conversion at various SPD values for sulfate and chloride media 119
Table 6.22 : Estimated leaching rates of chalcopyrite particles at different temperatures 121
Table 6.23 : Estimated initial leaching rates of chalcopyrite particles at the respective sizes 123
Table 6.24 : Effect of HQ concentration on leaching using high SPD systems 141
Table 6.25 : Effect of ferrous chloride addition on leaching using high SPD systems 141
Table 6.26 : Effect of metallic iron addition on leaching using high SPD systems 143
Table 6.27 : Effect of particle size on leaching using high SPD systems 145
Table 6.28 : Physical and thermodynamic properties of hydrogen sulfide 152
Table 6.29 : Oxidation reactions of hydrogen sulfide 153
Appendix II Tables
Table II. 1 : Sample material balance calculations for low SPD systems 196
Table II.2 : Sample material balance calculations for high SPD systems 200 LIST OF FIGURES
Fig. 1.1 : The crystal structure of chalcopyrite 5
Fig. 2.1 : The Eh- pH diagram for the copper-water system at 298.15 K 12
Fig. 2.2 : The Eh-pH diagram for the copper-sulfur-water system at 298.15 K 14
Fig. 2.3 : The Eh-pH diagram for the copper-iron-sulfur-water system at 298.15 K 15
Fig. 2.4 : The Eh-pH diagram for the copper-iron-chloride-sulfur-water system at
298.15 K 28
Fig. 5.1 : Schematic diagram of the reaction vessel during the kinetic study 61
Fig. 5.2 : Schematic diagram of the reaction vessel during the process study 61
Fig. 6.1 : Plot of conversion vs. time for demonstrating Wen's method (sulfate media, stoichiometric run, 25 °C) 72 Fig. 6.2 : Plot of acid consumption and conversion vs. time (sulfate media, stoichiometric run, 25 °C) 72
1/3 Fig. 6.3 : Plot of In t vs. In (1-(1-Xb) ) as per Wen's method (sulfate media, stoichiometric run, 25 °C) 73
Fig. 6.4 : Plot of fluid film diffusion control model fitting of conversion vs. time (unchanging size particles, sulfate media, stoichiometric run, 25 °C) 73
Fig. 6.5 : Plot of fluid film diffusion control model fitting of conversion vs. time (changing size particles, sulfate media, stoichiometric run, 25 °C) 74
Fig. 6.6 : Plot of chemical control model fitting of conversion vs. time (changing and unchanging size particles, sulfate media, stoichiometric run, 25 °C) 74
Fig. 6.7 : Plot of Product layer control model fitting of conversion vs. time (unchanging size particles, sulfate media, stoichiometric run, 25 °C) ; 75
Fig. 6.8 : Plot of conversion vs. time for demonstrating Wen's method (chloride media, stoichiometric run, 65 °C) 75
Fig. 6.9 : Plot of Product layer control model fitting of conversion vs. time (unchanging size particles, chloride media, stoichiometric run, 65 °C) 76
ix Fig. 6.10 : Plot of conversion vs. time at various agitation speeds (sulfate media, stoichiometric runs, 25 °C) 81
Fig. 6.11 : Plot of conversion vs. time at various agitation speeds (chloride media, stoichiometric runs, 25 °C) 81
Fig. 6.12 : Plot of conversion vs. time at various temperatures (0.1 M H2S04, stoichiometric runs) 86
Fig. 6.13 : Plot of conversion vs. time at various temperatures (0.1 M HC1, stoichiometric runs) : 86
Fig. 6.14 : Plot of product layer model fitting of conversion vs. time at various
temperatures (0.1 M H2S04, stoichiometric runs) 87
Fig. 6.15 : Plot of product layer model fitting of conversion vs. time at various temperatures (0.1 M HC1, stoichiometric runs) 87
Fig. 6.16 : Plot of reaction rates vs. inverse of temperature (Arrhenius plot) for sulfate
and chloride media 88
Fig. 6.17 : Plot of In (k/T) vs. inverse of temperature for sulfate and chloride media 88
Fig. 6.18 : Plot of conversion vs. time at various particle sizes (sulfate media, stoichiometric runs, 25 °C) 95 Fig. 6.19 : Plot of conversion vs. time at various particle sizes (chloride media, stoichiometric runs, 25 °C) 95
Fig. 6.20 : Plot of product layer model fitting of conversion vs. time at various particle sizes (sulfate media, stoichiometric runs, 25 °C) 96
Fig. 6.21 : Plot of product layer model fitting of conversion vs. time at various particle sizes (chloride media, stoichiometric runs, 25 °C) 96
Fig. 6.22 : Plot of reaction rates vs. inverse square of chalcopyrite mean particle diameter (sulfate and chloride media, stoichiometric runs, 25 °C) 98
Fig. 6.23 : Plot of conversion vs. time at various sulfuric acid concentrations (constant
CuFeS2 and Fe additions, 25 °C) 102
Fig. 6.24 : Plot of conversion vs. time at various hydrochloric acid concentrations
(constant CuFeS2 and Fe additions, 25 °C) 102
Fig. 6.25 : Plot of product layer model fitting of conversion vs. time at various sulfuric
acid concentrations (constant CuFeS2 and Fe additions, 25 °C) 105
x Fig. 6.26 : Plot of product layer model fitting of conversion vs. time at various
hydrochloric acid concentrations (constant CuFeS2 and Fe additions, 25 °C) 105
Fig. 6.27 : Plot of log k vs. pH (sulfate media, constant CuFeS2 and Fe additions, 25 °C) 106
Fig. 6.28 : Plot of log k vs. pH (chloride media, constant CuFeS2 and Fe additions, 25 °C) 106
Fig. 6.29 : Plot of conversion vs. time at various metallic iron additions (sulfate media,
constant CuFeS2 and H2S04 additions, 25 °C) Ill Fig. 6.30 : Plot of conversion vs. time at various metallic iron additions (chloride media,
constant CuFeS2 and HC1 additions, 25 °C) Ill
Fig. 6.31 : Plot of conversion vs. time at various chalcopyrite additions (sulfate media,
constant H2S04 and Fe additions, 25 °C) 116
Fig. 6.32 : Plot of conversion vs. time at various chalcopyrite additions (chloride media, constant HC1 and Fe additions, 25 °C) 116
Fig. 6.33 : Plot of iron released in solution vs. time at various chalcopyrite additions
(sulfate media, constant H2S04 and Fe additions, 25 °C) 118
Fig. 6.34 : Plot of iron released in solution vs. time at various chalcopyrite additions (chloride media, constant HC1 and Fe additions, 25 °C) 118
Fig. 6.35 : Plot of recorded final conversion vs. SPD (sulfate media, non-stoichiometric
metallic iron additions, constant CuFeS2 and H2S04 additions, 25 °C) 120
Fig. 6.36 : Plot of chalcopyrite leaching rates vs. temperature 122
Fig. 6.37 : Plot of chalcopyrite initial leaching rates vs. mean particle diameter 124
Fig. 6.38 : Schematic representation of the galvanic conversion of chalcopyrite using
metallic iron as a reductant (acidic media) 127
Fig. 6.39 : Sequential schematic representation of the leaching stages 132
Fig. 6.40 : SEM photograph of the fresh chalcopyrite concentrate (-325 mesh +400 mesh) 135
Fig. 6.41 : SEM photograph for the leached concentrate (-100 mesh +200 mesh) 135
Fig. 6.42 : SEM photograph for the leached concentrate (-270 mesh +325 mesh) 136
Fig. 6.43 : SEM photograph for the leached concentrate (-325 mesh +400 mesh) 136 xi Fig. 6.44 : SEM photograph for a polished section of the leached concentrate in Fig. 6.42 137
Fig. 6.45 : Back-scattered electron image for the same section in Fig. 6.44 137
Fig. 6.46 : Proposed flowsheet for processing chalcopyrite concentrates by the method of reductive leaching with metallic iron 140
Fig. 6.47 : Plot of conversion vs. chalcopyrite mean particle diameter (high SPD, 3 M
FeCl2.4H20 solution, 25 °C)..' 146
Fig. 6.48 : Plot of iron content in the enriched concentrate vs. chalcopyrite mean particle
diameter (high SPD, 3 M FeCl2.4H20 solution, 25 °C) 147
Fig. 6.49 : Plot of copper content in the enriched concentrate vs. chalcopyrite mean
particle diameter (high SPD, 3 M FeCl2.4H20 solution, 25 °C) 147
xii ACKNOWLEDGMENT
I would like to present thankfulness and express sincere respect and appreciation to my supervisor, Dr. David Bruce Dreisinger. His assistance and valuable suggestions together with thoughtful supervision and constructive discussion lead ultimately to the successful completion of this research. Dr. Dreisinger was also, for me, a very dear friend, who provided every moral support expected from courteous persons. It is the duty of this author and everyone who read this thesis or use it to thank Dr. Dreisinger for all what he offered and make supplication to our Lord to be always with him, wishing for him and his family a good health and prosperous life. May
God bless you all, and protect you from every wicked action.
Next I would like to thank the discussion committee for their interest in this work.
Thanks also to all my coworkers in the Hydrometallurgy group, and very special thanks to Dr.
Berend Wassink who helped me a lot in the chemical analysis and experimental set-up, and provided training and valuable suggestions for the completion of this work. Dr. Wassink was very patient and constructive for the huge number of inquiries overwhelmed him, which indeed reflects originality and politeness. He also helped in reviewing this work. Special thanks to the secretaries of this department, particularly to Ms. Joan Kitchen. The financial support for this research through an NSERC scholarship is highly appreciated.
I am very grateful to my dear friends Mr. Jameel Al-Sakaji (Abu Ayman) and Mr.
Mohammad Adili (Abu Habeeb), who provided every required financial and moral support for the completion of this work.
Last, but not least, I would like to thank my mother Haleema Mohammad Taha Al-
Amour, for her efforts and moral support that assisted in finishing this research. The supplications and prays she made to Allah, day and night, to help me were an essential part toward this achievement. I am totally indebted to you, Mama Haleema. "Thank you. Mom".
xiii 1. INTRODUCTION
The hydrometallurgical treatment of copper minerals is increasingly gaining attention in research and development, and establishing itself as a viable route of treatment against other practiced methods, like pyrometallurgy. It is expected that by the eve of the new millennium about 25% of world copper production will be from hydrometallurgical routes (Champagne
(1995)).
The extractive metallurgy of copper and its minerals is well studied and documented in the open literature. There are many textbooks and conference proceedings devoted to all aspects of this topic (Cooper et al (1995 and 1991), Biswas and Davenport (1994), and Yannopoulos and
Agarwal (1976)).
Copper sulfides, for instance, are upgraded by froth flotation from less than 0.7% copper in the feed to a concentrate containing around 30% copper (by mass). In general, such a concentrate is a mixture of the sulfides of copper, copper-iron, and iron with a smaller amount of gangue minerals (like silica). Depending on the mineral copper content, these concentrates can further be processed by different methods, viz. pyrometallurgical or hydrometallurgical processes.
Pyrometallurgical processes for copper recovery from sulfide ores are called conversion, while those from oxide ores are called reduction. Conversion or smelting designates the operations of melting the concentrate and extracting the copper by heat, flux and the addition of
oxygen. Sulfur is removed as S02 by the reaction at high temperatures in the presence of air and/or oxygen. It is later converted to sulfuric acid. Iron combines with silica either from the gangue or the added flux, and is removed together with some other impurities in the resulting slag. Precious and trace metals (such as selenium and tellurium) remain with the copper to be recovered as by-products in the final purification stage, namely, electrorefining.
Hydrometallurgical processing of copper concentrates has long been suggested as an alternative approach to avoid the environmental problems intrinsic to smelting.
Hydrometallurgical processes involve the direct dissolution of copper minerals in aqueous solutions with the subsequent recovery by chemical or electrochemical methods. In some cases
(Dreisinger (1997)), the sulfides are converted to a richer form by rejection of iron and sulfur, while retaining copper in the solid phase, producing a more amenable sulfide.
1 Hydrometallurgical processes can compete with pyrometallurgical processes on the basis of:
1) Potentially high metal recovery (at least as high as pyrometallurgical processes)
2) Suitability to treat complex and very low grades of copper and other metal ores
3) Suitability for direct processing of sulfide concentrates
4) Providing an advantage in by-product recovery and recycle (value-added metals)
The remaining fields of competition are innovative solutions to sulfur and iron problems. The fixation of sulfur in elemental form is a clear advantage from all points of view, while finding a suitable solution to iron control in a form other than the slag would also be attractive.
Many hydrometallurgical processes have been patented, but very few have achieved commercial application or demonstration plant operation because of the relatively high costs and some problems associated with waste disposal. However, the advancement in separation science and process technology should mitigate these difficulties and lead to some suitable solutions.
According to Peters et al (1981), for any copper hydrometallurgical process to prove itself, it should meet the following objectives :
1) High copper recoveries; at least as high as conventional smelting (> 98%)
2) High conversion of sulfidic sulfur to elemental sulfur
3) Separation of iron as a marketable product
4) Recovery of precious and minor metals, e.g.: gold, silver, PGM's, tellurium and selenium
5) Production of refined copper (without separate refining)
6) Complete closure of aqueous circuits, with control of purged materials and residues as
innocuous solids
7) Low energy consumption, by avoidance of any electrochemical steps, or electrowinning at
lower valency species
8) Selection of satisfactory and affordable material of construction
The main issues to be addressed in hydrometallurgy (specifically in a leach-solvent extraction- electrowinning process) are : selectivity, lixiviant regeneration and stream recycling.
When hydrometallurgical methods are used, copper minerals are characterized in two
groups (Table 1.1) : Principal oxides, such as cuprite (Cu20) and tenorite (CuO), and principal
sulfides, such as chalcopyrite (CuFeS2) and chalcocite (Cu2S). Copper extraction from the oxides
2 requires a lixiviant, while that from the sulfides requires a lixiviant and an oxidant, like oxygen or air.
Mineral Composition Copper Crystal Moh's Specific gravity % system hardness Primary sulfides
Bornite Cu5FeS4 63.3 Isometric 3 5.06-5.08
Chalcopyrite CuFeS2 34.5 Tetragonal 3.5-4.0 4.1-4.3
Tennantite Cuj2As4Sj3 51.6 Isometric 3.0-4.5 4.37.4.49
Tetrahedrite Cu12Sb4S13 45.8 Isometric 3.0-4.5 4.6 Secondary sulfides
Chalcocite Cu2S 79.8 Orthorhombic 2.5-3.0 5.5-5.8 Covellite CuS 66.4 Hexagonal 1.5-2.0 4.6-4.76 Oxides
Azurite 2CuC03.Cu(OH)3 55.1 Monoclinic 3.5-4.0 3.77-3.89
Atacamite Cu2Cl(OH)3 59.5 Orthorhombic 3.0-3.5 3.76-3.78
Brochantite Cu4S04(OH)6 56.2 Monoclinic 3.5-4.0 3.9
Chrysocolla CuSi03.2H20 36.0 Orthorhombic 2.4 2.0-2.4
Cuprite Cu20 88.8 Isometric 3.5-4.0 6.14
Malachite CuC03.Cu(OH)2 57.3 Monoclinic 3.5-4.0 3.9-4.03 Tenorite CuO 79.9 Monoclinic 3.5 3.8-4.03 Native copper Cu 100.0 Isometric 2.5-3.0 8.95
Table 1.1 : Common copper minerals (George (1992))
Chalcopyrite (CuFeS2) is the major copper-bearing sulfide mineral, and the world's most abundant copper mineral. Its color is frequently brass yellow, often tarnished and slightly iridescent. The molecular weight is 183.513 g/mol. It has a specific gravity of 4.1 to 4.3, hardness 3.5 to 4.0 on the Moh scale, and microhardness 199 to 245 kg/mm2, with very low solubility in water. Chalcopyrite can be dissolved using strong oxidizing agents, like concentrated sulfuric acid, or in bromine-water solutions. It can be found alone or in complex
ores with other minerals, like pyrite (FeS2) and sphalerite (ZnS).
Although chalcopyrite is usually written as CuFeS2, a better representation is Cu2S.Fe2S3, reflecting the fact that the ionic structure of its elements is as follows : Copper (with normal electronic configuration : [ls22s22p63s23p63d104s1]) in the cuprous state, iron (with normal electronic configuration : [ls22s22p63s23p63d64s2]) in the ferric state and sulfur (with normal
2 2 6 2 4 + 3+ 2 + electronic configuration : [ls 2s 2p 3s 3p ]) as negative S(II) state, giving Cu Fe (S ")2. The Cu state of copper is indicated by the lack of a magnetic moment (resulted from the mineral's
3 antiferromagnetic structure), and the Fe3+ state of iron is indicated by the small isomer shift in the
Mossbaur spectrum. Hence, the electron configuration of copper outer shell is [3d10] and that of iron is [3d5], leaving that of sulfur as [3p6] (Habashi (1978) and Donnay et al (1958)).
The structure of chalcopyrite is tetragonal. According to Hall and Stewart (1973), when alternate atoms in the diamond structure for carbon are replaced by Zn and S atoms, the isometric structure of sphalerite is obtained, in which every Zn atom is surrounded tetrahedrally by four S atoms, and every S atom in the same way is surrounded by four Zn atoms. If alternate Cu and Fe atoms replace Zn atoms, the tetragonal structure of chalcopyrite is now obtained. The atoms at the corner of the original ZnS unit cell are not all of the same kind, because of substitutions for
Zn by Cu and Fe. The resulting unit cell for chalcopyrite (Fig. 1.1) is therefore twice as large as that of ZnS. The sulfur atoms are in distorted cubic close-packing with metallic atoms in half the tetrahedral interstices. Above about 540 °C, the metal atoms disorder, and the isometric
sphalerite structure is recovered.
In terms of conductor type (Table 1.2) chalcopyrite is an n-type semiconductor (high electron conductivity, conduction via electrons), due to metal excess, with a fairly narrow range of resistivity near 0.02 ohm cm (equivalent to specific conductance of 50 S/cm). Chalcocite, compared to chalcopyrite, is a p-type semiconductor (conduction via holes) and has an orthorhombic crystal structure, with specific conductance of 125 S/cm. Chalcopyrite has a band gap of about 0.6 eV while that for chalcocite is about 1.1 eV. A review of the electrochemistry of chalcopyrite can be found in Hiskey (1993).
Unfortunately, chalcopyrite contains only 34.64 % copper and about one third sulfur (all
by mass). Chalcopyrite is closely related to bornite (Cu5FeS4), idaite (Cu5FeS6) and cubanite
(CuFe2S3). With the exception of cubanite, chalcopyrite contains more sulfur per unit of copper than any other sulfide mineral, hence seeking for a process to enrich chalcopyrite is attractive.
In addition, chalcopyrite is the most refractory, or hard to leach, among copper sulfide minerals. Even under conditions of elevated temperature and pressure, the rates of dissolution are relatively slow. In the past decades, there have been many contributions to the study of chalcopyrite leaching. The various studies in the open literature have demonstrated that it is not very responsive to chemical attack.
4 • Iron O Copper O Sulfur
System Tetragonal Space Group 122: Dj\ I42d Axial elements a:c = 1:1.9716 a=5.28A c=10.4lA Cell Content 4rCuFeS,1 CuFeS? per cm3 1.3 X 1022 Fe-S 2.25A Cu-S 2.30A S-S 3.68A. 3.80A Common faces and Angles roonArii2^ = 54° 21* (TOOWlOn = 26° 54* roonAno2) = 44° 36' fll2)A(T12) =109° 50'
Fig. 1.1 : The crystal structure of chalcopyrite (Hall and Stewart (1973))
5 Formula Name Resistivity, Usual Structure Ionic Structure Ohm - m Conductor Type IO"3 - 10-6 Tetragonal Cu5FeS4 Bornite P (Cu+)5 Fe3+ (S2-)4
4X10-2-8X10-5 Orthorhombic (Cu+)2S2- . Cu2S Chalcocite P
CuFeS2 Chalcopyrite 2X10-4-9X10-3 n Tetragonal Cu+ Fe3+ (S2-)2
CuS Covellite 8X10-5. 7X10-7 Metallic Hexagonal (Cu+)2(S2)2- PbS Galena IXIO-5 - 7X10-6 n and p Cubic Pb2+ S2- M0S2 Molybdenite 7.5-8X10-3 n and p Hexagonal Mo4+ S2-
3X10-2- 1X10-3 Fe2+ (S2)2- FeS2 Pyrite n and p Cubic ZnS Sphalerite 3X10-3 - 1X10-4 (Insulator) Cubic Zn2+ S2- 2+ 2+ 4- (Zn,Fe)S Marmatite N/A P Cubic Zn Fe S 2 lX10 -lX10-2 Sn4+(02)4- Sn02 Cassiterite n Tetragonal
10H-10 Cubic (Cu+)2 02- Cu20 Cuprite P
Fe203 Hematite 2.5X10-1-4X10-2 n and p Trigonal (Fe3+)2 (02-)3
Fe304 magnetite 2X10-4-4X10-5 n and p Cubic Fe3+ Mn0 Pyrolusite IO"1 - 10-3 n Tetragonal 2 2 Mn4*(d^-)2 "
Ti02 Rutile 104-10 n and p Tetragonal Ti4+(022-)2 Uraninite 20-4X10"1 N/A Cubic
Table 1.2 : Electronic and structural properties of selected sulfide and oxide minerals (Hiskey
and Wadsworth (1981)). N/A : not available.
Converting chalcopyrite to a more amenable form could solve this and other associated problems, and allow a better recovery of copper, and the accompanying precious metal values.
Such a conversion may be done in a single process, or be included in a total hydrometallurgical process. This method is called the reductive leaching or decomposition of chalcopyrite, that is
leaching chalcopyrite under reducing conditions or in the presence of a suitable reductant. This method might also be called cathodic conversion, the alteration or enrichment of chalcopyrite concentrates.
In the published literature, there are different studies on achieving this enrichment.
However, many of these studies were qualitative in nature, and few of them were concerned with the kinetics of reductive leaching or process development. Hence, it is necessary to study and establish the conditions under which such an option would be best suited for the treatment of chalcopyrite concentrates and, consequently, be considered as a viable route.
This research is concerned with studying the fundamental aspects of converting chalcopyrite to a richer copper sulfide on the basis of removing some of its iron and sulfur components. The basis of this reductive conversion is leaching a chalcopyrite concentrate in the
6 presence of the reductant : metallic iron. The objective of this research is to understand the physical chemistry (thermodynamics and kinetics) of such an option and determine, experimentally, the best leaching conditions, in terms of thermodynamic and chemical variables, that would eventually convert the refractory chalcopyrite to an amenable copper sulfide. To do so, a leaching mechanism was proposed, followed by a systematic physicochemical analysis that revealed the • conditions, or the combination of conditions, which would achieve the main objective.
The other objective of this research is to develop a simple process flowsheet for the reductive decomposition of chalcopyrite, that might find commercial application, while, also, searching for a suitable solution to mitigate any associated environmental problems (namely, the rejection of iron compounds and hydrogen sulfide).
This work will first present a literature survey of chalcopyrite leaching to demonstrate the advantages and disadvantages of different options, before discussing the proposed leaching mechanism for the studied method of reductive leaching. Then, a description of the experimental work and procedures that were followed throughout this research is given. The results of the systematic kinetic analysis are discussed in detail, before presenting the proposed process flowsheet and its viability. The last sections of this work include the final conclusions, recommendations for future research and bibliography. The various leaching models and other related topics of kinetic analysis, as well as the detailed chemical analysis, all are given in the
Appendices.
7 2. LITERATURE SURVEY
2.1 THERMODYNAMICS OF CHALCOPYRITE LEACHING
Before reviewing the published literature on chalcopyrite leaching, it is necessary to give a brief discussion on the aqueous chemistry and thermodynamics of copper and chalcopyrite under reducing conditions, since a good understanding of the basic principles of hydrometallurgical treatment is important for improving or developing new processes. Once the thermodynamic analysis of leaching is understood, the remaining issue to be addressed is an evaluation of three main factors, which are :
1) The conditions under which the system will go from the initial to the final state,
2) The mechanism or path(s) the reaction(s) would likely follow, and
3) The rate at which such reactions will proceed, or simply, a detailed kinetic study. The topics related to the kinetic analysis are given in
Appendix I. The reader is directed to refer to them, as appropriate, during the course of this thesis.
Table 2.1 summarizes the physical and chemical properties of copper and its group members (silver and gold). Although the ground state electronic configuration of copper
(ls22s22p63s23p63d104s1) implies a stable closed shell in the third energy level, i. e. 18 electrons or inert gas shell, the shell is not inert. Rather, the underlying d orbitals appear to participate in metallic bonding by promotion of at least one d electron into a higher energy orbital of the outermost principal quantum level (George (1992)). There, this electron is available for participation in electrical and thermal conduction, as well as chemical activity. Moreover, copper is capable of losing the single 4s electron, forming the copper (I) ion, or several electrons from the 3d level, by hybridization, leading to the formation of other oxidation states (+2, or the very rare +3) and the ability to form complex ions with several ligands (such as ammonia, cyanide, and others).
This unique nature of the electronic configuration of copper also provides chemical properties intermediate between transition and 18 electron-shell elements. The high ionization energy and small ionic radius of copper contribute to its forming oxides much less polar, less stable, and less basic than those of the alkali metals. This relative instability of its oxide is consistent with its occurrence in nature in the metallic form.
8 The standard reduction potentials of cuprous and cupric ions at 25 °C are :
+ Cu (aq) + e ^Cu(s) E° = +0.520 V (SHE) (1)
2+ Cu (aq) + 2e <-» Cu(s) E° = +0.337 V (SHE) (2)
Obviously, the cuprous ion is less stable than the cupric ion, and the same is likely to apply for their compounds or complex ions. As a note, the value shown in this research for E° is the standard reduction potential using the IUPAC convention. The symbol SHE stands for standard hydrogen electrode and all quoted potential values in this research are with reference to this electrode.
Property Copper Silver Gold Atomic number 29 47 79 Atomic weight 63.54 107.87 196.97 Oxidation states 1,2,3 1,2,3 1,2,3 Electronic configuration [Arpd'^s1 [Kr]4d105s1 [Xe]4f145d106s1 Standard reduction potential, E°, Cu2+/Cu = 0.337 Ag7Ag = 0.799 Au7Au= 1.692 V (SHE) Density, kg/m3 8960 10490 19320 First ionization energy, kJ/mol 745 732 891 Electronegativity 2.43 2.30 2.88 Metallic radius, nm 0.1276 0.1442 0.1439 Ionic radius, M+, nm 0.096 0.126 0.137 Covalent radius, nm 0.138 0.153 0.150 Thermal conductivity, W/m per K 394 427 289 Electrical resistivity at 20 °C, p,Q/cm 1.673 1.59 2.35 Linear coefficient of thermal 16.5 10.68 14.2 expansion X 106 per °C, at 20 °C Melting point, °C 1083 960.8 1063 Boiling point, °C 2595 2212 2970 Specific heat at 20 °C, J/kg per °C 384 233 131 Tensile strength (annealed metal), 230 280 170 MPa Modulus of elasticity, MPa 10.2-12 X 104 7.75 X 104 7.85 X 104 Crystal structure fee fee fee
Table 2.1 : Summary of physical and chemical properties of Group 11 (IB) metals (copper,
silver and gold), compiled from George (1992)
9 Species Solubility in cold Solubility in hot Comments water, g per 100 cm"3 water, g per 100 cm-3 Copper Insoluble CuCl 0.0062 N/A
CuCl2 70.6 at 0 °C 107.9 at 100 °C
2+ 2_ Cu2S04 Decomposes to Cu and S04
CuS04 14.3 at 0 °C 75.4 at 100 °C
14 Cu2S 1 X IO" N/A No temperature quoted CuS 3.3 XlO"5 at 18 °C N/A
5 CuFeS2 4.32 X 10 " at 110 degree Celsius Iron Insoluble
FeCl2. 64.4 at 10 °C 105.7 at 100 °C
FeCl2.4H20 160.1 at 10 °C 415.5 at 100 °C
Fe(OH)2 0.00015 at 18 °C N/A .
Fe(OH)3 Insoluble. Soluble in HC1
FeS04 Fairly soluble Fairly soluble No quoted figures
Fe2(S04)3 Fairly soluble Decomposes in hot water
4 FeS2 4.9 X IO" N/A No temperature quoted FeS 6.2 X 10"4 N/A No temperature quoted. Decomposes in hot water •s Insoluble. Soluble only in organic solvents, like alcohols
and CS2 or aqueous solutions containing sulfide ions, leading to the formation of polysulfides
H2S04 Infinite solubility with heat evolution HC1 82.3atO°C 56.1 at60°C
3 -3 H2S 437 at 0 °C 186 at 40 °C Figures are cm per 100 cm of water Pb Insoluble
PbCl2 Highly soluble due to formation of complex ions PbS 8.6 X 10"5at 18 °C N/A
3 3 PbS04 4.25 X 10" at25 °C 5.6 X 10" at40 °C
Table 2.2 : Selected solubility data for copper and related species (Weast (1976)). N/A :
not available.
10 There is a very strong tendency for the cuprous ion to disproportionate in aqueous solutions into cupric ion and metallic copper :
1 + 2+ 2Cu (aq)-Cu(s) + Cu (aq) K3=V (3)
The corresponding free energy of change is -35.32 kJ/mol, enthalpy of change is 87.8 ± 5.0
kJ/mol, and log10 K3 = -5.76 ± 0.06, all at 25 °C, which explains the disappearance of cuprous ion from the thermodynamic stability diagram of the copper-water system (Fig. 2.1). As given in
Table 2.2, cuprous sulfate decomposes to cupric and sulfate ions in aqueous systems.
The cupric ion is more stable than the cuprous ion. Ligands that form strong coordination bonds bind copper (II) ions readily to form complexes, in which the copper has a coordination
2+ number of 4 or 6. Examples of these are the tetraammine copper (II) complex, [Cu(NH3)4] , and
2+ the tetraaquocopper (II) complex, [Cu(H20)4] . Formation of copper (II) complexes in aqueous solutions depends on the ability of the ligand to compete with water molecules for coordination sites.
Thermodynamics can also be used to predict the stability or instability of several species in various leaching systems under fixed conditions of temperature, pressure and activity. This is
usually represented by the so-called stability diagrams (Eh-pH diagrams), which show the relation between solution potential (or chemical potential) and acidity level. Such diagrams, being thermodynamic, can indicate the initial and final stable species in a system, but they do not give any kinetic information, e. g. : the rate of decomposition of an unstable mineral or phase, or the path that would be followed by such reactions. These diagrams are useful in explaining the leaching chemistry of copper and chalcopyrite.
Fig. 2.1 represents the thermodynamic stability diagram for the copper-water system at 25 degree Celsius. The equations used for constructing this diagram are the same as those developed by Pourbaix (1966), except that the thermodynamic values were compiled from Robie and
Hemingway (1995).
11 Cu-H20 System at 25 C
2 i i t 1 I i 1 I PLOT LABELS
tw • 298. IS K 1C"I «= I H 1,5
^ ^ D 1 "** ± ^ R Cu 3 Cu <2*> IflOl 0.5 C Cu2 0 0 Cu 0
HJO STflBitlTT LIMITS £ 0 * ' "' " ••ii 1 OXTCEH 1 2 tlYORGtSN 7 * » -0.5 R *"~ ~- — ^
•1
•-1.5I
IIIIII -2 2 0 2 H 6 8 10 12 14 1 PH
Fig. 2.1 : The Eh - pH diagram for the copper - water system at 298.15 K. The diagram was prepared using the Commonwealth Scientific and Industrial Research Organization (CSIRO) software and based on the methods developed by Dreisinger and Peters (1992).
From such a figure, it can readily be seen that all the oxidized forms of copper should be reduced to metallic copper at 25 °C under a hydrogen pressure of 1 atmosphere. Elemental copper is resistant to aerated alkaline solutions, except in the presence of ammonia. Copper does not displace hydrogen from acid but dissolves readily in oxidizing acids such as nitric acid or in acid solutions that contain an oxidizing agent, such as sulfuric acid solutions containing ferric sulfate.
From Fig. 2.1, it can also be seen that copper is immune (will not corrode or dissolve) and has a large stability (dominance) region in the absence of substances with which it can form complexes or soluble substances. Copper can be produced from aqueous acidic solutions by direct electrolytic reduction from the divalent state, or from cuprous oxide at pH greater than 3.5.
12 By judicious cathodic polarization and according to Fig. 2.1, copper can be protected
from dissolution by bringing its potential (Eh) to below +0.1 V in acid solutions (for pH < 4), and below about -0.6 V in neutral or alkaline solutions, depending on the pH. At high pH values, the
2 cuprite and bicuprite ions (Cu02 " and HCu02") are found (but not shown on Fig. 2.1).
Table 2.3 gives some thermodynamic data for several species in copper aqueous chemistry. For a review on copper aqueous chemistry and related thermodynamic properties, the reader is referred to Plyasunova et al (1997), Dreisinger and Peters (1992) and Senanayake and
Muir(1988).
Species AGf°, kJ per mol AHf°, kJ per mol S°, J per mol per K
CU(Crystamne) 0 0 33.15+0.08
CuCl(crystal|jne) -121.16 + 0.73 -138.07+ 1.70 88.0+ 6.20 (aq) -131.22 ± 0.12 -167.08+ 0.10 56.6+ 0.20
Cu (aq) 48.99 + 0.24 76.35 ± 3.60 59.6+ 12.1
^U (aq) 65.10 + 0.05 64.9+ 1.00 -98+ 4.00
CuCl2-(aq) -245.88 + 0.84 -277.82+ 1.75 214.5+ 6.50
2 CuCl3 -(aQ) -373.41 ± 1.02 -457.88+ 1.76 214.9+ 7.20 -69.77 ± 0.36 -93.48+ 1.23 -0.2+ 4.300 cucr(aa)
CuCl2(neutral) -200.76 ± 2.12 -246.3+ 6.10 103.6+21.7
CuCl3-(aq) -315.05 -378.65 221.75
CuCl4 (aq) -433.46 -521.75 310.5
+ FeCl (aq) -210.87 -246.43 -46.03
2+ FeCl (aq) -143.93 -192.46 -154.808 Reaction logic K AHf°, kJ per mol
Cu(crystalline) + Cu (aq)
CuCl(crystalMne) Cu (aq) + Cl (aq) -6.82+ 0.12 47.3 ± 4.0
+ Cu (aq) + 2Cl-(aq) ^CuCl2-(aq) 5.68 ± 0.14 -20.0 ± 4.0
+ Cu (aa) + 3Cl-(aq) ~CuCl3\a) 5.02+ 0.12 -33.0 + 4.0
Cu (aq) + Cl (aq)
Cu (aq) + 2C1 (aq)
Table 2.3 : Thermodynamic values for some common species and reactions in copper aqueous chemistry (Wang et al (1997) and Bard et al (1985))
The stability diagrams can also be extended to chalcopyrite. Peters (1992, 1984 and 1976) has published several stability diagrams for copper and its sulfides, and, in fact, pioneered this
field. Fig. 2.2 represents the Eh-pH diagram for the copper-sulfur-water system, and Fig. 2.3
represents the Eh-pH diagram for the copper-iron-sulfur-water system, all at 298.15 K. The
13 equations used for constructing the Eh-pH diagrams are numerous (91 equations) and can be found in Table 2.4. The thermodynamic data for all the considered species in this research were taken from Wang et al (1997), Robie and Hemingway (1995), Pankratz et al (1987), Bard et al
(1985) and Weast (1976), and are given in Tables 2.5 through 2.7.
CU-S-H2Q System ot 25 C
PLOT LABELS 2. T "T T~ 1— 1 1 TW - 298.15 K ICul = 1 M 1.5 ISl' = 1 n -fi- _ STABLE FIBERS 1. -1 R Cu S B Cu C Cu2 5 • Cu <2*> IRQ I E MO F Cu 0
H20 STABILITY LiniTS 1 OXTCEN 2 HTOROGEN
Fig. 2.2 : The Eh - pH diagram for the copper - sulfur - water system at 298.15 K. The diagram was prepared using the Commonwealth Scientific and Industrial Research Organization (CSIRO) software and based on the methods developed by Dreisinger and Peters (1992).
A concise discussion on chalcopyrite leaching thermodynamics is given here, and more information can be found in Peters (1976). It should be noted that Fig. 2.3 was plotted assuming that, for laboratory conditions, unobserved chalcopyrite reactions are deleted, i. e. : hydrogen sulfide is not available as a reactant and/or pyrite can not form at a measurable rate; reactions involving these two species are not reversible.
14 Fig. 2.3 : The Eh- pH diagram for the copper-iron-sulfur water system at 298.15 K.
Species activities as indicated (Peters (1976)). Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3
Reaction no. Reaction and Equilibrium Equation
+ 1 2H (aq) + 2e~H2(g)
Eh = -0.0591 pH - 0.0295 log PHz
+ 2 02(g) + 4H (aq) + 4e-~2H20(1)
Eh = 1.229 - 0.0591 pH + 0.0148 log P02
3 HS (aq) + H (aq) H2S(aq)
pH = 7.00 + log
a H2S
+ 4 S (aq) H (aq) HS (aq)
pH= 13.99 +log
aHS"
S + 5 (S) + 2H (aq) + 2e" *-* H2S(aq)
Eh = 0.142 - 0.0591 pH - 0.0295 log aH2§
s H+ 6 (s) + (aq) + 2e" <-» HS"(aq)
Eh = -0.065 - 0.0295 pH - 0.0295 log aHS_
+ 7 H2S04(aq) + 6H (aq) + 6e- <- S(s) + 4H20(1)
Eh = 0.375 - 0.0591 pH + 0.00985 log aH2SOj
+ 8 HS04-(aq) + 7H (aq) + 6e «- S(I) + 4H20(1)
Eh = 0.338 - 0.0689pH + 0.00985 log aHSO_
2 + 9 S04 -(aq) + 8H (aq) + 6e" <- S(s) + 4H20(1)
Eh = 0.357 - 0.0788 pH + 0.00985 log a0„ 2_
+ 10 SO/(aq) + 9H (aq) + 8e «- HS"(aq) + 4H20(1)
a 2
Eh = 0.252 - 0.0665 pH + 0.00739 log
a HS04-
11 SO/(aq) + 8Ff(aq) + 8e «- S\aq) + 4H20(1)
a 2
Eh = 0.148 - 0.0591 pH + 0.00739 log
a§2.
+ 12 HS04"(aq) + H (aq) H2S04(aq)
pH = -1.91 + log
a H2S04
S( 13 H+ V"(aq) + (aq) ^ HS04"(aq)
pH =1.91+log
a HS04"
2+ 14 Fe (aq) + 2e-^Fe(s)
Eh = -0.440 + 0.0295 log aFe2+
16 Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3 (continued)
15 re (aq)^ e ^ re (aq)
Eh = 0.771 +0.0591 log ^ aFe2+
2+ 16 FeS(s) + 2FT(aq) <-> Fe (aq) + H2S(aq)
pH= 1.29-0.5 log aFe2+ aHiS
+ 17 FeS(s) + 2H (aq) + 2e" <- Fe(s) + H2S(aq)
Eh = -0.0364 - 0.0591 pH - 0.0295 log aH2S
18 FeS(s) + FT(aq) + 2e" «-> Fe(s) + HS"(aq)
Eh = -0.571 - 0.0295 pH - 0.0295 log aHg.
2 19 FeS(s) + 2e^ Fe(s) + S "(aq)
Eh = -0.985 - 0.0295 log ag2_
20 Fe304(s) + 8FT(aq) + 8e" ~ 3Fe(s) + 4H20(1)
Eh = -0.085 -0.0591 pH
2+ 21 FeS2(s) + 4FT(aq) + 2e «-> Fe (aq) + 2H2S(aq)
2 Eh = -0.057 - 0.1182 pH - 0.0295 log afe2+ a ^
+ 22 FeS2(s) + 2H (aq) + 2e «- FeS(s) + H2S(aq)
Eh =-0.133 - 0.0591 pH - 0.0295 log aH2S
+ 23 FeS2(s) + H (aq) + 2e- <- FeS(s) + HS"(aq)
Eh = -0.340 - 0.0295 pH - 0.0295 log aHS_
2 24 FeS2(s) + 2e-^FeS(s) + S -(aq)
Eh = -0.754 - 0.0295 log as2.
2+ 25 Fe (aq) + 2S(S) + 2e" -> FeS2(s)
Eh = 0.340 + 0.0295 log aFg2+
2 26 Fe304(s) + 3S "(aq) + 8Ff(aq) + 2e" <- 3FeS(s) + 4H20(1)
Eh = 2.615 - 0.2364 pH + 0.0887 log as2.
+ 2 27 3FeS2(s) + 4H20(1) - 8H (aq)+ 4e ~ Fe304(s) + 6S "(aq)
Eh = -2.349 + 0.1182 pH + 0.0887 log ag2_
2+ + 28 Fe (aq) + 2HS04"(aq) + 14H (aq) <- FeS2(s) + 8H20(1)
2 Eh = 0.339 - 0.0591 pH + 0.0042 log ac 2+ a _
2+ 2 + 29 Fe (aq) + 2S04 -(aq) + 16H (aq) + 14e" <-> FeS2(s) + 8H20(1)
2 Eh = 0.355 - 0.0675 pH + log afe2+ a Q 2.
2 + 30 Fe203(s) + 4S04 -(aq) + 38H (aq) + 30e <- 2FeS2(s) + 19H20(1)
Eh = 0.380 - 0.0749 pH + 0.0078 log a'o2_
2 + 31 Fe304(s) + 6S04 -(aq) + 56H (aq) + 44e <- 3FeS2(s) + 28H20(I)
Eh = 0.383 - 0.0752 pH + 0.0080 log a* 2_
+ 32 3Fe203(s) + 2H (aq) 2e <-> 2Fe304(s) + H20(1)
Eh = 0.211 -0.0591 pH
17 Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3 (continued)
+ 33 Fe304(s) + 2H20(1) - H (aq) + 2e" «- 3HFe02"(aq)
Eh = -1.819 + 0.0295 pH - 0.0887 log aL n -
" r o HFe02
34 Fe203(s) + H20(1) + 2e" <- 2HFe02"(aq)
Eh = -1.139 -0.0591 log a^.
+ 2+ 35 Fe203(s) + 6H (aq) + 2e" <-> 2Fe (aq) + 3H20(1)
Eh = 0.728 - 0.1773 pH - 0.0591 log aFe2+
+ J+ 36 Fe203(s) + 6H (aq) «- 2Fe (aq) + 3H20(1)
pH =-0.26 - 0.33 log aFe3+
+ 37 Cu2S(s) + 2H (aq) + 2e"~ 2Cu Eh = -0.305 - 0.0591 pH - 0.0295 log aHzS + 38 Cu2S(s) + H (aq) + 2e" <- 2Cu Eh = -0.512 - 0.0295 pH - 0.0295 log aRS_ 2 39 Cu2S(s) + 2e-^2Cu(s) + S -(aq) Eh = -0.926 - 0.0295 log ag2_ 40 Cu^ + e^Cu^ Eh = 0.337 + 0.0295 log aCu2+ 2+ + 41 2Cu (aq) + H2S(aq) - 2H (aq) + 2e" <- Cu2S(s) 2 Eh = 0.978 + 0.0591 pH - 0.0295 log a u2+ aH2S + 42 2CuS(s) + 2H (aq) + 2e" «- Cu2S(s) + H2S(aq) . Eh = 0.081 - 0.0591 pH - 0.0295 log aH2S + 43 CuS(s) + 2H (aq) + 2e <- Cu(s) + H2S(aq) Eh = -0.142 - 0.0591 pH - 0.0295 log aHjS 44 2CuS(s) + FT(aq) + 2e" <- Cu^ + HS"(aq) Eh = -0.126 - 0.0295 pH - 0.0295 log aH§. + 2+ 45 CuS(s) + 2H (aq) Cu (aq) + H2S(aq) pH = -7.58 - 0.5 log aCu2+ a„2S 2+ + 46 CuS(s) + Fe (aq) + H2S(aq)- 2H (aq)~ CuFeS2(s) pH =-1.564-0.5 log aFe2+aH2S 2+ 47 CuS(s) + Fe (aq) + S(s) + 2e" <- CuFeS2(s) Eh = 0.234 + 0.0295 log afe2+ 48 + 2+ 5CuS(s) + 2H (aq) + Fe (aq) + 4e ~ Cu5FeS4(s) + H2S(aq) E = 0.137 - 0.0295 pH + 0.0148 log ape2+ - 0.0148 log aH2§ 2+ 49 Cu (aq)+S(s) + 2e<-CuS(s) Eh = 0.590 + 0.0295 log aCu2+ 2+ 50 Cu (aq) + H2S04(aq) + 6FT(aq) + 8e" <- CuS(s) + 4H20(1) Eh = 0.415 - 0.0443 pH + 0.0094 log aCu2+ aH2S04 18 Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3 (continued) 51 Cu2S(s) + H2S04(aq) + 6Ff(aq) + 6e <- 2CuS(s) + 4H20(1) Eh = 0.377 - 0.0591 pH + 0.0098 log a^ + 52 Cu2S(s) + HS04"(aq) + 7H (aq) + 6e ~ 2CuS(s) + 4H20(1) Eh = 0.359 - 0.0690 pH + 0.0098 log aHS(V 2 + 53 Cu2S(s) + S04 "(aq) + 8H (aq) + 6e <- 2CuS(s) + 4H20(1) Eh = 0.377 - 0.0788 pH + 0.0098 log a 2_ 2+ + 2 54 Cu2S(s) + 2Fe (aq) + 24H (aq) + 3S04 "(aq) + 22e" <- 2CuFeS2(s) + 12H20(1) Eh = 0.340 - 0.0645 pH + 0.0081 log a 2_ + 0.0054 log aFg2+ 2+ + 55 Cu2S(s) + 2Fe (aq) + 21H (aq) + 3HS04"(aq) + 22e" <- 2CuFeS2(s) + 12H20(1) Eh = 0.325 - 0.0564 pH + 0.0081 log aHCn . + 0.0054 log ac 2+ + 2 56 Cu2S(s) + Fe203(s) + 30 H (aq) + 3S04 -(aq) + 24e <- 2CuFeS2(s) + 15H20(1) Eh = 0.373 - 0.0739 pH + 0.0074 log ac„ 2. 57 5Cu2S(s) + 2FeS2(s) + 2Ff(aq) + 2e <- 2Cu5FeS4(s) + H2S(aq) Eh = -0.029 - 0.0591 pH - 0.0295 log aH2§ + 58 5Cu2S(s) + 2FeS2(s) + H (aq) + 2e" <- 2Cu5FeS4(s) + HS"(aq) Eh = -0.178 - 0.0295 pH - 0.0295 log aHS_ 2 59 30Cu2S(s) + 4Fe304(s) + 18S04 "(aq) + 176FT(aq) + 140e <- 12Cu5FeS4(s) + 88H2O0) Eh = 0.380 - 0.0742 pH + 0.0076 log a 2 + 60 5Cu2S(s) + Fe203(s) + 3S04 -(aq) + 30H (aq) + 24e <- 2Cu5FeS4(s) + 15H20(I) Eh = 0.376 - 0.0738 pH + 0.0074 log acn 2_ 0O4 2+ 61 2Cu (aq) + H2S04(aq) + 6Ff(aq) + 10e" <- Cu2S(s) + 4H20(I) 2 Eh = 0.438 - 0.0355 pH + 0.0059 log a u2+ aH2S04 2+ + 62 2Cu (aq) + HS04"(aq) + 7H (aq) + lOe" ~ Cu2S(s) + 4H20(1) 2 Eh = 0.427 - 0.0414 pH + 0.0059 log a 2+ auc„. CU 0.SO4 2+ + 63 2Cu (aq) + SO/(aq) + 8H (aq) + lOe <- Cu2S(s) + 4H20(1) 2 Eh = 0.438 - 0.0473 pH + 0.0059 log a 2+ asn 2. 2 + 64 2Cu(s) + S04 "(aq) + 8H (aq) + 6e ^ Cu2S(s) + 4H20(1) • Eh = 0.506 - 0.0788 pH - 0.0098 log a 2. + 65 Cu20(s) + 2H (aq) + 2e" <- 2Cu(s) + H20(1) Eh = 0.471-0.0591 pH + + 66 2CV (aq) + H20(1) - 2H (aq)+ 2e- «-> Cu20(s) Eh = 0.203 +0.0591 pH +0.0591 log a£u2+ + 67 2CuO(s) + 2H (aq) + 2e" <-* Cu20(s) + H20(1) Eh = 0.669 - 0.0591 pH 2 + 68 Cu02 "(aq) + 4H (aq) + 2e <- Cu(s) + 2H20(1) Eh= 1.515-0.1182 pH +0.0295 log an n 2 * o CuU2 19 Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3 (continued) 2 69 2Cu02 "(aq) + 6Ff(aq) + 2e <- Cu20(s) + 3H20(1) Eh = 2.560 - 0.1773 pH + 0.0591 log a 2. + 2+ 70 CuO(s) + 2H (aq) «-> Cu (aq) + H20(1) pH-3.94-0.5 log aCu2+ 2 71 Cu02 "(aq) + 2Ff(aq) <- CuO(s) + H20(1) pH= 15.98+ 0.5 log aCuo2. 72 2Cu5FeS4(s) + 2Ff(aq) + 2e" <- 5Cu2S(s) + 2FeS(s) + H2S(aq) Eh = -0.295 - 0.0591 pH - 0.0295 log aHiS + 73 2Cu5FeS4(s) + H (aq) + 2e" <- 5Cu2S(s) + 2FeS(s) + HS"(aq) Eh = -0.502 - 0.0295 pH - 0.0295 log aHg. 2 74 2Cu5FeS4(s) + 2e <-> 5Cu2S(s) + 2FeS(s) + S "(aq) Eh =-0.917-0.0295 log ag2. + 2+ 75 2Cu5FeS4(s) + 6H (aq) + 2e «- 5Cu2S(s) + Fe (aq) + 3H2S(aq) Eh = -0.143 - 0.1773 pH - 0.0591 log afe2+ - 0.0887 log aHzS + 2 76 6Cu5FeS4(s) + 8H20(1) - 16H (aq) + 2e- <- 15Cu2S(s) + 2Fe304(s) + 9S "(aq) Eh = -7.980 + 0.4728 pH - 0.2660 log ag2_ + 2+ 77 Cu5FeS4(s) + 8H (aq) + 6e" <- 5Cu(s) + Fe (aq) + 4H2S(aq) Eh = -0.278 - 0.0788 pH - 0.00985 log afe2+ - 0.0394 log aH2S + 78 Cu5FeS4(s) + 4FeS2(s) + 4H (aq) + 4e <- 5CuFeS2(s) + 2H2S(aq) Eh = -0.020 - 0.0591 pH - 0.0295 log aH2$ + 79 Cu5FeS4(s) + 4FeS2(s) + 2H (aq) + 4e"» 5CuFeS2(s) + 2HS"(aq) Eh = -0.227 - 0.0295 pH - 0.0295 log a^ 2 + 80 2Cu5FeS4(s) + S04 "(aq) + 8H (aq) + 6e ^ 5Cu2S(s) + 2FeS2(s) + 4H20(1) Eh = 0.395 - 0.0788 pH + 0.0098 log a 2_ 2 + 81 3Cu5FeS4(s) + 4Fe304(s) + 18S04 "(aq) + 176H (aq) + 140e ^ 15CuFeS2(s) + 88H20(1) Eh = 0.377 - 0.0742 pH + 0.0076 log acn 2_ 2 + 82 Cu5FeS4(s) + 2Fe203(s) + 6S04 "(aq) + 60H (aq) + 48e <-> 5CuFeS2(s) + 30H2O(l) Eh = 0.381 - 0.0738 pH + 0.0074 log aso 2. + 2+ 83 5CuFeS2(s) + 12H (aq) + 4e - Cu5FeS4(s) + 4Fe (aq) + 6H2S(aq) Eh = -0.254 - 0.1773 pH - 0.0591 log aFe2+ - 0.0887 log aH2S + 84 5CuFeS2(s) + 4H (aq) + 4e" ^ Cu5FeS4(s) + 4FeS(s) + 2H2S(aq) Eh = -0.246 - 0.0591 pH - 0.0295 log aH2S + 85 5CuFeS2(s) + 2H (aq) + 4e *-» Cu5FeS4(s) + 4FeS(s) + 2HS"(aq) Eh = -0.453 - 0.0295 pH - 0.0295 log aHS_ 2 86 5CuFeS2(s) + 4e - Cu5FeS4(s) + 4FeS(s) + 2S "(aq) Eh = -0.867 - 0.0295 log a$2_ 20 + 2 87 15CuFeS2(s) + 16H20(1) - 32H (aq) + 4e" <- 3Cu5FeS4(s) + 4Fe304(s) + 18S "(aq) Eh = -7.834 + 0.4728 pH - 0.2660 log ag2_ 2 + 88 5CuFeS2(s) + 2S04 "(aq) + 16H (aq) + 12e" <- Cu5FeS4(s) + 4FeS2(s) + 8H20(1) Eh = 0.410 - 0.0788 pH + 0.0098 log asn 2_ 89 CuFeS2(s) + 2FT(aq) + 2e" <- Cu,s) + FeS(s) + H2S(aq) Eh = -0.280 - 0.0591 pH - 0.0295 log aHjS 2+ 90 CuFeS2(s) + 4Ff(aq) + 2e «- Cu(s) + Fe (aq) + 2H2S(aq) Eh = -0.204 - 0.1182 pH - 0.0295 log aFg2+ - 0.0591 log aH2S 2+ 91 2CuFeS2(s) + 6FT(aq) + 2e <- Cu2S(s) + 2Fe (aq) + 3H2S(aq) Eh = -0.309 - 0.1775 pH - 0.0591 log ape2+ - 0.08875 log aH2S Table 2.4 : Reactions and Eh-pH relations used in constructing Figs. 2.1-2.3 (continued). The complete derivation of these equations and other related topics can be found in Dreisinger and Peters (1992) and Peters et al (1972). The equations were written according to the convention adopted by the International Union of Pure and Applied Chemistry (IUPAC). 2 05 2 Species Specific heat: Cp = A! + A2 T + A3 T" + A4 T + A5 T , J per mol per K 1 H20(1) 7.523X10 2 5 2 2 05 6 2 H2(8) 4.783 + 1.335X10" T - 5.617X10 T +4.5 83X10 T -1.825X10" T °w 56.58 - 5.255X10-3T + 6.85 6X105T-2 - 5.78X102T05 + 1.113X10"° T2 1 S(s) (ortho-crystal) 2.270X10 1 FeS(s) (troilite) 5.049X10 F 1 e7S8(S) (pyrrhotite) 4.988X10 2 6 2 3 05 FeS2(s) (pyrite) -2.032X10 + 5.030X10" T - 3.200X10 T" + 1.787X10 T" Fe 0 3 7 2 4 05 3 2 3 4(s) 2.6591X10 - 2.5215 T + 2.0734X10 r - 3.64 5 5X10 T + 1.3677X10" T 3 8 2 4 05 Fe203(s) -1.0957X10 + 2.7267X10"' T - 1.0239X10 T - 3.396X10 T CuFeS 2 1 7 2 4 05 2(s) -5.8753X10 + 3.7073X10" T -1.4721X10 r +1.275X10 T" 2 Cu5FeS4(s) 2.429X10 1 2 5 1 CuS(s) 4.304X10 + 2.023X10" - 1.399X10" + 4.358X10" 1 Cu2S(s) 7.684X10 6.084X10' - 2.875X10"2 T + 3.331X105T2 - 5.671X102 T05 + 1.420X10"5 T2 2 1 6 2 3 s 5 2 Cu20(s) 4.26X10 - 2.508X10" T - 4.898X10 T" -6.078X10 T°- + 9.244X10" T CuO 2 6 2 2 05 (s) 30.97 + 1.374X10" T - 1.25 8X10 r + 3.693X10 T 2 5 2 05 2 H2S(R) 26.360 + 2.650X10" T + 2.660X10 T - 43.560 T -6.024X10° T Table 2.5 : Selected heat capacity values for different species (Robie and Hemingway (1995)) 21 Species AG°f, kJ per mol AH°f, kJ per mol S°, J per mol per K H20(1) -237.19 -285.83 69.91 H2(g) 0.0 0.0 130.68 TT+ n 0.0 (aq) 0.0 0.0 o™ 0.0 0.0 205.15 0H"(aq) -157.3 -230.0 -10.7 H 2S(g) -33.4 -20.6 205.8 F^Sfaq) -27.86 -39.33 122.17 HS"(aq) 44.8 16.3 67.08 H2S04(aq) -690.0 -814.0 156.9 HS04"(aq) -752.87 -885.75 126.86 SO4 (aq) -744.0 -909.3 18.5 S(s) (monoclinic) 0.0 0.0 33.03 S(s) (orthorhombic) 0.0 0.0 32.05 02- 3 (aq) 85.8 33.1 -14.6 Fe(s) 0.0 0.0 27.154 2+ Fe ' -78.87 -87.86 -113.38 rC 3+(aq ) Fe -4.6 -47.69 -293.29 re (aq) FeS(s) (troilite) -101.3 -101.0 60.3 Fe7S8(s) (pyrrhotite) -98.9 -97.5 80.932 FeS2(s) (pyrite) -160.2 -171.5 52.9 Fe304(s) -1012.7 -1115.7 146.44 Fe203(s) -744.4 -826.2 87.4 a-FeO(OH)(s) -491.8 -562.6 60.4 HFe02\a) 379.18 CuFeS2(s) -195.1 -194.9 124.9 Cu5FeS4(s) -394.7 -371.6 398.5 CuS(s) -55.3 -54.6 67.4 Cu2S(s) -89.2 -83.9 116.2 Cu(s) 0.0 0.0 33.14 Cu(aq) 48.99 76.35 59.6 ^U (aq) 65.10 64.9 -98 Cu20(s) -147.8 -170.6 92.4 CuO(s) -128.3 -156.1 42.6 2 Cu02 "(aq) -182.0 -211.315 -98.324 HCu02;ag) -256.98 -244.505 41.84 0.0 Pb(S, 0.0 64.8 2+ Ph -24.2 0.9 18.5 Table 2.6 : Standard thermodynamic data for different species (Robie and Hemingway (1995) and Pankratz et al (1987)) 22 Physical property Value and comments Reference Diffusion coefficient 3.4 X 10;10 cm2 s"1 at 75 °C in chalcocite Etienne (1970) D Cu+ 5 0.446XlO" cmVat25°Cinl.4MH2S04 Opekar&Beran(1975) DC>- 4.17 X 10 10 m2 s"1 at 25 °C in acidic solutions Dry &Bryson (1987) DFe- 5.5 X 10 "10 m2 s"1 at 25 °C in acidic solutions Dry &Bryson (1987) 9.3 X 10 "5 cm2 s"1 at 25 °C in acidic solutions Opekar&Beran(1975) D H+ 5 2 1 D 0.81 XlO" cm s" at 25 °C in 0.4 M H S0 Opekar&Beran(1975) cr 2 4 Solubility product 47 KSD forC^S 2 XlO" at 16-18 °C Weast(1975) 48 K3DforCu2S 1.6 XlO" at25°C Emmons (1913) 45 Ksp for CuS 8.5 XlO" atl8°C Weast(1975) K^forFeS, 3.7 XlO"19 atl8°C Weast(1975) 19 KSDforFeS2 1X10" at25°C Emmons (1913) 14 K.>rFe(OH)2 1.64 XlO atl8°C Weast(1975) 30 KSDforFe(OH)3 1.1 XlO" atl8°C Weast(1975) 28 KSD for PbS 3.4 XlO" atl8°C Weast(1975) 8 Kjp for PbS04 1.06 XlO" atl8°C Weast(1975) Density 5 3 H2 gas 8.99X 10" gem" at25 °C Weast(1975) 5 3 H2S gas 153.9X10" gem" at 25 °C Weast(1975) Iron 7.86 gem"3at25 °C Weast(1975) Copper 8.96 g cm"3 at 25 °C Weast(1975) Sulfur 2.07 g cm"3 at 25 °C Weast(1975) 3 CuFeS2 4.1 gem" at 25 °C Weast(1975) 3 Cu2S 5.8 gem" at25 °C Weast(1975) 3 FeCl2.4H20 1.3 g cm" at 25 °C for a 3 M solution Analytical grade reagent 3 H2S04 (96% solution) 1.84 g cm" at 25 °C for an 18.3 M solution Analytical grade reagent HC1 (37% solution) 1.19 g cm"3 at 25 °C for a 32.6 M solution Analytical grade reagent Molecular weight H2 gas 2.0158 gmol"' 1 H2S gas 34.0758 g mol" Iron 55.847 g mol"1 Copper 63.546 g mol"1 Sulfur 32.06 gmol"1 1 CuFeS2 183.513 gmol" 1 Cu2S 159.158 gmol" 1 FeCl2.4H20 198.8138 gmol" Table 2.7 : Selected physical constants for copper and other species 23 Depending on the presence of oxidizing/reducing agents, solution composition and other prevailing thermodynamic conditions, chalcopyrite can react anodically or cathodically in aqueous media. According to the stability diagrams, there are four different kinds of solutions in which chalcopyrite might decompose : 1) Oxidizing solutions, and hence oxidative leaching, leading to the formation of elemental sulfur or the release of sulfate ion in solution, depending on the oxidizing potential and the selected pH. The general leaching reaction in this region can be written as : 2+ 2+ . CuFeS2(s) + 40x Cu (aq) + Fe (aq) + 2S(S) + 40x(aq) (4) Here Ox represents an oxidant (any suitable oxidant) and this reaction is the most commonly observed reaction in the leaching of chalcopyrite under laboratory conditions. The formation of elemental sulfur is due to the fact that sulfur is oxidized to sulfate with great difficulty (at least in acid solutions) by a very irreversible path. Under certain oxidizing conditions, a new solid phase might form in addition to soluble species. Some researchers have speculated that such a solid phase is merely a defect structure of chalcopyrite with the general formula : CuxFeySz and the values of x, y and z are such that they do not refer to chalcopyrite (Hackl et al (1995b)). 2) Strong acid solutions, leading to hydrogen sulfide evolution and dissolved copper and/or iron. Examples are : + 2+ ' 2CuFeS2(s) + 2H (aq) -> Cu2S(s) + Fe (aq) + FeS2(s) + H2S(g) (5) 2+ CuFeS2(s) + 2FT(aq) -> CuS(s) + Fe (aq) + H2S(g) (6) + 2+ 2+ . CuFeS2(s) + 4H (aq)^Cu (aq) + Fe (aq) + 2H2S(g) (7) According to Peters (1976), reaction 6 is more probable than reaction 5, since it does not require the nucleation of pyrite. Reaction 7 was reported by Warren (1958) at pH < 1. 3) Strong alkaline solutions, leading to sulfide ions in solution and copper/iron oxides, or metal sulfides. The stability diagram indicates that chalcopyrite is unstable at high pH values, decomposing, for instance, to chalcocite and magnetite, by an equation of the form : 24 2 2 2CuFeS2(s) + 60H-(aq) -> Cu2S(s) + | Fe304(s) + S "(aq) + 3H20(1) + ^ S04 "(aq) (8) The rate of this reaction is expected to be very slow. Peters (1976) has indicated that there is some evidence that magnetite and sulfide ions are formed when chalcopyrite is digested with NaOH solution at 200 °C. The strong acid and base paths are non-oxidative decomposition paths for chalcopyrite and are normally beyond the reach of available leach solutions. 4) Reducing solutions, leading to hydrogen sulfide evolution or sulfide ion formation and simpler sulfides or a metal phase. As an example, chalcopyrite in acidic reduction region would yield dissolved iron in solution and chalcocite or copper as solids, but going to lower potentials, both metallic copper and iron would tend to form. Using the symbol R to designate an unidentified reducing agent, the possible reactions are : + 2+ + - 5CuFeS2(s) + 12H (aq) + 4R ^ Cu5FeS4(s) + 4Fe (aq) + 6H2S(g) + 4R (aq) (9) 2+ + 2CuFeS2(s) + 6PT(aq) + 2R Cu2S(s) + 2Fe (aq) + 3H2S(g) + 2R (aq) (10) Eq. 10 is written assuming R is a monovalent reductant like sodium. When it is a divalent reductant like lead, copper or iron it will take the form : + 2+ 2+ 2CuFeS2(s) + 6H (aq) + R -> Cu2S(s) + 2Fe (aq) + 3H2S(g) + R (aq) (11) When elemental copper is formed, the reaction will be : + 2+ + CuFeS2(s) + 4H (aq) + 2R -> Cu(s) + Fe (aq) + 2H2S(g) + 2R (aq) (12) and when both elemental copper and iron are formed the reaction will be : + + CuFeS2(s) + 4H (aq) + 4R Cu(s) + Fe(s) + 2H2S(g) + 4R (aq) (13) According to the stability diagram for chalcopyrite (Fig. 2.3), in acidic solutions and under cathodic polarization (the presence of a reducing agent) chalcopyrite can be converted to chalcocite at potentials less than 0 and greater than -440 mV, for instance with iron, with a narrow stability region for bornite (Cu5FeS4). At low potentials and when the solution pH is greater than 2, the stability region for chalcocite becomes very small, raising the possibility to produce metallic copper directly. In slightly acidic solution or under alkaline conditions, with little increase in cathodic polarization, chalcopyrite can be converted to metallic copper, while 25 pyrrhotite or troilite will likely be produced. At much lower solution potentials (for example with powerful reducing agents like magnesium) both metallic copper and iron would tend to form. The exact form of the new solid phase under reducing conditions is dependent on reaction conditions, both on kinetic and thermodynamic considerations. The enriched copper sulfide is normally written as chalcocite (Cu2S). However, a better and more accurate representation is to write it as Cu^S, where x is ranging from 1 to 2. Hence, in addition to chalcocite, other solid phases might form as a final or intermediate product, like djurleite (Cu, 95S), digenite (Cu, 80S), anilite (Cu, 75S) and covellite (CuS). Hiskey and Wadsworth (1981) have discussed the conditions (especially solution potential) under which such phases might exist. In the presence of a reductant, chalcocite is more likely to form than the other phases due to a catalytic effect imposed by such reductants. Under reducing conditions, there is no stability region for covellite (CuS), as can be seen from Fig. 2.3, explaining thermodynamically why covellite is not reported as a reduction product in many studied systems. Hackl et al (1987) have also indicated that chalcocite is a more favorable phase for solid state diffusional processes than covellite. Reduction under acidic conditions is most preferable, but similar equations can be written for alkaline conditions that tend to stabilize the bisulfide (HS_) or sulfide (S2) ions, and troilite 2+ (FeS) instead of hydrogen sulfide (H2S) and ferrous ion (Fe ), respectively (see Table 2.4). It should be noted that reductive decomposition can be done at different temperatures and pressures. Beyond 100 °C two concerns exist: the probability of thermal precipitation of copper salts, if formed, and usage of pressurized operations (in autoclaves) as the boiling point of water is exceeded. The advantages of reduction in acidic conditions are apparent from Eqs. 9 to 13, and will be addressed later. Copper is retained in the solid phase, sulfur is eliminated as a gas, and iron, the problematic species in hydrometallurgy, is dissolved and rejected alone in the solution. Under reducing and non-oxidative conditions, especially with metal reductants, hydrogen sulfide always forms as well as hydrogen gas. Hydrogen sulfide appears as a gaseous product, preferably, or as dissolved species. The formation of this gas is due to the release of S2_ ions in solution, which are unstable and can easily be protonated : + 2_ 2H (aq) + S (aq) H2S(g) (14) 26 The hydrogen evolution reaction is a serious side reaction to the main reduction reactions (Eqs. 9-13) and additional discussion will be given later. Also, hydrogen sulfide emerges as a drawback to such systems, unless suitable solutions are found. The mechanism(s) of CuFeS2 reductive decomposition in acidic solutions can be described as: 1) Anodic dissolution of the reducing agent, releasing electrons and its individual ions, either to solution or to undergo mediation reactions 2) Flow of these electrons and/or mediators through the coupling point to reach certain active reaction sites on chalcopyrite particles 3) Step 2 is accompanied by the diffusion of hydrogen ions from the aqueous phase to reach, again, the chalcopyrite surface 4) Reduction of chalcopyrite as per the above written reactions (Eq. 11 for example) It is important to recall that galvanic coupling is an essential step in completing the leaching reactions. The result is the collapse of chalcopyrite crystal structure, releasing iron and sulfur ions in solution. By this, it can be said that reductive leaching introduces changes in both the crystal structure and composition of chalcopyrite. In electrochemical principles, the cathodic process can convert a sulfide into a lower sulfide or a metal, depending on the potential and the amount of charge applied. The formation of the new phase (chalcocite or similar sulfide minerals) has been confirmed by different researchers through X-ray/AES analysis. So, the cathodic reduction of the mineral is due to the presence of a cathodic current distributor, i. e. : a strong reducing agent. A good review of the electrochemical aspects of non-oxidative leaching of chalcopyrite and other sulfides can be found inNicol(1983). The electrochemical nature of copper sulfide leaching is more observed in experiments using chloride media. Since leaching can be done in aqueous systems containing this halide ion, it is worth considering it in the stability diagrams. Upon usage of chloride ion, or similar specific reagents like bromide, iodide, cyanide ions and ammonia, new compounds or complex ions will be formed in solution. Fig. 2.4 represents the Eh-pH diagram for chalcopyrite when chloride ion is incorporated. It was compiled from Peters (1976) and a detailed analysis can be found in Duby (1977). The same discussion given earlier for reductive leaching applies here. 27 J i -j i ! i i i "tiAJ 0 2 4 6 8 10 12 14 PH Fig. 2.4 : The Eh - pH diagram for the copper-iron-chloride-sulfur-water system at 298.15 K. Species has the same activities as in Fig. 2.3 and the chloride ion presents at 1 M activity (Peters (1976)). It is clear from Fig. 2.4 that the chloride ion does not significantly alter the stability regions of sulfide phases, and almost all of the effects are in the regions where dissolved copper is found (the effect of complexation), or its relevant oxides are found (formation of new solid phases). In other words, its action is mainly in the leach solution, and because of this action, leaching in chloride media is expected to be more effective than in sulfate media. From this discussion on the thermodynamics of chalcopyrite leaching, it is evident that oxidative leaching in both acidic and alkaline media is the most common leaching method for 28 chalcopyrite and, consequently, has received substantial research from many investigators toward developing a new copper process. Acid or alkali decomposition does not look promising at all, but there may be a useful route for breaking down this refractory copper mineral in reductive processes that separate iron and some of the sulfur, producing the more reactive chalcocite (with any noble metal impurities) in quantitative yields, which is the subject of this thesis. The three categories of chalcopyrite leaching (oxidative, non-oxidative and reductive leaching) are widely studied and can be accomplished in a variety of media, like chloride and sulfate media. A literature survey will be given in the next sections. The intention here is not to give a comprehensive survey, rather, to give a brief review on various available studies in the published literature, with emphasis on leaching chemistry, industrial mode and relevance to this work. Dreisinger (1997) has given an excellent review on the recent developments in hydrometallurgical treatment of copper sulfides. There are also other published review articles on copper hydrometallurgy, like those published by Dutrizac and MacDonald (1974), Dasher (1973), Paynter (1973), Roman and Benner (1973) and Subramanian and Jennings (1972). More recently, Hackl et al (1995b) and Venkatachalam (1991) have reviewed some aspects of chalcopyrite hydrometallurgy. Pawlek (1976) has compared smelting to hydrometallurgy of copper sulfides from technical and economic points of view, while Paynter (1973) compared eight different hydrometallurgical processes for chalcopyrite. In addition to leaching chalcopyrite in sulfate and chloride media, there are other practiced methods like bacterial, dump, heap and in situ (or solution mining) leaching. Of course, there are other proposed or studied media, like perchloric acid under oxygen pressure (Peters and Loewen (1973)), which is a unique method in that the leaching medium is inert (no complexing), nitrate (Prater et al (1973)), cyanide (Lower and Booth (1965)) and ammonium media (Kuhn et al (1974)), but none are currently practiced on a commercial basis (although the latter was practiced for a while before shutdown). 29 2.2 OXIDATIVE LEACHING OF CHALCOPYRITE Most of the reported work in literature on chalcopyrite oxidative leaching is with sulfuric acid solutions containing ferric sulfate or under autoclave conditions at elevated temperatures (sulfuric acid pressure leaching). The selection of sulfuric acid is due to its merits, like : 1) Low cost 2) Wide availability 3 ) Adequate kinetics 4 ) Regeneration during electrowinning 5 ) Minimal corrosion and maintenance problems The following table summarizes the sulfate media reaction chemistry : Leaching Process Leaching chemistry method " temperature range Ferric CuFeS2(s) + 4Fe3+(aq) -> CU2+(aq) + 5Fe2+(aq) + 2S(s) sulfate 20 - 100 CuFeS ) + 16Fe3+ + 8H 0 i) -> C 2+ + 17Fe2+ + 2S0 2- leaching 2(s 2 ( U (aq) (aq) 4 (aq) +16H+(aq) Oxygen CuFeS2(s) + 4H+(aq) + 02(g) -+ Cu2+(aq) + Fe2+(aq) + 2S(s) + 2H20(i) pressure 100-220 CuFeS2(s) + 402(g) -> Cu2+(aq) + Fe2+(aq) + 2S042-(aq) + leaching 2Fe2+(aq) + 0.5O2(g) + 2H (aq) -» 2Fe3+(aq) + H20(i) Bacterial CuFeS2(s) + 402(g) -> CU2+(aq) + Fe2+(aq) + 2S042-(aq) leaching 20-70 2Fe2+(aq) + 0.5 02(g) + 2H+(aq) -> 2Fe3+(aq) + H20(i) CuFeS2(s) + 4Fe3+(aq) -> CU2+(aq) + 5Fe2+(aq) + 2S(s) S(s) +1-5 02(g) + H20(i) -> H2S04(aq) Table 2.8 : The oxidative leaching of chalcopyrite in sulfate media (Hackl et al (1995b)) The first extensive work on the rate of leaching of copper sulfides at elevated temperatures and pressures was reported by Warren (1958). The leaching of copper sulfide concentrates was studied in sulfuric acid, with oxygen as an oxidizing agent. The rate of oxidation of chalcopyrite was found to be independent of oxygen partial pressure above a certain value (~1 MPa) even at the highest temperature studied (180 °C). Surface chemical reaction was suggested as the major rate controlling step in this system. Elemental sulfur was produced, some 30 suggested as the major rate controlling step in this system. Elemental sulfur was produced, some being oxidized to sulfate at temperatures between 120 and 180 °C. When the pH was below 1, hydrogen sulfide (H2S) was formed at the lower oxygen pressures (that is lower potentials). The activation energy for the reaction was found to be 96 kJ/mol and this high value supports the postulated rate controlling step. Stanczyk and Rampacek (1963) studied the leaching of copper sulfides under acidic conditions using autoclaves. Several oxygen partial pressures were used, and the authors reported that chalcopyrite was the most difficult to leach under all the investigated conditions. Even so, almost complete dissolution of-325 mesh CuFeS2 could be achieved in 30 minutes at 230 °C for oxygen partial pressure greater than 0.62 MPa. Majima and Peters (1966) studied the kinetics of oxidation of several sulfide minerals (bomite, chalcocite, chalcopyrite, covellite, galena, pyrite, pyrrhotite, sphalerite and stibnite) by oxygen at elevated pressures and pH values between 2.7 and 14. The results obtained led to the conclusion that the crystal structure of the minerals did not play a role in the oxidation mechanism, except during the initial period. The authors reported an experimental activation energy of 46 kJ/mol for chalcopyrite. Vizsolyi et al (1967) studied the optimum conditions for the direct pressure leaching of CuFeS2 when elemental sulfur is obtained as a by-product. An increase in temperature enhanced the reaction up to the melting point of sulfur (M. P. = 119 °C), however, above this temperature, the formation of an impermeable liquid sulfur coating on the CuFeS2 surface reduced the rate sharply. Optimum conditions were reported which corresponded to 3.5 MPa oxygen pressure at 115.5 °C, copper concentrate ground to P995 , i. e. 99.5 % passes, minus 325 mesh and a leaching retention time of 2.5 hours. The idealized overall reaction under these conditions was suggested as : CuFeS2(s) + H2S04(aq) + 1.2502(g) + 0.5H2O(1) -> CuS04(aq) + Fe(OH)3(aq) + 2S(S) (15) although under acidic conditions, hydrated species of ferric ions only exist. The direct and strong dependence of the reaction rate on temperature indicates a chemical reaction control system. Later, Warren et al (1968) used the idea of heating prior to oxidation to enhance copper extraction. 31 Yu et al (1973) investigated the acid pressure leaching of chalcopyrite. The effects of temperature (125 to 175 °C), oxygen partial pressure (0.5 to 2.8 MPa), particle surface area (or diameter), and sulfuric acid concentration (0.1 to 1.0 N) were studied. The particle kinetics followed a shrinking core model, with an electrochemical surface reaction identified as the rate controlling step. The results indicated that the oxidation reaction was first order with respect to oxygen concentration and surface area. The oxygen dependence indicated that at high oxygen pressures, the reaction rate reached a limiting value due to surface saturation with oxygen. The observed enthalpy of activation for oxygen adsorption was 174.5 kJ/mol, while that for the surface reaction was 30.5 kJ/mol. Hackl et al (1995a) studied the effect of sulfur-dispersing surfactants during acid pressure leaching of chalcopyrite in the temperature range 125-155 °C. Such surfactants were used in different applications, but less work is reported on chalcopyrite acid pressure leaching. The leaching chemistry is written as : 2CuFeS2(s) + 5H2S04(aq) + 2.502(g) -> 2CuS04(aq) + Fe2(S04)3(aq) + 4S(S) + 5H20(1) (16) 2CuFeS2(s) + H2S04(aq) + 8.502(g) -> 2CuS04(aq) + Fe2(S04)3(aq) + H20(1) (17) The authors confirmed the findings by other researchers on temperature effect on leaching. At 110-120 °C, the rate of reaction was slow, as indicated by low consumption of oxygen, low copper extraction (< 50%) and low total sulfide oxidation (< 45%). Elemental sulfur yield was about 72%. At 130-170 °C, the reaction rate was found to increase rapidly, but then slowed dramatically or stopped completely, due to an inhibition effect caused by elemental sulfur wetting. At temperatures above 180 °C, the extraction increases with increasing temperature, and yield of sulfate ion increases, signaling finally complete oxidation of sulfidic sulfur to sulfate. Test results at -120 °C confirmed the inhibitory effect of liquid sulfur on leaching. Hackl and his coworkers stated that many of the tested surfactants decomposed too rapidly to be of benefit. The best obtained results were those with orthophenylene diamine (OPD) but the latter was added at a high dosage (-5%), resulting in increased copper extraction (80% with OPD, compared to 40% without addition) only after prolonged retention time (6 hours). Even, if elemental sulfur was prevented from wetting chalcopyrite, or with silver catalyzed pressure leaching, chalcopyrite still leached slowly, leading to the conclusion that the 32 reaction rate is controlled by another passivating mechanism, unrelated to elemental sulfur formation (see Hackl et al (1995b) for more discussion). King and Dreisinger (1995) discussed various aspects of copper concentrate autoclaving, and presented a detailed economic evaluation of such processes. As noted in Table 2.8, high temperature autoclaving of copper concentrates is done without an acid. Rather, sulfur is totally oxidized to sulfate ion. The obvious advantage of this option versus lower temperature options is the rapid and complete dissolution of copper and precipitation of iron. The production of quantitative amounts of weak acid make this options similar to smelting which produces quantitative amounts of sulfur dioxide, eventually converted to sulfuric acid. Nonetheless, under some conditions, this option may be desired. The overall chemistry can be written as : 4CuFeS2(s) + 1702(g) + 4H20(1) 4CuS04(aq) + 2Fe203(s) + 4H2S04(aq) (18) The steps for iron precipitation were discussed by Dreisinger (1997), and iron is shown here as the final residue, hematite. The precipitation of solubilized iron as hematite or basic ferric sulfate (depending on acidity) is an advantage, particularly when such precipitates do not appear to impede leaching. Copper concentration in leach liquor usually increases with increasing temperature, as solubility of iron compounds decreases. Weak sulfuric acid is generally produced, and the authors used 3 hours as retention time. At around 200 °C and 0.69 MPa oxygen pressure, 99% extraction of copper was possible. Under such severe conditions complete oxidation of sulfidic sulfur to sulfate is achieved. It is tempting to attribute the high reactivity of the refractory chalcopyrite at elevated temperature to the absence of elemental sulfur as a stable reaction product, but this has not yet been proven kinetically. The precipitated residue, depending on feed composition, usually contains other impurities and precious metals like gold and silver. These can be recovered by conventional cyanidation. As indicated earlier, the authors gave an economic evaluation of such an option versus other refining processes. The crucial factor is the fate of the acid, depending whether it is to be used in a nearby operation, neutralized with acid consuming gangue, or neutralized with a purchased agent prior to disposal. All these and other aspects can be found in their paper. Oxidative leaching was also studied under atmospheric conditions. Lowe (1970) studied the dissolution of ground and sized natural chalcopyrite in acidified ferric sulfate solutions over 33 the temperature range 32 to 50 °C. Linear kinetics were observed and the apparent activation energy was found to be 75.3 kJ/mol. The rate of dissolution was independent of variation in H2S04 concentration over the range 0.05 to 0.78 M. The rate was also insensitive to changes in ferric ion concentrations greater than 0.02 M. The author interpreted his results as being indicative of rate control by a surface reaction with surface saturation of Fe2(S04)3. Dutrizac (1989) and Dutrizac et al (1969) studied the dissolution of sintered discs of synthetic chalcopyrite between 50 and 94 °C in acidic ferric sulfate solutions. Stoichiometric yields of sulfur and ferrous ion were obtained according to the reaction : CuFeS2(s) + 2Fe2(S04)3(aq) -+ CuS04(aq) + 5FeS04(aq) + 2S(S) (19) The reaction displayed parabolic kinetics and the rate was approximately an order of magnitude greater than that observed for natural chalcopyrite. The experimental activation energy was found to be 71 ±3 kJ/mol. Below ferric ion concentrations of 0.01 M, the rate controlling step was attributed to ferric sulfate diffusion through a thickening sulfur layer formed on the surface of chalcopyrite, even though a reaction order of 2 was observed with respect to Fe2(S04)3 concentration. At higher ferric sulfate concentrations, the rate was independent of the ferric ion strength and was attributed to the outward diffusion of ferrous ion through the sulfur layer. They also found the rate to be insensitive to changes in acid concentrations and disk rotation speed. The unusual ferric ion dependence was attributed to the strong formation of ferric sulfate complexes in solution. Munoz et al (1979) also studied the acid ferric sulfate leaching of chalcopyrite using monosized particles and attritor ground concentrate in an intensely stirred reactor at ambient pressure. Their results indicated that although the initial stage of reaction appears to be controlled by an electrochemical surface reaction, it contributes little to the overall extent of reaction. It was found that the reaction is controlled by a transport process through the reaction product, which was identified as transport of electrons through the sulfur layer. The reaction rate was dependent on the inverse square of the initial particle diameter, and independent of Fe2+, Fe3+, Cu2+ and H2S04 additions. The apparent activation energy from experimental results was found to be 83.7 kJ/mol, which was shown to be approximately the same as the activation energy for transport of electrons through the elemental sulfur layer (96 kJ/mol) calculated from both conductivity and electron mobility measurements. 34 Beckstead et al (1976) studied the effect of particle size in the acid ferric sulfate leaching of attritor-ground chalcopyrite concentrates. According to the authors, the experimental results suggest that initial CuFeS2 particle diameter is the only controllable variable which has a significant effect on copper extraction. Different size production procedures were tested, and attrition grinding was found to be most useful. 90% copper extractions were possible by leaching particles of 0.5 um size at 93 °C in three hours. Enhanced leaching was not attributed to the "activation" or retained strain energy, rather, to the increase in surface area. The leaching data were analyzed and found to follow the product layer control model, but the apparent activation energy was somewhat high (83.68 kJ/mol). Jones and Peters (1976) studied the leaching of chalcopyrite with ferric sulfate and ferric chloride solutions. The chemistry of both systems is different, and a substantial amount of sulfur was oxidized to sulfate in the ferric sulfate system. The authors noted that increasing ferric ion concentration enhanced copper extraction in the range 0.01-0.1 M. Beyond these values, copper extraction decreased. Additions of ferrous sulfate retarded the reaction, which is unusual in such systems. The mixed potential measurements, around 0.61 V, indicated significant polarization of both the anodic and cathodic portions of the net leaching reaction (Eq. 19). This means that chalcopyrite dissolution in ferric sulfate leaching is under mixed control. The authors found that the chalcopyrite leaching rate is independent of particle sizes below 100 mesh (smaller than 149 um) in 0.1 M ferric sulfate solution at 90 °C. Later, Dutrizac (1989) attributed this unusual independence to the use of non-monosized particles. Some authors proposed a pretreatment step to be combined with oxidative leaching. The purpose was to introduce changes in the crystal structure and/or composition of chalcopyrite, rendering it more amenable for chemical attack. For instance, mechanical pretreatment was combined with ferric sulfate or acid pressure leaching of chalcopyrite. Rice et al (1990) investigated the effects of turbomilling parameters on the simultaneous grinding and ferric sulfate leaching of chalcopyrite. The authors were able to induce crystalline changes in chalcopyrite by a specially-designed turbomill. The results show that leaching at 25 degree Celsius was not enhanced by turbomilling. Rather, it was possible to completely dissolve the copper at around 90 degree Celsius by leaching for about two hours. 80% copper extraction was achieved with 1 hour leaching time. The leaching rate increases with operating mill speed and solid pulp density, due to scrubbing-attrition effect on chalcopyrite, exposing more fresh 35 surfaces to the lixiviant. However, the associated energy requirements were greater than other proposed leaching methods, but within the same order of magnitude. Gerlach et al (1973) presented a study of chalcopyrite leaching at 100 °C, after activation of the mineral by intensive grinding. The authors found that subjecting chalcopyrite to an oscillating milling action involving a strong impact, permits 99% of the copper to be extracted in a subsequent single pressure leaching stage with at least stoichiometric amounts of sulfuric acid. In addition, 85% of sulfide sulfur was converted to the elemental form. According to the authors, there will be a strong stressing, distortion or defect-formation in the crystalline particles making these particles more susceptible to leaching, which can not be obtained by simple comminution. 36 2.3 NON-OXIDATIVE LEACHING OF CHALCOPYRITE Non - oxidative leaching is referred to as a chemical metathesis process, in which chalcopyrite is converted to a richer copper sulfide (usually covellite). The general reaction can be written as : CuFeS2(s) + CuS04(aq) -> 2CuS(s) + FeS04(aq) (20) The formation of covellite may be followed by further enrichment and some sulfur oxidation to sulfate: 6CuS(s) + 3CuS04(aq) + 4H20(1) -» 5Cu,.gS(s) + 4H2S04(aq) (21) Sohn and Wadsworth (1980) found reaction 20 to be very slow, requiring sub-sieve size particles (-400 mesh) for significant conversion. The reaction itself was insensitive to the cupric ion concentration. Stirred ball milling was used to obtain very fine particles, which induced lattice strain. The reaction rate with strained particles was markedly higher than with annealed chalcopyrite. This enhanced rate was attributed to a recrystallization process which provides easy paths for diffusion along dislocations and grain boundaries. As a note, changes in the mineral crystal structure by any activation method would enhance leaching rates by the same reasoning. The reaction rate was found to be limited by the outward diffusion of ferrous ions through the product layers of covellite (CuS) and digenite (Cu, gS). In the Sherritt-Cominco copper process (Swinkels and Berezowsky (1978)), a pretreatment method was used. The previous metathesis reaction (Eq. 20) is the main reaction, along with the conversion of bornite into simpler copper sulfides, as : Cu5FeS4(s) + CuS04(aq) -> 2Cu2S(s) + 2CuS(s) + FeS04(aq) (22) An undesired side reaction is that between the formed covellite and any cupric sulfate still present, altering covellite to chalcocite (Cu2S): 5CuS(s) + 3CuS04(aq) + 4H20(1) 4Cu2S(s) + 4H2S04(aq) (23) Although the copper content of chalcocite is the highest among all other copper sulfides, and more amenable for further leaching, this side reaction will entail the increase of the required CuS04 recycle load by 20 to 30%. In addition, the process design was based on the main leaching reactions (Eqs. 20 and 22), which are in turn dependent on feed composition. 37 2.4 REDUCTIVE LEACHING OF CHALCOPYRITE Reductive leaching (decomposition) is a simple process in which chalcopyrite is converted to a "richer " copper sulfide. As indicated earlier, it involves reacting chalcopyrite with a common lixiviant, but in the presence of a good reducing agent, which is usually a less noble metal or a gas. Published studies have shown that cathodic conversion of chalcopyrite to other copper sulfides provides a means for pretreatment prior to anodic dissolution or normal leaching of minerals. Conversion of chalcopyrite to a simpler copper sulfide form could have many practical advantages in copper hydrometallurgy. The most important of these would be the rejection of iron and sulfur from the copper mineral. On the basis of copper content, chalcocite (a common product of the alteration process) has a much higher percentage of copper than does chalcopyrite. Thus, the percentage of copper in a concentrate can be increased by enrichment. Furthermore, processing of the enriched concentrate to produce metallic copper would be simpler and less expensive than the original mineral, once such a removal has been achieved. Iron and sulfur removal implies significant weight reduction. The weight of CuFeS2 required for 100 kg of contained copper is 288.8 kg while Cu2S requires only 125.2 kg for the same amount of contained copper. Also, chalcocite is denser than chalcopyrite, by about 30%. This weight reduction is specifically important where the cost of shipping a concentrate to a distant mill or smelter is high. There are other advantages besides the removal of iron and sulfur. Unlike most hydrometallurgical operations, any precious metal values will remain with the enriched concentrate rather than being lost with a discarded iron precipitate. It is important to retain precious metal values with the relatively high prices and demand. Finally, in a typical copper concentrating operation, when the recovery of the concentrate is increased, the grade of the collected product will decrease. However, a minimum grade is usually set by the smelter which receives the concentrate, and it is this value which dictates the recovery of any operation. By using the enrichment process, a higher overall recovery could be achieved and the lower grade concentrate could then be enriched to meet the smelter contract, with less penalties. 38 In summary, reductive decomposition or conversion of chalcopyrite to a more amenable form has some advantages such as : 1) The enriched product might further be treated hydrometallurgically or pyrometallurgically, at reduced costs and milder conditions, as major portions of impurities are removed 2) The removal of iron and sulfur will decrease shipping and material handling costs, specifically if the concentrate is to be transported to a distant plant 3) The enrichment process means high grade concentrates which imply higher copper recovery 4) Fixation of sulfur and iron upon leaching. Part of the sulfide sulfur is removed (usually as H2S) and iron enters the leach solution as ferrous ions. Once the selective separation of most of the iron is achieved, a troublesome operation in hydrometallurgy is avoided. For the purpose of comparison, the rejection of iron as iron-rich solutions and of sulfur as hydrogen sulfide is the counter part of the rejection of iron-rich slags and sulfur dioxide in base metal sulfide smelting. 5) Precious metal values present originally in the feed will be retained with the enriched sulfide phase, allowing better treatment and recovery While this application of enrichment process is technically feasible, there are some limitations that need to be mitigated. Such a process seems to be an expensive operation. The associated additional costs might make the process economically less attractive due to the need for many unit operations. The enrichment process is envisaged to require the usage of solid- liquid and/or other separation stages (flotation, filters, thickeners, etc.) which add to the expenses. The reductants used in leaching need to be regenerated (if expensive) in the process, hence a chemical or electrochemical reduction stage is needed. If the reagent is expensive and regeneration is by electrowinning, the latter is an energy intensive process, which makes the process, as a whole, unattractive. The generation of hydrogen sulfide, when using metallic reductants, requires collection and finding uses or treatment. There are also health concerns. Further, the gradual decrease of ore grades might make this option less applicable. Initial work to understand the enrichment reactions occurring in a porphyry ore body began as early as the turn of this century when various geologists speculated on the possible 39 chemical reactions involved in the enrichment process. Emmons (1913) proposed two possible reasons for the enrichment of sulfide ores and precipitation of cupric sulfate. These include : 1) A solubility series based on the solubility products (Ksp) of various metal sulfides 2) An electrochemical series based on the reduction potentials of these sulfide minerals The reductive decomposition is better explained using electrochemical terms since most sulfide minerals exhibit some electrical conductivity (see Table 1.2). A fundamental property of these semiconducting minerals is the characteristic "reduction or rest potential". This potential corresponds to the equilibrium electrode potential where no net cathodic or anodic current occurs, i. e. : no net reaction. A collection of the reduction potentials of some common minerals, and some other elements is presented in the following table : n+ - General reaction : M (aq) + xS(s) + ne <-> MSX(S) Material Chemical Formula Reduction Potential, E°, V Pyrite FeS2 0.63 Chalcopyrite CuFeS2 0.53 Chalcocite Cu2S 0.44 Covellite CuS 0.42 Copper Cu 0.34 Galena PbS 0.28 Hydrogen H2 0.00 Lead Pb -0.13 Sphalerite ZnS -0.24 Pyrrhotite FeS -0.28 Cadmium Cd -0.40 Iron Fe -0.44 Chromium Cr -0.55 Zinc Zn -0.76 Magnesium Mg -2.37 Table 2.9 : Reduction potentials of some metals and minerals at standard conditions (Hackl (1998) and Bard et al (1985)) The conditions for galvanic interaction, i. e. electron transfer from one material to the other, require materials exhibiting different rest potentials to be in contact. This galvanic interaction, or coupling, arises due to differences in the electrochemical reactivity of these materials, indicated by the rest potential. When two dissimilar materials are brought into contact, 40 the one with the highest reduction potential will be cathodically protected (cathodic behavior will be imposed), while the other one will undergo anodic dissolution. Similarly, when chalcopyrite is brought into contact with any less noble mineral or metal, it will undergo a cathodic reaction, where, for instance, its crystal structure will rearrange and form another solid phase, while the other mineral or metal will react anodically. Hence, reductive decomposition is achieved by utilizing the concept of metals and/or semiconductors in galvanic contact. From Table 2.9, it can be seen that all materials below chalcopyrite can be used for this decomposition in conjunction with proper solution conditions. Peters (1984) presented some explanation on the fundamental theory and some illustration of the galvanic interaction. He stated that such a process is dependent on : 1) The rest potential (E°) 2) The conditions of galvanic interaction, in particular the duration of electrical contact 3) The mineral conductivity It is important to remember that a mineral electrode system will establish and maintain a certain equilibrium potential that depends not only on the solution composition but also on the composition of the solid phase(s). Hiskey and Wadsworth (1975) studied the galvanic conversion of chalcopyrite using metallic copper in sulfuric acid solutions. The overall reaction was written as : + 2+ CuFeS2(s) + Cu(s) + 2H (aq) Cu2S(s) + Fe (aq) + H2S(g) (24) and found to occur by an electrochemical (galvanic) mechanism. The authors found that agitation has a detrimental influence on the conversion reaction, explained by an inhibition effect on the formation of particle to particle welds or galvanic coupling. The results for the effect of initial particle size of metallic copper (as shots) reflected a uniform enhancement of conversion with increased copper surface area, but the kinetics were found to be insensitive to chalcopyrite particle size. Finer sizes of copper resulted in better mixing and provided greater and more rapid conversion. The authors found that initially added amounts of cupric ions are detrimental to the decomposition process, by lowering the conversion, but for various ferrous ion additions, no adverse effect was noticed. No explanation was given for the latter finding. 41 Extensive X-ray diffraction and microscopic examination of the solid residue identified Cu2S as the predominant final reaction product, with a thin layer of bornite (Cu5FeS4), and some djurlite (Cu195S). The leaching kinetics were found to follow the chemical reaction control model, with an apparent activation energy, Ea, of 48.12 kJ/mol. The conversion was also affected by the initial sulfuric acid concentration, where a half order dependence was found. Under the experimental conditions, 96% recovery of copper was possible. The authors concluded that the kinetics may be explained by a process in which the mixed potential shifts from the anodic half cell potential to the cathodic half cell potential during the course of reaction. Sohn and Wadsworth (1980) investigated the reduction of chalcopyrite with sulfur dioxide in the presence of cupric ions. The overall reaction, which was electrochemical in nature, was written as : CuFeS2(s) + 3CuS04(aq) + 2S02(g) + 4H20(1) -> 2Cu2S(s) + 4H2S04(aq) + FeS04(aq) (25) This reaction was found to proceed by a corrosion mechanism rather than galvanic interaction of two solid phases. Consequently, agitation sufficient to suspend the solid particles is desirable which was confirmed from the experimental results. During the reaction, a defect structure of chalcopyrite (djurleite) followed by bornite were observed as intermediate reaction products, due to the preferential removal of iron from chalcopyrite lattice. The associated oxidation of S02 gas produces sulfuric acid. The experimental results showed that the reaction rate was sensitive to cupric ion concentration, but was not markedly affected by the partial pressure of sulfur dioxide. Hydrogen ion dependence was even less sensitive. Initial additions of ferrous ion favored more conversion, explained by the ferrous ion ability to exhibit a catalytic effect on leaching. The authors used the linear leaching model to fit the experimental data, which was satisfactory. An apparent activation energy of 77.5 kJ/mol chalcopyrite, and a linear dependence on the inverse of chalcopyrite particle size, supported the postulated surface chemical reaction control. For an attritor ground concentrate, greater than 95% conversion (decomposition of chalcopyrite to chalcocite) was obtained using twice the stoichiometric amount of cupric ions, in a solution containing 20 gpl ferrous ions. However, the authors did not report any information on pH values before and after the reduction. Nonetheless, they developed a process flowsheet that was claimed to be possible in specific areas of copper hydrometallurgy. 42 Hackl et al (1987) examined the autoclave reductive decomposition of chalcopyrite by additions of metallic copper, cuprous and cupric salts in a reducing atmosphere of hydrogen gas. They found that iron could be removed from both the chalcopyrite and pyrite in feed concentrates, to produce chalcocite or a low iron phase with the approximate composition Cu^S (1 < x < 2). The authors also found that the reduction leach took place at a rate comparable to the hydrogen reduction of copper sulfate solutions. An activation energy of approximately 67 kJ/mol chalcopyrite was determined. SEM testing indicated that the reaction product layer of chalcocite was expanding, cracking and spalling away from the unreacted particle core, due to positive volume changes. This form of product layer suggests a chemical reaction control mechanism. The following assumptions were made : a) Reverse leaching is based on a solid state mechanism b) Cu2S is a more favorable phase for rapid solid state diffusional processes than is CuS c) Solid state processes have high activation energies and are therefore preferred under autoclave conditions Based on these assumptions, the authors investigated the reduction leaching under autoclave conditions using molecular hydrogen and cupric ion additions. The general reaction was written as : 2+ 2+ + CuFeS2(s) + 3Cu (aq) + 2H2(g)^2Cu2S(s) + Fe (aq) + 4H (aq) (26) The results show that hydrogen reduction is effective in converting CuFeS2 to chalcocite with 99.97% iron extraction, with the codissolution of the associated pyrite mineral. The proposed leaching mechanism was based on hydrogen reduction of cupric ions followed by galvanic contact between copper and the mineral, with the latter being the rate determining step. The authors also investigated autoclave reduction leaching of chalcopyrite using metallic copper and cupric ion. The general reaction was written as : 2+ 2+ CuFeS2(s) + Cu (aq) + 2Cu(s) -> 2Cu2S(s) + Fe (aq) (27) 97% of iron was released while solution copper was almost quantitatively precipitated. The nearly complete removal of the chalcopyrite iron component means that there are other leaching mechanism(s) in addition to the galvanic contact. Hackl and his coworkers suggested a 43 mechanism in which leaching is mediated by cuprous ions. These ions will undergo disproportionation reactions (Eq. 3), that will eventually lead to the main leaching reaction, Eq. 27. The authors concluded that the outward diffusion of the ferrous ion is rate limiting as it has a small solubility in chalcocite, accounting for the high apparent activation energy (67 kJ/mol). According to the authors, this method of reduction leaching, while much faster, for instance, than non oxidative leaching, is much harder to be incorporated into a complete hydrometallurgical circuit for copper for different reasons. One of these reasons is the fact that 2 the reduction leach conserves sulfide sulfur (S ) during decomposition of CuFeS2 to Cu2S. This imposes a circulating load of copper in some form to the reduction leach, with 3 moles of copper required per mol of chalcopyrite, as per Eq. 27. In the case of other copper sulfides, this load may increase. Peterson and Wadsworth (1994) studied the autoclaving of chalcopyrite under reducing acidic conditions in the presence of copper sulfate. The isothermal kinetic study was made at temperatures above 100 °C using autoclaves. The effects of particle size, temperature, acidity, cupric ion concentration and ferrous ion concentration on the leaching kinetics were examined. It was found that chalcopyrite reacts in two stages. It first forms covellite with the accompanying rejection of iron. In the second stage, the covellite reacts to form digenite. A partially reacted particle was examined using X-ray diffraction and SEM and found to contain an unreacted chalcopyrite core surrounded by a thin layer of covellite. The covellite layer was surrounded by a thick digenite product layer which grows inward as the particle reacts. This two stage reaction was explained by a mixed kinetic control model consisting of surface reaction control and product layer diffusion control in series (Appendix I). The mechanism for the main leaching reaction : 3CuFeS2(s) + 6CuS04(aq) + 4H20(1) -> 5Cu,.8S(s) + 3FeS04(aq) + 4H2S04(aq) (28) was depicted to be composed of several stages. The first stage was a net metathetic reaction to form covellite, while the second stage was the cathodic reduction of this covellite, with additional cuprous ions, to form digenite. The second stage was found to impose a diffusion control mechanism, added to the metathetic reaction which imposed a chemical reaction control mechanism. Cupric and ferrous ions exhibited catalytic effects on leaching, while hydrogen ions affected only the second leaching stage. The estimated activation energy for the metathetic 44 reaction was 92 kJ/mol chalcopyrite, while for the diffusion control reaction it was 99.7 kJ/mol chalcopyrite. The latter is higher than the values normally associated with reactions controlled by diffusion through a product layer, but compares favorably with those for sulfide ion diffusion through metal sulfides (Hackl et al (1987)). Felker and Bautista (1990) studied the electrochemical dissolution of chalcopyrite using a fluidized bed electrochemical reactor (FBER). They proposed a three stage process in which chalcopyrite is firstly reduced to digenite, followed by the latter oxidation to form cupric ions and elemental sulfur. The last stage is the electrowinning of metallic copper from cuprous and cupric ions, in a conventional electrochemical cell. The net process reaction for the cathodic conversion of chalcopyrite can be represented as : + 2+ 2+ 1.8CuFeS2(s) + 5.2H (aq) -> Cu(s) + 0.8Cu (aq) + 1.8Fe (aq) + 2.6H2S(g) + S(s) (29) while the net moles of consumed electrons is 2.8 mol e" per 1.8 mol chalcopyrite. According to the authors, this proposed dissolution process represents a method for separating the copper, iron and sulfur from chalcopyrite by a series of reduction-oxidation reactions. The authors developed a mathematical model, based on Butler-Volmer equation, to describe the reaction mechanism and found the product layer to be porous. The estimated porosity, based on measurement of molar volume changes, was 65%. Chae and Wadsworth (1979) investigated the galvanic interactions between particulate chalcopyrite and lead in hydrochloric acid solutions. It was found that chalcopyrite could be converted into copper-rich sulfides followed by the sequential conversion to metallic copper. The effects of temperature, particle size, hydrochloric acid concentration and agitation were systematically examined. Agitation resulted in increased conversion. The reaction rate was insensitive to acid concentrations greater than 2 M, due to the formation of complex chloride ions. The reaction rates also appeared to be insensitive to initial particle sizes of chalcopyrite and metallic lead. Consequently, the authors developed a leaching model to account for these geometrical effects. According to the authors, 90% conversion of chalcopyrite can be achieved in 2 hours at an agitation speed of 350 rpm, temperature of 89 °C, stoichiometric molar ratio of lead to chalcopyrite and 1 N HC1 solution. An experimental activation energy of 28.1 kJ/mol was obtained. The initial kinetics were explained in terms of an ohmic electrical resistance across Pb- 45 CuFeS2 contact points inhibiting the flow of electrons. The conversion was depicted to occur by a combined corrosion-galvanic mechanism, leading to the net leaching reaction : 2CuFeS2(s) + 6HCl(aq) + P^ (30) or if rewritten for producing metallic copper : CuFeS2(s) + 4HCl(aq) + Pb(s) -> Cu SEM and X-ray diffraction tests showed that the reaction product was a porous defect structure of the form djurleite (Cu, 95S), but chalcopyrite grains were surrounded by this solid product. The lead particles were found to be free from solid products. Lee and Donofrio (1982) studied copper leaching from chalcopyrite in hydrochloric acid solutions using the reductants : chromium, zinc, cobalt, aluminum, nickel and iron. In their study it was found that chromium and zinc were the most effective metals for galvanic conversion of chalcopyrite, while the other reductants showed a little or insignificant conversion. Agitation was found to be detrimental to the reaction, explained by preventing galvanic coupling, but the reaction rate was dependent on the particle size of both chalcopyrite and the reductant. Reducing CuFeS2 particle size from 100/170 mesh to 325/400 mesh resulted in a 20% increase in conversion. The same was also found for the metallic reductants. The authors did not give a detailed kinetic study but from the available data in their article it is clear that the reactions are not very temperature sensitive. Increasing the temperature from 45 to 80 °C, resulted in a 10% increase in final conversion using chromium, and a 15% increase using zinc, after 60 minutes of leaching. Under the experimental condition, maximum conversion was less than 53% at 80 °C with IN HC1, CuFeS2 particle size of 147 um (-100 mesh) and molar ratio of chromium to chalcopyrite of 2:1. The same applies for zinc. The last reductant to be reviewed is metallic iron. In the published literature, there are few studies on using metallic iron as a reductant. These studies were concerned with the qualitative description of the leaching reactions rather than performing a systematic analysis. In sulfate media, the net leaching reaction is : 2CuFeS2(s) + 3H2S04(aq) + Fe(s) -> Cu2S(s) + 3FeS04(aq) + 3H2S(g) (32) and in chloride media, it is : 2CuFeS2(s) + 6HC1 (aq) + Fe(s) Cu2S(s) + 3FeCl2(aq) + 3H2S(g) (33) 46 Shirts et al (1974) conducted bench scale studies on a reductive hydrometallurgical procedure to convert chalcopyrite flotation concentrate to copper metal or a readily leachable sulfide, using metallic copper, iron or lead, in sulfate and chloride media. The parameters investigated were acid concentration, mole ratio of reductant to CuFeS2, temperature and agitation speed. The findings for copper and lead are in well agreement with those of Hiskey and Wadsworth (1975) and Chae and Wadsworth (1980), respectively. For metallic iron, the authors found that chloride medium is more effective than sulfate medium, based on the percentage of iron extraction. Greater than the stoichiometric amounts of iron were needed for good conversion, due to iron consumption by side reaction. The effect of chloride medium was that some of the enriched sulfide (labeled as Cu2S) was further reduced to copper by galvanic interaction with iron. The authors claimed that the new solid phase is porous and contains both cuprous and cupric species and a small amount of dissolved iron. It was noticed that leaching proceeds at two different rates, the first one being very fast (within less than 15 minutes), contributing to major conversion, while the second one, continued for around 6 hours and contributed very little to the total conversion. In the latter stage, inhibition of further conversion was attributed to lower transport rate of ferrous or sulfide ion, as well as electrons, through the copper sulfide product layer. The diminishing of active surface area of the solid particles was also indicated as a possible cause for the leveling off of conversion after 15 minutes of leaching. The authors found that increasing the temperature from 40 to 60 °C resulted in a 40% increase in final conversion, which was given no explanation. The assessment for best acid concentration showed that at 3.2 N HC1 and 1.6 N H2S04, best alteration is obtained, provided that twice the stoichiometric amount of iron is added, and a temperature of 95 °C is used. Utilizing these conditions for agitated systems, iron extraction from chalcopyrite of greater than 90% was achieved in three hours of leaching. No kinetic models were reported. In the electrochemical study done by Nicol (1975), iron was found to be more effective than copper, for instance, by about an order of magnitude. In addition, iron does not suffer from the limitations of mass transport through a boundary film, as was the case with copper. The author emphasized that in the case of iron, there is less dependence of the reaction on agitation speed, represented by the disc rotation speed used in the electrochemical experiments. Under 47 such conditions, mild agitation was found to cause an enhancement for proton transport to active reaction sites on chalcopyrite particles. The summary of this literature survey on chalcopyrite reductive leaching is that most of the researchers confirmed the electrochemical nature of the reactions, in that these reactions are composed of anodic and cathodic portions. There are two types of mechanisms for these reactions : a galvanic mechanism and a corrosion mechanism. Also, there are different research results on the composition of the new solid phase and the effects of different parameters, like agitation, temperature, particle size, acid concentration, molar ratios and the presence of certain ions. Finally, there are a variety of selected leaching models, and every researcher or a group of researchers justified their findings by the apparent conformance of leaching data with theoretical considerations. Table 2.10 summarizes these findings. 48 Researchers Reductant Remarks Hiskey and Copper • Chemical control with activation energy (EJ of 48.12 kJ/mol Wadsworth (1975) • Galvanic mechanism • Rate is dependent on hydrogen ion concentration • Rate is independent of initial particle size of chalcopyrite • Agitation is detrimental to conversion • Cu2+ is detrimental to conversion while Fe2+ is not • Products : Cu2S with some Cu5FeS4 and Cu,95S Sohn and S02 with • Chemical control with Eaof 77.5 kJ/mol Wadsworth (1980) CuS04 • Corrosion mechanism • Agitation is required • Rate is dependent on initial particle size of chalcopyrite • Rate is dependent on Cu2+ concentration • Fe2+ has a catalytic effect • Product: Cu, 95S followed by Cu5FeS4 2+ Hackl etal (1987) H2 with • H2 reduction of Cu followed by Cu reduction of CuFeS2 CuS04 • Galvanic mechanism • Complete dissolution of chalcopyrite iron component • Cu2S is the product 2+ Hackl etal (1987) Copper • Controlled by the outward transport of Fe , Ea = 67 kJ/mol under • Galvanic and Cu+ mediated mechanisms autoclave • 97% dissolution of chalcopyrite iron component conditions • Cu2S is the product Peterson and CuS04 • Mixed chemical-transport control Wadsworth (1994) under • Metathetic and electrochemical mechanism autoclave • Ea equals 92 kJ/mol for chemical control conditions • Ea equals 100 kJ/mol for transport control • Cu2+ and Fe2+ have catalytic effects • H+ only affects the transport mechanism • Cu, 8S is the product Felker and H2S04 • Three stages of reduction Bautista(1990) reduction • Copper is the final product with some elemental sulfur Chae and Lead in • Controlled by transport of electrons with Ea of 28.1 kJ/mol Wadsworth (1980) HC1 • Combined corrosion-galvanic mechanisms solution • Rate is independent of chalcopyrite and lead particle sizes • Stoichiometric amount of reductant is sufficient • Products : Cu, 95S followed by Cu Shirts etal (1974) Iron in • Controlled by a transport process H2S04 and • Two stages of leaching HC1 • Rate is dependent on acid concentration and reductant amount solutions • Twice the stoichiometric amount of reductant is required • Cu2S is the product Table 2.10 : Summary of reviewed research on chalcopyrite reductive leaching 49 2.5 HALIDE MEDIA LEACHING OF CHALCOPYRITE The same classification for leaching in sulfate media can be extended to halide media. Haver and Wong (1971) studied the ferric chloride leaching of chalcopyrite. The authors investigated the effects of particle size* temperature and ratio of ferric chloride to chalcopyrite on the rate of dissolution. For particles of P9g minus 325 mesh size, it was possible to extract 99.5% of the copper in 2 hours at the boiling point of solution (106 °C). Parabolic kinetics were reported and attributed to limited mass transport through a progressively thickening sulfur layer formed on the chalcopyrite surface. The authors found that at a ratio of ferric chloride to chalcopyrite of 1 to 2.7, virtually all the dissolved copper was in the cuprous state, and the overall reaction was : CuFeS2(s) + 3FeCl3(aq) -> CuCl(aq) + 4FeCl2(aq) + 2S(S) (34) Kruesi et al (1973) presented the Cymet process of the Cyprus Metallurgical Corporation. This was a process for converting the concentrates of base metal sulfides to the corresponding pure metals and elemental sulfur. The process uses two stages of leaching in a mixed FeCl2 - CuCl2 - NaCl solution to produce cuprous chloride, while iron is rejected as jarosite. Vacuum crystallization was used to recover copper as CuCl followed by hydrogen reduction in a fluidized bed electrochemical reactor. The last step is the smelting of the precipitate to produce copper wire. The main technical problems for this process were the generation of HC1 in the reactor, corrosion problems and high capital cost in the vacuum crystallization unit. For economic reasons the plant was shut down in 1982. Peters et al (1981) patented the UBC - Cominco process for copper recovery from sulfide concentrates using ferric chloride leach route. The process utilizes FeCl3 leaching followed by cementation with metallic copper to reduce Cu2+ ions to Cu+ ions, where CuCl is obtained via crystallization. The residual liquor is cemented with iron to produce cement copper seeds and ferrous chloride solution. The ferrous chloride solution is oxygen pressure oxidized to regenerate the lixiviant: FeCl3, whereas excess iron is precipitated as Fe203. Hydrogen reduction is used to produce metallic copper from CuCl crystals, while sulfur reports almost quantitatively as elemental form in the residue. Cominco diverted from this option to use a sulfate-based option in collaboration with Sherritt (Section 2.3). Schweitzer et al (1982) presented the CLEAR (copper leach, electrolysis and regeneration) process. This process of Duval corporation comprises four steps : concentrate 50 leaching in two stages using a mixed KCl-NaCl brine as the leaching solution, rejection of soluble iron as potassium jarosite in a pressure oxidation stage, copper electro winning from cuprous chloride brine, and oxidizing the depleted solution to cupric chloride followed by recycling to the leaching stage. Although a facility was built, technical problems associated with electrolysis caused the CLEAR operation to shut down in 1982. The electrolyte overvoltage associated with high current densities was a drawback over sulfate media. In addition, the purity of produced copper was not adequate, necessitating an electrorefining stage. Everett (1994) presented the INTEC copper process. This process may be the best one in halide media. The chemistry of this process is complex but innovative. The process consists of leaching copper concentrates in an NaCl-NaBr solution (four stages) at 80-85 °C with air blowing to precipitate iron as a goethite-type compound, followed by a two-stage purification process to remove impurities and recover precious metal values. The next stage is copper reduction in a diaphragm-type electrolytic cell. The process has some novel features like purification without solvent extraction and electrowinning at high current densities. In the opinion of this author, the most promising route for a halide-based process for chalcopyrite is that based on a cupric chloride (CuCl2) system, for there are several merits in process chemistry and metal recovery. These include lower propensity toward sulfate formation, faster kinetics, better utilization of solvent extraction stages and the possibility to electrowin copper from the cuprous state. 51 3. OBJECTIVES As can be seen from the previous detailed survey of the published literature, most of the published research focused on oxidative leaching, and relatively little attention was paid to reduction leaching. In addition, most of the reduction leaching studies were conducted with reductants other than metallic iron, and few of them were aimed at developing a process flowsheet. In the case of the reductant iron, no fundamental study was found in the open literature that quantitatively describes its physical chemistry (thermodynamics and kinetics) and there is neither a systematic analysis of the use of metallic iron as a reductant in chalcopyrite leaching nor a proposed process. It is the objective of this research to perform a fundamental study on the reductive decomposition (leaching) of chalcopyrite using metallic iron. Iron as a reductant has several incentives, including low price, availability and others, as will be addressed later. The thermodynamics of chalcopyrite leaching were given earlier. The remaining sections will describe the proposed leaching mechanism prior to analyzing in depth the kinetics of the leaching reactions, to establish the best leaching conditions. The other objective of performing this detailed study is to develop a simple process flowsheet that might find commercial applications. Consequently, the results of the fundamental study will be utilized in investigating the aspects of producing an enriched copper concentrate from a chalcopyrite concentrate, as a possible alternative to other systems. A discussion on the proposed mechanism of galvanic reduction of chalcopyrite with metallic iron will be given followed by a presentation of the experimental methodology for this research. 52 4. PROPOSED LEACHING MECHANISM WITH METALLIC IRON The proposed mechanism for chalcopyrite reductive decomposition with metallic iron is : Anodic reaction : dissolution of iron, 2+ 0 Fe(s) -4 Fe (aq) + 2e" AG = +84.91 kJ/mol E° = -0.44 V (35) Cathodic reaction : reduction of chalcopyrite by dissolving iron reductively, and converting copper(I) in CuFeS2 to copper(I) in Cu2S, + 2+ 0 2CuFeS2(s) + 6H (aq) + 2e -> Cu2S(s) + 2Fe (aq) + 3H2S(g) AG = +21.53 kJ/mol E° = -0.22 V (36) Giving the net leaching reaction : + 2+ 0 2CuFeS2(s) + 6H (aq) + Fe(s) -» Cu2S(s) + 3Fe (aq) + 3H2S(g) AG = -21.53 kJ/mol CuFeS, (37) Eq. 37 will take the form of Eq. 32 or 33, depending on the lixiviant. This equation explains the effectiveness of iron as a reductant compared to copper (Eq. 24 with AG0 of -6.37 kJ/mol 0 CuFeS2) and lead (Eq. 30 with AG of only -5.21 kJ/mol CuFeS2). Also, this net reaction suggests that a major portion of chalcopyrite sulfur might be removed as hydrogen sulfide (ideally -75%). It can be said that this mechanism is based on two parts : a corrosion mechanism and a galvanic mechanism. The reaction will proceed by a corrosion mechanism, which is the dissolution of iron, and this part is the main driving force behind the leaching reactions. It will contribute to a major portion of the overall mechanism, since it is more rapid (active). This part, however, can not work alone without being augmented by the second part which is a galvanic reaction (that is the flow of electrons or current). This part is complementary to the first one, and requires the presence of protons. This proposed mechanism is important to be remembered, because it will explain many of the experimental findings in this research. The conditions under which any component of this depicted mechanism will predominate can only be determined experimentally. Galvanic contact alone, if to be the only mechanism of reduction, is not really expected to achieve complete utilization of added metallic iron, and hence complete conversion (or iron rejection from chalcopyrite). This is due to different factors, among which are : the possible loss of galvanic coupling upon progress of reaction together with the consumption of the anodic 53 reagent (iron) due to side reactions, and the hindrance to some species transport caused by the product layer. The newly formed solid phase, mainly chalcocite, is envisioned to be thick and/or dense. It is also envisioned to cover the original chalcopyrite particle. This new phase is stable and less likely to be further reduced by iron for different reasons. The experimental reduction potential of chalcocite is 0.44 V and this potential will not be reached by iron reduction until the formation of an adherent chalcocite layer which will cause ohmic drop (overpotential) between the chalcopyrite surface and bulk of solution. Once this layer develops, it will serve as a bridge for electron transport. Upon the formation of this layer, it is assumed that iron will be depleted (due to side reactions) and the transport of different species through this layer becomes rate controlling. The proposed leaching model for this mechanism is based on these assumptions, and is discussed in detail in Section 6.2. Only experimental results could confirm or refute the model. One of the possible reduction reactions for the newly formed solid phase in the presence of iron is that according to : 2+ 0 Cu2S(s) + 2FT(aq) + Fe(s) 2Cu Although the value of the free energy of change for Eq. 38 is negative, it is less likely to occur for different reasons. First, Eq. 37 is thermodynamically more favorable. Second, in the presence of hydrogen sulfide, metallic copper is easily converted back to chalcocite due to the precipitating action of H2S. In this case the possible back reactions are : 2Cu(s) + H2S(aq) Cu2S(s) + 2Ff(aq) + 2e E° = -0.3 IV (39) + Cu(s) + H2S(aq)->CuS(s) + 2H (aq) + 2e E° = -0.14 V (40) The presence of hydrogen sulfide, particularly when it is evolving vigorously, would lessen the probability of copper production in quantitative yields. From Fig. 2.3, the formation of elemental copper is favored only at low hydrogen sulfide activity or low potentials, under which Eqs. 39 and 40 will be driven to the left hand side, assuming equilibrium conditions. Third, the driving force behind the reaction in Eq. 38, AG0, is small compared to that of Eqs. 39 and 40, and thus the high reducing power of iron is less likely to be utilized. As was found by other researchers (Table 2.10), metallic copper is formed from chalcocite only by manipulating the reduction conditions in the leach solution. 54 Seemingly, the main leaching reactions are less likely to be reversible, and, consequently, back reaction kinetics are of no importance. This implies that Eq. 37 is a non-equilibrium non- catalytic solid fluid reaction, which is important to recall in the remaining sections of this work. The formation of bornite from chalcopyrite as a reduction product is fhermodynamically possible. From Fig. 2.3, there is a narrow stability region for the formation of this mineral. The kinetic considerations in terms of the structural changes that accompany the collapse of the chalcopyrite lattice, and hence the associated molar volume changes, suggest that bornite might occur as an intermediate or final product. Also, this thermodynamic figure suggests that the formation of this mineral may occur prior to the formation of chalcocite. Someone might argue that chalcocite will form, on a thermodynamic basis, rather than covellite. Also, it might be argued that there is a possibility of removing dissolved iron by the action of the released hydrogen sulfide (since the latter is a good precipitating agent). Hydrogen sulfide, as a product, is extremely effective in removing dissolved ions from solutions. For example, the removal of cuprous and cupric ions from solutions is possible by the action of H2S as evident from the equilibrium constants for the reactions : + + 26 2Cu (aq) •+ H2S(g) ~ Cu2S(s) + 2H (aq) K41 = 8.7X 10 (41) 2+ + 5 Cu (aq) + H2S(g) «->.CuS(i) + 2H (aq) K42 = 1.7 X IO' (42) The precipitation of iron sulfides is less likely to be noticed because the requirement for their nucleation is extremely difficult under the considered reducing conditions (Peters (1976)). The preferred acidity range for iron sulfide precipitation is for pH values greater than 3. Such an acidity level was not used in this research. These sulfides are sufficiently soluble in acid solutions, for instance troilite, while copper sulfides are much less soluble (Table 2.2). Hence, cuprous and cupric sulfides will preferentially be precipitated or formed as reduction products instead of iron sulfides. The other implication of the equilibrium constants in Eqs. 41 and 42 is that under reducing conditions the probability to precipitate copper from solutions as rich copper sulfides (Cu2S) is greater than covellite, explaining the disappearance of a covellite stability region from Fig. 2.3 (see Section 2.1 for the thermodynamics of chalcopyrite reductive leaching). 55 A limitation on the use of iron as a reductant is the competitive cathodic reaction of hydrogen evolution. This reaction occurs at lower overvoltages and since the reactions are in acidic media, iron is susceptible to hydrogen evolution : + 2+ 0 Fe(s) + 2H (aq) -> Fe (aq) + H2(g) AG = -84.91 kJ/mol iron (43) which will compete with the net leaching reaction (Eq. 37), causing cathodic currents to occur at potentials less than -0.4 volts, i.e. : undesired corrosion of iron. This will limit the efficiency of iron as a reductant, and more than the stoichiometric amount of iron will be needed. In this research, to lessen the possibility of hydrogen evolution reaction, different techniques were used as will be shown in the experimental part (Section 5.2.1). Before concluding this part, it should be noted that the thermodynamic values for the net leaching reaction will vary in the presence of complexing reagents, e. g. : chloride media. The presence of this halide ion will entail, for instance, the formation of iron (II) chloro-complexes, especially when its concentration is greater than 3 M. The formation of these complexes should be considered if more research is needed in this direction. For example, Eq. 35 might read : Fe(s) + Cl-(aq)-> FeCf(aq)+ 2e" E° = - 0.412 V (44) and Eq. 43 would then read : + 0 Fe(s) + Cl"(aq) + 2H (aq) -> FeCf (aq) + H2(g) AG = - 79.50 kJ/mol iron (45) based on the thermodynamic values in Table 2.3. 56 5. EXPERIMENTAL METHODS 5.1 MATERIALS A chalcopyrite concentrate was obtained from the Gibraltar Mine, McLeese Lake, B. C., Canada. The detailed chemical analysis of the concentrate is given Table 5.1. Element Mass, % Copper 28.3 Iron 28.0 Sulfur (%)total 30.0 Sulfur, S2- (%) 32.0 Molybdenum 0.4 Insoluble (%)* 10.7 Element Mass, % Element Mass, % Aluminum 0.0253 Magnesium 0.0165 Barium 0.0011 Phosphorous 0.2001 Bismuth 0.0255 Sodium 0.0024 Calcium 0.0266 Tungsten 0.0022 Cobalt 0.0024 Zinc 0.0121 Lead 0.0232 * as siliceous gangue Table 5.1 : Detailed chemical analysis of the tested chalcopyrite concentrate This chemical analysis was performed using ICP (Inductively Coupled Plasma) method. The mineralogical composition of the concentrate is given in the following table : Mineral Content, mass % Molecular weight Chalcopyrite (CuFeS2) 63.8 183.513 Pyrite (FeS2) 17.0 119.967 Chalcanthite (CuS04.5H20) 9.3 249.677 Siliceous gangue and other refractory oxides 9.9 N/E* *N/E : not estimated Table 5.2 : The mineralogical composition of the tested chalcopyrite concentrate 57 This composition was obtained using X-ray diffraction (XRD) and verified by the Rietveld method (O'Connor et al (1992)). The concentrate was ground to P100 -325 mesh +400 mesh (-44 urn +38 um) and prepared as follows : the bulk concentrate was first split by coning and quartering. Then a sample weighing around 35 kg was split again using a riffle to obtain a representative sample for the fundamental study. The obtained sample, around 2 kg, was first wet screened to remove extremely fine particles which could bias initial leaching data, followed by rinsing with acetone and/or ethanol to allow fast drying. The remaining sample was then subjected to a careful dry screening into discrete size fractions (monosizing) to find the particle size distribution of the sub-samples to be used in the experiments. The results of this dry sieve analysis are given in the following table : Mesh range, # Size range, pm Mass % Cumulative % finer - 80 + 100 -180 + 149 10.49 100.00 -100 + 115 -149 + 125 18.48 89.50 -115 + 170 -125 + 90 34.86 71.02 -170 +200 - 90 + 74 14.53 36.16 -200 +240 - 74 + 63 8.42 21.63 -240 +270 - 63 + 53 5.88 13.21 -270 +325 -53+44 4.18 7.33 -325 + 400 - 44 + 38 2.58 3.15 -400 - 38 0.57 0.57 Total 100.00 Table 5.3 : Particle size distribution of the concentrate, using Tyler standard screen scale As can be seen from this table, the majority of the particles are in the size fraction -100 mesh +200 mesh (-149 urn +74 pm), hence most of the experiments were done using particles that have a P100 of this fraction. Again, a sample of this fraction was analyzed by the digestion procedure (Bennewith and Hackl (1998)) to compare the chemical composition with that of the bulk concentrate. The results were consistent, although some chalcanthite was washed off. The same procedure was repeated without wet screening to avoid the loss of chalcanthite. The results were again consistent. All the calculations were based on a concentrate containing 63.8% chalcopyrite, 28.3% copper and 28.0% iron. All samples were kept in tightly sealed plastic bottles under an inert atmosphere of nitrogen, to limit any surface oxidation of the mineral. 58 5.2. METHODS 5.2.1 THE KINETIC STUDY All the chemical reagents were of analytical grade, and were used as received. Analytical grade iron used in these experiments was in powdery form (-600 mesh) and produced by electrolytic reduction. Deionized water (DIW) was used in the experiments which were performed under atmospheric pressure. The required amount of the leach solution (in most cases 250 ml) was first placed in a 1- liter reaction vessel, fitted with a pH-probe and stirred continuously for at least 15 minutes to allow the pH reading to stabilize. The reaction vessel was always connected to an autotitrator instrument (Radiometer TTA 80 titrator, Radiometer Copenhagen, Denmark). Stirring was provided using a 3-blade impeller (diameter is 2.1 cm). Stirring was usually at 250 rpm and the agitation speed was monitored by the use of a tachometer. The reaction vessel was maintained at constant temperature (25 °C ± 0.01 unless otherwise specified) by being immersed in a circulating heating water bath fitted with a thermocouple. The temperature was monitored by a digital display and/or a mercury thermometer. The solid reactants (iron and the concentrate) were premixed in a suitable weighing pan by a spatula for at least 10 minutes to allow uniform distribution of particles, and promote coupling of solids (intimate mixing), thus decreasing the possibility of side reactions. Once the pH reading stabilized, the solids were added in one batch, and the reaction was monitored by keeping a constant pH reading throughout the reaction period. pH was always kept constant by slow titration with a solution that has a concentration of at least ten times that used in leaching. This is also important to maintain an almost constant solid pulp density. As soon as the solid mixture was introduced to the reaction vessel, the particles appeared to stick together at the bottom of the reaction vessel. As the reaction proceeded, solid particles started to float as if being joined (agglomerated) together and signs of the reaction were noticed (release of hydrogen sulfide and change of pH reading). Gas bubbles were clearly indicating the reaction was progressing. Agglomerated particles were noticed to float to the solution surface, prior to sinking again in the reaction mixture, and this continued for about 30 minutes. Flotation 59 of agglomerated particles was less pronounced in chloride media. The odor of hydrogen sulfide was always detected inside the fume hood, but virtually disappeared after 60 minutes. Sampling was done by taking 1-ml liquid samples from the reaction vessel at 10-min intervals. Because of the agitation, a slurry sample greater than 1-ml was first taken, allowed to settle for a while, then entrained solids, if any, were rejected back into the reaction vessel. The 1-ml sample was then diluted in a 10-ml centrifuging tube, and centrifuged for 10 minutes, prior to final dilution to the required volume. Centrifuging was done to ensure there were no entrained fine particles in diluted solutions, which might cause clogging in the atomic absorption instrument. After centrifuging, careful visual inspection revealed no solids in the tubes. Added acid was recorded whenever a sample was taken. After one hour of reaction time (see below), sampling was done every 30 minutes. Total allowed reaction time was 3 hours. At the conclusion of the experiment, the reaction mixture was vacuum filtered, and the filtrate volume was measured. The filtrate was always noticed to have a greenish color and in most experiments its volume was around 240 ml. A sample of the filtrate was also taken and analyzed as described below. The residue was rinsed with deionized water, and again a sample of the wash solution was taken and analyzed. Any amount of dissolved iron reported in the wash solution was included in material balance calculations. The rinsed residue was dried in an oven kept at 50 °C, to prevent any oxidation of sulfide sulfur. After drying, the residue was weighed and a sample was taken and analyzed by the digestion procedure. Detailed chemical analyses (material balance calculations) were done to study and confirm the reaction stoichiometry. Fig. 5.1 is a schematic drawing of the reaction vessel and the auxiliary equipment. To decrease the possibility of hydrogen evolution, different techniques were used such as : 1) Premixing of solids prior to addition to the reaction vessel 2) Slow titration in keeping constant pH. This also helped in avoiding any perturbation to reaction kinetics. 3) Running at high solid pulp density 60 Agitator Batch addition of premixed Fe- CnFeS-) solids acid addition pH meter Autoburette Digital TT80 titrator OCT Reaction vessel with solid mixture Fig. 5.1 : Schematic diagram of the reaction vessel during the kinetic study Nitrogen purging pH meter Rubber Acid addition stopper Auto burette Control unit Purge NaOH solution Reaction vessel with beaker fvented to aifl solid mixture (closed"* Fig. 5.2 : Schematic diagram of the reaction vessel during the process study 4) Purging with nitrogen gas. This technique was very helpful in the process study where the experiments were run at high solid pulp density. In addition, any entrained amounts of oxygen or air were removed, and reaction gases were easily expelled. 5.2.2 THE PROCESS STUDY The experiments performed in the process study were based on the results obtained in the previous section. The same materials were used, except high SPD (-35%) was used and wet screening was avoided. The leach solution in this part was 0.1 M HC1. 250 ml of this solution was placed in a 1-liter reaction vessel. To this solution, sufficient amount of analytical grade ferrous chloride tetrahydrate was added to produce a solution of 3 or 4 M ferrous chloride. The reaction vessel was sealed tightly with a rubber stopper, through which a pH-probe, a delivery-tip for titration, a plastic tube for purging and a stirrer were fitted. The leach solution was purged with a stream of nitrogen gas for at least 20 minutes. This purging was very necessary to remove any entrained air or oxygen. Under the experimental conditions, these entrained gases were found to severely oxidize the ferrous ions in solution and form ferric hydroxide. Precipitation of iron species was noticed to have a negative effect on leaching. At the same time, the solution was continuously stirred till the pH reading stabilized. Once the pH reading stabilized, the premixed solid mixture was added through a special hole, then plugged again. The reaction generates large amounts of H2S and other gases. These gases were collected and discharged against atmospheric pressure through a receiving tube, immersed in a 2 M NaOH solution, to absorb these gases and convert them into sulfides as sodium sulfide. The reaction assembly is shown in Fig. 5.2. The reaction vessel was also immersed in a circulating heating water bath kept at constant temperature. The reaction mixture temperature was monitored by the same method described above. The reaction was allowed to take place for 3 hours, with no samples being taken. Once the reaction time elapsed, the leach solution was vacuum filtered, to produce an almost dry residue. The filtrate was always observed to have a dark greenish color, and its volume was greater than 250 ml, since a considerable amount of titrating acid was added. Nonetheless, the re- 62 estimated SPD based on final solid weight and filtrate volume was always between 28-32%. A sample of the filtrate was taken and diluted for copper analysis by AAS. The residue was rinsed several times with DIW to remove any dissolved iron species, which would appear as iron precipitates upon drying. Wash solutions were collected and analyzed by the same procedure, and any amounts of dissolved copper were included in chemical analysis. The residue was dried at 50 °C and then a sample was taken and digested for copper and iron analyses. The residue from both studies, before being filtered, was observed to be of sintered or compact appearance, with a dark black color. Yellowish particles (pyrite) could be seen clearly, as observed in the fresh concentrate. These residues were analyzed for metallic iron by the use of a magnet and wet chemistry methods (Young 1971). Both methods showed that no metallic iron remained in the residue. In some cases, after 60 minutes, 120 and even 150 minutes, some metallic iron could still be detected by a strong magnet. To avoid any implications on kinetic analysis, a reaction time of three hours was used for all experiments performed in this research. On the other hand, random experimental samples of the generated residue were rinsed and vacuum filtered with organic solvents (alcohols and carbon disulfide) to test for elemental sulfur formation. The resulting filtrate was carefully vaporized to quantify any elemental sulfur, but only trace amounts were found, hence the main reaction products are those as per Eqs. 32 and 33. Different organic solvents were used because only rhombic and monoclinic sulfur are soluble in CS2, while amorphous sulfur is not. So, elemental sulfur was not detected. The generated solution from H2S purging is rich in sulfides, which can not be discarded directly without treatment. These solutions were oxidized with hydrogen peroxide to destroy the sulfides prior to discharge. As a safety precaution, some of the methods used in this research are potentially hazardous, and there are dangerous materials used and some of them are extremely toxic (H2S and bromine). So, it is recommended that those who do further study familiarize themselves with all safety aspects of the wet chemistry methods, and understand the procedures well. The digestion procedure can be found in Bennewith and Hackl (1998), and can be used for any sulfide or oxide mineral, except silica and similar minerals. 63 5.2.3 ANALYSIS TECHNIQUES Analysis of samples was done by atomic absorption spectroscopy (AAS) using a Unicam 929 AA spectrometer (Unicam Limited, Oxford, UK). In the kinetic study (low SPD), analysis of the leach samples was done for iron alone, as solution and chemical analyses in the early experiments show that little or no copper was dissolving from chalcopyrite. However, filtrate samples were analyzed for both copper and iron. For the process study (high SPD experiments) and its material balance calculations, filtrate samples were analyzed for copper only, since the amount of dissolved iron in the filtrate is quite high, and AAS will be inaccurate. Residue samples were analyzed for copper, iron and sulfur. For the residue digestion analysis and to avoid any interference, the standard solutions were subjected to the same conditions as the residue digestion solution (additions of aqua regia, bromine water and others), to obtain matrix matching for the samples and the standards. Chemical analysis show that there is a 10-20% discrepancy in results due to these interferences, caused by the effect of these analytical reagents and their chemical composition. Details of AAS analysis can be found in Harris (1991). 5.2.4 REACTION PRODUCT CHARACTERIZATION XRD was not extensively used in the identification or quantification of reduction solid products, rather, a detailed chemical analysis based on wet chemistry methods was done, as well as some SEM/EDX analysis. Chemical analysis by wet chemistry methods proved to be accurate and suitable for reaction product characterization. The results were found to be in agreement with the theoretical leaching reactions (Eqs. 32 and 33) and EDX analysis. Further, the results were considered to be satisfactory on the basis of the conformance of the leaching data with the filtrate and residue analysis. Sample material balance calculations are given in Appendix II. The reader is directed to refer to this appendix for detailed discussion on this and other related topics. Sample SEM/EDX results are given in Section 6.2 to demonstrate the validity of the leaching model. In the process study, it was assumed and found valid that the filtrate would be copper- free. In most cases, dissolved copper in solution was very small, although the concentrate 64 contains some chalcanthite (CuS04.5H20), which would release copper in solution by solvation. At high SPD and under the employed reducing conditions, with the action of H2S, most of the concentrate copper reported in the residue as Cu2S (major) and Cu (minor and assumed to be produced by the action of cementation). Chemical analysis showed no dissolved copper in solution for high SPD experiments. The gaseous stream was not analyzed to determine its composition or to base material balance calculations on it. Such calculations require the use of gas chromatography to find the exact composition of the gas stream, which was not done in this research (but may be a suitable recommendation for future work). 5.3 CALCULATIONS For the leaching experiments, the amounts of reactants used in the experiments are shown with every set of experimental data. The estimation of conversion from AAS data was based on the assumption that dissolved iron came only from the chalcopyrite and added powdery iron. The remaining iron bearing minerals; pyrite and other iron-bearing oxides (if any), were assumed not to contribute any dissolved iron in solution. Thermodynamically, pyrite can be decomposed reductively, but on a kinetic basis the rate of leaching of this mineral under reducing conditions is very slow, as was confirmed by Peters and Majima (1968). Hence, other possible sources for dissolved iron from the concentrate are minor and can be safely neglected. The concentration of ferrous ions in solution is a direct measure of the extent of the reduction reactions, since these ions are released from the crystal lattice of chalcopyrite. In turn, this will be a good tool to monitor the extent of reaction, depending on material balance calculations. The details of the methodology of conversion estimation is outlined in Appendix II. The reaction analyses showed that leaching continued for first 60 minutes and almost ceased later. Hence, all analyses were done solely for the first hour of reaction time. By accounting for dissolved iron from the metallic component, and by the assumption that this added metal will eventually dissolve by the effect of hydrogen evolution reaction, the iron balance is now established. Iron analysis in the filtrate showed that it is the sum of added metal and that released from chalcopyrite. 65 Early analysis of samples done with and without solid premixing, and with and without slow titration, showed that premixing and slow titration are more effective in achieving better conversion data, and more reliable in the analyses done in this research. The estimated conversion by the outlined methods in Appendix II is designated as XB or Xb, and will be used throughout the remaining sections of this thesis, unless otherwise specified. The reader is encouraged to refer, as appropriate, to the Appendices for further information. 66 6. RESULTS AND DISCUSSION 6.1 ANALYSIS OF REACTION KINETICS The rate at which the alteration reaction proceeds is an important parameter in analyzing the leaching data obtained from experimental work, in that longer leach times mean increased equipment and operating costs upon scaling-up. Consequently, the analysis of reaction kinetics is an important step toward understanding the nature and the mechanism of the leaching reactions, and then seeking the parameters that would favor the desired reactions and yield the maximum conversion. Eventually, the selection of a leaching model will depend on these findings. To do this, initial tests were run to observe the leaching behavior under some thermodynamic and chemical conditions. Next, such data were fitted to different leaching models to see which sort of control mechanism this system would likely follow. Upon deciding on the type of mechanism, other studies were done to confirm the control mechanism, and find the general leaching model. The proposed mechanism for this system has already been presented in Section 4, and the only way to prove or refute it is by a careful and systematic kinetic analysis. The selection of a leaching model to fit the experimental data is not an easy task. It requires different approaches prior to deciding on the model. The known models for solid-fluid reactions are reviewed in Appendix I. In the kinetic analysis of leaching data, several leaching models were tried to get a reasonable fit to the data. The leaching kinetics were assumed to be controlled by one mechanism, namely, control by a surface chemical reaction, control by transport through a product layer, or control by transport through a boundary fluid film, and for both changing and unchanging size particles. A combination of these controlling mechanisms was not used, since one of them proved to be satisfactory. The criteria for reaching a satisfactory data fit was based on different principles. Kinetically, the fit should show an acceptable level of dependency. For instance, it is not acceptable to say that a system is under product layer diffusion control, and its activation energy has a large value (say 100 kJ/mol or more). High values of activation energy are normally associated with chemically controlled systems. 67 The estimation of enthalpy of activation should be reasonably near to that of energy of activation, since most leaching systems will have a slight difference in the values of these two quantities. The difference is attributed to experimental errors and the methodology used in estimating these quantities (Levenspiel (1972)). A chemically controlled system should show a clear dependence on temperature, that is small increments in temperature, 5 to 10 degrees, result in significant improvement in reaction rates. These can finally be supported from testing the dependence of reaction rate on particle size, as explained in Appendix I. Statistically, linear regression fitting of the leaching data should have a value of r2 greater than 0.96 to be acceptable, although some authors gave a value of 0.80 to be satisfactory (Carnahan et al (1969)). r2, the coefficient of determination, signifies the improvement or error reduction due to the straight-line model. For r2 value of 1, the obtained line is a perfect fit and explains 100% of the variability. For r2 value of 0, the fit represents no improvement. It should be emphasized that, even with all these considerations, there might be some systems with irregularities, and even the experimental data themselves might be misleading or give a pseudo-real representation of the actual situation. The only way to overcome such difficult situations is to repeat the experiments with different laboratory techniques, and analyze them with different approaches. For the systems studied in this research, the collected leaching data were fitted using different methods. The first quantitative assessment was done using the methods of Wen (1968). In these methods, the leaching data were first fitted by Eqs. 1.27 and 1.29 (Appendix I). A sample of the data is given in Tables 6.1 and 6.2, and the fitting results are given in Figs. 6.1 through 6.9. In Fig. 6.2, both the conversion and amount of acid consumed (by titration) were plotted together, versus time, to show that the leaching process is in conformance with reaction stoichiometry. The data in Table 6.1 were manipulated according to Eqs 1.27 and 1.29, to assess the closest approximation to the controlling mechanism. From Fig. 6.3, the slope of the linear fitting is around 2, with r2 value greater than 0.96. Hence, it was concluded that the controlling mechanism is a diffusion process through the product layer. Then, the same data in Table 6.1 were fitted by all the known forms of leaching models (Figs. 6.4-6.7). It becomes apparent that 68 all the models are not satisfactory, except for the product layer model. This finding supports the proposed mechanism and the formation of a product layer. Experimental objective : To demonstrate Wen's method Experimental conditions Temperature 25 °C Agitation speed 250 rpm SPD 2.18% Iron added as % stoichiometric 100 Initial solution composition 0.1 MH2S04 Solution pH 1.08 Concentrate type Gibraltar chalcopyrite concentrate Particle size •149 um +74 pm Experimental results : m + 689.5 Maximum [Fe ]re,eased> PP Amount of iron used in CuFeS2 reduction, ppm 344.7 Overall efficiency of iron 0.185 Amount of iron used in H2 production, ppm 1516.7 Efficiency of H2 production 0.814 Time, min Acid consumed [F^2 Itotab Fractional [F^ ] released' Conversion (3M H2S04), ml ppm completion ppm 0 0 0 0.00 0.0 0.0 10 25.45 520 0.20 140.5 0.037 20 37.92 1125 0.44 304.1 0.081 30 46.85 1445 0.56 390.6 0.104 40 50.14 1730 0.67 467.6 0.125 50 53.84 1960 0.76 529.8 0.142 60 56.89 2100 0.82 567.6 0.152 75 58.35 2200 0.86 594.6 0.159 90 61.12 2300 0.90 621.7 0.167 105 64.47 2400 0.94 648.7 0.174 120 66.66 2400 0.94 648.7 0.174 135 68.72 2500 0.98 675.7 0.181 150 68.72 2500 0.98 675.7 0.181 165 68.72 2551 1.00 689.5 0.185 180 68.72 2551 1.00 689.5 0.185 Filtrate 68.72 2551 1.00 689.5 0.185 Table 6.1 : Sample experimental leaching data for selecting a leaching model by Wen's method (sulfate media, stoichiometric run, 25 °C) 69 From Fig. 6.3, the estimated leaching constant has the same order of magnitude compared to that estimated from the parabolic leaching model fitting (Fig. 6.7), and this, again, would be an evidence for the validity of the suggested mechanisms (Wen (1968)). The same procedure was repeated for chloride media, and the results are summarized in Table 6.2 and Figs. 6.8 and 6.9. More verification for the selected leaching model will be presented in the next sections. Experimental objective : To demonstrate Wen's method Experimental conditions Temperature 65 °C Agitation speed 250 rpm SPD 1.11% Iron added as % stoichiometric 100 Initial solution composition 0.1MHC1 Solution pH 0.82 Concentrate type Gibraltar chalcopyrite concentrate Particle size -149 um+74 um Experimental results : 2+ Maximum [Fe ]released, ppm 930.4 Amount of iron used in CuFeS2 reduction, ppm 465.2 Overall efficiency of iron 0.470 Amount of iron used in H2 production, ppm 524.4 Efficiency of H2 production 0.529 Time, min Acid consumed [^q2 ] total' Fractional P*"® ] released' Conversion (3M HC1), ml ppm completion ppm 0 0 0 0.00 0.0 0.0 5 15.84 550 0.28 266.5 0.134 10 21.26 700 0.36 339.2 0.171 15 35.05 880 0.45 426.4 0.215 20 42.45 1000 0.52 484.6 0.244 25 45.04 1176 0.61 569.9 0.287 30 50.82 1350 0.70 654.2 0.330 40 54.63 1550 0.80 751.1 0.379 50 54.63 1680 0.87 814.1 0.411 60 55.21 1800 0.93 872.3 0.440 90 56.32 1880 0.97 911.0 0.460 120 56.32 1893 0.98 917.3 0.463 150 56.32 1920 1.00 930.4 0.470 180 56.32 1920 1.00 930.4 0.470 Filtrate 56.32 1920 1.00 930.4 0.470 Table 6.2 : Sample experimental leaching data for selecting a leaching model by Wen's method (chloride media, stoichiometric run, 65 °C) 70 For the leaching curves presented in Figs. 6.1 and 6.8 (and later all leaching curves under a variety of conditions), it can be seen that rapid reaction rates occur in the first stage of leaching (the first 60 minutes of total reaction time), before leveling off or slowing dramatically at a certain value. Beyond this, little conversion is obtained, and in some cases even after two hours of leaching, conversion is barely improved. For systems under surface chemical reaction control, reaction rates are expected to be slow in the early stages because of the slow reaction, and will continue to progress with time, as long as sufficient reactants are supplied. For reactions under product layer control, the reverse applies, and no significant improvement in conversion is expected with prolonged leaching times. Since rapid kinetics occur in the initial stages, this means that as soon as the reactants reach the active reaction sites, chemical reaction will take place without any hindrance, releasing the products as per Eq. 37. If a surface chemical reaction is controlling, it would take a while for such products to form. Also, the electrochemical nature of the reactions supports the theory that a mechanism other than a chemical reaction mechanism is rate limiting, although there are some exceptions, as can be seen from Section 2.2. According to Wen (1968) for the diffusion control to be important, the reaction linear velocity, expressed as the rate of change of particle size, should be greater than 1X10"4 - 1X10"3 pm/s. Comparison of this value to any of the obtained leaching data confirms that this statement applies here (see Section 6.1.8 for more discussion). Seemingly, the surface chemical reaction model was not satisfactory for fitting the experimental data. Also, the control model for diffusion through a boundary film was not selected. Earlier, it was assumed that the formed product layer is thick and/or dense. According to Levenspiel (1972), when a thick solid product forms, the resistance to fluid transport through this product layer is usually much greater than through the fluid film surrounding the particle. Hence, in the presence of such a product layer, fluid film resistance can safely be ignored, and product layer resistance is unaffected by changes in fluid velocity (that is the rate of agitation in batch systems). In Section 6.1.1, it is shown that these statements are applicable to the studied systems in this research. Seemingly, the morphology of the new solid phase is as assumed. 71 o:2o -, 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.1 : Plot of conversion vs. time for demonstrating Wen's method (sulfate media, stoichiometric run, 25 °C). The data are as per Table 6.1. Time, min Fig. 6.2 : Plot of acid consumption and conversion vs. time (sulfate media, stoichiometric run, 25 °C). The data are as per Table 6.1. 72 4.500 4.000 - -2.6 -2.9 -3.1 -3.4 -3.6 -3.9 -4.1 -4.4 -4.6 In (1-(1-Xb)1/3) 1/3 Fig. 6.3 : Plot of In t vs. In (1-(1-Xb) ) as per Wen's method (sulfate media, stoichiometric run, 25 °C). The figure was plotted based on the data in Table 6.1. 0.18 Time, min Fig. 6.4 : Plot of fluid film diffusion control model fitting of conversion vs. time (Table 6.1, unchanging size particles, sulfate media, stoichiometric run, 25 °C) 73 0 10 20 30 40 50 60 Time, min Fig. 6.5 : Plot of fluid film diffusion control model fitting of conversion vs. time (Table 6.1, changing size particles, sulfate media, stoichiometric run, 25 °C) 0 10 20 30 40 50 60 TirrE,rrin Fig. 6.6 : Plot of chemical control model fitting of conversion vs. time (Table 6.1, changing and unchanging size particles, sulfate media, stoichiometric run, 25 °C) 74 0.009 4 Ash control 0.008 - (unchanging size particle) 0.007 - 2/3 l-3(l-Xb) +2(l-Xb) = 0.0001371 0.006 - ^ = 0.98 0.005 - + CS X) 0.004 - X 1 1 -3 ( 0.003 - 0.002 - o.oo r . 0.000 < 10 20 40 50 60 Fig. 6.7 : Plot of product layer control model fitting of conversion vs. time (Table 6.1, unchanging size particles, sulfate media, stoichiometric run, 25 °C) Fig. 6.8 : Plot of conversion vs. time for demonstrating Wen's method (chloride media, stoichiometric run, 65 °C). The data are as per Table 6.2. 75 0.09 0 10 20 30 40 50 60 Time, min Fig. 6.9 : Plot of product layer control model fitting of conversion vs. time (Table 6.2, unchanging size particles, chloride media, stoichiometric run, 65 °C) The figures presented so far show that incomplete conversion of chalcopyrite is occurring, and there are different possible reasons for this trend. The incomplete conversion of chalcopyrite may be attributed to : 1) Side reactions, especially the hydrogen evolution reaction that will lower available iron for reduction 2) Formation of dense product layers 3) Lack of uniform distribution during premixing of chalcopyrite and iron, resulting in isolated areas where, probably, the cathodic reaction is incomplete. For the premixing itself, nonetheless, improvement in conversion by this technique was realized. For instance, the conversion for a stoichiometric run using 0.14 M H2S04 (Table 6.11) was 18.52% without premixing, increased to 28.35% with premixing, which is an advantage. As will be shown later, the formation of a product layer accounts for the parabolic passivation shown in all conversion-time curves, and similar findings by other authors (Table 2.10) support this hypothesis. 76 The first 10 minutes of reaction time during this study were the most difficult to track and perform sampling. In this period, the reaction mixture was full of gas bubbles and the solids were combining together, floating on surface. For these reasons, the data collected during this period look to be out of place on the leaching curves, and later on model selection. It is certain that the metallic iron is almost completely dissolved after one hour, as reaction kinetics are not changing after this stage. This supports the arguments given in Section 4 in that the corrosion mechanism is more active and the main driving force behind chalcopyrite reduction. For this reason, and since the conversion is not significantly improving beyond this stage, leaching kinetics were analyzed only for the first hour. Having these findings, all the data in the kinetic study were fitted using the ash control model, and many of the fits were with acceptable accuracy. Any apparent deviation of the results using the product layer control model is attributed to experimental errors. As can be seen from Table 2.10, there are at least five parameters that have a direct effect on leaching kinetics, and consequently, it is expected to see a leaching model in which the rate dependence on such variables might be incorporated. The possible variables are : temperature, particle size, hydrogen ion concentration, metallic iron amount, chalcopyrite amount, solid pulp density, ferrous ion additions and others. The effect of catalytic additives, like lead or other possible metallic reductants, was not studied, but can be a recommendation for future research. Since the parabolic leaching model was chosen, it can be modified to include some parameters, that is Eq. 1.10 (Appendix I) becomes : 2/3 l-3(l-XB) +2(l-XB) = kt (46) and k, the parabolic leaching rate constant, is a function of different parameters : + k = f(T,d0,[H ],etc....) (47) Whenever any of these variables are changed, a new value for k is obtained, and, consequently, a new value for the diffusion coefficient will be obtained. No trial was made to estimate the latter from experimental kinetic data, but it might be of the order 1010 m2/s. The constant k will represent the velocity of changing or movement of particle-particle or particle-fluid interface and can be expressed in terms of the Arrhenius form to read : 77 [H+]a exp (48) KRTJ 2 and the constant k0 will include all the constants in Eq. 1.11 except R . The leaching model can now be rewritten as : 2/3 + l-3(l-XB) +2(l-XB) = ^[H f exp t (49) \KYJ k0 in this research is called the intrinsic parabolic leaching rate constant. The reaction kinetics will be analyzed in the next sections, and the estimated thermodynamic parameters should be in conformance with the selected leaching model. The experiments were performed with different reactant amounts as shown with every set of experimental data. Unless otherwise specified, leaching was always kept at room temperature and atmospheric pressure. The allowed reaction time was 3 hours. 78 6.1.1 EFFECT OF AGITATION The need for galvanic coupling is required to the extent needed to complete the two components of the leaching reactions. Recalling the proposed mechanism, the corrosion part is clearly dependent on the rate of agitation, while for the galvanic part such a dependence is less necessary in completing the reaction, as was noted with other reductants (like copper). As a result, agitation would be required to provide good mixing of the solids. This is augmented by other features of using iron, specifically, the high potential difference (the wide gap between E° of iron and E° of chalcopyrite aqueous reduction), as a driving force to complete the reduction reactions. Hence, it can be concluded that agitation would not impede or have a negative effect on leaching kinetics. On the other hand, the density difference between the solid particles (Table 2.7), tends to let iron particles settle down in the reaction vessel, without any extra chance of coupling with chalcopyrite particles, which would result in poor kinetics. As can be seen from the literature survey, there are contradictory findings on agitation effects by several researchers, and such effects cannot be deduced only from theory, without experimental evidence. To verify these points, experiments were done under agitating and non agitating conditions to understand the effect of agitation on conversion. These experiments were done in both sulfate and chloride media. Under non-agitated conditions, the stationary mixture displayed signs of reaction. Evolution of hydrogen sulfide gas being the diagnostic feature, it was easy to observe its formation visually and detect its pungent odor. Tests with agitation employing similar conditions resulted in the same feature. Also, aggregates of particles were observed with both stirred and non-stirred experiments. This discovery is a justification for the proposed mechanism. First it reflects the need for electrical or particle to particle contact as a condition for galvanic action. Second, it indicates that varying the agitation speed, more or less, has no significant effect on the leaching process. This agrees well with the findings by Nicol (1975) as was discussed in Section 2.4. Tables 6.3 and 6.4, and Figs. 6.10 and 6.11 summarize the results for agitation speed effect on the reaction kinetics. It is clear from these figures that the reaction rates are relatively insensitive to agitation speed. The effect of agitation is more pronounced in chloride media, probably due to experimental errors in sample collection and analysis, rather than a real dependence of leaching 79 CJ © CJ 'B 00 a T-H o o ON CO CO CO CO in CJ o co 00 CO © CO m VO r-- 00 ON I & o T-H CN CN CO CO CO CO CO CO CO © © T3 m ©' © © © ©• © © © © © a ff e ^° c3 u ,•—s o CN ©OO x> a a o o> 00 VO in Tt Tt t-- m T-H o ON o CN 00 CN in 00 ON ON © 1 CM CN O & T— T—' VO 1 Tt o CN CN CO CO CO CO CO CO CO >> o, © © © © © © © © © © © © CJ 3to '£ o CJ fl H o co o m iso d fl© Tt CN CN CN © ON ON 00 © CN r-- CN ON © m tio n tiv e ich i rv T—H T-H VO CO VO o CN CN CN CO CO CO CO CO CO o B S & © © © © © © © ©• © © CO fl o o o fl oCJ .n O X=> o O ox fl 1/3 GO *e*->3 ~3 _o +•» cO %-> s <—1 o o © © © © © © © © © T-H fl fl 3 H-» CN ON CN in oo CO VO T-H T—H 1—I np H CD CJ de d fl fl g 1 "o CJ cu o *cs +-> cO CJ cO <-> >H pe r pe r Q fl +-> 1 O 1 .1 X CJ CH Xfl 'fl CJ w w H JH HH GO Ha lu 1 Cl, X w (L> a 1 o m 00 CO in 00 ON in CN CN rv © ON T-H VO T-H CO in VO t> CJ o& © © CN CN CO CO CO CO CO CO CO T3 © © © © © ©' © © © © © CJ CJ in © fl CO O o 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.10 : Plot of conversion vs. time at various agitation speeds for the data in Table 6.3 (sulfate media, stoichiometric runs, 25 °C) 0.45 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.11 : Plot of conversion vs. time at various agitation speeds for the data in Table 6.4 (chloride media, stoichiometric runs, 25 °C) 81 kinetics on agitation rate. It was noticed that near an agitation speed of 250 rpm, the reaction mixture is well-mixed and the achieved conversion is comparable with all other experiments. No solids were noticed to settle down on the bottom of reaction vessel. According to the proposed leaching mechanism, there are different species, removal of which from reaction sites or transport toward these sites may facilitate reaction rates. Consequently, agitation at 250 rpm is satisfactory and justified for two purposes. First to provide good mixing in the reaction vessel, and second to enhance the transport of protons, sulfide or ferrous ions to and from reaction sites, respectively. The other benefit of mild agitation is to achieve improved particle to particle contact, and reproduce such contacts or increase their numbers. The result was that the reductant and chalcopyrite are in intermittent short-term contact, leading to continuous reaction. The corrosion mechanism is more important and active than the coupling mechanism, and so, mild agitation is desired to the extent sufficient to nearly suspend the solid particles and provide well-mixed solids. Later it was noticed that mild agitation would cause some sort of abrasion to the agglomerated solid particles, exposing more fresh surfaces of chalcopyrite to the lixiviant and the reductant. This in turn improved the integrity in interparticle welds and was of benefit for completing the leaching reactions. For the remaining experiments in this research good leaching rates were achieved by mild agitation and in the experiments performed at high solid pulp density (SPD), this phenomenon was quite useful in achieving even better results. The finding that reaction kinetics are independent of the rate of agitation is in conformance with the discussion given in Section 6.1 regarding the significant influence of product layer resistance on reaction rates compared to that of other resistances. Based on these results, agitation was used throughout this research, and in the kinetic study, where low SPD was used, an agitation speed of 250 rpm for both sulfate and chloride media was utilized. For the process study, the stirring speed needed to provide good mixing of solids was 610 rpm at -35% SPD. At such rates, agitation emerged as a necessity for good conversion under mild conditions. 82 6.1.2 EFFECT OF TEMPERATURE Temperature dependence is the most frequently analyzed parameter in studying reaction kinetics. Most researchers prefer to first analyze the temperature effect on reaction kinetics as this would allow better understanding of leaching reactions. The other benefit from this analysis is to estimate the related thermodynamic values like the activation energy which would explain the behavior of the system in terms of controlling mechanism (Appendix I). As presented in Table 2.10, most chalcopyrite reductive decomposition reactions are temperature dependent and are usually preferred under high temperatures or in autoclaves. Autoclaving of chalcopyrite using iron as a reductant was not studied in this research. The temperature range studied here was 25-85 °C. Experiments were performed to understand the effect of temperature on leaching kinetics, and the results are summarized in Tables 6.5 and 6.6, and Figs. 6.12 and 6.13. From these figures, an important feature appears. The reaction rates under the experimental conditions, and in both chloride and sulfate media, seems to be less sensitive to temperature variations. For sulfate and chloride media, up to 65 °C reaction kinetics are gradually improved with increasing temperature. Beyond 65 °C, no significant improvement is noticed, and even a drop in final conversion was noticed for some runs. The smaller sensitivity toward temperature variations, as explained in Appendix I, means that the reaction may not be under some sort of fluid film diffusion or chemical reaction control. It can be seen from the trend of the leaching curves in Figs. 6.12 and 6.13 that rapid and comparable reaction rates are occurring in the early stages. Hence, it can be deduced that these two types of mechanism are not rate controlling. Smaller sensitivity toward temperature variations is common for reactions controlled by a transport process in the product layer, which supports the findings obtained in Section 6.1. Shirts and his coworkers (1974) found that leaching in both systems is preferred at 65 degree Celsius. They reported a 40% increase in conversion when the temperature was increased from 40 to 60 °C. The authors did not explain their findings, but a rich chalcopyrite concentrate was used and all their experiments were run under non-stoichiometric conditions with agitation. 83 Experimental objective : To study the effect of temperature Experimental conditions Agitation speed 250 rpm SPD 1.82% Iron added as % stoichiometric 100 Initial solution composition 0.1MH2SO4 Solution pH 0.97 Concentrate type Gibraltar chalcopyrite concentrate Particle size -149 um +74 urn Experimental results : Conversion at temperature of Time, min 25 °C 35 °C 45 °C 55 °C 65 °C 75 °C 85 °C 0 0.000 0.000 0.000 0.000 0.000 0.000 0.000 10 0.046 0.062 0.084 0.104 0.129 0.071 0.061 20 0.086 0.107 0.145 0.166 0.187 0.123 0.103 30 0.115 0.132 0.179 0.212 0.233 0.154 0.126 40 0.132 0.164 0.217 0.246 0.276 0.185 0.155 50 0.152 0.189 0.252 0.284 0.316 0.214 0.175 60 0.169 0.203 0.270 0.311 0.355 0.231 0.183 75 0.179 0.206 0.273 0.324 0.361 0.235 0.197 90 0.188 0.211 0.274 0.327 0.367 0.240 0.198 105 0.191 0.212 0.276 0.327 0.370 0.244 0.204 120 0.194 0.214 0.276 0.330 0.372 0.246 0.209 135 0.194 0.215 0.282 0.330 0.375 0.248 0.211 150 0.197 0.217 0.284 0.331 0.377 0.249 0.213 165 0.197 0.221 0.286 0.331 0.378 0.250 0.214 180 0.198 0.221 0.286 0.332 0.378 0.250 0.214 Filtrate 0.198 0.222 0.286 0.332 0.378 0.250 0.214 Table 6.5 : Temperature effect on reaction kinetics (0.1 M H2S04 solution, stoichiometric runs, 25-85°C) The data given in Table 6.5 for sulfate media were re-manipulated using Eq. 46 to find the best linear fit. This resulted in a series of lines that have slopes equal to k, the parabolic leaching rate constant, which are presented in Fig. 6.14. As indicated earlier, the leaching reactions beyond 65 °C are less effective, which is attributed to the severe competition from the hydrogen evolution reaction. Hence, the data beyond this temperature were ignored. The same procedure was repeated for the chloride-medium leaching data (Table 6.6). The results are given in Fig. 6.15. 84 Experimental objective : To study the effect of temperature Experimental conditions Agitation speed 250 rpm SPD 1.11% Iron added as % stoichiometric 100 Initial solution composition 0.1 MHC1 Solution pH 1.1 Concentrate type Gibraltar chalcopyrite concentrate Particle size -149 um +74 um Experimental results Conversion at temperature of Time, min 25 °C 35 °C 45 °C 55 °C 65 °C 75 °C 85 °C 0 0.000 0.000 0.000 0.000 0.000 0.000 0.000 5 0.055 0.076 0.095 0.118 0.134 0.086 0.069 10 0.092 0.118 0.139 0.157 0.171 0.129 0.112 15 0.115 0.143 0.171 0.195 0.215 0.157 0.136 20 0.147 0.176 0.206 0.228 0.244 0.190 0.168 25 0.181 0.208 0.241 0.261 0.287 0.226 0.198 30 0.194 0.231 0.267 0.294 0.331 0.249 0.216 40 0.221 0.265 0.292 0.332 0.379 0.284 0.248 50 0.249 0.302 0.329 0.374 0.411 0.317 0.277 60 0.275 0.327 0.365 0.404 0.440 0.345 0.307 90 0.280 0.335 0.393 0.428 0.460 0.365 0.313 120 0.285 0.337 0.393 0.432 0.463 0.366 0.317 150 0.288 0.339 0.400 0.444 0.470 0.370 0.321 180 0.288 0.339 0.400 0.444 0.470 0.370 0.323 Filtrate 0.288 0.339 0.400 0.445 0.470 0.372 0.323 Table 6.6 : Temperature effect on reaction kinetics (0.1 M HC1 solution, stoichiometric runs, 25-85 °C) The slopes of the lines in Figs. 6.14 and 6.15 are summarized in Table 6.7. Fig. 6.16 is the Arrhenius plot of In k vs. T1, to estimate the energy of activation and other thermodynamic values. From this figure, the apparent activation energy is 33.9 kJ/mol in sulfate media, and 22.4 kJ/mol in chloride media, in the temperature range 25-65 °C, using particles within the size range -100 mesh +200 mesh (-149 um +74 um). The estimated activation energy for chloride media is smaller than that for sulfate media, suggesting that leaching kinetics in chloride media are more rapid than in sulfate media. This is also evident from Table 6.7, which shows that reaction rates in chloride media are almost twice 85 o 25 deg. C • 35 deg C A 45 deg. C m 55 deg. C e 65 deg. C o 75 deg. C A 85 deg. C 80 100 Time, min Fig. 6.12 : Plot of conversion vs. time at various temperatures (0.1 M H2S04, stoichiometric runs). The data are from Table 6.5. 0.50 0.45 0.40 o 25 deg. C 0.35 D 35 deg. C 0.30 A 45 deg. C 'w u 0.25 o 55 deg. C > O m 65 deg. C u 0.20 x 75 deg. C 0.15 * 85 deg. C 0.10 0.05 0.00 80 100 Time, min Fig. 6.13 : Plot of conversion vs. time at various temperatures (0.1 M HC1, stoichiometric runs). The data are from Table 6.6. 86 0.06 0.05 J X I 0 10 20 30 40 50 60 Time, min Fig. 6.14 : Plot of product layer model fitting of conversion vs. time at various temperatures (0.1 M H2S04, stoichiometric runs). The data are from Table 6.5. 0.09 0 10 20 30 40 50 60 Time, min Fig. 6.15 : Plot of product layer model fitting of conversion vs. time at various temperatures (0.1 M HC1, stoichiometric runs). The data are from Table 6.6. 87 -9.00 n -6.00 -| H , r , , , , , , 2.95 3.00 3.05 3.10 3.15 3.20 3.25 3.30 3.35 3.40 r'xio3, K"1 Fig. 6.16 : Plot of reaction rates vs. inverse of temperature (Arrhenius plot) for sulfate and chloride media. The data are from Table 6.7. -15.00 -, -14.50 J , | , , , , ! ! 1 2.95 3.00 3.05 3.10 3.15 3.20 3.25 3.30 3.35 3.40 r'xio3, K"' Fig. 6.17 : Plot of In (k/T) vs. inverse of temperature for sulfate and chloride media. The data are from Table 6.7. 88 as those in sulfate media for the same increment in temperature, supporting the thermodynamic statement given in Section 2.1 in that leaching in chloride media is more effective than in sulfate media. Fig. 6.17 summarizes the results for the estimation of enthalpy and entropy of activation. The estimated enthalpy of activation, from the slope of the linear fit for the plot of In (k/T) vs. inverse of temperature, is 31.3 kJ/mol for sulfate media and 19.8 kJ/mol for chloride media. As indicated earlier, the discrepancy between these values of enthalpy and energy of activation is common to leaching systems. A possible source for this is the experimental errors. The other related thermodynamic and kinetic data are given in Table 6.7. Temperature, °C Parabolic leaching rate constant, Parabolic leaching rate constant, k, min"1 (sulfate media) min"1 (chloride media) 25 1.63 X IO"4 4.62 X IO"4 35 2.45 X IO"4 6.78 X IO"4 45 4.51 X IO"4 8.61 X IO"4 55 6.01 X IO"4 11.04X 10"4 65 7.79 X IO"4 13.74 X 10"4 Term Value (sulfate media) Value (chloride media) Activation energy, Ea, kJ/mol 33.9 22.4 Enthalpy of activation, AH°, kJ/mol 31.3 19.8 Entropy of activation, AS°, J/mol per K -212.4 -242.0 Preferred leaching temperature, °C 65 65 Table 6.7 : Temperature dependence of reaction rates and related thermodynamic values for sulfate and chloride media. Experimental conditions are as per Tables 6.5 and 6.6. The positive value of the enthalpy or energy of activation suggests that heat had to be provided to the leaching solution (that is by increasing the temperature) to facilitate the reactions. This is evident from the performed experiments where some improvement in reaction kinetics was obtained by increments in temperature. The decrease in conversion beyond 65 °C is attributed to the severe competition from the hydrogen evolution reaction (Eq. 43). This reaction is endothermic, and increasing the temperature would favor its tendency. The estimated entropy of activation has a negative value. From thermodynamic principles, for a reaction to take place, it should have a net positive increase in entropy, i. e. : 89 AS° > 0. The negative number should not confuse the reader. The loss of entropy indicates that when the new solid phase is formed from chalcopyrite, the system's degree of freedom is decreased, and as a result the reaction disorder or randomness is lowered. The loss of degree of freedom is perceived by virtue of mass losses, or rejection of iron and sulfur, thus converting from a stable or refractory (hard to leach) solid phase, which is chalcopyrite, to an amenable solid phase, which is chalcocite, with the resulting collapse of the crystal structure of chalcopyrite. This loss of entropy also indicates negative molar volume changes in the solid particles which will be addressed later. Similar results were obtained when complex metal sulfides (like marmatite and pentlandite) were acid leached, as well as when chalcopyrite was leached by ethylene diammine tetra-acetic acid (Peters (1992)). This value of entropy loss, for both sulfate and chloride media, indicates that the alteration or the recrystallization process is directly related to the morphology of the solid phases, rather than the diffusing species, and explains how the crystal structure of the mineral affects the leaching kinetics, under reducing conditions. This agrees well with the findings by Majima and Peters (1966) discussed in Section 2.2. The formation of the new gas phase may also explain why these reactions take place, although negative entropy change is obtained, since a gas phase represents a high state of randomness (disorder). From a thermodynamic point of view, all systems or species seek for the most stable or favorable state by possessing a minimum energy level and having a high state of entropy (randomness). The final conclusion from this analysis, on kinetic basis, is that for a process to be selected to leach chalcopyrite reductively, it is advised to run at 65 °C whether a sulfate or a chloride medium is chosen. From the estimated activation energy, the first modification to Eq. 49 in sulfate media is : (50) and in chloride media is : (51) 90 Chae and Wads worth (1979) obtained an activation energy of 28.1 kJ/mol for the reductive decomposition of chalcopyrite with metallic lead, in strong HC1 solutions. The results for improvement in conversion are comparable to those of Shirts et al (1974). The estimated activation energy for both systems is common to those of reactions controlled by a transport process in the product layer. The reaction rates, also for both systems, appears to be less sensitive to temperature variations. 91 6.1.3 EFFECT OF PARTICLE SIZE Particle size plays an important role in most leaching kinetics, as was discussed in Sections 2.2-2.4. It is apparent from Table 2.10 that there are different findings regarding the effect of particle size of both the reductant and chalcopyrite on the reduction kinetics. For the systems studied in this research, particle size effects of the reductant, iron, are ignored, as all the experiments were performed with powdery iron. Only those related to chalcopyrite are analyzed. Since chalcopyrite is of refractory nature, it is expected that reaction rates will be improved upon using finer size fractions. In addition, ash control reactions are generally preferred using smaller particle sizes as discussed in Appendix I. Experiments were performed as required, to inspect the effect of chalcopyrite particle size on reaction kinetics. The results of the experiments are given in Tables 6.8 and 6.9, and in Figs. 6.18 and 6.19. From these figures, it is clear that the leaching kinetics are dependent on chalcopyrite particle size, and the final conversion increases by more than 50% when using the size fraction -100 mesh +200 mesh instead of -80 mesh +100 mesh. Also, the final conversion is doubled upon using the finest size fraction, -325 mesh +400 mesh, which clearly indicates such a dependence. The coarser the particle size of chalcopyrite means less reaction effectiveness, and was visibly noticed in the experiments as poor coupling and mixing, leading to poor leaching kinetics. On the other hand, finer chalcopyrite particles were observed to have better mixing and coupling with iron particles. Consequently, with fine particles better final conversion was obtained. The finding that the extent of conversion is directly related to the particle size of chalcopyrite makes this leaching system unique among other studied systems that were found to be less dependent on the particle size, whether of the reductant or the mineral (Table 2.10). The other implications of particle size dependence are the efficiency of utilizing iron as a reductant and kinetic selectivity. It is clear from Figs. 6.18 and 6.19 that reductive leaching is more efficient at finer chalcopyrite particle sizes, which means that a smaller proportion of added iron is used in side reactions (specifically the hydrogen evolution reaction). This means that selective reductive leaching of chalcopyrite is less achieved when coarser particle sizes are used. 92 It is concluded here that even very fine particles (smaller than 400 mesh) are best suited for this method of chalcopyrite enrichment. The particle size dependence in chloride media is very clear compared to that in sulfate media. The final conversion in chloride media appears to be greater than that in sulfate media, under nearly equal acid concentrations. This can also be a direct indication that leaching in chloride media is more efficient than in sulfate media, and is explained by the activity of the chloride ion compared to that of the sulfate ion. Experimental objective : To study the effect of particle size Experimental conditions Agitation speed 250 rpm Temperature 25 °C SPD 4.04% Iron added as % stoichiometric 100 Initial solution composition 0.21 MH2S04 Solution pH 0.80 Concentrate type Gibraltar chalcopyrite concentrate Experimental results : Conversion at chalcopyrite particle size of Time, min -180 um -149 um -90 um -74 um -63 um -53 um -44 um +149 um +125 um +74 um +63 um +53 um +44 um +38 um 0 0.000 0.000 0.000 0.000 0.000 0.000 0.000 10 0.066 0.091 0.119 0.155 0.186 0.217 0.258 20 0.104 0.143 0.175 0.219 0.244 0.306 0.351 30 0.137 0.177 0.242 0.264 0.311 0.389 0.433 40 0.161 0.206 0.282 0.316 0.365 0.444 0.496 50 0.189 0.246 0.313 0.361 0.413 0.495 0.546 60 0.217 0.281 0.345 0.413 0.455 0.526 0.588 90 0.253 0.332 0.376 0.439 0.482 0.553 0.615 120 0.267 0.361 0.407 0.453 0.511 0.581 0.618 150 0.269 0.363 0.429 0.458 0.518 0.586 0.631 180 0.271 0.368 0.436 0.462 0.524 0.591 0.636 Filtrate 0.272 0.372 0.438 0.463 0.527 0.593 0.639 Table 6.8 : Particle size effect on reaction kinetics (sulfate media, stoichiometric runs, 25 °C) The kinetic data in Tables 6.8 and 6.9 were fitted using the product layer model, and the results are given in Figs. 6.20 and 6.21. The slopes of the lines in these two figures, k values, are compiled in Table 6.10, which summarizes the reaction rate dependence on particle sizes. 93 Experimental objective : To study the effect of particle size Experimental conditions Agitation speed 250 rpm Temperature 25 °C SPD 2.47% Iron added as % stoichiometric 100 Initial solution composition 0.22 M HC1 Solution pH 0.88 Concentrate type Gibraltar chalcopyrite concentrate Experimental results : Conversion at chalcopyrite particle size of Time, min -180 um -149 pm -90 pm -74 pm -63 pm -53 pm -44 pm +149 um +125 pm +74 pm +63 pm +53 pm +44 pm +38 pm 0 0.000 0.000 0.000 0.000 0.000 0.000 0.000 10 0.074 0.089 0.106 0.138 0.166 0.197 0.231 20 0.121 0.134 0.191 0.227 0.277 0.313 0.353 30 0.172 0.192 0.233 0.2904 0.344 0.393 0.449 40 0.199 0.223 0.279 0.331 0.401 0.467 0.539 50 0.229 0.254 0.331 0.401 0.469 0.523 0.591 60 0.257 0.283 0.365 0.433 0.501 0.541 0.655 90 0.268 0.299 0.406 0.449 0.542 0.566 0.665 120 0.285 0.308 0.417 0.455 0.554 0.591 0.674 150 0.298 0.317 0.421 0.461 0.556 0.603 0.681 180 0.311 0.327 0.424 0.466 0.560 0.615 0.682 Filtrate 0.317 0.344 0.425 0.474 0.562 0.618 0.682 Table 6.9 : Particle size effect on reaction kinetics (chloride media, stoichiometric runs, 25 °C) Particle Mean CuFeS2 Parabolic leaching rate Parabolic leaching rate size, pm particle constant, k, min"1 constant, k, min"1 diameter, pm (sulfate media) (chloride media) -44/+38 41.0 2.695 X IO"3 3.321 X IO"3 -53 / +44 48.5 2.087 XI0"3 2.277 X 10"3 -63 / +53 58.0 1.396 X 10"3 1.773 X IO"3 -74 / +63 68.5 1.066 X 10"3 1.231 X IO"3 -90 / +74 82.0 0.761 X IO"3 0.821 X IO"3 -149/+125 137.0 0.452 X IO"3 0.481 X IO"3 -180/+149 164.5 0.261 X IO"3 0.386 X 10"3 Table 6.10 : Particle size dependence of reaction rates for sulfate and chloride media. Experimental conditions are as per Tables 6.8 and 6.9. 94 • -80# +100# • -100# +115# A -170# +200# 0 -200# +240# • -240# +270# A -270# +325# O -325# +400# 80 100 120 140 180 Time, min Fig. 6.18 : Plot of conversion vs. time at various particle sizes (sulfate media, stoichiometric runs, 25 °C). The data are from Table 6.8. 0.70 • -80# +100# 0.60 • -100# +115# A -170# +200# 0.50 • -200# +240# g 0.40 O -240# +270# > O -270# +325# cS 0.30 A -325# +400# Fig. 6.19 : Plot of conversion vs. time at various particle sizes (chloride media, stoichiometric runs, 25 °C). The data are from Table 6.9. 95 0.175 0.150 • -80# + 100# 0.125 2/3 2 l-3(l-X„) +2(l-Xb) = k0d0" t • -100# + 115# 1^ = 0.99 X A -170# +200# c. o.ioo * -200# +240# >< 0.075 m-240 # +270# o -270# ' +325# 0.050 • -325# +400# 0.025 0.000 Fig. 6.20 : Plot of product layer model fitting of conversion vs. time at various particle sizes (sulfate media, stoichiometric runs, 25 °C). The data are from Table 6.8. 0.250 0.225 -I 0.200 0.175 2/3 2 » -80# +100# >? 0.150 l-3(l-Xb) +2(l-Xb) = k0d0" t 1^ = 0.98 • -100# +115# ?f 0.125 -I A -170# +200# X i o.ioo A -200# +240# O -240# +270# a -270# +325# X -325# +400# Fig. 6.21 : Plot of product layer model fitting of conversion vs. time at various particle sizes (chloride media, stoichiometric runs, 25 °C). The data are from Table 6.9. 96 From Table 6.10 it can be seen that about ten-fold improvement in reaction rates was obtained upon decreasing the initial particle size from 164.5 pm to 41.0 pm, which clearly indicates that reducing the chalcopyrite particle size is an essential step toward improving the leaching kinetics, or for process development. This topic will be addressed later upon developing the process flowsheet. In the kinetic analysis of leaching data, it has been assumed that the solid particles are spheres having an average radius between respective sieve sizes and the particle sizes in this range were averaged. Since Eq. 48 is the general expression for the parabolic leaching rate constant, k, and since the reactant concentrations are kept constant, and the experimental temperature, 25 °C, is within the range of the estimated activation energy, it turns out that Eq. 48 has now the form : k = |y (52) and the intrinsic parabolic leaching constant, k0, now includes all the remaining constants. Consequently, a plot of k vs. -4- (or -4- , as in Appendix I) will give a straight line, with an R do intercept equal 0. Such a line will be another confirmation of the suggested theory of product layer control. Based on the data in Table 6.10, the required plot is given in Fig. 6.22 which proves that the system has parabolic leaching kinetics. The straight lines obtained in Fig. 6.22 have nearly a zero intercept, and may be accepted for such a verification. One can argue that the data points would have a better fitting if the intercept is not assumed to be zero, and this is correct. According to Dreisinger (1999), it seems that at large particle sizes, there would be a shift or deviation in the controlling mechanism of leaching, from parabolic to other forms, like chemical (linear) or electrochemical control, or may be a combination of these mechanisms. In fact, such a speculation may be accepted, especially for chloride media, but there are different authors who fitted their leaching data by this method, and allowed the intercept not to be zero, without any explanation (Sections 2.2-2.4). 97 3.50 3.00 Chtaride media: 2 k = M0- 2.50 r2 = 0.98 2 2.00 Sulfate media: 2 k = k0d0" B 1.50 c r2 = 0.98 ao % 1.00 oi 0.50 J 0.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 2 4 2 do' X10 , nm" Fig. 6.22 : Plot of reaction rates vs. inverse square of chalcopyrite mean particle diameter (sulfate and chloride media, stoichiometric runs, 25 °C). The data are from Table 6.10. Anyway, the reasons for such a trend are not clear, and in this author's opinion, there might be other reactions or kinetic events which are taking place independent of particle size. The mineralogical composition of the tested concentrate showed that it contains around 17% pyrite (by mass), and such a trend may be attributed to galvanic interactions between chalcopyrite and pyrite (Table 2.9). These galvanic interactions, that is the electron transfer from chalcopyrite to an active reaction site on pyrite, would allow electrochemical dissolution of the former, which would take place at such large particle sizes. In this case, the anodic reaction might be : 2+ 2+ 2 CuFeS2(s) + 8H20(1) -» Cu (aq) + Fe (aq) + 16FT(aq) + 2S04 "(aq) + 16e E° = 0.38 V (53) and the cathodic reaction in turn is oxygen reduction : + 402(g)+16H (aq)+16e-^8H20(1) E°= 1.23 V (54) At larger particle sizes, slower dissolution rates occur, and the solution transport of hydrogen, ferrous or sulfide ions is expected to face more resistance to reach or leave active 98 reaction sites, as required. More discussion regarding these possibilities will be given in Section 6.2 upon representing the schematic leaching model. The galvanic interactions were not investigated in this research, but may be a good recommendation for further studies. It is worth mentioning that the galvanic effect by pyrite might be impeded by the formation of dense product layers. From Fig. 6.22, the inverse second order dependence of reaction rates on particle radius is established, and Eqs. 50 and 51 are now confirmed with respect to particle size. Consequently, the galvanic conversion is again shown to be directly related to the initial particle size of chalcopyrite. The decision about which size fraction to be used will affect the power requirement for a selected process flowsheet. Such a parameter was not analyzed in this research, but can be studied within a complete assessment of leaching conditions and process cost effectiveness. 99 6.1.4 EFFECT OF INITIAL ACID CONCENTRATION Most of the reported work in the open literature includes investigating the effect of hydrogen ion concentration on leaching kinetics. This is because the amount and concentration of the acid are directly related to process control (by SPD) and economics. There is no obvious reason for claiming an "optimum" acid concentration, as Shirts et al (1974) did, because this should be determined in conjunction with all other variables, and is dictated by the particular system. The best thing that can be done is to determine the reaction dependence on hydrogen ion concentration, studying this within some reasonable range (normally 0.1-1.0 M) and applying this knowledge to system behavior using concentrated acid solutions. Experiments were done to assess the dependence of reaction rates on initial sulfuric or hydrochloric acid concentration. A sample of the kinetic data is given in Tables 6.11 and 6.12, and a graphical representation is given in Figs. 6.23 and 6.24. As expected from the reactions written in Eqs. 32 and 33, the hydrogen ion concentration has a direct effect on reaction kinetics. The extent and rate ofconversion increase gradually with increasing acid concentration. However, this increase disappears after a certain limit. As was found for a set of experiments done at concentrations greater than 1.0 M for sulfuric acid and 2 M for hydrochloric acid, the reaction rates start declining, due to the severe competition from side reactions (mainly the hydrogen evolution reaction). Under such concentrations, huge gas bubbles were forming which were efficient at breaking up any galvanic coupling required for the reactions to proceed. The other undesired result from using higher acid concentration is the possible tendency to start leaching chalcopyrite oxidatively, once iron is completely dissolved, or the leaching of other solid phases or even the new solid phase. If tiny amounts of metallic copper are formed, they will again redissolve in the leach solution, rather than remain as cemented copper. So, in the experiments done, it was preferred to start with lower acid concentrations and maintain the acidity with additions of concentrated acid. This was utilized in the process study, where the initial hydrochloric acid concentration used was 0.1 M and maintained with 5 or 10 M HC1 solution, as will be addressed in Section 6.4. 100 CD a o vo Tff CO CN ON CN o in Tt Tt o i-H 00 ON r-H Tt m vo NO r-- a •+-2» CO CO Tt Tt Tt Tt Tt Tt Tt o ir-HH o CN CN e n o o o t- ent r aci d alco p Tt o CD a o o + o o co in vo CN o vo vo NO NO r- a o Tt o in O Tt o CN CN CN CN CD MH -*-» T3 CO CO CO Tt Tt Tt Tt Tt a a CO o CN CN ra l CD u & ON ac i d d d d d d d d d d d d o o rO Tt d CD o . m in o in CN O CN o CN in CN o NO 00 00 m o o o rsio n a t s o r-H 00 CO o CO in 00 ON ON ON CD o r—1 <-H CN CN CO CO CO CO CO CO CO 0.24 : o stud y etri c B d d d d d d d d d d d d H SH o • • CO Oa U tiv e tio n ich i CO s o co CO in CO o o o o o el • i—t o -rn o oo CN m oo r—1 Tt NO 00 oo oo oo o o d d d CN CN CN CN CN CN CN 0.1 4 d d d d d d d d d 1 con d S % t typ e 1 resu l ee d * CD CO ta i obje < -fH &, CD CD -r-» a a CO T3 a a o © o o o o o o o o o CD 3 CD 'co CD r-H CO Tt NO ON 00 •a CN m CN r—m1 lH S CD" n ad d ncen t trat e peri m peri m mpera t jitatio i Q rticle; peri m CD PH *^H X roH o X a w W H C/2 U PH w H PH < o 00 CN CO m co NO Tt Tt Tt Tt o Tt Tt o m ON Tt in a o r—1 CN CO CO CO Tt Tt Tt r-- Tt Tt o 'i Tt 0.5 6 M d d d d d d d d d d d d rH CD a1 a CfH CD O O CD a o CO o Tt ON CN CO CN o o o in >H , o CO CO ON CO 00 CN Tt NO NO NO NO o CN CO CO Tt Tt Tt Tt Tt atio n o CN Tt o Tt CN d d d d d d d d d d d d l*S o d '8 I, entr ; r=3 Tt "S O -r-» + o Tt CN NO NO Tt CO CN NO NO NO o o r—1 O r—1 m o CN CN CN CN in r—1 CO Tt Tt Tt Tt Tt Tt i d con e o CN CO CO CN t+H s Ii CD d d d d d d d d d d d d CD 6- & d u o s ON o 00 Tt m -O •8 l| m CN CN o a CN o o ON o 00 ON in CN 00 NO o o o rH o oo in ON CN CO Tt in m m o CO CO CO CO CO CO 'tl 2 I CD CN o CN CN CN o %-» o NO m ON o CO ON NO NO NO CO CO ll o Tt t> co NO t> 00 00 O Tt o © r—1 CN •a O CN CN CN CN CN CN CN CD o "o CD d d d d d d d d d d d d CD d •¥ CD co ft CD co "3 cd -*-> "3 CH CD CD co | N a o o o o o o o o o o o a* CD CO Tt in m 00 I C , CN NO ON CN r-H '§ r-H CD A 0.35 M O 0.42 M • 0.56 M > a o U 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.23 : Plot of conversion vs. time at various sulfuric acid concentrations (constant CuFeS2 and iron additions, 25 °C). The data are from Table 6.11. 0.14M • -tr • 0.24 m -Q- A 0.33 M O 0.43 M n o • 0.53 M > c o U 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.24 : Plot of conversion vs. time at various hydrochloric acid concentrations (constant CuFeS2 and iron additions, 25 °C). The data are from Table 6.12. 102 In addition to minimizing the tendency toward hydrogen evolution reaction, other benefits of this procedure are stabilizing the pH reading and keeping smooth control of the leaching process, in terms of gas evolution and absorption, as well as preserving an almost constant SPD, which was found to affect the alteration process. It is clear from the indicated figures that hydrochloric acid is more efficient than sulfuric acid, with stoichiometric additions. This is attributed to the activity of the chloride ion which will increase the activity of the hydrogen ion. Also, the free chloride ion might play some role in affecting the leaching rates, because increasing the Cl" concentration would not only increase the hydrogen ion activity, but may also lead to a more direct participation of the chloride ion through specific adsorption or surface complexing, explaining the trend of the plot shown in Fig. 6.24. Recalling Eq. 45, it shows that the chloride ion is capable of decreasing the tendency for hydrogen evolution reaction to take place by the effect of complexing, as the standard free energy of change for this side reaction is lower than that of Eq. 43, and, consequently, better reaction rates in chloride media are obtained. The decrease in final conversion at a threshold acid concentration is due to the increased probability for reaction 43 to take place. Although maintained by acid addition, any losses of protons caused by consumption in side reactions would affect the reaction kinetics, as these protons are reactants. The vigorous evolution of the light H2 gas, as well as that of the heavy H2S gas, will affect the ability of particles to form or maintain galvanic couples, especially in non- stirred systems. Extensive hydrogen evolution is the main problem for the studied leaching systems, since it also consumes the added iron (or more precisely available electrons for reductive leaching). It is recommended here to make further investigation on the possible methods to suppress this side reaction (see Section 8). In the performed kinetic analysis, the activity of hydrogen ion was replaced by its initial concentration. In actuality, it would be more accurate to estimate and use the hydrogen ion activities rather than concentrations, in expressing reaction kinetics, but as many of the experiments were done at low acid concentrations, 0.1-0.5 M, and room temperature, the assumption to use the concentration instead may be justified. 103 In many cases, and as long as the solubility product of certain species is not exceeded, the incorporated marginal error of such an assumption is acceptable. Most of the known models of activity estimation result in a small discrepancy (normally ~5%) between activity and concentration at such thermodynamic conditions and low solution concentrations (Dreisinger and Peters (1992)). For sulfate media, the data for hydrogen ion dependence in Table 6.11 were fitted using the parabolic leaching model to establish the reaction kinetics dependence on hydrogen ion concentration. From Fig. 6.25, the linear fitting gives the values of the parabolic leaching rate constant (slopes of the lines). Since these experiments were done at constant temperature, iron and chalcopyrite amounts and particle size, Eq. 48 would now be rewritten as : + a k = k0[H ] (55) and k0, the intrinsic parabolic leaching rate constant, contains all other constants. This equation on logarithmic scale becomes a linear relation, as per : + logk = logk0 + alog[H ] (56) since the initial acid concentration can be expressed by the initial solution pH, then Eq. 56 becomes: logk = logk0-apH (57) and as per this equation, a plot of log k vs. pH would give a straight line. The slope of this line will reveal the reaction kinetics dependence on acid concentration. The kinetic data in sulfate media are summarized in the following table : Initial H2S04 , Solution pH Parabolic leaching rate Final conversion, % concentration, M constant, k, min"1 (sulfate media) 0.14 1.01 4.24 XI0"4 28.35 0.21 0.93 6.52 X 10"4 35.04 0.35 0.69 10.34 XI0"4 42.69 0.42 0.63 11.84 X 10"4 46.51 0.56 0.54 12.99 X 10"4 47.47 Table 6.13 : Hydrogen ion dependence of reaction rates (sulfate media). The data were compiled from Fig. 6.25. Experimental conditions are as per Table 6.11. 104 0.09 0.08 0.07 1-3(1-X f3+2(1-X ) = kt b b 0.14 M X 0.06 • r2 = 0.96 • 0.21 M ? 0.05 A 0.35 M 0 0.42 M X 0.56 M Fig. 6.25 : Plot of product layer model fitting of conversion vs. time at various sulfuric acid concentrations (constant CuFeS2 and Fe additions, 25 °C). The data are from Table 6.11. 0.09 0.08 - 0.07 - 2/3 l-3(l-Xb) +2(l-Xb) = kt _Q 0.06 - r = o. Xi of 0.05 - 0.14 M + • s 0.24 m x> 0.04 - • X A 0.33 M 1 0.03 - O 0.43 M 1-3( 1 X 0.53 M 0.02 - 0.01 - 0.00 , Fig. 6.26 : Plot of product layer model fitting of conversion vs. time at various hydrochloric acid concentrations (constant CuFeS2 and Fe additions, 25 °C). The data are from Table 6.12. 105 -3.4 . -3.3 -3.2 -3.1 -3.0 logk =-0.99 pH-2.31 -2.9 R2 = 0.96 -2.8 -2.7 0.50 0.60 0.70 0.80 0.90 1.00 1.10 pH Fig. 6.27 : Plot of log k vs. pH (sulfate media, constant CuFeS2 and Fe additions, 25 °C). The slope of line is -0.99 and the order of parabolic leaching rate constant with respect to [FT] is 1. The data are from Table 6.13. Fig. 6.28 : Plot of log k vs. pH (chloride media, constant CuFeS2 and Fe additions, 25 °C). The slope of line is -1.18 and the order of parabolic leaching rate constant with respect to [H+] is 1. The data are from Table 6.14. 106 Based on this table, Fig. 6.27 shows the plot of reaction rates vs. hydrogen ion concentration. The obtained straight line states a clear dependence of reaction kinetics on acid concentration, and compares well with the findings for other reductive leaching systems (Table 2.10). From this graph, the value of a is approximately 1. Hence, reaction rates are first-order dependent on sulfuric acid concentration. From Fig. 6.23, it seems that for sulfate media the reaction rates under stoichiometric additions of iron and chalcopyrite will level off at 50% conversion, which corresponds to the maximum acid concentration tested. It is therefore concluded implicitly that best sulfuric acid concentration to be used is around 0.6 M or 1.2 N. Shirts et al (1974) found that best conversion is obtained using sulfuric acid concentration of 1.6 N. The same procedure was repeated for chloride media and Fig. 6.26 is the ash control model fitting of the kinetic data in Table 6.12. Similarly, initial reaction rates are compiled in the following table : Initial HC1 Solution Parabolic leaching rate Final conversion, % concentration, M pH constant, k, min"1 (chloride media) 0.14 1.14 3.35 X10"4 28.15 0.24 0.88 7.21 X IO"4 39.06 0.33 0.83 9.13 X 10"4 42.77 0.43 0.70 11.96X10"4 45.55 0.53 0.59 14.70 XI0"4 47.42 Table 6.14 : Hydrogen ion dependence of reaction rates (chloride media). The data were compiled from Fig. 6.26. Experimental conditions are as per Table 6.12. and Fig. 6.28 is a plot of reaction rates vs. initial hydrochloric acid concentration". The slope of the straight line in the latter figure can be approximated to be like that in sulfate media, i. e. : a equals 1.0. Hence, reaction rates are also first-order dependent on hydrochloric acid concentration. The discrepancy between the values of the slope in Figs. 6.27 and 6.28 is attributed to experimental errors. Also, for the studied system in chloride media, it is concluded that best conversion may be obtained using 0.6 N HC1 solution, since reaction rates are almost indifferent near such an acid concentration (Fig. 6.24). 107 The linear fits in Figs. 6.27 and 6.28 gave an r2 value of 0.96 and 0.98, respectively. Some may claim this to be a less acceptable fit, compared to that obtained for other graphs. It should be pointed out that such fittings were based on the assumption that the concentrations are equivalent to the activities. As explained earlier, activities are more accurate to be used instead, and a better r2 value is expected to be obtained. The conclusion from this analysis is that Eqs. 50 and 51 are now rewritten as : (58) -22,423 2/3 l-3(l-XB) + 2(1-XB) = ^ [FT] exp t (59) V RT 9. and the reaction kinetics are first-order dependent on hydrogen ion concentration. Eqs. 58 and 59 are the final form of the leaching model for the systems studied in this research Finally, the benefit of sulfuric acid over hydrochloric acid is that for the former being a diprotic acid, it allows operating at higher solid pulp density at the same value of acid concentration, that is : lower solution volume. 108 6.1.5 EFFECT OF METALLIC IRON ADDITION As was pointed out in Section 2.4, the researchers who studied reductive leaching of chalcopyrite found that more than the stoichiometric iron amount is required to realize reasonable leaching rates. In this context, it is expected that leaching kinetics are dependent on metallic iron additions, since the anodic dissolution of the reductant is part of the leaching mechanism (Eq. 35) and for the reasons explained in Section 4. Experiments were performed to establish the leaching reaction dependence on iron additions. The required experiments were done at constant chalcopyrite and acid concentrations, and using monosized chalcopyrite particles. The results are given in Tables 6.15 and 6.16, and Figs: 6.29 and 6.30. It is very obvious from these graphs that initial metallic iron additions have a considerable effect on leaching kinetics. In sulfate media, doubling the amount of added iron has doubled the final conversion under equal additions of acid and chalcopyrite. For chloride media, the same trend can be deduced. For both systems, increasing the amount of iron beyond twice the stoichiometric requirements has little effect on final conversion, since the hydrogen evolution reaction will more likely take place. This finding agrees well with that obtained by Shirts et al (1974). Deciding on the amount of metallic iron to be added will allow leaching at higher SPD, which was found to be preferable for such systems. In addition, there are other benefits for such extra additions as explained below. The final conversion in these systems reaches its maximum value when twice the stoichiometric iron is added. It is evident from Figs. 6.29 and 6.30 that reaction rates are improved with such additions. Hence, it is recommended to use twice the required stoichiometric amount of iron. Beyond this value, no improvement in conversion is realized. This observation is important, because excess iron additions would favor more hydrogen evolution rather than chalcopyrite reduction, with the subsequent implications on leaching and operation. Increasing the amount of iron to the proposed limit would : 1) account for losses caused by the hydrogen evolution reaction 2) give more chances for repetitive coupling (reproducing or increasing their number) 109 CO CJ a o r-H VO 00 © Tt VO o Tt Os VO Os CM Tt Tt Tt o m (N CN o CM Tt Tt in m VO vo VO VO vq o •r-i cn II CN ©' © © ©' © © © © © © © © o >met r o o >H CJ o Tt VO CM o as Tt Tt o © © VO Tt 00 T-H CN CN CN in o rn Tt m VO VO VO VO VO oo CM Tt in cO T-H -r-> f stoic h © o © © © © © © © © © © 1-2 .2 rsio n a t i CJ CJ 'GO .> a :> o 00 O cn ON 00 s © CN "-fn 0 - © o o 1 © © VO © Tt 00 cn m m VO T3 Co n © © CN CN CN cn cn cn cn cn CJ fl I o cn O fl © © © © © ©' © © © © © © •? ha o OH GO CJ CJ CJ ^1 "os" cj <^ , fl *C0 OH o CJ fl" GO 2 © © © © © © © © © © fl" OH fl © CJ CJ C T-H CM cn Tt in VO Os CM m 00 2 GO fl T-H o II CJ O CJ CJ a O- CO CJ CJ a a -TH H CJ" U UH fl OH CJ ,'5bl o •s 1 x Ml St \< |5 O a f-H w 100% • • 135% • 155% A 185% O 215% 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.29 : Plot of conversion vs. time at various metallic iron additions (sulfate media, constant CuFeS2 and H2S04 additions, 25 °C). The data are from Table 6.15. The graph shows the gradual increase in reaction rates with increasing iron additions, then leveling off at values approximately twice the stoichiometric requirement. 0.70 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.30 : Plot of conversion vs. time at various metallic iron additions (chloride media, constant CuFeS2 and HC1 additions, 25 °C). The data are from Table 6.16. The graph shows the same trend as in Fig. 6.29. Ill 3) increase the surface contact with chalcopyrite or the available surface area for the reaction 4) provide more electrons for the cathodic part (Eq. 36) to be continued 5) allow higher solid pulp densities 6) not entail any retardation to reaction kinetics (whether in the forward or backward directions) Hydrogen evolution reaction (HER) and the high corrosion rate of iron in the presence of even small amounts of metal sulfides, i. e. E° or potential difference, will make the required metallic iron twice the theoretical estimate obligatory. Iron in fact is susceptible to HER and large quantities of hydrogen gas may be produced by this side reaction. The standard reduction potential, E°, of iron is -0.44 V, and is considered low compared to that of lead, for instance, which is -0.13 V. Hence, the higher tendency for HER is clear, and explains why for lead near stoichiometric amounts are sufficient for good reduction (Table 2.10), while for iron extra amounts are needed. It is recommended again to find a suitable solution for the severe competition from the hydrogen evolution reaction (see Section 8). It should be noted that the amount of metallic iron, whether as a solid reactant or a dissolved species, is constant by virtue of mass conservation law. What really matters in this context is the increase in available anodic surface area for chalcopyrite reduction upon increasing the amount of iron. This increase will significantly improve the leaching kinetics because it implies more electrons are available for reduction leaching and more active reaction sites. The direct effect of iron additions on leaching kinetics or chalcopyrite conversion can easily be seen from the presented figures. The crucial factor in this context is exploiting the synergistic effect of other parameters on reduction leaching in conjunction with increasing the reductant amount. It is noted from these figures that reaction rates in chloride media yielded better final conversion than that in sulfate media (compare the leaching data in Tables 6.15 and 6.16). This is attributed to the higher chloride ion activity compared to sulfate ion activity. This also implies that in chloride media better utilization of available electrons for reduction is gained, due to the chloride ion complexing ability, as was indicated in the previous section. As a final note for both hydrogen ion effect and metallic iron effect, the same trend in the leaching curves is noticed, and the leveling off is again clear after 60 minutes. Moreover, the recorded acid consumption in many of the experiments was proportional to the extent of 112 conversion. This again supports the argument given in Section 6.1 in that tracking the acid consumption will be another means for tracking the extent of reaction. From the previously presented leaching curves, it can be seen that there is an increase in reaction rates with increasing the amount of added iron, initial acid concentration and temperature, up to a certain point beyond which leaching kinetics become insensitive. The same trend is also observed with decreasing particle size, but no leveling off was noticed for the experimental conditions considered in this work. This latter observation again implies that leaching kinetics and final conversion are expected to be further improved upon decreasing the chalcopyrite particle size, which will be addressed again in Section 6.4. ; The leveling off of leaching curves can be attributed to different reasons. First, the depletion of metallic iron from the reaction mixture. Second, the rapid loss of available reaction sites on chalcopyrite surface, as they become covered with product layers. The rapid formation of these layers will in turn limit the reaction progress. All these possibilities and others will be discussed in detail in Section 6.2. 113 6.1.6 EFFECT OF CHALCOPYRITE ADDITION The leaching kinetics dependence on chalcopyrite is not expected, because if all other effects are neutralized, this refractory mineral is less likely to be reactive, as was shown in Section 1. In this context, it is not expected that chalcopyrite will have any significant effect on leaching reactions, because, in reality, the reactions are driven by iron dissolution (Eq. 35), which will provide the electrons, and by the acid, which will provide the protons to complete the cathodic component of the leaching mechanism (Eq. 36). However, an experimental evidence is required to confirm this assumption. A series of experiments were performed to uncover the effect of chalcopyrite on reaction kinetics. The results are given in Tables 6.17 and 6.18, and Figs. 6.31 and 6.32. The first impression from the leaching data is that chalcopyrite has a negative effect on reaction kinetics. In fact, this is a paradox. The method of estimating the conversion depends on the reaction stoichiometry and by accounting for the iron component of the metal in the dissolved iron in solution and that caused by hydrogen evolution, as was explained in Section 5.3. This is then divided by the original iron content in chalcopyrite. Since the latter is always increasing in every run, the estimated conversion will consequently decrease, whatever the amount of released iron. Thus, the fitting in the given figures for chalcopyrite dependence is misleading and a better and more careful approach is required. In Tables 6.19 and 6.20, and Figs. 6.33 and 6.34, the amounts of released iron in solution are plotted vs. time for both systems based on the kinetic data in Tables 6.17 and 6.18. From the graphs, it is very clear that such amounts are very close to each other, whatever the added quantity of chalcopyrite. It is therefore accurate to state that leaching kinetics are independent of chalcopyrite. 114 © cn T-H f- VO ON cn CM VO cn T-H Tf GO cu T-H "3 © m f- Ov cn in 00 ON © A © © © © CM CN Si UH © © © © © © © © © © © © I 0.049 3 ii o fl O III o H-CU» o fl © Tt CN ON Tt cn cn VO Oi VO © 00 m VO 00 CN CM cn >, ii t "3 © © © cn as CN CN CN OH o B O © © © O © © © © © © additi o © © © o o a. I a ••s ,"tf Tt o CM © Tt T-H t> o> T-H cn as CM in © Tt '•S © © T-H cn Tt 00 + cn "o o cn VO VO o "s Icopyr i © CN CN CM CM CM CN © A o CN o ca A © © © © © © © © © © © © o UH a, o o Tt o US o CM ON x> Tt" cu m in ts CN CM O fl T-H r-- © VO cn 00 VO 00 ON 00 ON © ON cn oo in © CN cn Tt CM "3 r- © © CN CM CN cn cn cn cn cn o © © © © ©' © © © © © © © © B © a CU oo Conversi c fl a, GO O cn Tt ON cn in o © © Ov CM CM © T-H OO m o Tt i> © T-H CM cn cn o '3. CM *3 T-H cu o A © T-H CM cn cn cn Tt Tt Tt ^r Tt fl UH I; fl © © © © © © © © © © © © o ii GO o '? o T3 CU VCU , GO §1 UH |l OH 0) 00 N © © © © © © © © © © © 1 OH | fl fl cu 1 CM cn Tt m VO ON CN in oo I T-H T-H cu O -fl C ii 'B I UH Cut G UH o ii H ca B ii CU OH & OH S OX) c B X o w cu •c E- a 15 o . x W CM cn VO © ©' T3 T3 ca ii -UJ CM 'UH VO >> in OH © o OH *eO3 o -fl JJ O "ca CtH -fl >> on o OH CM o O ON +-» JJ Tt O ~C&© -fl ii ii o t3 A _o CM 'GO CM UH Tt © CU fl o CD U _> '•+-» m o 13 m fl © cu O fl HO1 o GO o CU ~3 cu cu a ca OH X fl" GO fl" cu OH ii cu I C c cu UH o o a a CU a cu '»H ca UH "C CU '5b CU % OH s 1 cu PH X cu w H < , 0.0351 mol > 0.0422 mol 3 0.0492 mol ^ 0.0562 mol . 0.0632 mol 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.31 : Plot of conversion vs. time at various chalcopyrite additions (sulfate media, constant H2S04 and Fe additions, 25 °C). The data are from Table 6.17. 0.45 0.40 - 0.0201 mol • 0.35 - • 0.0271 mol 0.30 - • 0.0342 mol A 0.0417 mol s 0.25 - 'cn O 0.0493 mol 0.20 - inve r <-» u 0.15 - 0.10 - 0.05 _ 0.00 * 1 : | | I | 1 1 1 . 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.32 : Plot of conversion vs. time at various chalcopyrite additions (chloride media, constant HC1 and Fe additions, 25 °C). The data are from Table 6.18. 116 NO ON r-H t> Tt t+H o NO CN 00 Tt CN o cn d 00 00* Tt Tt in 00 cn Tt d Tt (H ON oo oo 00 cn 00 ON Tt CN o in rH cu Tt "o in r- o CN in NO 00 o r-H CN cn I to W r-H r-H r-H r-H r-H o a CN CN CN CN ! '•3 d oCJ fi o r-H 00 ON NO 00 CN NO m in Tt o • rH o d cn cn CN 00 ON r-" 00 d NO in Tt o CU NO NO Tt Tt CN CN cn NO m m or-H r-H PH Tt o in t> O CN Tt in 00 o US O o r-H r-H r-H r-H i—t r-H CN CN CN o d a o i, "-f3i o r-H r-H r-H CD ti o NO cn m 00 CN ON NO Tt U ect CN d r-H t> d Tt 00 CN cn in CN d Tt rH Tt NO in ON cn 00 o O o Tt fi cn rH m o o CN in NO f- ON o r-H si PH o r-H r-H r-H r-H i-H CN CN CN fi ON PH d a o Tt 2 u CN Tt fi" 1 o in in CN d o o o ON cn NO CN cn r-H Tt CN Tt oo Tt CN "I '1 r-H r-H r-H CN NO NO Tt NO NO cn NO ON* CD d t> 00 cn CN NO CN cn ON Tt CN a CO o m r-H 00 """^ cn NO 00 ON ON CO m m o mo l CN fl a CN l-i 0.0 2 CD o fl -C«D CO j-rH fi CN r-H CN CN cn 00 r-H cn Tt a 'co ea s o rr• © O r—1 if cu d ON NO in in in Tt CN in Tt o cn r-H CN Tt cn r- r-H 00 CN Tt m i-S T3 o CN O in 00 cn in NO 00 oo ON ON ON CD C 3 -r-» ^: o n co o _> O CO cu |H a *-rH o "O r-H d o CU o I PH EH CU CD o H» "3 9< £H 1 T3 PH| cu o o o O o o O o o o O co CU t3 CN cn Tt NO ON CN 00 '? cd fi m m r-H CU B fi r-H rH cu cu , mi l O o cu B 1 cd o cu 1 fi fi a I o a CU cu o O H Filt r w H P w o 00 in q r-- cn cn CN r- q cn in CN d ON cn CN cn 00 CN in ON r-H r-H in CD fi cn t-- CN Tt r~- Tt o 00 NO cn o NO r- CN m 00 cn NO c~~ ON o o kH o r-H i—i CN CN CN CN CN cn cn o I-H i—i d a CD fiO O CD o r-H 00 Tt Tt CN ON ON in 00 CN in o o CN d CN 00 d OO in ON 00 00 in CD NO OO r-H ON r-H in 00 CN cn NO ON r-H H-» PH in rH r-- CN m ON r-H cn NO r-- OO ON o US O o r-H r-H r-H CN CN CN CN CN CN cn rH d •a cu -r-» *3 CO o Ii rfi Tt O O: ,"3 o cn Tt r-H NO OO cn r-H NO Tt cn in 00 rfi t~- CN r-H -3 (NI CN Tt ON NO CN in rfi o •s ON d d d od cu ON CN O Tt NO NO ON cn m CrH + Tt CN NO ON i—i Tt NO r- ON ON ON c a^ r-H r-H r-H CN CN CN CN CN CN CN PH o . « o &l PH d a O 2 I, o o ON m O rP ON fi" CU in CN Tt NO r-H 00 ^H ON CN o o o CN CN m q in CN r-H d Tt Tt 00 Tt r>* Tt in od d in I: r-H CN 00 ON in cn in 00 00 CO Tt w C1—NI m ON CN Tt NO 00 ON ON fi O r-H r-H r-H CN CN CN CN O rH CN CN CN CD d a a fi ro ]-rH cCoD IEc2u g 'co cd o r-H r>- CN r- ON CN NO in cn cn in ,> O r-H Q o NO CN CN* ON NO ON in 00 '-rH d od d o rH in /-\ ON cn in 00 cn Tt Tt 00 r-H CN "0 cn rH CN NO ON cn in t> 00 ON ON I 3 r-H r-H cu 13 r—t CN CN CN CN CN CN CN <-> co o fi CD rH d a O CD rH '? 4> , CO fi cu CD §1 PH CD CD o "3 CO cd N A o o o o o o O o O O O 1 td 1 rH II &l 'co CN cn Tt in NO ON CN in 00 •s fi fi r-H 1 CD CD o CD cu 1 O CD a CD • i-H cd CD" ts "3 .1 s- I fi H-» a I ts (H cd U rH CD -4-> Si "o o , 1 PH E— "3)| CO U PH , x W 3500 3000 - _i :—a—" E B D. 2500 - Cl, c # 0.0351 mol o 2000 _ D 0.0422 mol oluti i A 0.0492 mol 1500 _ A 0.0562 mol ;ase d i n s mol 1000 - 0 0.0632 m rel < o 1— 500 - m ! ! ! ! ! ! ! ! ! 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.33 : Plot of iron released in solution vs. time at various chalcopyrite additions (sulfate media, constant H2S04 and Fe additions, 25 °C). The data are from Table 6.19. 2500 0 20 40 60 80 100 120 140 160 180 Time, min Fig. 6.34 : Plot of iron released in solution vs. time at various chalcopyrite additions (chloride media, constant HC1 and Fe additions, 25 °C). The data are from Table 6.20. 118 6.1.7 EFFECT OF SOLID PULP DENSITY (SPD) Some researchers (Hiskey and Wadsworth (1975)) claimed that SPD has no effect on leaching up to 50%, which is beyond commercial practice. To account for the effect of SPD on leaching, experiments must be performed in which the solid amounts (the concentrate and reductant) are increased, but at a constant concentrate/reductant ratio. This approach was not taken in this work, which in turn would establish the effect of SPD on the enrichment of CuFeS2. Nonetheless, it was pointed out in Section 6.1.6 that considerable improvement in the leaching process is obtained by increasing the amount of iron. This increment resulted in an increase in SPD. Qualitatively, it can be deduced here that the galvanic conversion of chalcopyrite by the studied method becomes more favorable upon increasing the SPD. Consequently, the enhancement in final conversion is attributed to the increased galvanic coupling upon increasing the quantities of solid reactants. The mild agitation would also increase the exposure of iron particles to active reaction sites on chalcopyrite particles, that are not yet covered with the newly formed solid phase. As long as there is metallic iron in the system, and as long as fresh reaction sites are available, the possibility for galvanic coupling and reduction will increase, thus enhancing the enrichment process. The following table was compiled from Tables 6.15 and 6.16 to demonstrate the improvement in final conversion by increasing the SPD : Sulfate media Chloride media Metallic iron SPD, Conversion, % Metallic iron SPD, Conversion, % addition as % % (sulfate media) addition as % % (chloride media) stoichiometric stoichiometric 100% 4.25 37.72 100% 2.47 41.95 135% 4.37 41.22 135% 2.58 47.46 155% 4.44 51.89 155% 2.68 55.02 185% 4.55 65.48 185% 2.76 62.47 215% 4.66 66.24 215% 2.85 64.67 Table 6.21 : Recorded final conversion at various SPD values for sulfate and chloride media. Experimental conditions are the same as those in Tables 6.15 and 6.16. Although the change in SPD is minor, the results can indirectly be used to judge on the incorporated improvement in the alteration process. As a demonstration, Fig. 6.35 is a plot of 119 recorded final conversion vs. SPD which shows that the leaching process is improved upon increasing the SPD. From this figure it can be concluded that higher pulp densities are preferred for this method of leaching. The other advantage of using high SPD is to decrease the possibility for hydrogen evolution reaction to take place as was indicated in Section 5.2.1. Since the experimental leaching data in Tables 6.15 and 6.16 were obtained for constant chalcopyrite and acid additions, it is concluded here that increasing the amount of metallic iron, which will entail an appropriate increase in the amount of chalcopyrite, is more favorable for this method of leaching, and eventually will lead to better conversion. It is preferred to keep the acid concentration low, to avoid the competition of the hydrogen evolution reaction, as was discussed in the previous sections. For these reasons, the experiments in the process study were performed at high solid pulp density, by increasing the amount of the solid reactants, while using a constant and low acid concentration. The results show that such a technique is quite fruitful in the reductive leaching of chalcopyrite under mild conditions (Section 6.4). 70 65 - 60 _ 55 - C 'w 50 - 40 - 35 - 30 4.10 4.20 4.30 4.40 4.50 4.60 4.70 4.80 SPD, % Fig. 6.35 : Plot of recorded final conversion vs. SPD (sulfate media, non stoichiometric metallic iron additions, constant CuFeS2 and H2S04 additions, 25 °C). The data are from Table 6.21. 120 6.1.8 LEACHING RATES OF CHALCOPYRITE PARTICLES The retention time in a selected reactor for this method of leaching can be set to be one hour, since the leaching kinetics were found to be indifferent after this period. However, for the purpose of flexible process design and industrial operation, further information can be obtained from the previously estimated parabolic leaching rate constants. These constants can be utilized to estimate the required leaching time for complete dissolution (reduction) of a chalcopyrite particle at the relevant physical or thermodynamic parameter, and under specific leaching conditions. The purpose here is to demonstrate how the results of a kinetic study can be used for such purposes. The parabolic leaching rate constants in Table 6.7 (temperature effects) are compiled in the following table and further expressed as the dissolution rate of CuFeS2 particle (pm/min) : Temperature, Parabolic leaching Rate of Parabolic leaching Rate of :°C rate constant, k, dissolution, kd0, rate constant, k, dissolution, 1 1 min" pm/min min" kd0, pm/min (sulfate media) (sulfate media) (chloride media) (chloride media) 25 1.63 X 10"4 1.82X10"2 4.62 X 10"4 5.15X10"2 ••35 • 2.45 X IO"4 2.73X10"2 6.78 X IO"4 7.56X10"2 45 4.51 X IO"4 5.03X10"2 8.61 X IO"4 9.60X10"2 55 6.01 X IO"4 6.70X10"2 11.04X 10"4 12.30X10"2 65 7.79 XI0"4 8.69X10"2 13.74 X 10"4 15.30X10"2 Table 6.22 : Estimated leaching rates of chalcopyrite particles at different temperatures. The experimental conditions are the same as those in Table 6.7 and d0 equals 111.5 pm. Fig. 6.36 is a plot of chalcopyrite leaching rates vs. temperature. It can be shown that under the same experimental conditions, and if the whole leaching process is controlled by the transport through product layer, the complete leaching of a 50 pm chalcopyrite particle in sulfate media would take around 995 and 575 minutes at 45 and 65 °C, respectively. In chloride media, the corresponding times will be 520 and 325 minutes, respectively. These numbers support the findings outlined previously in that the possible feasible changes to the leaching systems are either increasing the amount of metallic iron or using much 121 finer particle sizes. It is clear that these two parameters are most important in this system of reductive leaching. 0.16 n 20 25 30 35 40 45 50 55 60 65 Temperature, °C Fig. 6.36 : Plot of chalcopyrite leaching rates vs. temperature. The experimental conditions are the same as in Table 6.7. The data are from Table 6.22 The other implication of the plot in Fig. 6.36 is that a thermodynamic change to the system is a possible tool in improving the leaching kinetics as about one-third reduction in complete leaching time was incurred by a 20-degree increment in leaching temperature. This increment in temperature is expected to be of benefit, especially when a high recycle load of the leach solution is desired for the purpose of flexible process operation. Such a temperature increment is expected to improve the solubility of certain species in the leach solution. This will be addressed again in Section 6.4. From the estimated leaching rates in Table 6.22, it can be shown that the reaction linear velocity at 25 °C is around 3.03X10"4 and 8.58X10"4 pm/s for sulfate and chloride media, respectively, which signifies the importance of diffusion control in the leaching mechanisms. This is in conformance with the criteria outlined in Section 6.1. 122 Similarly, if the rate constants at different particle sizes are used to express the corresponding initial leaching rates, then for the reaction rates in Table 6.10 the required estimations are compiled in the following table : Mean particle Parabolic Initial rate of Parabolic Initial rate of diameter, leaching rate dissolution, kd0, leaching rate dissolution, kd0, 1 1 d0, pm constant, k, min" pm/min constant, k, min" pm/min (sulfate media) (sulfate media) (chloride media) (chloride media) 41.0 2.695 X IO"3 0.1.10 3.321 X IO"3 0.136 48.5 2.087 X 10"3 0.101 2.277 X 10"3 0.110 58.0 1.396 X 10"3 0.081 1.773 X 10"3 0.102 68.5 1.066 X 10"3 0.073 1.231 X10"3 0.084 82.0 0.761 X 10"3 0.062 0.821 X IO"3 0.067 137.0 0.452 X 10"3 0.061 0.481 X IO"3 0.065 164.5 0.261 X IO"3 0.042 0.386 X IO"3 0.063 Table 6.23 : Estimated initial leaching rates of chalcopyrite particles at the respective sizes. The experimental conditions are the same as those in Table 6.10. Although these rates were obtained at different particle sizes, but they can be used for the purpose of demonstration. Fig. 6.37 is a plot of chalcopyrite initial leaching rates vs. particle size. The graphs can be extrapolated beyond the tested particle size range to show the sharp dependence of reaction rates upon reducing the initial size of chalcopyrite particles. The sharp trend in this figure is similar for both chloride and sulfate media, and simulates the common dependence of leaching rates on particle size. Fig. 6.37 states clearly that significant improvements in leaching kinetics are achievable when smaller than 38 pm particles are used. For instance, the complete conversion of a 10 pm chalcopyrite particle, under the same experimental conditions and leaching mechanism, is estimated to take around 40 and 25 minutes in sulfate and chloride media, respectively. As will be addressed in Section 6.4, making these main changes to the leaching systems, that is reducing the initial particle size and increasing the amount of the reductant, would result in significant improvement in final conversion or enrichment of the chalcopyrite concentrate. 123 It is clear from Figs. 6.36 and 6.37 that leaching kinetics in chloride media are much faster and efficient than those in sulfate media, which makes leaching in the former a desirable choice for large scale operation. 0.150 0.125 CO 0.100 13 Chloride media C 0.075 CO Sulfate media 0.050 0.025 i i i i r 1 1 1 1 0 20 40 60 80 100 120 140 160 180 Mean particle diameter, urn Fig. 6.37 : Plot of chalcopyrite initial leaching rates vs. mean particle diameter. The experimental conditions are the same as in Table 6.10. The data are from Table 6.23. So far, the value of k0, the intrinsic parabolic leaching rate constant, was not estimated because, according to Eq. 47, whenever the temperature range or any reactant concentration range is changed or other particle sizes are used, a different value for the pre-exponential factor is obtained. As a demonstrative calculation, and according to Eq. 48, the parabolic leaching rate constant, k, is written as : k = |l [H+] exp [ (60) RT If the pre-exponential factor is called K0, then in the tested temperature range this factor will be : k„ + K0 = ^ [H ] (61) 124 1 The value of K0 from the intercept in Fig. 6.7 for sulfate media is 144.03 min" . Hence, Eq. 61 is now rearranged to read : K=^ (62) Taking the particle size to be the average radius for the size fraction -100 mesh +200 mesh (-149 2 um +74 um), i. e. R , then for the experimental conditions in Table 6.5 (sulfate media), k0 is estimated as : 144.03 min-1 x (3.46X10~9) m2 k„ - 0.1 mol liter"1 = 4.98X10",-66 mVmi™2/ n per (mol/liter). 125 6.2 SCHEMATIC REPRESENTATION OF THE LEACHING PROCESS After this detailed analysis of leaching kinetics it is now appropriate to develop a schematic representation of the leaching process and discuss in detail the possible rate determining steps in the studied systems. A schematic representation of the proposed leaching model is given in Fig. 6.38. In this figure it is depicted that iron particles will form a galvanic couple with a chalcopyrite particle. Two iron particles are shown to simulate the finding that twice the stoichiometric iron required is needed for desirable reaction rates. Immediately, dissolution of iron occurs and its electrons are released, which will transport through the coupling point to reach a suitable reaction site on the chalcopyrite particle. Simultaneously, a number of protons will diffuse from the bulk solution to this reaction site, and once these protons arrive, the reaction will take place instantaneously. This instantaneous reaction is justified from the rapid leaching kinetics as was shown earlier. As the reaction proceeds, the new solid phase will appear as a product layer which will cover the solid particles in the system. The exact nature of this layer was examined by SEM/EDX methods, and found to be chalcocite. The product layer can be said to be thick and/or dense, because reaction rates tend to level off in a short time, and the observed nature of the reaction residue (Section 5.2.2) supports this argument. It is speculated that the product layer would cover chalcopyrite particles, iron particles or both of them. Since the reaction kinetics are rapid and the chemical analysis of total iron in solution showed that it is the sum of added iron and that released from the chalcopyrite lattice, it can safely be assumed that passivation of iron particles by a product layer did not occur. The thick or dense chalcocite product layer surrounds the CuFeS2 particle and grows inward as the particle reacts, while the solution will barely be in continuous contact with the unreacted core of chalcopyrite. As this product layer will cover chalcopyrite, it is now clear why chalcopyrite, as a solid reactant with some sort of refractory (hard to leach) nature, has little or no real effect on reaction kinetics. The remaining chalcopyrite particle is assumed to retain the same geometrical shape as the parent particle and the dense Cu2S layer will result in a diffusion overpotential (diffusion control kinetics or parabolic leaching) and a decrease in the available electrochemical driving force for the main reaction, i. e. : a shift to favor the hydrogen evolution reaction or other side reactions. 126 Fig. 6.38 : Schematic representation of the galvanic conversion of chalcopyrite using metallic iron as a reductant in acidic media. The figure shows Fe-CuFeS2 couples assuming iron is not surrounded by the porous product layer, and two iron particles are sharing at once in the rapid reactions. The reacted particle retains the same geometrical shape of the parent particle. 127 The nature of the product layer will limit the transport of different species, some of which are important to complete the reaction. It can limit the inward diffusion of the protons, the outward diffusion of ferrous or sulfide ions, the rearrangement of the sulfide ion within the crystal structure of chalcopyrite, or the transport of electrons through the points of galvanic coupling. As the hydrogen ions are required to complete the leaching reactions, it is now clear why mild agitation caused some enhancement to leaching. This sort of agitation will enhance the diffusion of these protons from the bulk solution to the reaction site, as was found by Nicol (1975). As the leaching proceeds, the solid layer of chalcocite will build up and start hindering the diffusion of these protons to available reaction sites. The ferrous ion is a reaction product. From Hackl's earlier study (1987), it was concluded that this ion is much slower to diffuse in chalcocite layer than other species (like the sulfide ion), and has a small solubility in chalcocite. The presence of such a layer may also impede the outward solution transport of this ion. This possibility was not studied in this research, since it requires measuring the diffusivity of Fe2+ in the product layer by electrochemical techniques, but can be a suitable recommendation for future research. Reaction products of ferrous salt, like FeS04, are not expected to cause any passivation either on iron or chalcopyrite particles, for this salt is fairly soluble in sulfate media, compared, for instance, to lead sulfate (Table 2.2). So, reaction retardation by iron salts is not expected to occur and there are no published reports or findings on its passivation under similar experimental conditions. On the other hand, some researchers (Table 2.10) found this ion to have some catalytic effect on the alteration process. In chloride media, ferrous chloride is highly soluble as written in Table 2.2 and such a probability of passivation is of no sense. The possibility of back reactions (that is the formation of chalcopyrite again) is unlikely, however, the effect of initial additions of ferrous ions needs more investigation, as will be addressed later. The release of hydrogen sulfide into solution (dissolving or then evolving) is not expected to be rate limiting, although it was found that one of the slowest moving species in the solid state is the sulfide ion, S2", as was explained by Hackl et al (1987). The observed fast reaction kinetics exclude its possibility to be rate limiting. The sulfide ion would either rearrange itself from the chalcopyrite lattice to the chalcocite lattice, or diffuse to the bulk solution, and dissolve, forming 128 the bisulfide ion or being released as hydrogen sulfide. The formation of the latter is more acceptable, due to the noticed vigorous evolution of H2S gas, and acidic conditions prefer its formation. The only exception to this scenario is a thermodynamic change in the leaching system The formation of H2S bubbles may influence the coherence of the product layer (film) on chalcopyrite. In the initial stages of leaching, these bubbles are evolved vigorously and leave the reaction sites. In the later stages of leaching, their evolution is very small, due to the depletion of the reductant. Thus, such a possibility of influence is of no importance. The recrystallization process of chalcopyrite to chalcocite might not provide easy paths for the diffusion of products or reactants along dislocations and grain boundaries. Chalcopyrite has a tetragonal structure which upon iron and sulfur losses converts to the orthorhombic chalcocite, i.e. : lattice rearrangement. It is known that the molar volume of chalcocite is about half that of chalcopyrite, however, the molar volume changes that accompany this transformation or the alteration reactions were not measured in this research, since these changes require an electrochemical study. Once such changes are estimated, they can be compared to those of chalcopyrite, and afterwards in the estimation of the product porosity (Peters (1984)). This would help in the identification of any inhibitory effect due to lattice rearrangement. Nonetheless, a qualitative assessment can be done here to speculate on the porosity of the product layer. The molar volume of a species is defined as : _ MW Vmolar ~ p (63) 3 where Vmolar is the molar volume, cmVmol, p is the mass density, g/cm , and MW is the molecular weight, g/mol. The relevant values for chalcopyrite and chalcocite are given in Table 2.7. If all the vacated sites in a chalcopyrite particle were replaced by chalcocite, then, according to the stoichiometry of Eq. 37, the porosity (s) of the product layer can be deduced, qualitatively, from the relation : — , (Xnolar)cuFeS2 ^-5 (Vmolar)Cu s 8=1- nT~j— (64> V V molar/CuFeS2 3 for chalcopyrite, Vmolar = 183.513/4.1 = 44.759 cm mol"' 3 1 for chalcocite, Vmolar = 159.158/5.8 = 27.441 cm mol' and the porosity is : 129 s = (44.76-(0.5x27.44))/44.76 = 0.69 or -70%. That is the product layer is expected to be porous. This agrees well with the findings of Shirts et al (1974) regarding the morphology of the new solid phase and compares well with the results obtained by other reductive leaching methods, for example : electrochemical reduction (Felker and Bautista (1990)). This discussion on the role of the crystal structure of solid phases and the morphology of the product layer supports the findings in Section 6.1.2 regarding the temperature effects on leaching kinetics and the estimated loss of entropy. It is clear that the nature and composition of the solid phases (old and new) have direct effects on the galvanic conversion of chalcopyrite when iron is the reductant, which is in conformance with the theoretical considerations discussed by Peters (1984) and addressed earlier in Section 2.4. The presented qualitative assessment of product layer porosity shows that negative molar volume changes have taken place, by virtue of iron and sulfur removal from the chalcopyrite, which adds up to the previous discussion. As will be shown later, it became evident from SEM testing that these arguments are correct for the studied systems in this research, and consequently the proposed leaching mechanism is valid. The electrochemical nature of chalcopyrite reduction is clear from the proposed mechanism and the rapidity of the leaching kinetics means that charge transfer steps are not rate controlling (at least in the first hour of leaching). Upon developing the product layer, such steps might be inhibited. This can only be confirmed by an electrochemical study (see Section 8). With the formation of the product layer, mainly Cu2S, the electrical conductance across Fe-CuFeS2 coupling points will be affected, and the conductive transport of electrons through the formed bridge will also be affected. There are no experimental data on the conductivity of such a layer, but chalcocite is a relatively good conductor (Table 1.2). It is assumed here that electron transport through the product layer is not rate controlling. To account for such a possibility, the electrical conductivity and electron mobility in such layers need to be measured, as was done by Munoz and his coworkers (1979). From this discussion, it is concluded that the solid product of the leaching reaction, i.e. chalcocite layer, does limit the leaching rates due to an inhibitory effect on some transport steps. 130 According to Fig. 6.38, there are four possible rate determining steps, all of which are of transport nature: 1) Hydrogen ion diffusion 2) Ferrous ion diffusion 3) Transport of electrons 4) Sulfide ion diffusion and/or its rearrangement within the crystal lattice of CuFeS2 Consequently, the rate determining step will be any of these individual steps, or a combination of two or more. In the work done in this research, the only step that was proven to be rate determining, from the established leaching model in Eqs. 58 and 59, is the first step. The remaining steps still need more investigation, as outlined in Section 8. It should be noted that the overall reactions in Eqs. 32 and 33 may also include some adsorption or desorption steps, like any other solid-fluid reaction, in addition to the formation of product layers. By considering the proton diffusion as a rate limiting step, the reasons for leveling off of reaction rates can again be addressed. The rapid observed kinetics suggest that the. available iron for reduction is rapidly diminishing. As the amount of iron diminishes, slow discharge at particle surfaces is expected and tends to become rate controlling. Since iron corrodes quickly in acid solutions, the corrosion mechanism will prevail and will not be rate limiting in the early stages of leaching. On the other hand, the need for the galvanic couple as well as the diffusion of protons means that the rate at which the galvanic mechanism proceeds will be rate limiting in the early stages. In these early stages there is an abundance of both iron and chalcopyrite, and the only possible rate determining step will be the rate of diffusion of hydrogen ions to the chalcopyrite surface. Once the product layer is developed, the rate of electron transport might become an additional rate limiting step. Thus, the initial kinetics in this system can be explained in terms of some sort of resistance to flow. This resistance will be physical (for proton transport) and ohmic (or electrical, for electron transport). The ohmic resistance, if any, is depicted to take place across the Fe-CuFeS2 coupling points (Fig. 6.38). 131 Fe (a) Initial (b) Middle (c) Final Fig. 6.39 : Sequential schematic representation of the leaching stages, showing the build-up of reaction products on the parent chalcopyrite particle. The product layer is shown to have diffusional paths. The figure also indicates the systematic variation from galvanic mechanism control (initial) to corrosion mechanism control (final). 132 As the reaction proceeds, the solid products will appear as a dense layer, that will limit the rate of diffusion of any reactant or product, and the available iron in the system quickly diminishes by the effect of side reactions, so the corrosion mechanism will now be rate limiting (Fig. 6.39). X-ray diffraction was not extensively used to find the exact structure of the solid residue. Nonetheless, chemical analysis showed that the new solid phase has a composition close to Cu2S. There was a small difference for mass analysis, around 5-10%, which was assumed to be an experimental error. In actuality this difference might be due to the presence of intermediate and/or other solid phases, like bornite. On the other hand, some SEM/EDX testing was performed to assist in identifying the product layer and its approximate composition. SEM was also used to reveal the morphology of the product layer, which will give a physical evidence to support the proposed leaching mechanism and kinetics control. Figs. 6.40-6.43 give some idea about the morphology of the concentrate before and after leaching. The SEM photographs show that the newly formed solid phase is adherent to chalcopyrite particles, confirming the previous discussion. In addition, these photographs show that there are pores and open channels in the leached particles. The presence of these pores and channels supports the argument that negative molar volume changes have occurred. Rounded-off corners are also shown, supporting that the leaching reactions are controlled by a transport process in the product layer. It is clear from these figures that the leached particles retained the same geometrical shape of the parent particles. Figs. 6.44 and 6.45 are sample photographs for a polished section of the leached chalcopyrite concentrate. Fig. 6.44 is the SEM photograph while Fig. 6.45 is the counterpart back-scattered electron image for the same section. It is very clear from these two photographs that the product layer (shown as bright parts in Fig. 6.44) is covering the original chalcopyrite particle and grows inward. This layer is not growing on a particular point, rather, the whole active reaction sites are covered simultaneously. The product layer itself is porous, relatively thick and/or dense, explaining the leveling off of leaching curves presented in Section 6.1. This porosity is in conformance with the qualitative assessment estimated from Eq. 64. The photographs show that in the opening 133 channels there are also some new product layers indicating that the reaction was progressing through the diffusional paths. It should be recalled that leaching was done under mild agitation. As was stated in Section 6.1.1, this caused some attrition to the product layer, for which the product layer appears to be relatively thin in the presented images. EDX testing of the outside product layer and those in the pores and channels showed that the approximate composition in atomic percent is 61.22% copper, 32.10% sulfur and 6.68% iron, which is almost chalcocite. The latter is a clear evidence for the validity of the leaching mechanism and the schematic representation of the leaching process. These results are also in conformance with the reaction stoichiometry and those obtained by wet chemistry methods (Appendix II). Metallic iron was not detected in any of the scanned samples, neither by EDX analysis, supporting the statement given earlier regarding the certainty of its complete dissolution. By these points, with the previous findings of less temperature sensitivity, and large dependence on particle size, it can finally be concluded that this system is controlled by the solution transport of a species or a group of species through the product layer. Reductive decomposition of chalcopyrite with metallic iron is possible and this method of leaching can be utilized for the production of copper-super concentrates. Chalcocite is seemingly the new solid product (phase) of the leaching reactions. 134 Fig. 6.40 : SEM photograph for the fresh chalcopyrite concentrate (-325 mesh +400 mesh) Fig. 6.42 : SEM photograph for the leached concentrate (-270 mesh +325 mesh), as per Table 6.8 g. 6.44 : SEM photograph for a polished section of the leached concentrate in Fig. 6.42 Fig. 6.45 : Back-scattered electron image for the same section in Fig. 6.44 137 6.3 CONCLUDING REMARKS The schematic model in Fig. 6.38 is acceptable as a representation for the reductive decomposition of chalcopyrite using metallic iron. Seemingly, chalcocite is the main solid product. The leaching reactions are controlled by a transport process in the product layer, as was confirmed from the experimental findings. The parabolic leaching model was selected and found to be satisfactory for fitting much of the experimental data. The thermodynamic parameters obtained from the analysis of temperature and particle size dependence justified this selection. The leaching models are shown in Eqs. 58 and 59. Leaching kinetics are improved with increasing acid concentration and solution temperature up to a threshold value, while reaction rates are enhanced and final conversion is improved with decreasing initial particle size. Also, the leaching kinetics are much improved by increasing the iron to chalcopyrite molar ratio. The results from the kinetic study show that maximum conversion can be obtained under the following experimental conditions : 1) Mild agitation 2) A temperature of 65 °C 3) Fine particle sizes (smaller than 74 um) 4) An acid concentration of 0.6 M 5) Addition of twice the stoichiometric iron requirement It was shown that leaching in chloride media is more efficient than that in sulfate media, and the system is best run at high SPD. 138 6.4 PROCESS DEVELOPMENT The results of the detailed kinetic analysis are now utilized in developing a simple process flowsheet for the reductive decomposition of chalcopyrite using metallic iron. The process, shown in Fig. 6.46, was developed jointly with Dreisinger (1999) as a tentative process for chalcopyrite enrichment. A brief description is given below, but the process still needs more investigation to demonstrate its applicability and viability. 6.4.1 Leaching : The leaching stage is preceded by a premixing stage where sufficient amounts of iron and chalcopyrite are mixed for a while prior to being leached. Since the leaching reactions were found to be more favorable at high SPD, the leach mixture will be at -35% SPD. To this solid mixture, an acidic solution containing excess amounts of ferrous chloride is added. The selection of the chloride media is preferable because the sulfate ion is less active and for the reasons outlined previously. The reaction vessel is simulated to be a continuous mixed flow reactor, where the required amount of acid is added batchwise. The reaction takes place at room temperature. Once the reaction started, gaseous product will be formed and removed through the special collection unit for hydrogen sulfide by suctioning. The vessel itself is under continuous agitation to the extent required to promote interparticle integrity and obtain a well-mixed mixture, as simulated in the lab by stirring at-610 rpm. 6.4.1.1 Acid effect The leach solution is 0.1 M HC1. From the kinetic study, it was found that higher acid concentrations will improve leaching, up to a certain value beyond which there would be an increase in the tendency toward hydrogen evolution reaction, and the corrosion mechanism of iron will be of little benefit. In this context, the acid dependence under high SPD and additions of FeCl2.4H20 need to be established. Table 6.24 shows the effect of initial acid concentration on reaction kinetics, under these conditions. From the data presented, it is clear that starting at high acid concentration will severely hinder conversion, due to the abrupt evolution of large gas amounts. Rather, it was observed that 139 Cl O 2 2 CCS -*-» CD fl -fl CD CD fl .2 CD T3 --1 fl £ o -° -S s Cs+H O T3 fl cS CD o 3 CD fl o "o o >-> Vn fl CD o O -H CD >H lle c a col o co cu o ^ -fl o >> a tr) co 8 3 CO CD OO co CN >> •So ^1 s- CO T ccj O •a^ 9 CD "fl 8^ P co fl C (DO lH ^ S '-s cS H—> CD CO . . ,tN J+H CD fl Hfl CD fi U »1 « co Q "S S j§ u CD PH O O £J CD O 00 00 CD 00 O PH CD co CD O OH o PH CS fl NO 2 T* NO DO -fi I"2 CJ UH s 43 cu s 0 00 K VO o ON ON o UH 00 S-H © 00 14= CJ o vq O .fl 2 O ll VO ©' in r-- oo GO s rn oo fl • B o too ©' O •cj o fc cu cu s oo s Nr 0 CN .a s |CN o 0 CN m 00 r- vq c |oo VO CN I? JJ r-- Tt JI- -fi , in ©' VO Q ^t •f-i Tt id m 00 § 8 o fi 8 US '•3 •2 3s \= "s s s 0 CJ 6, 0 VO s : li, &|O N 0 in CU l VO Tt u in ON Tt • H-» o O CN 1 § | o CJ n t i nt h o n ( iel d ft CJ 13 '? "fi A CJ >-> s s cj o CJ CU cu H-» 13 ft -fl T3 C+Tt H 00 43 N c " ' >-> O CJ 00 t3 CJ 0 ~ei o CJ 0 CN 00 00 CJ fl -O 2 00 fl UH o CJ a UH vd fl ft CJ a fl" 'fi O g UH a CJ H-H CU cu ft n conte s i d cons n disso l UH fi fl lutio n p ft eoretic a CJ '00 Clconc e CJ o CJ 0 0 O 0 43 0 o s UH UH w UH ft E—' PC 00 r-H O HH H •8 CJ C/3 o , X U W H < starting with low acid concentration and titrating with high acid concentration will allow better manipulation of the reaction extent, and avoid both the abrupt evolution of gases and the increased tendency for side reactions to take place. Consequently, low acid concentration was used, and titration was done with concentrated HC1 solution (5 or 10 M solution). 6.4.1.2 Ferrous chloride effect: The leaching reactions were observed not to be affected by additions of ferrous chloride, which can also be generalized to sulfate solutions, since the ferrous ion as ferrous sulfate is fairly soluble. This is understandable by recalling that Eq. 37 is a non-equilibrium non-catalytic solid- fluid reaction, and thereby, Le Chatelier's principle is not expected to apply. Back reaction kinetics should also be negligible because of the driving force behind Eq. 37 (the highly negative free energy of change), which makes that reaction less likely to be reversible (Section 4). This was proven experimentally as per Table 6.25. According to this table, additions of ferrous chloride has no significant effect on reaction kinetics. The apparent decrease in conversion is attributed to experimental errors. The neutral effect of the ferrous ion was also proved by chemical analysis, since wet chemistry methods showed that there is no precipitated iron in the reaction residue that would impede the leaching. This was also ensured by nitrogen purging, as was discussed in Section 5.2.2. With the good solubility of ferrous chloride in acidic solutions (Table 2.2), up to 3 M of FeCl2.4H20 solution can be prepared and added to the system, and good conversion is achievable. In other words, under a heavy recycle load of ferrous solution, leaching is not hindered, and the process flexibility is now established. In addition, this implies that the process can straightforwardly be integrated into suitable hydrometallurgical flowsheets. The real reasons behind the observed neutral effect of dissolved iron are not known, but can be attributed to the solubility of ferrous chloride and the high corrosion rate of powdery iron. More investigation is needed to uncover the real reasons behind this effect (see Section 8). 6.4.1.3 Metallic iron effect The amount of iron added to the leaching stage is near stoichiometric. The excess iron has several purposes. First, it will improve the conversion of the system, as was explained in the 142 kinetic study. Second, it will improve the removal of dissolved copper from solution (by cementation). Third, it will allow running the system at high SPD without any troubles or increased possibility of side reactions. In the system studied, around 20% excess iron was used and found to be satisfactory. Table 6.26 shows the results for experiments done with different metallic iron additions. The best obtained conversion was that when almost twice the stoichiometric amount of metallic iron is added. If the experimental error is excluded, it is clear that iron increments beyond this amount has no significant effect on final conversion. These observations confirm the findings of the kinetic study. When near stoichiometric iron was used, almost half the iron component of chalcopyrite was released, and this result is better than that obtained under lower SPD (Table 6.21). The latter finding suggests that it can be improved by some thermodynamic change or physical change to the system. A physical change was later utilized when finer particles were used, which proved to be of benefit. Experimental objective : To investigate the effect of metallic iron using high SPD Experimental conditions Temperature 25 °C Agitation speed 650 rpm Ferrous chloride addition 149.1 g (3 M solution) Solid pulp density 35.59% Solution composition 0.1MHC1 Solution pH -0.26 Concentrate type Gibraltar chalcopyrite concentrate Particle size -149 um+74 um Experimental results : Term Run 1 Run 2 Run 3 Run 4 Iron added as excess stoichiometric 182.14% 212.50% 242.86% 121.43% SPD 35.02% 35.59% 36.14% 33.86% Iron content in the leach residue 18.82% 14.85% 20.38% 21.40% Copper content in the leach residue 38.10% 38.96% 35.40% 34.32% Iron dissolution (conversion) 68.89% 85.28% 77.00% 51.37% Theoretical yield for iron 74.25% 80.25% 62.53% 54.91% Acid consumption (10 M HC1) 162.9 ml 173.84 ml 123.52 ml 115.76 ml Table 6.26 : Effect of metallic iron addition on leaching using high SPD systems. The leach mixture contains 3 M FeCl2.4H20 solution. 143 The excess amount of iron can be premixed with the concentrate prior to being fed to the reactor, or part of it can be added to the slurry leaving to the separation unit to ensure no copper is lost with the leach solution. The other reason for this modification (stagewise addition of iron) is to give more flexibility to the proposed process, if another reductant is available for use, like lead, as was suggested by Dreisinger (1999). The cemented reductant can thus be recycled again to the reaction vessel. Also, this modification might lead to higher conversion or better utilization of iron as a reductant. Iron as a reductant has also some incentives, such as : 1) Low price. If scrap iron, as in copper cementation, or hot briquetted iron (HBI) could be used, this will offer good savings. 2) Availability 3) Iron will not report as an associated impurity in the production cycle of copper, except for any contained pyrite, since an iron-rich leach solution is produced, and so a troublesome impurity in hydrometallurgy is avoided. This solution can be discarded by a simple solid-liquid separation technique, with the possible subsequent treatment for regenerating different reagents. 4) The sulfides of iron are sufficiently soluble in acid solutions, and as metallic iron undergoes anodic dissolution, this will lead to the formation of soluble species (FeCl2 or FeS04, depending on the selected lixiviant) rather than gelatinous compounds, like Fe(OH)3. Apparently, there are different options for iron removal, depending on process economics. Iron, as a reductant, can be regenerated, be removed from effluents by manipulating the solution pH and utilizing the precipitation effect of hydrogen sulfide, or be treated by other methods to regenerate different reactants or for safe disposal. 6.4.1.4 Particle size effect The particle size of the chalcopyrite concentrate has a significant effect on the leaching reactions, as was confirmed from the kinetic study. Under the employed leaching conditions of high SPD and ferrous chloride additions, the same pattern of dependence is expected to remain. Experiments were performed for various size fractions as per the specified reactant amounts in Table 6.27. This table confirms the claim that the conversion is largely dependent on 144 the employed particle size fraction. For near stoichiometric additions of metallic iron, about 80% of the initial iron content in chalcopyrite can be removed using fine size fractions, smaller than 74 um, under mild conditions of temperature, with a large addition of ferrous salt. The table also shows that copper is retained in the leach residue. This retaining of copper is caused by the reduction reactions and is further enhanced by the precipitation effect of hydrogen sulfide and the cementation effect of excess iron. Experimental objective : To investigate the effect of particle size using high SPD Experimental conditions Temperature 25 °C Agitation speed 650 rpm Ferrous chloride addition 149.1 g (3 M solution) Solid pulp density 33.86% Iron added as excess stoichiometric 121.43% Solution composition 0.1 MHC1 Solution pH -0.26 Concentrate type Gibraltar chalcopyrite concentrate Experimental results Term Run 1 Run 2 Run 3 Run 4 Run 5 Chalcopyrite size fraction -149 um -74 um -63 um -53 um -44 um +74 um +63 um +53 um +44 um +38 um Iron content in the leach residue 21.40% 19.86% 17.50% 16,77% 16.11% Copper content in the leach residue 34.32% 39.11% 39.71% 40.61% 41.27% Iron dissolution (conversion) 51.37% 66.87% 76.59% 80.58% 83.93% Theoretical yield for iron 54.91% 69.22% 77.27% 80.64% 83.34% Acid consumption (10 M HC1) 115.76 ml 144.61 ml 166.32 ml 175.47 ml 185.29 ml Table 6.27 : Effect of particle size on leaching using high SPD systems. The reaction mixture contains 3 M FeCl2.4H20 solution. Also, it is noted that best yield for iron, that is the fraction utilized in reduction leaching, is obtained when finer particle sizes are used, suggesting that the severe competition from the hydrogen evolution reaction is minimized, as was stated in Section 6.1.3. Figs. 6.47-6.49 summarize the results for leaching tests in graphical form, which justify the conditions employed in the leaching stage. Fig. 6.47 is a representation of the common pattern of conversion curves with decreasing particle size, which is similar to that shown in Fig. 6.37. The reverse trend is seen in Fig. 6.48 for iron content in the enriched concentrate, while Fig. 6.49, for copper content in the enriched concentrate, retained the same pattern as Fig. 6.47. 145 Fig. 6.49 suggests that a large weight reduction can be obtained by this process (up to 30%), which makes this method of leaching attractive. 0 10 20 30 40 50 60 70 80 90 100 110 120 Mean particle diameter, |im Fig. 6.47 : Plot of conversion vs. chalcopyrite mean particle diameter (high SPD, 3 M FeCl2.4H20 solution, 25 °C). The data are from Table 6.27. It is concluded here that higher iron and sulfur release from the concentrate (enrichment) could be achieved using much finer particle sizes (smaller than 38 um or 400 mesh). The same is expected to occur if the amount of metallic iron and/or the leach temperature are increased, as was outlined in Section 6.1.8. Additional metallic iron will enhance the conversion as was confirmed from Section 6.1.5, while increasing the temperature would also allow higher recycle load of ferrous chloride since the latter solubility (Table 2.2) increases with increasing temperature. As a final note on Tables 6.24-6.27, the negative value of measured solution pH at different acid concentrations should not confuse the reader, since excessive amounts of ferrous chloride were added, and their hydration is expected to affect the solution pH reading. 146 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Mean particle diameter, nm Fig. 6.48 : Plot of iron content in the enriched concentrate vs. chalcopyrite mean particle diameter (high SPD, 3 M FeCl2.4H20 solution, 25 °C). The data are from Table 6.27. 33 J 0.0 20.0 40.0 60.0 80.0 100.0 120.0 Mean particle diameter, fxm Fig. 6.49 : Plot of copper content in the enriched concentrate vs. chalcopyrite mean particle diameter (high SPD, 3 M FeCl2.4H20 solution, 25 °C). The data are from Table 6.27. 147 6.4.2 Solid/liquid separation After a certain period of time, around one hour, part of the reaction mixture is removed to the solid-liquid separation unit (a thickener) where a stream containing ferrous chloride, as well as some enriched solids, is produced and recycled back to the main reactor. According to Dreisinger (1999) the need for recycling some enriched solids, i. e. chalcocite, is to enhance the conductivity of the solid mixture in the reactor, since chalcocite is a good conductor (Table 1.2), and maintain the high SPD that will continue to promote the galvanic mechanism as was explained in Section 6.1.7. 6.4.3 Iron removal From the solid-liquid separation unit, a stream of ferrous chloride is produced and envisaged to be fed to a spray roaster, where it is pressure oxidized to generate part of the lixiviant (HC1 solution) and dispose of iron as hematite. The spray roasting technique is widely practiced and there are several studies on the possibility of generating HC1 solution from spent acid liquors (Peters (1992)). In addition, the kinetics of the pressure oxidation of iron (II) in aqueous sulfate or chloride solutions are well studied and there are different published works on these topics (Peters (1992)). Other modifications to this stage are discussed in Section 6.5. 6.4.4 Solid washing unit Part of the recycled stream from the solid/liquid separation stage is depicted to be sent to a washing stage (a pressure filter or similar), since the solid mixture contains excessive amounts of chloride solution. In this washing unit, the enriched concentrate (copper super-concentrate) is recovered, while the spent wash water is recycled, either to the spray roaster or the thickener for mixing with fresh leach solution. The enriched concentrate can further be treated by a suitable hydrometallurgical or pyrometallurgical process to produce copper. Route selection is dependent on the process conditions and/or its integration with other process flowsheets. 148 6.4.5 Hydrogen sulfide collection unit A special collection unit for this gas is shown on Fig. 6.46. A detailed discussion on the available options for the treatment of this gas are given in the next section, but the most acceptable and feasible outlet for hydrogen sulfide is being utilized for the production of elemental sulfur, as indicated in Fig. 6.46, due to the intrinsic benefits. If the process is to be integrated with other flowsheets, this gas can be contained internally, especially when non- oxidative leaching units exist. 6.4.6 Process advantages The advantages of this proposed process are several. First, the process is simple. It comprises few treatment steps toward achieving the main goal, which is the rejection of iron and sulfur, and producing enriched chalcopyrite concentrates. Thus, the required number of unit operations is minimized. The rejection of iron and sulfur implies good weight reduction, which is further enhanced by recalling that chalcocite is seemingly the new solid phase. Second, its recyclability. Almost every stream is recycled for different purposes. The leach solution (supernatant) is oxidized to regenerate HC1 solution, which is returned back to the main reaction vessel. Hence, only make-up acid is needed (theoretically). Part of the slurry is recycled to the reaction vessel to improve solid conductivity and control the SPD. Wash water is also recycled to other separation units. Third, its environmental compatibility. Iron, after spray roasting, is rejected as hematite, as with other hydrometallurgical processes. Hydrogen sulfide gas is not released to the atmosphere, as the case with S02 release in some pyrometallurgical processes. H2S gas is envisioned to be collected in a special treatment unit as discussed in the next section. The possibility of recovering sulfur in elemental form will add to the competitiveness and attractiveness of this process. Fourth to be considered are the operating conditions. In the laboratory, mild conditions were employed. Experiments were performed with continuous stirring and at room temperature. Preparing ferrous chloride solution is very simple, due to its good solubility, which simulates a large recycle load of leach solution. This feature is desired as it implies flexible operation and process control, besides achieving the main objective, which is chalcopyrite enrichment. As with 149 other metallurgical processes, size reduction units are needed to give the required size fraction, which can range between -74 um to +38 um (-200 mesh to +400 mesh), or other finer size fractions. 150 6.5 HYDROGEN SULFIDE TREATMENT In Fig. 6.46, a special collection unit for hydrogen sulfide was attached to the leaching stage. The evolution of hydrogen sulfide forms a necessity for such collection units and this gas needs a suitable outlet, especially for large scale applications. The possibility for utilizing this product in hydrometallurgical applications or in a comprehensive process flowsheet needs to be further investigated. The possible alternatives for recovery and/or treatment are briefly discussed and more information can be found in Weil and Sandler (1992). Hydrogen sulfide, H2S, is a colorless gas having a characteristic rotten-egg (pungent) odor. Table 6.28 summarizes its physical and thermodynamic properties. Hydrogen sulfide is very soluble in alkanolamines, which are used as scrubbing solvents for its removal from gas streams. Among its various uses, H2S is known to be a good precipitation agent, and is mainly used in recovering metals from solutions, cleaning effluents by removing their heavy metal contents, and in chemical analysis. Other possible applications are sulfuric acid production and/or conversion to marketable elemental sulfur (by reduction with strong sulfuric acid solutions, bacterial or pressure oxidation). There are two general hydrometallurgical applications with hydrogen sulfide : the bulk precipitation of metal sulfides from weak aqueous solutions for the purpose of concentration, and the selective precipitation of some particular metal or metals in the presence of others for the purpose of purification or refining. This is controlled by the solution pH and the gas partial pressure. For instance, dissolved copper in solutions can selectively be precipitated in the pH range 0 to 1, and separated from solutions containing cobalt, nickel, iron, and manganese ions. As a note, this pH range was used in this research. Hydrogen sulfide can be used to produce elemental sulfur and hydrogen gas : 2H2S(g)<->2H2(g)+S2(g) (65) The decomposition enthalpy for this reaction at -850 °C is 16 kJ/mol. Below this temperature, a catalyst, like silica or cobalt molybdate, is needed. Nonetheless, hydrogen yield will be greater than theoretical estimates, due to the formation of sulfur species, other than S2. This option is important for refineries that require large amounts of hydrogen in the reforming/cracking units. Hydrogen sulfide can be utilized for the production of different reagents. The reaction of H2S with one molar equivalent of sodium hydroxide gives sodium hydrosulfide (NaHS); with 151 two molar equivalents of NaOH, sodium sulfide (NajS) forms, as was used in this research. NajS and NaHS have many applications in dyes, rubber chemicals, pesticides, pharmaceuticals, and in Kraft pulping. Property Value Molecular weight 34.08 Melting point, °C -85.53 Boiling point, °C -60.31 Latent heat of fusion, kJ/mol 2.375 Latent heat of vaporization, kJ/mol 18.67 Density at -60.31 °C, kg/m3 949.6 Sp. gr., gas (based on sp. gr. of air, 79 mol % N2 and 21 mol % 02, = 1) 1.182 Critical temperature, °C 100.38 Critical pressure, kPa 9006 Critical density, kg/m3 346.0 Standard free energy of formation (AGf°), kJ/mol -33.6 Standard free enthalpy of formation (AHf°), kJ/mol -20.6 Entropy of formation (S), at 25 °C, J/mol per K 205.7 Specific heat, Cp, at 27 °C, J/mol per K 34.2 Autoignition temperature in air, °C ca 260 Explosive range in air at 20 °C, vol. % Upper limit of ignition (ULI) 46 Lower limit of ignition (LLI) 4.3 Vapor pressure, kPa -60 °C 102.9 -40 °C 257.9 -20 °C 562.0 0 °C 1049 20 °C 1814 40 °C 2937 60 °C 4480 Solubility in water, g per 100 g solution at 1 atm 0 °C 0.710 10 °C 0.530 20 °C 0.398 Table 6.28 : Physical and thermodynamic properties of hydrogen sulfide (Weil and Sandler (1992)) Hydrogen sulfide can be oxidized by a number of oxidizing agents, to yield different products. As indicated in the previous section, the most feasible and desirable route is that for 152 elemental sulfur production. The following table summarizes some of these oxidation reactions, and the actual products are functions of the oxidant quantity and reaction operating conditions : Oxidizing agent Conditions Sulfur-containing products o2 Flame S02, some S03 Flame or furnace, catalyst (Claus process) Sulfur Aqueous solution of H2S, catalyst Sulfur 2 2 H202 Neutral; alkaline solution Sulfur, S203 \ S04 " Na202 Dry, elevated temperature Na2S, Na^ o3 Aqueous solution Sulfur, H2S04 S02 Elevated temperature, catalyst Sulfur Aqueous solution (Claus process) Sulfur, polythionic acids H2S04 Concentrated acid ( « 18 M) Sulfur, some S02 HN03 Aqueous solution H2S04 NO Silica gel catalyst Sulfur N02- Aqueous solution at pH 5-7 Sulfur, NO Aqueous solution at pH 8-9 Sulfur, NH3 Cl2 Gaseous mixture, excess Cl2 SC12 Gaseous mixture, excess H2S Sulfur Aqueous solution, excess Cl2 H2S04 Aqueous solution Sulfur (quantitative) 3+ 2 Fe Aqueous solution Sulfur, FexSv, S04 " Bacteria Aqueous solution and/or high SPD streams Sulfur and others Table 6.29 : Oxidation reactions of hydrogen sulfide (Weil and Sandler (1992)) In industry, the most widely used of these is the oxidation of hydrogen sulfide in a flame for producing sulfur dioxide, which can then be converted to sulfuric acid. The oxidation of hydrogen sulfide by sulfur dioxide, also in a flame, is the basis of the Claus process for sulfur production : St£pS 2H2S(aq) + S02(g) ^ > ! S8(s) + H20(1) (66) The Claus reaction can also take place at milder conditions in the presence of water, which catalyzes the reaction. The oxidation of H2S by S02 in water is a complex process, leading to the formation of sulfur and polythionic acids, the mixture known as Wackenroeder's liquid. Further, this mixture can be utilized to produce sulfuric acid by a series of steps which can be controlled 153 by proper temperature selection and additives. This, in turn, implies flexible operation, which explains the wide adoption of Claus process. Hydrogen sulfide can be converted to elemental sulfur by oxidizing with sufficient amounts of oxygen. The reaction will also lead to the formation of water, and this is another form of the Claus process : 2H2S(g) + 02(g)^2S(s) + 2H20(1) (67) This in fact was practiced in the Sherritt-Cominco process described in Section 2.3. By knowing the operating conditions of pressure and temperature, and other related parameters, flexible control of H2S gas is possible. There are numerous kinetic studies for the reactions to convert hydrogen sulfide to elemental sulfur by a Claus process. More information can be found in Weil and Sandler (1992). Hydrogen sulfide can also be decomposed by photochemical and electrochemical methods. This was suggested to be used in petroleum and/or nuclear industry, where photons or electricity will act as a catalyst for low temperature decomposition. A new area of investigation is the H2S bacterial oxidation to produce elemental sulfur. This is currently being investigated by several researchers, particularly those working in the waste water or manure treatment facilities. According to Fig. 6.46, iron in the proposed process is assumed to be removed as hematite, by spray roasting of the ferrous chloride solution. The collected hydrogen sulfide can further be used to produce more environmentally benign iron solution. It is known that various metal oxides and hydroxides react with H2S forming their sulfides (Weil and Sandler (1992)), and in the case of iron, a useful application can be obtained, where iron is removed as Fe2S3: Fe203.H20(aq) + 3H2S(aq) -> Fe2S3(s) + 4H20(1) (68) The produced iron sulfide can either be safely rejected to tailings, or, if desired, be recycled to the main reaction vessel, to adjust the SPD. Under certain conditions, it might find some commercial uses (Peters et al (1981)). In Fig. 6.46 it is shown that spray roasting is used to regenerate HC1 which is then recycled to the leaching vessel. Another variation to this option is to utilize the large amount of 454 hydrogen sulfide for some internal treatment, and this would be more convenient for sulfate media. Assuming that the supernatant from the solid-liquid separation stage is FeS04, then the spray roaster is now replaced by a pressure oxidation stage to oxidize ferrous ions to ferric ions : 2FeS04(aq) + 0.5O2(g) + II2S04(aq) -> Fe2(S04)3(aq) + 2H20(I) (69) The reason for this modification is to contain the generated hydrogen sulfide, as it is a good reducing agent. It can reduce the ferric sulfate back to ferrous sulfate : Fe2(S04)3(aq) + H2S(g)^FeS04(aq) + H2S04(aq) + S(s) (70) This variation has several advantages. First, the new variation will lead to the containment of large amounts of hydrogen sulfide (in the systems studied, 3 moles of the gas are generated per mole of iron added to the system). Second, it will generate some of the sulfuric acid which is required in the leaching stage. Third, it will regenerate the ferrous sulfate required for Eq. 69, which can further be oxidized to get rid of iron as goethite or hematite (by hydrolysis, and depending on the extent of acid consumption). Fourth, it can produce some elemental sulfur, which is desired due to interim storage simplifying and other benefits, although this may require another separation step. The only reagent needed here is oxygen, which is cheap, and the recycle nature of this new variation is established. The same modification can be extended to chloride systems. The corresponding reactions will be : 2FcCl2(aq) + 0.5O2(g) + 2HCl(aq) -» 2FeCl3(aq) + H20(1) (71) 2FeCl3(aq) •+H2S(g) -> 2FeCl2(aq) + 2HCl(aq) + S(s) (72) This modification to the flowsheet seems to be possible, both chemically and technically. The kinetics of Eqs. 69 and 71 are well studied in the open literature (Peters (1992)), and such an oxidation is practiced in industry, e.g. : zinc pressure leaching. However, the kinetics of Eqs. 70 and 72, and the subsequent difficulties of sulfur removal, are not well studied. It is recommended here to study the conditions which would affect the extent of these reaction, and in turn establish, for instance, the preferred circumstances to maximize the yield of elemental sulfur (conversion and morphology). It is worth mentioning here that hydrogen sulfide has a dispersive effect on sulfur, as it is a good solvent for the latter (Weil and Sandler (1992)). 155 } 7. CONCLUSIONS A fundamental study of the reductive decomposition of chalcopyrite using metallic iron was performed. This study has shown that chalcopyrite can be reductively leached to yield a new solid phase, chalcocite, that fixes copper. The iron component of chalcopyrite was shown to enter the leach solution, while part of the sulfide sulfur component was released as hydrogen sulfide. The net leaching reaction for this system is : 2CuFeS + Fe + 6H+ -> Cu^ + 3Fe2+ + 3H S 2(s) (s) (aq) (aq) 2 '(g( ) with a two-component mechanism (anodic and cathodic reactions). Its stoichiometry was established using detailed chemical analyses, supported by SEM/EDX analysis. The thermodynamic and kinetic analysis of the system show that the leaching reactions are controlled by a transport process through the new solid phase, which can comprise the diffusion of one or more species. Hydrogen ion diffusion was shown to be rate limiting. SEM/EDX testing confirmed the presence of a porous product layer and the validity of the schematic leaching model. The leaching reactions were found to follow the shrinking core model, and the parabolic leaching kinetics were established, that, under the experimental conditions, led to the following models : Sulfate media: Chloride media : The leaching kinetics were shown to be dependent on acid concentration, chalcopyrite particle size and, to a less extent, temperature. The reaction rates were found to be directly related to the reductant/chalcopyrite molar ratio but not dependent on the chalcopyrite amount. 156 For the tested chalcopyrite concentrate, the maximum conversion (decomposition of chalcopyrite to simpler copper sulfides) was obtained using the following leaching conditions, for both chloride and sulfate media : Temperature 65 °C Particle size smaller than 74 pm (200 mesh) Molar ratio of iron to chalcopyrite 2 to 1 Acid concentration 0.6 M with mild agitation. For high SPD systems, up to 4 M ferrous chloride solution can be added, with no severe adverse effect on the alteration reaction. Under mild conditions of ambient temperature and fine particles (< 74 pm), greater than 80% conversion or removal of the chalcopyrite iron component can be achieved, at around 120% stoichiometric iron addition. Further size reduction should allow improvement of conversion and iron utilization. A simple process flowsheet for the reductive decomposition of chalcopyrite using metallic iron was proposed that utilized the findings from the kinetic study. The process is simple and bears the common traits of hydrometallurgy. The proposed flowsheet envisaged a suitable solution to the generated hydrogen sulfide gas as well as iron removal. More investigation is required toward the complete development of a new process for the production of copper super-concentrates from chalcopyrite by the method of reductive decomposition with metallic iron. 157 8. RECOMMENDATIONS FOR FUTURE RESEARCH There are several suggestions for the continuation of this research, toward the complete development of a new process for the reductive decomposition of chalcopyrite. Also, there are some recommendations for other related fundamental studies. Examples of such recommendations are: 1) Studying the effect of very fine particles : According to Table 6.27, the conversion is increasingly improved when finer particle sizes are used. It is recommended here to use very fine particle sizes (< 10 um) as this implies two benefits. First, the severe competition from the side reactions is minimized, or selective leaching is achieved, as was shown in Section 6.1.3. Second, since the leaching is controlled by a transport process in the product layer, this means that finer particles result in shorter diffusional paths for the protons or other species, which is expected to significantly improve the enrichment of chalcopyrite. Such fine particles can be produced, for example, by attrition grinding, turbomilling or oscillating milling that includes high impact. These methods are also expected to induce some mechanical activation to the concentrate, which might prove to be of benefit to the alteration reactions. 2) Studying the galvanic interactions : So far, the galvanic interactions between chalcopyrite and other minerals in the concentrate, particularly pyrite, were not established. This requires an electrochemical study to understand the mechanism and contribution of such interactions to the overall conversion or leaching mechanism (if any). 3) A detailed electrochemical study of the system : A detailed electrochemical study for the leaching reactions is required for several purposes. An electrochemical study using a CuFeS2 electrode at cathode potential similar to that in a directly connected Fe-CuFeS2 couple would confirm the formation of chalcocite or similar copper sulfides, with iron entering the solution and hydrogen sulfide being released. Such a study should be able to predict the parabolic leaching dependence on hydrogen ions, or other species (if any), in that the leaching model should include such terms. 158 Charge transfer steps were assumed not to be rate limiting. This can be confirmed by such an electrochemical study. If the rate showed no voltage dependence then charge transfer steps are not rate limiting. The proposed electrochemical study will also quantify the molar volume changes that accompany the alteration reactions, which will be useful for different purposes, as was outlined in Section 6.2. Such a study will also account for the electrical conductivity and electron mobility through the product layers, as was stated in the same section. Finally, such a study will give more explanation for the finding that excessive additions of ferrous chloride do not have severe adverse effect on leaching kinetics. 4) Mixed kinetics model fitting : Experimental data were fitted by the product layer diffusion control model, which appeared to be satisfactory. However, more investigation is required, specifically in estimating the activity coefficients and their temperature dependence. An attempt to fit the experimental data by mixed kinetics models was not done, but may provide another tool in understanding the reaction kinetics. Mixed potential measurement will confirm or refute the statement given in Section 6.2 in that chalcopyrite reductive leaching by metallic iron is controlled by both the anodic and cathodic portions of the net leaching reactions. Moreover, the kinetic study and SEM testing showed that a layer or film of products form on the chalcopyrite particle. The electrochemical role of these films deserves further investigation which can be incorporated in the proposed electrochemical study. 5) Gas chromatography (GC): As was noted in this research, the gaseous products were not analyzed, but were absorbed in an alkali solution. Analysis of gaseous products by GC would help in confirming the stoichiometry of the leaching reactions, the extent of side reactions and any possible associated hazards or conversion improvement. Analysis for H2S gas would allow precise estimation of the extent of chalcopyrite conversion, since the source of the sulfide component will be only from the concentrate. 159 6) The effect of additives : As noted from the results of this research, the hydrogen evolution reaction (HER) is the main problem for this method of leaching. It is a severe competitor toward the successful exploitation of the added iron in chalcopyrite reduction. A possibility is to lessen the tendency of HER by seeking for certain additives that are capable of hindering or suppressing it. According to Tromans (1999), there are some organic materials that are capable of synergistic action on corrosion inhibition. Such surfactants were found to be useful on corrosion inhibition of mild steel in chloride solutions. Also, there are some additives that were found to selectively retard the hydrogen evolution reaction on copper, like benzotriazole. It is expected that similar chemicals may prove suitable for similar action on iron. The main issue to be addressed is the selectivity of the inhibitor. That is to allow iron corrosion, which is a part of the leaching mechanism, and decrease the tendency of hydrogen evolution reaction, which imposes the use of twice the stoichiometric amount of iron and limit the viability of this method of leaching. There are two known reagents that were found to be useful for this selective inhibition : acridine and straight-chain alkyl amines that contain one to four carbon atoms. Other reagents include : benzothiazole, mercaptothiazole, n-hexadecyl amine, alkyl paraffins or a combination of such reagents. 7) X-Ray Diffraction (XRD)/X-Ray Photoelectron Spectroscopy (XPS)/Auger Electron Spectroscopy (AES) studies : A detailed quantitative study using these methods to identify the nature and composition of the products is also recommended. This will help in the quantitative analysis of reaction products, as was found using wet chemistry methods and SEM/EDX analysis. These studies will assist in finding the causes and mechanism of any passivation and possible remediation. XRD should be capable of identifying the reaction products, and reveal the presence of other products or intermediate species. As explained in Section 4, in addition to chalcocite, there is a possibility for the formation of bornite or other intermediate solid phases. Hence, AES may be required to disclose the exact nature and/or composition of the product layer. 160 8) Kinetic and transport studies : The findings from this research can be utilized as a starting point in a fundamental kinetic study to investigate the conditions under which certain leaching mechanisms or control models will prevail. The purpose is to develop a systematic method for confirming the prevailing leaching mechanism and maximizing the extent of desired leaching reactions (that is selectivity): The published literature lacks data on the fundamental transport properties of different species, particularly in concentrated and mixed electrolytes, as well as semiconducting minerals. Examples include the activity and diffusivity. 9) Other modifications to the proposed process : The addition of excess amounts of ferrous chloride proved to have no severe adverse effect on backward or forward reaction kinetics, or on conversion. The addition of other reductants may help in improving the enrichment process. Examples are lead, zinc, cadmium and others. As was indicated in Section 6.4, a variation to the developed process flowsheet (Fig. 6.46) is the use of lead for chalcopyrite enrichment, followed by cementation with iron. This will further demonstrate the flexibility and recyclability of the process. Also, as was indicated in Section 6.5, a modification to the process flowsheet in Fig. 6.46 is replacing the spray roaster with a pressure oxidation unit for the ferrous supernatant. Hence, the kinetics of Eqs. 70 and 72 need extra investigation to demonstrate both the flexibility of the process and the viability of such a modification. 10) Complete assessment of the process cost effectiveness 161 BIBLIOGRAPHY 1. Atkins, P. W., "Physical Chemistry", 5th ed., Oxford University Press, Oxford, UK, 1994. 2. Bard, A.; Parsons, R. and Jordan, R., "Standard Potentials in Aqueous Solutions", 1st ed., International Union of Pure and Applied Chemistry (IUPAC), printed by Marcel Dekker, New York, NY, 1985. 3. Bennewith, K. and Hackl, R., "An Introduction to Copper Hydrometallurgy", 2nd ed., Laboratory Manual for MMAT 358 course : Metal and Materials Engineering 358 - Hydrometallurgy I, Metals and Materials Engineering Department, The University of British Columbia, Vancouver, BC, Canada, September 1998. 4. Beckstead, L. W.; Munoz, P. B; Sepulveda, J. L.; Herbst, J. A.; Miller, J. D.; Olson, F. A. and Wadsworth, M. E., "Acid Ferric Sulfate Leaching of Attritor-Ground Chalcopyrite Concentrates", pp. 611-632, in : Extractive Metallurgy of Copper, Eds. : J. C. Yannopoulos and J. C. Agarwal, Proceedings of an International Symposium, sponsored by the Metallurgical Society of AIME, held in Las Vegas, NV, February 22-26, 1976. 5. Biswas, A. K. and Davenport, W. G., "Extractive Metallurgy of Copper", 3rd ed., Pergamon Press, 1994. 6. Carnahan, B., Luther, H. and Wilkes, J., "Applied Numerical Methods", 1st. ed., John Wiley and Sons, Inc., New York, NY, 1969. 7. Chae, D. J. and Wadsworth, M. E. " Modeling of the Leaching of Oxide Copper Ores", Utah University Press, pp. 62, Dec. 1979. 8. Champagne, J. F., "Magma : Focused on Copper for the 21st Century", pp. 71-85, in : Copper 95 - Cobre 95 - Volume I, Eds. : C. Diaz ; G. Bokovay; G. Lagos; H. Larravide and M. Sahoo, Proceedings of an International Conference, sponsored by CIM, held in Montreal, Quebec, Canada, 1995. 9. Cooper, W. C; Dreisinger, D. B.; Dutrizac, J.; Hein, H. and Ugarte G. (Eds.) : Copper 95 - Cobre 95 - Volume III - Electrorefining and Hydrometallurgy of Copper, Proceedings of an 162 International Conference, sponsored by the Metallurgical Society of CIM, held in Montreal, Quebec, Canada, 1995. 10. Cooper, W. C; Kemp, D. J.; Lagos , G. E. and Tan, K. G. (Eds.) : Copper 91 - Cobre 91- Volume III - Hydrometallurgy and Electrometallurgy of Copper, Proceedings of an International Symposium, organized by the Metallurgical Society of CIM, held in Ottawa, Canada, August 18-21, 1991. 11. Dasher, J., "Hydrometallurgy for Copper Concentrates", CIM Bulletin, Vol. 66, No. 733, pp. 48-56, May 1973. 12. Donnay, G.; Corliss, L.; Donnay, J.; Elliot, N. and Hasting, J., "Symmetry of Magnetic Structures : Magnetic Structures of Chalcopyrite", Physics Review, Vol. 112, pp. 1917-1923, 1958. 13. Dreisinger, David Bruce (Prof.), Research supervision and private communication throughout the course of this research, Metals and Materials Engineering Department, The University of British Columbia, Vancouver, British Columbia, Canada, 1999. 14. Dreisinger, D. B., "The Hydrometallurgical Treatment of Copper Concentrates : Recent Developments", special report prepared at Metals and Materials Engineering Department, The University of British Columbia, Vancouver, BC, Canada, 1997. 15. Dreisinger, D. B. and Peters, E. : UBC Shortcourse on Hydrometallurgy, Chap. 1 : Introduction to Eh-pH Diagrams, Metals and Materials Engineering Department, The University of British Columbia, Vancouver, BC, Canada, 1992. 16. Dry, M. J. and Bryson, A. W., "Kinetics of Leaching of a Low Grade Fe-Ni-Cu-Co Matte in Ferric Sulfate Solution", Hydrometallurgy, Vol. 18, pp. 155-181, 1987. 17. Duby, P., "The Thermodynamic Properties of Aqueous Inorganic Copper Systems", INCRA Series on The Metallurgy of Copper, Monograph IV, New York, NY, June 1977. Library of Congress Catalogue Card Number : 77-71709. A copy on CD can be obtained from : International Copper Association (INCRA), Ltd., 260 Madison Avenue, New York, NY, 10016. 18. Dutrizac, J., "Elemental Sulfur Formation During the Ferric Sulfate Leaching of Chalcopyrite", Canadian Metallurgical Quarterly, Vol. 28, No. 4, pp. 337-344, 1989. 163 19. Dutrizac, J. E. and MacDonald, R. J. C, "Ferric Ion as a Leaching Medium", Minerals Science and Engineering, Vol. 6, No. 2, pp. 59-100, April 1974. 20. Dutrizac, J. E.; MacDonald, R. J. C. and Ingraham, T. R., "The Kinetics of Dissolution of Chalcopyrite in Aqueous Acidic Ferric Sulfate Solutions", Trans. AIME, Vol. 245, pp. 955- 959, May 1969. 21. Emmons, W. H., "The Enrichment of Sulfide ores", US Geological Survey Bulletin 529, pp. 260, US Government Printing Office, Washington, D. C, 1913. 22. Etienne, A., "Electrochemical Aspects of the Aqueous Oxidation of Copper Sulfides", Ph. D. Dissertation, The University of British Columbia, Vancouver, British Columbia, Canada, November 1970. 23. Everett, P. K., "Development of INTEC Copper Process by an International Consortium", pp. 913- 922, in : Hydrometallurgy '94, Symposium organized by IMM and SCI, Cambridge, UK, 1994. 24. Felker, D. L. and Bautista, R. G., "Electrochemical Processes in Recovering Metals from Ores", JOM, Vol. 42, No. 4, pp. 60-63, April 1990. 25. George, D. B., "Copper", pp. 381-428, in : Kirk Othmer Encyclopedia of Chemistry and Technology, 4th ed., Vol. 7, John Wiley and Sons, Inc., New York, NY, 1992. 26. Gerlach, J. K.; Gock, E. D. and Ghosh, S. K., "Activation and Leaching of Chalcopyrite Concentrates with Dilute Sulfuric Acid", pp. 403-416, in : Hydrometallurgy '72, Eds. : D. J. I. Evans and R. S. Shoemaker, Proceedings of the Second International Symposium on Hydrometallurgy, organized by the Metallurgical Society and SME of the AIME, Chicago, Illinois, February 25-March 1, 1973. 27. Habashi, F. : Chalcopyrite, Its Chemistry and Metallurgy, McGraw-Hill Book Company, New York, NY, 1978. 28. Hackl, R. P., "Introduction to Bioleaching", notes for MMAT 592F course : Metals and Materials Engineering 592F - Biohydrometallurgy, Metals and Materials Engineering Department, The University of British Columbia, Vancouver, BC, Canada, 1998. 29. Hackl, R.; Dreisinger, D. B; and King, J., "Effect of Sulfur-Dispersing Surfactants on the Oxygen Pressure Leaching of Chalcopyrite", pp. 559-576, in : Copper 95 - Cobre 95 - 164 Volume III - Electrorefining and Hydrometallurgy of Copper, Eds. : W. C. Cooper; David B. Dreisinger; J. Dutrizac; H. Hein and G. Ugarte, Proceedings of an International Conference, sponsored by CIM, held in Montreal, Quebec, Canada, 1995a. 30. Hackl, R.; Dreisinger, D. B; Peters, E. and King, J., "Passivation of Chalcopyrite during Oxidative Leaching in Sulfate Media", Hydrometallurgy, Vol. 39, pp. 25-48,1995b. 31. Hackl, R. P.; Dreisinger, D. B. and Peters, E., "Reverse Leaching of Chalcopyrite", pp. 181- 200, in : Copper 1987 - Volume 3 : Hydrometallurgy and Electrometallurgy of Copper, Eds. : W. C. Cooper; G. E. Lagos and G. Ugarte, Proceedings of an International Conference, organized by the Metallurgical Society of CIM, the Chilean Institute of Mining Engineers and the University of Chile, held at the University of Chile, Santiago, Chile, 1987. 32. Hall, S. R. and Stewart, J. M., "Crystal Structure Refinement of Chalcopyrite", Acta Crystallographica, Section B, Vol. 29, No. 3, pp. 579-585, 1973. 33. Harris, D. C, "Quantitative Chemical Analysis", 3rd ed., W. H. Freeman and Company, New York, NY, 1991. 34. Haver, F. P. and Wong, M. M., "Recovery of Copper, Iron and Sulfur from Chalcopyrite Concentrate using a Ferric Chloride Leach", JOM, Vol. 23, No. 2, pp. 25-29, February 1971. 35. Hiskey, J. B., "Chalcopyrite Semiconductor Electrochemistry and Dissolution", pp. 949-969, in : Extractive Metallurgy of Copper, Nickel and Cobalt, Volume I : Fundamental Aspects, Eds. : R. G. Reddy and R. N. Weizenbach, TMS, Warrendale, PA, 1993. 36. Hiskey, J. B. and Wadsworth, M. E, "Electrochemistry of Sulfide Minerals", pp. 303-325, in: Process and Fundamental Considerations in Selected Hydrometallurgical Systems, Ed. : M. C. Kuhn, SME/AIME, New York, NY, 1981. 37. Hiskey, J. B. and Wadsworth, M. E, "Galvanic Conversion of Chalcopyrite", Metall. Trans., Vol. 6B, pp. 183-190, March 1975. 38. Jones, D. L. and Peters, E., "The Leaching of Chalcopyrite with Ferric Sulfate and Ferric Chloride", pp. 633-653, in : Extractive Metallurgy of Copper, Eds. : J. C. Yannopoulos and J. C. Agarwal, Proceedings of an International Symposium, sponsored by the Metallurgical Society of AIME, held in Las Vegas, NV, February 22-26,1976. 165 39. King, J. A and Dreisinger, D. B., "Autoclaving of Copper Concentrates", pp. 511-534, in : Copper 95 - Cobre 95 - Volume III - Electrorefining and Hydrometallurgy of Copper, Eds. : W. C. Cooper; David B. Dreisinger; J. Dutrizac; H. Hein and G. Ugarte, Proceedings of an International Conference sponsored, by CIM, held in Montreal, Quebec, Canada, 1995. 40. Kruesi, P. R.; Allen, E. S. and Lake, J. L., "CYMET Process : Hydrometallurgical Conversion of Base Metal Sulfides to Pure Metals", CIM Bulletin, Vol. 66, No. 734, pp. 81- 87, June 1973. 41. Kuhn, M. C, Arbiter, N. and Kling, H., "Anaconda's Arbiter Process for Copper", CIM Bulletin, Vol. 67, No. 742, pp. 62-73, February 1974. 42. Lee, J. M. and Donofrio, J. R, "Copper Leaching from Chalcopyrite in Hydrochloric Acid", pp. 116-126, in : Processing of Energy and Metallic Minerals, Eds. : Sohn, H. Y.; Dill, S. D.; Wie, J. -M. and Sastry, K. V. S., AICHE Symposium Series, Vol. 216, No. 78, AICHE, New York, NY, 1982. 43. Levenspiel, O., "Chemical Reaction Engineering", 2nd ed., John Wiley and Sons, New York, NY, 1972. 44. Lowe, D. F., "The Kinetics of the Dissolution Reaction of Copper and Copper-Iron Sulfide Minerals using Ferric Sulfate Solutions", Ph. D. Dissertation, The University of Arizona, Tucson, AZ, 1970. 45. Lower, G. W. and Booth, R. B., "Recovery of Copper by Cyanidation", Mining Engineering, Vol. 17, No. 11, pp. 56-62, November 1965. 46. Majima, H. and Peters, E., "Oxidation Rates of Sulfide Minerals by Aqueous Oxidation at Elevated Temperatures", Trans. AIME, Vol. 236, pp. 1409-1413, October 1966. 47. Munoz, P. B.; Miller, J. D. and Wadsworth, M. E, "Reaction Mechanism for the Acid Ferric Sulfate Leaching of Chalcopyrite", Metall. Trans., Vol. 10B, pp. 149-158, June 1979. 48. Nicol, M. J., "Electrochemical Aspects of Non-oxidative Decomposition of Sulfide Minerals", pp. 177-195, in : Hydrometallurgy, Research, Development and Plant Practice, Eds. : K. Osseo-Asare and J. D. Miller, Proceedings of the 3rd International Symposium on Hydrometallurgy, sponsored by the Metallurgical Society of AIME and the Mineral 166 Processing Division of SME-AIME, held at the 112th AIME annual meeting, Atlanta, GA, March 6-10, 1983. 49. Nicol, M. J., "Mechanism of Aqueous Reduction of Chalcopyrite by Copper, Iron and Lead", Trans. IMM, Vol. 84, Section C, pp. 206-209, December 1975. 50. O'Connor, D. J.; Sexton, B. A. and Smart, R. St. C. (Eds.) : Surface Analysis Methods in Material Science, Springer-Verlag, Berlin, Germany, 1992. 51. Opekar, Frantisek and Beran, Premysl, "Rotating Disk Electrodes", Journal of Electroanalytical Chemistry, Vol. 69, pp. 1-105,1976. 52. Pankratz, L. B.; Mah, A. D. and Watson, S. W. : Thermodynamic Properties of Sulfides, USBM Bulletin No. 689, US Department of the Interior, Washington, D. C, 1987. 53. Pawlek, F. E., "The Influence of Grain Size and Mineralogical Composition on the Leachability of Copper Concentrates", pp. 690-705, in : Extractive Metallurgy of Copper, Eds. : J. C. Yannopoulos and J. C. Agarwal, Proceedings of an International Symposium, sponsored by the Metallurgical Society of AIME, held in Las Vegas, NV, February 22-26, 1976. 54. Paynter, J. C, "A Review of Copper Hydrometallurgy", J. S. Afr. IMM, Vol. 74, No. 4, pp. 158-170, November 1973. 55. Peters, E., "Hydrometallurgical Process Innovation", Hydrometallurgy, Vol. 29, No. 1-3, pp. 431-459, March 1992. 56. Peters, E., "Electrochemical Mechanisms for Decomposing Sulfide Minerals", pp. 343-361, in : Proceedings of the International Symposium on Electrochemistry in Mineral and Metal Processing, Eds. : P. E. Richardson, S. Srinivasan and R. Woods, The Electrochemical Society, Pennington, NJ, 1984. 57. Peters, E., "Direct Leaching of Sulfides : Chemistry and Applications", Metall. Trans., Vol. 7B, pp. 505-517, December 1976. 58. Peters, E.; Swinkels, G. M and Vizsolyi, A., "Copper Recovery from Sulfide Concentrates by the UBC - Cominco Ferric Chloride Leach Route", pp. 71-81, in : Process and Fundamental Considerations of Selected Hydrometallurgical processes, Ed. : Martin C. Kuhn, SME/AIME, New York, NY, 1981. 167 59. Peters, E. and Loewen, F., "Pressure Leaching of Copper Minerals in Perchloric Acid Solutions", Metall. Trans., Vol. 4, pp. 5-14, January 1973. 60. Peters, E.; Warren, I. H. and Veltman, H. (Eds.) : First Tutorial Symposium on Hydrometallurgy, The University of Denver, Denver, CO, 1972. 61. Peters, E. and Majima, H., "Electrochemical Reactions of Pyrite in Acid Perchlorate Solutions", Can. Met. Q., Vol. 7, No. 3, pp. 111-117, July-September 1968. 62. Peterson, R. D. and Wadsworth, M. E., "Solid, Solution Reaction in the Hydrothermal Enrichment of Chalcopyrite at Elevated Temperatures", pp. 275-291 in : EPD Congress 1994, Ed.: G. W. Warren, TMS, Warrendale, PA, 1994. 63. Plyasunova, N. V.; Wang, M.; Zhang, Y. and Muhammed, M., "Critical Evaluation of Thermodynamics of Complex Formation of Metal Ions in Aqueous Solutions - II. Hydrolysis and Hydroxo-complexes of Cu2+ at 298.15 K", Hydrometallurgy, Vol. 45, pp. 37-51, 1997. 64. Pourbaix, M. : Atlas of Electrochemical Equilibria in Aqueous Solutions, pp. 307-321 and 384-392, NACE edition, Houston, TX, 1966. 65. Prater, J. D.; Queneau, P. B. and Hudson, T. J., "Nitric Acid Route to Processing Copper Concentrates", Trans. AIME-SME, Vol. 254, pp. 117-122, June 1973. 66. Rice, D. A.; Cobble, J. R. and Brooks, D. R., "Effects of Turbomilling Parameters on the Simultaneous Grinding and Ferric Sulfate Leaching of Chalcopyrite", USBM RI 9351, US Department of the Interior, Washington, D. C, 1990. 67. Robie, R. A. and Hemingway, B. S. : Thermodynamic Properties of Minerals and Related Substances at 298.15 K and 1 Bar (105 Pascals) Pressure and at Higher Temperatures, US Geological Survey Bulletin 2131, US Government Printing Office, Washington, D. C, 1995. 68. Roman, R. J. and Benner, B. R., "The Dissolution of Copper Concentrates", Mineral Science and Engineering, Vol. 5, No. 1, pp. 3-24, January 1973. 69. Schweitzer, F. W. and Livingston, R. W., "DUVAL'S CLEAR Hydrometallurgical Process", pp. 221-227, in : Chloride Electrometallurgy, Ed. : Peter D. Parker, Proceedings of 111th AIME annual meeting, Dallas, TX, Feb. 15-16,1982. 168 70. Senanayake, G. and Muir, D. M., "Speciation and Reduction Potentials of Metal Ions in Concentrated Chloride and Sulfate Solutions Relevant to Processing Base Metal Sulfides", Metall. Trans., Vol. 19B, pp. 37-45, February 1988. 71. Shirts, M. B.; Winter, J. K.; Bloom, P. A. and Potter, G. M., "Aqueous Reduction of Chalcopyrite Concentrate with Metals", USBM RI 7953, US Department of the Interior, Washington, DC 20240, 1974. 72. Smith, J. M., "Chemical Engineering Kinetics", 3rd ed., McGraw-Hill Book Company, New York, NY, 1981. 73. Sohn, H-J and Wadsworth, M. E., "Reduction of Chalcopyrite with S02 in the Presence of Cupric Ions", JOM, Vol. 32 , No. 11, pp. 18-22, November 1980. 74. Spender, W. D. and Topley, B., "Chemical Kinetics of the System Ag2C03(s) «-> Ag20(s) + C02(g)", J. Chem. Soc, Vol. 50, pp. 2633-2650, 1929. 75. Stanczyk, M. H. and Rampacek, C, "Oxidation Leaching of Copper Sulfides in Acidic Pulps at Elevated Temperatures and Pressures", USBM RI 6193, US Department of the Interior, Washington, D. C, 1963. 76. Subramanian, K. N. and Jennings, P. H., "Review of the Hydrometallurgy of Chalcopyrite Concentrates", Can. Met. Q., Vol. 11, No. 2, pp. 387-400, April-June 1972. 77. Swinkels, G. M. and Berezowsky, R. M., "The Sherritt - Cominco Copper Process - Part I : The Process", CIM Bulletin, Vol. 71, No. 790, pp. 105-121, February 1979. 78. Tromans, D. (Prof.) : Private communication throughout the course of this research, Metals and Materials Engineering Department, The University of British Columbia, Vancouver, BC, Canada, 1999. 79. Valensi, G., "Kinetics of Oxidation of Chemical Wires", Compt. Rend., Vol. 201, pp. 602- 604,1935. 80. Venkatachalam, S., "Treatment of Chalcopyrite Concentrates by Hydrometallurgical Techniques", Minerals Engineering (UK), Vol. 4, No. 7-11, pp. 1115-1126, 1991. 169 81. Vizsolyi, A.; Veltman, H.; Warren, I. H. and Mackiw, V. N., "Copper and Elemental Sulfur from Chalcopyrite by Pressure Leaching", JOM, Vol. 19, No. 11, pp. 52-59, November 1967. 82. Wang, M.; Zhang, Y. and Muhammed, M., "Critical Evaluation of Thermodynamics of Complex Formation of Metal Ions in Aqueous Solutions- III. The System Cu (I,II)-C1" at 298.15 K", Hydrometallurgy, Vol. 45, pp. 51-72,1997. 83. Warren, I. H.; Vizsolyi, A. and Forward, F. A., "The Pretreatment and Leaching of Chalcopyrite", CIM Bulletin, Vol. 61, No. 673, pp. 637-640, May 1968. 84. Warren, I. H., "A Study of the Acid Pressure Leaching of Chalcopyrite, Chalcocite and Covellite", Australian Journal of Applied Science, Vol. 9, No. 1, pp. 36-51, January 1958. 85. Weast, R. C. (Ed.) : CRC Handbook of Chemistry and Physics, 56th ed., Sections B and D, CRC Press, OH, 1976. 86. Weil, E. and Sandlers, S., "Sulfur Compounds - Hydrogen Sulfide", pp. 275-284, in : Kirk Othmer Encyclopedia of Chemistry and Technology, 4th ed., Vol. 23, John Wiley and Sons, Inc., New York, NY, 1992. 87. Wen, C. Y., "Non-Catalytic Solid Fluid Reactions", Industrial and Engineering Chemistry Research, Vol. 60, No. 9, pp. 34-54, September 1968. 88. Yannopoulos, J. C. and Agarwal, J. C. (Eds.): Extractive Metallurgy of Copper, Proceedings of an International Symposium, sponsored by the Metallurgical Society of AIME, held in Las Vegas, NV, February 22-26,1976. 89. Young, R., "Chemical Analysis in Extractive Metallurgy", 1st ed., pp. 172-188 for iron, pp. 133-154 for copper and pp. 315-323 for sulfur, Charles Griffin and Company Ltd., London, UK, 1971. 90. Yu, P. H.; Hansen, C. K. and Wadsworth, M. E., "A Kinetic Study of the Leaching of Chalcopyrite at Elevated Temperatures", pp. 375-402, in : Hydrometallurgy '72, Eds. : D. J. I. Evans and R. S. Shoemaker, Proceedings of the Second International Symposium on Hydrometallurgy, organized by the Metallurgical Society and SME of the AIME, Chicago, Illinois, February 25-March 1,1973. 170 Appendices Appendix I : Leaching Kinetics Models Appendix II : Chemical Analysis APPENDIX I LEACHING KINETICS MODELS 1.1 KINETICS OF SOLID - FLUID REACTIONS The basic principles of chemical reaction kinetics and those for leaching models can be found in any textbook on physical chemistry or kinetics (see, for instance, Atkins (1994), Smith (1981) and Levenspiel (1972)). It is not the intention here to give a comprehensive description of leaching kinetics, rather, to review these kinetics with respect to the main leaching reactions studied in this research. The mathematical modeling of leaching kinetics has become an important tool in describing hydrometallurgical processes, and for the purpose of research and development. Leaching reactions are normally described as non-catalytic solid-fluid reactions, where one or more reaction steps are encountered. The early work on leaching mechanisms was applied for describing very simple laboratory reactions (Valensi (1935) and Spender and Topley (1929)). These models were developed assuming simple or elementary reactions, that is one step mechanism. With the progress and advances in leaching processes, such models were modified to account for various factors affecting leaching kinetics. The two main reactions studied in this research are : 2CuFeS2(s) + 3H2S04(aq) + Fe(s) -> Cu2S(s) + 3FeS04(aq) + 3H2S(g) (1.1) 2CuFeS2(s) + 6HC1 (aq) + Fe(s) -> Cu2S(s) + 3FeCl2(aq) + 3H2S(g) (1.2) More generally, these two equations can be described by a single reaction as : 2+ 2CuFeS2(s) + 6PT (aq) + Fe(s) Cu2S(s) + 3Fe (aq) + 3H2S(g) (1.3) The overall leaching reaction written in Eq. 1.3 includes both cathodic and anodic half reactions. Either or both of these reactions could be rate limiting. Also, each of these reactions might occur by a sequence or series of steps, any one of which in turn may be rate controlling. The only way to analyze the individual steps in every half reaction is to perform an electrochemical study on synthetic electrodes of chalcopyrite, under similar experimental conditions, and analyze the resulting electrochemical data, which was not done in this research. 172 Rather, the recorded kinetic data were fitted to the well-known leaching models to explain or assess the experimental kinetic behavior of the main reactions. As discussed in Section 6, both the anodic and cathodic reactions are rate controlling. For any solid-fluid reaction, the following mechanistic steps are considered : 1) Chemical reaction in solution 2) Boundary layer diffusion 3) Product layer diffusion 4) Chemical reaction at the solid-fluid interface with or without charge transfer 5) Diffusion of products away from the interface. 6) A combination of two or more of these steps Only step (4) represents an electrochemical process. For this step, Hiskey and Wadsworth (1975) developed a model for conversion occurring by an electrochemical phenomenon, where charge transfer processes are rate controlling. Sample geometry also manifests itself in kinetic effects, because of variation in surface area during the reaction. Flat plates and disks react with a minimum variation in area; whereas isometric shapes, such as cubes, cylinders and spheres, react with considerable change in area. According to Levenspiel (1972), there are different leaching models, for the remaining mechanistic steps, which would account for the rate determining step for a reaction of the type: A + bB(s) -» products (1.4) where A is the fluid phase and B is the solid phase. For monosized spherical particles, these models are : 1) Linear leaching model: When the rate determining step (RDS) is a chemical reaction occurring on the particle surface, the surface reaction model is used to describe experimental data. The general model (Spender and Topley (1929)) is : (1.5) 173 a is the fraction of CuFeS2 reacted at time t, and r0 is the initial particle radius, k, is the linear rate constant, that includes the specific rate constant for the surface reaction and a surface roughness factor, represented by ksc, volume and cross sectional area of molecular reactants, represented by the molar volume V, stoichiometric factor § (in the case of two solid reactants, like Fe and CuFeS2 or Cu and CuFeS2, <(> can be the molar ratio of iron or copper to chalcopyrite), and the concentration term(s),Cs, defined as : k.-^ d.6) The same model can be used to interpret the results in terms of chalcopyrite or the reducing agent (if two or more solids are encountered). If the reaction constant is written as f a ^ k= —, then from this model a plot of I 1 - (1 - —)3 I vs. time should yield a straight line having r0 V 4> ; the slope k. The estimated reaction constant at different temperatures can be used to prepare an 1 Arrhenius plot (a plot of reaction rate constants vs. - ) to evaluate the apparent activation energy, Ea. The new version of this model is written as : - = 1-(1-XB)' (1.7) T Where: x = -^- (1.8) bksCAf pB is the particle molar density, R is the initial particle radius, b is the stoichiometric factor, k^. is the surface chemical reaction rate constant and CAf is the bulk fluid concentration. XB is the conversion with respect to solids, while x represents the required time for complete conversion. This model is applicable to particles of changing and unchanging size. 2) Parabolic leaching model: When reaction products form on the reacting solid, the kinetics may be governed by the nature of the product layer. Generally, product layers will represent a resistance to reagent reaching the reacting interface. The rate controlling step for such a topochemical process may 174 involve the diffusion of one or more reactants through this layer. In some cases, the diffusion of reaction products through such a layer may be rate limiting. Valensi (1935) developed an equation relating diffusion paths to sample geometry for reactions proceeding topochemically in spherical shapes : 1 2 _(1_a)!)=iEit (L9) v 3 J r0 Again, a is the fraction reacted at time t, and r0 is the initial particle radius. The constant kd contains the effective coefficient of diffusion; which is a combination of diffusion coefficient, porosity and tortuosity, molar volume of reactant, stoichiometric factor, and concentration terms, and is defined as : kd = 3V(j)Def!ACA . ( 2 A Clearly a plot of |^1 - —a - (1 - a)3J vs. time should yield a straight line having the slope k • k = —f . In the same manner, the evaluated reaction constants at different temperatures can be used to prepare an Arrhenius plot (a plot of reaction rate constants vs. to estimate the apparent activation energy. The new version of this model is written as : - = 1-3(1 -XB)' + 2(1- XB) (1-10) T Where : P R2 B (LU) 6bDeCAf pB is the particle molar density, R is the initial particle radius, b is the stoichiometric factor, De is the effective diffusivity of the fluid in the product layer (in this case it is a function of molecular diffusivity, porosity, tortuosity, shape factor, and roughness) and CAf is the bulk fluid concentration, T and XB are the required time for complete conversion and the conversion of chalcopyrite, respectively, as above. This model is only applicable to particles of unchanging size. It is apparent from Eq. 1.11 that small particles require much shorter time for complete conversion than do large ones. Hence, diffusion controlled kinetics are generally improved when fine particles are used. 175 For these two models, surface reaction control and product layer diffusion control (or ash control model), one can distinguish reaction kinetics on the basis of the dependence of reaction rates on particle size. A plot of the initial rates vs. -7- (Eq. 1.8) gives a straight line for linear d° kinetics and a plot of the initial rates vs. —7- (Eq. 1.11) gives a straight line for parabolic do kinetics, where d0 is the initial particle diameter. In this research, all the kinetic analysis by the selected leaching model is done with reference to chalcopyrite particles, and all generated rate expressions or leaching graphs are with respect to these solid particles, unless otherwise specified. So, XB is related to chalcopyrite conversion (reduction). 3) Leaching model for diffusion through a boundary fluid film control: A less common case in leaching systems is when diffusion through a boundary fluid layer becomes the RDS. In this case the appropriate model is : " = XB (1.12) T here pBR z -^:t (L13) with kf being the mass transport coefficient between the fluid and solids (for the remaining symbols, the same definitions hold). This model is only applicable to particles of unchanging size. For particles with changing size, Eq. 1.12 is modified to read : t 2/3 = 1-(1-XB) (1.14) T where T, as usual, is the time required for complete conversion, and is defined as : PBR2 t-2bC^D (U5) with the same definitions for these symbols. This model is only applicable to small particles of changing size in the Stokes regime. 176 4) Mixed kinetics : Generalized rate equations can be developed to explain special cases of mixed kinetics, when one or more steps are rate determining. When both the interfacial area mechanism (reaction control) and the diffusion mechanism are contributing to the controlling mechanism of the reaction, the mixed kinetics model will be : l-ja-(l-a)']+p(l-(l-a)') = y Ct (1.16) Where a is the fraction reacted at time t, and C is the fluid concentration. The constants : 2kd 2kd P = —A- and (1.17) r0ks r0 p are determined empirically by an evaluation of selected data. The symbols are defined as above. ks is the surface reaction rate constant, and p is the particle molar density. It should be noted that the generalized expression shown above simplifies to linear kinetic model when ks « kd or k k •« 0. Furthermore, when ks » kd, or -f- « 0, the resulting expression is identical to that kd ks given for parabolic kinetic model. This model was used by Peterson and Wadsworth (1994) as discussed in Section 2.4. In the case where three controlling steps are involved, namely diffusion through boundary layer, surface chemical reaction and diffusion through product layer, the generalized model is: 0-8 2D;[i-!«-o-a)-]+k;(.- (1.18) Where : Stoichiometric factor 5 : Fluid boundary diffusion layer : Coefficient of diffusion D* a Conversion D' Effective diffusivity Initial particle radius r0 Surface reaction rate constant k0' 177 C : Bulk fluid concentration V : Molar volume of solid particle and equals MWB/pB' where MWB is the molecular weight of the solids B (chalcopyrite in this case) and pB' is their mass density t : Time For all the presented models, any convenient system of units can be used. More derivations for spherical and other geometrical shapes can be found in Levenspiel (1972). 178 1.2 THERMODYNAMIC DATA FROM THE KINETIC ANALYSIS After deciding on a suitable kinetic model, the rate data can be recorded at various temperatures. The activation energy, Ea, is estimated from the slope of the curve of a plot of initial rate data (reaction constants) vs. inverse of temperature. This plot (Arrhenius plot) will normally result in a straight line from which the activation energy and other information can be obtained. According to the Arrhenius equation, the reaction rate constant is defined as : k=k0exp(^) (1.19) where : k : Reaction rate constant, min"1 1 k0 : Frequency or pre-exponential factor, min" . Ea : Activation energy, J per mol R : Ideal gas constant, 8.314 J per mol per K T : Recorded temperature, K k0 in many physical chemistry textbooks is called A. Here, the selected units are for demonstration. In logarithmic form, Eq. 1.19 is written as : lnk = ^- + lnk0 (1.20) 1 -E and so, a plot of In k vs. — should give a straight line, of slope equals - and an intercept T R equals In k0. From the slope, the apparent activation energy, Ea, can be estimated, while the intercept (at T"1 = 0) will give the pre-exponential (frequency) factor. Unless otherwise noted, all logarithmic functions in this research are Naperian or Natural logarithms. Generally, it can be said that for chemical reaction control systems, activation energy is greater than 40 kJ per mol, while for diffusion control it is less than this value. Some authors 179 prefer to give an activation energy value between 20 to 40 kJ per mol for mixed kinetic systems, but their claim has little supporting evidence. The estimated activation energy can give different indications on the behavior of the system. If this energy is relatively small, then it will explain the rapidity of such a reaction, because, from the collision theory, if the activation energy is large, only a small fraction of molecules will have enough energy to surmount the reaction barrier and react rather than rebound upon collision. The other important indication is the temperature sensitivity. A reaction with low activation energy is not very sensitive to temperature variation. High activation energy means that small variations in temperature would result in considerable changes in reaction rates, which is suitable for chemical reaction controlled kinetics. Hence, it can be assumed that product layer diffusion controlled kinetics are less temperature dependent. The estimation of the activation energy would allow the estimation of the enthalpy of activation. According to the absolute rate theory, the enthalpy of activation of any reaction, is estimated from the following equation : (AS /R) (AH 0/RT) k = K0^e ' e- = (1.21) which results from replacing the activation energy, Ea, with the definition of free energy of activation, AG°,: AG,1 = AH,"-TAS: (1-22) Where: k : Reaction rate constant, min"1 K0 : Transmission coefficient, 0.5 < K0< 1.0 K : Boltzmann's constant, 1.38 X IO"23 J/K h : Planck's constant, 6.625 X 10"34 J s T : Temperature, K R : Ideal gas constant, 8.314 J/mol per K AS^ : Entropy of activation, J/mol per K AH^ : Enthalpy of activation, J/mol 180 AG° : Free energy of activation, J/mol In logarithmic form, Eq. 1.21 is written as : (AS /R ln(-) = ^^ + ln{K0-e - 4 (1.23) T RT 1 h j k 1 and so, the straight line obtained from plotting in (—) vs. — may be used to estimate both the enthalpy and entropy of activation (from the slope and the intercept, respectively). The value of the enthalpy of activation should be near to the energy of activation, and some discrepancy is normal in hydrometallurgical reactions. The entropy of activation should have a net positive increase, that is AS° ^ 0, for a reaction to take place. According to Wen (1968), in determining the rate controlling factor, experiments may be performed with different sized particles, and the time required to achieve a given conversion is measured. If R, and R2 are the radii of the two particles which have the same conversion, but have the corresponding reaction times of t, and t2, respectively, then, for qualitative assessment, it can be said that: for film diffusion controlling, 1.5-2.0 ft} (1.24) VtJ for ash diffusion controlling, , 2.0 V Rl I and for chemical reaction controlling, ro (1.26) vt2J The exponent in Eq. 1.24 is dependent on solution turbulence (or Reynolds number). Wen also stated that if the rate is very sensitive to temperature variation, surface chemical reaction may be considered the rate controlling factor. 181 In addition, utilizing the definition of effective diffusivity, De, through porous ash layer, based on the surface area of unreacted core, the equation for chalcopyrite leaching under ash diffusion control becomes : 1/3 2 kt = (l-(l-XB) ) (1.27) Where : 2bDeMWcpy[H2S04] k= ^^f-2—- (1.28) Pcpy R pcpy is the mass density of chalcopyrite (cpy), MWcpy is its molecular weight and R is the initial radius of chalcopyrite particle. The fluid concentration, [H2S04], can be replaced by [HC1] in chloride media. The significance of this form of controlling equation is that a plot of In t vs. 1/3 In (1-(1-XB) ), would give a straight line of slope equals 2, which can be used to assess the closest approximation to the controlling mechanism. Moreover, the value of the rate constant, k, estimated from the value of the line intercept, can be compared to that estimated from the predicted leaching kinetics model for further verification. If the reaction is chemically controlled, the slope would be 1 : 1/3 kt = (l-(l-XB) ) (1.29) Where : b ks MWCPY CPY k = * R (1.30) PCPYK Again, R is the initial radius of chalcopyrite particle and ks represents the surface chemical reaction rate constant. 182 APPENDIX II CHEMICAL ANALYSIS A detailed chemical analysis of both the leach solution and the reaction residue was performed to demonstrate the reaction stoichiometry. This analysis was based on wet chemistry methods and detailed material balance calculations were done to characterize the reaction products. Residue analysis was made by digestion with aqua regia and bromine water followed by analysis of digestion solution by AAS. Leach and filtrate samples were also analyzed by AAS. All weighing was done using an analytical balance. In Tables II. 1 and II.2, several calculations are presented. These calculations were obtained as explained below. II. 1 REACTANTS The amounts of reactants are shown in the first page of these tables. They contain the added concentrate, the leach residue weight, filtrate volume, digestion sample weight and digestion solution volume. The mineral and metal contents were estimated as per Tables 5.1 and 5.2. The solid pulp density is estimated from the relation : CuFeS, mass + Fe mass SPD % = —— 2- —— xlOO (II.l) CuFeS2 mass + Fe mass + Solution mass Initial leach solution has a sp. gr. of 1.0, while that with ferrous chloride additions has a sp. gr. of 1.3. The weight difference corresponds to the released amounts from the leached concentrate, and is estimated from the relation : Weight difference = Chalcopyrite concentrate weight - Leach residue weight (IL2) The residue weight after digestion (corrected) corresponds to the siliceous gangue content in the concentrate, based on residue weight after sample digestion, and is estimated from the relation : Residue weight after digestion (corrected) = Residue weight Siliceous gangue in the digestion sample x —; ; (II.3) Digestion sample weight and the estimated siliceous gangue content in the leached concentrate is estimated from the relation: 183 ., Residue weight after digestion (corrected) Siliceous gangue content in the leach residue = ———; x 100 Leach residue weight (II.4) This has to be comparable with that listed in Table 5.2, i.e.: ^ 9.9%. Stoichiometric iron needed for chalcopyrite is estimated according to Eq. 37, and is given as : Stoichiometric iron needed for chalcopyrite (only) = CuFeS2 content 1 momoll metallimetallicc FFe x MWFe (II.5) MWCuFeS2 2 mol CuFcS, and the excess iron used relative to the stoichiometric value is estimated from the relation : n/ Amount of metallic iron added Excess iron % = —-—— ;—; —— x 100 (II.6) Stoichiometric iron needed 184 II.2 MATERIAL BALANCE CALCULATIONS FOR IRON Page 2 of these tables shows the detailed iron analysis for the experiment. The required relation for every row is as follows : Iron content in the concentrate = Iron percentage from Table 5.1 (28.0%) x Chalcopyrite concentrate weight (IL7) Iron content can also be estimated for the chalcopyrite and pyrite since the concentrate content of these two minerals is known. For the chalcopyrite : ^ ^ r, CuFeS, weight 1 mol Fe Iron content as CuFeS2 = - — x x MWF (II.8) MWCuFeS2 ImolCuFeS, and for the pyrite : FeS, weight 1 mol Fe 1 Iron content as FeS2 = — x x MWF (II.9) Fe v MWFeS2 lmolFeS2 ' Atomic absorption readings were described in Section 5.2.3. The obtained AA readings are used to estimate chalcopyrite conversion as follows : 1 g Iron content in residue digestion = AA reading (ppm) x Digestion solution volume x 1000 mg (11.10) _ ...... Leach residue weight Iron content in the leach residue = Iron content in residue digestionx Sample weight (11.11) and iron released in solution from the concentrate will be : Iron released = Iron content as CuFeS2 + Iron content as FeS2 - Iron content in residue (11.12) This relation assumes that only the iron component of chalcopyrite is dissolving, and is justified theoretically as explained in Section 5.3 and from the discussion given below. From this value, the conversion of chalcopyrite is estimated as follows : 185 . , . , . Iron released in solution from the concentrate % iron dissolution (conversion) = xlOO Original iron content as CuFeS2 (11.13) and the resulting number from this equation should be near to that estimated from the leach data. The remaining iron in chalcopyrite is estimated as : Remaining iron in CuFeS2 = Original iron content as CuFeS2 - Iron released in solution (11.14) This amount of iron can also be used to estimate the corresponding amount of copper for this chalcopyrite Remaining iron in CuFeS2 1 mol Cu Corresponding Cu content = x -—: x MWCu (11.15) IVI^iVpg 1 mol Fe By these, iron amounts are tracked, and the percentage of iron in leach residue can be estimated from : n/ Iron content in residue Iron % in the leach residue = — x 100 (II. 16) Residue weight „..,,, ., . Iron content as FeS, Iron % in the leach residue as pyrite = — x 100 (II. 17) Iron content in residue . , , , ., • ' „ Remaining iron in CuFeS, Iron % in the leach residue as CuFeS2 = - x 100 (II. 18) Iron content in residue 186 II.3 MATERIAL BALANCE CALCULATIONS FOR COPPER The same procedure can be extended to copper, and Page 3 of these tables summarizes the required calculations. The relations are : Copper content in the concentrate = Copper percentage from Table 5.1 (28.3%) x Chalcopyrite concentrate weight (H-19) ... CuFeS, content 1 mol Cu Copper content as CuFeS2 (initial) = ——f- x , ^ ^ x MWCu (11.20) MW mo1 cuFes2 1 CuFeS2 The concentrate contains some chalcanthite, which will release copper by solvation : CuS04.5FLO content 1 mol Cu Solvated copper = * x —— — x MWCu (11.21) Cu K / MWCuS0)5Hi0 1 mol chalcanthite This solvated copper will also be called "Less cemented copper from chalcanthite". Atomic absorption readings are used to find the copper concentration in different solutions. For digestion solution : 1 g Cu in residue digestion = AA reading (ppm) x digestion solution volume x (11.22) 6 1000 mg v } „ ^ Leach residue weight Cu content in the leach residue = Cu content in residue digestion x Sample weight (11.23) Cu released from the concentrate in solution = Cu content in the concentrate - Cu content in the leach residue (11.24) and the copper content in leach residue as chalcocite will be the net difference between the total copper in leach residue and that remaining as chalcopyrite and/or cemented copper, Cu content in the leach residue as Cu2S = Cu content in the leach residue - Cu as chalcopyrite - solvated copper (11.25) and these equations state clearly that the copper component of chalcopyrite remains intact. The corresponding amount of chalcocite is : 187 Cu content in the leach residue as Cu,S 1 mol Cu,S Cu2S amount = ,„,r — x — —x MWCu s (11.26) 2 CU2S MWCu 2 mol Cu and copper percentage in the leach residue is : Cu content in residue Cu % in the leach residue = ——— — x 100 (11.27) Residue weight As explained in Section 6.1, acid consumption can be used to follow the reaction progress, and for every set of material balance calculations, the acid consumption is recorded and plotted separately. A sample combined plot of acid consumption and estimated conversion (chalcopyrite reduction, see Section II.5) is given in Fig. 6.2. In Tables II. 1 and II.2, the experiments were performed in chloride media. The amount of consumed acid (HQ) can be used to estimate the theoretical amount of chalcocite to be formed, based on reaction stoichiometry. The molarity of HC1 in low SPD experiments is 1.0, while for high SPD experiments it is 10.0. For chalcocite, and based on acid consumption : • . • , „ • Acid consumed (ml) xMm 1 mol Cu,S 2 Theoretical Cu2S = = — ^ x — x MWr„ „ (11.28) 2 1000 ml per 1L 6 mol HC1 CU2S ^ J and the yield for Cu2S is defined as : . ,, „ • „ Corresponding Cu,S for copper in leach residue v 2 vv Yield for Cu2S = ^ , ° — . : x 100 (11.29) Theoretical C^S based on acid consumption Hence, the amount of iron used for chalcopyrite decomposition based on actual chalcocite formed is : Fe for chalcopyrite decomposition = Corresponding Cu,S for copper in leach residue 1 mol Fe ~ ~T^TT X i 1 ^ o X MWFe (11.30) 10 MW(:„jS 1 mol Cu,S and the yield for iron will be : . ,, Fe for chalcopyrite decomposition based on Cu9S Iron yield = —^773-7-—77^ — * 100 (11.31) Added metallic iron 188 II.4 COMPARISON OF RELEASED IRON (CHALCOPYRITE CONVERSION) BASED ON COPPER READINGS The last page of the tables for material balance calculations contains additional information on conversion and released amounts of products, as well as an approximation to the residue composition. For instance, the amount of iron as chalcopyrite can be found by knowing the values of iron as pyrite and that in residue, Amount of Fe as CuFeS2 = Fe content in residue - Fe content as FeS2 (11.32) This value should compare well with the value found by estimating the remaining iron in chalcopyrite, from Section II.2 (Eq. 11.14). Next, both the estimated conversion from Section II.3 (Eq. 11.13) and that from the analyzed kinetic data (for instance, the data in Table 6.6) should be comparable with that estimated based on iron used for decomposition (Eq. 11.30), which in turn is estimated from actual chalcocite formed, Eq. 11.26, or copper material balance calculations. In other words, the estimated conversion based on iron yields, Eq. 11.31, should be comparable to that from chemical analysis, Eq. 11.13, and the leaching data (Table 6.6 or Table II.2): Corresponding CuFeS2 conversion = Fe for CuFeS2 decomposition based on formed Cu2S 2 mol CuFeS, x — x MWFe 1 mol metallic Fe Original CuFeS2 content (11.33) Copper readings in the filtrate should also be comparable with the calculations based on residue digestion : 1 g Corresponding Cu in the filtrate = AA reading (ppm) x Filtrate volume x (11.34) 1000 mg This can be used to estimate the fraction of copper in leach solution : Corresponding copper in the filtrate % copper in leach solution -—— x 100 (11.35) Original copper content in the concentrate 189 The amount of copper in filtrate should compare well with the estimated amount of copper released in solution, from Section II.3 (Eq. 11.24). From residue digestion : Copper released in solution = copper released from the concentrate - solvated copper (11.36) Moreover, the released sulfur from the concentrate as H2S can be estimated based on chalcopyrite conversion, according to the reaction stoichiometry. For sulfur : Released sulfur Iron released in solution from the concentrate 1 mol CuFeS, 3 mol H,S 1 mol S x x -— x x MW, s MWFe ImolFe 2 mol CuFeS2 1 mol H2S (11.37) and this amount of sulfur, together with all released species from the concentrate (from residue digestion calculations) should compare well with the weight difference estimated in Section II. 1 (Eq. II.2): Estimated total release from the concentrate = Iron released + Sulfur released + Copper dissolved (11.38) The last issue to be discussed is the theoretical composition of the leach residue. If solvated copper is assumed to be precipitated (which is correct when running at high SPD), then the only changing component is chalcopyrite, while a new solid phase is added, which is chalcocite. Remaining chalcopyrite can be estimated from the remaining iron (Eq. 11.14), since : Remaining iron in the chalcopyrite 1 mol CuFeS7 Remaining CuFeS2 = ^— x x MWCuFtS; (11.39) Fe The pyrite and siliceous gangue will remain intact, and the actual formed chalcocite can be estimated as shown in Section II.3 (Eq. 11.26). So, an approximate composition for the leach residue can be obtained, and in the ideal situation, the sum of the estimated amounts of remaining chalcopyrite, formed chalcocite, pyrite, cemented copper and siliceous gangue should equal the weight of the leach residue, stated in Section II. 1. Finally, by these relations, the material balance calculations are straightforward and will allow the demonstration of reaction stoichiometry and other related kinetic data. As explained in 190 Section 6, the results obtained by chemical analysis were generally in conformance with the kinetic analysis and SEM/EDX testing. A sample calculation is given in Tables II. 1 and II.2. In these tables, the calculations were done based on AA readings for the leach, filtrate and the digestion solution samples, as explained above. These calculations were capable of demonstrating the reaction stoichiometry. Table II. 1 represents the analysis for an experiment done at low SPD. The analysis was made for both the solid residue and the filtrate. The results are in conformance with the data obtained during the experiment which are summarized in Table 6.6. Thus, the assumption that dissolving iron from the concentrate is only from chalcopyrite is valid, and the assumed chemical composition of the leach residue is acceptable. Also, the copper released from the concentrate by solvation was not completely precipitated, because the reaction ceased within short period of time, and the advantage of the presence of H2S gas as a precipitating agent is lost. The criteria stated earlier can be seen valid for these calculations. The estimated conversion from iron analysis compares well with that from chalcocite. The theoretical yield for iron is close to the conversion (of course, the balance was consumed in side reactions). The estimation of copper released in solution compares well with the filtrate analysis, and the total release from the concentrate also compares well with the weight difference between fresh and leached concentrate. Another finding is the hypothesized composition of the leach residue, which agrees well with the residue weight after three hours of reaction time. Table II.2 summarizes the calculations for an experiment done at high SPD. It was difficult to take samples from the leach solution for two reasons. First, the reaction vessel was sealed and continuously purged with a stream of nitrogen. Second, the AAS for iron at such high concentrations will be somewhat incorrect, due to the incorporated errors in dilution. The addition of 3 M ferrous chloride is seen to be of great advantage, as reaction kinetics are not affected (Section 6.4). Once again, the continuous release of hydrogen sulfide is of great benefit for this method of reductive leaching. Any copper released in solution, whether by solvation, or less probably by oxidative leaching, will be precipitated in the residue. Next, the presence of some extra iron will enhance this effect by cementation. The criteria set above are now well suited, as revealed from 191 the detailed calculations. The analysis for the leach residue was again proven to be valid, and this method will result in the enrichment of chalcopyrite. At high solid pulp density, the probability of precipitating ferrous ions as troilite or pyrite is less likely to take place due to the preferential precipitation of copper sulfides and the solution pH (as explained in Section 4). Iron analysis of the residue showed that iron content in the concentrate is decreasing, eliminating any doubt about iron precipitation. In most residue analyses done, this methodology remained valid. For dissolved copper, the dissolved amount depends on the experiment. For low SPD experiments, some copper was released in solution, by solvation, rather than from the chalcopyrite. For high SPD experiments, no dissolved copper could be detected by AAS, as explained in Sections 4, 5.2 and 6.4. The detailed chemical analysis show that, especially under high SPD, copper is retained in the newly formed solid phase, while iron is entering the leach solution and sulfur is released as H2S. It should be noted that better results were obtained using finer size fractions (Table 6.27) and it is expected that the conversion will further increase by using much finer particles (less than 400 mesh). Also, increasing the temperature or the added amount of metallic iron is expected to increase the utilization of this method of reductive leaching, as is discussed in Section 6.4. The results from wet chemistry analysis showed that the stoichiometry of the leaching reactions in Eqs. 32 and 33 is correct, and the proposed leaching mechanism for this system is valid. Further, SEM/EDX analysis confirmed that chalcocite is the main solid product of leaching reactions. So, reductive decomposition of chalcopyrite with metallic iron is possible and will lead to the formation of chalcocite as the main copper sulfide in the leach residue. 192 II.5 ESTIMATION OF THE CONVERSION FROM AAS : There are two main reactions in the leaching system : Chalcopyrite reduction, + 2+ 2CuFeS2(s) + Fe(s) + 6H (aq) -> Cu2S(s; + 3Fe (aq) + 3H2S(g) (11.40) and hydrogen evolution, + 2+ Fe(s) + 2H (aq) -> Fe (aq) + H2(g) (11.41) 2+ At the end of an experiment the total dissolved iron in solution is [Fe ]total. Since all the added metallic iron dissolves completely as per Eqs. 11.40 and 11.41, the maximum amount of released 2+ iron from chalcopyrite lattice is [Fe ]released and : = 2 [Fe Leieased [Fe ]totai - [Fe ]metal (11.42) 2+ where [Fe ]metal refers to the concentration of iron due to complete metal dissolution. The final conversion is estimated as : rFe2+l released % Conversion = 2+ x 100 (11.43) e [^ JcuFeS2 2+ where [Fe ]CuFeS2 is the maximum concentration of iron for complete chalcopyrite conversion. Hence, the final extent of conversion of chalcopyrite is easily measured by analysis of dissolved iron in solution. As explained in Section 5.3, other possible iron sources in the concentrate are relatively minor and can be safely neglected. To estimate the conversion at a certain time during an experiment, the analyzed iron from the leach sample can also be used. At a certain time t, the extent of conversion can be written in a similar fashion to Eq. 11.43 : IFe2+l' V = i__ Released Vtt 44n B rFe2+l ( } rc f JCuFeS2 2+ t where XB is the fractional conversion of chalcopyrite and [Fe ] released is the solution iron 2+ t concentration arising only from chalcopyrite reduction at that time. To estimate [Fe ] released, new parameters need to be introduced to account for the effect of Eq. 11.41 on iron dissolution. 193 The parameter r|Fe, which is the overall efficiency of iron as a reductant toward chalcopyrite alone, is defined as : rFe2+l L IL45 The= rF 2::" ( ) I**6 Jmetal This parameter represents the fraction of added iron that was actually used in chalcopyrite reduction (Eq. 11.40). The term r|Fe also represents the final conversion of chalcopyrite in the system, which is derived from chemical analysis and the reaction stoichiometry. The term [Fe2]reduction represents the corresponding concentration from metallic iron consumed in chalcopyrite reduction, and can be estimated by utilizing the reaction stoichiometry. From Eq. 11.40 and 11.42 : 2+ 2+ 2+ 2+ [Fe ]reduction= ^x[Fe ]released= ^{[Fe ]total-[Fe ]metal} (11.46) The balance will be the amount of metallic iron used in hydrogen production. Similarly, the efficiency of hydrogen production, r\Hi, will be : MH2=l-%e (H.47) 2+ The analyzed total concentration of dissolved iron at time t, or [Fe ]|otal, can be used to estimate the fractional completion of the reactions in the system. In the following procedure, it is assumed that reactions 11.40 and 11.41, more or less, occur in direct proportion to each other. To a first approximation, it may be justified on the grounds that iron corrosion is the driving force behind these two reactions. While not strictly correct, this does allow estimating the extent of chalcopyrite conversion (reduction) with time. Defining the parameter © : lFe2+V 0=rr^rL (IL48> lre Itotal This parameter represents the degree of progress of the reactions (Eqs. 11.40 and 11.41) and hence can be used to estimate the amount of metallic iron reacted at time t, or [Fe2+ ]' metal 2+ 2+ [Fe ]me,a> = © x [Fe ]metal (11.49) 194 Then the amount of metallic iron reacted with chalcopyrite at time t (i.e. used in chalcopyrite 2+ t reduction) is [Fe ] rednBti0I1 and : 2+ , 2+ t [Fe ] reduction==[Fe ] me«a.><^Fe (H.50) From Eq. 11.40, the amount of iron released from chalcopyrite is : 2+ 2+ [Fe ]:eleased=2x[Fe ]:eduction (11.51) while the amount of dissolved iron due to reaction 11.41 is given as : 2 [Fe %=[Fe2+L,a> * ^ (H.52) 2+ where [Fe ]t,2 refers to the concentration of solution iron arising due to hydrogen evolution. Now Eq. 11.51 can be used to estimate the conversion of chalcopyrite at any time t as per Eq. 11.44. To simplify the calculations, by utilizing Eqs. II.45-II.52, Eq. 11.44 can be rewritten as : 2+ v 2 x 0 x [Fe ]metal x nFe *B~ rp 2+1 (11.53) 1 c t JCuFeS2 The calculation was used to provide an estimate of the extent of chalcopyrite conversion with time. 195 VO On u s o bo bO bO bO bO bO bo bO bO bO 6 m ^t U o pq ON O m in ON O g ON < ON > + ^ONt fi o CCJ ^)H cd N co a CD H-> CO > CO u Q CH CO 2 CO E-i o pq co § fi u HH o 2 H W O JD co "3 CD Q o CD HH HH O CD U K Q HH < O pq H O O Q co 2 E- Q pq < O '(H o CO O Oh HH CD Q •s K CO u O X O pq a H HH £ o Q "o, u H IBS Q CD i >H PH CO PH PH Q o O pq o pq S CD O HH Q Q H pq U U o a CO CO r-- Os •5 s NO I 60 bO 00 60 00 00 bO bO K oo s O OH 0 U o PQ Tt cn rn t m Tt VO oo r-- o ON Tt o> ON Tt m Tt ON Tt CN Tt Tt CN CN o" r- ON m ON m © U > un X Tt + s 3. ON %—» c* N C/3 U PQ Cu) C/3 H H—» 2 C/3 O PQ 00 HH OH Q C/3 Q O OH HH U C/3 C/3 HJ o O HH < § g Q X o HH u GO Oi PQ u HH < 2. C/3 HH oo HH' Q oi C oi HH o o oi PQ Pi o PQ o PH PQ PH > '3 H PH C/3 x >H PQ g H o O PQ HQ H HH H H HH 2 g H U Cu) 2 PQ g H < g O PH o HQ >< g HH o O U OH u o o O u o 13 g HH HQ HQ u H g »! E-i oi <: u HQ § PQ 13 2 H U HH HQ •c u w PQ u TERM ) UNIT cn o CN T-H COPPER CONTENT IN THE CONCENTRATE 60 © cn T-H in COPPER CONTENT AS CHALCOPYRITE (INITIAL) bO © CN m o bO CHALCANTHITE (CuS04.5H20) CONTENT IN THE CONCENTRATE o vo o cn © |BY SOLVATION, CORRESPONDING DISSOLVED COPPER bO cn O un AA READING FOR COPPER CONTENT IN RESIDUE DIGESTION ppm CN © m COPPER CONTENT IN RESIDUE DIGESTION bO © VO Tt m Tt COPPER CONTENT IN THE LEACH RESIDUE o o r- in ON |C0PPER RELEASED FROM THE CONCENTRATE IN SOLUTION bD © o o VO cn LESS CEMENTED COPPER FROM CHALCANTHITE bfi © CN Tt T COPPER CONTENT IN THE LEACH RESIDUE AS CHALCOCITE bO o cn o VO T-H CORRESPONDING CHALCOCITE FOR THIS COPPER bfl © o o O in |C0PPER PERCENTAGES THE LEACH RESIDUE ^-4- m ACID CONSUMED IN REACTION (AUTOTITRATION VIA 1.0 M HCl) XJ ml © m ON © THEORETICAL CHALCOCITE TO BE FORMED BASED ON ACID USED bO S 0 cn cn ON cn 0 sO YIELD FOR CHALCOCITE © T-H © Tt bD [AMOUNT OF IRON USED FOR CHALCOPYRITE DECOMPOSITION BASED ON Cu2S o CN Tt l> Tt METALLIC IRON ADDED bO S -t CN 0 Tt cn Tt T-H YIELD FOR IRON ON ON ON 00 vo l/-> CN O ^> Tf a. TT Tf 00 ON VO CN •S bfl 00 00 ^5 oo 60 bO bO bO bO © TT CN CN Tf rN ON K O tN VO vo b0| ON cn cn Tf O 00 U. o CN cn cn VO o cn o © 00 cn 00 cn o VO r- Tf Tf CO ON Tf cn o lO cd o CN Tf cn © © l S i Tf + s A ON Tf N • co a CL> co CO Q PH CO o co fl O W X o H S o K "c3 U o o < o H-l > u 2 H W o § 2 HH H E-i W HJ o Q W co U H < o HH 2 § § u H Q 2 o Pi HH 1 Z H o CO a| § § § W co CO pq 1 H O f— u U PH a a O HH u § § < Q o HH § w W Q o u co pq H pq a HD Q ID w g IS o Q HJ H HH § H I HH H1 H a O 2 Pi W i!C^O Q CO 2 § a PH I o W O >< co u : co Pi O HH pq CN HH O HH O u o HJ CO HH H O HJ h-1 w Q CO H? Q u Q HJ U 2 u Q co pq nq pq O < < < PH U O PH co H CO a CO X > E—1 a u W w bO bD PH bO bO bO bfi bO OH o VO oo VO o m o CN rn o CN 00 o 00 o 00 VO cn vo CN ION > TT CN CN Os g HH E-i oo W O O HH S3 g Q oo W F-H s Q E-i HH oi o CO O CZ) w a fe E-i w m E-i § E- E-i E-i HH 2 § H 2 § CH < OH E-i o E-i U O a W g u O o u E-i u E-i U < u u w <: 2 W w BC o u oi u H 0H g u 00 oo S3 w OH HH C/3 < < o £ OH H E-i o O fe QU CQ E-i § S3 § § o O E-i 3 o E-| 5 OH g 00 2 g g < w U g U o s cri o O § oi 3 o CN o O CN 00 00 OO 00 00 00 00 00 00 00 00 6s 00 00 o in CN o in o in cn vo o oo o o VO o CN vo X m 00 Tf Tf 00 cn in in Tf CN Tf Tf OS in m OS CN CN Tf cn vo CN CN cn cn cn © CN OS cn od vo O cn OS CD < PH > cn U X Q pq pq CO H U X Q HH O ed U pq O CH < Q o PH HH pq w co PL, o O > OH pq < o < X u CQ u H Q s Q ed pq HH pq O w > PH E—i O H CO pq co H § CO O 2 H HH § E—i HH PH E—i o O O H u H 5 u u pq X in pq U K u I H Q H co o HH pq HJ co H O Q < HH < fl U Q u U < o 1 Q u < w E— pq Z u H pq < O O HH X ed f—i ed H co u HH § ed y ed g U o o 5' PH CN 2 U ed U § PH ^q pq ed i~q Q CN u PH i fi _o %-» o *ed °w Vi B VI Vi Q PH 00 Xi .SP JrHH o CN Q HH < HH a _H 00 3H