ISSN: 1402-1757 ISBN 978-91-7439-XXX-X Se i listan och fyll i siffror där kryssen är

LICENTIATE T H E SI S

Department of Chemical Engineering and Geosciences Division of Extractive Metallurgy Samuel Ayowolea of Upgrading Tetrahedrite-RichAweHydrometallurgical Concentrate Copper Hydrometallurgical Upgrading of a ISSN: 1402-1757 ISBN 978-91-7439-163-3 Tetrahedrite-Rich Copper Concentrate Luleå University of Technology 2010

Samuel Ayowole Awe

HYDROMETALLURGICAL UPGRADING OF A TETRAHEDRITE RICH COPPER CONCENTRATE

Samuel Ayowole Awe

Division of Extractive Metallurgy Department of Chemical Engineering and Geosciences Luleå University of Technology SE-971 87 Luleå Sweden

December 2010 Printed by Universitetstryckeriet, Luleå 2010

ISSN: 1402-1757 ISBN 978-91-7439-163-3 Luleå 2010 www.ltu.se ABSTRACT

Removal of impurity elements in copper metallurgy is one of the major problems encountered today since pure copper ore reserves are becoming exhausted and the resources of unexploited ores often contain relatively high amounts of antimony, arsenic, mercury and bismuth, which need to be eliminated. The present work aims at pre-treating a tetrahedrite rich complex copper sulphide concentrate by selective dissolution of the impurities, therefore, upgrading it for pyrometallurgical processing. Characterisation of the complex concentrate was performed and the result shows that antimony and part of arsenic were present as tetrahedrite and bournonite. Dissolution kinetics of tetrahedrite in aqueous alkaline sodium sulphide solutions was investigated. It was found that the rate of dissolving tetrahedrite by the lixiviant increases with increase in reaction temperature, sodium sulphide concentration, sodium hydroxide concentration, and with decrease in mineral particle size. The kinetic study indicates that the rate of leaching tetrahedrite in the lixiviant under the selected conditions is chemically controlled through the particle surface reaction. The activation energies of the process were estimated as 81 kJ/mol and 75 kJ/mol, respectively, for antimony and arsenic dissolution from tetrahedrite. The estimated activation energies were within the range reported for a chemically controlled reaction process.

Besides, the alkaline sulphide lixiviant proves selective and effective to dissolve these impurity elements from the concentrate with good recoveries. Further investigations on the factors influencing the leaching efficiency of the lixiviant were studied. Analysis of the leach residue indicates that copper content of the tetrahedrite has transformed into copper sulphides with the average chemical formula Cu1.64S. The grade and economic value of the concentrate were improved greatly after sulphide treatment, and therefore, suitable as a feedstock for smelting. The impurities in the concentrate were found to have reduced to a level satisfactory for smelting operation.

Furthermore, modelling and optimisation of alkaline sulphide leaching of a complex copper concentrate containing 1.69% Sb and 0.14% Sn were conducted. Response surface methodology, in combination with central composite face-centred design (RSM-CCF), was used to optimise the operating parameters. The leaching temperature, sulphide ion concentration and solid concentration were chosen as the variables, and the response parameters were antimony and tin recoveries, and the time required to achieve 90% Sb dissolution. It was seen that the leaching process was strongly dependent on the reaction temperature as well as the sulphide ion concentration without any significant dependence on the solid concentration. Additionally, a mathematical model was constructed to characterise the leaching behaviour within the experimental range studied. The results from the model allow identification of the most favourable leaching conditions. The model was validated experimentally, and the results show that the model is reliable and accurate in predicting the leaching process.

iii

ACKNOWLEDGEMENTS

Firstly, I give thanks to God Almighty for His divine favour, protection, wisdom, love and mercy that endureth for ever.

Secondly, I would like to express my sincere gratitude to my supervisor, Professor Åke Sandström for his guidance and valuable suggestions/contributions during the course of this work. More power to your elbow. I am also indebted to Dr. Caisa Samuelsson for her valuable comments and role as co-supervisor of my programme.

Most importantly, I would like to give my profound appreciation to the following organisations: VINNOVA, Boliden Minerals AB, LKAB and the Adolf H. Lundin Charitable Foundation for their financial supports to carry out this research.

A special thank you goes to all the employees and colleagues at the Division of Extractive Metallurgy for their valuable supports and helps. A special recognition is deserving of Birgitta Nyberg for giving me some technical clues during my experimental work. I must sincerely thank the following people from Boliden Mineral AB: Dr. Nils-Johan Bolin, Jan-Eric Sundkvist, Paul Kruger and Andreas Berggren for their contributions and suggestions to the success of this research.

Further thanks to all my friends who in one way or the other have contributed meaningfully to my stay and study at LTU. May God bless you all.

Finally, my heartfelt gratitude is extended to my loving wife, Patricia, and my wonderful children for their understanding and perseverance when I returned home late in the night. I love you all.

iv

LIST OF APPENDED PAPERS

PAPER I

Dissolution kinetics of tetrahedrite mineral in alkaline sulphide media

Samuel A. Awe, Caisa Samuelsson and Åke Sandström

Hydrometallurgy 103 (2010) 167-172

PAPER II

Selective leaching of arsenic and antimony from a tetrahedrite rich complex sulphide concentrate using alkaline sulphide solution

Samuel A. Awe and Åke Sandström

Minerals Engineering 23 (2010) 1227-1236

PAPER III

Separation of antimony from a complex copper concentrate: Process optimization by response surface methodology

Samuel A. Awe, Mohammad Khoshkhoo, Paul Kruger and Åke Sandström

Manuscript under review in Separation and Purification Technology Journal

Other contribution not included

Leaching mechanism of tetrahedrite in alkaline sulfide solution

Samuel A. Awe and Åke Sandström

Proceeding from Minerals Engineering Conference in Luleå, February 2-3, 2010, pp13-24.

v

TABLE OF CONTENT

ABSTRACT...... III ACKNOWLEDGEMENTS...... IV LIST OF APPENDED PAPERS...... V TABLE OF CONTENT...... VI 1. INTRODUCTION...... 1

1.1 IMPURITY MINERALS IN COPPER CONCENTRATE ...... 2 1.2 INDUSTRIAL APPLICATIONS OF ANTIMONY AND ITS COMPOUNDS ...... 3 1.3 THE AIM AND SCOPE OF THE WORK...... 4 2. PROCESSING OF COPPER CONCENTRATES...... 5

2.1 PYROMETALLURGICAL TREATMENT ...... 5 2.2 HYDROMETALLURGICAL PROCESSING...... 6 2.2.1 Acidic chloride leaching of tetrahedrite ...... 6 2.2.2 Alkaline sulphide treatment of tetrahedrite ...... 7 3. MATERIALS AND METHODS...... 11

3.1 MATERIAL...... 11 3.2 MATERIAL CHARACTERIZATION TECHNIQUES ...... 11 3.2.1 X-ray diffractometer (XRD)...... 12 3.2.2 QEMSCAN and SEM-EDS analyses...... 12 3.3 LEACHING ...... 12 4. RESULTS AND DISCUSSION ...... 15

4.1 MINERALOGICAL STUDY ...... 15 4.2 FACTORS AFFECTING TETRAHEDRITE LEACHING (PAPERS I & II) ...... 18 4.2.1 Sulphide ion and hydroxide ion concentrations...... 18 4.2.2 Leaching time ...... 21 4.2.3 Mineral particle size...... 22 4.2.4 Reaction temperature ...... 23 4.3 LEACHING KINETICS AND ACTIVATION ENERGY (PAPER I) ...... 24 4.4 SELECTIVITY OF SULPHIDE LIXIVIANT TO DISSOLVE SB AND AS (PAPER II)...... 26 4.5 ALKALINE SULPHIDE PROCESS MODELLING AND OPTIMIZATION (PAPER III)...... 29 4.5.1 Data evaluation and response surface methodology (RSM) model analysis...... 29 4.5.2 Model validation and optimisation...... 35 5. CONCLUSIONS ...... 39 FUTURE WORK ...... 41 REFERENCES...... 43

vi

1. INTRODUCTION

From the dawn of civilization until today, copper has made, and continues to make, a vital contribution to sustaining and improving society due to its unique properties. Copper’s chemical, electrical, physical and aesthetic properties make it a material of choice in a wide range of domestic, industrial and high technology applications. Its ductility, corrosion resistance, malleability and excellent conductivity of heat and electricity make it to have diverse application potentials in modern material development and engineering [1]. The industrialization of developing economies in Asia, and the drive to improve standards of living in the region, fuelled the demand for copper over the last 10 years. Economic, technological and societal factors have influenced the demand and supply of copper in recent time. The increasing demand for copper worldwide has increased its production and therefore, new mines and plants are introduced and the existing ones are expanded [2].

Naturally, copper occurs in the Earth’s crust in a variety of forms. It can be found in sulphide deposits as chalcopyrite CuFeS2, bornite Cu5FeS4, chalcocite Cu2S & covelite CuS; in carbonate deposits as azurite Cu3(CO3)2(OH)2 & malachite Cu2(CO3)(OH)2; in silicate deposits (as chrysocolla (Cu,Al)2H2Si2O5(OH)4ʘnH2O & dioptase CuSiO2(OH)2) and as pure native copper. Copper is extracted from its ore by either smelting with refining [3, 4] or leaching with electrowinning [4]. Copper extraction through pyrometallurgical processing of concentrate relies heavily on copper sulphides and to a lesser degree on copper oxides, which are processed hydrometallurgically. Today, most of the new copper sulphide mineral deposits found are complex in nature and often found in association with minerals like tetrahedrite, enargite and tennantite [5] which render the concentrate unsuitable as a feedstock for smelting due to its antimony, arsenic and mercury contents which create serious environmental problems [6-8] and even affect the quality of the copper product [9]. Many of such complex sulphide mineral deposits are largely found in Sweden. Due to their association with the impurity elements (Sb, As and Hg) that make it difficult to produce clean and high grade concentrate, the deposits have been discovered for quite long time but have so far not been possible to mine economically. As a result of the global increase in demand for base metals, the mining companies are now giving more attention to these low grade and complex ore deposits to economically process them.

1 Arsenic and antimony have been considered as undesirable elements that cause serious toxicological and environmental problems [7, 10] which have forced the smelter to be conscious of the type of concentrate to be processed due to the stringent environmental law. Apart from this, it is reported that the presence of these noxious elements will significantly affect both the quality and the physical properties of the copper product [9-11]. However, as the global demand for copper continues to grow resulting in a decline in copper deposits with low antimony and arsenic impurities, the focus is now directed on the processing of copper ores with high tetrahedrite, tennantite and enargite contents. The high contents of the impurities in this type of copper resources would significantly reduce their economic values. Therefore, it is highly desirable to reduce the impurity from the ore concentrates to acceptable level prior to shipping to the smelter in order to eliminate smelter penalties and to reduce shipping weight. Furthermore, it is essential that the antimony and arsenic should be removed by a technique which minimises pollution control problems and produces arsenic as a waste product suitable for landfill while antimony is recovered as a marketable products; and consequently upgrades the ore concentrates for pyrometallurgical treatment.

1.1 Impurity minerals in copper concentrate

The common impurity minerals usually found in association with copper ores include tetrahedrite (Cu12Sb4S13), jamesonite (Pb4FeSb6S14), bournonite (PbCuSbS3), tennantite

(Cu12As4S13) and enargite (Cu3AsS4). These minerals are economically attractive; however, the content of antimony and arsenic reduces their economic values effectively due to the additional requirements for treatment options for the containment of hazardous emissions from pyrometallurgical processing [12]. Tetrahedrite is a complex sulphide of copper and antimony which can be an interesting and important resource of copper (40-46%), antimony (27-29%) [13] and or other non-ferrous metals as is evident from its general chemical formula

(Cu,Ag)10(Cu,Zn,Hg,Cd,Fe)2(Sb,As,Bi)4S13 [14]. Tetrahedrite belongs to a family of minerals usually called “fahlore” which is often described as the “sulphosalts”. This family indicates a certain type of un-oxidized sulphur mineral that is structurally distinct from a sulphide [15]. The tetrahedrite group is notable for the variety of elements which are stable in its structure.

The formula may be written as A12B4C13, where A = Ag, Cu, Fe, Hg, Zn; B = As, Sb, Bi; C = S, Se, Te. Naturally, it is very difficult to find pure tetrahedrite because its original constituents are always partially substituted by other elements. Thus, arsenic can substitute for antimony leading to tennantite, while copper can be substituted by silver, zinc, iron, mercury,

2 cadmium, lead etc, and some sulphur may be replaced by Se and Te [7, 15, 16]. Apart from the toxic elements (As, Hg) contained in tetrahedrite mineral, the substantial amount of copper, silver and gold usually present in the mineral makes its mining economically attractive.

1.2 Industrial applications of antimony and its compounds

Antimony is a brittle metal that is not readily fabricated and has no significant use in its unalloyed state. When alloyed with lead and other metals, it increases their hardness, mechanical strength, corrosion resistance, and electrochemical stability or decreases their coefficient of friction. Some antimony alloys expand slightly upon cooling, a valuable property for use in type metal and other castings. Besides, antimony and its compounds have a wide range of non-metallurgical industrial applications as shown in Fig. 1. Other uses 10%

Pigments 9%

Ceramics and glass 10% Flame retardants Plastics 57% stabilizers and catalysts 10% Lead-acid batteries 4%

Fig. 1: Applications of antimony [17]

The recent study report submitted by the Ad-hoc Working Group of the European Commission on defining the critical raw materials for the European Union (EU) has listed

3 antimony as one of the critical materials needed by the EU [18]. This implies that the price of antimony will continue to rise due to the risk of supply shortage of antimony to the EU economy. Based on this study, recovery of antimony as a marketable product will be of a paramount importance to the mining and metals industries. However, antimony which was regarded as a waste material from copper metallurgy will add values to the process economy.

1.3 The aim and scope of the work

In copper metallurgy, the pyrometallurgical processing route still accounts for almost 80% of copper production throughout the World. But this treatment method can only process relatively pure sulphide copper-concentrates efficiently and economically. As a result of the depletion of high-grade copper sulphide ore deposits, most of the copper ores mined today are complex in nature and even contain a number of impurity elements which attract smelter penalties. Besides, the high contents of the impurities usually accompanied with such ores significantly reduce their economic value. Therefore, the present research aims at improving the grade of the copper sulphide concentrate as a feedstock suitable for pyrometallurgical processing, by a selective removal of the impurities (As, Sb & Sn) in the concentrate via alkaline sulphide hydrometallurgical technology. The scope of this research includes x A mineralogical study of the various materials used x Investigation of tetrahedrite dissolution kinetics and the factors influencing its dissolution in alkaline sulphide media x Examining the effectiveness of the lixiviant to selectively solubilize the impurities from the concentrate and x Modelling and optimization of the alkaline sulphide leach process

4 2. PROCESSING OF COPPER CONCENTRATES

There are two major ways of processing copper concentrates. These include pyrometallurgical and hydrometallurgical routes. Copper sulphide ores are desirable feedstock materials for pyrometallurgical production of copper while the oxide and lean sulphide copper ores are treated via hydrometallurgical routes. Further discussion about the copper production routes is explained in the following sections.

2.1 Pyrometallurgical treatment

Pyrometallurgical processing of copper concentrates involves high temperature reactions which include roasting, smelting and conversion. The prevalence of antimony/arsenic minerals among the copper-bearing minerals, and as a result, the relatively high impurity content in the concentrates substantially reduces their economic value, owing to the hazardous emissions generated from pyrometallurgical processing (Eqs. 1-3). The amount of impurity elements removed during the roasting of the concentrate is very high. These impurities, as well as their oxides (Eqs. 2 and 3), are highly volatile and leave the reactor as off-gas constituents. Thus, in unfavourable metal market conditions, direct roasting of such impurity- containing copper concentrate is uneconomical because gas cleaning facilities demands are too expensive [19]. In order to minimize the problems associated with the processing of such concentrate, the impurity content must be reduced to low levels acceptable in the smelting plant.

2CuFeS2 + O2 ĺ Cu2S + 2FeS + SO2 (1)

4Cu3AsS4 + 7O2 ĺ 12CuS + 4SO2 + 2As2O3 (2)

Cu2SʘSb2S3 + 9/2O2 ĺ Cu2S + 3SO2 + Sb2O3 (3)

In addition, Wang [11] discussed that during copper smelting, some of the undesirable impurities like antimony, arsenic, bismuth and lead are only partially removed by oxidation. When white metal and blister copper are in equilibrium, these impurities are partly distributed into the copper phase, from which their removal is difficult. In an attempt to reduce the impurity contents, alkaline sulphide hydrometallurgical technology has been explored and considered suitable for dissolving these impurities selectively from the concentrate prior to pyrometallurgical processing.

5 2.2 Hydrometallurgical processing

As the global demand for copper steadily increases and the need for environmentally sound processing also grows, successful hydrometallurgical treatment of antimony and arsenic containing concentrates has come to the forefront. The amphoteric nature of antimony allows the use of both acidic and basic media for its solubilization [20, 21]. But, in reality only two solvent systems are utilized in antimony hydrometallurgy: acidic chloride [22-27] and alkaline sulphide [21, 28-30].

2.2.1 Acidic chloride leaching of tetrahedrite

Extensive work has been performed by several researchers on acidic oxidative leaching of tetrahedrite minerals in chloride [14, 16, 22-25, 27], and even sulphate media [5, 7]. According to the stability of tetrahedrite, it is necessary to use a strong oxidizing agent. The 2+ 3+ most commonly oxidizing agents used are Cu , Fe , O2, ozone and Cl2 [16]. The atmospheric leaching of tetrahedrite in chloride media has been studied in detail by Correia et al. [23, 24]. They found that at temperatures 60–104 °C and ferric concentrations 0.001–1 mol/L, leaching of tetrahedrite in FeCl3–NaCl–HCl solutions is controlled by a surface reaction. The process is extremely slow [23, 25]; it takes about 50 h to extract 90% of copper at 104 °C from tetrahedrite particles of size -100 + 63 μm [24]. Correia et al. [23] observed that tetrahedrites with high iron content dissolved faster than iron-poor tetrahedrites, a phenomenon that was interpreted in terms of the mineral semiconducting properties. The use of ozone (O3) can improve the rate of tetrahedrite dissolution [14] but, even under the best conditions tested (tetrahedrite particle size -45 μm; 9.0 g/h O3 per L of solution; ambient temperature), it takes about 6 h to reach 50% extraction of copper [7].

Havlik et al. [22] conducted a leaching investigation with native tetrahedrite concentrate in acidic ferric chloride in the temperature range of 40-90 °C. The results reveal slow leaching rate behaviour exhibiting apparent parabolic kinetics. It was observed that increasing the temperature improves the metal extraction and the apparent activation energy was 38 kJ/mol. This indicates that the process occurs in the mixed regime of both diffusion and chemical reaction. Riveros and Dutrizac [5] performed a series of leaching test on tetrahedrite

(Cu12Sb4S13) in acidic media using Fe2(SO4)3, FeCl3 or O2 as oxidizing agents. At temperatures <100 °C , the dissolution kinetics of disks of synthetic iron-bearing tetrahedrite

6 in Fe2(SO4)3 - H2SO4 media are linear; the leaching rates are slow but increase significantly with increasing temperature with an apparent activation energy of 120 kJ/mol. The dissolution of sized particles of natural Ag-bearing tetrahedrite in FeCl3-HCl solutions proceeds according to the shrinking core model and both Cu and Ag dissolve at about the same rate provided that the total chloride concentration of the solution is sufficiently high to solubilize the AgCl reaction product. The leaching rates were slow and the apparent activation energy was 116 kJ/mol.

By acidic oxidative leaching, e.g. in acidified ferric chloride solutions, copper and iron enter into solution, while antimony is partially precipitated as a compound with a composition similar to the mineral tripuhyite (FeSbO4) and the overall leaching reaction proceeds slowly with complicated kinetics [31].

2.2.2 Alkaline sulphide treatment of tetrahedrite

Leaching of copper-antimony-arsenic sulphide minerals in alkaline sulphide media have been studied [2, 9, 13, 21, 28-30, 32-38] and even applied industrially [39-41]. Antimony solubilizes under basic sulphide conditions while other metals like copper, silver, zinc, lead and iron remain in the residue [2, 16]. Usually, leaching is conducted in alkaline solution with about 100% stoichiometric excess of sodium sulphide (Na2S) for 4 -12 h at temperatures close to the solution boiling point. The lixiviant must contain enough sodium hydroxide (NaOH) to avoid the hydrolysis of sulphide ions (S2-) to hydrosulphide (HS-), which reduces the leaching efficiency [7, 42]. The tetrahedrite-tennantite mineral series are refractory to most common leaching processes [16, 22]. Consequently, Balaz et al. [13, 21, 28, 29, 34, 43, 44] carried out an intensive grinding on such mineral prior to its leaching in alkaline sulphide solution such that significant dissolution of antimony and arsenic can be obtained within a short period of leaching. However, this method is uneconomical due to the higher energy required for the grinding. Many arsenic-sulphur minerals solubilize in the lixiviant with the exception of arsenopyrite which is resistant to the reagent [7, 31, 41]. Filippou et al. [7] reported that the copper product obtained after sulphide leaching may be chalcocite (Cu2S) or covelite (CuS), 3- while arsenic and antimony are dissolved as thioanions: thioarsenate (AsS4 ), thioantimonite 3- 3- (SbS3 ), thioantimonate (SbS4 ) etc., depending on the feed composition and the reaction conditions. The chemistry of the reactions taken place during alkaline sulphide leaching is briefly described as follow:

7 Cu12Sb4S13(s) + 2Na2S (aq) ĺ 5Cu2S(s) + 2CuS(s) + 4NaSbS2 (aq) (4)

NaSbS2 (aq) + Na2S (aq) ĺ Na3SbS3 (aq) (5)

Sb2S3 (s) + 3Na2S (aq) ĺ 2Na3SbS3 (aq) (6)

2Cu3AsS4(s) + 3Na2S (aq) ĺ 3Cu2S(s) + 2Na3AsS4 (aq) (7)

As2S3 (s) + 3Na2S (aq) ĺ 2Na3AsS3 (aq) (8)

It is reported [7, 33] that when the lixiviant is made by dissolving elemental sulphur in sodium hydroxide solution, a solution containing a mixture of sodium sulphide, sodium thiosulphate and sodium polysulphides will be formed according to reactions 9 and 10.

0 4S (s) + 6NaOH(aq) ĺ 2Na2S(aq) + Na2S2O3(aq) + 3H2O (9)

0 (X-1)S (s) + Na2S(aq) ĺ Na2Sx, where x = 2 to 5 (10)

The S/NaOH lixiviant was used by the Sunshine Mining and Refining Company to dissolve antimony from tetrahedrite. Anderson et al. [33] reported that the lixiviant can leach gold and silver associated with the mineral due to the thiosulphate and polysulphides in the reagent which are good oxidants and complexing agents. The flowsheet describing the process at the Sunshine Mining plant is presented in Fig. 2a.

8 Fig. 2a: The Sunshine refinery flowsheet [7].

Similarly, the alkaline sodium sulphide leach process (Fig. 2b) was used at Equity Silver Mines Ltd., British Columbia, Canada, to remove soluble antimony and arsenic from a copper-silver concentrate of average composition 17% Cu, 7% Sb, 4% As, 23.5% Fe, 7,000 g/t (0.7%) Ag and 20 g/t Au [41, 45]. The process employed at Equity Silver plant (Fig. 2b) had many similarities with the process adopted in the Sunshine refinery (Fig. 2a), but it also had some significant differences. At the Equity plant, antimony was entirely recovered as sodium hydroxyl antimonate and arsenic was converted to calcium arsenate to be impounded at a site far from the mine [7, 40, 41] whilst antimony was solely electrowon at the Sunshine plant [46].

Above all, literatures [13, 14, 20, 28] have shown that alkaline sulphide technology is by far the most utilized because of its inherent selectivity to dissolve antimony, arsenic, mercury, tin [7], and its ease of full-scale application due to minimal corrosion problems associated with the acidic leaching system. Conversely, acidic leaching has so far been limited to laboratory

9 applications due to low metal recoveries and complications with the accompanying elements [14].

Fig. 2b: The Equity Silver Mining Plant flowsheet [7].

10 3. MATERIALS AND METHODS

3.1 Material

The materials used in this investigation were a tetrahedrite-rich sulphide concentrate originated from Casapalca, Departamento Lima, Peru and Rockliden complex copper concentrate supplied by the Boliden Mineral AB, Sweden. The elemental analyses of the concentrates are shown in Tables 1 and 2. (For details see paper I, II and III).

Table 1: Elemental analysis of a tetrahedrite-rich sulphide concentrate

Size fraction -106+75 μm -75+53 μm -53+38 μm Cu% 15.6 15.2 15.7 Fe% 11.1 14.2 13.3 Zn% 16.1 16.8 16.9 Sb% 5.78 5.43 5.71 As% 1.87 1.86 1.77

Table 2: Elemental analysis of Rockliden complex copper concentrate

Cu, % Fe, % Pb, % Zn, % Sb, % Sn, % As, % Se, % Hg, % Ag, % S, % 17.8 27.1 7.1 5.7 1.69 0.14 0.42 0.06 0.03 0.08 >15

3.2 Material characterization techniques

Mineralogical study is very important to the extractive metallurgist because it provides detailed information about the associated minerals in the material to be processed. Consequently, an appropriate processing route for the material can be designed. The various mineralogical techniques employed during this investigation are X-ray diffractometer (XRD), QEMSCAN (Quantitative Evaluation of Minerals by SCANning electron microscopy) and scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM- EDS) analyses.

11 3.2.1 X-ray diffractometer (XRD)

A Siemens D5000 automatic X-ray diffractometer equipped with a continuous scanning device was employed to characterize the concentrate as well as the leach residue. XRD patterns were collected between 2ș angles of 10–90° (at Cu KĮ radiation of 40 kV, 30 mA). Mineral phases were identified using the Joint Committee for Powder Diffraction Standards (JCPDS) file of the instrument.

3.2.2 QEMSCAN and SEM-EDS analyses

A representative sample was taken from the tetrahedrite-bearing concentrate as well as its leach residue for QEMSCAN analysis. SEM-EDS technique was used to characterise the Rockliden complex copper concentrate. In both cases, the samples were embedded in epoxy and a diamond cutting wheel was used to cut the hardened blocks to size and expose the sample grains. The samples were polished and carbon coated before the analyses were performed on them. The mineralogy of particles was inferred from data obtained using the QEMSCAN® advanced mineral analysis system. The particles were examined using the Field Scan method, involving spot analyses over the surface of the block, collection of the resultant X-ray spectra and comparison with a spectral data base.

3.3 Leaching

Alkaline sulphide leaching tests were performed by dissolving the concentrate in the lixiviant, to determine the effect of the following parameters: sodium sulphide (Na2Sʘ3H2O) concentration (0.76 M, 1.14 M and 1.89 M), sodium hydroxide (NaOH) concentration (0.75 M and 1.5 M), leaching temperature (357 K, 364 K, 371 K and 378 K) and particle size (-53 + 38 μm, -75 + 53 μm and -106 + 75 μm). All the leaching experiments, each lasting 6 h, were carried out with 0.5% solid, -75 + 53 μm particle size and 378 K reaction temperature (except where otherwise stated) in 500 mL leach solution containing 1.14 M Na2Sʘ3H2O + 1.5 M NaOH. A 500 mL four-necked round bottomed glass reactor was used which was stirred mechanically with a paddle stirrer at a constant stirring rate of 600 rpm and heated in an auto- regulated system (Fig. 3). The lixiviant was first added to the reactor and when the desired temperature was reached, the solid sample was added. At predetermined time intervals, 5 mL sample slurry was taken from the reactor for the analysis of dissolved metals. All the reagents used for leaching and chemical analysis were of analytical grade and used without further

12 purification. The influence of each variable was determined by keeping all other variables constant.

Fig. 3: Leaching experimental set-up.

In an attempt to optimise the leaching process, MODDE 8.0 Umetrics software was used in the design of the experiment and analysis of the test results. Response surface methodology in combination with central composite face-centred design (RSM-CCF) was used for the experimental design and analysis. Leaching tests were conducted according to Table 3 (Details about the procedure is reported in paper III). Leaching results were evaluated by

13 means of chemical determinations on the leach products using Inductively Coupled Plasma- Atomic Emission Spectrometry (ICP-AES) / Sector Field Mass Spectrometry (ICP-SFMS).

Table 3: Central composite face-centred design arrangements and responses for the complex concentrate leaching by alkaline sulphide solution

Exp.no. Run Experimental factors Response variables order Temperature S-2 conc. Solid conc. Sb Sn Leaching X1 (°C) X2 (g/L) X3 (g/L) recovery recovery, time, % % (T90) h N1 13 80 41 100 43.32 34.02 26.53 N2 2 100 41 100 78.11 39.43 8.89 N3 10 80 164 100 93.98 36.04 4.39 N4 3 100 164 100 97.25 68.28 3.48 N5 1 80 41 300 39.10 35.42 29.93 N6 12 100 41 300 75.77 31.03 9.76 N7 8 80 164 300 94.27 37.93 4.40 N8 4 100 164 300 97.55 82.98 3.46 N9 5 80 102.5 200 87.03 31.52 6.46 N10 16 100 102.5 200 90.98 39.22 5.34 N11 9 90 41 200 63.10 32.86 13.81 N12 6 90 164 200 95.88 48.55 3.97 N13 15 90 102.5 100 91.03 33.40 5.25 N14 14 90 102.5 300 88.00 33.74 6.13 N15 11 90 102.5 200 89.19 33.18 5.72 N16 17 90 102.5 200 91.51 34.29 5.18 N17 7 90 102.5 200 91.03 31.97 5.34

14 4. RESULTS AND DISCUSSION

4.1 Mineralogical study

The two concentrates used in the study were characterised by XRD, QEMSCAN and SEM- EDS analyses in order to understand the major mineral constituents of each concentrate. The major mineralogical phases identified in the diffractograms of both concentrates (Fig. 4) were tetrahedrite, sphalerite, galena, chalcopyrite, pyrite and iron sulphide (check papers II and III for further detail).

Fig. 4: XRD patterns of Rockliden and a tetrahedrite-rich concentrates

QEMSCAN analysis was conducted on the tetrahedrite-rich concentrate to confirm the mineralogy previously determined by the XRD technique. The result from this method is shown in Fig. 5 and Table 4. This result compares well with the XRD identification of the major phases present in the concentrate, but the method identifies a much wider range of minor minerals (Table 4) that were below the detection limit for XRD. This confirms that QEMSCAN analysis can provide much more mineralogical information than the XRD method for trace mineralogy [47]. No arsenic mineral was detected by the two techniques meaning that the arsenic in the mineral is hosted in the tetrahedrite structure. Also, the QEMSCAN result shows that the concentrate is rich in tetrahedrite (Table 4) and silver is

15 found bonded in the tetrahedrite (Fig. 5). This confirms the presence of silver and arsenic as a solid solution in the tetrahedrite crystal structure as widely reported in the literature [7].

Table 4: Mineralogy of a tetrahedrite-rich concentrate defined by QEMSCAN

Mineral Mass fraction (% w/w) Tetrahedrite 30.2 Sphalerite 20.6 Galena 19.0 Pyrite 15.7 Chalcopyrite 13.4 Quartz 0.4 Tetrahedrite with Ag 0.1 Andalusite (Al2SiO5) 0.1 Total 99.5

Background Hole x Covellite x Chalcopyrite x Tetrahedrite x Tetrahedrite with Ag x K feldspar x Quartz x Galena x Sphalerite x Andalusite x Pyrite Mineraler Pb Other Sb Other Zn Other Si Other As Other Ag Other Other

Fig. 5: QEMSCAN micrograph of a tetrahedrite-rich concentrate.

SEM-EDS technique was used to verify the mineralogical constituents of Rockliden concentrate. The result, as shown in Fig. 6, corroborates the one obtained from the XRD analysis, but the method further identifies the minor minerals (Table 5) that were not detected by XRD technique. The minerals recorded in Table 5 were estimated from the atomic weight percent obtained from the SEM-EDS analysis. Apart from tetrahedrite, SEM result (Fig. 6) shows that antimony was also found as bournonite (PbCuSbS3) while tin and arsenic was present as stannite (Cu2FeSnS4) and arsenopyrite, respectively. It is apparent from Table 5

16 that silver is bound in the tetrahedrite structure which may enhance the processing economy of the concentrate.

Fig. 6: SEM micrograph of Rockliden copper concentrate. A- tetrahedrite, B- bournonite, C- stannite, D- chalcopyrite, E- galena, F- sphalerite, G- pyrite, H- arsenopyrite and I- pyrrhotite.

Table 5: Mineralogy and atomic proportions of selected microanalyses of Rockliden copper concentrate defined by SEM-EDS analysis

Cu % Fe % Zn % Pb % Sb % As % Sn % Ag % S % Mineral 1 35.41 5.36 30.92 1.31 26.99 Tetrahedrite 2 33.71 5.77 1.62 31.34 2.03 25.55 Tetrahedrite 3 34.10 6.23 2.00 30.14 1.34 26.19 Tetrahedrite 4 34.10 6.05 30.08 3.17 26.60 Tetrahedrite 5 12.48 1.06 38.81 27.11 20.55 Bournonite 6 13.06 0.84 39.01 26.10 20.99 Bournonite 7 12.54 1.37 38.35 26.33 21.41 Bournonite 8 27.23 12.57 29.11 31.09 Stannite 9 28.62 12.51 27.96 30.92 Stannite 10 0.99 32.89 44.79 21.32 Arsenopyrite 11 33.53 46.07 20.40 Arsenopyrite 12 33.15 30.01 36.84 Chalcopyrite 13 32.72 30.17 37.11 Chalcopyrite 14 1.34 83.07 15.59 Galena 15 8.14 55.58 36.28 Sphalerite 16 44.17 55.83 Pyrite

17 4.2 Factors affecting tetrahedrite leaching (Papers I & II)

4.2.1 Sulphide ion and hydroxide ion concentrations During sulphide leaching, sodium sulphide dissociates according to reaction 11. The protonation of sulphide ions occur according to the equilibrium reactions 12 and 13 [36] which take place simultaneously during the leaching process. + 2- Na2S ļ 2Na + S (11) 2- + - S + H ļ HS log K1 = 11.96 (12) - + HS + H ļ H2S log K2 = 7.04 (13)

2- - Fig. 7 shows a plot of the molar concentrations of S , HS and H2S in the leach liquor as a function of the solution pH at a total sulphide concentration of 1.14 M. It can be inferred from Fig. 7 that sulphide ion concentration increases rapidly when pH is increased from 10 and upwards, therefore, pH is an important factor. This is because, at a pH > 12, the protonation of sulphide ions will be prevented and therefore available to solubilize antimony according to reactions 4 and 5. Also, a strong alkaline environment is needed to reduce the consumption of more expensive sodium sulphide due to the hydrolysis of sulphide ions (Eq. 14). 2- - - S + H2O ļ HS + OH (14)

1.2

1

0.8

S2- 0.6 HS- H2S Concentration, M 0.4

0.2

0 567891011121314 pH

Fig. 7: Equilibrium diagram of sulphide species as a function of pH at a total sulphide ion concentration of 1.14 M at 298 K.

18 According to Fig. 8, it is apparent that Na2S concentration has a significant effect on the extraction of antimony and arsenic from the complex sulphide concentrate. After 6 h leaching, approximately 21% of antimony and arsenic was extracted at a concentration of 0.76 M Na2S and when the Na2S concentration was increased to 1.14 M, the extraction increased with almost a factor of 3. Further increase in the concentration to 1.89 M Na2S resulted in an extraction of antimony and arsenic of 87% and 92%, respectively (Fig. 8).

100

90 Sb

80 As

70

60

50

40

Metal extraction, % extraction, Metal 30

20

10

0 0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1

Na 2S concentration, M

Fig. 8: Effect of sodium sulphide concentration on As/Sb leaching.

In addition, it is indicated from Fig. 9 that the natural logarithm of the rate constant (lnkr) is a linear function of ln[Na2S] when the kinetic data obtained from the study were plotted. The kinetic equations about the effect of the initial concentration of Na2S on the leaching rate of antimony and arsenic can be written as shown in Eqs. 15 and 16, respectively. lnkr = 2.2 ln[Na2S] – 7.77 (15) lnkr = 2.3 ln[Na2S] – 7.76 (16) The apparent reaction order for dissolving antimony and arsenic from tetrahedrite by the lixiviant was estimated to be approximately equal to 2, which implies that both elements reacted similarly with the lixiviant as should be expected since they were present in the same mineral.

19 -6 y = 2.3x - 7.76 R2 = 0.9748 -6.5 Sb As -7 y = 2.2x - 7.77 -1 R2 = 0.9742

min -7.5 r, lnk

-8

-8.5

-9 -0.4 -0.2 0 0.2 0.4 0.6 0.8

ln[Na2S], M

Fig. 9: Determination of the reaction order with respect to [Na2S].

Furthermore, the effect of hydroxide ion concentration on the extraction of antimony and arsenic was investigated and plotted in Fig. 10. It is illustrated in the figure that when a low concentration (0.75 M) of NaOH was used, the rate of extracting antimony and arsenic was slower and about 25% and 24% of antimony and arsenic were dissolved after 6 h leaching, respectively. As the concentration of NaOH increased to 1.5 M, the rate of metal extraction was accelerated and the percentage of metal extracted was enhanced. About 57% antimony and 60% arsenic was leached by the lixiviant after 6 h. The observed increase in metal extraction at high concentration of NaOH is due to the higher availability of S2- ion for dissolution action (as described in Paper II), which is promoted by the increase in the pH of the leach liquor. Consequently, the hydrolysis of sulphide ion is prevented which otherwise would reduce its leaching efficiency.

20 70

0.75 M NaOH, Sb 60 1.5 M NaOH, Sb

0.75 M NaOH, As 50 1.5 M NaOH, As

40

30

Metal extraction, % extraction, Metal 20

10

0 01234567 Time, h

Fig. 10: Effect of sodium hydroxide concentration on As/Sb leaching.

4.2.2 Leaching time

The effect of leaching time on antimony and arsenic dissolution from the complex sulphide concentrate was investigated. The experimental results are presented in Fig. 11. It is obvious from the figure that, by increasing the leaching time, antimony and arsenic extraction would be improved. Due to the narrow size fraction used in this investigation, the metal extraction by the lixiviant is linearly dependent on the leaching time. After 1 h of leaching, approximately 13% of both antimony and arsenic was extracted, and when the leaching time was increased to 6 h, antimony and arsenic extraction was 57% and 60%, respectively. By examining the extraction trend shown in Fig. 11, it can be assumed that higher extraction would be achieved if the leaching time was extended beyond 6 h.

21 70

60 Sb As 50

40

30

Metal extraction, % extraction, Metal 20

10

0 01234567 Leaching time, h.

Fig. 11: Effect of leaching time on As/Sb leaching.

4.2.3 Mineral particle size

The influence of particle size on antimony and arsenic leaching from the tetrahedrite-rich concentrate was studied using three particle sizes (-53 + 38, -75 + 53 and -106 + 75 μm) at

105 ºC in leach solutions containing 150 g/L Na2Sʘ3H2O + 60 g/L NaOH. A 2.5 g of the concentrate was leached in 500 mL of the solution. It is observed from Fig. 12 that the rate of dissolving antimony and arsenic decreases with increase in particle size. However, it can be concluded that the smaller the particle size, the faster was antimony and arsenic leaching by the lixiviant.

22 80

70 Sb -53+38 um Sb -75+53 um 60 Sb -106+75 um As -53+38 um 50 As -75+53 um As -106+75 um 40

% Recovery 30

20

10

0 0 50 100 150 200 250 300 350 400 Leaching time, min

Fig. 12: Effect of particle size on As / Sb leaching.

4.2.4 Reaction temperature

In order to investigate the effect of reaction temperature on antimony and arsenic dissolution from a tetrahedrite-rich concentrate, the experiments were performed under the following leaching conditions: mineral particle size of -75 + 53 μm, 150 g/L Na2S, 60 g/L NaOH and temperatures ranging from 84 ºC to 105 ºC. The amount of the concentrate leached was kept constant at 5 g/L. Fig. 13 illustrates the results obtained from the tests. It could be inferred from the figure that the reaction temperatures have a significant effect on the rate of antimony and arsenic dissolution. One can infer from Fig. 13 that the leaching rate increases rapidly with increasing time and temperature. After 360 minutes of leaching, about 57% and 60% of antimony and arsenic were dissolved, respectively, at a reaction temperature of 105 ºC.

23 70 Sb 84 0C Sb 91 0C 60 Sb 98 0C Sb 105 0C As 84 0C As 91 0C 50 As 98 0C As 105 0C

40

30 % Recovery

20

10

0 0 50 100 150 200 250 300 350 400 Leaching time, min

Fig. 13: Effect of leaching temperature on As / Sb leaching.

4.3 Leaching kinetics and activation energy (Paper I)

The important parameters of the leaching process are the overall rate and the variation in rate with leaching time. Since, these will determine the size of the reactors needed for a design criterion of plant capacity, and the degree of extraction that would be experienced in the plant when it is operated at a selected capacity, above or below, the design capacity. Leaching rates are usually investigated on a laboratory-scale, and the kinetic data obtained from the study are used in the scaling-up of the process [48]. Besides, a careful kinetic study is required to accurately establish the impact of variations in operating rate on metal’s extraction, so as to make wise choices of operating rates for variations in market conditions. When surface chemical reaction is rate controlling, the kinetics may be correlated graphically using Eq. 17.

1 k M C 3 c S A 1 1 X t kr t (17) U S Er0 where X is the fraction reacted, kc is the kinetic constant, MS is the molecular weight of the solid, CA is the concentration of the dissolved lixiviant A in the bulk of the solution, ȡS

24 density of the concentrate, ȕ is the stoichiometric coefficient of the reagent in the leaching reaction, r0 is the initial radius of the solid particle, t is the reaction time and kr is the rate constant. Application of Eq. 17 to the experimental data obtained, at different temperatures, resulted in linear plots presented in Fig. 14.

0.3 Sb 84 0C Sb 91 0C

0.25 Sb 98 0C Sb 105 0C

As 84 0C As 91 0C 0.2 As 98 0C As 105 0C 1/3 0.15 1- (1-X)

0.1

0.05

0 0 50 100 150 200 250 300 350 400 Leaching time, min

Fig. 14: A plot of 1-(1-X)1/3 vs. time at various reaction temperatures for antimony and arsenic dissolution in alkaline sulphide solution.

It is observed, from this plot, during the whole reaction time that the data in the figure are linear, which indicate that the rate of the reaction is chemically controlled through the particle surface reaction. The slopes of the straight lines from each plot in Fig. 14 were estimated which is equivalent to the rate constants, kr. The temperature dependence of the chemical reactions can be given by the Arrhenius equation [49, 50] as follows:

kr =A exp(íEa/RT) (18)

Where A is the frequency factor and Ea is the apparent activation energy. The Arrhenius plot shown in Fig. 15 was made by plotting the natural logarithm of the rate constant (lnkr) against the reciprocal of the temperature (1/T). Hence, the activation energies of antimony and arsenic

25 were calculated from the slopes of the Arrhenius plots presented in Fig. 15 as 81 kJ/mol and 75 kJ/mol, respectively, which is in agreement with the reported activation energy for a chemically controlled process [49]. Since the calculated activation energies are high, it can be concluded that the dissolution rate of antimony and arsenic from tetrahedrite by alkaline sulphide solution is very sensitive to the reaction temperature.

-6.5

2 Sb R = 0.9893 -7.0 As R2 = 0.9437

-7.5 -1

-8.0 , min r lnk

-8.5

-9.0

-9.5 2.62 2.64 2.66 2.68 2.70 2.72 2.74 2.76 2.78 2.80 2.82 103/T, K-1

Fig. 15: Arrhenius plot for tetrahedrite leaching in alkaline sulphide solution.

4.4 Selectivity of sulphide lixiviant to dissolve Sb and As (Paper II)

In order to identify the mineralogical changes brought about by leaching, the solid residue obtained after leaching was characterized by XRD and QEMSCAN analyses. The results were compared with the mineralogy of the original concentrate. The XRD pattern shown in Fig. 16 indicates that a new mineralogical phase, spionkopite (Cu39S28) is observed together with other major phases identified in the original material (Fig. 4). The XRD pattern in Fig. 16 shows that the tetrahedrite peaks found in the concentrate have disappeared completely implying that copper content of the tetrahedrite had been transformed to the new copper sulphide phase observed in the residue.

26 Fig. 16: XRD analysis of the leach residue after alkaline sulphide leaching (Paper II).

In addition, QEMSCAN analysis results on the leach residue are presented in Figs. 17 and 18. The major mineral phases observed in the residue were sphalerite, galena, pyrite and chalcopyrite, with the newly formed copper sulphide phase. In particular, observation of the polished sections of the leach residue (Fig. 17) showed a micro-porous structure of the new solid phase often enclosing a nucleus of untransformed tetrahedrite. The average crystal chemical formulae of the solid residue determined by QEMSCAN analysis on the transformed phase proves the conversion of tetrahedrite into a copper sulphide having stoichiometry of

Cu1.64S, which is close to the stoichiometry of the spionkopite observed by XRD but can also be a mixture of chalcocite and covellite. This observation confirms the expectation that tetrahedrite would decompose to either covellite or chalcocite in alkaline sulphide media [7]. However, if the leaching time had been prolonged further, the untransformed tetrahedrite would have been converted completely into the new copper sulphide phase observed in the residue.

27 Background Hole x Covellite x Chalcopyrite x Tetrahedrite x Tetrahedrite with Ag x K feldspar x Quartz x Galena x Sphalerite x Andalusite x Pyrite Mineraler Pb Other Sb Other Zn Other Si Other As Other Ag Other Other

Fig. 17: QEMSCAN micrograph of leach residue.

Furthermore, QEMSCAN analysis provides more interesting information about the selectivity of the lixiviant to solubilize antimony and arsenic from the concentrate containing them. Comparing the information given in Fig. 18 concerning the concentrate and the residue, one can see that the major phases found in the concentrate were upgraded and not leached with the exception of tetrahedrite phase which was reduced from 30.2% to 1.1% in the residue. This result confirms that other metal sulphides in the concentrate remains insoluble in the sulphide solution with the exception of antimony and arsenic that form soluble sulphide complexes.

This analysis shows that the Na2S/NaOH lixiviant is suitable for pre-treating, difficult to treat, complex antimony-arsenic-containing sulphide ores prior to their smelting and thereby avoiding the difficulty experienced when such ores are directly treated in smelting plant.

28 35.00

30.00

25.00

20.00

15.00 Mass fraction (%) fraction Mass 10.00

5.00

0.00 Tetra Chalc o Tetra Cu1.64S hedrite w ith Galena Sphalerite Pyrite Quartz pyrite hedrite Ag

Sb-bearing concentrate 0.02 13.35 30.23 0.11 18.99 20.55 15.71 0.41 Residue 6.89 19.30 1.12 0.00 21.27 28.45 21.87 0.48

Fig. 18: Mineral composition of the concentrate and leach residue.

4.5 Alkaline sulphide Process modelling and optimization (Paper III)

In this section, the study conducted on the modelling and optimisation of the alkaline sulphide leach process to selectively dissolve antimony and tin from the copper concentrate containing them is discussed. Tin removal from the concentrate was considered as one of the response variables because the concentrate under consideration contains about 0.14% Sn, which may be interesting to remove prior to smelting. Moreover, arsenic was present as arsenopyrite in the concentrate which is very refractory to the alkaline sulphide lixiviant.

4.5.1 Data evaluation and response surface methodology (RSM) model analysis

MODDE 8.0 software was used for the design and analysis of the experimental results. The experimental factors investigated were leaching temperature (X1), sulphide ion concentration

(X2) and solid concentration (X3), while the output variables were antimony recovery (Sb), tin recovery (Sn) and T90 (i.e. time required to leach out 90% Sb from the concentrate). The distributions of the data for all the responses were inspected for normality, and it was observed that the responses needed to be transformed in order to obtain improved models.

29 Negative logarithmic transformation was carried out for Sb recovery, and power transformation was performed for both Sn recovery and T90 as shown in Table 6.

Table 6: Response transformation

Response Transformation Sb (%) -10Log(100-Y) Sn (%) Y-1 -0.5 T90 (h) Y

Evaluation of the raw data revealed that the replicate errors were satisfactorily small for all of the output responses. The model was fitted using multiple linear regression (MLR) method, and the overall result of the model fitting is displayed in Fig. 19.

Fig. 19: Summary of fit plot of the initial modelling.

Moreover, a model can be judged as good, if R2- Q2< 0.2-0.3, Q2>0.5, model validity >0.25 and reproducibility is greater than 0.5 [51]. It is observed that the predictive power (Q2) of the 2 models ranges from excellent to good. The Q values for the responses Sb, T90 and Sn are 0.93, 0.91 and 0.55, respectively. All models also show high model validity, which is an

30 indication of no lack of fit. Fig. 19 shows that Sb recovery and T90 models satisfied the model performance indicator conditions while the Sn recovery model did not. The lower Q2 value of Sn recovery model could be due to the presence of irrelevant terms contained in the regression model. The regression coefficients plots of the models were examined and the statistically non-significant terms were eliminated. Thus, the models were refined and simplified. The summary of fit plot of the refined models is presented in Fig. 20. It can be seen from Fig. 20 that all three Q2 values have increased, and now amount to 0.97, 0.97 and 0.74 for Sb recovery, T90 and Sn recovery, respectively. The refined models show improved model validity.

Fig. 20: Summary of fit plot of the refined modelling.

The outcome of the normal probability plot of the residuals after refinement (Fig. 21) reveals that the models are satisfactory, except for the deviating behaviour of experiment 2 in Sn recovery model. Subsequently, experiment 2 was critically scrutinized for any possible error, but was found to be well fitted into the other models; therefore, the experiment cannot be removed from the model. Due to the satisfactory values of R2 and Q2 for Sn recovery model, it may be assumed that experiment 2 is a weak outlier which does not influence the model decisively.

31 Fig. 21: Normal probability plots of residuals after model refinement.

In order to obtain information concerning how the input variables affect the responses, regression coefficient plots of the refined models were made and interpreted (Fig. 22). Inspection of these plots shows that reaction temperature and sulphide ion concentration have a strong effect on all the three responses [52, 53]. It seems that solid concentration has insignificant effect on all the three responses within the selected range of solid concentration.

32 Fig. 22: Regression coefficient plots for the responses Sb, Sn and T90 after model refinement with confidence intervals.

Furthermore, examination of the analysis of variance (ANOVA) of the refined models for the responses Sb recovery, Sn recovery and T90 as shown in Table 7, reveal that all the regression models are statistically significant with a 95% confidence level in the range studied. For all the response variables, F0 values are greater than Fcritical values, and P-values are smaller than 0.05. According to the results displayed in Table 7, one can infer that the model error of the original model is of the same magnitude as the replicate error for all the responses, because their P-values are greater than 0.05 and F0

33 Table 7: ANOVA for the quadratic models predicted for each response variable

Response Source DF SS MS F0 Fcritical P-value SD variable (variance) Į = 0.05 Sb Total corrected 16 2.7026 0.1689 0.4109 recovery Regression 5 2.6619 0.5324 143.58 3.20 0.000 0.7296 Residual 11 0.0408 0.0037 0.0609 Lack of fit 9 0.0347 0.0039 1.28 19.37 0.514 0.0621 (model error) Pure error 2 0.0060 0.0030 0.0550 (replicate error) Sn Total corrected 16 0.0005 3.37e-5 0.0058 recovery Regression 5 0.0005 1.01e-4 30.61 3.20 0.000 0.0100 Residual 11 3.61e-53.29e-6 0.0018 Lack of fit 9 3.39e-5 3.77e-6 3.36 19.37 0.251 0.0019 (model error) Pure error 2 2.24e-6 1.12e-6 0.0010 (replicate error) T90 Total corrected 16 0.1823 0.0114 0.1068 Regression 5 0.1798 0.0360 155.78 3.20 0.000 0.1896 Residual 11 0.0025 0.0002 0.0152 Lack of fit 9 0.0023 0.0003 2.16 19.37 0.356 0.0160 (model error) Pure error 2 0.0002 0.0001 0.0109 (replicate error) DF-degree of freedom, SS-sum of squares, MS-mean square, SD-standard deviation

The illustration from Fig. 23 shows that the observed responses correlated very well with the predicted values. Hence, the models can be considered adequate for the predictions and for the optimisation process. The regression models describing the correlation between the transformed response variables and the parameters investigated were as written in Eqs. 19-21.

2 Sb = -1.002+ 0.168X1 + 0.485X2 – 0.079X 2 (19) 2 Sn = 0.030 – 0.003X1 – 0.004X2 – 0.005X 2 – 0.003X1 X2 (20) 2 T90 = 0.422 + 0.044X1 + 0.123X2 – 0.039X 2 – 0.020X1X2 (21)

In the regression models shown in Eqs. 19-21, all the experimental parameters are in coded values, where X1 is the reaction temperature, X2 is sulphide ion concentration, X3 is solid 2 concentration, and (X2) and X1X2 are the square and interaction of the main factors. The presence of significant square and interaction terms in the regression equations confirms the existence of quadratic behaviour and non-linear combining effects of the variables.

34 Fig. 23: Relationship between observed and predicted values for the response variables Sb, Sn and T90.

4.5.2 Model validation and optimisation

The validity of the model with regards to the response variables Sb, Sn and T90 was investigated by performing a separate experiment at the conditions shown in Table 8 for 24 hours. The results as shown in Table 8 indicate a close agreement with the values predicted by the model. Consequently, the model from a response surface methodology is considered to be accurate and reliable, for predicting the leaching of antimony and tin from the complex copper concentrate, and also good at predicting the time needed to recover 90% of the antimony from the concentrate.

Table 8: Validation of model

Temp. Sulphide Solid %Sb %Sn T90, h X1, °C conc. conc. Predicted Observed Predicted Observed Predicted Observed X2, g/L X3, g/L 80 45 290 36-58 52 30-37 26 19-30 20

35 Furthermore, optimisation of the factors affecting the leaching process can be carried out depending on what is expected from the process. If high recoveries of Sb and Sn are desirable in as short as possible leaching time and at a moderate cost, one might opt for leaching at maximum temperature and sulphide ion concentration as displayed in Figures 24 and 25.

Fig. 24: Contour plots of Sb recovery showing the interaction between sulphide concentration (g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

Fig. 25: Contour plots of Sn recovery showing the interaction between sulphide concentration (g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

36 If for practical and economic reasons, low production cost, less corrosion and electrowinning problems are desirable by the smelter. A lower temperature and less sulphide concentration can be used to set the factors in such a level to obtain high recovery of Sb (min 90%) by extending the leaching time as presented in Fig. 26.

Fig. 26: Contour plots for the response T90 showing the interaction between sulphide concentration (g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

37 38 5. CONCLUSIONS

Mineralogical investigation of the complex sulphide concentrate used in this investigation shows that the concentrate is rich in tetrahedrite. Antimony, arsenic and silver form a solid solution in the tetrahedrite crystal structure and other major minerals identified were sphalerite, galena, chalcopyrite and pyrite. Alkaline sulphide leaching of the concentrate shows that antimony and arsenic extraction from a tetrahedrite-rich concentrate strongly depends on the concentration of sulphide and hydroxide ions, reaction temperature, particle size and the leaching time. The kinetic study indicates that the rate of tetrahedrite leaching by the lixiviant under the selected conditions is chemically controlled through the particle surface reaction. The activation energies were found to be 81 kJ/mol and 75 kJ/mol, respectively, for antimony and arsenic dissolution from tetrahedrite, which are consistent with the values of activation energies reported for the chemically controlled reaction process. Due to the relative slow leaching kinetics of tetrahedrite, fine mineral particles are required in order to significantly improve antimony and arsenic extraction. Analysis of the leach residue indicates that the lixiviant is strongly selective to remove the impurity metals As and Sb. Tetrahedrite in the concentrate was found to be converted into a new species having the average chemical formula Cu1.64S which could be a mixture of chalcocite, covelite or spionkopite. The study shows that alkaline sulphide leaching can be a suitable hydrometallurgical upgrading process from technological and environmental point of view, to selectively and effectively reduce the antimony and arsenic content from the complex sulphide concentrate, and therefore, upgrading it for further pyrometallurgical treatment.

Modelling and optimisation of alkaline sulphide leaching of a complex copper concentrate containing 1.69% Sb and 0.14% Sn was conducted using response surface methodology- central composite face-centred design (RSM-CCF). Further, RSM was established to be effectual and reliable in formulating models for fitting the experimental data and predicting leaching behaviour of the copper concentrate in alkaline sulphide media. The experimental variables studied were reaction temperature, sulphide ion concentration and the concentration of solid leached. The response parameters include antimony recovery, tin recovery and the estimated time required to achieve 90% Sb dissolution. A strong model with no lack of fit was developed and the validity of the model was confirmed experimentally. The result shows that the model is reliable and accurate for predicting the leaching process.

39 40 FUTURE WORK

The following investigations are planned for the future work:

x Fundamental studies on antimony electrowinning from a modelled alkaline sulphide electrolyte with the factors affecting it

x Study the effect of arsenic, tin and mercury on the quality of the electrodeposited antimony

x Investigating the morphological and topographical structures of the electrodeposited antimony in order to ascertain its quality

41 42 REFERENCES

1. The world copper factbook. 2009; Available from: http://www.icsg.org. 2. Awe, S.A., Selective removal of impurity elements from Maurliden Västra complex sulphides flotation concentrate, in Process Metallurgy. 2008, Luleå University of Technology: Luleå. p. 42. 3. Habashi, F., Metals from ores : an introduction to extractive metallurgy. 2003, Sainte Foy, Québec: Métallurgie Extractive Québec. 4. Moskalyk, R.R. and A.M. Alfantazi, Review of copper pyrometallurgical practice: today and tomorrow. Minerals Engineering, 2003. 16(10): p. 893-919. 5. Riveros, P.A. and J.E. Dutrizac, The leaching of tennantite, tetrahedrite and enargite in acidic sulphate and chloride media. Canadian Metallurgical Quarterly, 2008. 47(3): p. 235-244. 6. Lattanzi, P., et al., Enargite oxidation: A review. Earth-Science Reviews, 2008. 86(1- 4): p. 62-88. 7. Filippou, D., P. St-Germain, and T. Grammatikopoulos, Recovery of metal values from copper-arsenic minerals and other related resources. Mineral Processing and Extractive Metallurgy Review, 2007. 28(4): p. 247-298. 8. Curreli, L., et al., Arsenic leaching from a gold bearing enargite flotation concentrate. Hydrometallurgy, 2009. 96(3): p. 258-263. 9. Gabb, P.J., et al., Hydrometallurgical processing of impurity streams generated during the pyrometallurgy of copper, in United States Patent and Trademark Office Granted Patent. 1995. 10. Grez, P., et al., Characterization of natural tetrahedrite/aqueous solution system by electrochemical techniques. Journal of Electroanalytical Chemistry, 2009. 628(1-2): p. 134-143. 11. Wang, S., Impurity control and removal in copper tankhouse operations. JOM Journal of the Minerals, Metals and Materials Society, 2004. 56(7): p. 34-37. 12. Tongamp, W., Y. Takasaki, and A. Shibayama, Arsenic removal from copper ores and concentrates through alkaline leaching in NaHS media. Hydrometallurgy, 2009. 98(3- 4): p. 213-218. 13. Baláž, P., et al., Leaching of antimony and mercury from mechanically activated tetrahedrite Cu12Sb4S13. Hydrometallurgy, 1998. 47(2-3): p. 297-307. 14. Ukasik, M. and T. Havlik, Effect of selected parameters on tetrahedrite leaching by ozone. Hydrometallurgy, 2005. 77(1-2): p. 139-145. 15. King, R.J., The tetrahedrite group. Geology Today, 2001. 17(2): p. 77-80. 16. Neiva Correia, M.J., J.R. Carvalho, and A.J. Monhemius, Study of the autoclave leaching of a tetrahedrite concentrate. Minerals Engineering, 1993. 6(11): p. 1117- 1125. 17. Butterman, W.C. and J.F. Carlin, Mineral commodity profiles: Antimony. 2004, U.S. Geological Survey. p. 35. 18. Critical raw materials for the EU, in Report for the Ad-hoc Working Group on defining critical raw materials. June 2010, European Commission-Enterprise and Industry. 19. Mihajlovic, I., et al., A potential method for arsenic removal from copper concentrates. Minerals Engineering, 2007. 20(1): p. 26-33. 20. Anderson, C., Hydrometallurgically treating antimony-bearing industrial wastes. JOM Journal of the Minerals, Metals and Materials Society, 2001. 53(1): p. 18-20. 21. Baláž, P., R. Kammel, and C. Villachica, As and Sb leaching from polymetallic sulfide concentrates. Metall(Berlin, West), 2001. 55(4): p. 196-200.

43 22. Havlik, T., et al., Extraction of copper and antimony from tetrahedrite by acid oxidative leaching. Metall, 1999. 53(7): p. 390-394. 23. Correia, M.J., J. Carvalho, and J. Monhemius, The effect of tetrahedrite composition on its leaching behaviour in FeCl3-NaCl-HCl solutions. Minerals Engineering, 2001. 14(2): p. 185-195. 24. Correia, M.J.N., J.R. Carvalho, and A.J. Monhemius, The leaching of tetrahedrite in ferric chloride solutions. Hydrometallurgy, 2000. 57(2): p. 167-179. 25. Guy, S., et al., Cupric chloride leaching of a complex copper/zinc/lead ore. Hydrometallurgy, 1983. 10(2): p. 243-255. 26. Padilla, R., D. Giron, and M.C. Ruiz, Leaching of enargite in H2SO4-NaCl-O2 media. Hydrometallurgy, 2005. 80(4): p. 272-279. 27. Havlik, T. and R. Kammel, Procedure for selective copper recovery from tetrahedrite. Metall, 2000. 54(1-2): p. 26-29. 28. Baláž, P. and M. Achimoviþová, Selective leaching of antimony and arsenic from mechanically activated tetrahedrite, jamesonite and enargite. International Journal of Mineral Processing, 2006. 81(1): p. 44-50. 29. Baláž, P. and M. Achimoviþová, Mechano-chemical leaching in hydrometallurgy of complex sulphides. Hydrometallurgy, 2006. 84(1-2): p. 60-68. 30. Frohlich, L. and V. Miklos, Leaching of As, Sb and Hg from tetrahedrite concentrate in the Na2S medium at increased temperatures. Metalurgija, 2001. 40(4): p. 213-218. 31. Baláž, P., Extractive Metallurgy of Activated Minerals. 2000, Amsterdam: Elsevier. 32. Achimoviþová, M. and P. Baláž, Kinetics of the leaching of mechanically activated berthierite, boulangerite and franckeite. Physics and Chemistry of Minerals, 2008. 35(2): p. 95-101. 33. Anderson, C.G. and L.G. Twidwell, Hydrometallurgical processing of gold-bearing copper enargite concentrates. Canadian Metallurgical Quarterly, 2008. 47(3): p. 337- 346. 34. Baláž, P., et al., Influence of mechanical activation on the alkaline leaching of enargite concentrate. Hydrometallurgy, 2000. 54(2): p. 205-216. 35. Baláž, P., et al., Non-oxidative leaching of mechanically activated stibnite. Hydrometallurgy, 1992. 31(3): p. 201-212. 36. Delfini, M., et al., Arsenic leaching by Na2S to decontaminate tailings coming from colemanite processing. Minerals Engineering, 2003. 16(1): p. 45-50. 37. Harrison, D., Process for recovery of antimony and metal values from antimony - and metal value-bearing materials, in Patent Cooperation Treaty Application. 2007. 38. Yang, T., et al., Sodium sulfide leaching of low-grade jamesonite concentrate in production of sodium pyroantimoniate. Journal of Central South University of Technology, 2005. 12(3): p. 290-294. 39. Ackerman, J.B., et al. Hydrometallurgy at the Sunshine mine metallurgical complex. in Hydrometallurgy Fundamentals, Technology and Innovations. 1993. Salt Lake City, Utah. 40. Edwards, C.R. Engineering the Equity concentrate leach process. in Complex Sulfides - processing of ores, concentrates and by-products. 1985. San Diego, Califonia. 41. Dayton, S., Equity silver on line with leach plant: How Sb and As are purged from a high-silver copper concentrate before smelting. Engineering and Mining Journal, 1982. 183(1): p. 78-83. 42. Nadkarni, R.M. and C.L. Kusik. Hydrometallurgical removal of arsenic from copper concentrates. in Arsenic Metallurgy Fundamentals and Applications. 1988. 43. Baláž, P. and E. Dutková, Fine milling in applied mechanochemistry. Minerals Engineering, 2009. 22(7-8): p. 681-694.

44 44. Baláž, P., et al., Application of attrition grinding in alkaline leaching of tetrahedrite. Minerals Engineering, 1995. 8(11): p. 1299-1308. 45. Coltrinari, E.L., Sodium sulfide leach process, in United State Patent 4,051,220. 1977, Equity Mining Corporation, Canada. 46. Nordwick, S.M. and C.G. Anderson. Advances in antimony electrowinning at the Sunshine mine. in Hydrometallurgy Fundamentals, Technology and Innovations. 1993. Salt Lake City, Utah. 47. Goodall, W.R., P.J. Scales, and A.R. Butcher, The use of QEMSCAN and diagnostic leaching in the characterisation of visible gold in complex ores. Minerals Engineering, 2005. 18(8): p. 877-886. 48. Peters, E., Leaching Kinetics, in Hydrometallurgy: Theory and Practice course notes, The University of British Columbia: Vancouver, B.C. p. 25. 49. Habashi, F., Principles of Extractive Metallurgy. Second ed. Vol. 1. 1980, New York: Gordon & Breach. 50. Ray, H.S., Kinetics of Metallurgical Reactions. 1993, New Delhi: Oxford and IBH publishing. 51. Eriksson, L., et al., Design of experiments-principles and applications. Third ed. 2008, Umeå: UMETRICS AB. 459. 52. Awe, S.A., C. Samuelsson, and Å. Sandström, Dissolution kinetics of tetrahedrite mineral in alkaline sulphide media. Hydrometallurgy, 2010. 103(1-4): p. 167-172. 53. Awe, S.A. and A. Sandström, Selective leaching of arsenic and antimony from a tetrahedrite rich complex sulphide concentrate using alkaline sulphide solution. Minerals Engineering, 2010. 23(15): p. 1227-1236.

45

PAPER I

Dissolution kinetics of tetrahedrite mineral in alkaline sulphide media

Samuel A. Awe, Caisa Samuelsson and Åke Sandström

Hydrometallurgy 103 (2010) 167-172

Hydrometallurgy 103 (2010) 167–172

Contents lists available at ScienceDirect

Hydrometallurgy

journal homepage: www.elsevier.com/locate/hydromet

Dissolution kinetics of tetrahedrite mineral in alkaline sulphide media

Samuel A. Awe, Caisa Samuelsson, Åke Sandström ⁎

Division of Extractive Metallurgy, Department of Chemical Engineering and Geosciences, Luleå University of Technology, SE-971 87 Luleå, Sweden article info abstract

Article history: The present study investigates the dissolution kinetics of tetrahedrite in aqueous alkaline sodium sulphide Received 16 February 2010 solutions. Effect of reaction temperature, mineral particle size, sodium sulphide and sodium hydroxide Received in revised form 16 March 2010 concentrations on antimony and arsenic dissolution rate from tetrahedrite were studied. It was found that Accepted 17 March 2010 the rate of reaction increases with increase in reaction temperature, sodium sulphide concentration, and Available online 30 March 2010 sodium hydroxide concentration and with decrease in mineral particle size. The kinetic study indicates that the rate of tetrahedrite leaching in the lixiviant under selected conditions is chemically controlled through Keywords: Dissolution kinetics the particle surface reaction. The estimated activation energies were found to be 81 kJ/mol and 75 kJ/mol, Alkaline sodium sulphide respectively, for antimony and arsenic dissolution from tetrahedrite, which are in agreement with the values Tetrahedrite of activation energies reported for the chemically controlled reaction process. Leaching © 2010 Elsevier B.V. All rights reserved. Lixiviant

1. Introduction (FeSbO4) and the overall leaching reaction proceeds slowly with complicated kinetics (Baláž, 2000). Havlik et al. (1999) conducted a

Tetrahedrite (Cu12Sb4S13) is a complex sulphide of copper and leaching investigation with native tetrahedrite concentrate in acidic antimony which can be an interesting and important resource of ferric chloride in the temperature range of 40–90 °C. The result reveals copper (40–46%), antimony (27–29%) (Baláz et al., 1998) and/or other that the leaching rate shows a slow behaviour exhibiting apparent non-ferrous metals as is evident from its general chemical formula parabolic kinetics. It was observed that increased temperatures improve

(Cu,Ag)10(Cu,Zn,Fe,Cd,Hg)2(Sb,Bi,As)4S13 (Ukasik and Havlik, 2005; the metal extraction and the apparent activation energy was estimated Baláž, 2000). Tetrahedrite belongs to a family of minerals usually to be approximately 38 kJ/mol, which indicates that the process occurs called “fahlore” which is often described as the “sulphosalts” which in the mixed regime of both the diffusion and chemical reaction control. indicates a certain type of un-oxidized sulphur mineral that is Also, it is reported (Correia et al., 2000) that the leaching of tetrahedrite structurally distinct from a sulphide (King, 2001). The tetrahedrite with ferric chloride, sodium chloride and hydrochloric acid solutions group is notable for the variety of elements which are stable in its involves the breakdown of the tetrahedrite crystal structure with the structure. The formula may be written as A12B4C13, where A=Ag, Cu, simultaneous liberation of all of its components without the formation Fe, Hg, and Zn; B=As, Sb, and Bi; C=S, Se, and Te. Naturally, it is very of sulphate in the leaching reaction. Antimony precipitates were difficult to find pure tetrahedrite because its original constituents are observed at chloride concentrations less than 3 M and the leaching always partially substituted by other elements. Thus, arsenic can rate was controlled by a surface reaction. They found that the apparent substitute for antimony leading to tennantite (Cu12As4S13), while rate constant is proportional to the inverse of the mean particle size. copper can be substituted by silver, zinc, iron, mercury, cadmium, lead The heterogeneous fluid–solid reaction systems have a number of etc, and some sulphur may be replaced by Se and Te (King, 2001; applications in chemical and hydrometallurgical processes. Many Filippou et al., 2007; Neiva Correia et al., 1993). Apart from the toxic studies have been conducted on these types of reactions and a lot of elements (As, Hg) contained in tetrahedrite mineral, the substantial mathematical models have been proposed and developed (Lapidus amount of copper, silver and gold (Havlik et al., 1999) usually present and de Lourdes Mosqueira, 1988; Ray, 1993). The shrinking particle in the mineral makes its mining to be more economically attractive. model and the shrinking core model are the most commonly used. It has been documented that in acidic oxidative leaching of The shrinking core model considers that the leaching process is tetrahedrite mineral, e.g. in acidified ferric chloride solutions, copper controlled either by the diffusion of reactant through the solution and iron enter into solution, while antimony is partially precipitated as a boundary layer, or through a solid product layer while the shrinking compound with the composition similar to the mineral tripuhyite particle model considers that the leaching process is controlled by the rate of the surface chemical reaction (Aydogan et al., 2005). Besides, during the leaching process, dissolution rate decreases with ⁎ Corresponding author. Tel.: +46 920 491290; fax: +46 920 491199. time and it is directly dependent on the activation energy. Habashi E-mail address: [email protected] (Å. Sandström). (1980) stated that if the rate of reaction in the bulk of the solution was

0304-386X/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.hydromet.2010.03.014 168 S.A. Awe et al. / Hydrometallurgy 103 (2010) 167–172 fast, the solution would be governed by the rate of diffusion of the ions from the surface of the solid through the boundary layer. On the other hand, if the rate of reaction is slow, it will control the overall process and the process will be chemically controlled, thus, diffusion through the boundary layer will not play any critical role (Aydogan et al., 2005). It is reported that a diffusion-controlled heterogeneous process was characterised by being slightly dependent on temperature, while the chemically controlled process was strongly dependent on temperature which is attributed to linear dependency of diffusion coefficients and exponential dependency of chemical velocity constants on temperature. However, the activation energy of a diffusion-controlled process is characterised as being 4–12 kJ/mol, while it is usually greater than 42 kJ/ mol for a chemically controlled process (Aydogan et al., 2005; Habashi, 1980). Recently, Riveros and Dutrizac (2008) performed a series of leaching Fig. 1. XRD pattern of the mineral sample. tests on tetrahedrite (Cu12Sb4S13) in acidic media using Fe2(SO4)3, FeCl3 or O2 as oxidizing agents. At temperatures b100 °C , the dissolution Departamento Lima, Peru, containing mainly sharp tetrahedrite kinetics of disks of synthetic iron-bearing tetrahedrite in Fe2(SO4)3– crystals accompanied by an excellent brilliance and striation growth H2SO4 media were linear; the leaching rates were slow but increased of sphalerite and partially covered by quarts. The tetrahedrite mineral significantly with increasing temperature with an apparent activation was crushed, ground and sieved into the following size fractions: energy of 120 kJ/mol. The dissolution of sized particles of natural Ag- −150+106, −106+75, −75+53, −53+38 and −38+20 μm. bearing tetrahedrite in FeCl3–HCl solutions proceeds according to the The particle size fractions of −106+75, −75+53 and −53+38 μm shrinking core model and both Cu and Ag dissolve at about the same rate were used in this study. provided that the total chloride concentration of the solution is sufficiently high to solubilize the AgCl reaction product. The leaching rates were slow and the apparent activation energy is 116 kJ/mol. The 2.2. Analytical and instrumentation techniques addition of sulphate ions to the FeCl3–HCl system reduces the tetrahedrite leaching rate to values similar to those realized in the Fe2 X-ray powder diffraction (XRD) was used to characterise the (SO4)3–H2SO4 system. tetrahedrite mineral using a Siemens D5000 automatic diffractometer Similarly, Anderson and Krys (1993) leached an argentiferous equipped with a continuous scanning device. CuKα radiation of 40 kV tetrahedrite concentrate from the Sunshine mine in sodium sulphide and 30 mA with a sample rotation of 30 rpm was used. Diffraction lixiviant produced by dissolving elemental sulphur in sodium hydroxide patterns (Fig. 1) for the sample were measured in the 2-theta range at temperatures close to the solution boiling point for 4–12 h. The from 5 °C to 90 °C and crystalline phases were identified using the chosen variables investigated were agitation rate, concentrate regrind, Joint Committee for Powder Diffraction Standards (JCPDS) file of the temperature, percent solids, caustic concentration, carbonate concen- instrument. The tetrahedrite mineral and the leach solutions were tration, and individual sulphur species concentration. Their findings chemically analysed by Inductively Coupled Plasma-Atomic Emission show that the process is sensitive to the amount of regrind performed on Spectrometry (ICP-AES)/Sector Field Mass Spectrometry (ICP-SFMS). the concentrate, reaction temperature and sulphide concentration. They The chemical composition of the fractions used in the study is shown also concluded that polysulphide, increased levels of thiosulphate and in Table 1. hydroxide, and increased percent solids enhance the process but the reaction is independent of sulphite, sulphate or carbonate concentration. 2.3. Experimental procedure In addition, alkaline sulphide leaching of copper–antimony–arsenic sulphide minerals has been extensively studied and documented in the The leaching experiments were conducted in batch-wise mode in a literatures (Baláz et al., 1998, 1992; Awe, 2008; Baláz and Achimovicová, four-necked round bottomed glass reactor which was stirred 2006a,b; Balaz et al., 2001; Delfini et al., 2003; Frohlich and Miklos, mechanically with a paddle stirrer at 600 rpm and heated with an 2001; Yang et al., 2005; Tongamp et al., 2010; Ackerman et al., 1993; autoregulated device. The leaching solution (500 mL) was added to Coltrinari, 1977; Dayton, 1982) but no detailed study was found the reactor and when the desired temperature was reached, 2.5 g of regarding the dissolution kinetics of natural tetrahedrite in alkaline solid sample was added in all the leaching experiments. At sulphide solutions. However, in the present study, the dissolution appropriate time intervals, 5 mL sample of the leach solution was kinetics of tetrahedrite in aqueous alkaline sodium sulphide solution collected for the analysis of the metals dissolved in the leach liquor. All was investigated. The effect of the reaction temperature, mineral reagents used for leaching and chemical analysis were of analytical particle size, and the concentration of sodium sulphide and sodium grade and used without purification. The experimental leaching hydroxide on the dissolution rate have been evaluated. The dissolution conditions used were: sodium sulphide (Na2S·3H2O) concentration kinetics of tetrahedrite was examined according to the heterogeneous (100 g/L, 150 g/L and 250 g/L); sodium hydroxide (NaOH) concen- reaction models. The activation energies of antimony and arsenic tration (30 g/L and 60 g/L); leaching temperature (84 °C, 91 °C, 98 °C dissolution from tetrahedrite by alkaline sulphide solution were and 105 °C), mineral particle size (−53+38 μm, −75+53 μm and determined from the experimental data. −106+75 μm) and reaction time of 360 min in all cases.

2. Material and methods Table 1 Chemical composition (%). 2.1. Material Size fraction Cu Fe Zn Sb As S The tetrahedrite used in this investigation was prepared from the −106+75 μm 15.6 11.1 16.1 5.8 1.9 28.0 tetrahedrite crystals bought from Gregory, Bottley & Lloyd Company, −75+53 μm 15.2 14.2 16.8 5.4 1.9 31.5 −53+38 μm 15.7 13.3 16.9 5.7 1.8 26.5 United Kingdom, which originated from Casapalca, Huarochiri, S.A. Awe et al. / Hydrometallurgy 103 (2010) 167–172 169

− − 1/3 Fig. 2. Effect of leaching temperature on tetrahedrite dissolution rate. Fig. 4. A plot of 1 (1 X) vs. time at various reaction temperatures for antimony and arsenic dissolution from tetrahedrite in alkaline sulphide solution.

3. Results and discussion

r0 is the initial radius of the solid particle, t is the reaction time, D is the 3.1. Effect of temperature diffusion coefficient in the porous product layer, kd and kr are the rate constants which are calculated from Eqs. (1) and (2), respectively. An The effect of reaction temperature on the rate of antimony and attempt was made to fit the obtained experimental data to the model arsenic dissolution from tetrahedrite at different reaction times are equation described in Eq. (1), and this is presented in Fig. 3 for both plotted in Fig. 2 using mineral particle size fraction of −75+53 μm, antimony and arsenic dissolution from tetrahedrite. It is observed that and concentration of sodium sulphide and sodium hydroxide of 150 g/ the data do not correlate to this model, since neither a straight line nor L and 60 g/L respectively at temperatures ranging from 84 °C to a zero point intercept was obtained in both figures. 105 °C. The amount of solid leached was kept constant at 5 g/L. The obtained results reveal that the reaction temperatures studied have a 3.2.2. Rate control by chemical reaction significant effect on the rate of antimony and arsenic dissolution from When surface chemical reaction is rate controlling, the kinetics tetrahedrite. It can be seen from Fig. 2 that the leaching rate increases may be correlated graphically using Eq. (2). rapidly with increasing time and temperature. After 360 min of 1= leaching, about 57% and 60% of antimony and arsenic were dissolved 3 k M C 1−ðÞ1−X = c S A t = k t: ð2Þ respectively at a reaction temperature of 105 °C. ρ β r S r0

3.2. Kinetic model Application of Eq. (2) to the experimental data obtained, at different temperatures, resulted in linear plots presented in Fig. 4.Itis 3.2.1. Rate control by diffusion through the product layer observed, from both plots, during the whole reaction time that the If the rate of the reaction is controlled by diffusion through the data in these figures are linear, which indicate that the rate of the product layer, the integrated rate equation is described as follows: reaction is chemically controlled through the particle surface reaction. The slopes of the straight lines from each figure were estimated which 2= 2 3 2M DC 1− X−ðÞ1−X = S A t = k t ð1Þ is equivalent to the rate constants, kr. ρ β 2 d 3 S r0 3.2.3. Calculation of the activation energy where X is the fraction reacted, kc is the kinetic constant, MS is the A process may be regarded either as diffusion rate controlled molecular weight of the solid, CA is the concentration of the dissolved through the porous layer when the activation energy of the process is lixiviant A in the bulk of the solution, ρS density of tetrahedrite, β is from 4 to 12 kJ/mol or chemical rate controlled through the particle the stoichiometric coefficient of the reagent in the leaching reaction, surface reaction when its activation energy is greater than 42 kJ/mol

Fig. 3. A plot of 1−2/3X−(1−X)2/3 vs. time at various reaction temperatures for tetrahedrite dissolution in alkaline sulphide solution. Fig. 5. Arrhenius plot for tetrahedrite leaching in alkaline sulphide solution. 170 S.A. Awe et al. / Hydrometallurgy 103 (2010) 167–172

Fig. 6. Effect of mineral particle size on tetrahedrite dissolution rate. Fig. 8. Dependence of rate constant on the reciprocal of mineral particle radius. (Habashi, 1980). The temperature dependence of the chemical reactions can be given by the Arrhenius equation (Ray, 1993; Habashi, in Fig. 7. The rate constant values increased with the decrease in the 1980) as follows: particle size, which may be attributed to the increase in surface area with the decrease in particle size. The values of the rate constants ðÞ− = ðÞ kr = A exp Ea RT 3 were plotted against the reciprocal of the mineral particle radii yielding a linear relationship with correlation coefficients of 0.99 where A is the frequency factor and Ea is the apparent activation and 0.98 (Fig. 8) respectively for antimony and arsenic dissolution energy. The Arrhenius plot shown in Fig. 5 was made by plotting the from the tetrahedrite. In addition, the linear dependence of the rate natural logarithm of the rate constant (lnkr) against the reciprocal of constant on the inverse particle radius, Fig. 8, is further evidence in the temperature (1/T). Hence, the activation energies of antimony and support of the surface reaction shrinking particle model proposed arsenic were calculated from the slopes of the Arrhenius plots for this process. presented in Fig. 5 as 81 kJ/mol and 75 kJ/mol respectively, which is in agreement with the reported activation energy for a chemically 3.4. Effect of sodium sulphide concentration controlled process (Habashi, 1980). Since the calculated activation energies are high, it can be concluded that the dissolution rate of Antimony and arsenic form complex compounds with sulphide ions, antimony and arsenic from tetrahedrite by alkaline sulphide solution therefore, the influence of sodium sulphide concentration on tetrahe- is very sensitive to temperature. drite leaching is an important parameter to investigate. When tetrahedrite is leached in alkaline sulphide solution, a variety of reaction 3.3. Effect of particle size products may result. This includes chalcocite (Cu2S) or covelite (CuS), while arsenic and antimony are dissolved as thioanions: thioarsenate The effect of particle size on the leaching of antimony and arsenic − − − (AsS3 ), thioantimonite (SbS3 ), thioantimonate (SbS3 ) etc depend- from tetrahedrite mineral was studied using three particle sizes 4 3 4 ing on the feed composition, sulphide ion concentration and the (−53+38, −75+53 and −106+75 μm) at 105 °C in leach reaction conditions (Filippou et al., 2007). The dissolution reaction of solutions containing 150 g/L Na S·3H O and 60 g/L NaOH. 2.5 g of 2 2 tetrahedrite in alkaline sulphide solutions is expressed as follows: tetrahedrite was leached in 500 mL of the solution and maintained throughout the tests. It was observed from Fig. 6 that the rate of Cu12Sb4S13ðsÞ þ 2Na2SðaqÞ→5Cu2SðsÞ þ 2CuSðsÞ þ 4NaSbS2ðaqÞ ð4Þ antimony and arsenic dissolution increases with increase in time and decreases with increase in particle size. The smaller the particle NaSbS þ Na S →Na SbS : ð5Þ size, the faster was tetrahedrite leaching by the lixiviant. Lineari- 2ðaqÞ 2 ðaqÞ 3 3ðaqÞ zation of the kinetic curves for antimony and arsenic dissolution from tetrahedrite was performed by means of Eq. (2) and presented Leaching tests were conducted to investigate the effect of varying the initial concentration of sodium sulphide in the range of 100–250 g/L at

Fig. 7. A plot of 1−(1−X)1/3 vs. time at various mineral particle sizes for antimony and arsenic dissolution from tetrahedrite. Fig. 9. Effect of sodium sulphide concentration on tetrahedrite dissolution rate. S.A. Awe et al. / Hydrometallurgy 103 (2010) 167–172 171

Fig. 12. Effect of sodium hydroxide concentration on tetrahedrite dissolution rate. Fig. 10. A plot of 1−(1−X)1/3 vs. time at various sodium sulphide concentrations for antimony and arsenic dissolution from tetrahedrite.

3.5. Effect of sodium hydroxide concentration

105 °C in solutions containing 60 g/L sodium hydroxide. During the Fig. 12 presents the influence of sodium hydroxide concentration tests, the amount of solid leached was kept constant at 5 g/L taking from on the leaching of tetrahedrite mineral. The tests were carried out − 75+53 μm mineral particle size fraction. It is apparent that the rate of under the following leaching conditions: 150 g/L Na2S·3H2O; particle dissolving antimony and arsenic from tetrahedrite (Fig. 9)bythe size of −75+53 μm; 2.5 g of tetrahedrite leached in 500 mL of the lixiviant increases steadily with increasing sodium sulphide concentra- lixiviant; reaction temperature and time of 105 °C and 360 min tion, and the dissolution rate at Na2S concentration of 250 g/L is much respectively. It is evident from this figure that tetrahedrite dissolution higher than that at 150 and 100 g/L Na2S·3H2O. The results show that in the lixiviant is strongly dependent on the concentration of NaOH. tetrahedrite leaching depends strongly on the concentration of sodium This can be observed from the plot that the percentage of antimony sulphide in the leach liquor. When the data from the experiment was and arsenic (Fig. 12) reported into the leach liquor after 360 min at a substituted into Eq. (2), Fig. 10 was constructed. By plotting the natural concentration of 60 g/L NaOH is greater than that obtained at sodium logarithm of the slope of each line (lnkr)inFig. 10 against the natural hydroxide concentration of 30 g/L. It is also found that increase in logarithm of initial concentration of sodium sulphide (ln[Na2S]), Fig. 11 NaOH concentration accelerates the rate at which antimony and can be drawn. It is indicated from Fig. 11 that lnkr is a linear function of ln arsenic solubilise in the leach solution. The kinetic data obtained from [Na2S]. The kinetic equations about the effect of the initial concentration the tests are plotted in Fig. 13. This plot shows a linear relationship of Na2S on the leaching rate of antimony and arsenic can be written as and therefore supports that tetrahedrite leaching by the lixiviant is Eqs. (6) and (7), respectively. controlled by the shrinking particle model for reaction-controlled process. lnkr ¼ 2:2ln½Na2S−7:77 ð6Þ 4. Conclusions lnkr ¼ 2:3ln½Na2S−7:76 ð7Þ The dissolution kinetics of tetrahedrite in aqueous alkaline sulphide solution was investigated. It was found that the rate of Therefore, the apparent reaction order for dissolving antimony and reaction increases with increase in reaction temperature, sodium arsenic from tetrahedrite by the lixiviant was estimated to be sulphide concentration, and sodium hydroxide concentration and approximately equal to 2, which implies that both elements reacted with decrease in mineral particle size. The kinetic study indicates that similarly with the lixiviant as should be expected since they were the rate of tetrahedrite leaching in the lixiviant under selected present in the same mineral. conditions is chemically controlled through the particle surface reaction. The activation energies were found to be 81 kJ/mol and

Fig. 13. A plot of 1−(1−X)1/3 vs. time at various sodium hydroxide concentrations for

Fig. 11. Determination of the reaction order with respect to [Na2S]. antimony and arsenic dissolution from tetrahedrite. 172 S.A. Awe et al. / Hydrometallurgy 103 (2010) 167–172

75 kJ/mol, respectively, for antimony and arsenic dissolution from Baláz, P., Achimovicová, M., Ficeriová, J., Kammel, R., Sepelák, V., 1998. Leaching of antimony and mercury from mechanically activated tetrahedrite Cu12Sb4S13. tetrahedrite, which are consistent with the values of activation Hydrometallurgy 47, 297–307. energies reported for the chemically controlled reaction process. Coltrinari, E.L., Sodium sulfide leach process, United State Patent No. 4, 051, 220, Equity Mining Corporation, Canada, 1977. Correia, M.J.N., Carvalho, J.R., Monhemius, A.J., 2000. The leaching of tetrahedrite in Acknowledgements ferric chloride solutions. Hydrometallurgy 57, 167–179. Dayton, S., 1982. Equity silver on line with leach plant: how Sb and As are purged from a The authors would like to thankfully acknowledge the financial high-silver copper concentrate before smelting. Engineering and Mining Journal 183, 78–83. support from VINNOVA, Boliden Mineral AB, LKAB and the Adolf H. Delfini, M., Ferrini, M., Manni, A., Massacci, P., Piga, L., 2003. Arsenic leaching by Na2Sto Lundin Charitable Foundation. decontaminate tailings coming from colemanite processing. Minerals Engineering 16, 45–50. Filippou, D., St-Germain, P., Grammatikopoulos, T., 2007. Recovery of metal values from References copper–arsenic minerals and other related resources. Mineral Processing and Extractive Metallurgy Review 28, 247–298. Ackerman, J.B., Anderson, C.G., Nordwick, S.M., Krys, L.E., 1993. Hydrometallurgy at the Frohlich, L., Miklos, V., 2001. Leaching of As, Sb and Hg from tetrahedrite concentrate in

Sunshine mine metallurgical complex. In: Hiskey, J.B., Warren, G.W. (Eds.), the Na2S medium at increased temperatures. Metalurgija 40, 213–218. Hydrometallurgy Fundamentals. Technology and Innovations, Salt Lake City, Habashi, F., 1980. Principles of Extractive Metallurgy, Second ed. Gordon & Breach, New Utah, pp. 477–498. York. Anderson, C.G., Krys, L.E., 1993. Leaching of antimony from a refractory precious metals Havlik, T., Ivanova, Z., Dvorscikova, J., Kammel, R., 1999. Extraction of copper and concentrate. In: Hiskey, J.B., Warren, G.W. (Eds.), Hydrometallurgy Fundamentals. antimony from tetrahedrite by acid oxidative leaching. Metall 53, 390–394. Technology and Innovations, Salt Lake City, Utah, pp. 341–363. King, R.J., 2001. The tetrahedrite group. Geology Today 17, 77–80. Awe, S.A., Selective removal of impurity elements from Maurliden Vastra complex Lapidus, G., de Lourdes Mosqueira, M., 1988. The effect of product solubility on the sulphides flotation concentrate, MSc Thesis, Luleå University of Technology, Luleå, leaching kinetics of non-porous minerals. Hydrometallurgy 20, 49–64. 2008, p. 42. Neiva Correia, M.J., Carvalho, J.R., Monhemius, A.J., 1993. Study of the autoclave leaching Aydogan, S., Aras, A., Canbazoglu, M., 2005. Dissolution kinetics of sphalerite in acidic of a tetrahedrite concentrate. Minerals Engineering 6, 1117–1125. ferric chloride leaching. Chemical Engineering Journal 114, 67–72. Ray, H.S., 1993. Kinetics of Metallurgical Reactions. Oxford and IBH publishing, New Baláž, P., 2000. Extractive Metallurgy of Activated Minerals. Elsevier, Amsterdam. Delhi. Baláz, P., Achimovicová, M., 2006a. Mechano-chemical leaching in hydrometallurgy of Riveros, P.A., Dutrizac, J.E., 2008. The leaching of tennantite, tetrahedrite and enargite in complex sulphides. Hydrometallurgy 84, 60–68. acidic sulphate and chloride media. Canadian Metallurgical Quarterly 47, 235–244. Baláz, P., Achimovicová, M., 2006b. Selective leaching of antimony and arsenic from Tongamp, W., Takasaki, Y., Shibayama, A., 2010. Selective leaching of arsenic from mechanically activated tetrahedrite, jamesonite and enargite. International Journal enargite in NaHS–NaOH media. Hydrometallurgy 101, 64–68. of Mineral Processing 81, 44–50. Ukasik, M., Havlik, T., 2005. Effect of selected parameters on tetrahedrite leaching by Balaz, P., Kammel, R., Villachica, C., 2001. As and Sb leaching from polymetallic sulfide ozone. Hydrometallurgy 77, 139–145. concentrates. Metall(Berlin, West) 55, 196–200. Yang, T., Jiang, M., Lai, Q., Chen, J., 2005. Sodium sulfide leaching of low-grade Baláz, P., Briancin, J., Sepelák, V., Havlík, T., Skrobian, M., 1992. Non-oxidative leaching jamesonite concentrate in production of sodium pyroantimoniate. Journal of of mechanically activated stibnite. Hydrometallurgy 31, 201–212. Central South University of Technology 12, 290–294. PAPER II

Selective leaching of arsenic and antimony from a tetrahedrite rich complex sulphide concentrate using alkaline sulphide solution

Samuel A. Awe and Åke Sandström

Minerals Engineering 23 (2010) 1227-1236

Minerals Engineering 23 (2010) 1227–1236

Contents lists available at ScienceDirect

Minerals Engineering

journal homepage: www.elsevier.com/locate/mineng

Selective leaching of arsenic and antimony from a tetrahedrite rich complex sulphide concentrate using alkaline sulphide solution ⇑ Samuel A. Awe, Åke Sandström

Division of Extractive Metallurgy, Department of Chemical Engineering and Geosciences, Luleå University of Technology, SE-971 87 Luleå, Sweden article info abstract

Article history: Removal of impurity elements in copper metallurgy is one of the major problems encountered today Received 24 May 2010 since pure copper ore reserves are becoming exhausted and the resources of unexploited ores often con- Accepted 18 August 2010 tain relatively high amounts of impurity elements like antimony, arsenic, mercury and bismuth, which Available online 16 September 2010 need to be eliminated. The present work is aimed at pre-treating a tetrahedrite rich complex sulphide concentrate by selective dissolution of the impurities, therefore, upgrading it for pyrometallurgical pro- Keywords: cessing. To accomplish this, dissolution of antimony and arsenic by an alkaline sulphide lixiviant from the Selective leaching concentrate were investigated. The lixiviant proved selective and effective to dissolve these impurity ele- Tetrahedrite ments from the concentrate with good recoveries. Further investigations on the factors influencing the Sodium sulphide Antimony leaching efficiency of the lixiviant were studied. The parameters considered were sulphide ion and Impurity element hydroxide ion concentrations, mineral particle size, reaction temperature and leaching time. Analysis Complex sulphide concentrate of the leach residue indicates that copper content of tetrahedrite has transformed into copper sulphides

with the average chemical formula Cu1.64S. The grade and economic value of the concentrate were improved greatly after sulphide treatment, and therefore, suitable as a feedstock for smelting. The impu- rities have been reduced to low levels which are tolerable in the smelting furnace and consequently reduce both the treatment and environmental problem encountered when such concentrate is processed pyrometallurgically. Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction cantly affect both the quality and the physical properties of the copper product (Filippou et al., 2007; Mihajlovic et al., 2007; Tong- Copper extraction through pyrometallurgical processing of con- amp et al., 2009; Viñals et al., 2003; Wang, 2004). Besides, arsenic centrate relies heavily on copper sulphides; and to a lesser degree and antimony have been considered as undesirable elements that on copper oxides, which are processed hydrometallurgically. The cause serious toxicological and environmental problems (Filippou majority of copper sulphide mineral deposits mined today are et al., 2007) which have forced the smelters to be selective of the complex in nature. These minerals are often found in association type of concentrate to process due to the stringent environmental with tetrahedrite (Cu12Sb4S13), enargite (Cu3AsS4) and tennantite laws. Wang (2004) explained further that during copper smelting, (Cu12As4S13) minerals (Riveros and Dutrizac, 2008), which render most of the undesirable impurities such as antimony, arsenic, bis- them less appropriate as a feed material for smelting due to their muth and lead are only partially removed by oxidation. When antimony, arsenic and mercury contents that can create serious white metal and blister copper are in equilibrium, these impurities processing and environmental problems (Curreli et al., 2009; are partly distributed into the copper phase, from which their Filippou et al., 2007; Lattanzi et al., 2008). These minerals are eco- removal is difficult. Furthermore, Navarro and Alguacil (2002) nomically attractive due to their substantial contents of silver and explained that during electro-refining of impure copper, the copper (Balázˇ et al., 1998; Havlik et al., 1999); however, the impurity elements would gradually increase in concentration to content of antimony and arsenic effectively reduces their economic their solubility limits if no electrolyte was eliminated from the value due to requirements for further treatment options for the electro-refining circuit. However, they would passivate the anodes containment of hazardous emissions from pyrometallurgical pro- and also contaminate the cathodes which cause a decrease in the cessing. Also, the presence of these noxious elements will signifi- efficiency of the electro-refining plant. Consequently, the electro- lyte is usually treated conventionally by withdrawing a bleed from the stream to recover its copper content and isolate, and recover ⇑ Corresponding author. Tel.: +46 920 491290; fax: +46 920 491199. the impurities afterwards. But, this conventional method of treat- E-mail address: [email protected] (Å. Sandström). ment has some disadvantages which include; difficult materials

0892-6875/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.mineng.2010.08.018 1228 S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236 handling, energy consumption, possibility of toxic arsine gas for- Cu12Sb4S13ðsÞþ2Na2SðaqÞ!5Cu2SðsÞþ2CuSðsÞ mation, and loss of high copper in a low value recycle product þ 4NaSbS2ðaqÞð1Þ (Navarro and Alguacil, 2002). As a result of the global increase in demand for base metals, NaSbS2ðaqÞþNa2SðaqÞ!Na3SbS3ðaqÞð2Þ mining companies are now paying more attention to the low grade and complex sulphide ore deposits to economically process them. Sb2S3ðsÞþ3Na2SðaqÞ!2Na3SbS3ðaqÞð3Þ However, this necessitates pre-treatment operations to be con- ducted to selectively remove the impurities prior to the smelting 2Cu3AsS4ðsÞþ3Na2SðaqÞ!3Cu2SðsÞþ2Na3AsS4ðaqÞð4Þ of the copper concentrates. Therefore, this paper reports and dis- cusses the studies carried out on the selective leaching of antimony As2S3ðsÞþ3Na2SðaqÞ!2Na3AsS3ðaqÞð5Þ and arsenic under various experimental conditions from a tetrahe- drite rich complex sulphide concentrate in alkaline sulphide Balázˇ et al. (1998) disclosed that sulphide lixiviant reacts with media. mercury to form a soluble salt complex which could easily be hydrolyzed if the alkali concentration is very low. Furthermore, alkaline sulphide leaching processes have been 2. Methods of pre-treating As/Sb rich sulphide ores/ utilized industrially at operations in Russia, China, the Sunshine concentrates antimony refinery, Idaho, USA (Anderson and Krys, 1993), and also at Equity Silver Mines Ltd., British Columbia, Canada (Dayton, In an attempt to upgrade sulphide concentrates rich in arsenic 1982; Filippou et al., 2007). At Sunshine, the lixiviant was pro- and antimony for pyrometallurgical processing, selective flotation duced on site by dissolving elemental sulphur, a major product of enargite, tetrahedrite and tennantite from other copper sulp- of the silver refinery, in sodium hydroxide and the resulting solu- hides such as covelite (CuS), chalcopyrite (CuFeS2), chalcocite tion contains the following chemical species; sodium sulphide, (Cu2S) and bornite (Cu5FeS4) has been studied (Curreli et al., thiosulphate and polysulphides according to reactions (6) and (7) 2005; Filippou et al., 2007). The outcome showed that these miner- (Ackerman et al., 1993). als are difficult to separate from other sulphide minerals (Filippou 4S ðsÞþ6NaOHðaqÞ!2Na2SðaqÞþNa2S2O3ðaqÞþ3H2O ð6Þ et al., 2007) because they exhibit similar flotation properties as cop- per sulphides and therefore report to copper concentrates. ðX 1ÞS ðsÞþNa2SðaqÞ!Na2Sx; wherex ¼ 2to5 ð7Þ Apart from this, a more cost effective and environmentally friendly process for the removal of antimony and arsenic from In addition, the leach processes at Sunshine Mining and Equity the so called ‘‘dirty concentrates” prior to pyrometallurgical pro- Silver Mines are somehow similar while the solution purification cessing has been considered to be achieved through hydrometal- processes are different. Antimony was electrowon from the leach lurgical process (Balázˇ and Achimovicˇová, 2006b; Balázˇ et al., solution at Sunshine Mining (Nordwick and Anderson, 1993) 1998; Nadkarni et al., 1975). The discussions about the possibility whereas Equity Silver Mines entirely recovered antimony in auto- of leaching antimony–arsenic bearing minerals in copper resources claves as sodium hydroxyl antimonate and arsenic was converted are available in literature. It is reported that the amphoteric nature to calcium arsenate to be impounded at a site far from the mine of antimony allows the use of both acidic and basic media for its (Dayton, 1982; Filippou et al., 2007). solubilization (Anderson, 2001; Balázˇ et al., 2001). In reality, only Due to the widespread occurrence of antimony and arsenic con- two lixiviant systems are utilized in antimony hydrometallurgy: taining minerals in copper concentrates, a suitable pre-treatment acidic chloride (Correia et al., 2001; Correia et al., 2000; Guy process to selectively remove antimony and arsenic from such con- et al., 1983; Havlik et al., 1999; Havlik and Kammel, 2000; Padilla centrates is a necessity. If the impurity content is reduced to an et al., 2005) and alkaline sulphide (Awe, 2008; Balázˇ and Achim- acceptable limit the penalties charged by the smelters would be ovicˇová, 2006a; Balázˇ and Achimovicˇová, 2006b; Balázˇ et al., alleviated. Thus, antimony can be recovered as marketable prod- 2001; Frohlich and Miklos, 2001). Previous studies have shown ucts while arsenic can be precipitated as a stable material (e.g. fer- that arsenic and antimony containing minerals dissolve slowly in ric arsenate) suitable for landfills. By so doing, pyrometallurgical acidic media below 100 °C(Correia et al., 2001; Padilla et al., treatment of the so called ‘‘dirty ore/concentrate” will be made 2005). Significantly faster leaching rates have been obtained in possible with more economical viability and environmental alkaline media containing sulphide (Awe et al., 2010; Awe and friendliness. Sandström, 2010; Balázˇ et al., 2000; Balázˇ et al., 2001) or sodium hypochlorite (Viñals et al., 2003) as well as in acid media at ele- 3. Materials and methods vated temperatures (Neiva Correia et al., 1993; Riveros and Dutri- zac, 2008). Due to the refractoriness of tetrahedrite, enargite and 3.1. Material characterization tennantite to most lixiviants; concentrated leaching agents, high reaction temperatures and long leaching times are required to effi- The antimony-bearing sulphide material used in this investiga- ciently dissolve antimony and arsenic (Balázˇ, 2000). In a recent tion was bought from Gregory, Bottley & Llooyd Company, United study conducted by the authors, it was shown that the rate of Kingdom and originated from Casapalca, Departamento Lima, Peru. leaching tetrahedrite in alkaline sulphide lixiviant was chemically The coarse material was first crushed and part of its associated controlled through the particle surface reaction (Awe et al., 2010). quartz was removed by handpicking prior to grinding. The ground Filippou et al. (2007) reported that alkaline sulphide solution is sample was sieved into the following size fractions: 106 + 75, selective in dissolving antimony and arsenic; and the copper prod- 75 + 53, 53 + 38 and 38 + 20 lm, which were used in the uct obtained after sulphide leaching may be chalcocite or covelite, study. The concentrate contains 5.43–5.78% Sb and 1.77–1.87% while arsenic and antimony are dissolved as thioanions: thioarse- As and other elements as shown in Table 1. 3 3 3 nite (AsS3 ), thioarsenate (AsS4 ), thioantimonite (SbS3 ), thioan- X-ray powder diffraction (XRD) patterns were collected from 3 timonate (SbS4 ), etc. depending on the feed composition and the pulverised solid samples using a Siemens D5000 Automated Pow- reaction conditions. The chemical reactions occurring during sul- der Diffractometer equipped with a continuous scanning device. phide leaching are briefly described as follow (Filippou et al., Patterns were collected between 2h angles of 5–90° (at Cu Ka radi- 2007): ation of 40 kV, 30 mA). Mineral phases were identified using the S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236 1229

Table 1 solution containing 1.14 M Na2S3H2O + 1.5 M NaOH. A 500 mL Chemical analysis of a tetrahedrite rich sulphide concentrate. four-necked round bottomed glass reactor was used which was Size fraction (%) 106 + 75 lm 75 + 53 lm 53 + 38 lm stirred mechanically with a paddle stirrer at a constant stirring rate Cu 15.6 15.2 15.7 of 600 rpm and heated in an auto-regulated system. The lixiviant Fe 11.1 14.2 13.3 was first added to the reactor and when the desired temperature Zn 16.1 16.8 16.9 was reached, the solid sample was added. At predetermined time Sb 5.78 5.43 5.71 intervals, 5 mL sample solution was taken from the reactor for As 1.87 1.86 1.77 the analysis of dissolved metals. All reagents used for leaching and chemical analysis were of analytical grade and used without further purification. The influence of each variable was determined Joint Committee for Powder Diffraction Standards (JCPDS) file of by keeping all other variables constant. Leaching results were eval- the instrument. uated by means of chemical determinations on the leach products A representative sample was taken from the antimony-bearing using Inductively Coupled Plasma-Atomic Emission Spectrometry sulphide concentrate and also from the leach residue, at which the (ICP-AES)/Sector Field Mass Spectrometry (ICP-SFMS). highest recovery of antimony was obtained, for QEMSCAN (a scan- ning electron microscope with four energy dispersive X-ray detec- 4. Results and discussion tors and a microanalyser) analysis. The samples were mounted in epoxy and a diamond cutting wheel was used to cut the hardened 4.1. Mineralogical composition blocks to size and expose the sample grains. The samples were pol- ished and carbon coated prior to QEMSCAN analysis. The mineral- The mineralogical composition of the antimony-bearing sul- ogy of particles was inferred from data obtained using the phide concentrate was investigated using XRD and QEMSCAN QEMSCANÒ advanced mineral analysis system. The particles were methods of analysis. Fig. 1 shows the result obtained from XRD examined using the Field Scan method, involving spot analyses analysis performed on the concentrate. The major mineralogical over the surface of the block, collection of the resultant X-ray spec- phases identified in the diffractogram were tetrahedrite, sphaler- tra and comparison with a spectral data base. ite, galena, chalcopyrite, pyrite and iron sulphide. QEMSCAN analysis was conducted on the concentrate sample to 3.2. Leaching confirm the mineralogy previously determined by the XRD technique. The result from this method is shown in Fig. 2 and Antimony and arsenic selective removal tests were performed Table 2. This result compares well with the XRD identification of by dissolving the antimony-bearing sulphide concentrate in alka- the major phases present in the concentrate, but the method iden- line sulphide solutions, to determine the effect of the following tifies a much wider range of minor minerals (Table 2) that were be- parameters: sodium sulphide (Na2S3H2O) concentration (0.76 M, low the detection limit for XRD. This confirms that QEMSCAN 1.14 M and 1.89 M), sodium hydroxide (NaOH) concentration analysis can provide much more mineralogical information than (0.75 M and 1.5 M), leaching temperature (357 K, 364 K, 371 K the XRD method for trace mineralogy (Goodall et al., 2005). No and 378 K) and mineral particle size (53 + 38 lm, 75 + 53 lm arsenic mineral was detected by the two techniques meaning that and 106 + 75 lm). All experiments, each lasting 6 h, were carried the arsenic in the mineral is hosted in the tetrahedrite structure. out with 0.5% solids, 75 + 53 lm particle size and 378 K reaction Also, QEMSCAN result shows that the concentrate is rich in tetra- temperature (except where otherwise stated) in 500 mL leach hedrite (Table 2) and silver is found bonded in the tetrahedrite

Fig. 1. XRD diffractogram of antimony-bearing sulphide concentrate. 1230 S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236

Fig. 2. QEMSCAN micrograph of Sb-bearing sulphide concentrate.

Table 2 Material mineralogy defined by QEMSCAN.

Mineral Mass fraction (% w/w) Tetrahedrite 30.2 Sphalerite 20.6 Galena 19.0 Pyrite 15.7 Chalcopyrite 13.4 Quartz 0.4 Tetrahedrite with Ag 0.1

Andalusite (Al2SiO5) 0.1 Total 99.5

(Fig. 2). This confirms the presence of silver and arsenic as a solid solution in the tetrahedrite crystal structure as widely reported in the literature (Filippou et al., 2007).

4.2. Eh–pH and speciation diagrams

In order to study the behaviour of sulphide ions during leaching in relation to the solution pH, an equilibrium Eh–pH diagram for S-H2O system was constructed using FactSage 6.1 computer soft- ware at the temperatures of 298 K and 378 K (Fig. 3a and b), respectively. The equilibrium diagram (Fig. 3) obtained shows that at lower redox potential, S2 ion is the dominant ion in the solution at pH P 13 at 298 K (Fig. 3a) and at pH P 11 at 378 K (Fig. 3b). This implies that higher concentration of sodium hydroxide will be re- quired at 298 K than at 378 K in order to prevent the protonation of S2 to HS in the solution and therefore, making it available to sol- ubilize antimony by forming soluble sulphide complexes. During Fig. 3. Eh–pH diagram of the S-H2O system at (a) 298 K and (b) 378 K. The dashed sulphide leaching of tetrahedrite, sodium sulphide dissociates line represents H2(g)/H2O. according to reaction (8) and antimony forms the complexes thi- oantimonite and thioantimonate (Filippou et al., 2007) with sul- such as polysulphides, thiosulphates and sulphates (Fig. 3) may phide ions as shown in reactions (1) and (2). Many other species form if the leaching is performed under an oxidising environment. S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236 1231

4.3. Influence of temperature on antimony and arsenic dissolution

The plot shown in Fig. 5 describes the dependency of antimony and arsenic extraction from a tetrahedrite rich sulphide concen- trate on the reaction temperature which was varied in the range between 357 K and 378 K. It is evident from Fig. 5 that, after 6 h, the extraction of antimony and arsenic from the mineral is highly influenced by the reaction temperature. At 357 K, 15% and 16% of antimony and arsenic were extracted respectively, and when the reaction temperature was increased to 378 K, about 57% and 60% of antimony and arsenic were extracted, respectively. This obser- vation supports the conclusion made by (Balázˇ and Achimovicˇová, 2006a) that temperature accelerates and enhances the extraction of metals into the leach liquor.

4.4. Effect of leaching time on extraction

The effect of leaching time on antimony and arsenic dissolution from the complex sulphide concentrate were investigated and the Fig. 4. Equilibrium diagram of sulphide species as a function of pH at a total experimental results are presented in Fig. 6. Increasing the leach- sulphide ion concentration of 1.14 M at 298 K. ing time resulted in increased antimony and arsenic extraction. Due to the narrow size fraction used in this investigation, the metal extraction by the lixiviant is linearly dependent on the leaching Protonation of sulphide ions occur according to the equilibria time. After 1 h of leaching, approximately 13% of both antimony (9) and (10) (Delfini et al., 2003) which take place simultaneously and arsenic was extracted, and when the leaching time was in- during the leaching process. creased to 6 h, antimony and arsenic extraction was 57% and þ 60%, respectively. Examination of metal extraction trend shown Na S $ 2Na þ S2 ð8Þ 2 in Fig. 6 reveals that higher extraction could be achieved if the

2 þ leaching time was extended beyond 6 h. S þ H $ HS log K1 ¼ 11:96 ð9Þ

þ 4.5. Influence of sulphide ion and hydroxide ion concentrations HS þ H $ H2S log K2 ¼ 7:04 ð10Þ

2 Fig. 4 shows a plot of the molar concentrations of S ,HS and H2S In the experiments to study the influence of Na2S concentration in the leach liquor as a function of the solution pH at a total sul- on antimony and arsenic dissolution, Na2S concentration was var- phide concentration of 1.14 M. It can be seen from Fig. 4 that sul- ied in the range of 0.76–1.89 M while the NaOH concentration was phide ion concentration increases rapidly when pH is increased kept constant at 1.5 M. The result is described in Fig. 7. It is appar- from 10 and upwards, therefore, pH is an important factor. This is ent from the figure that Na2S concentration has a significant effect because, at a pH > 12, the protonation of sulphide ions will be pre- on the extraction of antimony and arsenic from the mineral. After vented and therefore available to solubilize antimony according to 6 h leaching, approximately 21% of antimony and arsenic was ex- reactions (1) and (2). Also, a strong alkaline environment is needed tracted at a concentration of 0.76 M Na2S and when the Na2S con- to reduce the consumption of more expensive sodium sulphide due centration was increased to 1.14 M, the extraction increased with to the hydrolysis of sulphide ions (Eq. (11)). almost a factor of 3. Further increase in the concentration to 1.89 M Na S resulted in an extraction of antimony and arsenic of 2 þ $ þ ð Þ 2 S H2O HS OH 11 87% and 92%, respectively (Fig. 7).

Fig. 5. Influence of reaction temperature on As/Sb leaching. 1232 S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236

Fig. 6. Effect of leaching time on As/Sb leaching.

Fig. 7. Effect of sodium sulphide concentration on As/Sb leaching.

Fig. 8. Effect of sodium hydroxide concentration on As/Sb leaching. S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236 1233

Fig. 9. Effect of particle size on As/Sb leaching.

Fig. 10. XRD diffractogram of the residue after alkaline sulphide leaching.

In order to study the effect of hydroxide ion concentration on of sulphide ion which otherwise would reduce its leaching the extraction, the leaching conditions were maintained the same efficiency. as described above except that the concentration of Na2S was con- stant at 1.14 M while the concentration of NaOH were 0.75 M and 4.6. Particle size dependence on antimony and arsenic dissolution 1.5 M. Fig. 8 illustrates that when a low concentration (0.75 M) of The influence of particle size on the extraction of antimony NaOH was used, the rate of extracting antimony and arsenic was and arsenic from the complex sulphide concentrate was examined slower and about 25% and 24% of antimony and arsenic were dis- for three different size fractions (106 + 75, 75 + 53 and solved respectively. As the concentration of NaOH increased to 53 + 38 lm) at 378 K using solutions containing 1.14 M Na2S 1.5 M, the metal extraction rate is accelerated and the percentage and 1.5 M NaOH. It is illustrated in Fig. 9 that extraction of both of metal extracted was enhanced. About 57% antimony and 60% ar- metals increases with decreasing particle size. After 6 h leaching senic was leached by the lixiviant after 6 h. The observed increase of the 106 + 75 lm particle size fraction, about 43% and 49% of in metal extraction at high concentration of NaOH is due to the antimony and arsenic were extracted, respectively. When the same higher availability of S2 ion for dissolution action (as described leaching conditions were applied to the size fraction 53 + 38 lm, in Section 4.2 above), which is promoted by the increase in the antimony and arsenic extraction were 68% and 75%, respectively. pH of the leach liquor and consequently, prevents the hydrolysis Since the leaching kinetics of tetrahedrite is relatively slow 1234 S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236

Fig. 11. Mineralogical study of Sb-bearing concentrate and leach residue based on QEMSCAN analysis.

Fig. 12. QEMSCAN micrograph of leach residue.

there is much to gain by grinding the material thereby creating a NaOH, 0.5% solids, 378 K reaction temperature and 6 h leach time larger surface area which is advantageous for the leaching where antimony recovery of 87% was obtained. The XRD pattern efficiency. shown in Fig. 10 indicates that a new mineralogical phase, spi-

onkopite (Cu39S28) is observed together with other major phases 4.7. Selective removal of the impurity elements identified in the original concentrate (Fig. 1). Similarly, a non-stoi- chiometric copper sulphide (Cu3.9S2.8) phase was observed when In order to identify the mineralogical changes brought about by the residues from enargite leached in alkaline sulphide solutions leaching, solid residue from the leaching test at which the highest were analyzed (Filippou et al., 2007). The XRD pattern in Fig. 10 antimony extraction was obtained, was characterized by XRD and shows that the tetrahedrite peaks found in the concentrate QEMSCAN analyses. The results were compared with the mineral- (Fig. 1) have disappeared completely meaning that copper content ogy of the original concentrate. The residue used was obtained of the tetrahedrite had been transformed to the new copper sul- under the following leaching conditions: 1.89 M Na2S + 1.5 M phide phase observed in the residue. S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236 1235

The results of the QEMSCAN analysis on the leach residue are ing and Metallurgy (CAMM), Luleå University of Technology, Swe- presented in Figs. 11 and 12. The major mineral phases observed den, during the experimental work. in the residue were sphalerite, galena, pyrite and chalcopyrite, with the newly formed copper sulphide phase. In particular, obser- References vation of the polished sections of the leach residue (Fig. 12) showed a micro-porous structure of the new solid phase often Ackerman, J.B., Anderson, C.G., Nordwick, S.M., Krys, L.E., 1993. Hydrometallurgy at the Sunshine mine metallurgical complex. In: Hiskey, J.B., Warren, G.W. (Eds.), enclosing a nucleus of untransformed tetrahedrite. The average Hydrometallurgy Fundamentals, Technology and Innovations. Salt Lake City, crystal chemical formulae of the solid residue determined by QEM- Utah, pp. 477–498. SCAN analysis on the transformed phase proves the conversion of Anderson, C., 2001. Hydrometallurgically treating antimony-bearing industrial tetrahedrite into a copper sulphide having stoichiometry of wastes. JOM Journal of the Minerals, Metals and Materials Society 53 (1), 18–20. Anderson, C.G., Krys, L.E., 1993. Leaching of antimony from a refractory precious Cu1.64S, which is close to the stoichiometry of the spionkopite ob- metals concentrate. In: Hiskey, J.B., Warren, G.W. (Eds.), Hydrometallurgy served by XRD but can also be a mixture of chalcocite and covellite. Fundamentals, Technology and Innovations. Salt Lake City, Utah, pp. 341–362. This observation confirms the expectation that tetrahedrite would Awe, S.A., 2008. Selective removal of impurity elements from Maurliden Västra complex sulphides flotation concentrate. MSc Thesis, Luleå. University of decompose to either covellite or chalcocite in alkaline sulphide Technology, Luleå, 42 pp. solution (Filippou et al., 2007). However, if the leaching time had Awe, S.A., Samuelsson, C., Sandström, Å., 2010. Dissolution kinetics of tetrahedrite been prolonged further, the untransformed tetrahedrite would mineral in alkaline sulphide media. Hydrometallurgy 103 (1–4), 167–172. Awe, S.A., Sandström, Å., 2010. Leaching mechanism of tetrahedrite in alkaline have been converted completely into the new copper sulphide sulfide solution. In: Conference in Minerals Engineering, Luleå, Sweden, pp. 13– phase observed in the residue. Additionally, QEMSCAN analysis 24. provides more interesting information about the selectivity of Balázˇ, P., 2000. Extractive Metallurgy of Activated Minerals. Elsevier, Amsterdam. Balázˇ, P., Achimovicˇová, M., 2006a. Mechano-chemical leaching in hydrometallurgy the lixiviant to solubilize the impurity elements from the concen- of complex sulphides. Hydrometallurgy 84 (1–2), 60–68. trate. Comparing the information given in Fig. 11 concerning the Balázˇ, P., Achimovicˇová, M., 2006b. Selective leaching of antimony and arsenic from concentrate and the residue, one can see that the major phases mechanically activated tetrahedrite, jamesonite and enargite. International Journal of Mineral Processing 81 (1), 44–50. found in the concentrate were upgraded and not leached Balázˇ, P., Achimovicˇová, M., Bastl, Z., Ohtani, T., Sanchez, M., 2000. Influence of (Fig. 12) with the only exception of the tetrahedrite phase which mechanical activation on the alkaline leaching of enargite concentrate. was reduced from 30.2% to 1.1% in the residue. This result confirms Hydrometallurgy 54 (2), 205–216. that other metal sulphides in the concentrate remains insoluble in Balázˇ, P., Achimovicˇová, M., Ficeriová, J., Kammel, R., Sepelák, V., 1998. Leaching of antimony and mercury from mechanically activated tetrahedrite Cu12Sb4S13. the sulphide solution with the exception of antimony and arsenic Hydrometallurgy 47 (2–3), 297–307. sulphides which form soluble complexes in sulphide media. This Balázˇ, P., Kammel, R., Villachica, C., 2001. As and Sb leaching from polymetallic analysis shows that the Na S/NaOH lixiviant is suitable for pre- sulfide concentrates. Metall(Berlin, West) 55 (4), 196–200. 2 Correia, M.J., Carvalho, J., Monhemius, J., 2001. The effect of tetrahedrite treating, difficult to treat, complex antimony–arsenic-containing composition on its leaching behaviour in FeCl3–NaCl–HCl solutions. Minerals sulphide ores/concentrates prior to their smelting and thereby Engineering 14 (2), 185–195. avoiding the difficulty experienced when such ores/concentrates Correia, M.J.N., Carvalho, J.R., Monhemius, A.J., 2000. The leaching of tetrahedrite in ferric chloride solutions. Hydrometallurgy 57 (2), 167–179. are directly treated in smelting plant. Curreli, L., Garbarino, C., Ghiani, M., Orrù, G., 2009. Arsenic leaching from a gold bearing enargite flotation concentrate. Hydrometallurgy 96 (3), 258–263. Curreli, L., Ghiani, M., Surracco, M., Orrù, G., 2005. Beneficiation of a gold bearing 5. Conclusions enargite ore by flotation and As leaching with Na-hypochlorite. Minerals Engineering 18 (8), 849–854. Dayton, S., 1982. Equity silver on line with leach plant: How Sb and As are purged The mineralogical investigation of the complex sulphide con- from a high-silver copper concentrate before smelting. Engineering and Mining centrate reveals that the concentrate is rich in tetrahedrite. Anti- Journal 183 (1), 78–83. mony, arsenic and silver form a solid solution in the tetrahedrite Delfini, M., Ferrini, M., Manni, A., Massacci, P., Piga, L., 2003. Arsenic leaching by Na2S to decontaminate tailings coming from colemanite processing. Minerals crystal structure and other major minerals identified were sphaler- Engineering 16 (1), 45–50. ite, galena, chalcopyrite and pyrite. Alkaline sulphide leaching of Filippou, D., St-Germain, P., Grammatikopoulos, T., 2007. Recovery of metal values the concentrate shows that antimony and arsenic extraction from from copper-arsenic minerals and other related resources. Mineral Processing and Extractive Metallurgy Review 28 (4), 247–298. tetrahedrite strongly depends on the concentration of sulphide Frohlich, L., Miklos, V., 2001. Leaching of As, Sb and Hg from tetrahedrite and hydroxide ions, reaction temperature and the leaching time. concentrate in the Na2S medium at increased temperatures. Metalurgija 40 Due to the relative slow leaching kinetics of tetrahedrite, fine min- (4), 213–218. eral particles are required in order to significantly improve anti- Goodall, W.R., Scales, P.J., Butcher, A.R., 2005. The use of QEMSCAN and diagnostic leaching in the characterisation of visible gold in complex ores. Minerals mony and arsenic extraction. Analysis of the leach residue Engineering 18 (8), 877–886. indicates that the lixiviant is strongly selective to remove the Guy, S., Broadbent, C.P., Lawson, G.J., Jackson, J.D.J., 1983. Cupric chloride leaching of impurity metals As and Sb. Tetrahedrite in the concentrate was a complex copper/zinc/lead ore. Hydrometallurgy 10 (2), 243–255. Havlik, T., Ivanova, Z., Dvorscikova, J., Kammel, R., 1999. Extraction of copper and found to be converted into a new species having the average chem- antimony from tetrahedrite by acid oxidative leaching. Metall 53 (7), ical formula Cu1.64S which could be a mixture of chalcocite, covel- 390–394. ite or spionkopite. The study shows that alkaline sulphide leaching Havlik, T., Kammel, R., 2000. Procedure for selective copper recovery from tetrahedrite. Metall 54 (1–2), 26–29. can be a suitable hydrometallurgical pre-treatment process from Lattanzi, P. et al., 2008. Enargite oxidation: A review. Earth-Science Reviews 86 technological and environmental point of view, to selectively and (1–4), 62–88. effectively reduce antimony and arsenic content from difficult to Mihajlovic, I., Strbac, N., Zivkovic, Z., Kovacevic, R., Stehernik, M., 2007. A potential method for arsenic removal from copper concentrates. Minerals Engineering 20 treat complex sulphide concentrates, and therefore, upgrading it (1), 26–33. for further pyrometallurgical treatment. Nadkarni, R.M., Kusik, C.L., Heissner, H.P., 1975. Method of Removing Arsenic and Antimony from Copper Ore Concentrate. Assigned to Arthur D. Little, Inc. (Cambridge, MA), US Patent 3,911,078. Acknowledgements Navarro, P., Alguacil, F.J., 2002. Adsorption of antimony and arsenic from a copper electrorefining solution onto activated carbon. Hydrometallurgy 66 (1–3), 101– 105. Financial contributions from the Swedish Governmental Agency Neiva Correia, M.J., Carvalho, J.R., Monhemius, A.J., 1993. Study of the autoclave for Innovation Systems, VINNOVA, and the participating compa- leaching of a tetrahedrite concentrate. Minerals Engineering 6 (11), 1117– nies: Boliden Mineral AB, LKAB and the Adolf H. Lundin Charitable 1125. Nordwick, S.M., Anderson, C.G., 1993. Advances in antimony electrowinning at the Foundation, are gratefully acknowledged. The authors would like Sunshine mine. In: Hiskey, J.B., Warren, G.W. (Eds.), Hydrometallurgy to appreciate the help received from the Centre of Advanced Min- Fundamentals, Technology and Innovations. Salt Lake City, Utah, pp. 1107–1128. 1236 S.A. Awe, Å. Sandström / Minerals Engineering 23 (2010) 1227–1236

Padilla, R., Giron, D., Ruiz, M.C., 2005. Leaching of enargite in H2SO4–NaCl–O2 media. Viñals, J., Roca, A., Hernández, M.C., Benavente, O., 2003. Topochemical Hydrometallurgy 80 (4), 272–279. transformation of enargite into copper oxide by hypochlorite leaching. Riveros, P.A., Dutrizac, J.E., 2008. The leaching of tennantite, tetrahedrite and Hydrometallurgy 68 (1–3), 183–193. enargite in acidic sulphate and chloride media. Canadian Metallurgical Wang, S., 2004. Impurity control and removal in copper tankhouse operations. JOM Quarterly 47 (3), 235–244. Journal of the Minerals, Metals and Materials Society 56 (7), 34–37. Tongamp, W., Takasaki, Y., Shibayama, A., 2009. Arsenic removal from copper ores and concentrates through alkaline leaching in NaHS media. Hydrometallurgy 98 (3–4), 213–218. PAPER III

Separation of antimony from a complex copper concentrate: Process optimization by response surface methodology

Samuel A. Awe, Mohammad Khoshkhoo, Paul Kruger and Åke Sandström

Manuscript under review in the Journal of Separation and Purification Technology

Separation of antimony from a complex copper concentrate: process optimisation by response surface methodology

Samuel A. Awea, Mohammad Khoshkhooa, Paul Krugerb and Åke Sandströma*

aDivision of Extractive Metallurgy, Department of Chemical Engineering and Geosciences,

Luleå University of Technology, SE-971 87 Luleå, Sweden. bBoliden Mineral AB, SE-936 81 Boliden, Sweden

Manuscript submitted to Journal of Separation and Purification Technology

*Corresponding author: Åke Sandström, Division of Extractive Metallurgy, Department of

Chemical Engineering and Geosciences, Luleå University of Technology, SE-971 87 Luleå,

Sweden.

E-mail: [email protected],

Tel.: +46 920 491290

Fax: +46 920 491199 ABSTRACT

This paper reports the modelling and optimisation results for the alkaline sulphide leaching of a complex copper concentrate containing 1.69% Sb and 0.14% Sn. Response surface methodology, in combination with central composite face-centred design (RSM-CCF), was used to optimise the operating parameters. The leaching temperature, sulphide ion concentration and solid concentration were chosen as the variables, and the response parameters were antimony and tin recovery, and the time required to achieve 90% Sb dissolution. It was confirmed that the leaching process was strongly dependent on the reaction temperature as well as the sulphide ion concentration without any significant dependence on the solid concentration. Additionally, a mathematical model was constructed to characterise the leaching behaviour. The results from the model allow identification of the most favourable leaching conditions. The model was validated experimentally, and the results show that the model is reliable and accurate in predicting the leaching process.

Keywords: Alkaline sodium sulphide, Antimony, Leaching optimisation, Response surface methodology, Central composite face-centred design

2 1. Introduction

Antimony / arsenic-bearing minerals are commonly found in minor concentrations in many sulphide ore deposits. The tetrahedrite (Cu12Sb4S13) mineral is often found in association with copper sulphide ores, and though it is an important source of copper, but it is more significant as a source of silver and antimony. Enargite (Cu3AsS4) and tennantite (Cu12As4S13) are minor sources of copper, and they result in arsenic reporting to the processing circuit. These antimony and arsenic minerals contain significant concentrations of minor elements such as mercury, zinc, bismuth, selenium, etc., which result in penalties for copper concentrates sold to smelters [1]. However, processing of future copper ores and concentrates will most likely involve the treatment of more complex, fine-grained minerals containing increased levels of penalty elements, which are detrimental to the smelting process. Unfortunately, the prevalence of tetrahedrite, tennantite and enargite among the copper-bearing minerals will reduce their economic value due to their minor element content which generates hazardous emissions during pyrometallurgical processing [2].

In copper metallurgy, the removal of impurities is crucial to the production of high-quality copper. The ability to efficiently remove impurities will become even more important in the future, because copper ores show a decreasing ore grade and increasing levels of impurities.

Consequently, the penalties charged by smelters for treating minor elements in copper concentrates will increase because the copper smelters are forced to improve their technology for the efficient removal and disposal of the penalty elements, which are difficult and costly to handle [3].

In Sweden, many of the complex sulphide mineral deposits found in the Boliden mining area are found in association with impurities like arsenic, antimony, bismuth and mercury which make it difficult to produce clean and high grade concentrates [4]. During the pyrometallurgical treatment of the complex sulphide concentrate at the smelting plant, arsenic

3 and mercury in the concentrate end up in the roaster flue dust which is therefore regarded as hazardous metallurgical waste. The arsenic oxide generated in the waste material is highly soluble and consequently, creates a serious environmental problem, which needs to comply with the stringent legislation for its disposal as landfill. As a result of more stringent environmental legislation, the allowable amount of arsenic in the processing waste generated is progressively decreasing. Consequently, high financial penalties are imposed by smelters to treat copper ores containing more than 0.2 wt.% arsenic [3]. Besides, the antimony, arsenic and bismuth contents of the complex sulphide mineral renders it undesirable as pyrometallurgical smelter feed because these impurities affect the conductivity and ductility of the refined copper and also form a scum on electrolytic cells which allow the impurities to be carried over to the cathodes [5, 6]. These impurities are deleterious from the standpoint of metal quality. It is known that small amounts of bismuth and antimony in the copper cathode makes it brittle and renders it unsuitable for producing electric wires. Therefore, it is desirable to remove these impurities at an early stage of processing, since they are difficult to remove from metallic copper. However, in order to increase the economic value and metal grade of the copper concentrates, as well as decreasing the cost of the environmental measures needed when complex sulphide concentrates are processed pyrometallurgically, it is therefore imperative to choose a pre-treatment processing route which can selectively remove these impurities from the complex sulphide concentrates prior to their smelting. In view of this, the hydrometallurgical processing route using alkaline sulphide lixiviant has been proven to be the best alternative route to remove these impurity elements selectively from such concentrates [4, 7-10] and consequently, produce a suitable feed for pyrometallurgical operations .

4 The efficiency of the process strongly depends on factors such as reaction temperature, concentration of sulphide ion, leaching time, particle size, the amount of solid leached and so forth.

Conventionally, the study of the effects of aforementioned factors, on sulphide leaching of impurity-bearing copper concentrates, has been conducted using an approach with one variable at a time [4, 7, 11]. The effect of each experimental factor was investigated in this method by altering the level of one factor at a time while keeping the levels of the other factors constant. But, if the aim of conducting a study is to determine the optimum operating conditions, response surface methodology (RSM) will be the appropriate method of performing the task. Conversely, RSM is an efficient methodology where all the experimental factors are varied simultaneously over a set of experimental runs. RSM is one of the relevant multivariate techniques that can deal with multivariant experimental design strategy, statistical modelling and process optimisation [12, 13]. It is employed to study the relationship between one or more response variables and a set of quantitative or qualitative experimental factors. RSM is often used after the vital controllable factors are identified and to find the factor settings that optimise the response [14]. This type of experimental design is usually chosen when non-linear interactions between the experimental parameters and the response variables are suspected. Furthermore, application of RSM reduces the number of experiments required for the analysis of the main effects and interactions between factors [14-

16]. The objectives of the current research were to evaluate the influence of these experimental factors reaction temperature, sulphide ion concentration and solid concentration on the response variables and also model the behaviour of a complex copper concentrate in alkaline sulphide solution in order to determine the most favourable operating conditions.

5 2. Materials and methods

2.1 Material and characterisation techniques

Rockliden copper concentrate used in this investigation was a flotation concentrate obtained from Boliden Mineral AB, Sweden. Particle size distribution analysis performed on the concentrate shows that more than 80% of the mineral particles is smaller than 40 μm (Fig. 1).

Chemical analysis of the concentrate is presented in Table 1.

Table 1: Elemental analysis of the complex copper concentrate

Cu, % Fe, % Pb, % Zn, % Sb, % Sn, % As, % Se, % Hg, % Ag, % S, % 17.8 27.1 7.1 5.7 1.69 0.14 0.42 0.06 0.03 0.08 >15

Fig. 1. Particle size distribution of the concentrate.

A representative sample was taken from Rockliden complex sulphide concentrate for the scanning electron microscope (SEM) using a SEM JEOL JSM5900LV equipped with an

EDS-analyser. The sample was embedded in epoxy and the harden block was polished and coated with carbon prior to the analysis. X-ray powder diffraction (XRD) was used to characterize the complex copper concentrate using a Siemens D5000 automatic diffractometer

6 equipped with a continuous scanning device. Patterns were collected between 2ș angles of

10–90° (at Cu KĮ radiation of 40 kV, 30 mA). Mineral phases were identified using the Joint

Committee for Powder Diffraction Standards (JCPDS) file of the instrument.

2.2 Leaching procedure

Leaching experiments were performed by dissolving the Rockliden complex sulphide copper concentrate samples in 800 mL of lixiviant solutions. The lixiviant solution was prepared by dissolving sodium sulphide (Na2Sʘ3H2O) in sodium hydroxide (NaOH) solution. The concentration of sodium hydroxide used at each run of the test was 20 wt% of sodium sulphide concentration used. A total of 17 experiments were conducted in a batch mode using a 1 L five-necked round bottomed glass reactor. The contents of the reactor were mechanically homogenized using a paddle stirrer at a constant stirring rate of 300 rpm and heated on auto-regulated device. All the leaching experiments lasted for 6 h. The lixiviant was first added to the reactor and when the desired temperature was reached, the solid sample was added. At predetermined time intervals, slurry sample was taken from the reactor for the analysis of dissolved metals. All reagents used for leaching and chemical analysis were of analytical grade and used without further purification. Leaching results were evaluated by means of elemental determinations on the leach products using Inductively Coupled Plasma-

Atomic Emission Spectrometry (ICP-AES) / Sector Field Mass Spectrometry (ICP-SFMS).

2.3 Experimental design

The experiment was designed using response surface methodology (RSM) approach. RSM is a particular set of mathematical and statistical methods that includes experimental design, model fitting and validation, and condition optimisation [16]. Experimental runs were designed in accordance with central composite face-centred design (CCF), which allows a full quadratic model for each response under investigation. A detailed discussion of CCF design is documented elsewhere [12, 13].

7 MODDE 8.0 Umetrics software was used in the design and analysis of the experiment. The effect of the following three factors on the efficiency of the leaching process was investigated: reaction temperature (X1), sulphide ion concentration (X2) and solid concentration (X3). The interaction between the effects of the various variables was also examined. According to the previous work conducted by the authors [7, 17], the levels (low and high) of each of the factors were chosen as follow: temperature (80 and 100 °C), sulphide ion concentration (41 and 164 g/L) and solid concentration (100 and 300 g/L). The response variables were the recovery of antimony (Sb) and tin (Sn); and leaching time T90, which is the time required to leach out 90% of antimony from the concentrate by alkaline sulphide lixiviant. Recovery of antimony and tin was computed based on solid residues analysis and it was calculated from equation 1 whilst T90 was estimated from antimony leaching kinetic profiles which is fitted to the exponential equation (Eq.2).

f  r R u100% (1) f

kt Ct C0 e (2)

Where R represents percent recovery of Sb or Sn, f and r are the amount (g) of Sb or Sn in the feed and residue, respectively, C0 is the initial concentration of Sb in the feed material and Ct is the concentration of antimony in the residue after time t (h) of leaching while k is the rate constant. Since the weight difference between the feed and leach residue was negligible, Eq. 2 was used to simplify the relationship between the leaching time and the amount of metal leached.

Table 2 shows the actual values of the independent variables at which the experiments were performed to estimate the response variables. The table contains the measured values of antimony and tin recoveries, as well as the calculated T90 (h) from Eq. 2. The experiments were conducted in randomized order, to avoid any occurring systematic time trend which

8 could cause error in the model. The original design, that involved replicates of the reference mixture for estimating the pure error, was augmented with replicates of random runs so that the homogeneity of variance error could be checked. The model fitting was investigated by using multiple linear regression (MLR) method.

Table 2 Central composite face-centred design arrangements and responses for the complex concentrate leaching by alkaline sulphide solution

Exp.no. Run Experimental factors Response variables order Temperature S-2 conc. Solid conc. Sb Sn Leaching X1 (°C) X2 (g/L) X3 (g/L) recovery recovery, time, % % (T90) h N1 13 80 41 100 43.32 34.02 26.53 N2 2 100 41 100 78.11 39.43 8.89 N3 10 80 164 100 93.98 36.04 4.39 N4 3 100 164 100 97.25 68.28 3.48 N5 1 80 41 300 39.10 35.42 29.93 N6 12 100 41 300 75.77 31.03 9.76 N7 8 80 164 300 94.27 37.93 4.40 N8 4 100 164 300 97.55 82.98 3.46 N9 5 80 102.5 200 87.03 31.52 6.46 N10 16 100 102.5 200 90.98 39.22 5.34 N11 9 90 41 200 63.10 32.86 13.81 N12 6 90 164 200 95.88 48.55 3.97 N13 15 90 102.5 100 91.03 33.40 5.25 N14 14 90 102.5 300 88.00 33.74 6.13 N15 11 90 102.5 200 89.19 33.18 5.72 N16 17 90 102.5 200 91.51 34.29 5.18 N17 7 90 102.5 200 91.03 31.97 5.34

The experimental results obtained from the CCF model were described in the form presented in Eq. 3,

k k k1 k 2 Y E0  ¦ Ei Xi  ¦ Eii Xi  ¦¦Eij Xi X j (3) i 1 i 1 i 1 j i1

Where Y is the predicted response, ȕ0, ȕi, ȕii and ȕij are constant coefficients, linear coefficients, interaction coefficients and the quadratic coefficients, respectively, and Xi, Xj are the coded levels of the factors investigated. The model fitting was evaluated by checking the

2 2 2 coefficients of determination R , R adj and Q values, which indicate the fraction of the variation of the response explained by the model, the fraction of variation of the response

9 described by the model adjusted for degrees of freedom and the fraction of the variation of the response predicted by the model, respectively. The validity of the model was examined at

95% confidence interval.

3. Results and discussion

3.1 Mineralogy of Rockliden copper concentrate

The mineralogical composition of the concentrate was investigated by XRD and SEM-EDS analytical techniques. Fig. 2 presents the result of XRD analysis performed on the concentrate. The main mineralogical phases identified in the XRD pattern were tetrahedrite, chalcopyrite, sphalerite, galena, iron sulphide and pyrite.

Fig. 2. XRD pattern of the concentrate.

The result from SEM-EDS analysis as shown in Fig. 3 corroborates the result obtained from the XRD analysis. This method identifies a much wider range of minor minerals (Table 3) that were below the detection limit for XRD. Apart from tetrahedrite, the SEM result (Fig. 3) shows that antimony was also found as bournonite (PbCuSbS3) while tin and arsenic was present as stannite (Cu2FeSnS4) and arsenopyrite, respectively. It is apparent from Table 3

10 that silver is bound in the tetrahedrite structure, which may enhance the value of the concentrate.

Fig. 3. SEM-EDS micrograph of the concentrate. A- tetrahedrite, B- bournonite, C- stannite,

D- chalcopyrite, E- galena, F- sphalerite, G- pyrite, H- arsenopyrite and I- pyrrhotite.

11 Table 3 Mineralogy and atomic proportions of selected microanalyses of Rockliden copper concentrate defined by SEM-EDS analysis

Cu % Fe % Zn % Pb % Sb % As % Sn % Ag % S % Mineral 1 35.41 5.36 30.92 1.31 26.99 Tetrahedrite 2 34.43 5.55 2.51 30.64 1.29 25.58 Tetrahedrite 3 33.71 5.77 1.62 31.34 2.03 25.55 Tetrahedrite 4 34.10 6.23 2.00 30.14 1.34 26.19 Tetrahedrite 5 34.41 5.27 1.75 30.79 1.63 26.14 Tetrahedrite 6 34.10 6.05 30.08 3.17 26.60 Tetrahedrite 7 12.48 1.06 38.81 27.11 20.55 Bournonite 8 13.06 0.84 39.01 26.10 20.99 Bournonite 9 12.54 1.37 38.35 26.33 21.41 Bournonite 10 27.23 12.57 29.11 31.09 Stannite 11 28.62 12.51 27.96 30.92 Stannite 12 0.99 32.89 44.79 21.32 Arsenopyrite 13 33.53 46.07 20.40 Arsenopyrite 14 33.15 30.01 36.84 Chalcopyrite 15 32.72 30.17 37.11 Chalcopyrite 16 1.34 83.07 15.59 Galena 17 8.14 55.58 36.28 Sphalerite 18 44.17 55.83 Pyrite 19 58.81 41.19 Pyrrhotite

3.2 Data evaluation and RSM model analysis

The distributions of the data for all the responses were inspected for normality, and it was observed that the responses needed to be transformed in order to obtain improved models.

Negative logarithmic transformation was carried out for Sb recovery, and power transformation was performed for both Sn recovery and T90 as detailed in Table 4. Evaluation of the raw data revealed that the replicate errors were satisfactorily small for all of the output responses.

Table 4 Transformation of responses

Response Transformation Sb (%) -10Log(100-Y) Sn (%) Y-1 -0.5 T90 (h) Y

12 The model was fitted using multiple linear regression (MLR) method, and the overall result of the model fitting is displayed in Fig. 4. A model can be judged as good, if R2- Q2< 0.2-0.3,

Q2>0.5, model validity >0.25 and reproducibility is greater than 0.5 [13]. It is observed that the predictive power (Q2) of the models ranges from excellent to good. The Q2 values for the responses Sb, T90 and Sn are 0.93, 0.91 and 0.55, respectively. All models also show high model validity, which indicates that, no lack of fit. Fig. 4 shows that Sb recovery and T90 models satisfied the model performance indicator conditions while the Sn recovery model did not. The lower Q2 value of Sn recovery model could be due to the presence of irrelevant terms contained in the regression model.

Fig. 4. Summary of fit plot of the initial modelling.

The regression coefficients plots of the models were examined and the statistically non- significant terms were eliminated. Thus, the models were refined and simplified. The summary of fit plot of the refined models is presented in Fig. 5. It can be seen from Fig. 5 that

13 all three Q2 values have increased, and now amount to 0.97, 0.97 and 0.74 for Sb recovery,

T90 and Sn recovery, respectively. The refined models show improved model validity.

Fig. 5. Summary of fit plot of the refined modelling.

The outcome of the normal probability plot of the residuals after refinement (Fig. 6) reveals that the models look satisfactory, except for the deviating behaviour of experiment 2 in Sn recovery model. Subsequently, experiment 2 was critically scrutinized for any possible error, but it was found to be well fitted into the other models; therefore, the experiment cannot be removed from the model. Due to satisfactory R2 and Q2 of Sn recovery model, it may be assumed that experiment 2 is a weak outlier which does not influence the model decisively. In order to obtain information concerning how the input variables affect the responses, regression coefficient plots of the refined models were made and interpreted (Fig. 7).

Inspection of these plots shows that reaction temperature and sulphide ion concentration have a strong effect on all the three responses. It seems that solid concentration has insignificant effect on all the three responses.

14 Fig. 6. Normal probability plots of residuals after model refinement.

Fig. 7. Regression coefficient plots for the responses Sb, Sn and T90 after model refinement with confidence intervals.

15 Furthermore, examination of the analysis of variance (ANOVA) of the refined models for the responses Sb recovery, Sn recovery and T90 as shown in Table 5, reveal that all the regression models are statistically significant with a 95% confidence level in the range studied. For all the response variables, F0 values are greater than Fcritical values, and P-values are smaller than

0.05. According to the results displayed in the ANOVA table (Table 5), it can be inferred that the model error of the original model is of the same magnitude as the replicate error for all the responses, because their P-values are greater than 0.05 and F0

Hence, the models have small error and good fitting power, meaning that the models show no lack of fit.

Table 5 ANOVA for the quadratic models predicted for each response variable

Response Source DF SS MS F0 Fcritical P-value SD variable (variance) Į = 0.05 Sb Total corrected 16 2.7026 0.1689 0.4109 recovery Regression 5 2.6619 0.5324 143.58 3.20 0.000 0.7296 Residual 11 0.0408 0.0037 0.0609 Lack of fit 9 0.0347 0.0039 1.28 19.37 0.514 0.0621 (model error) Pure error 2 0.0060 0.0030 0.0550 (replicate error) Sn Total corrected 16 0.0005 3.37e-5 0.0058 recovery Regression 5 0.0005 1.01e-4 30.61 3.20 0.000 0.0100 Residual 11 3.61e-53.29e-6 0.0018 Lack of fit 9 3.39e-5 3.77e-6 3.36 19.37 0.251 0.0019 (model error) Pure error 2 2.24e-6 1.12e-6 0.0010 (replicate error) T90 Total corrected 16 0.1823 0.0114 0.1068 Regression 5 0.1798 0.0360 155.78 3.20 0.000 0.1896 Residual 11 0.0025 0.0002 0.0152 Lack of fit 9 0.0023 0.0003 2.16 19.37 0.356 0.0160 (model error) Pure error 2 0.0002 0.0001 0.0109 (replicate error) DF-degree of freedom, SS-sum of squares, MS-mean square, SD-standard deviation

Moreover, the observed responses correlated very well with the predicted values as illustrated in Fig. 8. Thereby, the models can be considered adequate for the predictions and for the

16 optimisation process. The regression models describing the correlation between the transformed response variables and the parameters investigated were as written in Eqs. 4-6.

2 Sb = -1.002+ 0.168X1 + 0.485X2 – 0.079X 2 (4)

2 Sn = 0.030 – 0.003X1 – 0.004X2 – 0.005X 2 – 0.003X1 X2 (5)

2 T90 = 0.422 + 0.044X1 + 0.123X2 – 0.039X 2 – 0.020X1X2 (6)

In the regression models shown in Eqs. 4-6 all the experimental parameters are in coded values where X1 is the reaction temperature, X2 is sulphide ion concentration, X3 is solid

2 concentration, and (X2) and X1X2 are the square and interaction of the main factors. The presence of significant square and interaction terms in the regression equations confirms the existence of quadratic behaviour and non-linear combining effects of the variables.

Fig. 8. Relationship between observed and predicted values for the response variables Sb, Sn and T90.

17 3.3 Interpretation of the data by response surface modelling

According to the empirical models explained in Eqs. 4-6, response surface plots were developed, which provide a better understanding of the effect of the experimental parameters on the response variables. Fig. 9 displays contour plots at various solid concentrations where antimony recovery is represented by varying simultaneously leaching temperature from 80 °C to 100 °C and sulphide concentration from 41 g/L to 164 g/L. The figure shows that antimony recovery increases with increase in both sulphide ion concentration and reaction temperature.

The effect of both leaching temperature and sulphide concentration on antimony recovery can be explained by the fact that tetrahedrite, the antimony mineral, is refractory in nature.

Therefore, high sulphide concentration and reaction temperature are needed to enhance antimony dissolution by the lixiviant. Besides, tetrahedrite dissolution in alkaline sulphide solution is controlled by chemical reaction with a relatively high activation energy [7], hence, higher lixiviant concentration as well as increasing leaching temperature would favour its dissolution.

Fig. 9. Contour plots of Sb recovery showing the interaction between sulphide concentration (g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

18 Fig. 10 presents the effect of leaching temperature and sulphide concentration on tin recovery.

It is seen that these independent variables have significant effects on tin dissolution. Like antimony, tin extraction is observed to increase with increasing temperature and sulphide concentration. The reason could be that the tin mineral, stannite (Cu2FeSnS4), may be refractory to the lixiviant, as is the case for the tetrahedrite mineral. Consequently, higher concentrations of sulphide and temperature would be needed to decompose the mineral for effective dissolution of tin. It is evident from Fig 10 that the concentration of the solid leached does not influence tin or antimony recovery (Fig. 9). For example, at 164 g/L sulphide concentration and 100 °C leaching temperature, tin recovery at 100 g/L, 200 g/L and 300 g/L solid concentrations were 64.7%, 69.5% and 68.5%, respectively.

Fig. 10. Contour plots of Sn recovery showing the interaction between sulphide concentration

(g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

19 Fig. 11 shows the influences of temperature and sulphide concentration on the time required to dissolve 90% Sb from the copper concentrate. It can be seen from the contour plots that the time needed to leach out 90% Sb decreases with increasing temperature and sulphide concentration, with an insignificant effect of solid concentration. The model reveals that it is possible to leach at lower sulphide concentration and temperature, which is beneficial to the process and still ensures high metal recovery over an extended leaching time. For instance, the lower left corners of the plots in Fig. 11 states the leaching conditions, at which 90% Sb can be obtained over 22 h duration of leaching, while the lixiviant concentration and temperature are at minimum levels.

Fig. 11. Contour plots for the response T90 showing the interaction between sulphide concentration (g/L) and temperature (°C) at various solid concentrations (a) 100 g/L (b) 200 g/L (c) 300 g/L.

20 3.4 Model validation

In order to test the validity of the model with regards to the response variables Sb, Sn and T90, a separate experiment was performed at the conditions shown in Table 6 for 24 hours. The results, as shown in Table 6, indicate a close agreement with the values predicted by the model. Consequently, the model from a response surface methodology is considered to be accurate and reliable, for predicting the leaching of antimony and tin from the complex copper concentrate, and also good at predicting the time needed to recover 90% of the antimony from the concentrate.

Table 6 Validation of model

Temp. Sulphide Solid %Sb %Sn T90, h X1, °C conc. conc. Predicted Observed Predicted Observed Predicted Observed X2, g/L X3, g/L 80 45 290 36-58 52 30-37 26 19-30 20

3.5 Process optimisation and economic implications

Furthermore, optimisation of the factors affecting the leaching process can be carried out depending on what is expected from the process. If high recoveries of Sb and Sn are desirable in as short as possible leaching time and at any cost, one might opt for leaching at maximum temperature and sulphide ion concentration as displayed in Figs. 9 and 10. On the other hand, if for practical and economic reasons, low production cost (due to heating energy and reagent costs); and less corrosion and electrowinning problems (due to high alkalinity and sulphide concentration respectively) are desirable. Consequently, lower temperature and sulphide concentration can be used (Fig. 11) to set the factors in such a level to obtain high recovery of

Sb (min 90%) by extending the leaching time.

The three parameters considered in this study affect the economics of the process in various ways. Increasing the leaching temperature accelerates the leaching kinetics, resulting in a smaller volume for the leach tanks. However, for temperatures above 90q C, filtration will

21 have high maintenance costs due to material constraints, necessitating the use of additional cooling equipment prior to filtration. Therefore the economical optimum leaching temperature is likely to be around 90q C. There exists a strong correlation between the soluble sulphide

(free ion) concentration and the leaching rate. However, raising the sulphide concentration could reduce the tank volumes at the expense of an increase in sulphide losses to the bleed stream. So any savings in capital expenditure is likely to be offset by an increase in operating cost. In practice, the process would have a large recycle stream and a small bleed stream, to minimise the loss of sulphide lixiviant. Another effective means of reducing the tank volumes is to increase the solids concentration. The result would be an increase in the solution metal concentrations reporting downstream, which would improve the recovery of antimony.

Further work should therefore be undertaken to determine the maximum solids concentration that is technically feasible.

4. Conclusions

Modelling and optimisation of alkaline sulphide leaching of a complex copper concentrate containing 1.69% Sb and 0.14% Sn was conducted using response surface methodology- central composite face-centred design (RSM-CCF). It was demonstrated that the leaching process was strongly dependent on the reaction temperature as well as the sulphide ion concentration with insignificant dependence on solid concentration. A strong model with no lack of fit was developed and the validity of the model was confirmed experimentally. The result shows that the model is reliable and accurate for predicting the leaching process.

Acknowledgement

The authors would like to appreciate the following organisations VINNOVA, Boliden

Mineral AB, LKAB and the Adolf H. Lundin Charitable Foundation, for their financial supports. The contribution from the Centre of Advanced Mining & Metallurgy (CAMM) is gratefully acknowledged.

22 References

[1] P.A. Riveros, J.E. Dutrizac, The leaching of tennantite, tetrahedrite and enargite in acidic sulphate and chloride media, Canadian Metallurgical Quarterly, 47 (2008) 235-244. [2] I. Mihajlovic, N. Strbac, Z. Zivkovic, R. Kovacevic, M. Stehernik, A potential method for arsenic removal from copper concentrates, Minerals Engineering, 20 (2007) 26-33. [3] P.C. Wilson, C. Chanroux, Revenue, calculation and marketing: copper, in: M. Noakes, T. Lanz (Eds.) Cost Estimation Handbook for the Australian Mining Industry, The Australian Institute of Mining and Metallurgy, Melbourne, 1993, pp. 325-368. [4] S.A. Awe, Selective removal of impurity elements from Maurliden Västra complex sulphides flotation concentrate, in: Process Metallurgy, Luleå University of Technology, Luleå, 2008, pp. 42. [5] P. Baláž, M. Achimoviþová, Selective leaching of antimony and arsenic from mechanically activated tetrahedrite, jamesonite and enargite, International Journal of Mineral Processing, 81 (2006) 44-50. [6] D. Filippou, P. St-Germain, T. Grammatikopoulos, Recovery of metal values from copper- arsenic minerals and other related resources, Mineral Processing and Extractive Metallurgy Review, 28 (2007) 247-298. [7] S.A. Awe, C. Samuelsson, Å. Sandström, Dissolution kinetics of tetrahedrite mineral in alkaline sulphide media, Hydrometallurgy, 103 (2010) 167-172. [8] C.G. Anderson, Nordwick, S. M., Krys, L. E., Processing of antimony at the Sunshine mine. In residues and enffluents: Processing and Environmental Considerations., The Minerals, Metals and Materials Society, (1991) 349-366. [9] P. Baláž, R. Kammel, C. Villachica, As and Sb leaching from polymetallic sulfide concentrates, Metall(Berlin, West), 55 (2001) 196-200. [10] Z. Tian-cong, The metallurgy of antimony, Central South University of Technology Press Changsha, the People's Republic of China, 1988. [11] S.A. Awe, A. Sandström, Selective leaching of arsenic and antimony from a tetrahedrite rich complex sulphide concentrate using alkaline sulphide solution, Minerals Engineering, 23 (2010) 1227-1236. [12] D.C. Montgomery, Design and analysis of experiments, 7th ed., John Wiley & Sons, Inc., Asia, 2009. [13] L. Eriksson, E. Johansson, N. Kettaneh-Wold, C. Wikström, S. Wold, Design of experiments-principles and applications, Third ed., UMETRICS AB, Umeå, 2008. [14] Z. Zhang, J. Peng, C. Srinivasakannan, Z. Zhang, L. Zhang, Y. Fernández, J.A. Menéndez, Leaching zinc from spent catalyst: Process optimization using response surface methodology, Journal of Hazardous Materials, 176 (2010) 1113-1117. [15] J.V. Mehrabani, M. Noaparast, S.M. Mousavi, R. Dehghan, A. Ghorbani, Process optimization and modelling of sphalerite flotation from a low-grade Zn-Pb ore using response surface methodology, Separation and Purification Technology, 72 (2010) 242-249. [16] V. García, J. Landaburu-Aguirre, E. Pongrácz, P. Perämäki, R.L. Keiski, Dehydration of water/dichloromethane/n-butanol mixtures by pervaporation; optimisation and modelling by response surface methodology, Journal of Membrane Science, 338 (2009) 111-118. [17] S.A. Awe, Å. Sandström, Leaching mechanism of tetrahedrite in alkaline sulfide solution, in: Conference in Minerals Engineering, Luleå, Sweden, 2010, pp. 13-24.

23