ALTA 2010 GOLD ORE PROCESSING SYPOSIUM MAY 27-28, 2010 SHERATON HOTEL PERTH, AUSTRALIA
ALTA Metallurgical Services Melbourne, Australia www.altamet.com.au ALTA Free Library www.altamet.com.au
PROCEEDINGS OF ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM 27-28 May 2010 Perth, Australia
ISBN: 978-0-9871262-9-0
ALTA Metallurgical Services Publications All Rights Reserved Publications may be printed for single use only. Additional electronic or hardcopy distribution without the express permission of ALTA Metallurgical Services is strictly prohibited. Publications may not be reproduced in whole or in part without the express written permission of ALTA Metallurgical Services. The content of conference papers is the sole responsibility of the authors. To purchase a copy of this or other publications visit www.altamet.com.au
Celebrating 29 years of service to the global mining and metallurgical industry.
ALTA Metallurgical Services was established by metallurgical consultant Alan Taylor in 1985, to serve the worldwide mining, minerals and metallurgical industries. Conferences: ALTA conferences are established major events on the international metallurgical industry calendar. The event is held annually in Perth, Australia. The event comprises three conferences over five days: Nickel-Cobalt-Copper, Uranium-REE and Gold-Precious Metals. Publications: Sales of proceedings from ALTA Conferences, Seminars and Short Courses. Short Courses: Technical Short Courses are presented by Alan Taylor, Managing Director. Consulting: High level metallurgical and project development consulting.
ALTA Metallurgical Services Level 13, 200 Queen Street, Melbourne, Vic, 3000, Australia T: +613 8600 6909 | F: +613 9686 3008 | www.altamet.com.au ALTA Free Library www.altamet.com.au
ALTA Free Library www.altamet.com.au
PUBLICATIONS AVAILABLE FROM ALTA CONFERENCES & SHORT COURSES Visit www.altamet.com.au for details or email alta@altamet.com.au GOLD 2010 Gold - 1 Proceedings URANIUM 2010 Uranium-6 Proceedings 2009 Uranium-5 Proceedings 2008 Uranium-4 Proceedings 2007 Uranium-3 Proceedings 2006 Uranium-2 Proceedings 1997 Uranium Ore to Yellowcake Seminar NICKEL/COBALT-COPPER 2010 Ni/Co/Cu -1 Proceedings NICKEL/COBALT 2009 Ni/Co-14 Proceedings 2008 Ni/Co-13 Proceedings 2007 Ni/Co-12 Proceedings 2006 Ni/Co-11 Proceedings 2005 Ni/Co-10 Proceedings 2003 Ni/Co-9 Proceedings 2002 Ni/Co-8 Proceedings 2001 Ni/Co-7 Proceedings 2000 Ni/Co-6 Proceedings 1999 Ni/Co Forum & Symposium Proceedings 1998 Ni/Co Forum Proceedings 1997 Ni/Co Forum Proceedings 1996 Ni/Co Forum Proceedings 1995 Ni/Co Laterites — The How To's of Project Development SEMINARS 2000 Ni/Co-6 Laterite Deposits Seminar 2000 Ni/Co-6 SX Seminar —Fundamentals, Contactor Design, Application COPPER 2009 Copper-13 Proceedings 2008 Copper-12 Proceedings 2007 Copper-11 Proceedings 2006 Copper-10 Proceedings 2005 Copper-9 Proceedings including SX Fire Protection World Summit 2003 Copper-8 Proceedings 2002 Copper-7 Proceedings 2000 Copper-6 Proceedings 1999 Copper Forum & Symposium 1998 Copper Hydrometallurgy Forum & Symposium Proceedings 1997 Copper Hydrometallurgy Forum Proceedings 1996 Copper Hydrometallurgy Forum Proceedings 1995 Copper Hydrometallurgy Forum Proceedings SX/IX 2003 SX/IX World Summit 2000 SX/IX-1 Proceedings MATERIALS OF CONSTRUCTION 2003 Materials of Construction Proceedings SLIDES & NOTES FROM ALTA SHORT COURSES SX & Its Application to Copper, Uranium & Nickel-Cobalt Copper Oxide Ore Heap Leaching Testwork & Scale-Up Treatment of Nickel-Cobalt Laterites Uranium Ore Treatment A-Z of Copper Ore Leaching Copper SX/EW Basic Principles, and Detailed Plant Design
ALTA Free Library www.altamet.com.au
CONTENTS
Technology Developments Gold Technology Developments and Trends Alan Taylor – ALTA Metallurgical Services - Australia Mineralogy & Testwork Pitfalls to Avoid When Undertaking Metallurgical Testwork on Gold Ores Damian Connelly – Mineral Engineering Technical Services Pty Ltd (METS) - Australia The Importance of an Effective Core Logging Data Storage and Retrieving System in the Design of a Metallurgical Testwork Programme During Feasibility and Select Study Leon Lorenzen - Snowden Mining Industry Consultants Pty Ltd - Australia Upgrading, Gravity Treatment & Comminution Sensor Based Sorting - Gold Applications Lütke von Ketelhodt - CommodasUltrasort (Pty) Ltd - South Africa Geko Systems New Developments with the Inline Pressure Jig and Inline Leach Reactor Systems Michael O. Braaksma & Tim Smith – Gekko Systems - Australia The Deswik Fine Grinding Mill, Making the Fine Grind of Refractory Gold Ores and Retreatment of Waste Tailings Simpler Than Ever Before Stephen Massey – Deswik Mining Consultants - Australia Leaching Processes Online Cyanide Measurement and Control for Complex Ores Paul Breuer - Parker Centre (CSIRO Minerals Down Under Flagship) - Peter Henderson – Orica Limited - Australia Alternative Lixiviants to Cyanide Mark G. Aylmore - Bateman Engineering Pty Ltd - Australia Current Status of Thiosulfate Technology, and Future Prospects for Its Use Matthew Jeffrey - Parker Centre (CSIRO Minerals Down Under Flagship) - Australia Albion Process For Treatment Of Refractory Ores
Duncan W. Turner - Core Resources Pty Ltd & Mike Hourn - Xstrata Technology Australia Carbon Adsorption & Ion Exchange Review of Applications of SuperLig® Molecular Recognition Technology Products for the Gold Industry Neil E. Izatt, Steven R. Izatt, Ronald L. Bruening & John B. Dale - IBC Advanced Technologies - USA Trace Elements Deportment in Gold Process Solutions Jim Kyle, Vera Gella and Peter May – Parker Centre (Murdoch University) - Australia Carbon Activity Measurement. Rate Limiting Mechanisms R. Pleysier, P. Austin & C. Wingate - Parker Centre (CSIRO Minerals Down Under Flagship) - Australia Gold Ion Exchange Marthie Kotze – Mintek – South Africa New Selective Strong Base Anion Exchange Resins with Promise for Commercial Gold Cyanidation C. R. Marston & D. J. Gisch - Dow Water & Process Solutions - USA
ALTA Free Library www.altamet.com.au
Bio-Oxidation Bioleaching – Problems for the Mining State Carla Zammit, Francoise Sandow, Lesley Mutch & Elizabeth Watkin – Parker Centre (Curtin University); Helen Watling - Parker Centre (CSIRO Minerals Down Under Flagship) Australia Tank Bio-Oxidation Agitation Systems B. Gigas & R. Kehn – SPX Flow Technology-LIGHTNIN - USA Pressure Oxidation Acid Brick Lining Optimization for the Twin Creeks Mine Kevin Brooks – Koch Knight LLC - USA Gold Pressure Oxidation – An Update Karel Osten – Amec Minproc - Australia
ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
TECHNOLOGY DEVELOPMENTS
ALTA Free Library www.altamet.com.au
GOLD TECHNOLOGY DEVELOPMENTS AND TRENDS
By
Alan Taylor ALTA Metallurgical Services, Australia
Presented by
Alan Taylor alantaylor@altamet.com.au
CONTENTS
1. 2. 3. 4. 5.
INTRODUCTION APPLICATION OF HIGH PRESSURE GRINDING ROLLS (HPGR) ORE UPGRADING CYANIDE CONTROL, DESTRUCTION AND REPLACEMENT TREATMENT OF ORES CONTAINING SOLUBLE SULPHIDES AND COPPER 6 TREATMENT TREATMENT OF HIGH SILVER ORES 7 TREATMENT TREATMENT OF REFRACTORY ORES 8. IN-PLACE LEACHING 9. CONCLUSIONS 10. REFERENCES
ALTA Free Library www.altamet.com.au
2 2 3 5 5 6 7 8 9 9
1. INTRODUCTION
Recent gold industry news has understandably focused on the meteoric price rise. However, although processing technology has been out of the limelight, it has not been standing still, and developments have been occurring in a number of key areas. Drawing from published information and data, this paper identifies and outlines some of the main developments. 2. APPLICATION OF HIGH PRESSURE GRINDING ROLLS (HPGR) Following successful application in the copper industry, high pressure grinding rolls are making inroads into gold ore treatment, typically to supplement or replace SAG milling for coarse grinding of relatively low moisture fresh ores. A key factor in the increased use of HPGR for hard rock applications has been the development of metal carbide studs and tiles designed to produce an (1) autogenous wear layer . The main attraction of HPGR is higher energy efficiency, especially if operated in closed circuit with fine screening(2). Other advantages include reduced grinding media consumption and greater flexibility. A recent example is the massive Boddington Expansion Project in Western Australia, designed to treat 35 mtpa of copper/gold ore. The flowsheet, shown in Fig. (3) 1 , involves primary and secondary crushing, HPGR (supplied by Polysius), flotation of a saleable copper concentrate, and cyanidation of the flotation tailings to produce gold bullion.
Figure 1: Boddington Expansion Flowsheet (Ref. 3: Newmont/Boddington Gold Paper at World Gold 2007)
2
ALTA Free Library www.altamet.com.au
High pressure grinding rolls are also being considered for fine crushing in gold heap leach projects in place of cone crushers and vertical shaft impactors. PotentIal advantages of HPGR for heap leaching applications include the ability to achieve a fine procuct size at high energy efficiency together with micro-fracturing of the ore and increased fines production, which have the potential to significantly increase the leaching rate and gold recovery (1) (4). On the debit side, the increased fines may increase the need for cement addition for agglomeration. A conceptual flowsheet is shown in Fig. 2 which includes gold recovery by standard carbon adsorption technology. The main HPGR suppliers are Polysius (Thyssen Krupp), Köppern, and KHD Humboldt Wedag, all of Germany.
Figure 2: HPGR Heap Leach Flowsheet (Ref. 4: Orway Paper at HR Crushing & Grinding 2006) 3. ORE UPGRADING
3.1 UNDERGROUND PROCESSING Gekko Systems have introduced the “Python” Underground Processing Plant which is designed to upgrade run-of-mine ore underground resulting in savings in haulage, ventilation, back fill, grinding, (5) staffing, tailings disposal and environmental costs . The modular system typically consists of jaw crusher, vertical shaft impactor or HPGR, in-line pressure jig circuit and flash flotation. The concentrate representing 5-35% of the feed weight is either pumped or dewatered and placed in skips or trucks for cartage to the surface for further processing. A prototype 20 tph system has been operated at the Central Rand Goldfield in South Africa, and two 50 tph units are being added. An outline flowsheet is depicted in Fig. 3. The process is most applicable to ores which respond effectively to gravity and flotation.
3
ALTA Free Library www.altamet.com.au
Figure 3: Gekko Underground Processing (Python) Flowsheet (Ref. 5: Gekko/Central Rand Gold Paper at First International Future Mines Conference 2008) 3.2 ORE SORTING Optical ore sorting was tested in an 82 tph pilot plant by Commodas Ultrasort in 2004 for upgrading a low grade waste rock dump at the Kloof Gold Mine in South Africa (6). More recent optical sorting tests were carried out for Central Rand Gold. The Kloof pilot plant, shown in Fig. 4, included a feed hopper, variable speed feeder, conveyors, washing screen, water circulation pump as well as the sorter.
Figure 4: Optical Ore Sorter Pilot Plant at Kloof Gold Mine (Ref. 6 Commodas Ultrasort Paper World Gold 2009) The program focused on the plus 16 mm size material. Pilot plant operating data is presented in Table 1. About 70% of the gold in the plus 16 mm material was recovered into a relatively low mass pull concentrate of 5–10% of the feed weight.
Table 1: Pilot Plant Operating Data Grades and Recovery (Ref. 6: Commodas Ultrasort Paper World Gold 2009)
4
ALTA Free Library www.altamet.com.au
CommodosUltrasort have also tested XRT sorting on rejects from the DMS treatment of sulphide ore at Simmer and Jack’s TGME operation in South Africa(21). 4. CYANIDE CONTROL, DESTRUCTION AND REPLACEMENT
4.1 ON-LINE CYANIDE ANAYSIS On-line cyanide analysis and automatic control are being increasingly used to optimize cyanide addition, and CNWAD on-line analyzers for weak acid dissociable cyanide have recently been (7) introduced . As more operations are adopting the International Cyanide Management Code, it has become increasingly important to monitor and log residual cyanide. 4.2 CYANIDE DESTRUCTION CyPlus have supplied two Cold Caro’s Acid systems for Jaguar Mining Inc. at their Turmalina and Paciencia gold operations in Brazil in 2009(17). The process is said to very economical with regard to hydrogen peroxide and sulfuric acid consumption. Another recent development is the SART Process (Sulphidization–Acidification–Recycle– Thickening) in which a reagent such as sodium hydrosulphide (NaHS) is used to precipitate copper and zinc as sulphides and convert cyanide to HCN. The precipitate is removed for possible sale or further processing, and the solution is neutralized with sodium hydroxide or lime and recycled back to the leaching process to re-use the cyanide. A simplified flowsheet is shown in Fig. 5. Biogenerated H2S gas can be used in place of sodium hydrosulphide, as discussed in Section 5.1 (8) below. A SART facility has been operated at the Newcrest Telfer plant in Western Australia .
Figure 5: Simplified SART Flowsheet 4.3 NON-CYANIDE LIXIVIANTS In the field of non-cyanide lixiviants, the Parker Centre (CSIRO) in Perth is developing ion exchange technology for recovering gold from thiosulphate leach solution(9). The process involves the use of a synergistic mixture of chloride and sulphite to elute the gold from the loaded resin. Thiosulphate leaching has potential applications for locations where cyanide is prohibited, and for processing preg-robbing ores for which it can yield a higher gold recovery than cyanide. The European parliament is seeking a European Union (EU) ban on using cyanide technologies in mining before December 2011. If this goes ahead, efforts to develop viable options to cyanide will likely increase.
5. TREATMENT OF ORES CONTAINING SOLUBLE SULPHIDES AND COPPER
5.1 COPPER SULPHIDE PRECIPITATION USING BIO-GENERATED H2S BioteQ in Canada and Paques in the Netherlands have developed BIoSulphide and THIOTEQ bioreactor technology respectively for the generation of H2S gas for the precipitation of copper sulphide from copper rich cyanide solutions from a sulphur source such as elemental sulphur. The H2S is used instead of sodium hydrosulphide in a SART type flowsheet to regenerate and recycle cyanide
5
ALTA Free Library www.altamet.com.au
and yield a copper sulphide by-product. The additional capital cost is said to be outweighed by the lower operating costs due to the cheaper sulphur source and reduced acid consumption. The benefits tend to increase with plant throughput and feed solution copper content. Paques’ THIOTEQ technology together Outotec’s OKTOP reactors is being installed at the Pueblo Viejo operation in the Dominican Republic, owned by Barrick and Goldcorp(10), while BioteQ’s BIOSULPHIDE technology has been applied at the Lluvia de Oro operation in Mexico. A flowsheet for the BioteQ Biosulphide Process is given in Fig. 6
Figure 6: Bioteq BioSulphide Process Flowsheet (Ref. 8: BioteQ Paper at Precious Metals 07) 5.2 SELECTIVE ION EXCHANGE Anglo Asian Mining has adopted Mintek developed IX technology using DOWEX MINIX strong-base resin for the Gedabek gold-copper heap leaching operation in Azerbaijan(11). This is the first application of the resin which is very selective for gold in the presence of high copper levels. It has been previously applied for the recovery of gold from carbonaceous preg-robbing ores at Avocet Mining’s Penjom mine in Malaysia. 5.3 ADDITION OF LEAD NITRATE The AngloAshanti Morila Mine in Mali has reported that the addition of lead nitrate for treating ore containing the reactive sulphide pyrrhotite improves gold recovery, enhances leach kinetics and (12) reduces cyanide consumption . Also, Mineral Engineering Technical Services (METS) of Perth have reported improved leach performance with the addition of lead nitrate when testing a gold(13) silver ore .
6. TREATMENT OF HIGH SILVER ORES CANMET, Canada, have introduced CELP (CANMET Enhanced Leaching Process) aimed at (14) reducing cyanide consumption and leaching time for ores with a silver content above 50 g/t . The first commercial installation came on stream in 2008 at Kupol in Far East Russia, now owned by Kinross Gold (15), treating 3,000 tpd high grade ore with initial grades of 28.2 g/t Au and 324 g/t Ag. The inclusion of CELP allows the residual cyanide concentration to be reduced to 410 ppm which eliminated the need for an AVR plant to recover and recycle cyanide. Gold and silver recoveries in 2008 were 95.4% and 85.6% respectively with a cyanide consumption of 1.3 kg/t. Performance data is given in Table 2. Refractory sulphide and antimony silver minerals are said to be efectively leached because CELP produces more oxidized mineral surface relative to conventional cyanidation. It is reported that CELP uses standard agitated leach tanks. The Kupol operation includes a calcium hypochlorite cyanide destruction facility as the efluent is regulated on total cyanide and thiocyanate concentrations.
6
ALTA Free Library www.altamet.com.au
Table 2: Summary of 2008 Kupol Performance with CELP (Ref. 15: Kinross/CANMET Paper World Gold 2009)
7. TREATMENT OF REFRACTORY ORES 7.1 LEACHOX PROCESS The Leachox Process has been has been adopted for Vasgold’s Vasilkovskoye project which is under construction in Kazakhstan, and for Banro’s planned Tangiza project in the DRC. Previous applications of Leachox technology include two refractory gold operations in South Africa. The Leachox Process is a development of Maelgwyn Mineral Services (MMS) in the UK, aimed (16)(17) particularly at lower grade refractory deposits . Gold recoveries are said to be lower than with roasting and pressure oxidation, but similar to to bio-oxidation, while capital and operating costs and plant footprint area are said to be reduced. In the process, a concentrate is produced by MMM centrifugal pneumatic flotation (“G-Cell”) cells or gravity concentration, followed by ultra-fine grinding in Deswick vertical stirred bead mill mills. Low pressure partial oxidation with oxygen is carried out in MMM (“Aachen) in-line type gas-liquid mass transfer reactors, and the gold is recovered in MMS cyanide leach columns and resin ion exchange columns.
Figure 7: Schematic of MMS Archen Reactors and Leach Columns (Ref. 17: Mining Magazine Jan./Feb. 2010)
7.2 ALBION PROCESS The Albion Process has been selected for two refractory gold projects - Enviro Gold at Las Lagunas in the Dominican Republic and European Goldfield at Certej in Romania. Albion is a relatively low capital cost process using ultrafine grinding with Isamills to produce an activated finely ground concentrate. This is followed by hot oxidative leaching at atmospheric pressure with oxygen
7
ALTA Free Library www.altamet.com.au
(18)
sparging in conventional agitated tanks in alkaline conditions and CIL . An outline flow diagram is shown in Fig. 8. The technology is owned by Xstrata (formerly MIM) and Highlands Pacific/OMRD (a Japanese consortium), with CORE Resources, Queensland, as the exclusive global marketing agent.
Figure 8: Albion Outline Process Flow Diagram (Ref. 18: Xstrata/Core/Aker Paper at Randol Forum Perth 2005) 7.3 ROASTING, PRESSURE OXIDATION & BIO-OXIDATION The more established roasting, pressure oxidation and bio-oxidation processes continue to find favour. Resolute Mining, Perth, commissioned the Syama Project in Mali in 2009 iwhich included roasting, and Agnico-Eagle of Canada brought the Kitilla pressure oxidation operation in northern Finland on stream in 2009. In the field of bio-oxidation the Kokpatas plant in Uzbekistan came into operation in 2008 using Gold Field’s BIOX technology, while the rival BacTech technology was used for the expansion of the BioGold operation in China in 2007.
8. IN-PLACE LEACHING The Parker Centre (CSIRO) in Perth is undertaking initial investigations for the in-place treatment of oxidized gold deposits(19. Because of environmental concerns, only non-cyanide lixiviants are being considered. Bench scale testwork has identified two promising systems - sodium thiosulphatethiourea-ferric EDTA (ferric ethylenediamine) in which thiourea is a catalyst for gold oxidation and
8
ALTA Free Library www.altamet.com.au
ferric EDTA an oxidant, and iodide–iodine, in which iodine is an oxidant. Testwork has shown that these lixiviant tend to break down in contact with pyrite, so the current focus is on pyrite free oxidized deposits. The permeability of these deposits is consifered to be generally too low for true in-situ leaching, and permeability enhancement methods are condidered to be necessary (which is why the process is designated as in-place rather than in-situ). Possible options include blasting and hydraulic fracturing. The results to date are sufficiently encouraging to warrant field studies, which could start in 2011 with the support from government and mining companies. “In-situ” processing has also been suggested for possible application in back-filled Witwatersrand gold mines in South Africa(20). (As with the Parker Centre program, in-place would be beter terminology.) 9. CONCLUSIONS
Gold ore processing technology is progressing in a number of key areas. Driving forces include the trend towards the treatment of lower grade, carbonaceous, copper bearing and refractory ores, and increasing environmental pressure against the use of cyanide including calls for a total ban in Europe and elsewhere. Developments include the application of HPGR, ore upgrading, gravity and flotation, cyanide monitoring and control, application of ion exchange, alternative lixiviants, ultrafine grinding, various oxidation processes and in-place leaching. These are being encouraged by the significant rise in the gold price. 10. REFERENCES 1. C. Morley, Fluor Signet Engineering, “HPGR in Hard Rock Applications”, Mining Magazine, Sept. 2003. 2. H. Von Michaelis, Randol International, “How Energy Efficient is HPGR?” Proceedings of World Gold 2009, Oct. 2009, Gauteng, South Africa. 3. R. Dunne et al, Newmont Mining and Boddington Gold, "Boddington Gold Mine – An Example of Sustaining Gold Production for 30 Years”, Proceedings of World Gold 2007, Oct. 2007, Cairns, Australia. 4. B. McNab, Orway Mineral Consultants, “Exploring HPGR Technology For Heap Leaching Of Fresh Rock Gold Ores”. Proceedings of HR Crushing & Grinding Conference 2006, Mar. 2006, Townsville, Australia. 5. T. Hughes, Gekko Systems and G. Cormack, Central Rand Gold, “Potential Benefits of Underground Processing for the Gold Sector – Conceptual Process Design and Cost Benefits”, Proceedings of First International Future Mines Conference, Nov. 2008, Sydney, Australia. 6. L. Von Ketelhodt, Commodos Ultrasort, “Viability of Optical Sorting of Gold Waste Rock Dumps”, Proceedings of World Gold 2009, Oct. 2009, Gauteng, South Africa. 7. T. Mulpeter, Sadiola Hill Gold Mine and D. Kotzen, Process Analytical Systems, “Online Cyanide Analysers Optimise Cyanidation and Ensure Cynide Code Compliance fro Plant Tailings”, Proceedings of World Gold 2007, Oct. 2007, Cairns, Australia. 8. M. Adams, and R. Lawrence, BioteQ, “Biogenic Sulphide For Cyanide Recycle and Copper Recovery In Gold-Copper Ore Processing, ” Proceedings of Precious Metals 2007, Brisbane, Australia, Aug. 2007. 9. R. Thyer, CSIRO, “Together Two Compounds Answer Gold Recovery Woes”, CSIRO Process Magazine, Feb. 2010. 10. L. Williams, “”Barrick’s $3 Billion Pueblo Viejo Gold Mine Buys Finish Copper Recovery Plant”, International Mining, Oct. 2009. 11. “Mintek ion exchange resin aids gold recovery in Azerbaijan”, International Mining, Dec. 2009. 12. T. Mahlangu et al., Univ. of Pretoria and Anglo Gold Ashanti, “Effect of Preoxidation and Lead Nitrate Addition on Sodium Cyanide Consumption and Gold Recovery at the Morila Mine Gold Plant”, Proceedings of World Gold 2007, Oct. 2007, Cairns, Australia.
9
ALTA Free Library www.altamet.com.au
13. A. Voung, METS, “”Enhanced Gold Leaching”, Mining Mirror, Oct. 2009. 14. “Help from CELP”, CIM Magazine, Feb. 2009. 15. J. Rajala, Kinross Gold and G. Deschenes, CANMET, “Extraction of Gold and Silver at the Hupol Mill using CELP”, Proceedings of World Gold 2009, Oct. 2009, Gauteng, South Africa. 16. J. Chadwick, “Bigger and Better”, International Mining, Apr. 2007. 17. “Management in Action – Gold Processing”, Mining Magazine, Jan./Feb. 2010. 18. M. Hourn et al Xstrata Technology, Core Resources and Aker Kvaerner (now Aker Solutions), “Benefits of Using The Albion Process for a North Queensland Project, and a Case Study of Capital and Operating Cost Benefits Versus Bacterial Oxidation and Pressure Oxidation”, Randol Innovative Metallurgy Forum, 2005, Perth. Australia. 19. P. Roberts et al, Parker Centre CSIRO, “In place leaching of oxidised gold deposits. A new method for recovering stranded gold resources?”, Proceedings of World Gold 2009, Oct. 2009, Gauteng, South Africa. 20. N. Sreven, Rockwater Consulting, “Potential in situ leach exploitation of back-filled Witwatersrand gold mines: parameters and flow-rate calculations from a Zambian Copperbelt analogue”, Proceedings of World Gold 2009, Oct. 2009, Gauteng, South Africa. 21. Simmer & Jack; 2008; Developing new Gold Opportunities in South Africa’s oldest Goldfield, Presentation at the Mining Indaba, 5th February 2008.
10
ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
MINERALOGY & TESTWORK
ALTA Free Library www.altamet.com.au
PITFALLS TO AVOID WHEN UNDERTAKING METALLURGICAL TESTWORK ON GOLD ORES By Damian Connelly Mineral Engineering Technical Services Pty Ltd (METS), Australia
Presented by Damian Connelly damian.connelly@mets.net.au
CONTENTS ABSTRACT
3
1. INTRODUCTION
3
2. ROLE OF METALLURGIST
3
3. THE DESIGN PROCESS
4
4. STUDIES
4
5. SCHEDULE AND BUDGET
4
6. SAMPLE SELECTION
5
7. GEOMETALLURGY
5
8. HEAD ASSAY
5
9. MINERALOGY
5
10. GOLD IS DIFFICULT
6
11. COMMINUTION
6
12. GRAVITY GOLD
6
13. EXTRACTION
7
14. OXYGEN UPTAKE RATE
8
15. HEAP LEACHING
9
16. FLOTATION
9
17. REFRACTORY GOLD
9
ALTA Free Library www.altamet.com.au
CONTENTS (CONT.)
18. ADSORPTION
10
19. STRIPPING OR ELUTION
11
20. ELECTROWINNING
11
21. RESINS
11
22. SITE WATER
11
23. VARIABILITY
12
24. COMMINUTION
12
24. BOTLE ROLLS
12
26. WHERE THINGS WENT WRONG
12
27. CONCLUSIONS
13
28. ACKNOWLEDGEMENTS
13
29. REFERENCES
13
2
ALTA Free Library www.altamet.com.au
ABSTRACT Gold is showing quite a surge in price and interest of late. While everyone is trying to cash in with the current high price, there are some issues to be aware of when evaluating a gold deposit. Mineralogy is the most important characteristic as it drives the whole process route. Free, coarse metallic gold is easily treated as opposed to complex sulphides. Gold tellurides or gold locked in arsenopyrite, commonly known as refractory deposits, require fine grinding and oxidative techniques to expose the gold. Head assays are important in terms of ore value (gold and silver); however trace elements like copper, nickel, zinc and mercury are also important to track and be aware of. High levels of coarse, free gold are particularly difficult to sample and can create sampling and representivity issues. Bulk leaches or screen fire assays technique may be of consideration. The ore characterisation is very important. Ores may range from highly weathered oxides through to fresh primary sulphides, with a transition material in between. Each ore type may require a significantly different process to treat efficiently. Comminution deals with crushing and grinding to the target size to liberate or expose the gold so it can be extracted. Selecting the correct crushing and milling circuit for a particular ore, or range of ores, is critical for the success of the project. Leach extraction will determine the overall recovery and success of the project. The use of leaching enhancers (PbNO3, O2, peroxide) may be required in the presence of some reactive sulphide ore types to promote the leaching kinetics. Reactive sulphides (arsenopyrite, pyrrhotite, marcasite) can consume oxygen from the slurry, which is essential for the gold leaching reaction. Once gold has been dissolved it needs to be adsorbed onto carbon before it can be electro won and smelted into dorÊ bullion. Cyclic or Sequential Carbon Loading is required to determine the loading capacity and rate. This test can also highlight potential issues with competing ions (Cu, Ni, and Zn) or fouling potential (organic oxalate and humates) which may impact on the amount of carbon required to recover the amount of gold leached. Tailings characterisation is required to maximise the amount of tailings that can be contained in a tails dam. Sulphide ore types may require acid generation and acid neutralising capacity tests to ensure acid mine drainage issues do not occur within the tailings dam. The water source is important, especially when using hyper-saline water which will buffer at a lower pH than is optimum for cyanide leaching (typically from 8.5 – 9.0). As well as resulting in high lime consumption, the formation of HCN in the leaching process requires careful management from an operational OHS. The high lime usage may also create scale (CaCO3 and CaSO4) formation which will need to be managed in the circuit. These important issues, if not addressed at an early stage, may cause critical problems during the processing of the gold ores being evaluated. Some examples will be presented where issues were not addressed and problems occurred.
1. INTRODUCTION There are a number of gold projects where the underperformance or failure could be traced back to poorly executed metallurgical testwork. This paper outlines the minimum standards of fundamental areas of testwork that must be evaluated in order to diminish the likelihood of process issues.
2. ROLE OF METALLURGIST The Metallurgist adds value to a project by using extraction and beneficiation technology which involves developing Test plans, project managing the testwork and interpreting the results. This covers the unit operations of crushing, grinding, gravity, leaching, adsorption, stripping and electrowinning.
3
ALTA Free Library www.altamet.com.au
3. THE DESIGN PROCESS No two plants are the same and each ore is unique. The design of a process plant is time consuming and complex and includes many disciplines such as civil, structural, mechanical, electrical, piping, instrumentation and control. The design evolves based on testwork results, the flowsheet, mass balance, equipment list, piping and instrumentation diagrams (P&ID’s), process control philosophy etc. There are crucial early decisions that need to be made and plant safety depends on the basic design and layout.
4. STUDIES The typical progression timeline is a Scoping Study followed by Pre Feasibility Study (PFS) and a subsequent Feasibility Study (See Figure 1). A more detailed Definitive Study can follow if the owner requests higher project definition requiring significantly more engineering. This occurs when a project is known to be feasible and is well defined. The Feasibility Study has one clear goal which is to demonstrate that the project is technically sound and economically viable if designed, procured, constructed and operated in accordance with the parameters outlined in the study. All major decisions concerning the project are made during the Feasibility Study. Typical costs for undertaking Greenfield Project Feasibility Studies are 3% of the project capital cost (CAPEX) for large projects and up to 15% for Brownfield Projects where the cost is much less. Feasibility Studies should be accurate to within +/-10%.
Figure 1: Decreasing ability to impact on the project
The Stage Gate process allows clearly defined study milestones to re-evaluate strategic directions and development options for the project (Figure 1). Geology is usually well advanced compared to the metallurgy because the initial exploration effort examines resources, mineralogy and assay data very early in the project. Unless a resource is established no study can commence. While undertaking studies is costly it adds value to the project by a multiplier effect. For example selling a project with a positive Feasibility Study will provide a much higher return than selling at the Scoping Study stage.
5. SCHEDULE & BUDGET A budget should be based on actual quotes from at least two laboratories and include a contingency for additional testwork. Typically, a schedule should be developed up front and it is quite common for PFS level gold testwork to take at least three months to complete.
4
ALTA Free Library www.altamet.com.au
6. SAMPLE SELECTION
METALLURGY SAMPLES DDH CROSS SECTION
DDH3
DDH2
DDH1
Surface v1
v2v
3
v4
Oxide
Transition
Primary Include Dilution
Waste
Include dilution waste
Oxide Ore Zone Transition Ore Xone
VariabilitySamples 1 metre sections v1,v2,v3
Primary Ore Zone
Figure 2: Drill Hole Locations The samples must be representative with separate composites for oxide, transition and primary ore. Sample selection should be undertaken with the Geologist and should also include 10% waste to cover dilution. (Figure 2)
7. GEOMETALLURGY During feasibility studies, the improved ore characterisation combined with the spatial modelling of critical physical characteristics that forms the basis of a geometallurgical approach provides a much improved basis for operational and mineral processing plant design. This approach reduces the risk associated with developing new operations or expanding an existing business. By increasing ore body knowledge and then using it to design the entire operational flowsheet (from in situ rock to end product) a more reliable and realistic model of the production capacity of the system can be developed. Key bottlenecks, product constraints and low-cost/high value opportunities can be identified. The result can be increased total metal recovery and improved asset utilisation. This in turns allows better decision making, in terms of capital expenditure, to optimise project economics.
8. HEAD ASSAY The head assay should include gold, silver, sulphur and a full elemental scan. The presence of copper, mercury, arsenic or antimony should warrant further investigation of their impact.
9. MINERALOGY The difficulty with gold is to find visible gold and determine if it is typical. The Scanning Electron Microscope (SEM) is very good for detecting and locating gold. Similarly, diagnostic leaching with optical microscopy to determine the nature and occurrence of the gold will provide deportment of the gold which is free, associated with sulphides or locked in gangue.
5
ALTA Free Library www.altamet.com.au
10. GOLD IS DIFFICULT The ppm levels of gold present makes gold assaying difficult. Similarly the “nuggett effect� can make achieving a mass balance very problematic. Bulk leaches, Leachwell and screen fire assays are useful where this is a problem.
11. COMMINUTION A number of circuit alternatives have been put forward for possible inclusion. These are listed below: 1
Three stage crushing/single stage ball milling
2
Rod/ball milling
3
Semi Autogenous Grinding (SAG)/ball milling
4
SAG mill/ball mill/crusher (SABC)
5
Autogenous mill/ball mill/crusher (ABC)
6
Two stage autogenous grinding
7
Autogenous mill/pebble milling
8
High pressure grinding rolls (HPGR)/ball milling
9
Lump mill/pebble mill/ crusher (to control pebble make).
10
Two-stage crush/SAG mill/ball mill
11
Single Stage SAG milling
The tests required are Crushing Work Index, Rod Mill Work Index, Bond Ball Mill Work Index, Abrasive Index, Unconfined Compressive Strength and SMC or Drop Weight test.
12. GRAVITY GOLD
Figure 3: Knelson Concentrator The standard Gravity Recoverable Gold (GRG) test is mandatory. Interpreting the results can be a challenge. Maximising gravity gold recovery results in higher overall recovery, faster cash flow and lower costs per ounce to produce. Figure 3 reveals a Knelson Concentrator.
6
ALTA Free Library www.altamet.com.au
13. EXTRACTION CIL TEST DETAILS AND RESULTS TEST OBJECTIVE
TEST PARAMETERS
To show the rate and extent of gold extraction.
Grind size Water type Water S.G.
m
75 Tap 1.00
g/mL
TEST DATA
CONTENTS/ADDITIONS
SOLUTION pH
TIME
Solids
Water
NaCN
Lime
Oxygen
hours
g
mL
g
g
mg/L
Initial
1000
1000
0 1
0.50 0.22
0.5 0.2
2.9 5.1
2 4
0.18 0.18
0.2 0.1
6 24 48
0.18
0.1 0.1
RESIDUE NaCN
Au
Au
Final
%
mg/L
g/t
Au Ext. %
7.5 9.0
9.6 9.6
0.050 0.024
0.00 1.71
26.60 14.50
0.0 45.8
4.8 5.0
9.2 9.3
9.7 9.7
0.034 0.036
1.52 0.67
10.45 4.92
61.0 81.6
5.3 5.9
9.4 9.1
9.8 10.0
0.042 0.012 0.008
0.32 0.07 0.07
3.97 2.72 2.71
85.2 89.8 89.9
GOLD BALANCE
MATERIAL
Solids Solution Carbon
g mL g
QTY
1000 1000 20.34
GOLD ppm
g
%
2.71 0.070 1180
2705 70 24001
10.1 0.3 89.6
26776
100.0
Total Calc.'d Head
26.78
Assay Head
26.60
REAGENTS Cyanide Consumption
kg/t
1.10
Lime Addition
kg/t
1.23
Figure 4: Cyanide Leach Test Sheet A standard cyanide leach test is included in Figure 4 and measures the leach kinetics, final gold recovery and reagent consumption. Lead nitrate is worth testing on primary ores at low ppm levels to establish if there is any benefit.
7
ALTA Free Library www.altamet.com.au
14. OXYGEN UPTAKE RATE
TEST DATA - Dissolved Oxygen (DO), mg/L Interval
Monitoring Time ( hr ) 2 3 4 5
min
0
1
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
0.7 0.5 0.5 0.4 0.4 0.4 0.4 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.5
5.4 5.3 5.4 5.4 5.4 5.4 5.4 5.4 5.3 5.3 5.3 5.3 5.2 5.2 5.2
5.3 5.5 5.5 5.4 5.4 5.4 5.3 5.3 5.3 5.3 5.3 5.2 5.2 5.2 5.2
5.9 5.8 5.7 5.8 5.8 5.8 5.7 5.7 5.6 5.7 5.7 5.7 5.6 5.6 5.5
5.9 5.9 5.8 5.8 5.8 5.7 5.7 5.7 5.6 5.7 5.6 5.6 5.6 5.6 5.7
0.0143
0.0189
0.0196
0.0204
Oxygen Uptake Rate : 0.0100 ( mg/L/min )
6
24
6.3 6.3 6.2 6.1 6.1 6.1 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0 6.0
6.5 6.2 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1 6.1
5.9 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8 5.8
0.0207
0.0121
0.0025
OXYGEN DEMAND CURVE
Oxygen Uptake ( mg/L/min )
0.03
0.02
0.02
0.01
0.01
0.00 0
1
2
3
4
5
6
7
Time ( hr )
Figure 5: Oxygen Uptake Rate Results The oxygen demand of some ores will affect the leach kinetics. Measuring the oxygen uptake rate is mandatory (Figure 5).
8
ALTA Free Library www.altamet.com.au
15. HEAP LEACHING Heap Leaching is high risk. Preliminary crush size sensitivity tests followed by percolation tests and a thirty day column test will give a preliminary indication of the likely gold recovery, reagent consumption, slumping and leach kinetics.
16. FLOTATION Flotation is a method to selectively concentrate and separate valuable materials from gangue minerals on the basis of their surface properties (See Figure 6). Reagents are utilised to vary the surface chemistry (wettability) properties of the minerals. The hydrophobic (water repelling) particles will attach to air bubbles and separate into a concentrate, whilst the hydrophilic (water attracting) particles will stay in suspension. Flotation is a common method used for treating sulphide minerals.
100 80
Cum Au rec (140 um)
60
Cum Au grade (140 um)
40
Cum Au rec (75 um)
20
Cum Au grade (75 um)
%
0 0
5
10
15
20
Time, min
Figure 6: Flotation Results
17. REFRACTORY GOLD The typical options to consider and relative advantages are shown in Table 1.
9
ALTA Free Library www.altamet.com.au
Table 1: Refractory Gold Alternatives ASPECT
ROASTING
BACTERIAL LEACHING
PRESSURE OXIDATION
Technical Risk Environmental Risk
Low Highdepending on gas clean up Prohibitive Low Not with gas cleaning low
High Low
High Lowest
High High Low High High Low No complex No materials of Simple plant construct High High Low
Medium Low
High Very High
High High
Modest Medium
Typical
Exotic
Exotic
Typical
Simple
Complex
Complex
Simple
Sulphur level
Power cost/Au:S ratio
History
Tried & true
Poor, failures
Operability
Simple
Difficult, learning Curve applies
Recovery
Very good
Subject to “bug Health”
Operating Power cost pressure & temperature Difficult Good, growing Fine, getting Good, wear better all the issues time Power cost impt Excellent Grind sensitive
Capital Cost Operating Cost Simplicity Acid neutralisation Cyanide use Power consumption Materials construct Process Chemistry Major Driver
ULTRA FINE GRINDING Low Low
The options to test are roasting, bacterial leaching, pressure oxidation, Albion and ultrafine grinding.
18. ADSORPTION
CARBON KINETIC ACTIVITY FLEMING MODEL EQUATION
8.0
Ln [Au]c
[Au]c=k x [Au]s x t
n
FLEMING MODEL CONSTANTS k (hr-1) n r
188 0.36 0.993
intercept @ 1hr slope
0.3610688 7.6819613 0.0180367 0.0133715 0.9925696 0.0258896
7.0 -2.0
-1.0
0.0 Ln t
1.0
k - indicates the emperical rate constant n - indicates the carbon loading capacity
Figure 7: Cyclic Carbon Loading Results Cyclic carbon loading testing to determine k and n is mandatory with gold ores to size the CIL or CIP tanks (See Figure 7).
10
ALTA Free Library www.altamet.com.au
Assaying of the loaded carbon by full elemental scan is also recommended to establish what species are on the carbon.
19. STRIPPING OR ELUTION Stripping is not normally tested. If carbon management by way of acid washing and regeneration is used this is not likely to be an issue. The choice of pressure Zadra, Anglo or Integrated Pressure Strip (IPS) is more critical depending on water quality and quantity available.
20. ELECTROWINNING Electrowinning is not normally tested; however an analysis of the pregnant liquor or loaded carbon will highlight likely issues. There has been a trend to stainless steel cathodes and self sludging cells.
21. RESINS While resins have been used for uranium for many years their use in the gold industry has seen both success and failure. While this is emerging technology it would not be recommended for new projects.
22. SITE WATER
10 9 pH
8 7 6
0
5
10
15 20 kg Lime/t water
25
Figure 8: Buffering Response Site Water Water hardness can impact on alkali consumption. This can result in high cyanide gas levels over the tanks. It is not uncommon for hyper saline water to be used. Site water must be tested using the final flowsheet as it can impact on viscosity, leach kinetics, thickening and adsorption characteristics. (Figure 8)
11
ALTA Free Library www.altamet.com.au
23. VARIABILITY
Figure 9: 50 Bottle Rolls Throughout The Ore body
24. COMMINUTION Undertake BWI on samples selected spatially throughout the ore body to confirm the range of BWI to be encountered.
25. BOTTLE ROLLS The samples should have designated coordinates so the recovery can be assigned to a location in the block model and used in the mine schedule. (Figure 9) This data will be used in conjunction with the variability data and support the recovery predictions used for a particular ore body. By storing the data electronically we can manipulate the data and evaluate various scenarios based on comminution and ore recovery. Samples should be obtained throughout the oxide, transition and primary zones and If with greater definition such as coffee rock, pisolites, saprolite etc. The impact of recovery with depth would be useful to understand if this can be tested in the primary core.
26. WHERE THINGS WENT WRONG The following list identifies failures or omissions with the gold ore testwork: Mistake 1 - Failure to obtain representative ore samples. Mistake 2 - Failure to fully comprehend the mineralogy and variability Mistake 3 - Insufficient bench scale or pilot metallurgical testwork. Mistake 4 - Failure to characterise the ore and interpret the results Mistake 5 - Testwork did not cover the limits Mistake 6 - Skipping Stage Gate Studies
12
ALTA Free Library www.altamet.com.au
Mistake 7 - Failing to run the “what happens if scenarios” Mistake 8 - Failure to address operability and maintainability Mistake 9 - Failure to optimise comminution Mistake 10 - Failure to optimise flotation Mistake 11 - Failures with leaching Mistake 12 - Failures with solid liquid separation Mistake 13 - Failures with roasters and calciners Mistake 14 - Failures with pumping, materials transport, drying and storage Mistake 15 - Tailings disposal, water balance Mistake 16 - Lack of continuity with project management, data management Mistake 17 - Failures with material of construction Mistake 18 - Failures start up training and commissioning Mistake 19 - Failure to undertake a reasonableness check Mistake 20 - Thinking problems can be solved once production starts Mistake 21 - Failure to define Scope of Work 27. CONCLUSIONS It is impossible to assume all risks can be eliminated; however, if we are to progress, the “lessons learned” need to be carried forward for the benefit of future projects. Metallurgists have their limitations with solving technical issues and the classification of “good ore” and “bad ore” highlights this. Complex sulphide ore, refractory gold ores are not simple problems to solve; hence the JORC code which correctly applies modifying factors to ores that may in fact be waste because of process metallurgical difficulties. This paper has examined some of the common metallurgical mistakes in developing resource projects. There is no simple solution and each project brings its own unique challenges. The successful metallurgical programme requires geometallurgy with respect to sample selection, a thorough understanding of the mineralogy and carefully thought through metallurgical test work to ensure all aspects are covered. Several iterations of ore characterisation and interpretation of the test results are required to develop the Process Design Criteria and Process Flowsheet. A “what happens if” scenario should be applied and mitigation solutions developed for all resource projects at every study stage gate.
28. ACKNOWLEDGEMENTS The author would like to thank various companies, all colleagues, engineers at various sites, METS staff and other consultants for their contribution and the management of METS for their permission to publish this paper and constructive criticism of various drafts.
29. REFERENCES Angove, J, 2005. Metallurgical Testwork: Gold Processing options, physical ore properties and cyanide management, In Developments in mineral processing, Volume 15, Elsevier. Deschênes G and Guo H, 2005 Cyanidation of Gold, Ottawa, Mining and Mineral Sciences Laboratories CANMET. Fleming, C A, 2005 CIP/RIP/Electrowinning, SGS Lakefield Research Limited, Lakefield, Canada. Lastra, R, 2005. Mineralogy of Gold Ores, In Gold Extraction Short Course, COM. Lunt, D and Weeksm, T, 2005. Process Flowsheet selection, In Developments in mineral processing, Volume 15, Elsevier. Merrill-Crowe Plants, Denver Mineral Engineers, [online] Available from: <http://www.denvermineral.com/merril~1.html> [Accessed 26 March 2010]. Mineralogy Database, [online] Available from: <http://www.mindat.org/> [Accessed 26 March 2010.
13
ALTA Free Library www.altamet.com.au
PSA Oxygen Generator, MSA Engineering Ltd, [online] Available from: <http://www.mvsengg.com/psa-oxygen.htm> [Accessed 26 March 2010]. Ryan, A et. Al, 2005. Feasibility study plant design, In Developments in mineral processing, Volume 15, Elsevier. Spry, P, Chryssoulis, S and Ryan, C, 2005. Process Mineralogy of Gold: Gold from TellurideBearing Ores, Journal of Metals, 56(8). The Mineral Sylvantite, Amethytst Galleries, [online], Available from: <http://www.galleries.com/minerals/sulfides/sylvanit/sylvanit.htm> [Accessed 26 March 2010]. Vaughan, J P, 2004. The Process Mineralogy of Gold: The Classification of Ore Types, Journal of Metals, 56(7). Vaughan, J Pand Kyin, A, 2004. Refractory gold ores in Archaean greenstones, Western Australia: mineralogy, gold paragenesis, metallurgical characterization and classification, Mineralogical Magazine, 68(2), pp. 255-277. Yannopoulos, J C, 1991. The Extractive Metallurgy of Gold, Van Nostrand Reinhold, New York.
14
ALTA Free Library www.altamet.com.au
THE IMPORTANCE OF AN EFFECTIVE CORE LOGGING DATA STORAGE AND RETRIEVING SYSTEM IN THE DESIGN OF A METALLURGICAL TESTWORK PROGRAMME DURING FEASIBILITY AND SELECT STUDY By Dr Leon Lorenzen Snowden Mining Industry Consultants Pty Ltd, Australia
Presented by Dr Leon Lorenzen LLorenzen@snowdengroup.com
CONTENTS
1. 2. 3. 4. 5.
INTRODUCTION GOLD CASE STUDY URANIUM CASE STUDY CONCLUSION REFERENCES
ALTA Free Library www.altamet.com.au
2 3 8 22 22
1. INTRODUCTION
1.1
BACKGROUND
It is the purpose of the paper to discuss the effect correct mineralogical information from core logging has on the subsequent metallurgical testwork programme and process selection using some case studies in the gold and uranium and industries. Factors such as type of rock, rainfall and topography are thought to influence the minerals that ultimately formed in various ore bodies. Different ore bodies can therefore react differently during metallurgical testwork programmes. It is therefore important to obtain accurate mineralogical and elemental deportment data which also indicate where the valuable metal containing minerals are located within the ore body. The latter is especially important for low grade ores, where the valuable component phases maybe “locked” or encapsulated within an inert phase. The correct and accurate logging of both quantitative as well as observed mineralogical data will thus be of utmost importance. On a practical basis, powder X-ray Diffraction (XRD), Scanning Electron Microscopy (SEM) as well as Mineral Liberation Analyser (MLA) are the general methods currently used to determine the mineral wt% values and the average composition of the minerals in the ore. This information together with some other developed methods such as diagnostic leaching can then be used to calculate valuable metal and related deportment data. However, mineralogical analysis of certain ore bodies can be very challenging and further work is required to accurately identify or to ensure QA/QC is effective during this analysis process. XRD for example is unable to accurately quantify some of the economically important phases and metals/minerals and additional techniques are required to obtain a more detail understanding of the mineralogy. A fundamental understanding of process chemistry and mineralogy can also be obtained through other techniques such as synchrotron-based techniques (of which most need to be developed for specific ore body). Then there is also a need to develop methods that allow targeted data to be obtained in shorter time frames, i.e. methods such as near-infrared spectroscopy for example. So in general the quality, quantity and availability of accurate and complete mineralogical and chemical data from core logging are of fundamental importance to metallurgists in selecting samples for metallurgical testing. Samples are usually selected to statistically cover all of the alteration types in each deposit. To accurately statistically select sufficient samples and samples representing the alteration types within the future ore body, the metallurgist needs quality, accurate and as much core logging data as possible from the geologist and mineralogist. Not only is the quality and accuracy important, but also the way these drill hole data is stored and presented as for each meter in depth (drill holes) all relevant information needs to be stored in one file referring to that specific meter and drill hole. 1.2
GEOMETALLURGY
“Geometallurgy helps in understanding the effects and interactions between geology and mineralogy on metallurgical performance” Geometallurgy is the understanding of the variability in metallurgical parameters based on geological information such as mineralogy, grade and lithology. To understand the geometallurgy is to provide the practitioner a means to optimise the schedule and the metallurgical processing. The distribution of the lithology, alteration and mineralogy of the ore body influences the metallurgical parameters derived from the metallurgical samples, and it is essential to use this information in order to optimise the process modelling and scheduling. The selection of samples for metallurgical testwork is generally focused on providing geologically representative samples. One of the major problems experienced with metallurgical testwork using drillhole data, is the small sample sets selected for metallurgical testing. Major process and financial decisions are being made from results obtained using snapshot samples. Snapshot samples means that samples are basically selected only by grade and spatial location with the assumption that the rest of the ore body or portion of ore body is similar and will respond metallurgically the same way – i.e. the assumption is that the mineralogy and lithology are basically the same as the test sample (which is obviously is not the case). From the testwork, algorithms were usually developed and used to describe throughput and recovery using some explanatory variables including grade (mineral chemistry) and ore hardness. These algorithms were then used in the recovery models to determine tonnes produced and plant throughput. However, the
2
ALTA Free Library www.altamet.com.au
metallurgical parameters are a function of many geological factors including grade, lithology, alteration, mineralogy, spatial relationships and SG. The limited mineralogical information available was previously deemed enough for the metallurgist to make decisions with regard to how many samples are needed to evaluate the metallurgical performance of the ore during processing, and subsequently the basis for the plant design. If the ore body is consistent and not complex or refractory (thus, relative simple lithology), then only spatial and grade samples will be needed to evaluate the throughput and recovery. Whilst limited samples were previously considered enough for metallurgical testwork, there is an increasing realisation that the mineralogy and geology change, and so do the metallurgical characteristics. Increasingly complex and refractory ore bodies are being mined and treated, but there are a lot of associated process problems. Recent metallurgical testwork performed on new mineral resources in the base metal and gold industries (with complex mineralogy), has shown that the number of samples required for metallurgical testing increases dramatically. The number of samples needed for testing is a function of the many geological factors including grade, lithology, alteration, mineralogy, spatial relationships and SG. The more complex these characteristics are, the more scientific you have to be in selecting the quantity and source location of metallurgical samples to be tested. Factorial designs and statistical methods can then be employed to determine the amount of samples needed to develop an algorithm statistically or how many you need to develop the geological/metallurgical (geometallurgical) framework. The information derived from these samples can be used with advanced geostatistical techniques to estimate the metallurgical parameters within the resource model. Ideally a geometallurgical approach will incorporate the collection of spatially and geologically representative metallurgical information, incorporate the metallurgical information into the resource model, generate quality metallurgical parameters, and build a modelled metallurgical performance into the mining and processing schedules. These modelled parameters will then be used in process modelling and scheduling, thus enhancing the quality and accuracy of the decision making process for process design, financing, production forecasting, optimization as well as life of asset calculations. 2. GOLD CASE STUDY
2.1.
BACKGROUND
The interrelationship between mineral liberation and leaching behaviour of a gold ore is ill defined, mainly due to the complexity of both leaching and mineral liberation. This study presents a neural network approach to modelling the liberation of gold bearing ores. A complete mineralogical analysis of unmilled and milled ores, including gold deportment and gangue content are used as inputs to a self-organising neural net which generates order preserving topological maps. The arrangement and shapes of these clusters are coupled to unmilled free gold data to predict gold liberation in milled ores (absolute error: 8.1%). Moreover, the self-organising maps were diagnostic of the quality of data used, indicating that the relationship between particle size and gangue material content requires further investigation. The grinding of ore to a size that is conducive to the extraction of minerals by leaching or flotation is an essential and expensive component of most mineral processing operations. Identifying correct comminution practices is of major importance since overgrinding results in reduced floatability of minerals, while undergrinding leads to a product in which minerals have not been exposed, and cannot be leached or floated effectively. Milling control is commonly applied on the basis of producing a product within a narrow size band and at a high through-put (Metzner, 1993). With frequent changes in ore types, beneficiation setpoints that produce optimal liberation are usually poorly determined. Thus, great importance is assumed by models able to give a detailed description of the ore liberation - comminution process. Mineral liberation is discussed extensively in the literature (Gaudin, 1939; Wiegel and Li, 1967; Andrews and Mika, 1975; Barberty, 1985, Wiegel, 1975; King, 1979; Bonifazi and Massacci, 1995; Lorenzen and van Deventer, 1994; Lorenzen 1994, Annandale et al., 1996), although further work is evidently required. Progress in liberation technology is best described by comparing the work of Lynch (1984) who expected in 1984 for a mineral liberation monitor to “become available in the near future”, to that by Hales and Ynchausti (1993) in 1993, “it is expected to become commercially available in the future”. In spite of the attention given to liberation, the practical benefits in industry have not materialised. The drawback of current liberation models rests with the manual, laborious nature of determining model parameters, such as intercept lengths, particle size effects, etc., in order to predict the 3
ALTA Free Library www.altamet.com.au
liberation of the valuable mineral. Even in view of complex mathematical approaches such as stereology (Barbery, 1985) which are used to infer volumetric particle properties, the reported advantages are limited. Diagnostic leaching has aided efforts to rationalise these models (Lorenzen and van Deventer, 1994; Lorenzen, 1995; Annandale et al., 1996), enabling a more “hands on” approach, but is application specific. The existing gap in our understanding of the liberation phenomenon is further frustrated by poor use of available data or limited quantities of data. Other techniques such as neural computing have emerged in recent years, demonstrating tremendous application flexibility in the areas of classification, simulation and optimisation of nonlinear processes. The ability of a neural network to learn complex relationships makes it ideal for mineral processing, in that process disturbances are frequent and commonly the result of a multivariable state space change. This paper is a continuation of previous work (Annandale et al., 1996) where only the relationship between gold liberation and particle size was considered. The current investigation extends this to include a texture characterisation of the ore, which other than size, is a major factor in liberation behaviour. In this investigation a self-organising neural net with a Kohonen layer was used to generate order-preserving topological maps of gold leaching results for both unmilled and milled ore characteristics. The arrangement and shapes of these clusters could then be used to develop simple neural net models which were capable of predicting the degree of gold liberation, i.e. gold leaching results, more accurately than other presently used models. 2.2.
EXPERIMENTAL DATA
Seven different ores obtained from South African gold mines were used in the experiments, namely ores from the Beatrix, St. Helena, Unisel, Harmony, Harties, Kinross and Leslie mines. All the ores were fed directly to autogenous mills. Ore samples were initially screened into three size intervals, viz. +6700 µm, +1500-6700 µm, and -1500 µm. The -1500 µm fraction was classified into six size fractions, viz. -1500+300, -300+150, -150+106, -106+75, -75+53 and -53 µm. This set of samples represents the unmilled ore. Representative samples from the eight size fractions were then fed to a laboratory ball mill with iron balls of different sizes, to produce the milled sample of 70% -75 µm. This milled sample was also screened into the above-mentioned six size fractions. Diagnostic leaching (Lorenzen, 1995) was performed on each of the particle size fractions for each of the ore types (unmilled and milled). The deportment of gold (free, base metal sulphides, pyrite silicates and carbon) and the gangue content (base metal sulphides, silicates and carbon) are shown in Tables 1 and 2 where the pairs of numbers are percentage values for the unmilled and milled ores respectively. Table 1: Mineralogical analyses of gold deportment (1. unmilled, 2. milled)
Beatrix 1 2 St. Helena 1 2 Unisel 1 Kinross 1 2 Leslie 1 2 Harties 1 2 Harmony 1 2
Head Grade (g/t) 7.49 7.70 13.08 7.96 8.21 9.20 3.76 3.61 9.46 7.12 12.4 12.23 2.65 2.76
FREE (%)
17.40 64.05 23.21 68.76 40.73 74.09 36.49 84.91 32.08 75.82 45.70 67.40 18.58 89.43
BMS (%)
PYRITE (%)
10.22 8.03 6.31 9.97 9.14 8.62 13.31 3.64 8.25 5.08 8.30 7.70 6.47 4.81
9.93 8.58 6.44 7.36 16.40 10.18 6.93 3.84 11.02 3.23 8.30 4.90 4.56 1.30
4
ALTA Free Library www.altamet.com.au
SILICATE (%) 55.30 11.45 53.72 8.04 22.41 3.65 38.87 5.24 46.03 13.83 37.70 20.00 36.94 1.98
CARBON (%) 7.14 7.89 10.31 5.88 11.32 3.47 4.40 2.38 2.62 2.04 33.45 0.54
2.3.
DATA ANALYSIS
The percentage of gold deportment (Aup, Aus, Aub, Auc) and percentage of gangue (Gp, Gs, Gb, Gc) (p-pyrite, s-silicates, b-base metal sulphides and c-carbon) in the various milled and unmilled ore minerals, the percentage of free gold in each of the particle size fractions (Auo), the head grade (HG), as well as the mass distribution (m) were projected to a two-dimensional topological map with a self-organising neural net. The eleven inputs, viz. x = { Auo, Aup, Aus, Aub, Auc, HG, m, Gp, Gs, Gb, Gc }, were fed to a two-dimensional 10x10 Kohonen layer which yields a two-node output layer (x- and y-coordinates for graphical visualisation). Figures 1 and 2 show the maps generated for the unmilled and milled ore samples respectively. Table 2: Mineralogical analyses of gangue content (1. unmilled/2. milled) Ore type Beatrix St. Helena Unisel Kinross Leslie Harties Harmony
BMS
PYRITE
14.93/ 16.16 3.06/ 2.55 3.47/ 5.03 1.94/ 2.82 15.99/ 6.44 4.00/ 3.44 10.39/ 3.96
SILICATE
CARBON
74.61/ 73.96 90.99/ 91.82 89.69/ 88.60 92.68/ 90.55 75.90/ 87.17 89.80/ 89.85 78.09/ 90.18
3.17/ 2.86 1.24/ 1.15 2.62/ 1.59 3.21/ 1.22 2.20/ 1.81 2.26/ 2.35 1.25/ 1.25
7.28/ 7.02 4.71/ 4.48 4.21/ 4.78 2.17/ 5.42 5.90/ 4.58 3.85/ 4.36 10.26/ 4.62
1 Harmony Beatrix
0.5
Leslie Y
Kinross 0 -1
-0.5
St. Helena 0
0.5
1
Harties Unisel
-0.5
C-O-G
-1 X
Figure 1: Topological feature map of unmilled gold ores. The clustering of ore type is reasonably defined (except for Unisel). For the unmilled ores (Figure 1), a reasonable level of clustering is evident. The bottom right quadrant shows the mapping of the largest size fraction for all ore types, indicating a particle size dependence for liberation characteristics. The main feature in this SOM is the clusters that are grouped together; for example, Beatrix and Leslie. These two ore types are very similar in both their gold deportment and mineralogical content, as shown in Tables 1 and 2. Similarly, the Harmony cluster is also in close proximity, but obviously separated. Data from Tables 1 and 2 show that gold deportment for Harmony in the form of free gold and carbon differ to that of Beatrix and Leslie. Similar deductions regarding the position of all ore type clusters can be made by analysing Tables 1 and 2. The other significant feature in Figure 1 is that a single size fraction (+300Âľm) data 5
ALTA Free Library www.altamet.com.au
point from each ore type is located in the bottom right quadrant, and as a consequence, most of the centres-of-gravity (C-O-G) do not lie within close proximity of the majority of the data points for each ore type. For this size fraction for all ore types, the free gold available for leaching is much less; for example, the three largest size fractions for Beatrix (unmilled) yield free gold percentages of 13.4% (+300µm), 72.3% (-300µm+150µm) and 83.5% (-150µm+106µm). Clearly, there is a particle size dependency which is not accounted for in the current set of data. With only 42 exemplars available and a large number of parameters involved, it is difficult to obtain a comprehensive picture of liberation properties. It would be reasonable to suggest that geographical location of each mine would be strongly related to the relative position of the clusters on the SOM. However, this is not the case, indicating that mineralogical content of an ore is a highly localised phenomenon. 1 Harmony 0.5
Beatrix
Y
Leslie Kinross
0 -1
-0.5
0
0.5
1
St. Helena Harties
-0.5
Unisel
-1 X
Figure 2: Topological feature map of milled gold ores
Measure of Clustering (normalised)
1 0.9
Unmilled
0.8
Milled
0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Harmony
Beatrix
Leslie
Kinross
St Helena
Harties
Unisel
Ore Types
Figure 3: Measure of clustering for unmilled and milled ores As shown in Figures 2 and 3, the level of clustering in the milled SOM is considerably less than its unmilled counterpart (Figure 1), indicating that the ores behave quite differently when comminuted. The shift in relative positions of the centres-of-gravity is well illustrated in Table 3. For example, Harties maintains its degree of clustering (1.00unmilled - 0.81milled) and only shifts a distance of 0.5. In contrast, St. Helena undergoes a dramatic increase in dispersion (0.67unmilled 0.00milled) and a significant translation in its centre-of-gravity (0.91). This information gives an indication of the complexity involved in modelling the liberation process. While particle size information is an attractive basis for determining liberation characteristics, as in the case of the King liberation model (Lorenzen and van Deventer, 1994), Figures 1,2 and 3 and Tables 1, 2 and 3 demonstrate that gold deportment and mineralogical content are the main determinants of liberation.
6
ALTA Free Library www.altamet.com.au
Table 3: Euclidean distances between the centres-of-gravity of unmilled and milled ores Ore Types
Euclidean distances
Harmony
0.65
Beatrix
1.23
Leslie
0.72
Kinross
1.18
St. Helena
0.91
Harties
0.50
Unisel
0.40
Free Gold Predicted, % (Milled, BPNN)
100 90
Absolute error: 25%
80 70 60 50
+300Âľm
40 30 20 10 0 0
20
40
60
80
100
Experimental Free Gold Data, % (Milled)
Figure 4: Experimental milled free gold data versus neural net predictions
The prediction of free gold liberation resulting from the comminution process is achieved by linking the topological map features (i.e. x- and y- coordinates) in Figure 1 to a simple back propagation neural net. The assumption is that similar ores (i.e. ores projected to the same area of the SOM map) will behave similarly with regard to the liberation of gold during comminution. The neural net is trained using {x, y, Auounmilled} data as the inputs and {Auomilled } data are the outputs. The hidden layer is omitted due to a limited number of exemplars. As shown in Figure 4, this single layer neural net model was able to predict the data with an absolute error of 25%. This is significantly better than the 63% achieved by Annandale [11]. As indicated by the highlighted square in Figure 4, the +300Âľm data is the most significant contributor to prediction error, a fact which was identified in analysing the unmilled self-organising map. Figure 5 represents the neural net predictions of the average liberated free gold. The 8.1% absolute error is a good improvement on the 15% by Annandale et al, (1996) and the 19% by the Lorenzen model (Lorenzen, 1993).
7
ALTA Free Library www.altamet.com.au
Average Free Gold Content (%)
80
abs. error: 8.1%
70 60
Experimental
50
SOM(NN) 40 30 20 10 0 Harmony
Beatrix
Leslie
Kinross St Helena
Harties
Unisel
Ore Types
Figure 5: Neural net prediction of the average free gold (milled) for each ore type. Absolute error: 8.1%. The major contributor to error is due to poor prediction of +300 Âľm size fraction 2.4.
DISCUSSION AND SUMMARY
The grinding of ores to a size that is conducive to the optimal extraction of valuable mineral is the single most critical aspect of comminution practices. Although mineral liberation is discussed thoroughly in the literature by various researchers, there is still much work to be done. In this investigation it was shown that neural net modelling is an excellent tool for understanding the relationship between gold liberation and ore gangue content. In particular, self-organising maps were used as a diagnostic tool to analyse the significance of gold deportment and mineralogical content in comminution activities. It was found that the relationship between particle size and gangue mineral content needs further investigation, particularly in view of the fact that the proportions of base metal sulphides, silicates, pyrite and carbon are directly responsible for gold liberation. The error of 8.1% in predicting milled free gold belies the complexity of the modelling process, since the quantity of free gold in the larger size fractions was poorly determined. It is postulated that a larger data base will bring clarity to many of these issues. Furthermore, it is expected that an extensive analysis of particle breakage mechanisms will bring to light the role of fracturing in the liberation phenomenon. Thus, if more representative samples and statistically enough drill hole samples can be collected at an early stage of a project, characterisation will be more accurate and extraction and processeing predictions using tools scuch as these explained in this section, will be of utmost improtance. 3. URANIUM CASE STUDY
3.1.
BACKGROUND
Gold deposits in South Africa are invariably associated with uranium bearing minerals. Although the uranium content of these ores is generally low by international standards, averaging around 0.3 kg U3O8 per tonne, the overall economics of processing these deposits can be substantially improved by recovery of uranium as by-product. Moreover, the current high uranium prices as well as increased demand due to renewed interest in nuclear power provide significant incentives for dedicated uranium production. An understanding of the factors responsible for the leaching behaviour of uranium-bearing ores is critical in achieving optimal uranium recoveries. This is particularly important in light of the fact that dissolutions higher than 90% are very difficult to achieve under the normal operating conditions employed on the South African acid leaching plants. To achieve optimal extraction, an important consideration is the mineralogical characteristics of the ores in terms of a plantâ&#x20AC;&#x2122;s flow/operational perspective. Typically a diagnostic leaching approach is followed. In this paper, a mineralogy-leachability explanation is presented to rationalise the difficulty in exceeding 90 % dissolution from low grade uranium ores on the basis of a novel diagnostic leaching method. More specifically, to determine the interrelationship between mineralogy, mineral liberation and the leaching behaviour of uranium, a methodology was developed for unlocking
8
ALTA Free Library www.altamet.com.au
uranium by a combination of chemical (drastic leaching of minerals associated with the residual uranium) and physical (fine grinding to increase area exposure and liberation) methods. The findings from these investigations will form guidelines for developing economically viable flow sheets for improved uranium recovery, greater than 90%. A diagnostic leaching method is typically used to characterise the leaching behaviour of a specific mineral. It is used to describe limitations in achieving optimum recoveries and conditions to overcome these limitations. An understanding of the factors responsible for the leaching behaviour of uranium-bearing ores is critical in achieving optimal uranium recoveries. This analytical method was developed in 1986 by Anglo American Research Laboratories (AR), originally for gold leaching (Tumilty et al.,1986; Lorenzen, 1992; Lorenzen, 1995). It involves a series of sequential leaches, developed to elucidate the deportment of a specific mineral within various matrices. The differences in the kinetic and thermodynamic stability of various minerals allow leach selectivity. It is therefore possible to eliminate the least stable mineral present in the ore matrix (into an aqueous media) through selective oxidative leaching. The measured concentration of the mineral of interest will represent the amount associated with specific constituent minerals. Based on this information problem areas can be identified in plants and unit operations. This is particularly important in light of the fact that dissolutions higher than 90 % are very difficult to achieve under the normal operating conditions employed on the South African acid leaching plants, as seen in literature (Ford and Gould, 1994; Lottering et al, 2007). To establish the reasons for the difficulty in achieving higher uranium dissolutions, a diagnostic leaching method was used to analyse the unlocking and dissolution of refractory uranium. The development of this method will be explained in detail in this paper. With some current technologies it is possible to determine the amount of an element associated with a specific mineral using a mineral liberation analyser. Therefore, for this application the question was not necessarily to determine the deportment of the element in the mineral matrix but rather conditions necessary to leach the specific minerals, being able to improve the recovery of the element of interest. The uranium mineralogy as well as the uranium leaching response of three different ore types from the Vaal River area during atmospheric sulphuric acid leaching was investigated and reported on in a previous paper (Lottering et al., 2008). The aim of this paper involves a diagnostic leaching method for the development of a treatment process for the different Vaal River ores types (further referred to as ores A, B and C) that will give insight as to whether it is possible, within practical boundaries, to leach residual uranium minerals. The information (mineralogical and diagnostic leaching data) can be used as baseline for the design of alternative flow sheets. The mineralogy of the ore investigated is crucial in designing a diagnostic leaching sequence. The step-by-step method presented in this paper can also be applied to evaluate and fully characterise any uranium containing ore body within the Witwatersrand Basin. The information will be essential regarding choice of leaching technique, operating conditions and expected uranium recovery. 3.2.
EXPERIMENTAL
Mineralogy Like gold, uranium minerals are concentrated in a matrix of pebble-supported conglomerates (Smit, 1984). The bulk mineralogies of the three different ores investigated are fairly similar and consist primarily of quartz (70 – 80 %), with lesser amounts of muscovite (8 – 11 %). From Table 4 it can be seen that samples of ore A are slightly different from, and contain less quartz, pyrite and chlorite and more pyrophyllite, as compared to the other ores. The uranium concentrations in the different ores also varied within the following ranges: ore A (270 – 330 ppm), ore B (290 – 450 ppm) and ore C (390 – 540 ppm). A detailed uranium mineralogical characterisation of the ore was conducted and the uranium mineral distributions are shown in Table 5. As expected from earlier studies by Smit (1984), it was found that uraninite as well as branneritetype minerals (U1-xTi2+xO6) are jointly responsible for the major portion of uranium carriers in ore from the Witwatersrand basin. Table 5 shows that 80 – 90 % of the uranium in the ores is contained as uraninite, 8 – 19 % as brannerite, and the balance as traces of coffinite and uranium phosphates.
9
ALTA Free Library www.altamet.com.au
Table 4: Bulk mineralogy of the three Vaal River ores Mineral
Formula
ORE A (%)
ORE B (%)
ORE C (%)
Quartz
SiO2
70.2
73.3
79.8
Muscovite
K-Al-silicate
10.1
11.3
8.2
Chlorite
Mg-Fe-Al-silicate
2.0
3.2
3.6
Pyrophyllite
Al4(Si8O2)(OH)4
9.7
2.5
1.1
Pyrite
FeS2
1.3
2.5
2.8
Albite
Na-Al-silicate
4.8
5.1
1.9
Table 5: Uranium mineral distribution (area %) of feed samples. Values in brackets represent uranium metal deportment, in %, calculated from ideal mineral uranium contents, assuming spherical shapes to convert area to volume and then to mass of 1 mineral using their ideal densities Mineral
Formula
ORE A (%)
ORE B (%)
ORE C (%)
Uraninite
UO2
47.6 (84.9)
42.1 (79.7)
52.8 (89.2)
Brannerite
(U,Th,Ca)[Ti,Fe)2O6
42.2 (12.9)
49.6 (18.5)
32.3 (7.7)
U-Phosphate
(U, Cl)PO4
3.1 (0.2)
2.5 (0.1)
5.7 (0.4)
Coffinite
U(SiO4)1-x(OH)4x
7.1 (2.0)
5.8 (1.7)
9.2 (2.7)
Uraninite dissolves readily in the presence of a lixiviant provided that the required conditions of extraction are met. Brannerite-type minerals, unlike uraninite, are not readily leachable in sulphuric acid and therefore are referred to as refractory. Liebenberg (1955) distinguished between two uraniferous titanates in Witwatersrand ore: uraniferous leucoxene and brannerite which have UO2:TiO2 mole ratios of <1 and >1, respectively. One would expect variability in the leaching response amongst brannerite-type minerals, but at this stage of the study different brannerite types were not investigated. Previous work done by Glatthaar and Duchovny (1979) indicated that Vaal River ores mostly consist of brannerite associated with leucoxene and other titaniferous minerals (termed uraniferous leucoxene) which have a more loosely knit appearance and probably are more readily available for dissolution as compared to brannerite associated with silicates (termed brannerite), which occurs as minute, compact crystals intergrown in the siliceous material. This however, is not indicative that the different type of brannerite minerals will dissolve. Small amounts of coffinite were present. Coffinite is generally more reactive than brannerite to oxidative sulphuric acid leaching, but less reactive as compared to uraninite. In ore from the Elliot Lake district, which also consists mostly of brannerite and uraninite, secondary coffinite intergrowths are enhancing the overall kinetics of brannerite by accelerating leaching pit formation (Ifill et al., 1996). Further, in the ore samples investigated there were also traces of uranium associated with monazite, which may be inert or reactive, depending on whether uranium is a substitutional impurity or is adsorptively associated with monazite. Ford and Gould (1994) found that the amount of inert uranium is, in absolute terms, fairly similar for all Witwatersrand ores, at about 0.015 kg/t to 0.030 kg/t, suggesting that there is always an amount which is very inert. The mineralogy investigations focused on uraninite and brannerite, as they form the bulk of the uranium-bearing minerals. Table 3 shows the uranium-mineral associations. Most of the unliberated uraninite is associated with silicates, carbon, or has a ternary association. Ore C also has a high percentage uraninite associated with phosphates. Association with carbon is quite low for brannerite.
1
There is no â&#x20AC;&#x2DC;standardâ&#x20AC;&#x2122; U-phosphate mineral, so its uranium concentration was assumed to be the same as 3 that of brannerite and its density that of apatite. Densities used were 10.88, 5.2, 3.19 and 5.44 g/cm for uraninite, brannerite, apatite and coffinite respectively. Uranium contents used were 88.15, 33.54, and 72.63 % for uraninite, brannerite, and coffinite respectively (data mainly from www.webmineral.com).
10
ALTA Free Library www.altamet.com.au
Table 6: Brannerite and Uraninite associations Brannerite
Uraninite
ORE A
ORE B
ORE C
ORE A
ORE B
ORE C
Binary Association
67.8
85.8
67.7
86.0
85.2
86.6
Liberated
11.8
11.1
13.4
30.6
30.9
42.3
Uranium Minerals
1.7
7.5
15.7
11.8
7.1
6.5
Silicates
46.8
59.0
28.7
27.4
20.6
24.0
BMS
1.0
2.1
4.1
0.3
0.5
0.7
Oxides
3.6
4.0
4.6
0.1
0.1
0.7
REE-Phosphate
0.0
0.0
0.2
0.0
1.2
8.0
Carbonate
0.0
0.0
0.0
0.0
0.0
0.0
Carbon
2.9
2.1
1.0
15.8
24.8
4.3
Ternary Association
32.2
14.2
32.3
14.0
14.8
13.4
Total
100
100
100
100
100
100
Uranium grain sizes were found to be very small, with 50 % of the particles passing 19.4, 21.3 and 23.2 Âľm for ore A, ore B and ore C, respectively. The degree of liberation of the uranium-bearing minerals was low (see Table 6), between 11 and 45 %, and expectedly increased as particle size decreased. It is important to realize that surface area exposure may be a more useful indicator of leachability of uranium minerals, as opposed to intrinsic liberation. Minerals with exposed surface area are technically leachable as they can be accessed by a lixiviant. Tables 7(a) and 7(b) show that between 87 and 93 % of the uraninite particles and 71 to 86 % of the brannerite particles have more than 10 % of their surfaces exposed, and even higher proportions have more than 5 % of their surfaces exposed.
Table 7(a): % uraninite liberation and % of uraninite with 5 and 10 % surface exposure. ORE A (%)
ORE B (%)
ORE C (%)
Liberated
30.6
30.9
42.3
Middlings
26.1
31.6
31.2
Locked
43.3
37.5
26.5
Total
100
100
100
5 % Surface Exposure
96.4
96.4
98.3
10 % Surface Exposure
88.9
87.4
93.2
Table 7(b): % brannerite liberation and % of brannerite with 5 and 10 % surface exposure ORE A (%)
ORE B (%)
ORE C (%)
Liberated
11.8
11.1
13.4
Middlings
38.7
53.4
29.6
Locked
49.5
35.5
57.0
Total
100
100
100
5 % Surface Exposure
93.9
94.2
87.8
10 % Surface Exposure
79.3
86.1
71.3
11
ALTA Free Library www.altamet.com.au
Figure 6 shows the distinction between liberation and surface exposure. It is suggested that the measurement of area exposure of the uranium minerals (especially the fraction with > 10 % of surface exposed) is a very good indicator of their leachability, because only intrinsic inertness to leaching reagents can cause an exposed mineral to remain unleached.
Figure 6: Distinction between liberation and surface exposure
Uranium minerals seem to have a high percentage surface area exposure, despite poor liberation. This suggests that the breakage of the ore particles occurs near the uranium grains. It needs to be established whether this is due to association with soft gangue components. This means that coarser grinds can be tolerated for uranium leaching than one would suspect from their grain sizes alone. Residence Time The dependency of total uranium recovery were evaluated at different residence times supplying the necessary potential as well as acid concentration over the leaching period. The leaching tests o were carried out at aggressive plant conditions (Temperature: 60 C, H2SO4: 16.3 kg/t, MnO2: 4 kg/t). The leaching experiment was carried out in a 2L water-jacketed batch reactor, which was mechanically agitated. Sulphuric acid was added as a 647 g/L H2SO4 solution. Pyrolusite containing 29.3 % MnO2 and 36.5 % Fe was used as a solid oxidant. The absolute dosages were calculated on the basis of a pulp RD of 1.55 and a solids SG of 2.7. Stirring speed was kept constant at 6 rpm. The experiment was divided into 3 tests which run sequentially over 72 hours. The experiment start with a pulp mixture consisting of 1310 g ore and 1015 ml water was preheated for 1 hr to the test temperature, before the start of the experiment. Sulphuric acid was added to the mixture at time t=0, signifying the start of the experiment. After 1.5 hours, the solid oxidant (pyrolusite) was added. After 24 hours the experiment was stopped and filtered. The solids 6+ 2+ 2+ were dried and a sample was taken. The solutions were analysed for [U ], [Fe], [Fe ], [Mn ] and + [H ], while solids were analysed for U and Fe only. Solids were also sent for mineralogical analysis. The remaining solids were used for the next experiment. This was repeated 3 times up to a total of 72 hours. Each time the same procedure was used as described above. Enough water was added to ensure the same solid: water ratio and the acid and manganese dioxide amount were calculated accordingly. Potential Specified potentials were managed by using an auto titrator ensuring a constant potential throughout the process eliminating the possible effect of variation in potential. The leaching experiment was carried out in a 2L water-jacketed batch reactor, which was mechanically agitated. An auto titrator as well as a data collector was used to control the oxidant addition and capture the relevant information. Note that an anti oxidant was not used therefore if the potential overshoot only the reaction will bring it back to the set potential. The experiment starts with a pulp mixture consisting of 1310 g ore and 1015 mL water. Sulphuric acid (33 mL of 647 g/L H2SO4) was added to the mixture before heating the solution to 60oC. A 30 minutes delay after the addition of the acid 12
ALTA Free Library www.altamet.com.au
and before the oxidant was added allowing time for the potential of the system to stabilise before the oxidant was added to pre-dose the solution to the set potential. Thereafter, the experiment started for the duration of 24 hours. After 24 hours the experiment was stopped and filtered. The 6+ 2+ 2+ solids were dried and a sample was taken. The solutions were analysed for [U ], [Fe], [Fe ], [Mn ] + and [H ], while solids were analysed for U and Fe only. Solids were also sent for mineralogical analysis. Extreme Leaching conditions Using the residue from sulphuric acid leaching with no addition of oxidant, extreme conditions can be used to determine if it is possible to dissolve the remaining uranium minerals (specifically brannerite). Over dosing of nitric acid combined with high temperature were investigated. The following method was used as proposed by Lorenzen (1995): Nitric acid mostly reacts with sulphates and the reaction is highly volatile and envolves dangerous brown fumes of NO2. The leach must be carried out in a well ventilated fume cupboard and the nitric acid should be added slowly to the agitated slurry before heating. A stock solution of 1:1 HNO3 (55%)/distilled water by volume was added to a beaker containing the filtered wet solids to make up to roughly 10:1 L/S ratio. The mixture was boiled for 6 hours or until no further reaction occurs (i.e. no brown fumes of NO2 evolve). Additional HNO3 was added to make up for evaporation losses. At the end of the leach the solids were filtered and washed using distilled water (to a L/S of about 2:1). Repeat this step at least two times. The filtrate of the NO3 leach was sent for analysis to determine the uranium concentration as well as the residue sample to determine if there is any uranium left. 3.3.
RESULTS AND DISCUSSION
The bulk mineralogies of the three different ores investigated are fairly similar and consist primarily of quartz (70 – 80 %), with lesser amounts of muscovite (8 – 11 %). The uranium concentrations in the different ores varied within the following ranges: ore A (270 – 330 ppm), ore B (290 – 450 ppm) and ore C (390 – 540 ppm). The main uranium carriers are uraninite and brannerite with smaller amount of coffinite and uranium phosphates. For uranium leaching, the bulk mineralogy studies have shown that the reason for not achieving optimum dissolution is not necessary due to entrainment like it is the case for gold leaching (very high surface exposure of the different uranium minerals). The residue analysis (after standard sulphuric acid leaching) revealed that basically all the uraninite had dissolved in the leaching experiments and that most of the unleached uranium is brannerite which is at least exposed with some even fully liberated (Lottering et al., 2008). Based on that, it is not really necessary to use a diagnostic leaching method to unlock the refractory uranium mineral since it is already exposed. In this context it is more important to determine the conditions necessary to leach the uranium – by employing increasingly aggressive leaching conditions. The conditions to dissolve uraninite are well known from literature but, at this stage there is little knowledge with reference to other uranium minerals. Tests were conducted without the addition of an additional oxidant primarily to remove all the “easy” leaching uranium minerals (i.e. mostly uraninite). In addition to the results obtained it was possible to determine the true influence of an additional oxidant. Uranium dissolution results for sulphuric acid leaching at natural potential (without the addition of an additional oxidant – potential ranges are defined in Table 8) are as follows (note that the potential was measured at reaction temperature):
Table 8: Results of sulphuric acid leach without an additional oxidant (Acid: 16.3 kg/t; Temp: o
60 C) Minimum Eh (mV vs. SCE)
Maximum Eh (mV vs. SCE)
Average Eh (mV vs. SCE)
U Dissolution (%)
Ore A
390.9
450.1
415.9
83.6
Ore B
371.7
465.7
411.7
82.1
Ore C
287.1
457.5
359.2
81.7
From Table 8 it is clear that the initial potential (provided by the Fe that is leached from the ore) is sufficient for uranium leaching to take place under aggressive acid leaching conditions. It is
13
ALTA Free Library www.altamet.com.au
3+
important to note the Fe in the ore serves as an oxidant, the absence of MnO2 only results in no regeneration of Fe2+ to Fe3+. Therefore, dissolutions over 80 % are not unusual regardless of the absence of an additional oxidant. For Ore A and Ore B ore the potential drop over a 24 hour period to below 400mV but average over the leaching period 416 and 412 mV, respectively. For Ore C the change in Eh was more severe averaging only 359 mV. From the Nernst equation, potential is a function of the Fe3+:Fe2+ ratio therefore, the lower Eh maybe attributed to the fact that Ore C contains more chlorite and pyrite (Lottering et al., 2008). It is found that the actual Fe3+ concentration maybe more important than the actual potential that are illustrated in Table 8. To compare the absolute contribution of the addition of an oxidant, dissolution results with oxidant addition and without are presented in Table 9.
Table 9: Leaching results comparing the difference with and without the addition of an oxidant U Dissolution (%) with oxidant
U Dissolution (%) without oxidant
Ore A
90.3
83.6
Ore B
83.5
82.1
Ore C
80.5
81.7
Adding the same amount of acid, results from Table 9 indicate that the iron (Fe3+) supplied by the 3+ ore is sufficient and that the addition of a oxidant (increasing the Fe concentration and directly increasing the potential) did not have a significant effect on the total dissolution of uranium from Ore B and Ore C. However, the addition of the oxidant did seem to have an effect on the uranium dissolution form Ore A. These results can be explained in terms of the bulk mineralogy of the different ores (Lottering et al., 2008). Ore B and Ore C contain a significant amount more chlorite 2+ 3+ 3+ 2+ which provide both Fe and Fe . It also contains more pyrite, consuming Fe and produces Fe . Therefore, as indicated in Table 9, the overall potential maybe lower (due to a higher concentration of Fe2+) but maybe the absolute amount of Fe3+ is sufficient for uranium leaching. Interpreting the results it became evident that the addition of an oxidant may not be that critical in leaching Ore B and Ore C but the opposite is true for Ore A. This indicates that there is definite scope for reduction of additional oxidant and that further investigation in this regard is needed. Leaching with sulphuric acid and manganese dioxide in operating conditions that are attainable on plant scale and independent of conditions used, it was not possible to increase the uranium dissolution beyond 90 % (Lottering et al., 2007). To establish if it is possible to increase the total uranium yield it is necessary to move outside operating boundaries. Previous work done by Lorenzen (1992) showed that there is not a big difference in the total uranium dissolution between leaching at 400 mV and 600 mV (vs. SCE). Since normal operating conditions average between 400 and 500 mV (vs. SCE), it was decided to do tests at 500 mV and 700 mV (vs. SCE). Leaching with sulphuric acid and manganese dioxide, it is impossible to increase the overall potential above 500 mV (vs. SCE). Therefore, it was necessary to use other oxidants to reach higher potentials (i.e. H2O2 and HNO3). For the H2O2 tests sulphuric acid was also added to maintain a reasonable pH within the thermodynamic window. Nitric acid is both a strong acid as well as a high oxidising agent; therefore, for the nitric acid test no sulphuric acid was added. The pH values after 24 hours are reported in Table 10. Since different oxidants were used it will only make sense to compare end dissolutions. Table 10: pH values of constant Eh tests 500mV
700mV
Ore A
1.08
0.55
Ore B
1.57
1.07
Ore C
1.52
1.21
From Figure 7 below it can be seen that there is a significant increase in the uranium dissolution for Ore A moving form 300 mV to 500mV but no increase upon increasing the solution potential to
14
ALTA Free Library www.altamet.com.au
700mV. It is possible that maximum leachable uranium has leached. However, for Ore B and Ore C there were improvements of between 8 and 10 basis points leaching at 700 mV. Interestingly it seems that variation of potential has a great effect on the dissolution of uranium for Ore A at lower potentials (see Table 8) while it only starts to show to have a significant effect on dissolution at higher potential for Ore B and Ore C. Evaluating the bulk mineralogy it is possible to partly explain these results in terms of the sulphide mineral composition of these ores. Ore A contains little pyrite while Ore B and Ore C contain significant higher amounts as previously mentioned. Increasing potential for 500 mV to 700 mV destroys these sulphide minerals liberating the uranium associated with these minerals. The effect thereof can also be noted in the acid concentrations in Table 10. The distribution of the residue uranium indicated that the main uranium mineral left is brannerite. However, since the uranium concentrations left are very low the results are rather expressed as absolute uranium deportment of the different uranium minerals left (in ppm) to put the results more in perspective (see Table 11 â&#x20AC;&#x201C; Table 13). Since the uranium concentration in uraninite is much higher compared to brannerite, higher absolute uranium concentrations left as uraninite compared to brannerite is expected. Also note that the mineralogy results are relative and only give an indication since the samples only contain trace amounts of uranium. 100
U Dissolution ( % )
95
90
85
80
75
70 Ore A
Ore B Natural Eh
Constant 500mV
Ore C Constant 700mV
Figure 7: Comparing uranium dissolution at different constant potentials
From Table 11 â&#x20AC;&#x201C; Table 13 the following can be observed: The uranium contribution in the leaching solution from brannerite is very little compare to uraninite. The increase in uranium dissolution by an increase in the potential is mostly due to an increase in the dissolution of uraninite. Brannerite and uraninite are associated with each other. It is speculated that destroying gangue minerals as well as dissolving some brannerite, liberates remaining uraninite particles, thus increasing uraninite dissolution. Brannerite kinetics are expected to be very slow but there is a definite increase in brannerite dissolution for Ore B and Ore C at more aggressive leaching conditions moving to 700 mV (see Table 12 and Table 13). Leaching at more oxidising conditions resulted in the dissolution of sulphate minerals which is most probably the reason for the increase, since brannerite is highly associated with pyrite. These experiments proved that even under such extreme conditions it was not possible to get a total dissolution. Table 14 and Table 15 indicate that the remaining uraninite and brannerite are to a large degree exposed to the leaching environment. To determine the effect on the leaching kinetics of brannerite, it is recommended to do tests at longer leaching times at 700 mV since the recovery of Ore B and Ore C still on the increase.
15
ALTA Free Library www.altamet.com.au
Table 11: Uranium mineralogy (expressed as ppm) for the residue samples Ore A after tests at different potentials
Left (ppm)
Start (ppm)
Left (ppm)
Dissolved (%)
Start (ppm)
Uranium
234.1
38.3
83.6
357.3
24.9
93.0
243.6
21.2
91.3
Uraninite
198.8
25.1
87.4
303.3
18.1
94.0
206.8
17.0
91.8
Brannerite
30.2
10.9
63.8
46.1
6.2
86.5
31.4
4.2
86.7
U_Phosphate
0.5
0.5
0.0
0.7
0.6
10.3
0.5
0.0
100.0
Coffinite
4.7
1.1
76.0
7.1
0.0
100.0
4.9
0.0
100.0
Ore A
Dissolved (%)
Dissolved (%)
700mV
Left (ppm)
500mV
Start (ppm)
Natural Potential
Table 12: Uranium mineralogy (expressed as ppm) for the residue samples Ore B after tests at different potentials
Left (ppm)
Start (ppm)
Left (ppm)
Start (ppm)
Left (ppm)
Dissolved (%)
700mV
Dissolved (%)
500mV
Start (ppm)
Dissolved (%)
Natural Potential
Uranium
320.5
57.2
82.1
306.6
62.8
79.5
353.8
38.5
89.1
Uraninite
255.4
47.1
81.6
244.4
47.0
80.8
282.0
29.5
89.5
Brannerite
59.3
10.1
83.0
56.7
15.8
72.2
65.5
9.0
86.3
U_Phosphate
0.3
0.0
100.0
0.3
0.0
100.0
0.4
0.0
100.0
Coffinite
5.4
0.0
100.0
5.2
0.0
100.0
6.0
0.0
100.0
Ore B
Table 13: Uranium mineralogy (expressed as ppm) for the residue samples Ore C after tests at different potentials
Left (ppm)
Dissolved (%)
Start (ppm)
444.9
81.6
81.7
475.3
65.2
86.3
649.4
34.5
94.7
Uraninite
396.9
56.2
85.8
424.0
42.2
90.1
579.3
26.0
95.5
Brannerite
34.3
24.3
29.0
36.6
22.1
39.7
50.0
8.5
82.9
U_Phosphate
1.8
1.1
37.4
1.9
0.9
52.2
2.6
0.0
100.0
Coffinite
12.0
0.0
100.0
12.8
0.0
100.0
17.5
0.0
100.0
Left (ppm)
Left (ppm)
Dissolved (%)
700mV
Start (ppm)
Dissolved (%)
500mV
Uranium
Ore C
Start (ppm)
Natural Potential
16
ALTA Free Library www.altamet.com.au
Ore A Natural Potential
Ore A 500mV
Ore A 700mV
Ore B Natural Potential
Ore B 500mV
Ore B 700mV
Ore C Natural Potential
Ore C 500mV
Ore C 700mV
Table 14: Percentage brannerite liberation and percentage of brannerite with 5 and 10 % surface exposure for the residue samples for different leaching potentials for Ore A, Ore B and Ore C
Liberation
7.91
9.49
17.21
33.16
26.69
20.11
42.94
14.03
14.49
Free surface >5%
98.06
97.03
91.79
97.76
97.88
87.65
92.13
96.75
95.28
Free surface >10%
97.97
89.76
67.71
96.45
90.26
70.86
84.00
91.64
67.10
Ore A 700mV
Ore B Natural Potential
Ore B 500mV
Ore B 700mV
Ore C Natural Potential
Ore C 500mV
Ore C 700mV
Liberation
33.53
7.37
19.30
6.14
46.05
5.32
10.38
10.12
14.01
Free surface >5%
92.44
99.10
77.65
79.69
94.12
94.22
91.07
87.80
85.47
Free surface >10%
90.46
97.56
69.33
72.62
91.10
90.86
72.96
58.39
82.62
Ore A 500mV
Ore A Natural Potential
Table 15: Percentage uraninite liberation and percentage of brannerite with 5 and 10 % surface exposure for the residue samples for different leaching potentials for Ore A, Ore B and Ore C
Based on the mineralogical analysis that indicated that most of the unleached uranium exists as brannerite ,and since it is known that the leaching kinetics of brannerite is slow, tests were carried out for 24, 48 and 72 hours to determine the effect of time on leaching thereof (Zhang et al., 2003). As explained in the experimental setup (see Section 2) fresh acid and oxidant were added after every 24 hours. To indicate that sufficient acid and potential were available throughout the tests, the end pH and Eh measurements are included in Table 16. Note that measurements were taken at room temperature. Table 16: Eh and pH measurements for different leaching times 24h
48h
72h
mV (vs. SCE)
pH
mV (vs. SCE)
pH
mV (vs. SCE)
pH
Ore A
438.2
2.51
416.0
1.58
414.8
1.43
Ore B
386.1
2.78
414.0
1.84
409.0
1.70
Ore C
401.7
2.41
417.0
1.60
405.9
1.63
17
ALTA Free Library www.altamet.com.au
From Figure 8 (as expected) an increase in residence time resulted in increasing uranium recovery. This confirms the fact that branneriteâ&#x20AC;&#x2122;s kinetics are slow and that increasing the leaching period improves the total uranium recovery. However, the uranium leaching of Ore A and Ore C flattened off after 48 hours while that of Ore C was still in an upward slope. Before making any conclusions regarding the leaching characteristics of Ore B it is still necessary to increase the residence time to see if it is possible to increase recovery beyond 95%. However, based on the results for Ore A and Ore C it seems not possible to increase uranium recovery beyond 95 % even leaching over a 72 hour period with a over supply of acid and oxidant compared to the amount of uranium in the ore. This confirms the results discussed after Table 11 about the leaching of uraninite and brannerite from these ores.
100
U Dissolution ( % )
95
90
85
80 Ore A Ore B
75
Ore C 70 24
48
72
Time ( hours )
Figure 8: Dependency of residence time on uranium dissolution
Table 17: Uranium mineralogy (expressed as ppm) for the residue samples of Ore A after tests at different leaching times
Dissolution (%)
Aggressive H2SO4 72h
Left (ppm)
Dissolution (%)
Aggressive H2SO4 48h
Left (ppm)
Dissolution (%)
Aggressive H2SO4 24h
Left (ppm)
Ore A
Start (ppm)
Head Grade
Uranium
328.5
31.9
90.3
21.9
93.3
20.9
93.6
Uraninite
278.9
25.0
91.0
17.0
93.9
14.4
94.9
Brannerite
42.4
5.6
86.7
4.9
88.6
6.6
84.5
U_Phosphate
0.7
0.3
54.8
0.0
100.0
0.0
100.0
Coffinite
6.6
0.9
85.7
0.0
100.0
0.0
100.0
Interestingly it can be seen from Table 17 that there is not much difference in uranium dissolution for Ore A ore at various leaching times. However, it did have a major influence on the uranium dissolution for Ores B and C (Tables 18 and 19). The remaining uraninite and brannerite is mostly exposed to the leaching environment and is either liberated or associated with quartz (SiO2) etc.
18
ALTA Free Library www.altamet.com.au
Table 18: Uranium mineralogy (expressed as ppm) for the residue samples of Ore B after tests at different leaching times
Left (ppm)
Dissolution (%)
Aggressive H2SO4 72h
Dissolution (%)
Aggressive H2SO4 48h
Left (ppm)
Dissolution (%)
Ore B
Aggressive H2SO4 24h
Left (ppm)
Start (ppm)
Head Grade
Uranium
389.8
64.3
83.5
39.4
89.9
21.4
94.5
Uraninite
310.7
45.1
85.5
27.4
91.17
16.5
94.68
Brannerite
72.1
18.3
74.6
11.9
83.45
4.9
93.18
U_Phosphate
0.4
0.4
0
0.0
100
0.0
100
Coffinite
6.6
0.0
100
0.0
100
0.0
100
Table 19: Uranium mineralogy (expressed as ppm) for the residue samples of Ore C after tests at different leaching times
Left (ppm)
Dissolution (%)
Aggressive H2SO4 72h
Dissolution (%)
Aggressive H2SO4 48h
Left (ppm)
Dissolution (%)
Ore C
Aggressive H2SO4 24h
Left (ppm)
Start (ppm)
Head Grade
Uranium
602.5
117.5
80.5
52.4
91.3
43.8
92.7
Uraninite
537.4
86.4
83.9
50.0
90.7
30.3
94.4
Brannerite
46.4
29.6
36.3
2.5
94.7
12.8
72.5
U_Phosphate
2.4
1.6
34.9
0.0
100.0
0.7
70.0
Coffinite
16.3
0.0
100.0
0.0
100.0
0.0
100.0
To attempt achieving 100% dissolution, experiments were done with nitric acid (in excess) as o lixiviant at evaluated temperature (90 C). This method destroys practically most minerals except quartz. Therefore, the remaining uranium is expected to be locked in quartz and will not be leachable. The residue from the tests not using additional oxidant was used. Nitric Acid digestion improves dissolution with between 15 â&#x20AC;&#x201C; 17 percentage points. End dissolution percentages are presented in Table 20. The mass loss during these test were Âą 10%. This is the maximum attainable recovery for the different ores under extreme leaching conditions. Note that tests using HF were not performed due to technical difficulties. It is recommended to do sulphuric acid tests using NaF as an oxidant since F- has a high affinity for uranium species in solution to form a complex with the respective anions. o
Table 20: Uranium dissolution with excess HNO3 at 90 C Ore Type
U Dissolution (%)
Potential (mV vs. SCE)
pH
Ore A
98.0
885.1
-1.04
Ore B
98.5
888.4
-1.23
Ore C
98.6
900.7
-1.23
19
ALTA Free Library www.altamet.com.au
To put it all into perspective a summary of the results obtained are presented in Figure 9. For Ore A the uranium seems to react differently compared to the other ores due to differences in mineralogy. From the results it is clear that the optimum uranium dissolution of Ore A is reached at a constant potential of 500 mV. For Ore B and Ore C more aggressive conditions are required either increasing the residence time or increasing the potential to 700 mV. If uraninite and brannerite are not the dominating minerals then the method needs to be re-evaluated. Based on these results the following procedure can be concluded from a plant operation perspective (see Table 21):
Table 21: Recommended Plant Operations Conditions Ore Type
Leaching Technique
Ore A
Sulphuric Acid Leaching
(sulphide minerals < 2%)
(At least 11.2 kg/t H2SO4 , 60oC, 24h, 4kg/t MnO2)
Ore B and Ore C
If the brannerite concentration is between 0-20%-Aggressive Sulphuric acid leaching (At least 13.9 and 15.9kg/t H2SO4 respectively, 60oC, 24h, 4kg/t MnO2)
(sulphide minerals > 2 %)
If brannerite concentration is >20% consider more aggressive leaching methods (i.e. pressure leaching or alternative leaching agents)
100 98 96
U Dissolution ( % )
94 92 90 88 86 84 82 Ore A
80 78
Ore B
76 74 72
Ore C
70 Natural Eh Aggressive Most Constant Aggressive Aggressive Constant Plant Aggressive Eh = Plant Plant pH = 1 Conditions Plant 500mV Conditions Conditions (24h) Conditions (48h) (72h) (24h)
Constant Eh = 700mV
Nitric Acid Overdose
Figure 9: Summary of uranium dissolution results under different conditions
3.4.
STEP-BY-STEP DIAGNOSTIC LEACHING METHOD
Table 22 is an easy 5 step method to be followed to fully characterise any Witwatersrand ore investigated with regards to uranium leaching. The first step to evaluate the ore, would be to do a standard sulphuric acid leaching test without the addition of an oxidant (H2SO4: 16.3 kg/t; Temp: 60oC) over 24 hours to determine the amount of uranium leachable under natural potential circumstances. The residue from step 1 will then be used at aggressive conditions over 24 hours to determine the extra amount uranium leachable subjecting o the ore to the maximum attainable plant conditions (H2SO4: 16.3 kg/t; MnO2: 4 kg/t; Temp: 60 C) for
20
ALTA Free Library www.altamet.com.au
24 hours. The residue for step 2 must then be leached for 72 hour leaching as similar conditions to determine the effect of time on the leaching of the remaining uranium. To increase the aggressiveness of the leaching conditions step 4 involve leaching of the residue of o step 3 at aggressive leaching conditions but at an evaluated temperature of 90 C. The residue from step 4 is then leached at 700 mV for 24 hours and the unleached uranium from step 5 is then finally exposed to very aggressive leaching conditions using nitric acid digestion also at 90 oC. The remaining uranium is locked in gangue minerals (i.e. silica) and therefore not expected to leach. Based on results from these tests economical evaluations need to be done to justify process routes.
Table 22: Diagnostic Leaching method developed to evaluate any uranium containing ore for the Witwatersrand Basin Step
Method
Conditions
1
Leaching without the addition of an oxidant for 24 hours
H2SO4 addition: 16.3 kg/t
Aggressive leaching conditions over 24 hours on residue of step 1
H2SO4 addition: 16.3 kg/t
Aggressive leaching conditions over 72 hours on residue of step 2
H2SO4 addition: 16.3 kg/t
2
3
Minerals Dissolve o
Temperature: 60 C
MnO2 addition: 4 kg/t Temperature: 60oC
MnO2 addition: 4 kg/t
Most of the uraninite and coffinite, small amounts of brannerite, chlorite Remaining uraninite and coffinite, small amounts of brannerite Mostly brannerite, small amounts of U-Phosphates
o
Temperature: 60 C 4
5
6
3.5.
Aggressive leaching conditions at 90 oC for 24 hours on residue of step 3
H2SO4 addition: 16.3 kg/t
Leaching at 700 mV on the residue of step 4 for 24 hours
Eh: 700 mV
o
Nitric acid digestion at 90 C on the residue of step 5
MnO2 addition: 4 kg/t
Mostly brannerite, small amounts of U-Phosphates
Temperature: 90oC o
Temperature: 60 C HNO3 addition: Access o
Temperature: 90 C
Sulphate minerals (i.e. Pyrite) All exposed uranium minerals left
SUMMARY AND RECOMMENDATIONS
Diagnostic leaching tests show that the uranium dissolution optimum for Ore A is reached at 500 mV for 24 hours. For Ore B and Ore C it is reached after 48 - 72 hours of aggressive leaching at potentials of 500 mV or at potentials of 700 mV for 24 hours. I t is recommended to do sulphuric acid tests adding NaF as an oxidant since F has a high affinity for uranium species in solution as well as doing leaching tests at even higher constant potential as 700 mV to determine the true effect of potential on uranium mineral dissolution. Pressure leaching tests for the three ore are recommended to compare results with sulphuric acid tests. The absolute necessity for an additional oxidant need to be investigated for all three ores since the addition of an oxidant proved not to be necessary working at aggressive acid leaching concentration (16.3 kg/t) for Ore B and Ore C since iron leached for ore seems to be sufficient. The lixiviant can also be evaluated, at the same time, especially with regards to the leaching kinetics of brannerite Mineralogical analysis indicates that it is possible to leach brannerite but the leaching kinetics thereof are very slow. Based on the diagnostic leaching tests the following is recommended: Sulphuric acid leaching must be used for treating Ore A. If the brannerite concentration of Ore B and Ore C is < 20 % sulphuric acid leaching is recommended but if the brannerite concentration > 20 % other leaching methods will be recommended (i.e. pressure leaching or using a different leaching reagent).
21
ALTA Free Library www.altamet.com.au
4. CONCLUSION
From these case studies it is evident that good mineralogical information during the early stages of a project, for example using drill core samples, is of utmost importance to correctly identify: •
The type and quantity of other mineralogical analysis to be conducted.
•
The type and quantity of metallurgical testwork to be conducted.
•
Geometallurgical information to be collected.
To select such samples an effective core logging data storage system with all relevant information should be available to the metallurgist to statistically select appropriate samples and sufficient samples for testwork. The results from a well designed geometallurgical programme can then be used for: • • • • • • • •
Better flowsheet design (more flexible), Better use of algorithms for throughput and recovery in resource and reserve models, Better use of the mining schedule to optimise plant performance, Better plant and equipment design and sizing, Optimise plant performance , Optimise forecasting, Maximise NPV, and Reduce risk in subsequent phases.
In short, geometallurgy is an integrated part of the scoping, pre-feasibility and feasibility phases of a project. The better you plan and execute your metallurgical testwork programme during the early phases, the more accurate the decision making process will be during the selection phase. The better this integration between geology and metallurgy data gathering during early phases, the more accurate the predictions with regards to throughput and recovery during modelling and process and equipment selection will be in later phases, and the better the mine and process scheduling and subsequent NPV.
5. REFERENCES
1. Andrews, J.R.G., Mika, T.S., 1975, Comminution of a heterogeneous material: development of a model for liberation phenomena, Proceedings of the 11th International Mineral Processing Congress, Cagliari. 2. Annandale, G.J., Lorenzen, L., Van Deventer, J.S.J. and Aldrich, C., 1996, Neural net analysis of the liberation of gold using diagnostic leaching data, Minerals Engineering, 9(2), 195-213. 3. Barbery, G., 1985, Mineral liberation determination using stereological methods: a review of concepts and problems, In: W.C. Park, D.M. Hausen and R.D. Hagni (Editors), Applied Mineralogy, AIME, New York. 4. Bonifazi, G., Massacci, P., 1995, Ore liberation modelling by minerals topological evaluation, Minerals Engineering, 8(6), 649-658. 5. Ford, M.A. and Gould, D.G., 1994, The leaching of uranium on the Witewatersrand, Mintek Report No. M410, Mintek, South Africa. 6. Gaudin, A.M., 1939, Principles of mineral dressing, McGraw Hill, New York. 7. Glatthaar G.W., Duchovny M., 1979, Mode of occurrence of uranium in a Western Deep Levels and Vaal Reefs sample, AR Report, Project no. R54, Ref no. M/79/304, August 1979. CONFIDENTIAL. 8. Hales, L., Ynchausti, R., 1993, Futuristic look at process control in the minerals industry, Emerging techniques in the minerals industry, ed. B. Scheiner et al., SME, pp. 147-151. 9. Ifill, R.O., Cooper, W.C, Clark A.H., 1996, Mineralogical and process controls on the oxidative acid leaching of radioactive phases in the Elliot Lake, Ontario, uranium ores: II Brannerite and allied titaniferous assemblages, CIM Bulletin, June 1996.
22
ALTA Free Library www.altamet.com.au
10. King, R.P., 1979, A model for the quantitative estimation of mineral liberation by grinding, International Journal of Mineral Processing. 11. Liebenberg, W.R., 1995, The occurrence and origin of gold and radioactive minerals in the Witwatersrand System, the Dominion Reef, the Ventersdorp Contact Reef and the Black Reef. Trans. Geol. Soc. S.Afr., vol. 58, pp108 -227. 12. Lorenzen, L., "An electrochemical study of the effect of potential on the selective dissolution of base metal sulphides in sulphuric acid", Minerals Engineering, vol. 5, nos. 3-5, 1992, pp 535-546. 13. Lorenzen, L., 1993, A fundamental study of the dissolution of gold from refractory ores, PhD thesis, University of Stellenbosch, Stellenbosch, South Africa. 14. Lorenzen, L, 1995, Some guidelines to the design of a diagnostic leaching experiment, Minerals Engineering, 8(3), 247-256. 15. Lorenzen, L. & Tumilty, J.A., 1992, Diagnostic leaching as an analytical tool for evaluating the effect of reagents on the performance of a gold plant, Minerals Engineering, vol. 5, nos. 3-5, pp 503-512. 16. Lorenzen, L., Van Deventer J.S.J., 1994, Interrelationship between mineral liberation and leaching behaviour, International Journal of Mineral Processing, 41, 1-15. 17. Lottering, M, Lorenzen, L., Phala, N.S., Smit, J.T and van Schalkwyk, G.A.C., Mineralogy and uranium leaching response of low grade South African Ores, Bio – and Hydrometallurgy ’07, Falmouth, England, 1 and 2 May 2007, 16 pages. 18. Lottering, M, Lorenzen, L., Phala, N.S., Smit, J.T and Schalkwyk, G.A.C., Mineralogy and uranium leaching response of low grade South African Ores, Minerals Engineering, 2008, vol. 21, no. 1, pp 16-22. 19. Lynch, A., 1984, The automatic control of mineral preparation and concentration circuits, Control ‘84, ed. J.A. Herbst et al., SME,, pp. 3-12. 20. Metzner, G., 1993, The control of milling, SAIMM School: process simulation, control, and optimization, The South African Institute of Mining and Metallurgy, Randburg. 21. Smit. G., 1984, Some aspects of the Uranium minerals in the Witwatersrand Sediments of the early Proterozoic, Precambrian Research 25, pp37 – 59. 22. Tumilty, J.A., Sweeney, A.G. and Lorenzen, L., "Diagnostic leaching in the development of flow sheets from new ore deposits", Proceedings of the International Symposium on Gold Metallurgy, Editors: R.S. Salter, D.M. Wyslouzil & G.W. McDonald, Pergamon Press, 1987, pp 157-168 [CIM, 26th annual Conference of Metallurgist, Gold Symposium, Winnipeg, Canada, 23-26 August 1987]. 23. Wiegel, R.L., Li, K., 1967 A random model for mineral liberation by size reduction, Trans. AIME, 1967, pp. 238. 24. Wiegel, R.L., 1975, Liberation in magnetic iron formation, Trans. AIME, 1975. 25. Zhang, Y, Thomas, B.S., Lumpkin, G.R., Blackford, M., Zhang,Z., Colella, M. and Aly, Z., 2003, Dissolution of synthetic brannerite in acidic and alkaline fluids, J. Nucl. Mater., 321, pp 1-7.
23
ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
UPGRADING, GRAVITY TREATMENT & COMMINUTION
ALTA Free Library www.altamet.com.au
SENSOR BASED SORTING - GOLD APPLICATIONS By Lütke von Ketelhodt CommodasUltrasort (Pty) Ltd, South Africa
Presented by
Lütke von Ketelhodt lvonketelhodt@commodas.com
CONTENTS 1. 2. 3. 4. 5.
INTRODUCTION XRT SORTING OF NARROW-VEIN SULPHIDE TYPE GOLD ORE OPTICAL SORTING OF WITWATERSRAND CONGLOMERATE REEF CONCLUSION REFERENCES
ALTA Free Library www.altamet.com.au
2 4 10 17 17
1. INTRODUCTION 1.1 SOUTH AFRICAN GOLD ORES Gold mining activities have been key to the rapid industrial development in South Africa. The first gold rush in 1873 took place in the Pilgrim's Rest/Sabie goldfields approximately 450 kilometres North East of Johannesburg. Here the gold is contained in narrow-vein suIphide type deposits. In 1886 the Australian miner Geoge Harrison discovered the world’s largest gold bearing orebody, the Witwatersrand complex. This led to the incredible growth of the Johannesburg metropolitan area and other towns such as Welkom, Klerksdorp and Carletonville. The mining activities have been ongoing for over 120 years. The Witwatersrand ore bodies are tabular conglomerate reefs ranging in width from a few centimetres to 3 – 4 meters. Mining narrow reefs especially those of less than 1m in channel width, it is inevitable that there will always be waste rock diluting the grade of the run-of-mine ore going to the mill and gold processing plant. This waste rock stems from shaft sinking and development operations as well as hanging and footwall contamination from the stopes. A range of benefits could be achieved if such waste rock is removed as early as possible in the process stream. A number of mines have tried hand-picking of gold ore, for example at Buffelsfontein Gold Mine and also at Kloof Gold Mine. Usually these operations failed due to the inconsistencies and inefficiencies of such an operation. Particularly in the finer size ranges of -50mm hand picking is labour intensive and difficult to produce adequate tonnages. In some Witwatersrand ore a relationship between gold and uranium, related to a particular reef, has been exploited by using radiometric sorting to selectively pre-concentrate coarse gold- (and uranium) bearing rock from waste rock [1]. In the 1970’s and 1980’s radiometric sorters were used at Buffelsfontein Gold Mine with limited success. Density separation techniques are being used on sulphide type ores from the Pilgrim’s Rest area, but this has not worked for the Witwatersrand conglomerates where all the rock types have a similar specific gravity. In this paper two sensor based sorting applications for two different gold ore types will be discussed. •
Dual-Energy X-Ray Transmission (XRT) for sorting narrow-vein sulphide type gold ore, and
•
For Witwatersrand conglomerate reef optical sorting, using colour and brightness properties of the different rock types.
1.2 SENSOR BASED SORTING TECHNOLOGY Ore sorting itself is not a new concept, with hand sorting being one of the first methods of minerals processing. Electronic ore sorting equipment was first produced in the late 1940’s (Wills 1992) [2]. Although still a relatively small industry, ore sorting equipment can be applied to a variety of different applications. “Ore Sorting involves the appraisal of individual particles and the rejection of those particles that do not warrant further treatment” (Wills 1992) [2]. Salter and Wyatt (1991) [3] discuss that the sorting process can be divided into four interactive sub-processes 1. Particle presentation 2. Particle examination 3. Data analysis 4. Particle separation Feed preparation is more critical for sorters due the importance of surface characteristics and physical size of the particles most sorters need a 3:1 or 2:1 ratio between the largest and smallest particle to be efficient. Once the particles have been properly prepared for sorting they must be presented to the sensor. To operate efficiently the sensor must be able to analysis each single particle. As a result, feed rate and the materials handling methods are the critical components, with this most commonly being done by a conveyor belt or chute (Wotruba, 2006) [4].
2
ALTA Free Library www.altamet.com.au
Belt System
Chute System
Figure 1: Two types of sorting systems â&#x20AC;&#x201C; Chute vs. Belt (Harbeck, Kroog, 2008) [6] The critical stage of examining the particle and determining whether material is valuable or barren, is done by a combination of sensor and processing unit. Once the decision of has been made as to accept or reject a given particle, a mechanical device is required to physically make the sort. High pressure jets of air or water and mechanical arms or paddles are generally used to make this separation. Of all the components in a sorter, it is the choice of sensor that controls the design of a sorter (Weatherwax, 2007) [5]. A multitude of different sensors are available and the choice is generally driven by the mineralogy of a given ore. Optical sensors are the most common sensor type, which has been very successfully used in the industrial minerals industry (Wotruba, 2006) [4].
Table 1: The electromagnetic spectrum and the different sensors available for sensor based sorting in mineral processing (Harbeck, Kroog, 2008) [6]
3
ALTA Free Library www.altamet.com.au
2. XRT SORTING OF NARROW-VEIN SULPHIDE TYPE GOLD ORE 2.1. TGME [8] Test work using Dual-Energy-X-Ray-Transmission (XRT) on a number of sulphide type ores, for example nickel, have shown good sorting potential. In this example we describe the XRT sorting application at Transvaal Gold Mining Estate Limited (TGME) to upgrade waste rock dump material producing a concentrate of quartz-sulphide reef. Gold mineralisation is accompanied by various sulphides of iron, copper, arsenic and bismuth. The TGME operation, a wholly-owned subsidiary of Simmer & Jack, is situated in the Pilgrim's Rest/ Sabie goldfields approximately 450 kilometres North East of Johannesburg. The run-of-mine (ROM) ore is trucked from different satellite operations to TMGEâ&#x20AC;&#x2122;s central processing site which is located approximately 3km by road from the historic town of Pilgrimâ&#x20AC;&#x2122;s Rest. Some 200t/day are fed onto the crusher and pre-concentrated with dense-medium-separation (DMS) and then milled and processed further. Figure 1 shows an idealised cross section of the narrow vein sulphide deposit in Pilgrims rest and a picture of a stope.
Figure 2: Idealised cross section of the Narrow Vein Sulphide-Type Gold deposit, Pilgrims Rest [7] 2.2. DUAL ENERGY X-RAY TRANSMISSION SORTING The x-ray transmission method works according to the DE-XRT principle (Dual Energy X-Ray Transmission) as it is known from airport baggage inspection scanners. Commodas/UltraSort has drawn upon this basic principle and has developed a sensor system suitably adapted to sorting techniques. The broad-band x-ray radiation of an electrical x-ray source is applied to the material while it is moved along the scanning area at a rate of 3m/s (see Figure 3). The x-ray sensor system, which works like a line scan camera, picks up the x-rays penetrating the material and converts them into digital image data.
4
ALTA Free Library www.altamet.com.au
Figure 3: XRT Sorting Principle Figure 4 shows the sensor system consisting of two channels , each capturing the image of material in different x-ray energy spectra. The material attenuates the x-ray radiation, thus decreasing the modulation amplitude of the sensor, so that dark image areas appear. Attenuation depends on both the thickness and atomic density of the material. An image transformation of the density images of the two spectral radiation ranges makes it possible to classify each pixel according to the atomic density of the material. This is accomplished almost regardless of the thickness of the material. The example in Figure 3 shows pieces of iron ore with variable iron content with low, moderate, or high ore content (from left to right). The image on the right displays a color-encoded density classification. The different iron grades of ore pieces are clearly distinguishable.
Figure 4: XRT â&#x20AC;&#x201C; the Dual-Energy X-Ray sensing principle
5
ALTA Free Library www.altamet.com.au
XRT sorting has been chosen for this application because it is the most suited sensor to identify interlocked, non-liberated sulphides and small sulphide inclusions in waste rock particles when compared to other sensors and other coarse particle separation techniques. Figure 5 shows an example of how well sulphide bearing rocks can be distinguished from barren waste rocks. The dark blue color depicts the dense sulphide inclusions of each particle.
Figure 5: XRT images of different sulphide grade particles and waste rocks [8]
2.3. TEST WORK OF SULPHIDE SAMPLES Test work was conducted in March 2009 on 60kg ROM ore from the Frankfurt operation and on 100kg DMS floats to assess the amenability of these samples to upgrading with XRT sorting. The machine used for this test work is the Commodas PRO Secondary XRT. Assay results of the DMS floats test below clearly show the significant upgrade from a feed grade of 1.45 g/t to a concentrate of 3.83 g/t. The carbon content in the concentrate had also been halved in this sorting process.
Table 2: DMS Floats Test Work Results [8] DMS Floats DMS Float XRT Conc DMS Float XRT Tails Total
Mass [kg] 33.5 72.0 105.5
Mass [%] 31.8% 68.2% 100.0%
6
C [%] 1.5% 3.8% 3.0%
Au [g/t] 3.83 0.34 1.45
ALTA Free Library www.altamet.com.au
Au Recovery Feedrate [t/h] 84.0% 6.4 16.0% 13.6 100.0% 20.0
These results show that good value gold can be recovered from the discarded DMS tailings. With a gold recovery of 84% and a mass pull of 31.8% to concentrate and a significant decrease of the carbon content, this application would enhance the overall recovery of the TGME operation. The figures 6 and 7 display the separated fractions of the DMS floats sorting test. It can be seen that massive pieces and big inclusions of sulphides have been recovered. Small interlocked and finely disseminated sulphides have been lost into the tailings.
Figure 6: XRT concentrate grade ~ 3.8 g/t [8]
Figure 7: XRT waste grade ~ 0.2 g/t [8]
7
ALTA Free Library www.altamet.com.au
Table 3 hows the results of the coarse run-of-mine sample of + 9.5mm size fraction used in the XRT test run. It showed an upgrade from 0.49 g/t feed grade to 1.16 g/t. Combining the -9.5mm fines at 3.85 g/t with the sorted concentrate 1.16 g/t, an overall grade of 3.19 g/t was achieved at 97.3% recovery with a total mass-pull of 65.6%.
Table 3: ROM Test Work Results [8] ROM Overview ROM -9.5mm ROM +9.5mm XRT Conc ROM +9.5mm XRT Tails Total
Mass [kg] 30.6 10.0 21.3 61.9
Mass [%] 49.4% 16.2% 34.4% 100.0%
C [%] 1.6% 3.5% 2.3% 2.2%
Au [g/t] 3.85 1.16 0.17 2.15
Au Recovery Feedrate [t/h] 88.6% 19.6 8.7% 6.4 2.7% 13.6 100.0% 39.6
ROM XRT Sorting ROM +9.5mm XRT Conc ROM +9.5mm XRT Tails Total
Mass [kg] 10.0 21.3 31.3
Mass [%] 31.9% 68.1% 100.0%
C [%] 3.5% 2.3% 2.7%
Au [g/t] 1.16 0.17 0.49
Au Recovery Feedrate [t/h] 76.2% 6.4 23.8% 13.6 100.0% 20.0
ROM XRT Sorting +Fines ROM +9.5mm XRT Conc + ROM -9.5mm ROM +9.5mm XRT Tails Total
Mass [kg] 40.6 21.3 61.9
Mass [%] 65.6% 34.4% 100.0%
C [%] 2.1% 2.3% 2.2%
Au [g/t] 3.19 0.17 2.15
Au Recovery Feedrate [t/h] 97.3% 25.9 2.7% 13.6 100.0% 39.6
2.4. ON SITE OPERATION Based on these results, TGME decided to install an XRT-sorter to scavenge the gold from the discarded DMS tailings to fill the spare capacity that is available in the downstream processes. The XRT sorting circuit will be evaluated at TGME over a 6 month trial period under real operating conditions. Data on the existing DMS circuit indicates that this dump contains an average grade of 0.7g/t gold. Assuming an overall recovery of 80%, a feed rate of 25/h and an availability of 80%, this would result in a recovery of 50kg of Gold from the current stockpile of DMS floats. Figure 8 below shows the XRT-sorter installation. On the left side of the picture the DMS floats dump material is loaded into the feed bin. The feed belt discharges the material onto a scalping screen with 10mm screen apertures before a vibratory feeder feeds the XRT-sorter via an acceleration chute. The waste is conveyed to the left of the sorter house while the concentrate is fed directly onto the mineâ&#x20AC;&#x2122;s mill feed belt. To the right of the XRT-sorter the existing DMS-cyclone installation can be seen.
8
ALTA Free Library www.altamet.com.au
Figure 8: XRT-sorter plant at TGME [8] Note: Results discussed here are from the first months commissioning phase and are influenced by low running times due to finalising of the installation. Figure 9 shows an overview of the sorterâ&#x20AC;&#x2122;s latest performance.
Figure 9: Overview XRT-sorter performance [8]
9
ALTA Free Library www.altamet.com.au
The operational results show very low feed grades due to the composition of the dump which was used in the initial commissioning phase of the sorter plant. In the first month an average gold feed grade of 0.3g/t was assayed. The XRT-sorter was able to recover 20% of the gold in 10% of the mass. The XRT-recoverable-gold is only 26% due to the fact that the background value of the material is 0.1g/t. This was determined by assaying optically identified and hand sorted barren waste. The concentrate had an average Gold content of 0.6g/t and a waste of 0.2g/t. The grade was improved by a factor of at least two even at low feed grades. Higher upgrade ratios are achieved when the feed material has a higher grade. The undersize -10mm of the scalping screen was assayed at 0.4g/t of Gold. The average sorter feed during operation was 25t/h. The highest average sorter feed rate per day achieved was 34t/h with an accumulated sorter feed of 644t on this day. 2.5. CONCLUSION OF XRT SORTING AT TGME XRT sorting shows great potential for the pre-concentration of massive sulphide ores. Even though the grade of the XRT-sorter feed was a very low 0.3g/t it doubled the grade to 0.6g/t with a mass pull of only 10% to the concentrate fraction. The plant is expected to operate on higher grade run of mine material in the near future, where significant upgrades are expected as the initial test work showed. XRT sorting is a dry separation process with little infrastructure requirements and low operating cost. It is certainly an excellent new alternative beneficiation process for future applications. 3. OPTICAL SORTING OF WITWATERSRAND CONGLOMERATE REEF Optical sorting of Witwatersrand conglomerate reefs has been tested and proven as a viable technology over the last 7 – 8 years. During the period October 2003 to June 2004 a containerised optical sorting pilot plant was operated at Kloof Gold Mine on a waste rock dump. The paper “Viability of Optical Sorting of Gold Watse Rock Dumps” describes the test work and the operation of pilot plant in detail [9]. This paper is focussed on a new run-of-mine waste rock sorting project by Central Rand Gold (CRG). 3.1. CENTRAL RAND GOLD CRG is a gold development and mining company situated south of the city of Johannesburg. As its name indicates it operates on the central rand goldfield of the Witwatersrand basin. The wits basin is the largest multi-layer pebble reef gold orebody with exceptional continuity over hundres of kilometers. It has yielded 49,332 tons of gold so far, which is 39% of historical world gold production.
Johannesburg
East Rand Goldfield
Carletonville
Central Rand Goldfield West Rand Goldfield
Evander Goldfield
Potchefstroom
Klerksdorp
Klerksdorp Goldfield
Key Goldfields
Freestate Goldfield Welkom
Central Rand Goldfield (Study Area)
50
0
Kilometres
Carletonville
Figure 10: The Witwatersrand Basin
10
ALTA Free Library www.altamet.com.au
Cities / Towns
CRG has large contiguous mining and prospecting tenements of over 40km of strike. This relates to a resource base of 35.6 million ounces of gold. 3.2. GOLD REEFS MINED AT CRG [10] In the early days of gold mining in Johannesburg, the high grade Main Reef Leader (MRL) was mined out in most places. The lower grade Main Reef (MR) 5 g/t to 13 g/t which is lying 1 to 1,5m below the MRL is still in tact. At todayâ&#x20AC;&#x2122;s gold price levels mining the main reef has become very attractive. The diagram in Figure 11 shows the MR in relation to the MRL.
Figure 11: Schematic Diagram of Main Reef (MR) and Main Reef Leader (MRL) [10]
The picture in Figure 12 gives an idea of what the pebble watrix of the Main Reef looks like.
Figure 12: Main Reef [10]
11
ALTA Free Library www.altamet.com.au
The main challenge for CRG however is to mine the MR as cleanly as possible to avoid waste dilution of the run of mine going to the mill. It is impossible to mine only the MR channel as the hanging wall would only be a thin middlings layer of 1 – 1,5m beneath the void which used to be the MRL. CRG have develped a 2 stage mining model where the middling is mined and trammed first as waste and in a second stage the MR is mined and hoisted as run of mine ore. This is illustrated in Figures 13 and 14.
Figure 13: Selective mining stage 1. – removing middling low-grade/waste rock [10]
Figure 14: Selective mining stage 2. – removing high-grade Main Reef [10]
12
ALTA Free Library www.altamet.com.au
This mining method requires carefull management, control and discipline to ensure little losses of high grade MR. It is also time consuming and expensive to mine in two cuts. Therefore, CRG had also considered Hand sorting as a method to ensure high grade production. 3.3. CRG TEST WORK - OPTICAL SORTING OF MAIN REEF An automated sorting system would improve this complex mining method considerably. For this reason CRG proceeded with a comprehensive test work campaign to test the viability of optical sorting as part of their feasibility study. 3.3.1. Bulk Test 1 In November 2009 CRG Gold mine sent a ~30t run of mine sample to Mintek for optical sorting test work on the Commodas PRO Secondary COLOR sorting plant, shown in Figure 15. A 30 ton sample arrived pre-screened at 20mm and 75mm. Screened fractions -75+50mm (5.7t) and 50+20mm (9.2t) were used for optical sorting test work.
Figure 15: CommodasUltrasort PRO Secondary COLOR sorting plant - Mintek, Randburg, South Africa
The optical sorter was set up with a ~1t training sample. For the sorting algorithm several different colour classes were set up representing red material (oxidised MRL), pebble material (MR) and all remaining colours. Concentrate was defined as having a certain percentage of the red and pebble colour classes. This percentage was varied to produce different mass pulls to concentrate during set up and was later fixed to produce the final products, resulting in masspulls to concentrate of ~15% for both size fractions. For the test runs, the washing screen was turned on and the feed rate set to ~40t/h for the fines fraction (-50+20mm) and ~60t/h for the coarse fraction (-75+50mm).
13
ALTA Free Library www.altamet.com.au
Figure 16: Optical Sorting Test Flow Sheet
The sample used for this test work turned out to be a low-grade of between 0.5 g/t to 0.7 g/t which was taken from a particularly low grade section of the mine. The sorting results prove conclusively that separation between Main Reef conglomerate and waste quartzite is possible.
Table 4: Optical Sorting Test Work Results Feed Size
Calc. Head Conc Grade Tails Grade Feed Mass Conc Mass Tails Mass Conc Au Recovery [kg] [kg] [kg] Masspull Grade [g/t] [g/t] [g/t]
-75+50mm
5,582
810
4,772
14.5%
0.47
1.71
0.26
52.7%
-50+20mm
4,343
742
3,601
17.1%
0.67
1.67
0.46
42.8%
Figures 17 and 18 show images of the concentrate Main Reef fraction which shows the predominately pebble quarzites. The colours of tailings fraction is mainly grey.
Figure 17: -50+20mm Concentrate
14
ALTA Free Library www.altamet.com.au
Figure 18: -50+20mm Tailings
The assay results show good recognition and differentiation of Gold reef and waste material. The sorter improved the Gold grade in the coarse fraction by a factor of 3.6 and in the fines fraction of 2.5. In both size fractions the mass was reduced to below 20% of the feed. 3.3.2. Bulk Test 2 Following the good results of the large low grade sample a second round of test work was conducted on higher grade material. The objective was to beneficiate three different 5 ton samples with different waste-reef ratios in the size fraction -75+50mm on an optical sorter. In February 2010 CRG delivered -75+50mm run-of-mine which was hand sorted into reef and waste. The reef and waste was then blended at Mintek into 3 sets of samples according to the table 5.
Table 5: Optical Sorting Test Work Results Test No. 1 2 3
Reef (t) 2.50 1.67 1.25
Bulk Sample Testwork W aste (t) Total (t) 2.50 5.00 3.33 5.00 3.75 5.00
Ore:W aste Ratio Size Fraction 1:1 -75mm +50mm 1:2 -75mm +50mm 1:3 -75mm +50mm
The samples were artificially made up by hand picking and mixing/blending reef and waste into these ore/waste ratios. The optical sorter algorithm of the previous test work was used as a basis to separate reef and waste of the different samples. T he sorting algorithm was based on the reef having bright spots in form of pebbles. These were identified by a brightness setting and a filter enhancing large bright spots (pebbles) as opposed to random small bright spots. Concentrate was then defined as having a certain percentage of the bright (pebble) colour class. T he washing screen was turned on for the test runs to present material with a clean surface to the detection system. The resulting sorter products were then hand sorted (as illustrated in Figure 19) to determine: â&#x20AC;˘
The amount of misplaced reef in the waste fraction (or low-grade fraction)
â&#x20AC;˘
The amount of waste diluting the reef concentrate fraction (or high-grade fraction)
15
ALTA Free Library www.altamet.com.au
Figure 19: Evaluation of Sorted Fractions by Hand Sorting
Table 6: Summarises the results of the combined runs for the three mixed ratios:
Ratio
1:1
Ratio
1:2
Ratio
1:3
Stream
Total [kg]
Reef [kg]
Waste [kg]
Reef Content
Waste Content
Reef Recovery
Mass Stream Distribution Grade [g/t]
Reef Grade Gold Content [g/t] [g]
Concentrate
2,048
1,310
738
64.0%
36.0%
77.6%
57.7%
2.6
4.0
5.2
Tailings
1,501
379
1,122
25.2%
74.8%
22.4%
42.3%
1.0
4.0
1.5
Feed
3,549
1,689
1,860
47.6%
52.4%
1.9
4.0
6.8
Stream
Total [kg]
Reef [kg]
Waste [kg]
Concentrate
1,204
554
Tailings
1,068
Feed
2,272
Stream
Total [kg]
Reef Content
Waste Content
Reef Recovery
Mass Stream Distribution Grade [g/t]
Reef Grade Gold Content [g/t] [g]
650
46.0%
54.0%
77.7%
53.0%
1.8
4.0
2.2
159
909
14.9%
85.1%
22.3%
47.0%
0.6
4.0
0.6
713
1,559
31.4%
68.6%
1.3
4.0
2.9
Reef [kg]
Waste [kg]
Reef Content
Waste Content
Reef Recovery
Mass Stream Distribution Grade [g/t]
Reef Grade Gold Content [g/t] [g]
Concentrate
2,101
1,282
819
61.0%
39.0%
85.1%
53.7%
2.4
4.0
5.1
Tailings
1,809
224
1,585
12.4%
87.6%
14.9%
46.3%
0.5
4.0
0.9
Feed
3,910
1,506
2,404
38.5%
61.5%
1.5
4.0
6.0
The results show the following trends and observations: â&#x20AC;˘
The test work shows that run of mine ore can definitely be upgraded to a higher reef content. The lower the reef content in the feed, the more favourable the upgrade ratio becomes. This is also reflected in the calculated concentrate and tailings grades.
â&#x20AC;˘
Reef recoveries lie between 77 and 85% at masspulls between 53 and 58% to the highgrade fraction.
3.4. CONCLUSION OF OPTICAL SORTING OF MAIN REEF The test work has demonstrated that optical sorting is definitely suitable to upgrade run of mine ore from the Central Rand Gold operation in various reef to waste ratios. The sorting process will produce a high-grade and a low-grade product stream. The tests have shown that the more waste is present in the sorter feed, the higher the upgrade ratio to concentrate and the less the gold content in the low-grade fraction.
16
ALTA Free Library www.altamet.com.au
CRG’s feasibility study concluded that optical sorting will be implemented at the mine and it is expected to achieve the following: - Increased throughput (bulk mining - no resue stoping required) - Reduce operating costs (Reduce waste processing through the mill) - Increase ounces recovered (Capture all sweepings and vampings in the form of fine gold)
4. CONCLUSION In today’s and in the future’s mining and mineral processing industry we face increasing challenges such as lower grade ore reserves, high energy cost, water shortages and increasing environmental legislature. Sensor based sorting forms a very significant innovation in mineral processing technology and up-front beneficiation, even though this new technology is only at the beginning of its market introduction. This paper described two different gold beneficiation applications for sensor based sorting technology: •
XRT sorting for narrow vein sulphide type gold ores
•
Optical sorting for Witwatersrand qurz pebble reefs
Both examples have proven to upgrade run-o- mine feed material significantly with a low mass pull to concentrate. By placing a sorter between the mine and the mill will provide numerous benefits by not having to spend effort and energy on crushing, grinding and treating barren material. Mill the ounces, not the waste! 5. REFERENCES [1]
John O. Marsden and C. Iain House. “The Chemistry of Gold Extraction” Published by SME in 2006.
[2]
Wills B. A., 1992 Camborne School of Mines, Cornwall, UK, “Mineral Processing Technology – An Introduction to the Practical Aspects of Ore Treatment and Mineral Recovery” 5th Edition, Pergamon Press.
[3]
Salter, J. D., Wyatt N.P.G.,1991 “Sorting in the Minerals Industry: Past, Present and Future”, Mineral Engineering, Vol. 4, Nos 7-11, pp. 779-796, Pergamon Press, Great Britain.
[4]
Wotruba, H., 2006, “Sensor Sorting Technology – is the minerals industry missing a chance?”, XXIII International Mineral Processing Congress, Istanbul, Turkey, pp. 21-30.
[5]
T. W. Weatherwax, 2007, “Integrated Mining and Preconcentration Systems for Nickel Sulphide Ores”, The University of British Columbia.
[6]
Harbeck H., Kroog H. “New Developments in Sensor Based Sorting”, Montan University Loeben, Austria, January 18, 2008.
[7]
Simmer & Jack; 2008; Developing new Gold Opportunities in South Africa’s oldest Goldfield, Presentation at the Mining Indaba, 5th February 2008.
[8]
Kleine C., Riedel F., von Ketelhodt L., Murray R., “XRT Sorting of Massive Quartz Sulphide Type Gold Ore”; RWTH Aachen; Sensorgestützte Sortierung 2010.
[9]
von Ketelhodt L., “Viability of Optical Sorting of Gold Waste Rock Dumps”; South African Institute of Mining and Metallurgy; World Gold 2009.
[10]
Central Rand Gold, Investor & Analyst Roadshow; September 2009.
17
ALTA Free Library www.altamet.com.au
_
GEKKO SYSTEMS NEW DEVELOPMENTS WITH THE INLINE PRESSURE JIG AND INLINE LEACH REACTOR SYSTEMS By M O Braaksma1 and T Bell2 1
Gekko Systems, Ballarat, Australia 2 Gekko Systems, Perth, Australia
Presented by
Tim Bell TimB@gekkos.com
CONTENTS
1. 2. 3. 4.
INTRODUCTION IPJ DEVELOPMENT ILR DEVELOPMENTS REFERENCES
ALTA Free Library www.altamet.com.au
2 2 7 12
_
1. INTRODUCTION New developments with the Gekko systems InLine Pressure Jig (IPJ) and InLine Leach Reactor (ILR) have allowed the technologies to prosper and provide positive outcomes for customers globally.
A strong focus is placed on continuous product improvement at Gekko Systems and this is particularly so with the IPJ and ILR. Design improvements through past learnings from both a physical design perspective and increased levels of operability through increased automation, have ensured these products provide customers with what they require.
The paper highlights these key areas of product development and how they have contributed to the on going success of the respective technologies.
2. IPJ DEVELOPMENT
2.1.
INLINE PRESSURE JIG (IPJ)
The InLine Pressure Jig (IPJ) is unique in its design and use of jigging concepts. The unit is fully encapsulated and pressurised and combines a circular bed with a moveable sieve action. The encapsulation allows the IPJ to be completely filled with slurry and water. As a result, slurry velocity is slowed and water surface tension eliminated which leads to increased recovery potential. The design allows a wide range of operating conditions to be employed, each specifically targeted to maximise the gravity separation of components in a given feed material.
The screen is pulsed vertically by a hydraulically driven shaft. Length of stroke and speed of up and down stroke can be varied to maximise the gravity separation of components in a given feed material. Screen aperture and ragging dimension and material can also be altered for the application.
Separation of mineral, gem and ore particles occurs based on relative density as well as particle size and shape. High specific gravity particles are drawn into the concentrate hutch during the suction stroke of the bed and are continuously discharged. The lighter gangue is discharged over the tailboard to the outer cone. Both concentrates and tailings are discharged under pressure.
Figure 1: IPJ cross sectional view
2 ALTA Free Library www.altamet.com.au
2.2.
PJ DEVELOPMENTS
CFD Modelling The CSIRO has recently developed a two-dimensional single phase mode of the IPJ to investigate its performance. It was found that slurry recirculates within the deceleration chamber, spreading solid material more evenly across the screens and increasing residence time of the recirculating solid material within the chamber. In the future the model could also be used to investigate the effects of parameters such as bed pulse rate and wave pattern, feed rate, feed properties and ragging makeup, and has currently proved to be a useful tool in the understanding and design of the InLine Pressure Jig.
3
3
(a) Run 1: 3000 kg/m particles
A
B
C
D
(c) Run 3: 3000 kg/m particles
A
E
3
100
B
C
258
D
364
D
E
(d) Run 3: 8000 kg/m particles
A
E
470
C
3
(b) Run 1: 8000 kg/m particles
A
B
576
681
B
787
C
D
893
E
1000
Particle size ( m)
Figure 2: CFD modelling output example â&#x20AC;&#x201C; particle tracking
IPJ Internal Scalping Screen The installation of jigs in the circulating load of a milling circuits, allows the continuous removal of heavy valuable minerals from the circuit, however the presence of circulating steel often requires additional screening equipment ahead of the gravity device to remove coase steel fragments. In order to reduce the complexity of the circuit and the capital required, Gekkos have developed an internal scalping screen that allows the larger steel fragments to bypass the jigging bed.
As shown in figure 3, the scalping screens consist of wedge wire segments that form a sloping screen around the circumference of the bed. The aperature of the screen is typically set at 4 mm which prevents the majority of particles above 2 mm from entering the jigging bed. Maintenance of the screen is faciliated by the ease of removal with just three screws required to be removed to remove all the screen panels.
3 ALTA Free Library www.altamet.com.au
Figure 3: IPJ Scalping Screens
IPJ Automation Since the development of the IPJ, Gekkos has steadily added more supporting equipment to the client packages, in order to improve its operability. Some of the improvements include: •
Flowmeters on all inputs and output streams. This ensures that flows can be monitored and/or controlled, to ensure that the flow velocities are maintained at design levels, reducing the risk of pipeline blockages.
•
Pressure transmitters for internal IPJ pressure. This allows for corrective action to be taken if pressures go outside the normally operating range.
•
Automatic dump valves on tails and concentrate lines. This allows a control system to quickly clear blockages and facilitates start and stop sequences.
•
Sequence start. A one button start to flush the jig, and restart feed, tails and concentrate lines.
•
Sequence stop to fully flush unit of accumulated solids, which allows for a smoother restart.
Figure 4: View of IPJ controls and instrumentation
4 ALTA Free Library www.altamet.com.au
Multiple IPJ systems Multiple IPJ systems have been developed to allow for applications that require dry solids throughputs greater than 100 tonne per hour, with concepts up to 400 tonnes per hour developed. The fully automated operating system applied to multiple jigging systems, controls the even distribution of feed to each primary jig, and allows for easy start-up and shutdown procedures.
As shown in Figure 5, the Jigs are positioned in a light weight modular frame, which is pre-assembled in the factory prior to packing, which significantly reduces the construction time required on-site. The factory assembly also allows for pre-commissioning work to be carried out before shipment, which reduces the commisisoning time required on-site.
Potential applications of the large jigging systems: •
Pre-concentration of either coarse or finely crushed ore to reduce the size and hence capital and operating cost of downstream processing equipment
•
Installed in the circulating load of a large grinding circuit, the jigs can be used to remove coarse free gold and sulphide, which can then be treated separately to increase the overall gold recovery of a circuit. Ideally a dedicated pump at the grinding mill discharge, would feed the gravity module, which reduces the complexity of operations by allowing it to run independently of the cyclone feed pump. The concentrate can be either upgraded in batch centrifugal concentrators such as the InLine Spinner, leached directly in a large InLine Leach Reactor or reground to obtain higher recoveries downstream .
•
For ores that have a high coarse gold/sulphide content, the large jigging circuit can act as the primary recovery unit, in the circulating load of a fine crushing circuit.
Figure 5: Modular pre-concentration plant Python IPJ System The plant, called “the Python Processing Plant” (Gekko, 2007), was the result of a four year research and development program part-funded by the Australian Government and built on Gekko’s 10+ years of experience with high mass pull gravity concentration. This experience included the development of laboratory procedures to characterise orebodies, engineering know how to ensure the gravity circuit operated at its optimum and improvements in the design and control of the key gravity recovery component, the InLine Pressure Jig (IPJ).
The design of the Python is based (Hughes and Cormack, 2008) on two key factors:
5 ALTA Free Library www.altamet.com.au
1.
Concentrate the ore as soon as it is liberated.
2.
Combining continuous high mass pull gravity (the InLine Pressure Jig) with flotation to recover both the fine and coarse particles as soon as possible.
The Python relies on the use of coarse and fine crushing, wet screening, continuous gravity concentration, flash flotation and water recycling (refer to Figure 6) to concentrate greater than 90% of the gold into a high mass pull concentrate of 10 to 40% of the mass.
Figure 6: Python processing plant flow diagram
The Python processing plants have been designed to treat nominally 20 or 50 tonnes per hour of ore to a particle size of approximately 500 Âľm and recover gold and sulphides into the concentrate. The plant is only 3 metres wide (including bolt on platforms) and can be installed on a 1:50 slope making for very simple installation. (refer to Figure 7 and 8). Test work carried out by Gekko Systems Pty Ltd has indicated up to 50% of ores tested to date would recover over 90% of the gold using this method. The projected benefits of this strategy are numerous and include low overall capital and operating cost, significantly lower power consumption compared with traditional milling circuits and low environmental footprint.
Figure 6: Python Flowsheet
Figure 7: Python 500 installed on surface in a â&#x20AC;&#x153;Uâ&#x20AC;? shape
ALTA Free Library 6www.altamet.com.au
Figure 8: Python 500 Gravity Skid
3. ILR DEVELOPMENTS
The Gekko InLine Leach Reactor (ILR) is a modular, skid mounted intensive cyanidation reactor. The unit can be used to leach gold and/or silver from medium to high grade gravity and flotation concentrates. The design is based on the same principles as the laboratory bottle roller. In its batch form (ILR-BA), the units available can process between 0.9 to 24 tonnes per batch. While in its continuous form, the largest unit currently available can process up to 1.9 t/hr (based on a residence time of 6 hours).
The intensive cyanidation reaction is typically performed at 2% NaCN levels, 8-20ppm O2, 30% solids and requires no exotic chemicals or materials. The oxygen is supplied to the unit as gaseous oxygen or as 25-50%w/w hydrogen peroxide.
In the batch ILR concentrates are leached and pregnant solution is produced and electrowon in discrete batches before being returned to the CIP plant. In the continuous ILR the concentrate is fed, leached and discharged continuously while a recirculating stream of leach solution passes continuously from the reactor to solution recovery â&#x20AC;&#x201C; typically electrowinning â&#x20AC;&#x201C; and back to the reactor to be reused.
7 ALTA Free Library www.altamet.com.au
3.1.
ILR DEVELOPMENTS
Batch ILR System The ILRBA has evolved over the course of the past 8 years since the first units were installed at the Target Mine in RSA and Kelian Equitorial Mining in Indonesia. The level of automation has increased and the PLC control sequence has been refined to produce a truely hands free system.
Figure 9: Batch Reactor Flow Diagram ILRBA - Single Solution Cone One of major design improvements to take place with the batch automatic unit was the change to a single solution cone design in 2002, which allowed all the solids, including fines, to be efficiently recaptured in the drum during the dedicated clarification stage. Earlier batch manual units ran with a 2 stage solution cone approach which added operational complexity without providing any genuine technical benefit. The single solution cone design and the associated control sequence allows the ILR to prepare, circulate and clarify pregnant solution in the one vessel. It also provides ideal control from a measurement perspective because the solution mass is measured on a single load cell which provides optimum reagent and gold accounting measurements.
Figure 10: Batch ILR within bunded area
8 ALTA Free Library www.altamet.com.au
ILRBA-Single Pump System The new generation batch automatic syetem was also refined into a one pump design in 2002. The design ensures absolute minimum pump maintenance and is made possible by the PLC controlled valve distribution network that directs the product flow depending on which stage the program is in. It is now proven across more than 50 installations and has been a key factor in the batch automatic units continued growth and market dominance.
ILRBA - Custom oxygen sparger The refined and improved oxygen sparging unit is standard with all ILRBA installations. A major advantage of the ILR’s rolling drum technology is it’s ability to generate accelerated kinetics as a result of the mechanical agitation within the drum. The reaction zone is supplied with oxygen by sparging gaseous oxygen into the recirculating leach solution as a dispersion of fine bubbles. This is achieved using a spring loaded ceramic tipped sparger. This principal of metallurgical operation has ensured that chemical oxidants are generally not required and the refinement of the sparging unit has further enhanced this. Oxygen is the most common oxidant in use with the BA units globally and recoveries >98% are typical.
Continuous ILR System The continuous ILR was first pioneered by Gekko Systems for the purpose of treating IPJ concentrates, which are typically high mass, medium grade, coarse concentrates. However the majority of early applications were to treat low mass high grade gold concentrates produced by BCCs (Falcons, Knelsons and InLine Spinners) which typically are low mass, high grade and can contain significant levels of slimes. The new generation batch ILR replaced the original continuous units in these applications from about 2002, however the ILR “continuous” remains a critical unit within Gekko Systems product range and has now found its design target as the treatment of high mass gravity and flotation concentrates becomes more prevalent.
Figure 11: Continuous ILR flowsheet The new generation continuous ILR’s are specifically designed around the treatment of medium to high mass applications and generally treat lower grades than the batch automatic units. They provide the ideal solution for those operations wishing to leach flotation concentrates and or concentrates collected from continuous gravity circuits. The multiple leaching system shown in Figure 12, allows for the treament of up to 20 tonnes per hour of concentrate based on a residence time of 4 hours.
9 ALTA Free Library www.altamet.com.au
Figure 12: Continuous ILR Leaching System
CCD-Filter circuits A key development in the ILR”CA” range has been the addition of high efficiency solid/liquid seperation options. The early continuous units that treated predominantly low mass high grade concentrates were not required to obtain high solution recovery since all effluent streams were returned to a CIP plant where residual dissolved gold was recovered. This could be up to 10% of gold dissolved in the ILR. The new continuous ILR’s are usually the primary gold leaching operation and therefore soluble gold recovery becomes critical for overall plant recovery. To obtain the necessary high soluble gold recovery continuous ILRs are normally packaged with either: •
A 3 stage counter current decantation process which not only ensures maximum gold solution recovery through multiple wash steps but provides a clear solution for the “in circuit” electrowinning process.
•
Automated plate and frame filters, which can produce much higher wash efficiencies than CCD’s, resulting in higher soluble gold recovery, however usually at the price of a higher capital cost.
Super ILR As the continuous ILR has evolved there has been a higher degree of interest in utilising the benefits of the ILR technology for higher and higher mass yields of concentrate. It is for this reason that Gekko have now finalised design of the super ILR technology. It allows for the treatment of far larger mass treatment rates (8 tonnes per hour) whilst reducing the number of individual drums. This ensures that capital costs remain competitive with other options and expands the application of the technology.
Gravity Box Whether it be a greenfields site designing a gravity circuit from scratch or brownfields opportunity where a customer is looking to optimise or upgrade a curcuit, there is far more to a gravity circuit than may initailly appear. Traditionally the various pieces of the gravity circuit puzzle are purchased independently and designed by a site projects team and/or engineering company. Increasingly in recent times Gekko Systems have engineered complete “gravity box” systems to provide a complete solution for the unit processes required.
10 ALTA Free Library www.altamet.com.au
Gekko utilise our extensive in house gravity expertise to provide not only the equipment package but a fully engineered, designed and constructed package which incorporates steel frames, piping, electrics, automation, installation and commissioning. It is fully engineered, factory pre-assembled, and only requires the connection of reagents and services, power input and product feed. It can be incorporated into a new plant design or simply bolted onto an existing plant structure.
We supply and install the best option according to: •
Gravity Recoverable gold potential test work
•
Throughput requirements
•
Client preference on technology
•
Budget requirements
•
Client preferred scope of work (ie, gravity box package or just equipment package supply only)
Current standard “gravity box” screening/gravity concentration/intensive cyanidation options available are: •
50 tph free Au
•
100 tph free Au
•
200 tph free Au
•
300 tph free Au
Figure 13: Gravity Box System – Magscreen, BCC, ILR
11 ALTA Free Library www.altamet.com.au
4. REFERENCES 1. Gray, S., and Hughes, T., 2007. “The Modular Python Processing Plant – Designed for Underground Pre-Concentration”, paper presented to SME2009. 2. Solnordal, C., Hughes, T., Gray, S., and Schwarz, P., “CFD Modelling of novel gravity separation device”, Seventh International Conference on CFD in the Minerals and Process Industries, CSIRO Melbourne 9-11 Dec 2009.
12 ALTA Free Library www.altamet.com.au
DESWIK MILLS
THE DESWIK FINE GRINDING MILL MAKING THE FINE GRIND OF REFRACTORY GOLD ORES AND RETREATMENT OF WASTE TAILINGS SIMPLER THAN EVER BEFORE By Stephen Massey stephenm@deswik.com
Deswik Mining Consultants
1
About Deswik • Consulting, Software (Mine 2-4D Mine planning), Milling Technology • Offices Cardiff, Johannesburg, Brisbane, Vancouver • Joint Venture Agreement with Knelson Concentrator Technology • Manufacturing to now be with Knelson Concentrator Technology in Canada
1 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Deswik Experience • Refractory Gold Ore – Fine grind to unlock finely trapped particles • Concentrate regrinds – Copper-Gold float cons put back through the Deswik and grinding from F80 50 to P80 20 microns • Tailings retreatment – Find a waste dump and regrind it followed by a CIL process (for example) • Fine Grind Iron Ore Magnetite
Deswik operating principals
Power transfer
Open to atmosphere open circuit
Product
RPM’s 200 – 280 11 – 14m/s Nm 4,000 – 10,000
Attritional grind
Media Fill 60 – 85% Feed Why fine grind
2 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Grinding Media •
The Bead media is based on (TZP). Tetragonal Zirconia Polycrystals.
•
The Beads have superior microstructure with a fine uniform grain of consistent size distribution.
•
These properties offers a grinding media with a very high fracture toughness, hardness and density.
Feed and Recovery Size • Feed Size?
• Recovery Size?
• <300 Microns
• 40 microns • 25 microns • 15 microns • 5 microns • Controllable and possible in a single, self classifying pass
3 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Fine grind within PGM circuit
Deswik 1000 power consumption
An average of 5.0 kWh/t over the eight month period Nov 07 to Jun 08
4 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Working Costs Over an 18 month production period approximately 300, 000 dry tons produced Cost per ton Grinding media consumption (R900,440) R3.00/$0.45 Impellor replacement 34 Disks & 5 spacers (R109,950)
R0.37/$0.06
Spare lined barrel (R304,000)
R1.01/$0.15
Wedge wire replacement (R15,400)
R0.05/$0.01
Energy costs at an average of 6.5 kWh/t
R0.88/$0.65
Ongoing monthly support and maintenance (R14,400pm)
R0.53/$0.08
Total working costs over 18 months R1, 477, 500 @ a cost per dry ton produced of R5.84 = AUD $1.40 per dry tonne throughput
5 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Example of Fine Grind Circuit
Rougher Con
Deswik Mill
Flotation Cell
$$$
Example of Find Grinding Circuit
6 ALTA Free Library www.altamet.com.au
DESWIK MILLS
Mill Deliveries • • •
Deswik can deliver and commission within 6 months of a firm order. Deswik can mobilise to almost all remote and difficult locations in the world Deswik are currently installing 3 large Deswik 2000 Mills in Kazakstan
Mill Estimate/Testing
7 ALTA Free Library www.altamet.com.au
DESWIK MILLS
The Deswik 25 Pilot Mill
References
- Kazakhstan 3 x Deswik 2000s
- Brisbane Deswik 25 Pilot - South Africa 1 x Deswik 500 - Perth Deswik10 AMMTEC - South Africa 2 x Deswik 2000s - South Africa 1 x Deswik 1000
8 ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
LEACHING PROCESSES
ALTA Free Library www.altamet.com.au
ONLINE CYANIDE MEASUREMENT AND CONTROL FOR COMPLEX ORES
By
P Breuer* and P Henderson** * Parker CRC for Integrated Hydrometallurgy Solutions CSIRO Minerals Down Under National Research Flagship CSIRO Process Science and Engineering, Australia ** Orica Australia Pty Ltd
Presented by
Paul Breuer paul.breuer@csiro.au
CONTENTS
1. 2. 3. 4. 5. 6. 7.
INTRODUCTION CYANIDE MEASUREMENT ONLINE CYANIDE MEASUREMENT CYANIDE CONTROL CONCLUSION ACKNOWLEDGEMENTS REFERENCES
ALTA Free Library www.altamet.com.au
2 2 8 11 11 12 12
1. INTRODUCTION The gold industry in the past couple of decades has looked to implement better control of cyanide addition to the leach circuit in order to optmise cyanide usage, subsequent cyanide destruction costs and the environmental risks. The implementation of online cyanide measurement as a feedback control parameter has proven problematic in several cases. The issues can be categorised as: Analyser reliability/acceptance which includes: •
limited output,
•
filtration issues,
•
maintenance, and
•
differences between laboratory and analyser readings.
Solution complexity/interferences such as: •
sulphides,
•
metal cyanides, and
•
pH.
Control strategy/parameters: •
cyanide measurement variability,
•
tank residence time, and
•
cyanide demand.
With the gold industry looking to process more complex ores, the control of cyanide addition to the gold leaching circuit has become more crucial. On-line cyanide measurement provides a rapid response to changes in ore mineralogy, however poor understanding of the methods and/or reliability issues described above have impaired their uptake within the gold industry. This paper discusses the potentiometric end-point silver nitrate titration method, the most common on-line analysis technique adopted for “free” cyanide, and its comparison to the common site laboratory rhodanine endpoint silver nitrate titration method. In particular the influence of metal cyanides and the effect of titration pH on the “free” cyanide measurement using these two methods are presented and discussed. At some operations, the presence of clay minerals and/or fine particle size (e.g. fine grinding used in order to expose the gold), results in difficulties in maintaining filtered solution to the analyser. Examples of these issues are presented and discussed in this paper along with some findings and knowledge gained from investigations into many of these issues. Some solutions are proposed whilst areas needing further investigation are highlighted.
2. CYANIDE MEASUREMENT Methods commonly used to measure cyanide concentration determine either: •
free (titratable) cyanide,
•
weak acid dissociable (WAD) cyanide, or
•
total cyanide.
The cyanide concentration of interest in the leaching of gold is the cyanide available in solution which can leach gold. The free cyanide methods measure the free and very weakly complexed (such as the fourth cyanide complexed with copper) cyanide ions and thus provides a cyanide determination that best reflects the available cyanide for gold leaching; some weak acid dissociable cyanide ions, such as the third cyanide complexed with copper, are also available to leach gold, though the gold leach rate is much lower than for free cyanide (Breuer et al., 2005). However, the free cyanide measurement can be misleading in the treatment of more complex ores where achieving a leach pH above 9.5 is uneconomical or copper is present in solution (Breuer and 2
ALTA Free Library www.altamet.com.au
Rumball, 2007). The most commonly used methods for determining the free cyanide involve titration with silver nitrate using either rhodanine indicator or the potential of a silver wire to ascertain the endpoint. To interpret the free cyanide titration results it is vital to know the solution species present and the influence these may have on the free cyanide determination. This not only includes the metal cyanides, but also other anions in solution which can complex with silver (for example sulfide and thiosulfate). Traditional analyses for elements such as AAS and ICP can quantify the metal ions, but techniques such as HPLC are required to provide information on the various anion species (Breuer et al., 2009). 2.1.
SILVER NITRATE TITRATION WITH RHODANINE ENDPOINT DETECTION
For simple solutions the colour change of the rhodanine indicator is quite sharp. This is shown in Figure 1 where a sharp increase in UV absorbance of the titration solutions occurs at the endpoint for a solution containing only sodium cyanide. Also shown is the much slower colour change observed when copper is present, which makes identifying the endpoint more difficult and as such interpreting the end-point can produce significant variation in results by different operators.
UV absorbance (530 nm)
0.12 No Cu 0.1
Low Cu High Cu
0.08 0.06
Visual EP Region 0.04 0.02 0 0
2
4
6
8
10
12
14
AgNO3 (mL of 0.0103 M) Figure 1: UV absorbance of rhodanine indicator during titration of 250 mg/L NaCN solution with silver nitrate; low Cu – 115 mg/L Cu as Cu(CN) 32-, high Cu – 530 mg/L Cu as Cu(CN)32-. As a solution containing cyanide is titrated with silver nitrate, the added silver ions complex the free cyanide ions (Equation 1). The silver ions form a more stable complex with cyanide than rhodanine and thus a colour change from silver complexing with the rhodanine (Equation 2) does not occur until all the free cyanide has been complexed with silver. Further addition of silver nitrate then results in precipitation of silver cyanide (Equation 3).
Ag + + 2CN − → Ag(CN) -2
(1)
Rh (yellow) + Ag + ↔ Rh - Ag (Pink)
(2)
Ag + + Ag(CN)-2 → AgCN
(3)
With copper present in the cyanide solution the titration becomes more complex as the cyanide associated with the copper can also be released and be complexed with silver ions. The fourth cyanide complexed with copper is weak and thus is titrated (Equation 4) along with the free cyanide 3
ALTA Free Library www.altamet.com.au
(see potentiometric endpoint determinations below). The cyanide ions complexed with copper tricyanide can also complex with silver ions (Equation 5), which according to Figure 1 is a competing reaction with rhodanine (Equation 2). Hence, the rhodanine colour change becomes broader and the endpoint more difficult to identify as the copper concentration increases.
2.2.
Ag + + 2Cu(CN) 34- → Ag(CN) -2 + 2Cu(CN) 23-
(4)
Ag + + 2Cu(CN)32- → Ag(CN)-2 + 2Cu(CN) -2
(5)
SILVER NITRATE TITRATION WITH POTENTIOMETRIC ENDPOINT DETECTION
A potentiometric endpoint determination uses the measured potential of a silver wire in the titration solution as an indicator of changes in solution speciation as silver nitrate is added. The change in the potential of a silver wire during the silver nitrate titration of cyanide solutions in the absence and presence of copper is shown in Figure 2. For the solution without copper, a sharp change in the measured potential of the silver wire, and hence a maximum in the potential change, occurs when all the cyanide has been complexed with silver ions. Notably, in the presence of copper the endpoint is not changed (cf. Figure 1 for rhodanine), though the potential change is less sharp. This potentiometric endpoint corresponds with the free cyanide ions and the fourth cyanide associated with copper (i.e. each copper at the end-point has three cyanide ions remaining complexed with it). Thus, in copper cyanide solutions the rhodanine method will give higher cyanide results than the potentiometric method, with the extent dependent on the copper concentration.
100
200 Potential
0 -100
Potential change
-200 -300
180 160 140 120
Cu(CN)43- → Cu(CN)32- + Ag(CN)2-
CN → Ag(CN)2
-400
Cu(CN)32- → Cu(CN)2- + Ag(CN)2-
-
100
-500
80
-600
60
-700
40
-800
20
-900
Potential change
Potential (mV)
No Cu Low Cu High Cu No Cu Low Cu High Cu
0 0
1
2
3
4
5
6
7
8
9
10
AgNO3 (mL of 0.0103 M) Figure 2: Silver wire potential during silver nitrate titration of the same cyanide solutions as in Figure 1. Figure 3 shows a comparison of potentiometric (off-site laboratory) and rhodanine (operator) titrations for plant solutions over a 12 month period; copper concentrations in the leach solutions were typically 200 – 500 mg/L. These results are consistent with the effect of copper described above. The rhodanine titration results are around 200 mg/L higher than those determined by potentiometric titration. Further investigations are currently being conducted to establish a correlation between the two cyanide measurements.
4
ALTA Free Library www.altamet.com.au
1000
NaCN (mg/L) - Potentiometric
900
Leach Tank 1
800
Leach Tank 2 Leach Tank 8
700 600 500 400 300 200 100 0 0
200
400
600
800
1000
1200
NaCN (mg/L) - Rhodanine Figure 3: Comparison of potentiometric (off-site laboratory) and rhodanine (operator) titrations for plant solutions over a 12 month period (typically 200-500 mg/L copper).
2.3.
OTHER INTERFERENCES TO THE SILVER NITRATE TITRATION
In the extraction of gold from sulfide ores using cyanide, there is the potential for sulfide ions to form in solution at the beginning of the leach. If sulfide ions are present in the cyanide solution, silver sulfide precipitates rapidly upon the first addition of silver nitrate (Equation 6). This masks the rhodanine colour and thus the sulfide ions must first be removed by the addition of lead(II) ions with the pH maintained above 11 (Breuer and Rumball, 2007). The silver sulfide forms preferential to silver cyanide and thus doesn’t have to be removed for a potentiometric titration (Figure 4); this also allows quantification of the sulfide ion concentration.
2Ag + + S 2− → Ag 2 S
(6) -
Depending on the pH of the cyanide solution being titrated, both CN and HCN(aq) can be present. The CN-/HCN(aq) distribution as a function of pH is described by Equation 7 (pKa = 9.2). HCN(aq) is not measured, but is included in the determination if there is sufficient buffer present (Breuer and Rumball, 2007). The addition of NaOH to the titration is used where there is insufficient buffer to remove the influence of pH on the cyanide measurement; low pH and insufficient buffer are indicated by a difference in measured cyanide values with and without the addition of NaOH.
[CN- ]
= 10 (pH−pKa)
(7)
[HCN(aq) ]
5
ALTA Free Library www.altamet.com.au
100
200 Potential
80
Potential change
60
-200
Ag+ + 2CN- ⇒ Ag(CN)2-
-400 +
40
2-
2Ag + S ⇒ Ag2S
-600
Potential change
Potential (mV)
0
20
0
-800 0
2
4
6
8
10
12
14
AgNO3 (mL of 0.0103 M)
Figure 4: Potentiometric silver nitrate titration of cyanide solution (2 mL, 4000 mg/L NaCN) containing 270 mg/L sulfide ions. Zinc, if present in the gold ore, may also dissolve during cyanidation being complexed by cyanide ions at pH’s less than 12, or hydroxide ions at high pH’s. Rhodanine titration of zinc cyanide solutions of various pH values in the range 9 – 13 found that the measured cyanide was independent of pH and equivalent to the total cyanide in solution. This indicates that the cyanide ions complexed with zinc are only weakly bound and are easily released and complex with silver ions in preference to rhodanine. However, for the potentiometric titration, the presence of zinc ions complicates the results if the pH is below 12 (Figure 5). At high pH (NaOH added to the titration) the zinc is complexed with hydroxide ions and thus a single titration endpoint is observed corresponding to the total cyanide. Without NaOH addition (pH = 10.5) the zinc is complexed with cyanide and two additional transitions (maxima in the rate of potential change) are observed at the addition of 2.5 and 3.8 mL of silver nitrate solution. These respectively correspond closely to four and three cyanide ions remaining complexed with the zinc. Depending on solution composition these peaks independently become more or less distinguishable and thus in some cases just a single inflection is observed. Thus, to avoid incorrect or varying automated endpoint identification, the determination is best simplified with the addition of NaOH to the titration. Also, as the cyanide complexed with zinc is only weakly bound (weaker than the third cyanide associated with copper which is known to leach gold), it is most likely available for gold dissolution and thus with the addition of NaOH to the titration provides a cyanide measurement applicable to that available for gold leaching.
No NaOH
NaOH added
100
80
100
100
Potential
-100
-200 40 -300
-400
Potential
20
80
Potential change
0 -100
60
-200 40
-300
Potential change
60
Potential change
Potential change
Potential (mV)
0
Potential (mV)
200
-400 20
-500
-500
-600
0 0
2
4
6
8
-600
10
0 0
AgNO3 (mL of 0.0103 M)
2
4
6
8
10
AgNO3 (mL of 0.0103 M)
Figure 5: Potentiometric silver nitrate titration of cyanide solution (5 mL, 1500 mg/L NaCN) containing 320 mg/L zinc ions without (left) and with (right) NaOH added.
6
ALTA Free Library www.altamet.com.au
The effect of dissolved zinc on the online potentiometric cyanide measurement at a gold operation is shown in Figure 6. The Orica OCM5000 cyanide analyser used in this case identified the endpoint within a defined potential window as the largest change in the measured potential of the silver wire in the titration solution. This potentiometric titration curve has low changes in the measured potential (differential) and two peaks in the differential curve are observed within the potential window. The relativity of these two differential peaks was observed to vary between measurements and on occasions the first peak had the higher potential change. This may account for some sharp changes and lower cyanide concentrations determined by the online analyser measurement previously.
0
40
Potential Potential change
Potential (mV)
30 -200
-300
20
-400 10
Potential change ("D")
-100
-500
-600
0 0
0.5
1
1.5
2
2.5
AgNO3 (mL)
Figure 6: Orica OCM5000 online potentiometric silver nitrate titration of a cyanidation leach solution (12 mL sample, 5.199 g/L silver nitrate). Online analysers typically donâ&#x20AC;&#x2122;t store the titration data or display the titration curve for interpretation, and connection to a PC is thus required to obtain this information; hence this information may only be accessible by the vendor. This obviously impedes the troubleshooting of the analyser and has no doubt contributed to the lack of acceptance and implementation at operations. Manual potentiometric titration curves for the same solution without and with the addition of NaOH are shown in Figure 7. Without the addition of NaOH a similar titration curve is observed to that shown in Figure 6: Orica OCM5000 online potentiometric silver nitrate titration of a cyanidation leach solution (12 mL sample, 5.199 g/L silver nitrate). , however only one endpoint is observed with NaOH added to the titration. This single peak with NaOH added corresponds closely with the second peak without NaOH addition. The concentration of zinc in the solution also supported that the first of the two peaks was due to cyanide ions complexed with zinc. In this case the potential window of the online analyser was reduced (starting potential increased) such that the first differential peak observed at around -375 mV was excluded from the endpoint window. An alternative is to retrofit the on-line analyser to add some caustic to the titration (as is done at some operations where the titration pH is raised to avoid unmeasured HCN, the quantity of which is dependent on pH). The accuracy of the endpoint with 0.1 mL additions used in the online titration is Âą15 mg/L NaCN for a 10 mL titre; other errors such as variability in titre volumes would add to this. The accuracy of the cyanide measurement may be improved by decreasing the silver nitrate concentration, though limited by the maximum cyanide concentration to be encountered and the titration vessel volume.
7
ALTA Free Library www.altamet.com.au
NaOH added 100
70
0
60
0
60
-100
50
-100
50
-200
40
-200
40
-300
30
-300
30
-400
20
Potential Potential change
-500
-600 0
0.5
1
1.5
2
2.5
3
Potential (mV)
-400
10
-500
0
-600
3.5
20
Potential Potential change
Potential change
70
Potential change
Potential (mV)
No NaOH 100
10
0 0
0.5
1
AgNO3 (mL)
1.5
2
2.5
3
3.5
AgNO3 (mL)
Figure 7: Manual potentiometric titration curves for the same solution analysed in Figure 6 without (left) and with (right) NaOH added (10 mL sample, 5.199 g/L silver nitrate).
3. ONLINE CYANIDE MEASUREMENT
3.1.
SAMPLING VARIABILITY
A survey was conducted at a gold operation over the top of the first cyanide leach tank at the points identified in Figure 8. Cyanide is added into this tank in the downcomer with the inlet feed. The variability in the free cyanide measurements by potentiometric titration (Table 1) was 50 mg/L NaCN between the highest and lowest across the points surveyed with a much lower variability of less than 5 mg/L NaCN observed in the majority of the duplicate assays. The observed variability between sample points gives a good indication of the variability expected in the operator and online analyser measurements from the variability within the tank (a similar assessment could also be made by sampling the same point several times at one minute intervals and analysing). If it is assumed operator and analyser variability is around 15 mg/L NaCN then the variability of cyanide measurements will be at best Âą40 mg/L of the average tank concentration.
Table 1: Cyanide concentrations at the top of the leach tank for the positions shown in Figure 8.
Position
1 2 3 4 5 6
Cyanide (mg/L NaCN) A B 733 732 694 693 705 706 685 685 712 709 727 717
8
ALTA Free Library www.altamet.com.au
2
Sample points across top of tank
3
Agitator
Outlet 4
Operator sample point
1
5 6
Inlet Figure 8: Plan view of the location points where samples were taken at the top of the first leach tank.
3.2.
ANALYSER PERFORMANCE
The long term successful use of an Orica OCM5000 on-line cyanide analyser comes down to the implementation of an appropriate site specific preventative maintenance scheme. It should be noted that the daily preventative tasks are not onerous and that for operations that do have high reliability and accuracy, activities take no more than about 20 - 30 minutes per day. The most common causes observed in recent times for poor reliability and accuracy of on-line cyanide analysers can be reduced down to three key areas: 1. Poor filtration. 2. Poor condition of electrodes in titration cell. 3. Incorrect settings and more importantly knowledge of how the settings can influence the accuracy of the result. The same or similar issues are encountered with all the commercial online cyanide analysers. 3.2.1.
Filtration
Experience has shown that most sites which experience poor short term reliability and accuracy with the Orica OCM5000 free cyanide analyser have well below optimum filtration rates and should always be the first place to investigate when it is believed there are errors with the analyser readings. Operations have been observed to experience reduced readings by as much as 40 %, believed to be due to volatilisation of the cyanide in the sampling line and reservoir. Much lower readings are obtained if insufficient filtrate is attained in the reservoir for the titre volume needed. Blocking of the filter is highly dependant on the slurry particle size distribution and properties, particularly clay content. The low rate of filtration and excessive aeration also promotes the formation of scale which has shown to affect both the accuracy and the precision of the analyser by partially or totally blocking filtrate lines. As a general rule of thumb, to maintain optimum filtration the filters should be cleaned of any solids build up at the beginning of each shift by the operators, and on a weekly basis removed for further cleaning in acid followed by cleaning in an ultrasonic bath. Operations using analysers in an hyper saline environment, solutions that have a high scaling tendency or highly viscous slurries will need to at least double the frequency of cleaning to maintain optimum rates of filtration, which is typically >1 mL/sec. As a general guide the filter of choice is a stainless steel sintered filter with a nominal aperture of 10 Âľm. This filter in all cases provides a very clean filtrate however is prone to blockages more
9
ALTA Free Library www.altamet.com.au
easily than the alternative, which is a cloth filter with nominal aperture of 20 Âľm. Cloth type filters have been shown to be more beneficial in viscous slurries. Trials of alternative cloth material in different configurations and filter mediums are currently being explored with initial success being experienced with a very cost effective sock type system. Other key variables for optimum filtration to be mindful of are: â&#x20AC;˘
Keep the sample lines as short as possible, i.e. <20 m. If greater than 20 m then investigate the use of a pumping system to transfer slurry to a dedicated hopper adjacent to the analyser hut.
â&#x20AC;˘
Air back flush valve is operational and operating at >200 kPa.
3.2.2.
Titration Cell
The biggest cause observed for poor longer term (weeks) accuracy and/or precision in online potentiometric analysis is poor condition of the electrodes. Typically, it is the reference electrode that is in a poor condition with some cases highlighting that the reference electrode has had no maintenance performed on it for over twelve months. The reference electrode is a Ag/AgCl reference that is located in a double junction reference electrolyte. In nearly all cases of poor accuracy and precision, the silver chloride coating has been lost and the electrolyte has been contaminated. These conditions result in the standard potential drifting and a decrease in potential readings observed between the reference and indicator electrode. The end result is that the analyser may no longer detect an endpoint within the potential window and if a default potential is used then there will be a drift in the endpoint. As a general guide, the titration cell and electrode/electrolyte conditions should be checked on a daily basis and maintained in optimum condition. The reference electrode electrolyte levels should be maintained and regularly (approxiamtely bimonthly depending on conditions) replaced to avoid the build up of contaminants. Alternative reference electrodes (gel and solid state) being trialled for WAD cyanide analysis may prove to be a more reliable alternative. The silver wire needs to be lightly polished on a regular basis to avoid the build up of contaminates and scale. 3.2.3.
Titration settings
The Orica OCM5000 cyanide analyser (like most potentiometric on-line cyanide anlysers) has the ability to detect an endpoint based on a peak in the potential change above a given threshold being identified within a defined potential window. There is also the ability for an endpoint to be defined based on a set default potential. If set, this default potential endpoint is reported by the analyser when no endpoint is detected by the peak detection method. Typical practice is to set parameters for both peak detection and a default potential based on the potential reading observed for the peak detection method. However, if the peak detection method is unable to identify a peak then the validity of a default potential endpoint would be highly questionable. For example, the most common reason this situation occurs is when the reference electrode potential drifts and hence the set default potential is no longer correct, yet the analyser will still be reporting a result that is now inaccurate and can create significant issues if being used to control cyanide addition. Thus, when using peak detection it is recommended that a default potential should NOT be set as the analyser will then not report a result if no peak is identified and thus will flag an issue with the analyser. Some operations set the peak threshold high such that the default potential is used. Where there is a difference in the rhodanine titration to the potentiometric endpoint (ie. when copper is present) the rhodanine potential can be used to set the default potential such that comparable results are obtained. However, a clear understanding of the solution speciation and the effect changes in the solution speciation have on the determination are required. There is also no alert from the default endpoint detection method of any issue with the analysis unless the measured starting potential is above the default potential. Thus, close monitoring and cross checking of the online analyser results with laboratory analyses is required to identify when there may be an issue with the online analyser. Hence, operation of analysers using the default potential is not recommended.
10
ALTA Free Library www.altamet.com.au
4. CYANIDE CONTROL The residence time of the leach tank in which the cyanide measurement is being taken and the error associated with this measurement, establishes the minimum response time in which changes to the system that produce a change in the cyanide concentration greater than the measurement error can be detected. This minimum response time creates a time lag in providing feedback control to the system. For example, if a leach tank has a 24 hour residence time, the theoretical change in cyanide concentration with a step change in cyanide addition will take several days to be fully realised (Figure 9). Most notable is that in this case only 50 % of the resultant change is observed in the cyanide concentration after 16 hours. If there is an error of Âą30 mg/L in the cyanide measurement (for complex solutions using the rhodanine method the error would be higher) then at least 12 hours is needed for the change to be outside the operator range. In such a case, it can be seen that changes between individual cyanide analyses conducted on a 3 to 6 hourly basis could be misleading. Thus, it is recommended that a moving average is used to follow the cyanide concentration in the tank to modulate the variability in individual values and establish a clear trend. The control strategy should utilise the cyanide trend and have a minimum of 12 hours pass between making adjustments to the cyanide addition rate (at this point ~40 % of the resultant change will have been observed when all else has remained constant).
700
Cyanide (mg/L NaCN)
600 500 400 True value 300
Operator range
200 100 0 0
1
2
3
4
5
Time (days)
Figure 9: Leach Tank (24 hour residence time) response to a step change in cyanide addition that would result in a change in the measured cyanide concentration from 400 to 600 mg/L NaCN with everything else remaining constant. If a large portion of the cyanide addition to the leach circuit is due to ore consumption, small changes in the ore consumption will result in large changes in the measured free cyanide. Thus, with only cyanide addition into the first leach tank it is difficult to control the cyanide addition in order to minimise the cyanide concentration and variability in the leach discharge (and thus minimise the cost of cyanide destruction if implemented). The major cyanide consumption of an ore typically occurs in the first leach tank, thus implementing cyanide measurement along with controlled cyanide addition on the next or subsequent leach tank can substantially reduce the cyanide variability in the leach discharge. In this situation the cyanide addition to the first leach tank can be reduced which has the economic benefit of essentially eliminating cyanide overdosing (increased reagent costs). This also minimises gold losses to tails that would otherwise occur when cyanide is under dosed in the first tank.
5. CONCLUSION This paper has discussed a number of issues that appear to have impaired the uptake of potentiometric on-line cyanide measurement to control cyanide addition within the gold industry. 11
ALTA Free Library www.altamet.com.au
These include a poor understanding of the potentiometric and rhodanine analysis methods, poor reliability of the analyser, and/or an inadequate control strategy. Understanding the effect of solution composition (eg. presence of copper or zinc) on the silver nitrate titration with potentiometric or rhodanine endpoint detection is vital in interpreting the difference between these two deteminations for complex solutions. In such solutions the potentiomtric endpoint provides an accurate determination of the â&#x20AC;&#x153;freeâ&#x20AC;? cyanide concentration available to readily leach gold. Most importantly, a maintenance schedule must be established for the online analyser and adhered to if reliable results are to be maintained. Peak detection is recommended with no default potential set, in which case any issues with the analyser become apparent when no result is output. Using the default potential is not recommended unless close monitoring and cross checking is conducted with laboratoty analysis. Using an online cyanide measurement for the control of cyanide addition requires a good understanding of the analysis error/variability and system residence time relative to the frequency of analysis. Where a large portion of the cyanide added is consumed, then a second cyanide measurement and addition point will allow for better control and optimal cyanide addition.
6. ACKNOWLEDGEMENTS The support of the CSIRO Minerals Down Under National Research Flagship and Parker CRC for Integrated Hydrometallurgy Solutions (established and supported under the Australian Governmentâ&#x20AC;&#x2122;s Cooperative Research Centres Program) is gratefully acknowledged. The authors would like to thank Barrick Australia Pacific for their support and provision of plant data presented in this paper.
7. REFERENCES Breuer, P. L., Dai, X. and Jeffrey, M. I., 2005. Leaching of gold and copper minerals in cyanide deficient copper solutions. Hydrometallurgy, 78 (3-4), 156-165. Breuer, P. L., Hewitt, D. M., Sutcliffe, C. A., and Jeffrey, M. I., 2009. The quantification of cyanide and its reaction products during leaching and cyanide destruction processes. In: The Southern African Institute of Mining and Metallurgy (Ed.), World Gold Conference 2009. The Southern African Institute of Mining and Metallurgy, pp. 287-293. Breuer, P. L. and Rumball, J. A., 2007. Cyanide measurement and control for complex ores and concentrates. In: Ninth Mill Operators' Conference, AusIMM, Carlton, Victoria, Australia, pp. 249254.
12
ALTA Free Library www.altamet.com.au
ALTERNATIVE LIXIVIANTS TO CYANIDE
By
M G Aylmore Bateman Engineering Pty Ltd, Australia
Presented by
Mark Aylmore mark.aylmore@batemanengineering.com.au
CONTENTS
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
INTRODUCTION THIOSULFATE LEACHING THIOUREA LEACHING HALIDE LEACHING OXIDATIVE CHLORIDE LEACH PROCESSES SULFIDE/BISULFIDE/SULFITE LEACHING AMMONIA LEACHING BACTERIA AND NATURAL ACID LEACHING THIOCYANATE LEACHING RECOVERY PROCESSES ECONOMIC EVALUATION ENVIRONMENTAL CONCERNS CONCLUSIONS ACKNOWLEDGEMENTS REFERENCES
ALTA Free Library www.altamet.com.au
2 4 9 11 14 16 18 19 20 21 25 28 28 29 29
ABSTRACT Over 25 alternative lixiviant processes to cyanide have been tested in the laboratory; some of which have been successful for niche applications. The process conditions, applications and current status of the most attractive are reviewed. Most work has focussed on thiosulfate, thiourea and halide processes. The key to making these processes attainable is in decreasing the quantity of reagents used, reducing reagent consumption and improving gold recovery. Pressure oxidative chloride, sulfide and ammonia leaching processes are generally more applicable for the extraction of gold or platinum group metals as a by-product from base metal sulfide concentrates. The most advanced is the thiosulfate leaching of carbonaceous preg-robbing ores which has been largely developed by Newmont Mining Co, Placer Dome and Barrick Gold Corporation. Some chemicals used such as ammonia also pose health, safety and environmental concerns. Consequently proper disposal of wastes and sustainable development issues will have to be addressed by mining companies.
1. INTRODUCTION The major impetus in seeking alternative lixiviants to cyanide arises from the environmental hazards posed by the toxicity of cyanide, with numerous environmental groups throughout the world actively pursuing a ban on its use. Approval for any new gold project using cyanide is extremely unlikely in some areas around the world. Elsewhere, increasing regulatory scrutiny of new gold projects and a lowering of acceptable levels of cyanide discharge are of considerable concern to mining companies. Over the past two decades a significant amount of literature has examined alternative extraction processes to the use of cyanide for recovering gold from different ores. The chemistry of these alternative processes has been reviewed by Avraamides (1982), Nicol et al. (1987), Hiskey and Atluri (1988), Sparrow and Woodcock (1995), and more recently by Aylmore (2005). The 27 possible solvents can be grouped into 11 categories (Table 1). Table 1: Alternative lixiviants to cyanide 1/ Thiosulfate (Cu(II)-NH3-S2O3) 2/ Thiourea (Fe(III), CS(NH2)2) 3/ Halide (Cl2, Br2, I2) 4/ Oxidative chloride processes - Aqua regia - Acid ferric chloride - Haber process 5/ Sulfide systems - Sodium sulfide - Polysulfide - Biocatalysed bisulfate - Bisulfide/sulfur dioxide - N cat press proc 6/ Ammonia/O2 or Cu(II)
7/ Bacteria/natural acids 8/ Thiocyanate/Fe(III) 9/ Nitriles /O2 or Cu(II) 10/ Cyanide + other combination - Ammonia-cyanide - Alkali cyanoform - Calcium cyanamide - Bromo cyanide 11/ Others - Electrolysis of ore slurry - CSUT - DMSO,DMF - BioD leachant
Most work has focussed on thiosulfate, thiourea and halide leach systems. The oxidative chloride, sulfide and ammonia leaching processes have generally been used for extraction of a wide range of elements (e.g. base metal, PGMâ&#x20AC;&#x2122;s) - with gold as a by-product. Thiocyanate, nitriles, and combined cyanide lixiviant systems contain cyanide or derivatives of it, and therefore may not be considered by some to be different to the use of cyanide. Most of the other lixiviants discussed by Sparrow and Woodcock (1995) are solely of academic interest or have received limited publication. This presentation
2 ALTA Free Library www.altamet.com.au
will be concerned with the first 4 categories, concentrating on the most recent literature. However for completeness the first 8 processes are summarized in the following document.
1.1 OVERVIEW OF ALTERNATIVE PROCESS OPTIONS Despite the research interest on non-cyanide gold lixiviants, alternative gold processes are still at the developmental stages. A key factor affecting ultimate commercial success is the stability of the lixiviant and the gold complex in solution. In some cases there is a limited understanding of solution and pulp chemistry. This is partly associated with i) the difficulties in measuring reliable equilibrium data for various gold(I/III) complexes ii) the lack of knowledge on mixed-ligand complexes, and iii) different kinetic stabilities of gold(I) complexes with respect to disproportionation. (Senanayake, 2004). The equilibrium data for complex formation, dissolution, precipitation, hydrolysis and disproportionation reactions of gold(I/III) compounds for a range of non-cyanide ligands has recently been re-evaluated (Senanayake, 2004). The stability constants (β2 or β4) for various gold(I) and (III) complexes, together with their standard reduction potentials are shown in Table 2. Clearly the cyanide complex is more stable than any of the other reagents, with thiosulfate, thiourea and bisulfide several orders of magnitude less. o
Table 2: Stability constants and standard reduction potentials for gold complexes at 25 C Ligand CNS2O32CS(NH2)2 Cl Br-
I
HSNH3 Glycinate SCNSO3
2-
Au(I) or Au(III) Complex Au(CN)2Au(S2O3)23+ Au(NH2CSNH2)2 AuCl2 AuCl4AuBr2AuBr4 AuI2 AuI4Au(HS)2+ Au(NH3)2 Au(NH2CH2COO)2 Au(SCN)2 Au(SCN)43Au (SO3)2
Log β2 or β4 38.3 28.7 23.3 9.1 25.3 12.0 32.8 18.6 47.7 29.9 13 18 17.1 43.9 15.4
E0 Au(I or III) /Au (V vs SHE) -0.57 0.17 0.38 1.11 1.00 0.98 0.97 0.58 0.69 -0.25 0.57 0.632 0.66 0.66 0.77
pH range >9 8-10 <3 <3 5-8 5-9 <9 >9 9 <3 >4
As a result of the wide range of values for the stability constants of the gold complexes, the standard reduction potentials for the different gold ligand species vary by almost 2 volts (Ritchie et al., 2001). For many of the ligands like thiosulfate and thiourea, oxidation of the ligand occurs at a potential below that for the corresponding Au(I) complex, while the reverse is true for ligands SCN and Cl . Therefore there is a competing reaction to gold dissolution with most alternative lixiviants which increases reagent consumption. The presence of Fe(III) catalyst in acid thiourea solutions and Cu(II) catalyst in alkaline thiosulfate solutions, also results in rapid oxidation of the ligand. Oxygen itself is often a poor oxidant due to low rates of mass transport and slow rates of reduction on gold surfaces in non-cyanide systems. With the exception of the halides, the alternative lixiviants are clearly more complex to operate than cyanide. Most reagents have a small operating window where the alternative lixiviants effectively dissolve gold compared with cyanide (Figure1). The high oxidising potentials involved with some lixiviants inevitably lead to high reagent consumptions due to reaction with any sulfide minerals as well as oxidation of the reagent itself (Nicol, 1980). This applies particularly to thiocyanate and thiosulfate. Consequently leaching conditions have to be better controlled than those used for cyanide leaching. Equally important, although not always considered, is the adsorption of reagents and/or precipitation of gold onto some clay and gangue minerals which will be detrimental to overall gold recovery.
3 ALTA Free Library www.altamet.com.au
Oxidising
1.5
Hypochlorite Chloride
1.0
Bromide
Eh (Volts)
Thiocyanate
Ammonia
Iodide
0.5 Thiourea
bisulfide
0.0 Ammonium thiosulfate
Sodium cyanide
-0.5
-1.0 Reducing
0
2
4
6
Highly acidic
pH
8
10
12
14
Highly alkaline
Figure 1: Eh-pH diagram showing typical operating regions for gold lixiviants
2. THIOSULFATE LEACHING A review of the literature (Aylmore and Muir 2001a) shows that thiosulfate had been used on several occasions to leach gold and silver ores since 1900, but serious interest did not take place until around 1980 with an emphasis on copper-gold and carbonaceous ores that give poor gold recoveries using cyanide. The earlier studies, using relatively high concentrations of reagents, gave reasonably high gold extraction but consumed up to 40 kg/t thiosulfate. It was recognised that it was necessary to limit the concentration of copper in solution because Cu(II) degrades thiosulfate leading to high reagent losses. Later studies by Langhans et al. (1992) established that comparable extraction of gold could be achieved with dilute ammoniacal thiosulfate and identified the catalytic role played by trace amounts of copper. In recent years, thiosulfate has been considered the most attractive alternative to cyanide for leaching gold, with many investigations taking place around the world. This is primarily based on its low toxicity and its potential use on ‘preg-robbing’ carbonaceous ores that cannot be readily treated by conventional cyanidation. The ammoniacal thiosulfate leaching process for gold and silver extraction has recently been reviewed in terms of the leaching mechanism, thermodynamics, thiosulfate stability and gold recovery options (Aylmore and Muir, 2001b; Muir and Aylmore 2004; Molleman and Dreisinger, 2002; Grosse et al., 2003). Recent advances in understanding the thiosulfate process are presented by Muir and Aylmore (2005). 2.1 PROCESS CONDITIONS The chemistry of the ammonia-thiosulfate - copper system is complicated. However, by maintaining suitable Eh and pH and by controlling the concentrations of thiosulfate, ammonia, copper and oxygen in the leach solution, high gold extraction can be achieved with low reagent consumption for some ores (Wan, 1997, West-Sells and Hackl., 2005). +
The oxidation of metallic gold to the aurous Au ion in 0.10 M ammoniacal thiosulfate in the presence of copper(II) occurs at a potential of about 0.0V (pH between 8 and 10) and can be simply represented by the following reaction: Au + 5S2O32- + Cu(NH3)42+ → Au(S2O3) 23- + 4NH3 + Cu(S2O3)35-
4 ALTA Free Library www.altamet.com.au
However studies by Senanayake (2004) suggest that the mechanism involves the formation of a mixed copper(II)-ammonia-thiosulfate complex and its absorption onto the gold surface with simultaneous oxidation of gold and thiosulfate. The redox equilibrium between the cuprous-cupric couple in ammoniacal thiosulfate solution, with oxygen acting to re-oxidise copper(I) to copper(II), is as follows (Wan, 1997). 5-
2+
2Cu(S2O3)3 + 8NH3 + ½O2 + H2O = 2Cu(NH3)4
-
2-
+ 2OH + 6S2O3
To leach gold the copper(II) catalyst for this reaction is normally present at concentrations around 10-3 – -4 10 M (60 –6 ppm) (Lam and Dreisinger, 2003). Unfortunately, both oxygen and copper(II) also oxidize thiosulfate to trithionate and tetrathionate, resulting in high reagent consumption and the production of other sulfur species that impair gold leaching (Chu et al., 2003). Ammonia is not directly involved in the mechanism of gold leaching except as a complexant for copper(II) (Senanayake, et al. 2003). Ammonia also inhibits the formation or absorption of various gold passivating sulfur-containing species on the gold surface as a result of the oxidation of thiosulfate and decomposition of polythionates (Bagdasaryan et al., 1983; Chen et al. 1996; MacDonald 1990; Wan and LeVier, 2003; Breuer and Jeffrey, 2003a; Zhang and Nicol, 2003;). Surface enhanced Raman scattering spectroscopy (SERS) studies on the anodic dissolution of gold in thiosulfate solutions has recently identified sulfur, polythionate and copper sulfide species on gold surfaces that passivate the gold surface and prevent leaching (Jeffrey et al., 2008). The leaching rate of gold with 0.1 M thiosulfate containing 10 mM Cu(II) and 0.4 M NH3 has been shown to be similar to the rate with air saturated 5 mM cyanide (250 ppm NaCN), and significantly slower than with 5 mM hypochlorite solution (Jeffrey et al., 2001; Breuer et al. 2001; Breuer and Jeffrey, 2002; Jeffrey (2001). The rate of gold leaching is dependent upon the presence of copper(II), thiosulfate and ammonia and is faster with higher reagent concentrations. Unfortunately reagent degradation is also much faster, and leads to a decrease in the leach rate due to passivation of the gold surface (Breuer and Jeffrey 2002). Low gold recoveries from ores have been attributed to the coarse nature of the gold, the precipitation of gold on iron from the grinding media (Ji et al. 2003; Feng and Deventer, 2010) and the preg-robbing of gold onto sulfide and clay minerals (Feng and van Deventer 2001, Aylmore and Rae 2003; ). Sulfide minerals such as arsenopyrite, pyrrhotite and pyrite are also leached in ammoniacal thiosulfate (Feng and van Deventer 2002) and hence consume oxygen/copper(II) and lower Eh. In some leach tests over an extended time there has been evidence of re-precipitation of gold, silver and copper and lower gold recovery. Briones and Lapidus (1998) found precipitation of Ag2S occurred at low thiosulfate concentrations. Once Ag2S is formed, it appears that dilute thiosulfate solutions can only partially releach the sulfide when the Eh is increased. In addition the decomposition of thiosulfate and passivation of the gold surface also leads to unacceptably long leach times and the poor recovery of coarse and occluded gold. The presence of lead species in solution has been demonstrated to have a negative impact on gold dissolution, with the formation of lead hydroxide condensing on the gold surface which is believed to passivate the gold surface (Xia and Yen, 2008). Anions such as chloride and phosphate significantly lower the rate of copper(II) reduction by thiosulfate; whilst polythionates such as tetrathionate greatly increase the rate of reduction (Breuer and Jeffrey 2003c). 2.2 DECOMPOSITION OF THIOSULFATE AND POLYTHIONATES One of the major problems in thiosulfate leaching of gold ores is the high consumption of thiosulfate reagent during gold leaching and the generation of polythionates which readily absorb on anion exchange resins and impair the recovery of gold from solution. Ji et al. (2003) demonstrated that to achieve gold loadings on resin similar to those obtained on carbon from the cyanide system, the level of polythionates has to be reduced to below 0.1 g/L. In typical leach solutions containing about 1 g/L polythionates the ratio of gold on resin to gold in solution was less than 1000 (Ji et al. 2003; Nicol and O’Malley 2001, 2002). The form and quantity of these degradation products are dependent on reagent concentrations, dissolved oxygen concentrations, pH, Eh and temperature. High consumption of thiosulfate is mainly caused by its oxidation in solution in the presence of copper acting as a catalyst, although losses by absorption onto ore minerals or bacterial degradation occur. Some of the many reversible reactions in which thiosulfate is either consumed or regenerated are
5 ALTA Free Library www.altamet.com.au
discussed by Aylmore and Muir (2001a) and by Grosse et al. (2003). The stability region of S2O32- is restricted to a narrow range of Eh in neutral to basic pH solutions and it oxidises to tetrathionate between pH 4 â&#x20AC;&#x201C;10 when catalysed by Cu(II). 8 S2O32- + 4 Cu(NH3)42+
2 Cu(S2O3)35- + S4O62- + 8 NH3
Tetrathionate itself is unstable at pH > 10. However in the presence of oxygen, direct oxidation of thiosulfate to sulfate and trithionate also occurs. 4 S4062- +
6 OH- 5 S2O32- + 2 S3O62- + 3 H2O
Fundamental studies by Byerley, et al. (1973a, 1973b, 1975) on the kinetics and mechanism of oxidation of thiosulfate ion in aqueous ammonia solution, established that the rate of oxidation and uptake of oxygen was dependent upon the concentration of Cu(II) and thiosulfate, and was inversely proportional to ammonia concentration. Other studies on the kinetics of decomposition of tetrathionate and trithionate in alkaline solution (Naito et al., 1970; Rolla and Chakrabarti 1982; Zhang and Dreisinger, 2002b) showed that whilst tetrathionate can be readily decomposed to thiosulfate and trithionate by raising the pH to 11, trithionate was more stable and required unacceptably high pH or temperature to effect decomposition. This has practical significance because even trithionate alone impairs the recovery of gold using resins. The addition of ammonium sulfide under reduced conditions has been used successfully to convert partially oxidized thiosulfate species, trithionate and tetrathionate back to thiosulfate (West-Sells and Hackl, 2005). Clearly the mechanism and decomposition pathways of thiosulfate and tetrathionate species are complex and depend upon Eh and pH. There is evidence that some copper and silver sulfide can precipitate from solution, and such precipitates could also contain some gold; but the exact mechanism for gold losses from solution is not simple. Thus much closer control is required in the copper catalysed ammoniacal thiosulfate system to prevent undesirable side reactions. 2.3 ADDITIVES A detailed study on the effects of various additives on the rate of electrochemical oxidation of gold has been carried out by Chandra and Jeffrey (2004). Several organic sulfur- or nitrogen- containing species such as xanthate, pyridine, dithiocarbamate and imidazole completely passivated gold oxidation in thiosulfate solutions, whilst thiourea and thioacetamide enhanced oxidation. Similarly ammonium and alkyl ammonium cations increased the rate of gold oxidation compared to sodium and other alkali cations. The presence of sulfite ions has been claimed to prevent the formation of any free sulfide ion and avoid the precipitation of gold or silver from solution. In early research, Kerley (1983) reported that maintaining a level of 0.05% sulfite ion stabilizes the thiosulfate in solution. Recent studies by Fleming et al. (2003) and Ji et al. (2003) found that sulfite addition after leaching, reacted with the tetrathionate that had built up to form trithionate, sulfate and thiosulfate. However excessive sulfite addition inhibits gold leaching by consuming oxygen and reducing Cu(II). The addition of phosphates and carbonates has been shown to suppress the catalytic effect of pyrite on oxidation of thiosulfate (Zhang and Jeffrey, 2008) and improve gold extraction in the presence of lead minerals (Xia and Yen, 2008). 2.4 ALTERNATIVE THIOSULFATE SYSTEMS In order to overcome the environmental and economic problems associated with ammonia and the chemical problems of copper(II) degradation of thiosulfate, Ji et al. (2003) examined a novel ammonia and copper-free, oxygen-thiosulfate system. It was found that in <6 hours >80% Au was readily leached from the preg-robbing carbonaceous ore tested at elevated temperatures (60-80oC) and overpressures of oxygen (10-100 psig). Under these conditions there was relatively little degradation of thiosulfate and trithionate was the predominate product which could be significantly reduced by the addition of sulfite or o by anaerobic decomposition at 90 C. Unfortunately some gold re-precipitated during such treatments whilst other work showed that gold passivation was more severe in the absence of ammonia (Zhang and Nicol 2003). Thus more work is required on this potentially attractive system to establish its applicability to other ores. Arima, et al., (2004) examined nickel as an alternative catalyst to copper in ammoniacal thiosulfate solution. It was found that up to 95% Au could be leached from a silicate gold ore using 10-4 â&#x20AC;&#x201C; 5.10-3 M Ni(II) as catalyst. Thiosulfate consumption was significantly lower (1-5 kg/t) compared to Cu(II) as
6 ALTA Free Library www.altamet.com.au
catalyst, but the ammonia concentration was high (0.50 M, pH 9.5). It was proposed that Ni3O4 is produced on the gold surface and acts as the oxidant under the Eh and pH conditions of the leach. Other fundamental studies on alternative copper(II) ligands found that amino acids, polyamines and pyridine compounds inhibited thiosulfate degradation compared to ammonia (Brown et al. 2003; Michel and Frenay 1999), but their application to ores is clearly limited. More recently a new thiosulfate process has been reported which involves the use of a small concentration of thiourea as a catalyst for the oxidation of gold and the complex of ferric-EDTA as an effective oxidant (Zhang et al 2005; Zhang 2005). The reduction of FeEDTA is reversible and regenerated by dissolved oxygen. Another oxidant tested is the ferric oxalate complex although the reduction to ferrous oxalate is less reversible (Chandra and Jeffrey, 2005). However sulfide minerals present in sulfide ores catalyse the oxidation of thiosulfate resulting in limited gold extraction (Heath et al 2008). Hence while degradation of thiosulfate can be controlled to a degree by changing the oxidate the influence of minerals components in the ore still need to be overcome. 2.5 APPLICATIONS Optimum conditions for leaching appear to vary depending upon the mineralogy of the ore treated and the deportment of gold. Preferred conditions reported by Muir and Aylmore (2004) for treating oxidised ore are 0.050M thiosulfate (6.6 g/L), 0.40M total ammonia (6.8 g/L), and 60 mg/L Cu(II) at pH 9.5 and o 40 C. The Eh is maintained around 0.3 V by addition of air to give DO levels between 1-3 mg/L. Under the preferred leach conditions, thiosulfate consumption was found to range from 2 to 5 kg/t ore, representing ~30% loss over an 8 hour period. Ammonia losses were approximately 1 kg/t at 25oC but it appears that most was absorbed on the ore itself. While gold and silver extraction yielded up to 90% in 2-4 hours, a fraction of the cyanide-soluble gold (5 to 10%) always remained in the thiosulfate leach residue. Although changes in reagent composition and leach conditions affected the kinetics and percentage gold recovery, no conditions were found that could match gold recovery by cyanide. For heap leaching carbonaceous ores, Newmont (Wan, LeVier and Clayton 1995; Wan and LeVier 2003) showed that it was possible to heap leach low grade preg-robbing carbonaceous ores using dilute ammoniacal thiosulfate solutions, and a leach solution containing 0.1-0.2 M (NH4)2S2O3, at least 0.1 M NH3 and up to 60 mg/L of Cu(II) has been found suitable (Wan, et al., 1994). Thiosulfate consumption was generally around 3 to 5 kg/t ore. Similar reagent concentrations were used by Lakefield Research at 40-60oC and pH 8.0 to leach Barrick carbonaceous double refractory ores after pressure oxidation pre-treatment (Thomas et al.1998; Fleming et al 2003). Up to 95% Au was extracted from the finely divided gold left in the oxidised residue. In this process ore which is pressure oxidised, leaves the autoclave at about 35% solids and is directed to a leaching operation where it is contacted with ammonium thiosulfate (5g/L) and copper sulfate (25ppm Cu). The slurry of gold-bearing leachate and solid residue leaving the leaching circuit contains in the range of 1-5ppm gold and is directed to a Resin-in-pulp (RIP) circuit where gold and copper are loaded onto a strong base resin to about 1-5kg/t Au and about 10-25 kg/t Cu. Copper is eluted from the resin using ammonia thiosulfate (200g/L) and gold is eluted using potassium thiocyanate (200g/L). The copper-bearing eluate is returned to the leaching circuit while the gold eluate is either electro-won or precipitated. Several research groups have studied copper-gold ores but generally find it difficult to control the extent of copper leaching and the degradation of thiosulfate (Molleman and Dreisinger, 2002). Typically between 20-50% Cu is extracted together with 50-85% Au whilst reagent consumption varies between 10-40 kg/t. Control of aeration is regarded as critical to minimise reagent consumption. Studies by Frietas, et al., (2001) on three Brazilian copper-gold ores found that each ore responded differently, and each required different leach conditions for optimum recoveries of between 70-90% Au. In some cases, copper partially precipitated from solution and gold extraction reached a maximum after 8 hours. Studies by Xia, Yen and Deschenes (2003) on a mildly refractory copper-gold ore achieved gold extractions around 90% but with a reagent consumption of around 30 kg/t. Lower reagent consumption (<10 kg/t) was achieved by increasing the pulp density, lowering the oxygen supply, adding sulfite ion, by complexing Cu(II) with ethylene diammine tetra-acetate (EDTA) or ammonium triacetate; or by replacing copper sulfate with nickel sulfate. The most significant factors were pulp density and oxygen flow. Placer Dome developed and successfully piloted a copper ammoniacal thiosulfate process for treating carbonaceous ore that mimicked a Merrill â&#x20AC;&#x201C; Crowe process (West Sells and Hackl, 2005). The process involved the regeneration of the principle leaching reagent to allow for low ammonium thiosulfate consumption. The process involved tank leaching, a CCD circuit to separate the pregnant gold solution
7 ALTA Free Library www.altamet.com.au
and a precipitation process to recover the gold. For a carbonaceous ore containing 7.5 Au g/tore, and organic carbon and total sulfur contents of 1.6% and 0.35% respectively, gold extractions of 65% were obtained at low operating costs. Ammonium thiosulfate was maintained at 15 g/L with copper added to the aerated leaching circuit at a concentration of 50 mg/L in solution, with pH values between 8.7 and 9.5. After leaching, separation of the pregnant leach solution from the ore occurred by CCD and the gold was recovered from the pregnant leach solution by precipitation. The precipitation involved the addition of ammonium sulfide to reduce the potential of the solution allowing gold to precipitate to the metallic state. Additionally during precipitation, partially oxidized thiosulfate species, trithionate and tetrathionate, were reduced back to thiosulfate. Ammonium sulfite was added to the filtered precipitated product to react with elemental sulfur present in the precipitate to produce ammonium thiosulfate which was recycled back to the process. The resultant gold sludge can be smelted. Several patents have employed either the use of sulfide directly from the ore or the addition of sulfur and sulphite to produce thiosulfate in situ to allow direct leaching of gold (e.g., Kerley 1981,1983; Aylmore, 2004). Barrack Gold Corporation in recent times has patented a process where precious metals are recovered by thiosulfate from double refractory preg-robbing ore. In this process thiosulfate lixiviant is generated in situ by reacting a sulfite source (ammonium sulfite or generated from SO2 gas) with elemental sulfur generated from partial oxidation of sulfuric ores during low temperature pressure leaching (Choi et al., 2007). Utilising a resin in leach process gold recoveries observed were between 81 and 85%. A mini pilot plant testing an integrated resin in pulp process using a novel sulfite enhanced chloride solution elution of gold thiosulfate process for the recovery of gold from resins, together with electrowinning of gold, and recycle of eluent and resin shows promise (Jeffrey et al., 2008). The tests were carried out on a pressure oxidised residue, under undisclosed conditions with an additive which appear to favour gold recovery by thiosulfate than cyanide. Gold recoveries of >95% were achieved. The favoured thiosulfate leaching conditions were 5 mM copper sulfate, 50 mM ammonium thiosulfate at a pH of 8.5 which allowed reduction in the generation of polythionates and there subsequent loading on resins (0.1 kg/t). Gold and copper loadings onto a strong base resin were about 2.5kg/t Au and about 2 kg/t Cu. Copper was eluted from the resin using 0.5M ammonia thiosulfate and gold is eluted using sodium chloride and sulfite mixture. The copper-bearing eluate is returned to the leaching circuit while the gold eluate was electro-won. 2.6 CURRENT STATUS In recent years considerable research has gone into understanding and developing the thiosulfate leaching process work. Thiosulfate remains as an attractive alternative to cyanide for processing carbonaceous ores where the recovery of gold is poor. It may also be attractive for processing some refractory copper-gold concentrates and ores that consume significant cyanide. However for simple oxide ores further research is required on alternative oxidants and additives to overcome passivation to improve gold recovery and lower reagent costs to match that of cyanide. Improvements in recovery processes following leaching have reduced the impact of recycling degradation products and made thiosulfate a more attractive option as a replacement for cyanide. Continued work by the industry is required to continue developing robust overall process flow sheets incorporating gold recovery, reagent recycling or destruction and impurity control.
8 ALTA Free Library www.altamet.com.au
3. THIOUREA LEACHING
Interest in the thiourea process for leaching gold occurred mainly during the 80’s and early 90’s (Groenewald, 1977, Hiskey, 1988, Lan et al., 1993). Considerable research has been carried out by CANMET and MINTEK for gold extraction in underground (Tremblay et al 1997) and ‘in-stope’ applications (e.g. Van Staden and Laxen, 1989). In addition Newmont Mining and Barrick Gold Corporation looked closely at thiourea leaching as a potential process for treating refractory ores. In recent years interest in this leaching approach has decreased due to the potential carcinogenic properties of thiourea. However there is still laboratory research being carried out in Eastern Europe (e.g., Gönen, N., 2003, Örgül and Atalay, 2000, 2002). Also a novel leaching process was patented by Dublin University in Ireland (Kavanagh et al., 2000). Some recent work on the oxidation of gold and thiourea in acidic thiourea solution has been carried out by Zhang et al (2001), Li and Miller (2002), and Lapidus et al., (2008). A number of papers have also been published by a group of researchers in China on the dissolution of gold electrodes in an alkaline thiourea system (e.g. Chai et al 1999); however thiourea is generally unstable in alkaline media and its application to ores has not been demonstrated. 3.1 PROCESS CONDITIONS In practice, thiourea leaching of gold is typically performed at thiourea concentrations of 0.13M (10g/L), ferric ion concentrations of 0.09 M (5g/L), pH values of 1 to 3 and potentials between 0.4 and 0.45 V (vs SHE). The overall reaction that describes gold dissolution in thiourea and ferric ion solutions is Au + NH2CSNH2 + Fe3+ = [Au(NH2CSNH2)2]+ + Fe2+ Various oxidants, including ferric ion, hydrogen peroxide, manganese dioxide, mono-per-oxysulfate compounds and ozone have been examined, but ferric ion is the most common (Sparrow and Woodcock, 1995). The gold leaching reaction is very sensitive to pH and the redox potential (Kenna and Moritz, 1991; Li and Miller, 2007). Furthermore, thiourea is intrinsically unstable and decomposes rapidly to substances that are unable to leach gold. Tremblay et al (1996) concluded that to limit the thiourea decomposition and optimise gold extraction it was necessary to maintain the leaching potential between 0.42 and 0.45 V (vs SHE). Thus careful control of solution potential in a commercial process is necessary. Leaching rates as high as 10 times faster than for cyanide have been reported with silver and gold reacting differently implying that the dissolution mechanisms are not the same. Gold in contact with pyrite or chalcopyrite also exhibits an enhanced gold dissolution rate (Van Deventer et al., 1990). However thiourea also forms strong complexes with some base metals such as copper and to a lesser extent lead and zinc (Deschenes et al 1994; Fang and Muhammed, 1992: Alodan and Smyrl, 1998), which can increase thiourea consumption. High potentials during leaching produce an oxidative degradation of thiourea that proceeds via formamidine disulfide (NH2(NH)CSSC(NH)NH2), which eventually decomposes to thiourea (CS(NH2)2), cyanamide (NH2CN), and elemental sulfur (S0). Acid hydrolysis also forms urea (NH2CONH2) and hydrogen sulfide (H2S) as follows: 2NH2CSNH2 + 2 Fe
3+
+
2+
= NH2(NH)CSSC(NH)NH2 + 2H + 2Fe
NH2(NH)CSSC(NH)NH2 = NH2CSNH2 + NH2CN + S0 NH2CSNH2 + H2O = NH2CONH2 + H2S Elemental sulfur and hydrogen sulfide are undesirable species to have present during gold leaching. It is believed that both these species cause a decrease in the leaching rate due to surface passivation, and hydrogen sulfide may cause re-precipitation of the gold (Sparrow and Woodcock, 1995; Munoz and Miller, 2000). For optimum leaching the oxidant must be added at such levels as to oxidise approximately 50% of the thiourea to formamidine disulfide. Gold dissolution can be accelerated effectively by the presence of formamidine disulfide or by using a higher temperature (Zhang et al., 2001). However, excess oxidant will increase thiourea consumption significantly. In practice it is necessary to use a stabiliser to convert formamidine disulfide back to thiourea, or to complex the oxidant.
9 ALTA Free Library www.altamet.com.au
Spectroelectrochemical investigations of gold leaching in thiourea media has recently shown that sorption of thiourea occurred concomitantly with co-adsorption of the electrolyte, whether chloride or sulfate, and that continued oxidation results in the formation of a Au2S-like phase (Parker and Hope, 2008). Formamidine disulfide, which has been alleged to possess catalytic properties in the Au– thiourea leaching system, was not observed at the gold electrode. A mixed thiourea and thiocyanate system has been shown to promote a higher dissolution rate than either lixiviant alone (Yang et al., 2010). The passivation of gold occurs in a thiourea only solution, but when thiocyanate is mixed with thiourea, passivation is significantly alleviated. The dissolution rate of gold in the mixed lixiviant system increases with increasing thiourea and thiocyanate concentration. Results from linear sweep voltammetry and electrochemical impedance spectroscopy indicated that gold dissolution was controlled by a combination of charge transfer and diffusion in the mixed lixiviant system. The optimum concentration for thiourea and thiocyanate is about 5 mM and 0.05M, respectively for the rate of gold dissolution. Surface enhanced Raman spectroscopy results suggest a + − possible formation of a mixed ligand complex involving the interaction of Au(Tu)2 and SCN . 3.2 STABILISERS To combat the loss of thiourea several workers have used additions of sulfur dioxide or substantial quantities of sulfite, and iron(III) complexing acids with mixed success (see Sparrow and Woodcock, 1995 and papers therein). Deng et al. (2001; 2002) reported on test work involving the thiourea leaching of a bio-oxidised primary gold sulfide ore to which 4.5 g/L Na2SO3 had been added. The consumption of thiourea decreased significantly from 12 kg/t to 3 kg/t and extraction time was shortened from 6 hour to 1 hour at a pH of 2. Thiourea substituted compounds such as N,N’-ethylenethiourea (Schulze 1985) is more stable to oxidation and minimises reagent consumption (Kenna and Moritz, 1991). Kenna (1991) patented a thiourea gold extraction method in which di- and tri-carboxylic acids, fluorides, fluosilicic acid and fluosilicate salt, EDTA and EDTA salts are used to complex the ferric ion to reduce thiourea consumption. Cysteine (HSCH2CH(NH2)COOH) has also been found to be an effective reagent for the stabilisation of thiourea (Ardiwilaga, 1999). However, cysteine prevented the formation of formamidine disulfide required to maintain a high gold leaching rate and consequently gold recovery was reduced. Washing with water and with varying concentrations of sulfuric acid prior to thiourea leaching has been found to be advantageous by several workers (e.g Kavanagh et al 1994; Trembley et al., 1997; LacosteBouchet et al. 1998; Deng et al. 2001). The value of this approach varied from ore to ore, however pretreatment with acid has been shown to not only remove easily leachable metals, but to allow the thiourea system to stabilise more quickly and to prevent unnecessary reagent consumption due to the precipitation of base metals. Activated carbon, some clay and gangue minerals adsorb thiourea and its gold complex, thus reducing the overall gold recovery (Quach, 1993; Zhang, 1995). 3.3 APPLICATIONS Applications of thiourea leaching have been demonstrated after fine grinding (Kusnierova et al., 1993), mechanochemical milling (Balaz et al. 2003), bacterial oxidation (Caldeira and Ciminelli, 1993; Wan et al., 1995; Deng et al 2001; Deng and Liao, 2002), after pressure oxidation (Yen and Wyslouzil, 1986; Bilston et al., 1990; Murthy et al, 2003) and after roasting (Bilston et al., 1990; Moussoulos et al., 1984). The perceived advantage is that the acidic pre-oxidised sulfide ore can be directly leached with thiourea without a neutralisation step that would be required for leaching with cyanide. Thiourea has also been applied to oxide ores with mixed results (Ubaldini et al., 1998, McNulty 2000). Most successful applications have been carried out on ores that have a high content of cyanicides such as antimony, or sulfide ores which have undergone bacteria oxidation or pressure leaching. Thiourea concentrations have ranged from 2 to15 g/L, acid from 1-150 g/L (or higher), pH 1-3 in laboratory and pilot scale test work. The oxidant has varied from 0-20 g/L where in some cases the iron from the ore has been used. Interestingly gold extraction has not always been as high as that obtained using cyanide. In general, reagent consumption is dependent on reagent concentration and values reported in the literature seem to be considerably higher (ranging from 4 to 47 kg thiourea/t) than those for cyanide, thus resulting in higher extraction costs. The high consumption of thiourea, acid (for pH control), peroxide and sulfur dioxide (for Eh control) make the overall cost of the process between 1.5 and 2 times the cost of cyanidation for the same material.
10 ALTA Free Library www.altamet.com.au
A 450 tonne pilot test heap leach using acidic thiourea after biooxidation was conducted at the Carlin mine by Newmont (Wan et al., 1995). The biooxidation test heap was washed with fresh water following biooxidation and drained before thiourea leaching. The thiourea solution at a concentration of 10 g/L was pumped onto the heap at a flow rate of 30-45 litres per minute with the whole operation running over a 110 day period at pH below 2.5. The redox potential was generally in the range of 0.43-0.5 V (vs SHE). No chemical control in terms of pH or Eh adjustments, was required during the thiourea leach. Outcomes on this work were that thiourea yielded poor gold extraction kinetics because of the large particle size used and the cold temperature experienced during the pilot run. A maximum of 29% gold was recovered. Gold recovery from the pregnant solution using activated carbon or cationic ion exchange resin was ineffective. Continuous recirculation of the solution caused elemental sulfur to form, which coated the carbon and resin, impeding the recovery process. 3.4 CURRENT STATUS With thiourea labelled as a potential carcinogen it is difficult to see it providing a replacement for cyanide in the near future. Extensive investigations on all aspects of leaching have been evaluated, and while there would be some potential niche applications it appears that the sensitivity of leaching conditions would not make it an obvious choice compared with other alternatives. Lower thiourea concentrations would be required to make it economic. Recovery of gold from thiourea solution also requires further development. The addition of thiourea in thiosulfate and thiocyanate system has been shown to enhance the anodic oxidation of gold as discussed in other sections, hence the application of thiourea may lie in improving other alternative lixiviant processes.
4. HALIDE LEACHING Chlorine, bromine and iodine are well known lixiviants for leaching gold (Tran et al., 2001). Chlorination was applied extensively in the late 19th century before the introduction of the cyanidation process. Bromine/bromide for leaching gold from ores was reported as early as 1846. Chlorination was used extensively for pretreating refractory and carbonaceous ores in several plants in the USA in the 1980’s (Marsden and House, 2006). Renewed interest in halides as a lixiviant for leaching gold occurred in the 1990’s after several patents based on the bromine/bromide systems were lodged (see Pesic et al, 1992 and papers there in; Tran et al.,2001). For completely oxidised materials, the chloride-based leaching processes have a clear advantage in applications where a high dissolution rate is required. The use of chlorine is a proven technology in gold refining and electroplating processes. Most notable in recent years has been the development of the Minataur process by Mintek. 4.1. PROCESS CONDITIONS Typical conditions used for leaching gold by halogens are listed in the Table 3. Table 3: Typical leaching conditions used in leaching gold with halides Reagent
Ligand
Chlorine
Cl
Cl2 or HClO
[AuCl4]
Bromine
Br-
Br2
[AuBr4]-
Iodine
I
I2
[AuI2]
-
-
Oxidant
Gold complex -
-
Typical leaching conditions 5-10 g/L Cl2 5-10 g/L NaCl 2-5g/L Br2, 0-10 g/L NaBr 1g/L I2, 9 g/L NaI
pH <3 5-8 5-9
In all halogen processes, high oxidation conditions are required. The general equation describing the reaction of gold with chlorine or bromine is as follows: -
-
2Au + 3X2 + 2X = 2[AuX4]
where X= Cl, Br
The complex [AuCl2]- is formed initially which is rapidly oxidised to [AuCl4]- (Nesbitt et al., 1990; Sun and Yen , 1992). In the iodine system, iodine reacts with iodide in aqueous solution to form I3- ions with Au(I) rather than forming the Au(III) complex (Eo for [AuI2] < [AuI4] ,Tran et al., 2001). The Au(III) complex is not stable because gold(III) oxidises iodide ion to iodine and the Au(III) complex is reduced to Au(I) complex.
11 ALTA Free Library www.altamet.com.au
2Au + I3- +
I- = 2[AuI2]-
Unlike gold cyanide, which is very stable and does not decompose easily in most applications, the stability of gold halide is dependent on the solution pH, composition (with respect to halide concentration), solution potential and the presence of reductants (such as metals and sulfidic minerals) in the ores. A residual amount of oxidant is required to maintain a high solution potential to avoid the precipitation of metallic gold from solution (Tran et al., 2001). The gold dissolution rate can be very high and is dependent on solution pH, lixiviant and oxidant concentrations (Sergent et al., 1992; Tram et al., 2001). Overall the stability of halides is in the order of I>Br>Cl, whereas the rate is Cl>Br>I. 4.2. CHLORINE The fundamental electrochemical kinetics of gold dissolution in chloride media and the chemistry of the chlorine process have been described by Finkelstein (1972), Nicol (1976) and Avraamides (1982). Whilst the dissolution of gold in chloride media, using hypochlorite as an oxidant, was evaluated by Yen et al. (1990) and Tran et al. (1992). The weight loss of gold strips immersed in different chloride2 hypochlorite mixtures (up to 20 mg/cm h) was much faster than that achieved by cyanidation under 2 similar test conditions (2.5 mg/cm h for 2g/L NaCN) (Tran et al., 2001). The stability of the Au(III) chloride complex (AuCl4-) is strongly dependent on the solution pH and requires high chloride and chlorine levels, increased temperature and high ore surface area. The complex is only stable at pH <3.0 unless a sodium chloride concentration higher than 100 g/L is maintained. As the dissolved gold complex is unstable and re-precipitates upon contact with a reductant such as sulfidic materials or metals, application of the chloride-chlorine systems is limited to extraction of gold from oxidised materials. Attempts have been made to reduce the reactivity of sulfides in halide systems using compounds such as flotation collectors used to coat sulfides (Stace, 1984). This was found effective for some metals sulfides (e.g. copper sulfides) where the coating significantly reduced their reactivity without altering the reaction rate of gold dissolution. Other sulfides such as pyrite were unsuccessful. High silver contents in ores possibly dissolve slowly in low-chloride solutions because of the formation of a passivating film of insoluble silver chloride (Sparrow and Woodcock, 1995). Consequently higher concentrations of chloride in solution are required to solubilise the relatively insoluble silver chloride. Therefore the chloride system is not ideally suited to treatment of ores in which silver is of primary value. 4.3. BROMINE Bromine leaching results in rapid gold dissolution at near neutral pH conditions. Due to the high potential required for dissolution (E0 = 0.97 V), compared with the formation of gold cyanide at (-0.57V), gold bromide is unstable and requires the addition of a strong oxidant such as bromine. Extensive research on the bromide-bromine system was carried out in the 1990â&#x20AC;&#x2122;s. Electrochemical techniques were applied by Pesic and Sergent (1992) and by van Meersbergen et al. (1993) to determine the complex reaction mechanism for gold dissolution in bromine-bromide systems. The dissolution of gold was shown to depend on the bromine-bromide ratio and the associated minerals in the ore. The presence of copper, zinc, and aluminium as sulfates have no effect on dissolution, but iron(II) and manganese(II) are oxidised and consume bromine. Alternative oxidants to bromine to eliminate the problems associated with high vapour pressure and corrosive reactions of bromine have been examined with limited success. These are ferric ion, hydrogen peroxide and sodium hypochlorite (Trindade et al., 1994; Sparrow and Woodcock, 1995). The use of commercially available organic bromides (e.g. N-halo hydantoins such as Geobrom 3400) also reduces problems with vapour loss. 4.4. IODINE Of all the halogens, the gold iodide complex is the most stable in aqueous solution, presumably because of its lower redox potential value compared with the other halogens. Hiskey and his workers have extensively evaluated the leaching behaviour and chemistry of gold and silver dissolution in various iodide-iodine mixtures (Hiskey and Atluri 1988; Hiskey and Qi, 1991; Qi and Hiskey, 1991, 1993; Angelidis et al., 1993). The gold dissolution rate was directly proportional to iodine and iodide concentrations and was not greatly affected by change in pH over the range pH 2-10. Several early patents were granted on the application of this system for deep lead â&#x20AC;&#x153;in-situâ&#x20AC;? leaching (McGrew and Murphy 1985) or for processing gold from electronic scraps (Falanga and MacDonald, 1982). The iodide-iodine system does not normally oxidise metal sulfides such as chalcocite and pyrrhotite, therefore avoiding excessive reagent consumption in a potential gold processing plant (Hiskey and Qi,
12 ALTA Free Library www.altamet.com.au
1991). Consequently it might be appropriate for treating ores containing sulfidic minerals. However recovery and regeneration of the active species is important, as iodide and iodine are very expensive. Alternative oxidants have been suggested in the iodine system such as hypochlorite which permits a lower iodine concentration in solution, and so minimise losses of iodine by evaporation. (Davis and Tran, 1991; Davis et al., 1992). The conditions for extraction including oxidant/iodide molar ratio, concentration and pH, have to be optimised to avoid the passivation of gold by gold iodide (AuI) and to maximise the gold extraction rate. In addition, hypochlorite concentration has to be properly optimised as an overdose of this oxidant can destroy the iodide lixiviant used for complexing the gold. A maximum gold dissolution rate was achieved at a [OCl-]/[I-] molar ratio of 0.25 (Davis and Tran, 1991). In this system the active species dissolving gold is I3 which is formed from the reaction between hypochlorite and iodide. The same applies for the hypochlorite-bromide system in which the active species dissolving gold is Br3- (Tran et al., 2001). 4.5. APPLICATIONS Pretreatment of sulfidic or carbonaceous ores by roasting or pressure oxidation is normally required prior to chlorine or bromine leaching to render ores relatively inert, and consequently reduce reagent consumption (Li et al., 1992; Sparrow and Woodcock, 1995). For example, the selective recovery of gold and silver from a chalcopyrite concentrate was carried out by Puvvada and Murthy (2000). Gold and silver grades in the concentrate were 11 g/t and 140 g/t, respectively. Laboratory scale tests were conducted at room temperature on 20% solids slurry containing 25 g/L NaOCl and 0.35 M HCl. Increasing the NaCl concentration increased the rate as well as the extent of gold and silver extraction. Gold and silver recoveries of 42.7% and 45.0% respectively were obtained with 200 g/L NaCl. Dissolution of silver was found to be independent of NaOCl concentration. However, pressure oxidising the copper concentrate and then leaching with NaOCl concentration of 25 g/L, 200 g/L NaCl and 0.35 M HCl for 1 hour, resulted in gold and silver recoveries of 90.0% and 92.5%, respectively. A pre-feasibility test on hypochlorite leaching McDonald Gold Mine oxide ore in Montana gave 68% gold recovery compared with the 73% obtained with cyanide (McNulty 2000). Under the conditions used in their bottle-roll test work (pH 6.5, Eh 1.14 V, 33% solids for 96 hours leaching time) reagent consumption was 5.55 kg/t NaOCl and 3.25 kg/t HCl compared with 0.15 kg/t NaCN and 0.55 kg/t CaO for the cyanide process. Interest in the use of chloride-based leaching for treating gold from copper anode slime has continued with several studies on its kinetics (Donmez et al., 1999; Herreros et al., 1999). The technique was tested or practised in several copper refineries in the early 1990â&#x20AC;&#x2122;s (see Tran et al., 2001 and reference there in). More recent work by Nam et al (2008) has examined conditions to leach a mine tailing containing gold and silver. Using chlorinated sea water (approximately 0.5 M NaCl and pH5.5 and Eh of less than 1.00 V vs Agâ&#x20AC;&#x201C;AgCl reference), both gold and silver could not be fully extracted due to the formation of gold hydroxide and silver chloride. Nevertheless, the gold and silver extractions reached 80% and 50% in reactor leaching. There are several chloride leaching process variations for treating high gold grade slags which have been reported on various websites such as the Clakdale process and Hyperleach process, but limited details are available on the efficiency of such processes for general gold leaching. 4.6. CURRENT STATUS Halide leaching provides greater flexibility than cyanidation as reagent dosages can be controlled to enhance the dissolution rate and is particularly useful for processing coarse gold in oxidised materials. However, gold halides are unstable and critics remain skeptical as a great deal of chemical and process control is required during processing to maintain gold in solution. Furthermore losses of bromide or iodide absorbed on gangue minerals or precipitated as insoluble copper, silver or lead salts can lead to high reagent consumption on low grade ores. The technique is suitable for extracting gold from goldrich materials such as anode slimes and oxidised gold gravity concentrates. Chlorination processes are being utilised as a gold and silver refining technique to replace smelting. The lack of a suitable recovery process to match the cyanide CIP plant technology has been the main limitation of the use of chloride systems.
13 ALTA Free Library www.altamet.com.au
4.7. OTHER HALIDE PROCESSES 4.7.1. The Minataur process The Minataur process developed by MINTEK in South Africa comprises oxidative leaching of the feed material, followed by selective solvent extraction of the gold from the leach liquor to reject impurities, and precipitation of high-purity gold (Feather, et al., 1997). Suitable feeds include silver-refining anode slimes, gold electrowinning cathode sludge, zinc precipitation filtrates, gold gravity concentrates and the residues from mill liners in gold plants. Impure gold feed material is leached for 2-3 hours in 5M HCl under oxidising conditions with chlorine continuously added into the leach reactor. The leach solution is then purified by solvent extraction. Gold is selectively extracted into the organic phase (not specified) over silver as well as the platinum group and base metals which report to the raffinate. Gold is recovered as a metal powder by direct reduction in sulfur dioxide or oxalic acid from the loaded strip liquor. In addition the Gravitaur process has been developed which incorporates gravity concentrates as the feed, without the need for an intermediate solvent-extraction step to upgrade the gold tenor of the solution. The Chinese (Li et al., 1996) have also evaluated a chloridising process for gold extraction from silver anode sludge, where gold is recovered by reduction of liquor with oxalic acid, possibly similar to the Minataur process. 4.7.2. Intec and N-Chlo Copper and Gold processes The Intec Copper process is primarily used for processing copper and other base metal sulfides; however the application to the simultaneous extraction and recovery of gold/silver has also been evaluated (Moyles et al., 1999; Moyes et al 2005). In this process, copper sulfide feeds are typically leached at 85oC using a chloride electrolyte (280 g/L NaCl and 28 g/L NaBr) containing a chlorinebromine complex (BrCl2-, named Halex) produced from a copper electrowinning cell downstream. Gold and silver are leached by this leachant which achieves a solution potential (Eh) of around 1.2 V (vs SHE). Silver is recovered by cementation, whereas gold is recovered by precipitation as the solution potential is dropped to less than 0.80 V (vs SHE) after leaching (Severs, 1999). High gold extractions have been achieved from a range of refractory gold concentrates at laboratory scale, which has led to a continuous locked-cycle pilot plant program. The pilot plant operated at >99% availability, with a maximum of 96.5% gold extraction from a concentrate containing 58.6 g/t gold. >99% of the dissolved gold was loaded onto carbon at up to 1% w/w, with no loss of carbon activity detected over five loading/washing/elution cycles. The N-Chlo process utilises a similar approach to the Intec process where fine ground copper and gold sulfide concentrate are leached in a chloride bromine medium (Abe and Hosaka, 2007). Gold is recovered on carbon whereas solvent extraction is used to recover silver. A demonstration plant is currently in operation in Perth. The Intec gold process, an extension of the halide based Intec Copper Process, has also been developed. Sulfide concentrate is ground to a P80 of 10-40 and fed to a leach train where the refractory sulphides are oxidised along with the liberated gold by oxygen from direct air injection. Operating conditions involve temperatures between 85 to 95°C with a 6 to 8 M chloride/bromide electrolyte 2+ containing 20 to 60 g/L Cu , leaching for up to 10 hours. Copper is used as a catalyst to assist the transfer of oxygen from air. Soluble iron is controlled by the addition of limestone into the last leach reactor. The slurry is then sent to the solid/liquid separation stage, where the leach residue is washed and discharged. Gold in the clear liquor is adsorbed onto activated carbon or resin packed fixed-bed columns and subsequently eluted using thiourea, thiosulphate, or cyanide. Gold is subsequently recovered as a metal by electrowinning or cementation. Gold-depleted liquor is sent to the purification circuit where by-products are precipitated with slaked lime. The precipitated solids are separated by filtration where they are washed, and the filtrate is recycled to leaching operations. Ferrous and cuprous reaction products are subsequently oxidised by further air sparging. The process for treating refractory concentrate has been claimed to be more economical than biooxidation or pressure oxidation processes. 5. OXIDATIVE CHLORIDE LEACH PROCESSES Besides chlorine, several other oxidants such as oxygen or nitric acid dissolve gold in the presence of chloride. Oxidative chloride leaching may have possible applications in treating silicate ores if only small
14 ALTA Free Library www.altamet.com.au
amounts of high-gold sulfide minerals are present, or it can be used as a total dissolution process for treating sulfide ores to recover base and precious metals. Most oxidative chloride leach processes, including ferric chloride leaching, have been mainly used as a pre-treatment process rather than a gold leaching process. However the Platsol Process has been developed to leach and recover base metals, gold and PGM’s from sulfide ore. 5.1. PROCESS CONDITIONS Some conditions used for leaching gold by oxidative chloride systems are listed in the Table 4. Table 4 Typical oxidative chloride conditions used in leaching gold Process
Ligand
Oxidant
Aqua regia Ferric chloride Platsol
ClCl Cl
HNO3 3+ Fe O2 (689 kPa)
Gold complex in Typical leaching solution conditions [AuCl4]3:1 HCl:HNO3 [AuCl4] 3-6% FeCl3 [AuCl4] 5-10g/L NaCl, T>220°C
pH pH<0 pH <2 ~2
Except for special processing of high grade materials, such as the refining of precious metal concentrates, gold bullion or platinum metals, aqua regia is not considered practical for use at the ore treatment plant-scale level. Suitable oxidants include oxygen under pressure, ferric salts, hydrogen peroxide, sodium hypochlorite and persulfates (Sparrow and Woodcock, 1995). Acid ferric chloride solutions have been used primarily as a process to oxidise sulfide concentrates and complex lead-zinc sulfide ores and concentrates prior to cyanidation (e.g. Jain and Hendrix,1996). However recent electrochemical investigations by Liu and Nicol (2002) have demonstrated the effectiveness of iron(III) as the oxidant for leaching of gold at high temperatures under typical pressure oxidation conditions used for treating refractory gold ores. Measurements of the equilibrium potentials of the gold(III)/gold and iron(III)/iron(II) couples over a temperature range from 25°C to 200°C in acidic sulfate solutions containing various concentrations of chloride ions showed that the equilibrium solubility of gold increases with increasing temperature, chloride concentration and iron(III)/iron(II) ratio. 5.2. PLATSOL PROCESS The Platsol process was developed in collaboration with the University of British Columbia, Kane Consultants Ltd and Lakefield Research in Canada for the treatment of flotation sulfide concentrate for Polymet Mining Company in Minnesota. This process involved dissolution in one step of the base metals (copper and nickel) as well as the gold and PGM’s. This was followed by solid/liquid separation, gold and PGM’s recovery and conventional Cu SX/EW and recycling of the copper raffinate to the autoclave. Pilot plant work has recently been carried out. The fundamental difference between the Platsol process and the conventional high temperature pressure oxidation processes is that a small concentration of chloride ions is added to the autoclave with ~25g/L sulfuric acid. The chloride favours the oxidation of gold and platinum group metals and stabilises them as dissolved chloro complexes. Grinding the ore with ceramic rather than iron balls was required to prevent cementation of gold chloride. (Ferron et al., 2001). The concentrate tested was a flotation concentrate assaying 14.7% Cu, 3.05% Ni, 0.14% Co, 26.7% S, 1.4 g/t Au, 2.2 g/t Pt and 9.9 g/t Pd. Pressure oxidation conditions were 225°C, pulp density 11%, retention time 120 minutes and oxygen overpressure was 689 kPa. The ore treated had a P80 of 15 microns. After solid/liquor separation the gold and PGM’s were recovered by sulfide precipitation using NaSH or by activated carbon. The copper was recovered using conventional solvent extraction and electrowinning techniques. Overall recoveries were Cu 99.6%, Ni 98.9%, Co 96%, Pd 94.6%, Pt 96% and Au 89.4% (Ferron et al., 2001). Recent work in treating a variety of refractory gold concentrates under optimum conditions (225oC, NaCl 10-20 g/L, 2-6 hours O2 at 700 kPa) achieved gold extractions of the order of 90-96% compared with direct cyanidation where gold extraction was less than 20%. (Ferron et al., 2003). Examination of a variety of recovery options showed that loading of gold onto carbon from clear liquors and pulps was rapid and did not require prior neutralisation. Zadra elution of the loaded carbon recovered more than 90% of the gold, however further work in investigating carbon regeneration is required. Gold could easily be precipitated from acidic Platsol leach liquors with NaSH, but minimisation of the co-precipitation of impurities such as copper needs to be addressed. In addition some tests on
15 ALTA Free Library www.altamet.com.au
gold recovery by ion exchange resins showed promise. Currently the gold chloride is precipitated using a synthetic covellite produced in residual copper recovery process. T he Platsol process is currently in detailed engineering phase following the success of the DFS and FEED studies (Wardell-Johnson et al., 2009) 5.3. CURRENT STATUS Most of these processes are restricted to high grade concentrates and have been used as a pretreatment process for leaching gold with cyanide. Similar to the halide systems, the stability of gold in solution and the lack of a suitable recovery option appear to restrict them as a simultaneous leaching and recovery process. For sulfide ores containing a combination of base, precious and platinum group metals, the Platsol process appears to be an economical and viable process where polymetallic ore is high in abundance.
6. SULFIDE/BISULFIDE/SULFITE LEACHING The sulfur based gold lixiviants other than thiosulfate and thiourea are sulfide, bisulfide, bisulfite (or sulfur dioxide) and polysulfides. Of these possible lixiviants only the bisulfide and bisulfite ions appear to have any practical use under ambient conditions. YES Technologies described a process using bisulfide generated from sulfate reducing bacteria as a lixiviant for gold, although bisulfide ions can readily be generated from hydrogen sulfide gas. Alkaline sulfide lixiviation of sulfur residues from the Nitrogen Species Catalysed Pressure Leaching process has been demonstrated to leach and recover silver successfully. 6.1 PROCESS CONDITIONS The possible processes investigated using sulfur based lixiviants for gold are listed in Table 5. The mechanism of gold dissolution and precipitation from aqueous sulfide solutions under a range of conditions have received considerable attention in the early geological literature (e.g. Hanninton and Scott, 1989; Seward, 1973; Tan and Bell, 1990). Neutral bisulfide solutions dissolve gold as follows: -
-
Au + H2S + HS = Au(HS)2 + ½H2 Under weak acid conditions the gold complex [Au(HS)]0 forms, whilst under strongly alkaline conditions [Au2(HS)2S]2- is formed (Seward, 1973). Because of the similar stabilities of gold and silver bisulfide complexes, bisulfide leaching may be suitable for leaching gold ores with a significant silver content (Hunter et al., 1998). Table 5: Typical sulfur based conditions used in leaching gold Reagent
Ligand
Oxidant
Sodium sulfide
HS
-
H2S generated
Gold complex in solution [Au(HS)2]
YES technology 2(bacteria SO4 reducing process) Bisulfite or SO2 Polysulfide Lime sulfur synthetic solution/ Phase transfer catalysts Nitrogen Species Catalysed Pressure Leaching process
HS-
H2S generated
[Au(HS)2]-
Around 2.5g/L H2S, HS-, 2and S . pH 6-9
HSO3Sx222S2O3 , Sx
O2 S0 O2, Cu(II) and + NH4
[Au(HSO3)2]Au/Sx 3Au(S2O3)2 , Au/Sx(?)
15-50 kg/t SO2, pH 4-5 >2M polysulfides ?
S52-, S2O32
+
Au S5 , Au S2O3
0
-
NO , S
-
16 ALTA Free Library www.altamet.com.au
Typical leaching conditions 50 g/L Na2S, pH >12
20-175 g/L H2SO4, 620-975 o kPa, 125-170 C, 2.0 g/L Nitric acid
The YES process uses naturally-occurring, sulfate-reducing, bacteria for the recovery of gold and silver from ores. A full description of the process is outlined in the patent by Hunter et al. (1998). Conventional bio-oxidation of ore particles is carried out to free the precious metals dispersed or occluded within the ore. A portion of the acidic, base metal sulfate leach solution produced from biooxidation is introduced to an anaerobic reactor. A non-toxic donor such as acetate or methanol (which does not bind to activated carbon) is added to the anaerobic reactor to enrich within it a mixed culture of sulfate-reducing bacteria. Bisulfide ions are generated biologically in the process with an electron donor such as acetate by the following reaction: -
2-
-
-
CH3COO + SO4 → 2HCO3 + HS
Additions of acidic sulfate solution may be required to maintain a neutral pH in the reactor. In a second process step, the oxidised ore (covered and submerged to exclude oxygen) is leached by recirculating the neutral bisulfide lixiviant saturated with H2S. Precious metals are recovered from the pregnant bisulfide solution by contact with activated carbon or other conventional techniques. -
Since HS ions and H2S molecules diffuse more slowly than cyanide ions and oxygen molecules, slower gold dissolution can be expected than with cyanide leaching if the same concentrations of reactants are used. Possible problems with regard to obtaining high gold recoveries include passivation and adsorption effects, which are not well understood. Investigations into using polysulfides to leach gold have been carried out by Chen et al. (1996). Gold dissolution can occur as a result of adsorption of polysulfide on gold surfaces accompanied by oxidation of polysulfide as follows: Au/Sx → AuS- + (x-1)S0 + eAu/Sx → Au/Sx + 2eS0 + 2e- → S2Chen et al (1996) reported 90% gold extraction from a sulfide concentrate at 50°C without the addition of an oxidant. However a relatively high polysulfide concentration is required for high gold extraction. Similar processes where polysulfide was generated in situ such as the “Lime/Sulfur synthetic solution” and “Phase Transfer Catalysts” have also been described (Deng et al., 1984; Zhang et al., 1993) but in these cases thiosulfate is also present. In mixtures of thiosulfate and polysulfides, the polysulfides only act as a lixiviant when no other oxidant is present. In the presence of copper, polysulfides precipitate with copper to form CuS. 6.2 NITROGEN SPECIES CATALYSED PRESSURE LEACHING PROCESS In the “nitrogen species catalysed pressure leach process” developed at Montana Tech, the sulfur residue is converted into a gold extracting lixiviant (Anderson, 2003,2008). The principles of the reaction of a sulfide mineral with nitric acid in conjunction with sulfuric acid are shown below. It is postulated that the actual reaction species is NO+, and not NO3- , which reacts with the mineral and oxidises the sulfide to sulfur at an Eh of around 1.45V (vs SHE). 2 MeS(s) + 4NO
+ (aq) →
2+
2Me
(aq)+
o
2S + 4NO(g)
By partially oxidising the sulfide to elemental sulfur instead of sulfate, the gold can be accumulated in the elemental sulfur, recovered from the other leached solids, and then leached via alkaline sulfide lixiviation whereby the sulfur containing the gold is dissolved in an alkaline solution. The combination of 2sodium hydroxide and elemental sulfur results in the formation of sulfide (S ), sodium polysulfide (Na2SX) and sodium thiosulfate (Na2S2O3) as follows: 4S0 + 6 NaOH → 2Na2S + Na2S2O3 + 3H2O (X-1)S0 + Na2S → Na2Sx (where X= 2 to 5) The gold is leached by polysulfide and thiosulfate as follows 0 2Au + S5 → AuS- + e0 Au + S2O32- → AuS2O3 + e-
17 ALTA Free Library www.altamet.com.au
The gold can be recovered by electrowinning, chemical precipitation, cementation, solvent extraction or ion exchange. The barren alkaline sulfide solution can be recycled for further gold leaching or processed with low temperature oxidation to produce sodium sulfate. As such, it is claimed that there is no environmental or toxicological issue in the use of alkaline sulfide gold recovery. 6.3 APPLICATIONS Bisulfide leaching is clearly more applicable to sulfidic ores rather than oxidised ores. Touro and Wiewiorowski, (1992) described a process of steam heating finely ground ore (50% solids) to 35-45째C and then conditioning with 2-5 kg/t of H2S gas to make a gold-sulfide complex. A chelating agent such as 0.5 kg/t EDTA may be added to complex the calcium. Sulfur dioxide (15-50 kg/t, depending on the lime content of the ore) was then injected into the pulp to reduce the pH to the optimum range of pH 4-5. The pulp was then air agitated to maintain an oxidising atmosphere for 16-20 hrs in the presence of an ion exchange resin to simultaneously dissolve the gold and transfer it to the resin. The loaded resin was finally screened from the pulp and treated to recover the gold and regenerate the resin. Gold recovery obtained was 80% compared with 30% by CIL cyanidation in some ores. Initial bisulfide leaching tests using the YES process attained 25-31% gold extraction from a bio-oxidised Nevada ore compared with 88% by conventional cyanide leaching (Hunter et al., 1998). At the same time, 39-81% silver was recovered compared to 86% with cyanide. However tests conducted in a pressure vessel to allow higher bisulfide concentrations gave 75% gold extractions. Some applications of the nitrogen species catalysed pressure leach process to treat auriferous copper sulfide concentrates have been reported (e.g. Anderson, 2001, 2003). Pre-oxidation of 100 g/L concentrate (ground to 10 microns) was carried out with 175g/L H2SO4 at 620 kPa and 125째C, in the presence of 2 g/L nitrogen species. This was followed by alkaline leaching, resulting in 98.3% gold extraction. 6.4 CURRENT STATUS Further fundamental investigations on gold dissolution processes involving sulfur chemistry are required. A better understanding of the adsorption and precipitation reactions, which reduce gold extraction, also require further investigation. The advantages the process offers over cyanidation include lower reagent costs and the ability to leach preg-robbing ores and other ores not amenable to cyanidation. It may also selectively leach precious metals from base-metal concentrates (Hunter et al., 1998). One of the limitations would seem to be the bioreduction of sulfate ion using organic substrates for bisulfide regeneration. The bisulfide process is recyclable, but theoretically oxidation to sulfate is possible. A major drawback is that H2S, which is also generated, has an occupational health standard very similar to HCN. Extremely long retention times and a closed system would be required. The nitrogen species catalysed pressure leach process is less hazardous and has been successfully demonstrated at laboratory scale on silver ores. Further investigations on treating gold ores are required. The process is most probably better suited for the extraction of multi elemental systems such as base and precious metals.
7. AMMONIA LEACHING Ammonia is most commonly used as an additional reagent in cyanidation for copper-containing ore bodies. However the laboratory use of ammonia as a lixiviant for refractory gold ores at high temperatures has been reported. The process was recently reviewed by Han (2001). 7.1 PROCESS CONDITIONS Ammonia leaching of gold in the presence of an oxidant is carried out at temperatures between 100 and 300 째C and 600-1000 kPa of pressure using 5-10 g/L copper(II) as oxidant, 5.5M free ammonia and 0.5M ammonium sulfate. Leaching times are short, around 1 to 4 hours. At ambient temperatures gold is passivated and gold dissolution in ammonia solutions is observed only above 80째C (Meng and Han, 1993). The most effective oxidant for gold dissolution in ammoniacal solutions is Cu(II) as represented in the following reaction (Guan and Han, 1996): Au + Cu(NH3)42+ = Au(NH3)2+ + Cu(NH3)2+
18 ALTA Free Library www.altamet.com.au
The dissolution rate of gold increases with increasing copper and ammonia concentrations providing there is sufficient ammonia to complex copper. Optimum pH conditions appear to be around 9.5. Increasing temperature also increases the gold dissolution rate. Alternative and less effective oxidants include oxygen, hypochlorite, peroxide and Co(III). The oxidation of sulfides in the ore produces some thiosulfate which complicates the leaching process. The use of combined halogen and ammonia as a gold leaching process has also been examined. High gold o recoveries are achieved at <100 C with iodine providing the best oxidant in terms of the rate of dissolution of precious metals (Han, 2001). Iodine, I2 is a very effective oxidant (Peri et al., 2001) where gold can react with ammonia in the presence of iodine as an oxidant to form gold-ammine. +
Au + 2I2 + 4NH3 = 2Au(NH3)2 + 4I
-
One of the advantages of such a reaction is the ability to regenerate iodine by reaction with oxygen. However no applications on the use of iodine as an oxidant have been published. 7.2 APPLICATIONS Sulfidic and carbonaceous refractory ores have been successfully treated by the ammonia process (Han and Fuerstenau, 2000) with >95% gold extraction achieved in 2 to 4 hours compared with <70% by conventional cyanidation. 7.3 CURRENT STATUS The capital and operating costs are the limiting factors in the implementation of this process, since high pressures and temperatures are required. Consequently it would only be viable in isolated cases (e.g. spent catalysts or very high grade concentrates).
8. BACTERIA AND NATURAL ACID LEACHING Gold has been reported to be solubilised in biological or natural systems by a combination of microorganisms and amino acids such as glycine, in the presence of an oxidant. High gold recoveries have been reported using both amino and humic acids. However the rate of gold dissolution is extremely slow and therefore could only be used in treating low grade ores by a heap leach process (Sparrow and Woodcock, 1995). 8.1. CONDITIONS Typical conditions used for leaching gold with bacteria and natural acids are listed in Table 6. Table 6: Typical bacterial leaching conditions used for gold Reagent
Ligand
Oxidant
Bacteria and natural acids
Amino acids
O2 or KMnO4
Humic acids
KMnO4
Gold complex in solution [Au(CH2NH2COO)2] and (possibly gold glycine)
-
Typical leaching conditions
3-5g/L amino acids 3-5 g/L KMnO4 and 1g/L merthiolate , Fermentation fluid obtained from cultivation of Bacillus subtilis strain III-5 5g/L humic acid 10-15 g/L alkali, 2-3 g/L KMnO4
pH
9.5
acidic
Dissolution of gold is enhanced by the addition of an oxidising agent such as sodium peroxide and by the selective breeding of more active strains of organisms. The reaction of amino acids with gold in the presence of permanganate produces a complex anion containing the salt of the relevant acid linked to gold (I) through gold-nitrogen bonds. An example of a gold dissolution reaction using amino acid glycine as the ligand is as follows: 4Au + 8NH2CH2COOH + 4 NaOH + O2 = 4Na[Au(NH2CH2COO)2] + 6H2O
19 ALTA Free Library www.altamet.com.au
Bacteria, most notably Bacillus subtilis, in the presence of amino acids such as glycine and an oxidant, have been shown to dissolve gold from ore samples. Groudev and Groudeva (1990) found that Bacillus subtilis was the most active solubilizing bacterium among 18 varieties tested. The aurous gold complex is an anion containing the salt of the relevant organic acid. Groudev and Groudeva (1990) found that under optimum leaching conditions (5 g/L amino acid, minimum 3-10 g/L potassium permanganate, 1.0 g/L sodium merthiolate, at pH 9.5, Eh >0.5 V, 30°C) nearly 90% gold extraction was obtained in 3 days by agitation leaching, and 70% extraction in 150 days by percolation leaching. Humic acid and other naturally-occurring organic acids have been studied by the U.S. Geological Survey, as mobilizing agents for gold in acidic swampy environments. Low gold solubility and slow kinetics militate against commercial use. Mineev and Syrtlanova (1984) reported that a solution containing 5 g/L humic acid and 10-15 g/L of alkali in the presence of 2-3 g/L KMnO4 leached 44% of the gold from a quartz-carbonate ore by percolation leaching in 45 days. Leaching the ore in pachuca tanks gave 69% extraction in 96 hours. Reagent consumptions were 2 kg/t humic acid, 0.7 kg/t KMnO4 and 8 kg/t NaOH. They estimated from cost data that the system is economically viable for ores containing 12.5 g/t Au. Fan et al (1992) also used humic acid and some unspecified additive when column leaching several ores. They appear to have obtained 40-80% gold extraction with humic acid over a 20 day period compared with 60-90% with cyanide. The ability of plants to take up gold has been long recognised (Girling and Peterson, 1980). Gold can be taken up actively, by utilising the plant metabolism, or passively, by means of the carbonyl functional groups on plant tissues. Gold ions from aqueous solutions were recovered and reduced to elemental gold colloids by alfalfa biomass (Gardea-Torresdey et al. (1999ab). The adsorption of gold(III) ions from solution by dead alfalfa tissues (Medicago sativa) is almost independent of pH, but increases with temperature (Gamez et al., 2000). 8.2. CURRENT STATUS The processes are not well understood and consequently no practical process exists. This is an area where further studies should be carried out, particularly as the substances are generally non-toxic. Much work has been carried out on the use of micro-organisms in the treatment of minerals and effluents and this is an area where active research is increasing. The uptake of gold by plants and the mobilisation of gold in soils could have limited commercial application for a secondary recovery process.
9. THIOCYANATE LEACHING Thiocyanate has been known for a long time to act as a lixiviant for gold (White 1905). Early work examined acidic thiocyanate solutions to recover gold and uranium simultaneously from South African gold ores (Fleming, 1986). However the majority of studies on leaching rates, mechanisms and thermodynamics of the thiocyanate system have been published by Monhemius and co-workers (Barbosa-Filho and Monhemius, 1994a,b,c; Ball and Monhemius, 1995). Further studies on thiocyanate leaching and gold recovery are being carried out at the University of Utah, USA (Li et al., 2008). 9.1. CONDITIONS Gold can be leached by 0.01-0.05 M thiocyanate at potentials of around 0.4-0.45 V at pH 1-3 in the presence of either ferric ions (2-5 g/L) or peroxide as oxidant. The simplified reaction can be written as follows: -
3+
Au + 4SCN + 3Fe
-
2+
= [Au(SCN)4] + 3 Fe
Thermodynamic analyses indicated that at the upper limit of the potential range [Au(SCN)4]- is believed to be the gold species formed but with decreasing potential, [Au(SCN)2] is formed. It was proposed that intermediate species such as (SCN)2 and (SCN)3 act like iodine both as oxidants and complexants. In addition various iron thiocyanate complexes can also be present under certain conditions. More recently, Barbosa-Filho and Monhemius (1997) showed that thiocyanate can leach gold in the pH range of 1 – 3 at temperatures up to 85 °C. However, increasing the pH of the thiocyanate leaching solution has a detrimental effect on the gold dissolution rate (Kholmogorov et al., 2002), presumably because of Fe(III) hydrolysis.
20 ALTA Free Library www.altamet.com.au
Addition of small amounts of iodine or iodide has been shown to increase the rate of gold dissolution in the iron (III) thiocyanate system (Barbosa-Filho and Monhemius, 1994d). This is most probably due to a synergistic effect as a result of the formation of iodine-thiocyanate species (such as I2SCN and I(SCN)2 ) (Monhemius and Ball, 1995). Besides being destroyed by oxidation during leaching, thiocyanate is known to form complexes with several metal cations. For example, a precipitate is formed with copper in the leach solution (probably [Cu(CNS)2]), and a ferric ferri-thiocyanate complex (blood red) is formed with iron. While it may be possible to recirculate the thiocyanate in the iron complexes the thiocyanate in the precipitated copper complexes would be lost from the circuit. Thus it can be expected that reagent consumption would be significant. The sulfide minerals significantly accelerate the oxidation rate in the order of pyrite, chalcopyrite and then galena. Activated carbon has been shown to not only catalyse the redox reaction between the thiocyanate and ferric ion but also adsorbs free thiocyanate as well, contributing to thiocyanate consumption. 9.2 APPLICATIONS Kholmogorov et al. (2002) have investigated the use of thiocyanate solutions to extract gold from a chemically treated arsenopyrite concentrate. The concentrate, containing approximately 110 g/t gold, was leached at room temperature for 3-5 hours with 39 g/L KSCN and 0.8 g/L Fe(III) at pH 2.2 and a pulp density of 5 to 10% solids. Gold extraction of 89 to 93% was achieved under these conditions. Comparative tests using 2.45 g/L NaCN at pH 11.3 resulted in 87% gold recovery in 89-96 hours. Work on the pregnant leach solutions demonstrated that gold could be recovered by adsorption onto activated carbon and ion exchange resins. Complete desorption of gold from resins was achieved by using sulfuric acid/thiourea solutions at room temperature and from activated carbon by using basic thiourea o solution at 150 C. Recently both bottle roll leaching tests and column extraction tests were performed to assess the effect of thiocyanate and ferric ion concentrations on gold recovery from a bio-oxidised low-grade (2.13 g Au/t) refractory sulfidic ore from Nevada (Wan et al., 2003). In bottle roll tests conducted on ground biooxidised ore, the gold extraction after 24 hours was 64% using 2.9 g/L SCN- and 5.6 g/L Fe(III) (pH 2, 20% solids, P80 = 75 µm). However comparable cyanidation test work gave 69% gold extraction. Increasing the initial ferric ion concentration had little effect on the gold dissolution and thiocyanate degradation was observed to be rapid in the initial contact with ore. For the column tests, the biooxidised ore was crushed to –38.2 mm. After 16 days of thiocyanate treatment, gold extraction was 52% and 10% higher than that obtained with cyanide. However, thiocyanate consumption (0.6-0.8 kg/t of ore) was twice as high as cyanide consumption (0.33 kg/t of ore). 9.3 CURRENT STATUS This lixiviant may be suitable for most ore types and recycling of the leach solution could be possible if the temperature is not too high to cause significant decomposition. Unfortunately, in practice, a high temperature around 85°C is necessary to achieve satisfactory leach performance. The low pH and higher temperatures would require high capital costs and high operating costs compared to cyanidation. The limited availability of thiocyanate is also a restriction and if thiocyanate had to be detoxified by oxidation to cyanate and sulfate, it would further increase the operating costs. One potential benefit is in the simultaneous leaching and recovery of both gold and uranium in ores containing uranium and gold.
10. RECOVERY PROCESSES Equally important to leaching is the selection of a method to recover gold. A major disadvantage of replacing cyanide with an alternative lixiviant remains that no suitable recovery of gold from solution has been conclusively demonstrated as economic. Most research has been carried out on thiosulfate, thiourea and halide leaching systems with gold recovery by cementation, solvent extraction, carbon, ion exchange resin or electro winning techniques being investigated. A comprehensive review of options for gold recovery from non-cyanide solutions was first presented by Wan et al. (1993). Possible recovery options from thiosulfate solutions have also been reviewed by Grosse et al., (2003) and by Aylmore and Muir (2001). Application of recovery for many of the lixiviants has already been discussed, but is now summarized collectively.
21 ALTA Free Library www.altamet.com.au
10.1 PRECIPITATION METHODS In general, precipitation methods have been considered for clarified leach solutions, particularly heap leaching, whilst carbon and resins have been considered for adsorption from slurries. Table 7 shows chemical species that are capable of reducing the various gold lixiviant complexes to gold metal. Table 7: Electrode potentials for gold complexes and possible reductant systems for gold precipitation (Marsden and House, 2006) Gold complex/Gold metal + 0 Au /Au AuCl2 /Au0 -
0
Au(SCN)2 /Au + 0 Au(NH2CSNH2)2 /Au 30 Au(S2O3)2 /Au
-
0
Au(CN)2 /Au
0
E 1.69 1.11 0.77 0.66 0.38 0.17 0.15 0.14 0 -0.39 -0.48 -0.57 -0.75 -0.76 -1.66
Reduction system Fe3+/Fe2+ (Iron) SO42-/H2SO3 (sulfur dioxide) Sn4+/Sn2+ 0 S /H2S + H /H2 (Hydrogen) H2CO3/COOH)2 So/S2- (Sulfide) 3-
BO3 /BH4 (borohydride) Zn2+/Zn (Zinc) Al3+/Al (Aluminium)
Aluminum can be an effective reductant for chloride, thiocyanate and thiosulfate complexes, however the purity of precious metals products obtained by this method is generally low. Zinc is unsuitable for acidic solutions because of its high solubility and generation of hydrogen gas resulting in excessive reagent consumption. Applications of zinc and copper have been effective in the thiosulfate leach system particularly for heap leach applications but again high concentrations of copper (10 to 20%) in the product requiring further processing. Detailed kinetic studies by Guerra and Dreisinger (1999) on the copper cementation process that increased temperature (30-50°C) and a higher pH/ammonia concentration enhanced cementation performance, whereas the presence of sulfite and copper ions in solution negatively affected cementation performance. Sodium borohydride can also be used as an efficient agent for reducing gold and silver in clarified acidic solutions of thiourea and thiocyanate, thiosulfate at room temperature. The Au(I) ion is reduced to metallic gold in the form of very fine crystals. A complete reduction of gold can occur with a sodium borohydride to gold molar ratio of 0.625 at a pH of 6 over a one hour time period (Awadalla and Ritcey, 1991, 1993; Groves and Blackman, 1995). However, the presence of ferrous ion, cobalt, nickel or in particular, copper in solution decrease the efficiency of borohydride to reduce gold because of extensive co-precipitation of other metals. Gold recovery has also been achieved by sparging or pressurising thiosulfate or thiourea solutions containing gold with hydrogen, but a catalyst of nickel or platinum is required and the pressured equipment used is more expensive than other recovery approaches (Deschenes and Ritcey, 1990; Deschenes, 1987). Finally dissolved gold and silver have been recovered by the addition of a sulphide, bisulfide or hydrogen sulphide solution with regeneration of thiosulfate (Kerley, 1981; Flett et al., 1983). West-Sells and Hackl (2005) successfully demonstrated the application of sulfide precipitation following clarification in the copper ammoniacal thiosulfate tank leach process which enabled the recycling of reagents. 10.2 ADSORBENT MATERIALS Activated carbon and resins have been investigated on most of the alternative lixiviants to allow gold recovery directly from pulp. These are summarised in the following Table.
22 ALTA Free Library www.altamet.com.au
Table 8: Adsorbent materials used for the recovery of gold from various lixiviant systems Adsorbent
Lixiviant
Carbon
Thiourea
Strong base resins
Cation ion exchange resins polyacrylic ester adsorbents
Competing elements Formamidine disulfide elemental sulfur catalyses the oxidation/ degradation of thiourea Thiourea
Elution process(es)
References
sodium cyanide sodium sulphide organic solvents
Yen et al, (1992) Wan et al., (1995) Fleming (1987) Deschene (1986)
Halides Chloride Cl-/Br-
heavy metals chloride halides
Burn carbon Cyanide/thiosulfate/ Thiourea
Intec Gold process/N-Chlo
Thiocyanate
thiocyanate
Cyanide/ethanol
Thiosulfate
Base metals (Cu, Ni, Hg) Polythionates
• • • •
Thiocyanate Trithionate Sodium nitrate Sulfite salt mixture
Kholmogorov et al., (2002); Monhemius and Ball (1995) Wan (2003) Fleming et al., (2003). Nicol and O’Malley, 2001; Jeffrey, (2007)
Thiocyanate
Thiocyanate Ferric ion Base metals
Alkaline thiocyanate/cyanide
Kholmogorov et al. (2002) Fleming (1986)
Thiourea
Base metals (Cu, Ni, Hg) Sulfur coatings
Ammonium thiosulfate
Chloride
high selectivity for gold Chloride
alkaline solution
Simpson et al., 1984; Nakahiro et al., 1992; Bjerre et al., 1989 Wan et al., (1995). Ball and Wan, (1990)
Activated carbon is preferred for gold recovery from cyanide solutions because it can be added to the 3pulp and avoids soluble losses in tailings. Unfortunately it has a low affinity for Au(S2O3)2 (Gallagher et al., 1989). While some workers have achieved high gold recovery, the concentration of gold loading on carbon was too low (5mg Au/g carbon) to be considered practical or economic (Jiexue and Qian, 1989; Abbruzzese et al., 1995; Yen et al. ,1998,; Kononova et al., 2001). To circumvent the problem of low 3loading/low affinity for Au(S2O3)2 , gold can be adsorbed from a thiosulfate solution by carbon after adding a small (stoichiometric) amount of cyanide to the system (Lulham and Lindsay, 1997). Gold recovery from acidic thiourea solutions by activated carbon adsorption is reported to be very high (Deschene, 1986) with reported maximum recoveries of 90% after 1 hour from a solution containing 27 mg/L gold and a carbon concentration of 20 g/L, with loadings as high as 15-17 wt/wt% possible. However high losses of thiourea (30%) have been reported in the leach solution due to adsorption by activated carbon (Kavanagh et al (1994), but can be washed out and collected in the absence of oxygen with hot water (Schulze, 1984). The loading mechanism is similar to that observed in the cyanide system (Fleming, 1987). Formamidine disulfide, an oxidation product of thioureas, strongly adsorbs onto carbon and subsequently further oxidises to elemental sulfur. The elemental sulfur physically deposits in the carbon macropores, thereby impeding the diffusion and adsorption of the gold thiourea complex. Furthermore, activated carbon catalyses the oxidation/degradation of thiourea resulting in the formation of elemental sulfur. Elution of the gold thiourea complex from the loaded carbon is found to occur readily. About 90% of the gold can be eluted from the carbon with inorganic stripping solutions (e.g. sodium cyanide or sodium sulfide) or solutions with organic solvents (e.g. Butanol, Acetonitrile, Diethyl ether, ethanol, acetone in decreasing effectiveness, Yen et al, 1992; Wan et al., 1995). All elution methods generally recover precious metals, and a further scrubbing step is necessary if adsorbed impurities have to be removed in order to regenerate the capacity of activated carbon recycle on the efficiency of gold recovery. Carbon was used at the New England mine in NSW to recovery gold and copper from thiourea solutions. The carbon product was sold to the smelter.
23 ALTA Free Library www.altamet.com.au
Gold recovery from halide solutions using activated carbon has two major problems. First, reduction of gold by activated carbon makes conventional stripping difficult. Second, by the finely precipitated metallic gold can be abraded from the activated carbon and lost in the tailings. Several researchers have investigated the fundamentals of gold extraction by activated carbon from chloride solutions under various conditions. It has been found that metallic gold deposits superficially on the carbon surface and that the uptake of gold by activated carbon follows a mechanism whereby gold (III) is reduced to the metallic state (Siegel and Soto, 1984; McDougall et al., 1987; Hiskey et al., 1990; Wan et al., 1995). The performance of a carbon in chlorine leaching process for different ores was evaluated by Greaves et al (1990) which showed that this was not as effective as cyanide in leaching gold, however optimum conditions were not accessed. The gold-chloride complex loads very highly (up to ~100 kg/Au/tonne) onto activated carbon. Loading of ~50kg Au/tonne is also achievable on Australian char (with cost of char approximately 1/10 of the cost of activated carbon). At these loadings, the carbon (or char) can be economically burnt to allow recovery of the gold bullion (La Brooy et al., 1994). Work by Cashion and co-workers at Monash University have developed and demonstrated carbons with properties which load gold from halide systems as the halide complex and not as metallic gold (Cashion et al., 1997). Unfortunately the carbon used is soft and therefore not abrasion resistant. In addition, no research in real pulp slurries or on the effect of impurities and recycling of carbon has been reported. In the Intec Process pilot plant gold was successfully loaded onto carbon at up to 1% w/w, with no loss of carbon activity detected over five loading/washing/elution cycles. The adsorption behaviour of gold in containing the chloride bromide complex clearly changes the loading characteristics compared with direct chloride solutions. Some carbon adsorption and elution tests have been reported for thiocyanate (Kholmogorov et al., 2002; Monhemius and Ball; 1995; Jinshan et al., 2008). A dsorption of gold has been observed in some cases to be very rapid onto activated carbon with little iron adsorption. In contrast the rate of adsorption of gold-thiocyanate complex onto activated carbon has elsewhere been reported to be very slow (Wan 2003). Presumably other factors such as the presence of other anions present in solution may compete with gold thiocyanate adsorption. Significant reduction of the ferric ion occurs in the presence of carbon and has been attributed to the carbon catalysing the reduction. Attempts to elute the gold from the loaded carbon in thiocyanate solutions have been unsuccessful. Attempts to strip gold from carbon in 311% thiourea and 3-4% H2SO4 solutions were also unsuccessful at room temperature with only 15% of the gold desorbed, but increased to 94% at temperatures up to 150 °C (Kholmogorov et al., 2002). However conventional cyanide/ethanol stripping of the loaded carbon was possible. Cation exchange resins such as AG-50W-X8 or Amberite 200 have been tested by several investigators (Simpson et al., 1984; Nakahiro et al., 1992) for gold recovery from thiourea solutions. Although the results reported suggested that cation exchange resins would be effective, results from more recent testing indicate that effective gold recovery from acidic thiourea using cation exchange resins is not possible. The dissolution of a variety of metals in the acid leach results in these metal cations being loaded on to the cation exchange resin, thereby substantially reducing the gold loading capacity. In addition elemental sulfur coated the resin and significantly reduced resin capacity. Until a selective cation exchange resin is developed, it is unlikely that gold recovery from acidic thiourea solutions will be a viable process alternative. Resins for recovering gold from thiosulfate leached pulps have been a challenge to control the loading and competitive nature of thiosulfate degradation products (Nicol and Oâ&#x20AC;&#x2122;Malley, 2001; Fleming et al., 2003). More recently the development of a sulfite based resin elution process has enabled the elution and subsequent electrowinning process to be more effective in reducing reagent consumption costs and improving resin recycling qualities (Jeffrey, 2008). The loaded resin from the adsorption circuit is transferred to an elution column where the copper present on the resin is stripped in an ammonia and ammonium sulphate mixture (or ammonia/ammonium thiosulfate). The copper containing eluent is then returned to the leach circuit. For gold elution a mixture of sodium chloride and sulfite is used to strip the gold. The chloride itself is not effective to elute gold thiosulfate complex without the presence of sulfite ions. The addition of sulfite forms a mixed complex with gold (Nimal et al., 2005) and allows elution using lower ionic strength solutions. Both strong and weak base resins can load gold from acidic chloride solutions, however most commercial resins are not stable under the oxidising conditions that exist in halide/halogen systems. Gold adsorption from neutral polyacrylic ester adsorbents (XAD-8) appears more favourable. The
24 ALTA Free Library www.altamet.com.au
extent of gold loading is a function of pH. Therefore gold can be adsorbed in acidic solutions and be removed by stripping with an alkaline solution. The alkaline solution can be thiosulfate or cyanide solution. In addition degradation by oxidising agents is expected to be less severe. Consecutive loading and stripping studies carried out by Ball and Wan (1990) over 16 days showed no degradation of the resin. Also polyacrylic ester adsorbents appeared to provide a high selectivity for gold with respect to base metals ions (Wan et al., 1995). Gold is successfully recovered on resin in the Intec process from a chloride bromide solution. Details of process are not published. Early work by Fleming (1986) evaluated the use of a strong base resin (A101Du supplied by Sentrachem Pty Ltd) to recover gold and uranium from acid solutions. Selectivity for silver and gold over all base metals is high except for copper. However extraction of gold was poor at low resin flow rates. Selectivity of gold (and uranium) increases with decreasing thiocyanate concentration. Thiocyanate ions are strongly adsorbed on to strong base resins and are not readily removed. Fleming (1985) patented a process for removing thiocyanate from resin. The regeneration process involves a stripping solution of either 5BV of 1M ferric sulfate or 0.5M ferric nitrate solution to displace about 90% of the thiocyanate from the resin. Solvent extraction can only be applied to clarified solutions containing relatively high concentrations of gold and silver. Processes have been described by
11. ECONOMIC EVALUATION To justify a mining project using an alternative lixiviant to cyanide, a complete financial analysis covering capital investment, operating expenses, revenue, water treatment and monitoring is required. However it is difficult to obtain a complete picture of how different reagents respond to different ores due to the lack of detailed information in the published literature and limited pilot plant data. Table 7 shows comparative investigations carried out by several authors illustrating the variation in conditions, reagent consumptions and gold extractions observed in treating specific ores with different reagent systems. Clearly the high reagent concentrations associated with the use of alternative lixiviants result in higher reagent consumptions compared with cyanidation. In cases where cyanide consumption is high, e.g. the reaction of cyanide with limonitic gangue, alternative lixiviants can be more favourable than cyanide for leaching gold (Monhemius and Ball, 1995). Although some alternative reagents can give better gold extractions with carbonaceous or refractory gold ores, cyanidation often achieves the highest gold recovery. For high grade ores, the economic benefit of increased gold recovery usually outweighs reagent cost and increased recovery is the main driver for choosing a particular system.
25 ALTA Free Library www.altamet.com.au
Table 7: Comparison of leach investigations on selected ores Reagent
pH
concentrations (g/L)
Reagent
Au , Ag
consumption (kg/t)
Extraction (%)
Oxide ore (0.9 g/t Au, 5 g/t Ag) Bottle roll tests (McNulty 2000) Cyanidation
Thiourea
NaCN =
CS(NH2)2 =
0.5
2
10.5-11.0
1.1-1.3
NaCN =
0.15
CaO =
0.55
CS(NH2)2 = 3.05
73, 23
57, 22
Fe2(SO4)3 = 9.0 H2SO4 = Thiosulfate
(NH4)2S2O3 = 15 NH4OH =
9.5
3.5
48
(NH4)2S2O3 = 14 NH3 =
37,16
2
CuSO4.5H2O = 0.06 Chloride
NaOCl = NaCl =
Bromide
1
6.4-6.5
100
Br2 =
1
NaBr =
10
1.3-2.0
NaOCl =
5.55
HCl =
3.25
Br2 =
2.85
H2SO4 =
6.8
68, 22
57, 13
High grade ore (68.2 g/t Au, 2 g/t Ag)(Monhemius and Ball, 1995) Cyanidation
NaCN =
5
10.5
NaCN =
7.80
CaO = Thiourea
CS(NH2)2 = 3.8
1.5
86
2.0 -
<60
Fe2(SO4)3 = 10 Thiocyanate
NaSCN =
8.1
2
NaSCN =
1.3
94
7.42
97.5
Fe2(SO4)3 = 11 I2 =
0.5
Low grade ore (4.8 g/t Au, 2 g/t Ag) (Monhemius and Ball, 1995) Cyanidation
NaCN = 5
10.5
NaCN = CaO =
Thiourea
CS(NH2)2 = 40
2.5
5.0
CS(NH2)2 = 12.8
83
NaSCN =
95
Fe2(SO4)3 = 10 Thiocyanate
NaSCN =
8.1
2
0.54
Fe2(SO4)3 = 11 I2 =
0.5
Sulfide ore (7.8 g/t Au, 13.4 g/t Ag, 11.2 %S) (Munoz and Miller, 2000) Cyanidation
NaCN =
0.98
11
NaCN =
0.36
93.7
Thiourea
CS(NH2)2 = 11.4
2
CS(NH2)2 = 38.3
27.6
2
NaSCN =
49.5
Fe2(SO4)3 = 4 Thiocyanate
NaSCN =
8
1.1
Fe2(SO4)3 = 16 Table 8 shows this authorâ&#x20AC;&#x2122;s estimate of reagent cost for leaching a low grade oxide ore (3.2 g/t Au, 0.3 g/t Ag, 0.5% pyrite in quartz/albite/clinochlore/biotite) in the Kalgoorlie region of Australia using cyanide
26 ALTA Free Library www.altamet.com.au
and three other lixiviants based on optimum reagent compositions and estimated range of reagent consumption. Table 8 Comparative reagent costs for treating a Kalgoorlie oxide ore Cyanide Thiosulfate Thiourea Assuming mining/grinding costs fixed
Chlorine
Leaching (40% solids) #
Reagents costs (A$) Lixiviant
Lime Acid Oxidant
NaCN $1.61/kg
(NH4)2S2O3 $0.75/kg
CS(NH2)2 $3.5 /kg
NaCl $0.50/kg
$140/t 3 $0.80/m (O2)
$140/t $0.43/kg (CuSO4) $0.50/kg (NH3)
$150/t $0.30/kg (Fe2(SO4)3) $1.00/kg Na2SO3
$150/t $2.00/kg (NaOCl)
9.5 6.6 g/L S2O32-
3 3.7-8 g/L
5-6 30 g/L NaCl 3.1 g/L HCl 3.15 g/L NaOCl
Others Reagent concentrations pH 10 Lixiviant 300-350 mg/L Lime 9.1 Acid Oxidant 15 mg/L Others Temperature Reagent Consumption(kg/tore) Lixiviant 0.54 kg/t Lime (Quicklime 2.2 81%CaO) Acid 3 Oxidant 0.06 m /t Others
Reagent costs
H2SO4 64mg/L Cu as Fe from ore CuSO4 6.8 g/L total 0- 9 kg/t + NH3/NH4 NaHSO3 Ambient 2-
2-3.4 kg/t S2O3 2
3-8.4 kg/t -
0.05-0.06 kg/t Cu
12-52 kg/t 8.04 kg/t Fe 1-1.5 kg/t 0-5 kg/t NH3/NH4+ NaHSO3 Costs based on reagent consumption (A$/tore) 1.23 2.5-3.6 14.4-39.2 # Costs base on transporting chemicals to Kalgoorlie in 2001
75 kg/t 13.4 kg/t 13.6 kg/t Cl-
29.2-66.7
Without taking into account any differences in gold extraction, the reagent costs for thiosulfate are about double those for cyanide. While actual costs may vary, reagent costs for thiourea and chloride are estimated to be an order of magnitude higher than for cyanide or thiosulfate. These estimates indicate that in treating a low grade ore, alternative reagent costs can be more expensive than the value of the gold. In many cases reduced reagent consumption is achieved when a pre-treatment process has been applied. Fleming et al (2003) considered that pressure leaching followed by thiosulfate leaching was possibly economic, although in this case the variability of the ore caused significant differences in gold recovery. High reagent consumptions can obviously be tolerated where ores are refractory to cyanide, such as preg-robbing carbonaceous ores. Gold recoveries of the order of 60-80% can be achieved using thiosulfate, whereas those for cyanide extractions range from 0-15% (Wan et al, 1994; Aylmore et al unpub; West-Sells et al., 2005). An economic evaluation of all processes considered by McNulty (2000), which took into account the transport of reagents required for the leaching operations, revealed cyanide as the only option where a profit could be made by heap leaching at the McDonald gold mine. In addition, from the experience of Newmont Mining, additional costs would be associated with continuously shifting the heap to maintain chemical control for alternative lixiviants, which would increase both capital and operating costs. By
27 ALTA Free Library www.altamet.com.au
comparison, consecutive heaps can be placed on top of one another in cyanide heap leaching operations. The capital costs associated with construction materials can be significantly higher than those used in a conventional cyanide plant. The highly corrosive nature of solutions used in chloride leaching systems requires titanium impellers and corrosive resistant pumps. In thiosulfate and thiourea leaching metal surfaces need to be coated to prevent cementation of gold from gold bearing solutions. 12. ENVIRONMENTAL CONCERNS The environmental properties of the alternative lixiviants to cyanide have also been considered in a number of articles (e.g. Avraamides, 1982; Swaminathan et al 1993; DeVries and Hiskey, 1992). However, the process development work involving alternative reagents to cyanide has mostly lacked any environmental focus in terms of consideration of regulatory factors, employee health and safety, environmental protection, proper disposal of wastes and sustainable development issues that mining companies will have to address. DeVries and Hiskey (1992) first reviewed environmental implications of some of the alternative reagents to cyanide. Worker and environmental risk of reagents used in alternatives to cyanide have been evaluated in detail by Gos and Rubo (1999). De Voto and McNulty (2001) have also emphasised the less than favourable environmental aspects of the alternative lixiviants. Many issues relating to environmental concern with respect to the use of any of the lixiviants, including cyanide, will be based on local climatic conditions. Also acid production in tailings dams (especially after decommissioning) is possibly the real threat to the environment irrespective of leaching process used. However the acidic leaching processes such as thiourea and halide will be more severe than others. While studies have investigated the effect of cyanide in the environment, there are no studies on the effect of alternative lixiviants on the environment. This is primarily as a result of the fact that there are no commercial alternative processes. However investigations on other mineral processes are available which provide sufficient information to indicate that some of these alternatives, such as those using ammonia, are more of a worker and environmental concern than cyanide. This has been illustrated by Gos and Rubo (1999). A major aspect of minimising the worker and environmental risk would be selection of a process that minimises the quantity of chemicals used. Since many of the alternative lixiviants use around 50 times the concentration used in cyanide leaching, it is important to be able to recycle as much reagent as possible, not only to reduce costs to the leaching process, but also to prevent build up in tailings dams. Methods for reagent recycling or destruction of cyanide have been demonstrated to be successful on economic and environmental grounds, and many have been utilised in the mining industry. However, while some alternative reagents to cyanide can be recycled further investigations are required.
13. CONCLUSIONS Of all the processes available thiosulfate and chloride based leaching appear to be the most favourable options to replace cyanide. However, in developing an alternative process, reductions in reagent consumption and improvements in recovering gold from solution are required. Feasibility studies into the worker and environmental risks associated with the process have to be taken into account before commercialisation, particularly in environmentally sensitive areas. Extensive investigations carried out by Newmont, Barrick and Placer Dome to semi commercial scale using thiosulfate or thiourea as alternative lixiviants to cyanide has occurred not because of any particular concerns with the environment or health and safety concerns while using the cyanide system, but because cyanide is unable to effectively extract gold from a proportion of their carbonaceous ore resources. Oxidative chloride leaching and chlorination processes have been used extensively as pre-treatment processes to oxidise refractory or carbonaceous ores prior to conventional cyanidation and carbon in pulp technology, rather than as gold leach process. While high gold dissolution by halides has been demonstrated, improvements in gold recovery methods are required.
28 ALTA Free Library www.altamet.com.au
Despite pilot plant trials on many of the alternative options to cyanide, none have been successfully commercialised. Many of the pilot scale studies have highlighted problems associated with scaling up a new technology from laboratory experiments to pilot plant or commercial stage. These include: •
Halides, thiourea and thiosulfate leaching are very susceptible to variations in mineralogy of the feed being leached, and may require constant adjustments in regulating reagent concentrations to maintain optimum leaching conditions.
•
Except for halogens, most alternative reagents do not exhibit fast leach kinetics compared with cyanide at similar concentrations.
•
Shortfalls in reagent recovery and in gold recovery are evident when running a continuous leaching operation due to adsorption or reaction with other ore minerals.
•
Conditions in heap leach operations on a large scale are more difficult to maintain compared with cyanide, and are not easily scaled up from column leach experiments in the laboratory.
•
Higher reagent concentrations compared to that used in cyanide leaching, make recycling important to make gold extraction by alternative lixiviants economical.
•
The complexity of the process is substantially greater than cyanide leaching.
14. ACKNOWLEDGEMENTS The author acknowledges the input in the past of former colleagues. In addition the approval of my current employer Bateman Engineering to attend and to contribute to this conference is appreciated.
15. REFERENCES Abe, Y., Hosaka, H., Takebayashi, K., Ishiguro, Y. and Yoshimura, A. (2009) Process of leaching gold, US US20090241735, Oct 2009. Alodan, M and Smyrl W. (1998) Effect of thiourea on copper dissolution and deposition, Electrochimica Acta, 44, 299-309. Anderson C. G. (2003) The industrial non-cyanide hydrometallurgical recovery of silver and gold utilizing nitrogen species catalyzed pressure oxidation. in Hydrometallurgy 2008 Proceedings of the 6th International Symposium. (Eds. Young, C.A., Taylor, P.R., Anderson, C., and Hio Y.), Minerals, Metals and Materials Society, Warrendale, PA, pp. 858-884. Anderson C. G. (2008) NSC Pressure Leaching: Industrial and potential applications in Hydrometallurgy 2008 Volume 1: Leaching and Solution Purification, (Eds. Young, C.A., Alfantazi, A., Anderson, C., James, A., Dreisinger, D. and Harris, B.), Minerals, Metals and Materials Society, Warrendale, PA, pp. 89-104. Angelidis, T.N., Kydros, K.A. and Matis, K.A. (1993) Fundamental rotating disc electrode study of gold dissolution in iodide-iodine solutions, Hydrometallurgy, 34(1) 49-61. Ardiwilaga, S. (1999) Effects of cysteine and oxygen on recovery of cemented gold from leach liquors in a thiourea system, Minerals Engineering, 12 (6), 645-653. Arima H., Fujita T. and Yen W-T., (2004). Using nickel as a catalyst in ammonium thiosulfate leaching for gold extraction. Materials Transactions 45(2), 516-526. Avraamides, J. (1982) Prospects for alternative leaching systems for gold: A review. in Proc. Symposium on Carbon-in-Pulp Technology for the Extraction of Gold, Aus.I.M.M. Melbourne, pp 369391. Aylmore, M.G. (2004) Process for Recovering Precious Metals, Patent WO2004005556
29 ALTA Free Library www.altamet.com.au
Aylmore M.G. and Muir D.M., (2001a), Thiosulfate leaching of gold - A review. Minerals Engineering 14(2), 135-174. Aylmore M.G. and Muir D.M., (2001b), Thermodynamic analysis of gold leaching by ammoniacal thiosulfate using Eh/pH and speciation diagrams. J. Minerals and Metallurgical Processing, 18(4), 221227. Aylmore M.G. and Rae C., (2001), Factors affecting the rate of leaching of gold in ammoniacal thiosulfate and the stability of the leach solution. CSIRO Report DMR-1493, 36pp. Aylmore, M.G., (2005) Alternative lixiviants to cyanide, Chapter 21, Advances in Gold Ore Processing, Vol 15, Edited by Mike Adams, Elsevier Pty Ltd. Bagdasaryan K.A., Episkoposyan M.L., Ter-Arakelyan K.A., Babayan G.G., (1983). The kinetics of the dissolution of gold and silver in sodium thiosulfate solutions Sov. J. Non-Ferrous Met. 376, 64-68. Balaz, P., Ficeriova, J. and Leon, C. V. (2003) Silver leaching from a mechanochemically pretreated complex sulfide concentrate, Hydrometallurgy, 70 113–119. Barbosa-Filho, O. and Monhemius, A.J. (1994a) Leaching of gold in thiocyanate solutions – Part 1: Chemistry and thermodynamics, Trans.Inst.Min. &Metall., Section C: 103, C105-C110. Barbosa-Filho, O. and Monhemius, A.J. (1994b) Leaching of gold in thiocyanate solutions - Part 2: Redox processes in iron(III) thiocyanate solutions, Trans.Inst.Min. &Metall., Section C, 103, C111-C116. Barbosa-Filho, O. and Monhemius, A.J. (1994c) Leaching of gold in thiocyanate solutions – Part 3: Rates and mechanism of gold dissolution, Trans. Inst. Min. & Metall., Section C: 103, C117-C125. Bilston, D.W., Bruckard, W. J., McCallum, D. A., Sparrow , G. J. and Woodcock, J. T. (1990) A comparison of methods of gold and silver extraction from Hellyer pyrite and lead-zinc flotation middlings, in sulfide deposits – their origin and processing, Inst. of Min. and Metall., London, pp 207-221. Breuer P.L. and Jeffrey M.I., (2002). An electrochemical study of gold leaching in thiosulfate solutions containing copper and ammonia. Hydrometallurgy, 65, 145-157. Breuer P.L. and Jeffrey M.I., (2003a). A review of the chemistry, electrochemistry and kinetics of the gold thiosulfate leaching process. In Hydrometallurgy 2003. Vol 1 Leaching and Solution Purification (Eds C.A.Young, A.M.Alfantazi, C.G.Anderson, D.B.Dreisinger, B.Harris and A.James) TMS, Warrendale, pp 139-154. Breuer P.L. and Jeffrey M.I., (2003b). Copper catalysed oxidation of thiosulfate by oxygen in gold leach solutions. Minerals Engineering, 16(1), 21-30. Breuer P.L. and Jeffrey M.I., (2003c). The reduction of copper(II) and the oxidation of thiosulfate and oxysulfur anions in gold leaching solutions. Hydrometallurgy, 70, 163-173. Breuer P.L., Jeffrey M.I. and Choo W.L., (2001). Fundamental aspects of the gold thiosulfate leaching process. Cyanide: Social, Industrial and Economic Aspects, (Eds C.A Young, L.G.Twidwell and C. A.Anderson) TMS, Warrendale, pp 455-468. Briones R. and Lapidus G.T., (1998). The leaching of silver sulphide with the thiosulfate-ammoniacupric ion system. Hydrometallurgy, 50, pp 243-260. Brown T., Fischmann A., Spiccia L. and McPhail D.C., (2003). Alternative copper(II) catalysts for gold leaching: use of multi-dentate ligands to control thiosulfate oxidation. In Hydro-metallurgy 2003. Vol 1 Leaching and Solution Purification (Eds C.A.Young, A.M.Alfantazi, C.G.Anderson, D.B.Dreisinger, B. Harris and A.James) TMS, Warrendale, pp 213-226. Byerley J.J., Fouda S. and Rempel G.L., (1973a). Kinetics and mechanism of the oxidation of thiosulfate ions by copper (II) ions in aqueous ammonia solution, Journal of Chemical Society: Dalton Transactions, 889-893.
30 ALTA Free Library www.altamet.com.au
Byerley J.J., Fouda S. and Rempel G.L., (1973b). The oxidation of thiosulfate in aqueous ammonia by copper (II) oxygen complexes. Inorganic Nuclear Chemistry Letters, 9, 879-883. Byerley J.J., Fouda S. and Rempel G.L., (1975). Activation of copper(II) ammine complexes by molecular oxygen for the oxidation of thiosulfate ions, Journal of Chemical Society: Dalton Transactions, 1329-1338. Chandra I. and Jeffrey M.I., (2004). An electrochemical study of the effect of additives and electrolyte on the dissolution of gold in thiosulfate solutions. Hydrometallurgy, 73, 305-312. Chandra I. and Jeffrey M.I., (2003). Can a thiosulfate leaching process be developed which does not require copper and ammonia. In Hydrometallurgy 2003. Vol 1 Leaching and Solution Purification (Eds C.A.Young, A.M.Alfantazi, C.G.Anderson, D.B.Dreisinger, B. Harris and A.James) TMS, Warrendale, pp 169-182. Caldeira, C. L. and Ciminelli, V, S. T. (1993) Thiourea leaching of a refractory gold ore, in XVIII International Mineral Processing Congress (Vol 5: Gold Processing, Hydrometallurgy and Dewatering and Miscellaneous), Aus.I.M.M., Melbourne, pp 1123-1128. Cashion, J. D., Brown, L.J. and Jay, W.H. (1997) Mossbauer spectroscopy applied to mineralogy and minerals extraction. Resourcing the 21st Century, AusIMM Annual Conf. Australas. Inst. Mining and Metall., Melbourne, pp285-9. Chai, L., Okido, M. and Wei, W. (1999) Effect of Na2SO3 on electrochemical aspects of gold dissolution in alkaline thiourea solution, Hydrometallurgy, 53, 255- 266. Chen, J., Deng, T. Zhu, G. and Zhao, J. (1996) Leaching and recovery of gold in thiosulfate based system – a research summary at ICM, Trans. Indian Inst. Metall., 49(6), 841-849. Choi Y, Kondos, P., Aylmore M.G., Jacques McMullen and Van Weert, G. (2006) Thiosulfate insitu in precious metal recovery US 11/558,639. Chu C.K., Breuer P.L. and Jeffrey M.I., (2003). The impact of thiosulfate oxidation products on the oxidation of gold in ammonia thiosulfate solutions. Minerals Engineering, 16, 265-271. Davis A. and Tran T. (1992) Fundamental aspects of gold leaching in halide media, Proc. EDP’92 Congress, Ed: J. Hager, TMS, Warrendale, PA, pp 99-114. Davis, A., Tran T. and Young, D. (1992) Solution chemistry of iodide leaching of gold, Hydrometallurgy, 32, 143-159. Deng, T., Ke, J.J. and Chen, J. (1984) Kinetics of disproportionation of elemental sulfur in aqueous solutions, J. Chem. Ind. Engng., 3(4), 128-334. Deng, T.L., Liao, M.X., Wang, M.H., Chen, Y.-W. and Belzile, N. (2001) Technical Note: Enhancement of gold extraction from biooxidation residues using an acidic sodium sulfite-thiourea system, Minerals Engineering, 14(2), 263- 268. Deng, T. and Liao, M. (2002) Gold recovery enhancement from a refractory flotation concentrate by sequential bioleaching and thiourea leach, Hydrometallurgy, 63, 249-255. Deschenes G., Bernard D., Prud'Homme P.J.H. and Laforest D. (1994) A preliminary techno-economic evaluation of the extraction of gold from a chalcopyrite concentrate using thiourea. Minerals Engineering, 7(4), 435–448. De Voto, R.H. and McNulty, T.P. (2001) Evaluation of non-cyanide technologies for processing ore of the McDonald Gold Deposit, Montana, in Cyanide: Social, Industrial and Economic Aspects, Eds. Young, C.A., Twidwell, L.G. and Anderson, C.G., TMS, Warrendale, PA, pp. 83-95. DeVries, F.W. and Hiskey, J.B. (1992) Environmental impact of lixiviants: An overview that includes noncyanide chemistry, in Proc. of the Randol Gold Forum, Randol Intl., Golden, CO, pp. 89-92.
31 ALTA Free Library www.altamet.com.au
Dönmez, B., Ekinci, Z.., Çelik, C. and Çolak, S. (1999) Optimisation of the chlorination of gold in decopperised anode slime in aqueous medium. Hydrometallurgy, 52, 81-90. Falanga, B.J. and MacDonald, D.I. (1982) Recovery of Gold and/or Palladium from an Iodide-Iodine Solution”, US Patent 4,319,923. Fan, Z., Peng, J., An, Z. and Wang, D. (1992) Gold extraction with humic acid, in ICHM ’92 (Proc. 2nd Intern. Conf. on Hydrometallurgy) (Ed., Chen, J., Yang, S and Deng, Z.), International Academic Publishers: Beijing, pp 487-492. Fang Z., Muhammed M. (1992) Leaching of precious metals from complex sulfide ores. On the chemistry of gold lixiviation by thiourea. Min. Pro. & Extr. Metall. Reviews, 11, 39-60. Feather, A., Sole, K.C. and Bryson, L.J. (1997) The Minataur™ process: an alternative gold-refining technology. In: Randol Gold Forum ’97, Randol Intl., Golden, CO, pp. 181–185. Feng D. and van Deventer J.S.J., (2001). Preg-robbing phenomena in the thiosulfate leaching of gold ores. Minerals Engineering, 14(11) 1387-1402. Feng D. and van Deventer J.S.J., (2002). Leaching behaviour of sulphides in ammoniacal thiosulphate systems. Hydrometallurgy, 63, 189-200. Feng, D. and van Deventer, J.S.J. (2010) The effect of iron contaminants on thiosulphate leaching of gold, Minerals Engineering, 23 (5), p.399-406. Ferron, C.J., Fleming, C.A., Dreisinger, D.B., O'Kane P.T. (2000) One Step Leaching of Gold, PGM's and Base Metals from Various Ores and Concentrates, Proc. of Randol Au and Ag Forum - Vancouver, April 25-28, Randol International (Golden), 243-252. Ferron, C.J., Fleming, C.A., Dreisinger, D. and O’Kane, T. (2003) Chloride as an alternative to cyanide for the extraction of gold – going full circle? in Hydrometallurgy 2003 Volume 1: Leaching and Solution Purification, Young, C.A., Alfantazi, A., Anderson, C., James, A., Dreisinger, D. and Harris, B. (Eds), The Minerals, Metals and Materials Society, Warrendale, PA, pp. 89-104. Finkelstein, N.P. (1972) The chemistry of the extraction of gold from its ores, in Gold Metallurgy in South Africa (ed R. J. Adamson), Chamber of Mines of South Africa, Johannesburg. pp284-351. Fleming, C.A. (1986) A process for simultaneous recovery of gold and uranium from South African ores, in Gold 100, Volume 2: Extractive Metallurgy of Gold, South Afr. Inst. Min.& Metall., Johannesburg, pp. 301-319. Fleming C.A., McMullen J., Thomas K.G. and Wells J.A. (2003). Recent advances in the development of an alternative to the cyanidation process: thiosulfate leaching and resin in pulp. Minerals and Metallurgical Processing, 20(1), 1-9. Freitas L.R, Trindade R.B.E. and Carageorgos T., (2001). Thiosulfate leaching of gold-copper ores from Igarape Bahia mine (CVRD), Proc. VI SHMMT / XVIII ENTMH, Rio de Janeiro. Gamez, G., Dokken, G. K., Herrera, I., Parson, J. G., Tiemann, K. J. and Gardea-Torresdey, J. L. (2000) Chemical processing involved in Au(III) binding and bioreduction by alfalfa biomass, Proc. of the 14th Conference on Hazardous Waste Research, HSRC, pp 46-52. Gamez, G., Dokken, K., Tiemann, K.J., Herrera, I., Yacaman, M.J., Renner, M.W., Furenlid, L.R. and Gardea- Torresdey, J.L. (1999) Spectroscopic Studies of Gold (III) Binding to Alfalfa Biomass, In: L.E. Erickson and M.M. Rankin (Eds.), Conference 14th on Hazardous Waste Research: Gateway to Environmental Solutions, Proc., St.Louis, Missouri, pp. 78-90. Gardea-Torresdey, J. L, Tiemann, K. J., Gamez, G., Dokken, G. K., and Pingitore, N. E. (1999) Recovery of gold (III) by alfalfa biomass and binding characterisation using X-ray microfluoresence, Advances in Environmental Research, 3, 83-93. Girling, C. A. and Peterson, P. J. (1999) Gold in plants, Gold Bull., 13, 151-157.
32 ALTA Free Library www.altamet.com.au
Greene, B., Hosea, M., McPheason, R. M., Henzl, M., Alexander, M. D. and Darnell, D. W. (1986) Interaction of gold (I) and gold (III) complexes with algal biomass. Environmental Science and Technology, 20, 627-632. Guan Y. C. and Han, K. N. (1996) The electrochemical study on the dissolution behaviour of gold in ammoniacal solutions at temperatures above 100oC, J. Electrochem Soc, 143(6), 1875-1879. Guerra, E. and Dreisinger, D. B. (1999) A study of the factors affecting copper cementation of gold from ammoniacal thiosulfate solution. Hydrometallurgy 51, 155-172. Greaves, J. N., Palmer, G. R and White, W. W. (1990) the recovery of gold from refractory ores by the use of carbon-in-chlorine leaching, JOM, 12-14. Groenewald, T. (1977) Potential applications of thiourea in the processing of gold, J. South. Afr. Inst. Min. Metall., 77, 217-223. Grosse A.C., Dicinoski G.W., Shaw M.J. and Haddad P.R. (2003) Leaching and recovery of gold using ammoniacal thiosulfate leach liquors. A review, Hydrometallurgy, 69, 1-21. Groudev S.N. and Groudeva V. I. (1990) Small scale feasibility study of the biological leaching of gold from ores, in Minerals Materials and Industry, 14th Congress of the Council of Mining and Metallurgical Institutions, Inst.Min.& Metall.: London, p221-227. Gönen, N. (2003) Leaching of finely disseminated gold ore with cyanide and thiourea solutions, Hydrometallurgy, 69, 169-176. Gos, S. and Rubo, A. (2000) Alternative lixiviants for gold leaching – A comparison. in Proc. Randol Gold & Silver Forum 2000, Randol Intl., CO, pp. 271-281. Han, K.N. (2001). Electrochemical behavior of the dissolution of gold in ammoniacal solutions, in Cyanide: Social, Industrial and Economic Aspects, Eds. Young, C.A., Twidwell, L.G. and Anderson, C.G., TMS, Warrendale, PA, pp. 485-499. Han, K.N. and Fuerstenau, M. C. (2000) Factors influencing the rate of dissolution of gold in ammoniacal solutions. Intl. J. Mineral. Proc., 58, 369-381 Hannington, M. D. and Scott, S.D. (1989) Sulfidation equilibria as guides to gold mineralisation in volcanogenic massive sulfides: evidence from sulfide mineralogy and the composition of sphalerite, Econ Geo., 84, 1978- 1995. Heath, J.A., Jeffrey, M.I., Zhang, H.G. and Rumball, J.A. (2008) Anaerobic thiosulfate leaching: Development of in situ gold leaching systems, Minerals Engineering, 21(6). Herreros, S. (1999) Dissolution kinetics of copper, white metal and natural chalcocite in chlorine/chloride media, Hydrometallurgy, 51, 345-357. Hiskey, J.B. (1981) Thiourea as a lixiviant for gold and silver, in Gold and Silver Leaching, Recovery and Economics (Eds. Schlitt, W. J. Larson, W.C. and Hiskey, J. B.), A.I.M.E., New York, pp 83-91. Hiskey, J.B. (1988) Thiourea leaching of gold and silver-technology update and additional applications, Minerals & Metallurgical Processing, 1, 173-179. Hiskey, J. B. and Atluri, V.P. (1988) Dissolution chemistry of gold and silver in different lixiviants, Min. Proc. & Extr. Metall. Reviews, 4, 95-134. Hiskey, J. B. and Qi, P.H. (1991) Leaching behaviour of gold in iodide solutions”, in: World Gold ’91, Aus.I.M.M., Melbourne, pp 115-120. Hiskey, J.B., Jiang, X. H. and Ramadorai, G. (1990) Fundamental studies on the loading of gold on carbon in chloride solutions, Gold ’90, D. M. Hausen SME, pp369-376.
33 ALTA Free Library www.altamet.com.au
Hunter, R.M., Stewart, F.M., Darsow, T., Fogelsong, M.L., Mogk, D.W., Abbott, E.H. and Young, C.A. (1998) New alternative to cyanidation: biocatalysed bisulfide leaching, Min. Proc.& Extr. Metall. Reviews, 19(1-4), 183-197. Jain, G. X. and Hendrix, J. L. (1996) Ferric chloride leaching and metals recovery of polymetallic gold ore, in Proc. EPD Congress (Ed G. W. Warren), Minerals, Metals and Materials Society, Warrendale, PA, pp 289-296. Jeffrey M.I., (2001). Kinetic aspects of gold and silver leaching in ammonia-thiosulfate solutions. Hydrometallurgy, 60, 7-16. Jeffrey, M. I. (2007) Process for recovering metals from resins, PCT Patent WO/2007/137325. Jeffrey M.I., Breuer P. and Choo W.L., (2001). How rapidly do alternative lixiviants leach gold? In Cyanide: Social, Industrial and Economic Aspects, (Eds C.A.Young, L.G.Twidwell and C.A.Anderson) TMS, Warrendale, pp 441-454. Jeffrey M.I., Breuer P.L. and Chu C.K., (2003). The importance of controlling oxygen addition during thiosulfate leaching of gold ores. Int.J. Mineral Processing, 72, 323-330. Jeffrey, M.I. Hewitt, D.M. Dai, X. and Brunt. S.D. (2010) Ion exchange adsorption and elution for recovering gold thiosulfate from leach solutions Hydrometallurgy, 100(3-4), p 136-143 Jeffrey, M.I., Hewitt, D.M,. Dai, X. and Brunt, S.D. (2010) Ion exchange adsorption and elution for recovering gold thiosulfate from leach solutions Hydrometallurgy, 100(3-4), p 136-143. Jeffrey, M.I., Watling, K., Hope, G.A. and Woods, R. (2008) Identification of surface species that inhibit and passivate thiosulfate leaching of gold. Minerals Engineering, 21(6). Ji J., Fleming C., West-Sells P.G. and Hackl R.P., (2003). A novel thiosulfate system for leaching gold without the use of copper and ammonia. In Hydrometallurgy 2003. Vol 1 Leaching and Solution Purification (Eds. C.A.Young, A.M.Alfantazi, C.G.Anderson, D.B.Dreisinger, B.Harris and A.James) TMS, Warrendale, pp 227-244. Kavanagh, P. F., Foley, B. F., McLoughlin, J., Owens, P., Curran, S. A., McNamee, C. and Mullins, K. (1994) Leaching and recovery of gold by use of acido-thioureation on copper-mine wastes: laboratory and pilot-plant tests and processing modelling, in Hydrometallurgy ’94, Chapman and Hall, London, pp 441-461. Kavanagh, P. F., Foley, B. F., Smalley N. (2000) Recovery of noble metals lixiviation with thiourea controlled acidic solution Intenat. Pat. PCT 00/68446. 18pp. Kenna C. (1991) Extraction and recovery of gold, US Patent 5,260,040. Kenna, C. and Moritz, P. (1991) The extraction of gold from bioleached pyrite using novel thiourea leaching technology, in World Gold ’91, Aus.I.M.M., Melbourne, pp 133-137. Kholmogorov, A.G., Kononova, O.N., Pashkov, G.L. and Kononov, Y.S. (2000) Thiocyanate solutions gold technology, Hydrometallurgy, 64, 43-48. Kusnierova, M. Sepelak, V. and Briancin, J. (1993) Effects of biodegradation and mechanical activation on gold recovery by thiourea leaching. J. Metals, 45(12), 54-56. La Brooy S. R., Linge, H. G. and Walker, G. S, Review of gold extraction from ores. Minerals engineering, 7, 1213-1241 (1994). Lacoste-Bouchet, P., Deschênes, G. and Ghali, E. (1998) Thiourea leaching of a copper-gold ore using statistical design, Hydrometallurgy,. 47, 189-203. Lan X., Zhang J. and Zang J., (1993) 50 years of the extracting gold and silver by thiourea-a literature review, in Precious Metals 1993 (Ed R. K. Mishra), International Precious Metals Institute, Allentown, pp 435-448.
34 ALTA Free Library www.altamet.com.au
Langhans J.W., Lei K.P.V. and Carnahan T.G., (1992). Copper-catalysed thiosulfate leaching of low grade gold ores. Hydrometallurgy, 29, 191-203. Lapidus, G.T Gonzalez, I, Nava, J.L. Benavides, R. and Lara-Valenzuela, C. (2008) Integrated process for precious metal extraction and recovery based on electro-oxidised thiourea. Hydrometallurgy 2008 Proceedings of the 6th International Symposium Honour Robert Shoemaker, Ed Young Ca, Taylor, PR, Anderson, CG and Choi, Y, SME., Littleton, pp837-842. Li X., Ke, J., Meng X. and Li, B. (1992) Chlorine leaching of gold bearing sulfide concentrate and its calcine, Hydrometallurgy, 29, 205-215. Li, J, Miller, J. D., Wan, R. Y. (1996) Importance of solution chemistry. Factors that influence the ammonium thiosulfate leaching of gold, paper presented at SME Annual Meeting, Feb, 1996, Phoenix, Arizona, pp 159-162. Li, J. and Miller. J.D. (2002) Reaction kinetics for gold dissolution in acid thiourea solution using formamidine disulfide as oxidant. Hydrometallurgy, 63(3), 215-223. Li, J. and Miller, J.D. (2007) Reaction kinetics of gold dissolution in acid thiourea solution using ferric sulfate as oxidant, Hydrometallurgy, 89 (3), p.279-288, Dec 2007Li, J. and Miller, J.D. (2006) A review of gold leaching in acid thiourea solutions, Mineral Processing and Extractive Metallurgy Review, 27(3), p177 – 214. Li, J, Wan, R. Y. LeVeir K.M and Miller, J. D. (2008) Thiocyanate process chemistry for gold recovery. Hydrometallurgy 2008 Proceedings of the 6th International Symposium Honour Robert Shoemaker, Ed Young Ca, Taylor, PR, Anderson, CG and Choi, Y, SME., Littleton, pp824-836. Liu, J. Q. and Nicol, M. J. (2002) Thermodynamics and kinetics of the dissolution of gold under pressure leaching conditions in the presence of chloride, Can. Metall. Quart., 41(4) 409-415. Marsden, J. and House, I. (2006) The Chemistry of Gold Extraction, Ellis Horwood, London. MacDonald D.D., (1990). Review of mechanistic analysis by electrochemical impedance spectroscopy, Electrochimica Acta, 35, p 1509. McGrew, K.J. and Murphy, J.W. (1985) Iodine leach for the dissolution of gold”, US Patent 4,557,759. McNulty, T. (2001) Cyanide substitutes, Mining Magazine, May 2001, 256-261. Michel D. and Frenay J., (1999). Integration of amino acids in the thiosulfate gold leaching process. Proc. Randol Gold and Silver Forum, Randol Intl, Golden CO., pp 99-103. Molleman E. and Dreisinger D.B. (2002) The treatment of copper-gold ores by ammonium thiosulfate leaching, Hydrometallurgy, 66, 1-21. Monhemius, A.J. and Ball, S.P. (1995) Leaching of Dominican gold ores in iodide- catalysed thiocyanate solutions, Trans. Inst. Min.& Metall. Section C, 104 , C117-C124. Moussoulos, L., Potamianos, N and Kontopoulis, A. (1984) Recovery of gold and silver from arseniferous pyrite cinders by acidic thiourea leaching, in Precious Metals: Mining Extraction Processing (Ed Kudryk, V. Corrigan, D A. and Liang, W.W.), AIME, Warrendale, pp 323-335. Muir D.M. and Aylmore M.G. (2004) Thiosulfate as an alternative to cyanide for gold processing – issues and impediments. Mineral Processing and Extraction Metallurgy, Trans. Inst. Min. & Metall. Section C, 113(1), C2-C12. Muir, D. M. and Aylmore, M. G., (2005) Thiosulfate leaching of gold Chapter 22, Advances in Gold Ore Processing, Vol 15, Edited by Mike Adams, Elsevier Pty Ltd. Munoz, G.A. and Miller, J.D. (2000) Non-cyanide leaching of an auriferous pyrite ore from Ecuador, Minerals & Metallurgical Processing, 17 (3), 198-204.
35 ALTA Free Library www.altamet.com.au
Murthy, D.S.R and Prasad, P.M. (1996) Leaching of gold and silver from Miller Process dross through non-cyanide leachants, Hydrometallurgy, 42, 27-33. Naito K., Sheh C. and Okabe T., (1970). The chemical behaviour of low valence sulfur compounds V. Decomposition and oxidation of tetrathionate in aqueous ammonia solution. Bulletin of the Chemical Society of Japan, 43, 1372-1376. Nam, K., Jung, B H., An, J W., Ha, T. J., Tran T. and Myong Jun Kim (2008) Use of chloride– hypochlorite leachants to recover gold from tailing International Journal of Mineral Processing 86(1-4), p 131-140. Nesbitt, C. C., Milosavljevic, E. B. and Hendrix, J. L. (1990) Determination of the mechanism of the chlorination of gold in aqueous solutions, Ind. Eng. Chem. Res., 29, 1696-1700. Nicol, M. J. (1976) An electrochemical and kinetic investigation on behaviour of gold in chloride solutions, National Institute for Metallurgy, Johannesburg, Reports 1844 and 1846. Nicol, M. J. (1980) The anodic behaviour of gold. Part 1 –oxidation in acidic solutions, Gold Bull., 13(2), 46-55. Nicol M. and O’Malley G.P., (2001). Recovery of gold from thiosulfate solutions and pulps with ionexchange resins. In Cyanide: Social, Industrial and Economic Aspects, (Eds. C.A.Young, L.G.Twidwell and C.A.Anderson) TMS, Warrendale, pp 469-483. Nicol M. and O’Malley G.P., (2002). Recovery of gold from thiosulfate leach pulps via ion exchange. J. of Metals, Oct 44-46. Nimal Perera, W., Senanayake, G. and Nicol, M. J. (2005) Interaction of gold(I) with thiosulfate–sulfite mixed ligand systems Inorganica Chimica Acta, 358( 7), p 2183-2190. Örgül, S. and Atalay, Ü. (2000) The Prospects for an alternative gold leach reagent: thiourea, in Mineral Processing on the Verge of the 21st Century, in: Proc. of the Eighth International Mineral Processing Symposium, Antalya, Turkey, 16–18 October 2000, Balkema, Rotterdam, p. 271. Örgül, S. and Atalay, Ü. (2002) Reaction chemistry of gold leaching in thiourea solution for a Turkish gold ore, Hydrometallurgy, 67, 71-77. Perera, W. N, Senanayake, G and Nicol, M J. (2005) Interaction of gold(I) with thiosulfate–sulfite mixed ligand systems Inorganica Chimica Acta, 358(7), p 2183-2190. Peri, K. Guan, Y.C. and Han K.N. (2001) Dissolution behaviour of gold in ammoniacal solutions with iodine as an oxidant, Minerals & Metallurgical Processing, 18, 13-17. Pesic, B. and Sergent, R.H., (1992) Dissolution of gold with bromine from refractory ores pre-oxidised by pressure oxidation”, in Proc. EDP’92 Congress, Ed: Hager. J., TMS, Warrendale, PA, pp 99-114. Pesic, B., Smith, B. D. and Sergent, R. H., (1992) Gold recovery from pressure oxidised ores with bromine, in Randol Gold Forum Vancouver ’92, Randol Intl., Golden, CO, pp 287-291. Puvvada, G.V.K. and Murthy, D.S.R., (2000) Selective precious metals leaching from a chalcopyrite concentrate using chloride/hypochlorite media, Hydrometallurgy, 58, 185-191. Qi, P. H. and Hiskey, J. B. (1991) Dissolution kinetics of gold in iodide solutions, Hydrometallurgy, 27, 47-62. Qi, P. H. and Hiskey, J. B. (1993) Electrochemical behaviour of gold in iodide solutions, Hydrometallurgy, 32, 161-179. Ritchie, I.M., Nicol, M.J. and Staunton, W.P. (2001) Are there realistic alternatives to cyanide as a lixiviant for gold at the present time?, in Cyanide: Social, Industrial and Economic Aspects, Eds. Young, C.A., Twidwell, L.G. and Anderson, C.G., TMS, Warrendale, PA, pp. 427-440.
36 ALTA Free Library www.altamet.com.au
Sanuki G. K. (1990) Oxidative leaching behaviour of copper anode slime in a nitric acid solution containing sodium chloride, J. Japan. Inst. Min. & Metal., 54, 442-447. Schulze, R.G. (1985) Recovery of noble metals from ores (Chem. Abstracts 103, 90914y). Severs, K. (1999) Technological advances in treating copper concentrates – The Intec Copper Process” th Paper presented at the 128 SME Meeting, Denver, Colorado, March 1-3,. Seward, T. M. (1973) Thio complexes of gold and the transport of gold in hydrothermal ore solutions., GeoChim. Cosmochim. Acta, 37, 379-399. Senanayake, G. (2004) Gold leaching in non-cyanide lixiviant systems: critical issues on fundamentals and applications. Minerals Engineering, 17, 785–801. Senanayake G., Perera W.N. and Nicol M.J., (2003). Thermodynamic studies of the gold(III)/gold(I) redox system in ammonia-thiosulphate solutions at 25 C. In Hydrometallurgy 2003. Vol 1 Leaching and Solution Purification (Eds. C.A.Young, A.M.Alfantazi, C.G.Anderson, D.B.Dreisinger, B.Harris and A.James) TMS, Warrendale, pp 155-168. Sergent R. H., Dagar, A., Shin, C. C. and Ried, K. J. (1992) A comparison of bromine and cyanide for refractory gold concentrate in Precious and Rare Metal Technologies, ed A.E. Torma and H.I.H. Gundlier, Elsevier, New York, pp 149-162. Sparrow, G.J. and Woodcock, J.T. (1995) Cyanide and other lixiviant leaching systems for gold with some practical applications, Min. Proc.& Extr. Metall. Reviews, 14, 193-247. Stace, C. R. (1984) Selective passivation of sulfides, CRA Services Ltd Research - Cockle Creek. Technical Report no R84/043. Sun T. M. and Yen, W. T. (1992) Chemistry of gold dissolution in acidic hypochlorite solution, in ICHM ’92, Proc. Second Intl. Conf. on Hydrometallurgy (Eds. Chen J., Yang S. and Deng Z.), International Academic Publishers, Beijing, pp 475-480. Swaminathan C., Pyke, P. and Johnston, R. F. (1993) Reagent trends in the gold extraction industry, Minerals Engineering, 6, 1-16. Tan, S. W. and Bell, P. R. F. (1990) A study on the dissolution of gold in sodium bisulfide solutions, in Chemeca 90,University of Auckland, Auckland, pp 962-968.. Thomas K.G., Fleming C., Marchbank A. R. and Dreisinger D.B., (1998). Gold recovery from refractory carbonaceous ores by pressure oxidation, thiosulfate leaching and resin-in-pulp adsorption. U.S. Patent 5,785,736. Tran, T., Davis, A. and Song, J. (1992) Extraction of gold in halide media, in Proc. Symposium on Extractive Metallurgy of Gold and Base Metals, Eds. Misra, V., Halbe, D. and Spottiswood, D.J., Aus.I.M.M., Melbourne, pp. 323-327. Tran, T., Lee, K. and Fernando, K. (2001) Halide as an alternative lixiviant for gold processing – an update, in Cyanide: Social, Industrial and Economic Aspects, Eds. Young, C.A., Twidwell, L.G. and Anderson, C.G., TMS, Warrendale, PA, pp 501-508. Tremblay, L., Deschênes, G., Ghali, E., McMullen, J. and Lanouette, M. (1996). Gold recovery from a sulfide bearing gold ore by percolation leaching with thiourea, Intl. J. Min. Proc., 48, 225-244. Trindade, R. B. E., Rocha, P. C. P. and Barbosa, J. P. (1994) Dissolution of gold in oxidised bromide solutions, in Hydrometallurgy ’94, Chapman and Hall, London, 1994, pp 527-540. Touro, F. J. and Wiewiorowski, T. K. (1992) Recovery of gold from ores using a complex pre-treatment and leaching with sulfurous acid, US Patent 5,147,617. Ubaldini, S., Fornari, P., Massidda, R. and Abbruzzese, C. (1998) An innovative thiourea gold leaching process, Hydrometallurgy, 48, 113-124.
37 ALTA Free Library www.altamet.com.au
Van Meersbergen, M. T., Lorenzen, L. and van Deventer, J. S. J. (1993) The electrochemical dissolution of gold in bromine medium, Minerals Engineering, 6, 1067-1079. Van Staden, P. J. and Laxen, P. A. (1989) ‘In stope’ leaching with thiourea, J. South. Afr. Inst. Min. Metall., 89, 221-229. Von Michaelis, H. (1987). Non-cyanide lixiviants for gold and silver, Randol Phase III, Randol Intl., Golden, CO, USA, Vol. 4, Ch. 29, pp. 2423-2552. Wan, R.Y. (1997) Importance of solution chemistry for thiosulfate leaching of gold, in Proc. World Gold ’97, Singapore, SME., Littleton, pp 159-162. Wan, R.Y., Brierley, J.A., Acar, S. and LeVier, K.M. (2003). Using thiocyanate as lixiviant for gold recovery in acidic environment, in Hydrometallurgy 2003 Volume 1: Leaching and Solution Purification, Young, C.A., Alfantazi, A., Anderson, C., James, A., Dreisinger, D. and Harris, B. (Eds), The Minerals, Metals and Materials Society, Warrendale, PA, pp 105-121. Wan, R. Y., Le Vier, M. and Miller, J. D. (1993) Research and development activities for the recovery of gold from non-cyanide solutions in Hydrometallurgy Fundamentals, Technology and Innovations, Eds., Hiskey, J. B and Warren, G.W., SME., Littleton, pp 415-436. Wan, R.Y., Luinstra, L. and Brierley, J.A. (1995) Gold recovery from refractory sulfidic-carbonaceous ore. Part II. Thiourea leaching following biooxidation-heap pretreatment, in EPD Congress 1995, Ed. Warren, G.W., Minerals, Metals and Materials Society, Warrendale, PA., pp. 165-173. Weissberg, B. C. (1970) Solubility of gold in hydrothermal alkaline sulfide solutions. Econ. Geol., 65, 551-556. Wassink, B.; West-Sells, P.; Dreisinger, B. and Fisher, N., “Leaching of a Gold Ore Using the Hydrogen Sulfide-Bisulfide-Sulfur System,” Treatment of Gold Ores, Proceedings of the International Symposium on the Treatment of Gold Ore, Calgary, Alberta, Aug. 21-24, 2005, 229-241. West-Sells, P. and Hackl, R.P., (2005) A Novel Thiosulfate Leach Process For The Treatment Of Carbonaceous Gold Ores. Treatment of Gold Ores, Proceedings of the International Symposium on the Treatment of Gold Ore, Calgary, Alberta, Aug. 21-24, 2005, 241-255. Wardell-Johnson, M., Steiper and Dreisenger, D. (2009) Engineering Aspects of the Platsol Process, ALTA 2009 Nickel-Cobalt , Copper and Uranium Conference. Xia C., Yen W-T. and Deschenes G., (2003). Improvement in thiosulfate stability in gold leaching. Minerals and Metallurgical Processing, 20(2), 68-72. Xia C. and Yen W-T. (2008) Effect of lead ion and minerals on thiosulfate gold leaching. Hydrometallurgy 2008 Proceedings of the 6th International Symposium Honour Robert Shoemaker, Ed Young Ca, Taylor, PR, Anderson, CG and Choi, Y, SME., Littleton, pp760-768. Yang X, Michael S., Moats, M. S. and Miller J. D. (2010) Gold dissolution in acidic thiourea and thiocyanate solutions Electrochimica Acta, 55(11), p3643-3649. Yen, W. T. and Wyslouzil, D. M. (1985) Gold extraction from refractory ore by pressure oxidation and th thiourea leach, Paper presented at the 17 Canadian Mineral Processors Operators Conference, Jan 22-24, Ottawa, Ontario, 15pp Yen, WT. Pindred, RA. and Lam, M.P. (1990) Hypochlorite leaching of gold ore. in Hydrometallurgy Fundamentals, Technology and Innovations, Eds., Hiskey, J. B and Warren, G.W., SME, Littleton, pp 415-436. Yen, W. T., Qiu, D., Deschenes, G., and Jin, S. (1992) Gold elution from carbon loaded in thiourea leach solution, Nonferrous Metals, 44(2), pp 45-54.
38 ALTA Free Library www.altamet.com.au
Zhang H., Ritchie, I. M. and La Brooy S.R. (2001) Electrochemical oxidation of gold and thiourea in acidic thiourea solutions, J. Electrochem. Soc., 148, D146-D153. Zhang, J., Xinzhe, L., Nengwen, Y. and Feng, D. (1993) Leaching gold and platinum group metals by the LSSS method, in Precious Metals 1993 (Ed Mishra, R.K.) International Precious Metals Institute: Allentown, pp 281-288. Zhang S. and Nicol M.J., (2003). An electrochemical study of the dissolution of gold in thiosulfate solutions. Part I Alkaline solutions. J Applied Electrochemistry, 33, 767-775. Zhang X.M, Senanayake G. and Nicol M.J., (2004). A study of the gold colloid dissolution kinetics in oxygenated ammoniacal thiosufate solutions. Hydrometallurgy, in press. Zhang H., Nicol, M., J. and Staunton, W.P (2005) An electrochemical study of an alyterantive process for the leaching of gold in thiosulfate solutions. Treatment of gold ores, ed, Deschenes G, Hodouim, D. and Lorenzen, L, Canadian Institute of Mining Metallurgy and Petroleum, pp 243-257. Zhang H.,(2005) Improved thiosulfate process International Patent Publication WO 2005/017215 A1. Zhang H., and Jeffrey, M (2008) A study of pyrite catalysed oxidation of thiosulfate, Hydrometallurgy 2008 Proceedings of the 6th International Symposium Honour Robert Shoemaker, Ed Young Ca, Taylor, PR, Anderson, CG and Choi, Y, SME., Littleton, pp769-778.
39 ALTA Free Library www.altamet.com.au
CURRENT STATUS OF THIOSULFATE TECHNOLOGY AND FUTURE PROSPECTS FOR IT’S USE By Matthew Jeffrey
CSIRO, Parker Centre, CSIRO Minerals Down Under National Research Flagship, Australia Presented by Matthew Jeffrey Matthew.Jeffrey@csiro.au
Overview • Introduction • Review of thiosulfate leaching / chemistry • Prospects for use for: • “Conventional” ores – leach + RIP • Refractory / Cu-Au ores • In-situ leaching
• Conclusions
ALTA Free Library www.altamet.com.au
1
Introduction • Parker Centre • • • •
“Virtual” research centre CSIRO / Murdoch / Curtin / UQ as core participants Govt funding supports “core capability” development Large number of industry participants support 1:1 projects
• CSIRO • Research provider with ~6500 staff • CSIRO Minerals -> Process Science and Engineering • Minerals Down Under Flagship is branding for delivery of majority of minerals based research in CSIRO. Themes in Exploration, Mining, Processing, Carbon Steel, On-line analysis, and Sustainability
• Come and see the booth!
Thiosulfate leaching technologies • Thiosulfate leaching an alternative hydrometallurgical process to cyanidation • “Classic” copper-ammonia-thiosulfate system • Well researched since 1970s • Wide variety of conditions utilised
• Other non-ammonia systems have been developed more recently • Oxygen based systems • Ferric based systems
ALTA Free Library www.altamet.com.au
2
Cu-NH3-TS leaching • Oxidant is Cu(II) ammine complex Au + 5S 2O32− + Cu ( NH 3 ) 24+ → Au ( S 2O3 ) 32− + Cu ( S 2O3 ) 53− + 4 NH 3 0.5mM Cu
• NH3 affects stability of Cu(II)
2mM Cu
8 7 6 Leach rate
• Leach rate increases with copper and ammonia
1mM Cu
5 4 3 2 1 0 0
50
100
150
200
250
Ammonia Concentation (mM)
Cu-NH3-TS solution chemistry • Cu(II) oxidation of S2O32• Simplified reaction:
2Cu(II)+ 2S2O32- → 2Cu(I)+ S4O62• Rate increases with • ↓ Ammonia, ↑ Thiosulfate, ↑ Copper
• Oxygen oxidises Cu(I) to Cu(II)
4Cu(I) + O 2 + 2H 2 O → 4Cu(II) + 4OH • Regenerates Cu(II) in leaching • But leads to redox cycle with TS oxidation
ALTA Free Library www.altamet.com.au
3
Extraction of Gold from Ores • Numerous studies published • Widely varying conditions
• Little regard to oxygen addition • Belief that high DO good
• Summary of data sorted by date:
Researcher
S2O32-
Cu
Temp.
(M)
(mM)
(oC)
Ore type
pH / NH3
Berezowsky et al.(1978)
CuS
0.5-1*
30-60
25-50
9-10
Kerley (1983)
Cu-MnS
12-25 % *
15-60
40-60
7-9
Block-Bolten et al. (1985)
Zn-PbS
0.1-0.5*
-
21-50
0.75M NH3
Perez and Galaviz (1987)
Mn-CuS
5-15 % *
15-60
25-40
10-11
Zipperian et al. (1988)
Rhyolite, Mn
>0.1
200
50
10
Hemmati et al. (1989)
Carbonaceous
0.7
150
35
10.5
Ji and Yu (1991)
Oxidised Cu
0.1-2
8-300
30-65
0.5-5.2M NH3
Hu and Gong (1991)
Sulfide
1
16
40
2M NH3
Murthy (1991)
Pb-ZnS
0.1-0.5*
-
21-70
6.9-8.5
Cao et al. (1992)
Sulfide conc
0.2-0.3*
47
60
10-10.5
Langhans et al. (1992)
Oxidise Cu
0.2
1
room
11
Huge range of conditions for earlier work
ALTA Free Library www.altamet.com.au
4
Researcher
S2O32-
Cu
Temp.
(M)
(mM)
(oC)
Ore type
pH / NH3
Gong et al. (1993)
Sulfide conc
0.7*
60
50
1M NH3
Wan et al. (1994)
Carbonaceous sulfidic
0.1-0.2*
0.5
room
9-10
Abbruzzese et al. (1995)
Gold ore
2
100
25
8.5-10.5
Groudev et al. (1995)
Oxide
0.1*
8
room
8.5-9
Groudev et al. (1996)
Fe-Cu-Pb-ZnS
0.14*
8
room
9-10
Marchbank et al.(1996)
Sulfide-carbonaceous
0.025-0.1*
0.8-8
25-80
7-8.7
Yen et al. (1996)
Au-Cu ore
0.4
30
room
11
Wan (1997)
Sulfide-Carbonaceous
0.1*
0.5
room
9
Yen et al. (1998)
Au-Cu ore
0.5*
100
room
10
Thomas et al. (1998)
Oxidised sulfide
0.03-0.05*
0.1-2
80-85
7-9
Fleming (2000)
Sulfide-Carbonaceous
0.05
0.5
40-60
7.5-8
Low pH / low ammonia works for some ores
Researcher
S2O32-
Cu
Temp.
(M)
(mM)
(oC)
Ore Type
pH / NH3
Balaz et al. (2000)
Sulfide Conc.
0.5*
62
70
6-7
Aylmore (2001)
Au-CuS Conc.
0.1-0.8
12-62
room
10.2
Schmitz et al. (2001)
Sulfide-Carbonaceous
0.4
16
room
1.9M NH3
van Z. de Jong et al. (2001)
Various
0.25-0.5*
6-12
25
1-2M NH3
Feng and van Deventer (2002)
Oxide, sulfide ore; pyrite, cu conc.
0.1*
3.1
20
0.5M NH3
Navarro et al. (2002)
Au Conc.
0.3*
50
25
10
Molleman and Dreisinger (2002)
Cu-Au
0.2*
15.7
35
10
Ficeriova et al. (2002)
Cu-Pb-Zn Sulfide Conc.
0.5*
157
70
-
Aylmore (2004)
Oxide
0.075*
1.5
room
9.5 / 0.4M NH4+
Bhakta (2003)
Carbonaceous sulfidic
0.07-0.09 *
0.5-1
-
8.8-9.2
West-Sells and Hackl (2005)
Carbonaceous
0.1*
0.8
room
9
Approaching an â&#x20AC;&#x153;optimalâ&#x20AC;? system??
ALTA Free Library www.altamet.com.au
5
S2O32-
Cu
Temp.
(M)
(mM)
(oC)
Quartz and Feldspar
0.125 – 1
0 – 100
room
0.1–0.8M NH3
Ficeriova et al. (2005)
Sulfide Conc.
0.5*
157
70
6
Meng (2005b, 2005a)
Arsenosulfide Conc.
0.05 – 0.1
10 – 20
room
0.2-0.25M NH3
Feng and van Deventer (2007b)
Pyrite conc. & Sulfide ore
0.1 - 0.2*
6 - 12
room
0.8-1M NH3
Feng and van Deventer (2007a)
Sulfide ore
0.165*
0.8
room
0.5M NH3
Arslan and Sayiner (2008)
Quartz and Feldspar
0.3 – 1.2
0 – 50
20 – 60
0-4M NH3
Researcher
Ore type
Tanriverdi et al. (2005)
pH / NH3
• “Typical” conditions for most ores • 50 – 100 mM ATS • 1 – 5 mM Cu (60 – 300 mg/L) • pH 9 – 10 (or at least 50 mM free ammonia)
Oxygen thiosulfate system • Analogous to cyanide
4 Au + 8S 2 O32− + O2 + 2 H 2 O → 4 Au ( S 2 O3 ) 32− + 4OH − • Potential benefits • Cu and NH3 not a necessity • Low rate of thiosulfate oxidation • Can utilise ATS, sodium thiosulfate (STS), or calcium thiosulfate (CaTS)
• Placer process • 60 to 80 °C, 150 to 1500 kPa Oxygen • Gold extracted in 6 hours
• Issues • Usually a very slow rate of leaching • Copper as catalyst (but doesn’t act as oxidant)
ALTA Free Library www.altamet.com.au
6
Iron Systems • Fe(III) medium strength oxidant • E0 = 0.77V • Uncomplexed form rapidly oxidises thiosulfate
• Can be stabilised by ligands • Oxalate: log β Fe(C2O4)33– = 18.6 • EDTA: log β Fe(EDTA) = 25.1
• Minimises thiosulfate reaction Early testwork for ligands: • Oxalate • Leach kinetics faster than Cu-NH3 • Works with only 15 mM thiosulfate • Homogeneous thiosulfate loss negligible
• EDTA • >90% recovery from simple ores
Issues with Fe System • Largely untested • Require thiourea (or alternative) catalyst • To obtain reasonable leach rates
• Presence of sulfides problematic • Catalyse thiosulfate oxidation by oxygen, which occurs more readily in the neutral solutions
ALTA Free Library www.altamet.com.au
7
Potential applications of thiosulfate • Legislative issues • Use of cyanidation limited in a number or regions (most recently the EU) • Heap or tank leach • E.g. Thiosulfate leaching being tested in French Guyana
• Carbonaceous ores • Gold cyanide complex is “preg-robbed” • Traditional tank or heap leach, with RIP (+electrowinning) or CCD (+ either cementation or precipitation • Double refractory ores? POX / TS tank leach vs. BiOX / TS heap. Others, e.g. Galvanox (esp when S generated?)
• Copper/gold ores (high CN consumption due to nuisance copper concentrations) • Copper needed for TS leaching process
• New applications, e.g. In-situ leaching
Tank leaching • Cu-ATS leaching operated using continuous pilot plants by a number of companies / researchers • Gold recovery a key issue • CCD (MATS process) operated by Placer
• Resin in pulp • Resins strongly adsorb gold • But polythionates compete
-1
10000
Gold on resin / mg kg
• Need to ensure total stability of gold complex through circuit (including washing). • Usually have some gold loss
1000
100
2.35 mM polythionates 3.7 mM polythioantes 7.7 mM polythionates
10 0.1
1
Gold in solution / mg L
ALTA Free Library www.altamet.com.au
10 -1
8
Sulfite enhanced elution of gold • Elution of gold from resins problematic • Recent development at CSIRO: Sulfite enhanced elution • Makes RIP more attractive, regeneration of resin not required • No large pH swings 1200 2M NaCl 2M NaCl + sulfite
Au Concentration (mg L -1)
1000
1600 1 M nitrate
Au Concentration (mg L -1)
1400
1M nitrate + sulfite
1200 1000 800
800
600
400
200
600 0
400
0
5
10 15 Bed Volum es
20
25
200 0 0
5
10
15
20
25
Bed Volum es
Closed circuit adsorption / elution demonstrated 7 adsorption tanks tails
Loaded resin 5BV ATS Pre-eluent
6.75 BV
2 BV water NaCl
8 BV eluent 2 BV SO32wash
E1
EW
2.25 BV purge
E2 8 BV
Barren resin Recycle for next elution
ALTA Free Library www.altamet.com.au
9
Heap leaching • Demonstrated at pilot scale by Newmont (Cu-ATS system) • Coupled with BiOX for refractory ores, and copper cementation for gold recovery • Maintaining solution chemistry a key aspect
• Sulfide ppt an alternative technology for recovery • Also regenerates thiosulfate
• More recent work with resin in column also shows potential
In-situ / in-place leaching • Used since the mid 1970’s • Uranium, copper, soluble salts (halite, trona, boron), potash from phosphate rock
• Not been adopted for the recovery of gold • Number of deposits with favourable characteristics for true in-situ application • e.g. Victorian Deep Leads
• Alternative to cyanide required • Environmental concerns • Availability of oxidant
CRA studies in Vic
• Potential for thiosulfate based system • But Cu-ATS system unlikely to work in anaerobic systems
ALTA Free Library www.altamet.com.au
10
Application: Anaerobic column leaching • Ferric – EDTA leach systems • Sealed, column flooded with up-flow
Systems tested include: • Synthetic Columns • Sand + Au • Sand + 1% pyrite + Au
• Ancient (buried) paleochannel Au river gravel from St. Ives • Oxide gold ore
180
8
160
7
140
6
120
5
100
Au
4 2
80 tot Au 60 Fe 40
1
20
3
0
[Fe] / ppm
9
0 0
10
20
30
Time / day
Synthetic sample without pyrite 1
180 160
0.8
Au
140
0.6
tot Au Fe
120 100 80
0.4
60
[Fe] / ppm
• Column leach testwork on paleochannel material failed • Testwork on synthetic samples revealed problems with sodium thiosulfate system • Pyrite catalysed oxidation of thiosulfate by FeEDTA
Au / ppm or mg
• Technically possible – anaerobic thiosulfate leaching of gold demonstrated in lab • Why oxide gold?
Au / ppm or mg
In-place leaching for oxide gold - rationale
40
0.2
20 0
0 0
10
20 Time / day
30
40
Synthetic sample with 1% pyrite
ALTA Free Library www.altamet.com.au
11
Samples selected for testing 0.1M TS+10mM tu+3mM FeEDTA
• Site and ore type selection • Bottle roll results
100
Au%
80 60
Tunkillia Moolart Well Glendower Cornishman
40 20 0 0
20
40 time / h
60
80
Cyanide leaches 100
Au%
80 60 Tunkillia Moolart Well Glendower
40 20 0 0
20
40 time / h
60
80
Column thiosulfate leaching 120
100
[ Fe ] / ppm
a 50
80
0 40
▲ [Fe] ◊ EH
0
-100 0
5
10
120
20 30 Time / day
40
10
40
1 0
0 10
20
30 Time / day
40
b Au / ppm or mg
2-
2
80
2-
trithionate tetrathionate pentathionate hexathionate thiosulfate
3
[S 2O3 ] / mM
4 [S xO6 ] / mM
-50
EH / mV (Ag/AgCl)
• Anaerobic column leaching of drill chips from Tunkillia • Stable chemical system • Good gold leaching
[Au]
8
tot Au 6 4 2 0 0
10
20
30
40
Time / day
ALTA Free Library www.altamet.com.au
12
Future research program: in-situ leaching • Hydrometallurgy • Column leach testwork on progressively coarser samples • Leaching with drill core (ca. Brent Hiskey Cu ISL)
• Ore characterisation • QEMSCAN (gold search) • Microtomography for predictive permeability models
• Reactive transport and hydrology • Air permeability tests • Modelling
• Permeability enhancement: • Financial analysis tool
Conclusions • In past decade, significant advances in understanding and improving thiosulfate processes • Prospects for use. Shorter term? • • • •
Cyanide will be used if it can, it is just too effective! Technology available for situations where cyanide cant be used But, not “off the shelf” Risk vs reward?
• Prospects for use. Longer term? • Technology for in-situ leaching under development • High risk / High reward
Acknowledgements • Research team (D Hewitt, H Zhang, X Dai, P Roberts + others)
ALTA Free Library www.altamet.com.au
13
ALBION PROCESS FOR TREATMENT OF REFRACTORY ORES By Duncan W. Turner PhD General Manager - Albion Process Core Resources Pty Ltd 44 Corunna Street, Albion, QLD 4010, Australia Ph +61 7 3637 8105 Fax +61 7 3637 8199 Email dturner@coreresources.com.au Mike Hourn Manager Hydrometallurgy Xstrata Technology Level 4, 307 Queen St, Brisbane, QLD 4001 Ph +61 7 3833 8539 Fax +61 7 3833 8555 Email mhourn@xstratatech.com.au Website www.albionprocess.com Presented by Duncan Turner
CONTENTS
1. 2.
INTRODUCTION ............................................................................................................. 2 ISAMILL TECHNOLOGY .................................................................................................3
3. 4.
ALBION PROCESS OXIDATIVE LEACH.........................................................................7 CAPITAL AND OPERATING COSTS ............................................................................ 10
5.
PROJECT UPDATES .................................................................................................... 12
6. 7.
PROJECT DEVELOPMENT PROGRAM ....................................................................... 13 CONCLUSIONS ............................................................................................................ 14
8.
REFERENCES .............................................................................................................. 14
ALTA Free Library www.altamet.com.au
1. INTRODUCTION
1.1 OVER VIEW OF THE ALBION PROCESS The Albion Process was developed in 1994 by Xstrata PLC and is a combination of ultrafine grinding and oxidative leaching of sulphide concentrates at atmospheric pressure. The process is [1,2] patented worldwide . The Albion Process uses ultrafine grinding in stirred mills to produce a highly activated, finely ground concentrate at low energy inputs. The finely ground concentrate is then leached with oxygen at atmospheric pressure in conventional agitated tanks. The use of the atmospheric pressure leach offers considerable capital cost savings over alternative technologies, such as pressure or bacterial leaching. The key to the Albion Process is the ultrafine grinding stage. The process of ultrafine grinding results in a high degree of strain being introduced into the mineral lattice. As a result, the number of grain boundary fractures and lattice defects in the minerals increase by several orders of magnitude, relative to un-ground minerals, which lowers the activation energy for the oxidation of the sulphides, facilitating leaching. The rate of leaching is also enhanced, due to the increase in the mineral surface area. The most common barrier to leaching of mineral sulphides is passivation of the leaching surface by precipitates â&#x20AC;&#x201C; most commonly elemental sulphur. Passivation occurs when precipitates form on the surface of a leaching mineral as the mineral dissolves. These precipitates passivate the mineral, by preventing the access of chemicals to the mineral surface. Passivation is normally complete once this precipitated layer is 2 â&#x20AC;&#x201C; 3 microns thick. Ultrafine grinding of a mineral to a particle size of 80% passing 8 â&#x20AC;&#x201C; 12 microns will prevent passivation, as the leached mineral will disintegrate prior to the precipitate layer becoming thick enough to passivate the mineral. The temperatures and operating pH employed in the Albion Process leach also results in low density precipitates, which are slow to passivate. This is illustrated in the following diagram.
Figure 1: Passivation of Leaching Minerals
Unleached 9 micron particle
Leached Particle with over 96 % volume dissolution
After the mineral has been finely ground, the oxidative leaching stage is then carried out in conventional agitated tanks operating at close to the boiling point of the slurry. Oxygen is introduced to the leach slurry for oxidation.
2 ALTA Free Library www.altamet.com.au
Leaching is carried out autothermally, in that the temperature of the leach slurry is set by the amount of heat released by the leaching reaction. Heat is not added to the leaching vessel from external sources. Temperature is controlled by the rate of addition of oxygen, and by the leach slurry density, with excess heat removed by direct evaporative cooling. In general, mineral leaching in the Albion Process will occur via ferric iron intermediates, in two steps. In the first step, the mineral sulphide is oxidised to a soluble sulphate and elemental sulphur. In the second step, the elemental sulphur is then oxidised to form sulphuric acid: Step 1 - MS + Fe2(SO4)3 = MSO4 + 2FeSO4 + S
o
o
Step 2 - S + 3Fe 2(SO4)3 = 4H2 SO4 + 6FeSO4 The Albion Process leach conditions can be controlled to vary the relative extents of steps 1 and 2, so that the leaching stage can be operated without the addition of acid from external sources. Alternatively, elemental sulphur formation can be avoided entirely if the leach residue is to be treated for precious metal recovery.
2. ISAMILL TECHNOLOGY
2.1 ULTRAFINE GRINDING In most grinding mills, mineral breakage occurs as a result of the impact of grinding media on the mineral to be ground. These impacts impart tensile stresses to the mineral, and when the resulting shear strain is sufficient, the mineral will break. This process is known as fracture. Grinding to fine sizes then requires fine media. Conventional milling occurs in tumbling mills, such as a ball mill, which are suitable for use with media sized at 25 mm or above. The rotating action of the mill lifts the media charge to a point where gravitational effects cause the media to fall back off the side of the mill, to the base of the mill. At media sizes below 25 mm, however, the media charge will tend to centrifuge, as the centrifugal forces acting on the media exceed the gravitational forces. When this occurs, little breakage occurs and the mill will no longer grind efficiently. Ultrafine grinding, therefore, requires a different type of milling action than found in a conventional ball mill. In most ultrafine grinding mills a stirrer is used to impart momentum to the media charge. Media is agitated through stirring, and the resulting turbulent mixing overcomes the tendency of fine media to centrifuge. Abrasion is the major breakage mechanism in a stirred mill. The common aspects of a stirred mill are a central shaft and a series of stirrers attached to the shaft. These stirrers can range from pins to spirals and discs. Furthermore two configurations are common in stirred mills. In the first, the mill shaft and grinding elements are set up vertically within the mill (shown in the Vertimill schematic below). This type of configuration is limited in size to typically 750 kW of installed power or less. This limitation is brought about by the large break out torque imposed on the impeller located at the base of the media charge, due to the compressive load of media sitting vertically on the impeller. In the second configuration, the mill shaft is aligned horizontally within the mill chamber (shown in the IsaMill schematic below). This configuration is the most cost efficient at motor sizes in excess of 500 kW, as there is very little break out torque required to begin to agitate the media charge. This limits the motor size to that required for grinding only.
3 ALTA Free Library www.altamet.com.au
Figure 2: Stirred Mill Technologies
(a)
Vertical Mill (Vertimill)
(b)
Horizontal Mill (IsaMill)
The IsaMill, marketed by Xstrata Technology, is the most advanced ultrafine grinding technology available, with over 80 mills installed in 16 countries with in excess of 130 MW installed power. The IsaMill circuit consists of the IsaMill, gearbox and drive as well as feed and product pumps, pump boxes and the media addition system. Industrial IsaMills are available in a range of sizes from 350 kW up to 3.5 MW. The media size for the IsaMill is within the range 1 â&#x20AC;&#x201C; 3 mm. Media can come from a range of sources, such as an autogenous media screened from the feed ore, to silica sands and ceramic beads. The ability to agitate such a fine media allows the IsaMill to grind to particle sizes as fine as 80 % passing 5 microns at very low specific energy inputs compared to Tower mills or Ball mills. The IsaMill also utilises nonferrous media, and produces a ground slurry that is not impacted by the adverse ORP effects commonly seen with steel grinding media. The IsaMill contains up to eight discs on the shaft, with each disc acting as a separate grinding element. The slurry residence time distribution through the mill therefore approaches perfect plug flow with virtually no short circuiting. This allows the IsaMill to be operated in open circuit without the need for cyclones. The resulting particle size distribution is the narrowest of any commercially available fine grinding mill, resulting in very high leach recoveries, with no coarse minerals that would otherwise travel through the leach train without complete oxidation. An interesting feature of the IsaMill is its patented internal classification system ('Product Separator') that retains media and coarser unground particles in the mill, while allowing ground product to exit. The Product Separator is a centrifugal separation device that generates very high Gforces, resulting is a sharp classification zone. Coarser particles entering this classification zone are centrifuged toward the mill shell and then backwards toward the mill feed port, returning to the grinding zone for further size reduction. The Product Separator is more efficient than a conventional hydrocyclone due to the very high G-forces generated by the high tip speed of the disc. The Product Separator eliminates the need for fine screens from the grinding process that can block and wear quickly. It also allows open circuit operation of the IsaMill at high slurry densities which may be important to the process water balance.
4 ALTA Free Library www.altamet.com.au
2.1.1 ULTRAFINE GRINDING AND MECHANICAL ACTIVATION Ultrafine grinding of a mineral will result in a loss of crystallinity of the mineral and increases the amorphous character of the mineral. During milling, the minerals are impacted by the grinding media, a process that will cause plastic deformation of the mineral grain, leading to breakage and an increase in grain boundary and lattice defects in the mineral structure. Mechanical activation is relatively well understood for the production of alloy metal powders, but is not well understood when applied to minerals. However key contributors to improved mechanical activation are believed to be mill pressure and mill energy intensity. The milling intensity in an IsaMill is of the order of 300 3 kW/m , which is the highest energy intensity available in commercial grinding mills. The process of milling results in the breaking of bonds in the crystalline lattice of the mineral and thereby brings about a decrease in the activation energy. The decrease in activation energy after high intensity milling, ∆E, is equal to the difference between the activation energy of the original highly crystalline mineral, E, and the activation energy of the finely ground mineral, E*. The reduction in the activation energy will lead to an increase in the rate of mineral leaching, according to the Arrhenius relationship. 2.2 TYPICAL ENERGY REQUIREMNTS IN THE ISAMILL Grinding power requirements for a selection of refractory gold concentrates are presented in Figure 3 below. For an Albion Process application, the target grind size is 80 % passing 10 – 12 microns, and a specific energy consumption typically in the range 25 – 90 kWh/tonne is required. Figure 3: Refractory Gold IsaMill Fine Grind Window
110
Specific E nergy (kWh/t)
90
High - 90 kWh/tonne
70
Mean - 60 kWh/tonne 50
30 Low - 25 kWh/tonne
10 1
Size (um)
10
100
2.3 ISAMILL CASE STUDIES IN REFRACTORY GOLD 2.1.1 KCGM (AUSTRALIA) The KCGM operation located in Western Australia run by Newmont Mining Corp and treats 12 tonnes per hour pyrite concentrate with IsaMill technology. At this operation an M3000 unit grinds concentrate to a P 80 of 10 to 12 micron to liberate gold ahead of CIL. The following photograph shows the IsaMill unit installed at KCGM.
5 ALTA Free Library www.altamet.com.au
Photograph 1: KCGM IsaMill Installation
2.1.2 MACRAES (NEW ZEALAND) The Macraes operation located on the South Island of New Zeland run by Oceana Gold Limited treats 12 tonnes per hour pyrite concentrate with IsaMill technology. At this operation an M1000 unit grinds pyrite concentrate to a P80 of 15 micron ahead of feeding to pressure oxidation. The following photograph shows the IsaMill unit installed at Macraes. Photograph 2: Macraes IsaMill Installation
2.1.3 KUMTOR (KYRGYZSTAN) The Kumtor operation located in Kyrgyzstan run by Centerra treats up to 35 tonnes per hour of pyrite concentrate with IsaMill technology. At this operation an M10000 unit grinds concentrate to a P 80 of 12 micron to liberate gold ahead of CIL. The following photograph shows the IsaMill unit installed at Kumtor.
6 ALTA Free Library www.altamet.com.au
Photograph 3: Kumtor IsaMill Installation
3. ALBION PROCESS OXIDATIVE LEACH
3.1 ALBION PROCESS CHEMISTRY In the Albion Process, finely ground minerals are leached under atmospheric conditions in a specially designed Albion Process Leach Reactor. The Albion Process Leach Reactor consists of a leaching tank with a centrally mounted impeller, typically fitted with dual hydrofoil agitators, a ring of supersonic oxygen spargers at the base of the tank and baffles to ensure adequate mixing. The combination of agitation power and sparger pressure is the key to achieving low cost and efficient oxygen mass transfer in the reactor. Xstrata Technology has extensive experience in the design of these systems. A schematic diagram of a typical Albion Process Reactor and Xstrata Technologyâ&#x20AC;&#x2122;s supersonic oxygen sparger are shown below. Figure 4: Key Albion Process Technology Equipment
(a)
Xstrata Technology Albion Process Leac h Reactor
(b)
Xstrata Technology Oxygen Sparger
No internal temperature control is required for the Albion Process leach tanks. The tanks are allowed to operate autothermally, with no internal cooling coils or flash stages required. This significantly simplifies the construction of the leach circuit.
7 ALTA Free Library www.altamet.com.au
The temperature of the leach operates at or close to the boiling point of the slurry and typically lies o in the region 80 â&#x20AC;&#x201C; 105 C depending on the precise process conditions and elevation of the plant. As mentioned previously no dedicated cooling is required for the leach, with temperature controlled by simple humidification of leach tank off gasses. The principle gold bearing mineral in most refractory gold concentrates is pyrite. During the oxidative leaching of pyrite, the sulphide matrix is broken down to release ferric ions and sulphuric acid, according to the following reaction: FeS2 + 15/4O 2 + 1/2H 2O = 1/2Fe2 (SO4)3 + 1/2H2 SO 4 Oxidation of the pyrite matrix liberates precious metals for recovery by cyanidation. The precious metals remain in the oxidised residue. In conventional refractory gold treatment circuits, the oxidised residue is separated from the acidic leach solution prior to cyanidation. This results in a high capital cost solid/liquid separation circuit. The acidic liquor is then neutralised in a dedicated neutralisation circuit, involving additional capital cost. However in the Albion Process leach, the liberated ferric and acid are neutralised in-situ by the continual addition of limestone slurry. The neutralised iron oxides and gypsum are then sent to the CIL circuit along with the liberated gold and silver, resulting in a single tailings stream and a single tailings impoundment. The differences in these respective flowsheets are shown in the following figure.
Figure 5: Comparative Sulphide Oxidation Flowsheets
(a)
POX/BIOX Leaching Flowsheet
(b)
Albion Process Flowsheet
In-situ neutralisation in the Albion Process leach eliminates the need for dedicated solid/liquid separation circuits and neutralisation plants , resulting in substantial capital cost savings. Limestone is dosed into all Albion Process Leach Reactors off a central ring main, to maintain the leach pH within the range 5 â&#x20AC;&#x201C; 5.5. Under Albion Process leach conditions, goethite will be the major iron precipitate formed on neutralisation with limestone and the overall pyrite leach reaction becomes: FeS2 + 15/4O 2 + 9/2H 2O + 2CaCO 3 = FeO.OH + 2CaSO4 .2H 2O + CO2
8 ALTA Free Library www.altamet.com.au
Operation of the oxidative leach under conditions of neutral pH is only possible due to the finely ground nature of the concentrate. Without fine grinding, the iron oxides and gypsum formed during neutralisation of the leach solution would passivate the sulphide, and prevent further leaching. Furthermore the neutral operating pH in the Albion Process leach results in very low background salt levels in solution. For example, calcium levels in solution will typically be 600 â&#x20AC;&#x201C; 620 mg/L, compared to a saturated level under leach operating conditions of 600 â&#x20AC;&#x201C; 605 mg/L. The degree of calcium super-saturation is typically 5 % or less, and crystal growth will be favoured. That coupled with the high relative surface area of gypsum crystals in the leach will result in the majority of the calcium precipitating within the slurry as gypsum, with minimal scale formation. This is in contrast to acid neutralisation systems, where the super-saturation level for calcium can exceed 20 %, due to the alkali addition rates employed, with nucleation and scale formation strongly favoured. The net result is that Albion Process conditions tend to miminze formation of scale, simplifying operation and reducing maintenance requirements. A further advantage of low dissolved salt levels is enhanced oxygen solubility. This results in high oxygen mass transfer rates in the leach tank, reducing the installed power requirements in the leach over other systems. Another feature of the Albion Process is that the oxidised residues have low cyanide consumption. In traditional acidic leach residues cyanide losses are significant, with the formation of thiocyanate and ferri/ferro cyanides phases. The neutral operating pH in the Albion Process leach ensures complete oxidation of sulphide to sulphate. No elemental sulphur is formed in the leach, and so no thiocyanate is formed in the CIL circuit. Goethite and hematite are the only precipitates formed in the oxidative leach and both iron precipitates have extremely low solubility in cyanide relative to sulphated iron precipitates. Due to the alkaline nature of the Albion Process leach slurry, low cost materials of construction can be used for the leach circuit. Both 304 and 316 stainless steel are suitable for the leach tanks, along with low cost duplex steels, such as LDX 2101. These alloy grades typically cost 50 % or less of the more expensive 2205 and 2507 duplex steels commonly employed for acidic leach circuits. Atmospheric oxidation of sulphide concentrates is not a new technology, and has been practiced for over 20 years. Some examples of the atmospheric sulphide oxidation plants, employing purely oxygen and sulphate leaching conditions, in the Western world and China are shown in Table 1. Table 1: Atmospheric Oxidative Sulphide Leaching Plants Company
Plant
Capacity
Statis
New Boliden New Boliden New Boliden Korea Zinc Birla/Western Metals Cobre Las Cruces Zhuzhou Smelter Corp
Kokkola 1 Kokkola 2 Odda Onsan Mt Gordon Las Cruces Shandong
50,000 tpa zinc 50,000 tpa zinc 50,000 tpa zinc 225,000 tpa zinc 45,000 tpa copper 40,000 tpa copper 180,000 tpa zinc
Operational Operational Operational Operational Decommissioned Commissioning Operational
9 ALTA Free Library www.altamet.com.au
3.2 TYPICAL ALBION PROCESS LEACHING RESULTS Sulphide sulphur oxidation versus gold recovery data for a selection of refractory gold concentrates are presented in Figure 6 below. This plot illustrates that sulphide oxidation requirements can vary from relatively low (~30%) to high (~90%) levels of oxidation as the gold mineralogy moves from semi refractory to refractory in nature. Based on bench and pilot scale testwork, refractory gold concentrates typically require sulphide oxidation in the range of 50% to 85% to achieve greater than 90% gold recovery under CIL conditions.
Figure 6: Typical Albion Process Sulphide Oxidation versus Gold Recovery Curves Low Oxidation
100
High Oxidation
90 80
Gold Recovery (%)
70 60 50 40 30 20 10 0 0
10
20
30
40
50
60
70
80
90
100
Sulphide Oxidation (%)
4. CAPITAL AND OPERATING COSTS
4.1 CASE STUDY Order of magnitude capital and operating costs have been estimated (based on recent plants that have progressed to construction), for a Neutral Albion Process plant treating 600 tonnes per day of a generic pyrite concentrate grading 20 g/t gold and 40%w/w sulphide sulphur. Applying testwork data from a sample having similar gold and sulphide levels, a sulphide oxidation requirement of 70% have been arbitarily selected to achieve 90% gold recovery in CIL assuming a IsaMill power requirement of 60 kWh/t. Based on the above, the Albion Process plant would oxidise 61,320 tonnes per year of sulphide sulphur in order for the project to produce 126,752 ounces per year of gold. Assuming project input costs for power, limestone, oxygen and labour (weighted average) of US¢10/kWh, US$50/t, US$45/t and US$50,000/person/year respectively, the Albion Process operating costs have been calculated at around US$19.7 million per year equating to approximately US$320 per tonne of sulphide sulphur oxidized. A breakdown of the main operating categories is shown in the following figure.
10 ALTA Free Library www.altamet.com.au
Figure 7: Albion Process Operating Cost Breakdown Maintenance supplies & consumables Power 3% (IsaMill) 7%
Power (Rest of plant) 2%
Labour 6%
Oxygen 32%
Limestone 50%
Capital investment requirements for the Albion Process (battery limit: concentrate feed to IsaMill and cooled slurry discharge, excluding limestone milling handling and storage) and Oxygen plants have been estimated at US$38.2M and US$24.5M making a total of US$62.7M. Converting these capital costs to an operating cost using a discount rate of 12.5% over 10 years, the amortised capital costs equate to approximately US$185 per tonne of sulphide sulphur oxidized. Combining the direct operating costs and the amortised capital costs the overall sulphide sulphur oxidation costs using Albion Process have been estimated at US$505 per tonne of sulphide sulphur oxidized. This is equates to a figure of US$245 per ounce of gold recovered for this particularly case study. 4.2 COMPARATIVE CAPITAL COSTS The capital cost benefits of the neutral operating pH in the Albion Process leach relative to acidic processes such as pressure oxidation or bacterial leaching are pronounced. The figure below shows the relative capital cost of an Albion Process plant compared to a pressure or bacterial leach [3] plant, using data published by Aker Solutions . The installed capital cost of an Albion Process oxidative leach plant is approximately 30 % that of a bacterial or pressure leach plant. Figure 8: Relative Capital Costs
The layout of an Albion Process leach circuit is simple, and closely mirrors a conventional CIL plant. This ensures simple maintenance and allows for commonality of spares and design.
11 ALTA Free Library www.altamet.com.au
A conceptual layout is presented below, with the IsaMill circuit shown to the left of the diagram, followed by slurry storage tanks and five Albion Process Leach Reactors. Slurry cooling towers are shown in the forefront, to cool the hot slurry prior to CIL. In some cases slurry cooling may not be required depending on the overall gold recovery flowsheet.
Figure 9: Typical Albion Process Plant Layout Leaching Train Slurry In
IsaMill
Slurry Out
The cooled slurry is suitable for direct feed to the CIL/CIP circuit, and no CCD or additional neutralisation circuit is required.
5. PROJECT UPDATES
5.1 LAS LAGUNAS PROJECT The Las Lagunas Project is located in the Dominican Republic and is being developed by EnviroGold Limited. The process plant incorporates a flotation, Albion Process and CIL flowsheet designed to treat 800,000 tonnes per year of tailings grading 3.8 g/t Au and 38.6 g/t Ag producing in excess of 80,000 ounces per year of gold plus silver. Detailed engineering has been completed by Lycopodium Minerals and Xstrata Technology both based in Brisbane, Australia. Commissiong of the plant expected in early 2011. Stainless steel Albion Process tanks have been fabricated in Taiwan ready to be shipped to site in early May 2010 to be installed on foundations already complete, as shown in Figure 10. Figure 10: Albion Process Leach Tank Foundations
12 ALTA Free Library www.altamet.com.au
5.2 CERTEJ PROJECT The Certej Project is located in the Romania and is being developed by European Goldfields Limited. The process plant incorporates a flotation, Albion Process and CIL flowsheet designed to treat nominally 3,000,000 tonnes per year of ore grading 1.6 g/t Au and 11.5 g/t Ag producing in excess of 160,000 and 800,000 ounces per year of gold and silver respectively. Basic engineering has been completed by Aker Solutions (Stockton, UK) and Xstrata Technology with the project moving into detailed design now that all permits are in place. The project is already financed and commissioning of the plant expected in late 2011/early 2012. 5.3 OTHER PROJECTS Several projects are entering or moving towards pilot programs in Q2 to Q4 of 2010 including the Joanna Project being developed by Aurizon Mines Limited which is an arsenopyrite refractory gold project expected to generate around 110,000 ounces per year of gold. Other pilot programs planned include refractory pyrite, copper and cobalt projects. Numerous testwork programs are underway supporting process studies including enargite Cu/Au/Ag applications (projects in South America, South East Asia), pyrite gold applications (projects in Queensland, Armenia, Mexico, Kazakhstan) and arsenopyrite applications (projects in Western Australia, New Zealand). Internally within Xstrata PLC projects in advanced states of testing and engineering through to commercialisation include copper/cobalt, nickel and zinc applications. A full scale Albion Process plant is due to be commissioned on fine ground material from an existing Xstrata PLC operation in July/August 2010, thus demonstrating the Albion Process at commercial scale.
6. PROJECT DEVELOPMENT PROGRAM The recommended steps involved to advance a project through Albion Process testing and engineering are summarized below. 1. Define and collect preliminary key projects and metallurgical data about the likely feed and range of feed material to be treated via the Albion Process - Complete questionnaire 2. Undertake a high level economic assessment of the process (indicative order of magnitude costing exercise) 3.
Phase 1 testing - Initiate an orientation testwork program to broadly demonstrate the technical feasibility of the process and generate key process design parameters to enable a preliminary process feasibility study to be completed
4. Commence a preliminary process study to establish the capital and operating costs for the process specific for project and target feed material 5.
Phase 2 testing - Optimize process conditions and undertake testwork on variability samples
6. Phase 3 testing - Continuous pilot trial work 7. Commence bankable feasibility study (BFS) design and costing work leading into basic and detailed engineering and finally plant construction and commissioning The approximate costs and timing for these steps are shown in Table 2.
13 ALTA Free Library www.altamet.com.au
Table 2: Approximate Albion Process Program Costs and Duration Step No. 1 2 3 4 5 6 7
Dscription
Cost
Duration
Data collection Highlevel OOM Capex/Opex Phase 1- testing Process Study Phase 2- testing Phase 3- testing Feasibility study etc…
~A$30-40K ~A$20K ~A$50-100K ~A$500-750K ?
2-4 weeks 2-4 weeks 6-8 weeks 4-6 weeks 8-12 weeks 3-6 months ?
Sample
10-20 kg 20-50 kg >250 kg
7. CONCLUSIONS In conclusion the Albion Process is a cost effective technology for treating refractory gold material with capital cost savings of move than half that of alternative processes. The technology is robust and simple to operate with IsaMill fine grinding and atmospheric oxidative leaching well proven technologies that have been demonstrated at commercial scale. 8. REFERENCES [1] Hourn et al, “Atmospheric mineral leaching”, Australian Patent No. 700850, 1996 [2] Hourn et al, “Method for treating precious metal minerals”, Australian Patent No. 744356, 1999 [3] Bartsch et al, “Benefits of Using the Albion Process for a North Queensland Project, and a Case Study of Capital and Operating Cost Benefits Versus Bacterial Oxidation and Pressure Oxidation Randol Gold Conference, Perth, 2005
14 ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
CARBON ADSORPTION & IX
ALTA Free Library www.altamet.com.au
REVIEW OF APPLICATIONS OF SUPERLIG® MOLECULAR RECOGNITIONTECHNOLOGY PRODUCTS FOR THE GOLD INDUSTRY
By
Neil E. Izatt, Steven R. Izatt, Ronald L. Bruening and John B. Dale IBC Advanced Technologies, Inc, USA
Presented by
Neil E. Izatt nizatt@ibcmrt.com
CONTENTS
1. INTRODUCTION
2
2. STATUS OF THE GOLD MARKET
2
3. COMPETITIVE TECHNOLOGIES FOR GOLD EXTRACTION AND RECOVERY FROM SOLUTION
2
4. MOLECULAR RECOGNITION TECHNOLOGY
3
5. KEY ADVANTAGES OFFERED TO THE GOLD INDUSTRY BY THE MOLECULAR RECOGNITION TECHNOLOGY PROCESS
4
6. POTENTIAL APPLICATIONS TO THE GOLD INDUSTRY FOR THE MOLECULAR RECOGNITION TECHNOLOGY PROCESS
5
7. CONCLUSIONS
12
8. REFERENCES
12
ALTA Free Library www.altamet.com.au
1. INTRODUCTION Cyanide processing of Au ores was originally developed over 100 years ago. Prior to this time, amalgamation with Hg was the preferred technology [1]. The driving force for the change to cyanide processing was purely economic. With amalgamation only about 60% of the Au content could be recovered, whereas cyanidation provided a Au recovery of about 90%. Cyanide is essentially a universal solvent for Au, and can be used to treat almost any type of Au ore. Other solvents such as bromides, chlorides, thiourea, and thiosulfate will generally work only on very specific ore types. Au usually occurs at low concentrations in ores, typically less than 10 g/ton. At these concentrations the use of aqueous chemical hydrometallurgical extraction processes represents the only economically viable method for Au extraction [2]. Consequently, cyanide remains the preferred, usual, method for extraction of Au from its ores. Beginning some twenty years ago, the use of cyanide became an important environmental issue and remains so today. This increased the pressure to develop substitutes for cyanide, and considerable effort was also put forth to investigate various processes for cyanide recovery once the Au had been extracted. However, to the best of our knowledge, cyanide is still used almost exclusively in any significant Au mining operation, although considerable efforts are still being undertaken to develop commercial substitutes. Activated C is the preferred method in the Western World for the recovery of Au from cyanide solution. The Merrill Crowe Process (precipitation from pregnant solution with Zn powder) has limited application (when the Ag to Au ratio exceeds ~4:1). Various ion exchange (IX) resins have been used successfully for Au extraction and recovery from solution, particularly, in Asia, and to a very limited extent, in South Africa. However, IX resins have not been used in the Western Hemisphere in any large scale Au mining operation. Molecular Recognition Technology (MRT) offers a viable, competitive, and environmentally friendly alternative for Au extraction and recovery from mine cyanide streams. Several examples will be presented and discussed that support this proposition. 2. STATUS OF THE GOLD MARKET Presently, the Au price is hovering around $1,130 per troy ounce, and is highly volatile, trading in a range of almost $20 per troy ounce for the day. Market commentators have dramatically conflicting views on the forward outlook for the Au market and price. However, a few facts are reasonably apparent. The world jewelry industry has historically been the major physical market for Au, accounting for roughly 70% of total demand. Consumption in jewelry has now dropped to only about 40% of total demand, and actual consumption in the jewelry industry has dropped by 25% [3]. The quantity of Au being recycled from various scrap sources now exceeds new mine production. At present prices this situation is likely to continue. Au investment demand in 2009 exceeded Au consumption in jewelry for the first time since 1980 [3]. The GFMS World Gold Limited is of the opinion that the current Au market is not in long term equilibrium, and, there is a sizeable Au surplus. Consequently, the current prices are not sustainable as the market faces considerable forward headwinds [3]. However, it is entirely possible that short term prices may rise further, perhaps to peak at $1,300 per troy ounce, depending on day to day events. In any case, it is difficult to foresee a situation in which the market declines to a level of $200 per troy ounce, which prevailed only a few years ago. According to the GFMS, the world average cost of Au mine production in the first half of 2009 was $457 per troy ounce [4]. The production cost in the second half of 2009 was expected to approach $500. The outlook for the world Au mining industry would seem to be quite bullish for sound projects. Thus, the interest in new process and flow sheet technologies in the industry will continue and finances should be available.
3. COMPETITIVE TECHNOLOGIES FOR GOLD EXTRACTION AND RECOVERY FROM SOLUTION
3.1 USE OF ACTIVATED CARBON AND ION EXCHANGE RESINS Granular, coconut shell activated C is widely used for recovery of Au from cyanide solutions. In this process, the Au is actually adsorbed in pores in the activated C. A wide range of process flow sheets can be employed including Carbon-in-Column (CIC), Carbon-in-Pulp (CIP), and Carbon-in-
2 ALTA Free Library www.altamet.com.au
Leach (CIL). In spite of the widespread usage, there are many disadvantages to the use of activated C. Au loading rates are relatively slow, meaning a long dwell time in the leach process to achieve a high Au recovery percentage, a relatively low Au loading capacity, resulting in large quantities of activated C to achieve an equivalent Au removal quantity, and the need for complex, lengthy and expensive thermal activation and regeneration. The presence of naturally occurring carbonaceous or organic matter can dramatically reduce the activated C efficiency. This phenomenon is known as “preg-robbing” [1]. The activated C does not recover 100% of the Au in solution. Other impurity elements such as Cu and Zn in the cyanide solution will interfere with the Au recovery and co-load on the activated C. The efficiency of recovery from solutions with very low Au concentrations is very low and this efficiency decreases as Au content decreases. Considerable quantities of Au-bearing activated C can be lost downstream due to physical attrition. This necessitates the periodic use of messy, cumbersome activated C-recovery procedures and then outside refining of the spent activated C to recover the Au. In any case, some of the Au on the activated C will be physically lost. The spent activated C must always be periodically replaced. Work began on the potential use of IX resins to replace the CIP process some 30 years ago as it was felt that IX resins could prove to be superior adsorbents [5]. It was felt that the resins have potentially higher loading capacities, higher loading rates, are less likely to be poisoned by organics, and do not require thermal regeneration. The IX resins were more expensive than activated C. One particular resin used in the Resin-in-Pulp (RIP) process to replace the Resin-inCarbon (RIC) process was less selective for Au than the activated C, but can have a higher loading capacity [6]. This resin is said to be more effective than activated C through a range of Au concentrations using RIP to replace CIP [6]. IX resins have also found niche applications, for example, in the treatment of “preg-robbing” ores to replace the CIL process, using the Resin-inLeach (RIL) process, where the Au recovery can be significantly improved [6]. In South Africa, another unique, automated, IX process was developed to recover Au powder from activated C eluate [6]. The use of IX fibers for Au recovery has also been reported [7]. 3.2 NEED FOR A CLEAN, ENVIRONMENTALLY FRIENDLY GOLD EXTRACTION AND RECOVERY PROCESS Current methods for Au extraction and recovery (Section 3.1) have many limitations. First, these methods are not designed to be environmentally friendly. A large number of chemicals are introduced into the process. Some of them, such as cyanide ion, are toxic and any discharge, accidental or intentional, can have severe ecological and environmental consequences. Second, the processes are lengthy and inefficient. Precipitation and adsorption processes require much manual handling of material and are labor-intensive. Third, current methods become much less efficient as the Au concentration decreases and fail at ppb and low ppm levels. This is an important limitation because many streams contain levels of Au in this range. Lack of an economic-recovery option results in these streams or solids going to landfill together with other toxic metals present with the Au. Fourth, a large Au inventory is unavailable for extended periods during processing. This cost can be substantial. Fifth, there are usually no effective ways to remove and safely dispose of contaminants either from the Au ore or from species introduced during processing. The result is that these wastes are used for landfill or discharged into heaps. These methods of disposal can have dire ecological and environmental effects. IBC’s MRT process overcomes these limitations and provides, at many places in the Au processing flow sheet, a means to improve the Au extraction and recovery processes. In the following sections, MRT is described, examples are given of its use in Au recovery, and its potential use in Au extraction recovery flow sheets is described. 4. MOLECULAR RECOGNITION TECHNOLOGY 4.1 DESCRIPTION OF MRT MRT is a highly selective, non-ion exchange system, using specially designed organic chelators or ligands that are chemically bonded to solid supports such as silica gel or polymer substrates. The MRT process [8,9] utilizes “lock and key,” or “host guest” chemistry as a basis for its high selectivity. The solid-phase SuperLig® system consists of small particles to which the selective ligand is attached. The SuperLig® product is packed into fixed bed columns that, in commercial operation, can be built in skid-mounted modular form, and can be fully automated for continuous
3 ALTA Free Library www.altamet.com.au
operation. The feed solution is passed through the column and the target ion is removed from the solution. SuperLig® products are designed to bind selectively with ions based on multiple parameters such as size, coordination chemistry, and geometry. In contrast, conventional separation methods such as precipitation, IX, and solvent extraction (SX) generally recognize differences between ions based only on a single parameter (i.e., charge, solubility, size). SuperLig® products can bind ions even when they are present in highly acidic or highly basic solutions and/or in solutions containing high concentrations of competing ions. The MRT process exhibits high selectivity, high binding factors, and rapid reaction kinetics, resulting in a very efficient separation. The simple elution chemistry uses small volumes of eluate. Thus, dilute solutions, such as those expected from the dissolution of the low-grade resources under discussion, are concentrated and can be treated to produce, with minimal environmental impact, either marketable products of high-added value or pure products that can be disposed of in an environmentally safe manner. Due to high selectivity, high loading capacities, and rapid loading and release kinetics of the SuperLig® materials, application of the IBC MRT procedure to commercial scale operations results in substantially lower capital and operating costs than are found for other technologies like IX, SX, and chemical precipitation. Because relatively small quantities of the appropriate SuperLig® product are required, the scale of the installation can be smaller; solution wash and elution chemical requirements and volumes are substantially less; and high feed solution flow rates are possible. Higher efficiencies are attained due to single-pass high percentage removal of the target species. SuperLig® products have a long life expectancy and do not introduce contaminants into the separation process. MRT can be used to accomplish metal separations at low ppb/ppm levels that are not possible using traditional technologies. The effectiveness of traditional technologies decreases sharply as the metal content in the feed stock decreases toward the ppm level. The commercially pure products produced from such feed stocks using MRT can be sold or recycled. This is an important factor from the standpoints of cost, the environment, and waste disposal. A wide choice of eluent formulations is usually available to ensure compatibility with particular plant requirements. Highly concentrated eluent solutions can be produced from which the simple recovery of a high purity, high value-added product is possible. The use of MRT has wide applicability to the removal of impurities directly from low-grade feed stocks. SuperLig® materials are available for a wide range of cationic and anionic species [9-13]. The examples in Sections 4.1.1 and 4.1.2 illustrate the capabilities of MRT. In both cases, the target metal is selectively extracted at ppm levels directly from the process solution without the need for a pre-concentration step. The resulting elution of the target metal produces a concentrated solution that can be easily converted to a valuable commercial product. 4.1.1 Recovery of ppm levels of palladium from spent catalysts SepraMet, Ltd., a wholly owned subsidiary of IBC, operates a metals refinery in Houston, Texas. Low-grade spent petroleum/petrochemical catalysts are digested, yielding an acidic solution containing ppm levels of Pd. This solution is passed directly through the MRT system which selectively binds the Pd. The Pd is eluted from the MRT columns. The eluent contains high purity Pd at a high concentration that can be sold directly or reduced to Pd metal [14]. 4.1.2 Extraction of ppm levels of bismuth from copper tank house electrolyte Bi impurity levels must be maintained at target concentrations in Cu tank house solution to prevent Cu embrittlement. The Cu electrolyte is passed directly through the MRT system which selectively binds the ppm level bismuth impurity. The bismuth is eluted from the MRT column, resulting in a highly concentrated bismuth sulfate solution. Following a liquid/solid separation, the bismuth sulfate can be marketed to offset the cost of its removal [15]. 5. KEY ADVANTAGES OFFERED TO THE GOLD INDUSTRY BY THE MOLECULAR RECOGNITION TECHNOLOGY PROCESS The MRT process for Au extraction and recovery offers a number of specific advantages over both the activated C and IX processes, as described below:
4 ALTA Free Library www.altamet.com.au
• • •
•
• •
• •
• • • •
Extremely high selectivity for Au over other contaminants such as the base metals Cu, Zn, Co, Ni, and Fe. Essentially quantitative 100% Au extraction and recovery from solution. Higher Au-loading capacities than are attainable with competitive technologies such as activated C or IX. Loading capacities of up to 100 g Au per kg SuperLig® product are achievable. Faster metal loadings than are possible with competitive technologies such as activated C or IX. Loading rates of up to 0.2-0.4 L/kg of resin per minute are achievable. Wider concentration range for Au extraction including extremely low ppb/ppm levels of concentrations at which no other technology is effective. Availability of compatible, low-cost reagents for elution, which readily strip the precious metal at ambient pressure and temperature, and from which the precious metal can be readily recovered in high purity form. For example, SuperLig® 127 uses a simple water elution. Inexpensive regeneration procedures. The SuperLig® product can be readily and rapidly eluted and recycled for multi-cycle operation. Availability of a wide range of potential ligand support systems, and SuperLig® containment systems to provide necessary durability and ability to operate in a wide range of media from clear solutions to slurries and pulps. Economic to install and operate. Capability to handle high solution volumes and high flow rates. Minimum possibility of blinding or fouling of the SuperLig® product with base metals or fine particles. Environmentally friendly process with minimal carbon footprint. Treated solutions can be used for agricultural purposes or discharged into culinary water systems.
These inherent properties of the SuperLig® products allow MRT plants for Au recovery to produce a high purity, directly marketable Au product at the mine site. The high purity Au eluent solution generated in the MRT process can be directed to Au recovery using conventional technology. Other benefits to the mine include: major reductions in process Au inventory in the mine process flow sheet/recovery pipeline resulting in improved cash flow, lower interest costs, and reduced security risk. There is a reduction in Au loss in the MRT process. Direct sale of 99.99% Au as ingot, sponge, or grain is possible to the consuming markets by the mining company, thus bypassing traditional outside toll refiners. 6. POTENTIAL APPLICATIONS TO THE GOLD INDUSTRY FOR THE MOLECULAR RECOGNITION TECHNOLOGY PROCESS 6.1 GENERAL Some examples of the potential process applications for the MRT SuperLig® materials in the Au mining industry are as follows: • Au from pregnant alkaline Merrill-Crowe NaCN leach solution [16,17]. • Au from dilute alkaline NaCN leach solution (replacement for CIC and CIL processes). • Au from dilute alkaline NaCN leach solution (replacement for CIP and RIP processes). • Au from aqueous solutions containing Au at low (ppb/ppm) concentrations (Au scavenger from mine tailings ponds, waste water streams, etc.) [16,17]. • Au from dilute alkaline NaCN leach solution containing high levels of Cu, base metals, and/or Ag [16,17]. Recovery of Cu and/or base metals from dilute alkaline NaCN leach solution following Au removal [16,17]. • Recovery and recycle of free CN in alkaline NaCN leach solution [16,17]. An example of Au extraction from an aqueous dilute clarified alkaline NaCN stream using SuperLig® 127 is provided in Section 6.2. SuperLig® 127, on a polystyrene support, is highly 0 selective for Au and will reject base metals or Fe. The elution used is water at ~70 C. Au is present in the eluent product as NaAu(CN)2 or KAu(CN)2 . The SuperLig® 127 is washed with a dilute salt solution and is then available for reloading. The particle size of this SuperLig® product + + is ~0.5 mm. Total available capacity of the SuperLig® product for (Na /K )Au(CN)2 is ~1 mole/kg.
5 ALTA Free Library www.altamet.com.au
-
+
+
Actual loadings depend on the concentrations of Au(CN)2 and either Na or K , depending on which is present. The eluate can be directed to either Au electrowinning or chemical Au reduction to produce a 99.99% Au product. The above properties make SuperLigÂŽ 127 appropriate for use in a number of process applications including Au recovery and purification from clarified pregnant Merrill Crowe solutions, or Au recovery and purification from dilute alkaline streams. 6.2 GOLD EXTRACTION FROM AN AQUEOUS DILUTE ALKALINE SODIUM CYANIDE STREAM CONTAINING SILVER Figure 1 shows a schematic diagram of a typical flow sheet for Au extraction and recovery from a mine-agitated tank leach solution, or heap leach solution, using a CIC system. The clarified sodium cyanide alkaline leach solution containing Au, Ag, and base metals is passed through a column system in series loaded with activated C. The Au, Ag, Cu, and most base metals will also load on the activated C. Due to the base metals present, the loading capacity of the activated C for Au and for Ag can be greatly reduced, thus requiring a much larger system to ensure that all of the Au and Ag are recovered. Hot NaOH and sodium cyanide solution are periodically used as the eluent to strip the activated C of the precious and base metals. The eluent solution containing NaAu(CN)2, NaAg(CN)2, and base metal cyanides is then passed to an electrowinning system based on NaOH at pH around 12. The Au, Ag, Cu, etc. are all deposited on the cathode. The cathode material is then melted and cast into impure Au dore bars. The bars require refining and purification to minimum 99.99% Au. The Ag is recovered in the Au refining process and refined separately.
6 ALTA Free Library www.altamet.com.au
Gold Ore + Ag + Base Metals
Comminution and Pre-Concentration NaCN Leach
Periodic Gold and Silver Strip or Elution with Hot NaOH and NaCN Solution
NaAu(CN)2- + NaAg(CN)2-+ Base Metals
Clarification
L
S
Tail s
Activated Carbon- in -Column (CIC) Units in Series
Raffinate Solution to Cyanide Recovery and Disposal
NaAu(CN)2 NaAg(CN)2 NaOH Electrolyte -
Gold Electrowinning
Gold Dore Bullion (May Contain Ag + Au + Base Metals)
NaOH Addition
Figure 1: Schematic activated Carbon-in-Column flow sheet showing extraction and recovery of gold from aqueous alkaline cyanide mine feed solution. Base metals = Cu, Zn, Co, Ni, and Fe.
Figure 2 provides a schematic diagram of a typical flow sheet incorporating an MRT system to extract and recover Au and Ag from the mine leach stream on a selective basis. The MRT column system can directly replace the CIC column system (Figure 1) in the flow sheet. The three-stage
7 ALTA Free Library www.altamet.com.au
MRT column system, loaded with SuperLig® 127, loads the Au. The Au eluent solution at high purity is passed to the conventional Au electrowinning cell where 99.99% Au is deposited at the
Gold Ore + Ag + Base Metals
Comminution and Pre-Concentration NaCN Leach
3 Stage MRT Column System
NaAu(CN)2 + NaAg(CN)2+ Base Metals
Clarification
L NaAu(CN)2 in H2O Elution Stage 2 MRT Column System Stage 3 MRT Column System
S
Stage 1 MRT Column System
Tail s
SuperLig® 127 in Each MRT Column System Base Metals + Ag
Raffinate Solutions to Further Treatment
Ag
Ag
NaAu(CN)2 Gold Electrowinning
High Purity Gold Bullion
Figure 2: Schematic MRT flow sheet showing extraction and recovery of gold from aqueous alkaline cyanide mine feed solution. Base metals = Cu, Zn, Co, Ni, and Fe.
cathode, melted, and shipped. Ag can also be selectively recovered at high purity by treatment of the Au system raffinate with a separate MRT system also using SuperLig® 127. The Ag eluent solution can then also be treated by electrowinning to produce 99.99% Ag. However, any Cu and other base metals that are not loaded pass through the system.
8 ALTA Free Library www.altamet.com.au
The MRT column system portion of the flow sheet shown in Figure 2 has been tested using an aqueous-based mine solution sample containing ~15 mg/L Au as Au(CN)2- with significant Na, base metals, and Ag present. The Ag content was ~200 mg/L. The sample was tested as received, except for filtration to remove extraneous solids. The flow rate used was 0.2 L/min/kg of dry SuperLig® 127. A single column was used for the test. All flows for the test were conducted in a down-flow mode. The NaCl wash solutions were made from reagent grade NaCl and de-ionized water. The de-ionized water elution was performed at 65 OC. The loading and pre-elution wash o steps were performed at room temperature (~25 C). The feed solution was passed through the column until the feed solution was completely polished (within detection limits of 0.05 mg/L). In commercial practice a number of columns would be used in series to ensure that the entire feed solution volume is fully polished, and that the lead column is fully loaded. The Au extraction and recovery tests using SuperLig® 127 were successful. The SuperLig® 127 product for Au(CN)2- binding and separation during the loading step operates via a mechanism + + involving the binding of the Na and/or the K and Au(CN)2 species present in the as received solution as part of the binding of a salt pair, i.e., NaAu(CN)2 or KAu(CN)2 to a neutral covalently + bound ligand on the SuperLig® 127 product. The presence of the highly concentrated Na in the NaCl wash allows for the binding of the salt complex to be maintained as, or entirely converted to, the NaAu(CN)2 salt pair. The fact that both the cation, Na+, and the anion, Au(CN)2-, are needed for binding to the SuperLig® 127 is also what allows the Au to be eluted in the de-ionized water elution since binding is reduced or eliminated by the absence of sufficient concentration of either the cation or anion involved in the binding. The elevated temperature aids significantly in the elution since the binding constant for the NaAu(CN)2 salt pair with the covalently bonded ligand on the SuperLig® 127 decreases with increasing temperature. The results indicated that three columns of the same size would be required to maintain a constant polish of the Au in the feed solution to a maximum of 0.05 mg/L, while fully loading the first column. The Ag in the feed was initially partially bound to the SuperLig® 127 in the column. However, as the Au loading increased, most of the Ag was pushed off the column. The base metal content in the elution was below detectable levels. The Ag in the elution was present at a much lower ratio to the eluent than in the feed. The concentrated part of the Au elution was collected at an Au concentration of ~0.75 g/L Au for the approximate 0.03 mole/kg Au capacity bound as NaAu(CN)2, from the feed solution followed by the NaCl wash. The elution was essentially complete at ~15 Bed Volumes of eluent solution. Approximately two thirds of the elution can be recycled in commercial practice, thus yielding a final concentrated solution of ~0.8 to1 g/L Au content. The Ag was present in the eluent in a mass ratio of just under 2:1 Au to Ag for the first stage separation. Thus the Au to Ag selectivity level on a mass basis is about 22. Second and third stage SuperLig® separations can be added to increase the Au purity versus that of Ag. A second stage separation would increase the Au to ~98% purity versus Ag, and a third stage would produce an approximate 99.9% Au purity. Second and third stage separations would also produce Au concentrates near 10 g/L. The eluent from each prior stage must be at an NaCl concentration of at least 0.1 M. The second and third stage separations would have Au binding capacities of ~0.3 mole/kg. 6.3 EXTRACTION AND RECOVERY OF GOLD FROM ELECTROPLATING SOLUTIONS In technical and jewelry plating, KAu(CN)2 (PGC) is the most commonly used base matrix for the Au plating electrolyte. The specifications for the plating electrolyte solutions are particularly tight in the electronics sector and the baths must be periodically replaced due to the build up of various impurities. Jewelry plating baths must also be periodically replaced for the same reason. The conventional process for production of PGC is to recover the Au from the plating solution in metallic form, refine it, and use it for the production of new PGC. The MRT process enables recovery of the PGC directly from the spent plating solution and direct conversion back into PGC, or to metallic Au. This eliminates a number of unit operations and greatly reduces the processing time, resulting in substantial cost savings. IBC has developed and commercialized a gold-cyanide-selective SuperLig® product, which can be used in the Au plating industry for selective removal of Au cyanide from a wide range of solutions including plating drag out rinse solutions, spent electrolytes, stripping solutions, and immersion Au ® solutions. The MRT separations material, SuperLig 127, selectively binds either PGC or NaAu(CN)2. Tanaka Kikinzoku Kogyo K.K. (TKK), Japan, has previously reported on the
9 ALTA Free Library www.altamet.com.au
速
completion of successful pilot tests using SuperLig 127 to recover PGC from a plating drag-out rinse solution and to recycle it directly in the form of PGC [18]. The commercial application of SuperLig速 127 for the recycling of PGC from spent Au-plating solutions at Wieland Dental + Technik GmbH&Co, Germany, was recently described [19]. An integrated conceptual flow sheet for extraction and recovery of Au from both spent plating bath and rinse solutions is provided in Figure 3. Au recovery from various mine cyanide solutions has been described in Section 6.2 and in the literature [16-18]. The binding and elution mechanisms for the MRT Au extraction system have been explained in Section 6.2. A good example of the relative simplicity and efficiency of the MRT process in the surface finishing industry is the use of SuperLig速 127 for extraction and purification of Au from the PGC spent plating solution and from the aqueous rinse solution stream. Au feed concentrations tested by IBC ranged from ~20 to 800-900 mg/L. The K concentration was ~12 g/L. Tests were successful in reducing the Au concentration in the MRT system raffinate stream which is recycled back to the rinse tank to <0.05 mg/L. Two wash and elution systems are available. The o SuperLig速 127 can be readily eluted with de-ionised water at 90 C and using a wash of 0.5 M to 1.0 M KCl. Alternatively, an elution with deionised water at 65-70 oC combined with a wash of 5 M NaCl can be used. The advantages of the NaCl pre-elution wash are the lower temperature of the elution required and the slightly higher Au concentration in the Au elution concentrate compared to the KCl pre-elution wash case. In all cases, the Au can be polished down to Au levels of <0.05 mg/L from the feed samples tested. The levels of the base metals (those found to be present in the feed sample) in the purified Au concentrate are also below detection levels leading to a high Au purity (99.99%) in all cases tested. The typical system design shown in Figure 2 is fully integrated to ensure maximum, efficient Au recovery at minimum cost. 6.4 RECOVERY OF GOLD FROM AQUEOUS ALKALINE CYANIDE SOLUTIONS ORIGINATING FROM ELECTRONIC WASTES The generation and disposal of massive quantities of electronic waste (e-waste) has become of major concern to many jurisdictions worldwide. According to one recent report, the e-waste crisis may peak in 2015 when an estimated 73 million tons of electronics are at risk of ending up in landfills and global solid waste streams [20]. This mounting crisis has highlighted the need for programs that efficiently handle collection and disposal of used electronic products in a way that is environmentally sound. Electronic product stewardship programs are rapidly sprouting in many countries. Under this concept, responsibility for environmentally sound disposal of the electronic product is shared by both the producer and the consumer. On average, e-waste has the following composition: plastics ~30%, refractory oxides ~30%, and metals ~40% [21]. The primary economic driving force for the recycling for e-waste is recovery of the metal value, particularly the precious metal value. The average metal composition of the e-waste metal-bearing components is approximately: Au: 0.1%, Ag: 0.2%, Pd: 0.005%, Cu: 20%, Ni: 2%, Sn: 4%, and Pb: 2% [21]. In addition to consumer-generated e-waste, there is also electronic manufacturing e-waste which is generated in considerable quantities. Other minor metals of value such as In, Ge, and Ga are also present in certain e-waste components.
10 ALTA Free Library www.altamet.com.au
Gold Cyanide Plating Bath ~ 1 g/L Au
Rinse Tank ~ 3 mg/L Au
Depleted Au Solution Spent Bath Solution
Printed Circuit Board Flow
Rinse Solution
Raffinate Recycle to Rinse Solution
Distilled Water Elution KCl or NaCl Wash
Washes to Waste Treatment SuperLig® 127 MRT Column System
Gold Plating Cells
Eluent Product Solution Au ≥4 g/L
High Purity Gold Cathodes
Figure 3: Schematic MRT flow sheet showing extraction and recovery of gold from aqueous alkaline cyanide electroplating solution and rinse solution. There are powerful ecological, environmental, and economic incentives to extract, recover, and recycle from e-waste. Au is the highest value component of e-waste and the emphasis has always been on Au recovery from these materials. Historically, cyanide has been the primary leachant for dissolution of Au. During the Au dissolution process, Ag, Pd, Pt, and Cu also dissolve [21]. The optimum dissolution conditions are apparently at a pH of 10-10.5 [21]. Many processes have been tried to recover these metals but the major problems have always been chemical consumption and chemical selectivity for target elements. For example, IX systems and activated C have been used, but neither of these processes is sufficiently selective. Cementation and electrolysis methods have been used for Au; however, the efficiency of these processes is low and they are not selective [21]. Favorable properties of the SuperLig® materials, as described earlier (Section
11 ALTA Free Library www.altamet.com.au
4.1) and in the references provided, are high selectivity for the target element, high-loading capacity, rapid loading and elution, ability to handle a wide range of target element concentrations including concentrations too low for other technologies to handle, minimal chemical consumption, and negligible environmental and ecological impact. These properties make MRT SuperLig® materials logical candidates for use in the economic extraction and recovery of Au as well as of other metals from the cyanide-leach solution generated from the e-waste. The Au extraction MRT process that would be used here is the same as that described for plating solutions (Section 6.3) and for Au mine cyanide solutions (Section 6.2). MRT systems can be designed and constructed for any size, concentration, and volume throughput requirement. With its proven process for recovery of ppb/ppm quantities of precious and other metals, MRT is ideal for the task of recovering these species from end-of-life (EOL) e-waste. Effective means for recycling metals of value from e-waste are available on a very limited basis in the U.S.A. Elsewhere, smelting is widely used for recovery of the precious metals and Cu. However, smelting is fraught with environmental and ecological problems and has many limitations. Many metals are lost in the massive amounts of slag and other waste products produced. For this reason, hydrometallurgical processes as described above are becoming the technology of choice. Disposal in landfills is still common. Of great concern is the practice of sending EOL products to non-OECD nations where they are dismantled and the metals of value recovered in mom-and-pop shops and backyards without any safety precautions or concern for certain adverse environmental and ecological consequences [22-27]. Implementation of procedures for efficient and clean recovery of value from these low-grade resources near their point of use is of high international priority. MRT has the potential to be such a procedure. 7. CONCLUSIONS In the mining industry, MRT has the ability to selectively extract and recover Au through a wide concentration range from a variety of process streams including heap leach, cyanide pressure leach, Merrill Crowe, tailings dam effluent, and agitated tank leach. The MRT process that selectively and separately extracts and recovers Au, and then Cu, from cyanide solution and can recycle a portion of the cyanide offers the potential to improve the economics of many ore bodies that contain high Cu and Au concentrations that were not previously economical due to high chemical consumption. In the decorative and electronics plating industries, MRT has the ability to extract and recycle Au from a variety of streams including spent plating baths, rinse baths, stripping solutions, immersion Au solutions, and PGC manufacturing streams. The treatment of ewaste for recovery of Au, as well as other precious and base metals, offers an extremely promising opportunity for the application of MRT. Au and other metals present in e-waste can be extracted at high-purity. MRT system economics can be very attractive since the MRT process can greatly simplify the process flow sheet, offer high selectivity, have low chemical consumption, and be ecologically and environmentally friendly. 8. REFERENCES 1.
“The basic process of gold recovery, introduction-amalgamation.” Retrieved April 30, 2010, from Denver Mineral Engineers, Inc. website: http://denvermineral.com/basicp~1.html
2.
“Use of cyanide in the gold industry.” Retrieved April 30, 2010, from International Cyanide Management Code for the Gold Mining Industry website: http//www.cyanidecode.org/cyanide_use.php
3.
Kitco, The Bullion Report (April 14, 2010), “Gold’s final, mature stage.” Retrieved April 30, 2010 from Coin Link website: http://www.coinlinkbullion.com/?p=680
4.
L. Burgess, “Gold mining production costs” (December 3, 2009), GO ARTICLES.Com. Retrieved April 30, 2010, from website http://www.goarticles.com/cgi-bin/showa.cgi?C=2301602
5.
V. St.Ciminelli, “Ion Exchange Resins in the Gold Industry,” JOM, Vol. 54, October 2002, 35-36.
12 ALTA Free Library www.altamet.com.au
6.
B.R. Green, M.H. Kotze, and J.P. Wyethe, “Developments in Ion Exchange: The Mintek Perspective,” JOM, Vol. 54, October 2002, 37-43.
7.
R.M. Kautzmann, C.H. Sampato, J.L. Cortina, V. Soldatov, and A. Shunkevich, “The Use of Fibrous Ion Exchangers in Gold Hydrometallurgy,” JOM, Vol. 54, October 2002, 47-51.
8.
R.M. Izatt, J.S. Bradshaw, and R.L. Bruening, “Ion Separations in Membrane and Solid Phase Extraction Systems,” In Supramolecular Materials and Technologies (Series: Perspectives in Supramolecular Chemistry, Vol. 4), D.N. Reinhoudt, Ed., John Wiley & Sons: New York, NY, USA, 1999, 225-243.
9.
N.E. Izatt, R.L. Bruening, K.E. Krakowiak, and S.R. Izatt, “Contributions of Professor Reed M. Izatt to Molecular Recognition Technology: From Laboratory to Commercial Application,” Ind. Eng.Chem. Res., Vol. 39, 2000, 3405–3411.
10.
S.R. Izatt, J.B. Dale, and R.L. Bruening, “The Application of Molecular Recognition Technology (MRT) to Refining of Platinum and Ruthenium,” International Precious Metals st Institute, 31 Annual Meeting, Miami, FL, USA, June 9-12, 2007.
11.
S. Bélanger, M. Malone, N.E. Izatt, S.R. Izatt, J.B. Dale, and R.L. Bruening, “Selective Removal of Nickel from Cadmium and Zinc-Rich Sulphate Electrolyte in the Zinc Industry,” Hydrometallurgy 2008, Phoenix, AZ, USA, August 7-8, 2008.
12.
S.R. Izatt, N.E. Izatt, R.L. Bruening, and J.B. Dale, “Update on the Application of Molecular Recognition Technology (MRT) to Separations of Interest in the Nickel and Cobalt Industry,” Nickel Processing ’10 Conference, Falmouth, U.K., June 15-18, 2010.
13.
S.R. Izatt, N.E. Izatt, R.L. Bruening, and J.B. Dale, “Update on the Application of MRT to Separations of Interest in the Zinc Industry,” in Proceedings of the COM 2010, Forty-ninth Conference of Metallurgists, Vancouver, British Columbia, Canada, October 3-6, 2010.
14.
S.R. Izatt and D.M. Mansur, “Environmentally Friendly Recovery of Precious Metals from Spent Catalysts,” International Precious Metals Institute Petroleum Seminar, Houston, TX, USA, November 13–14, 2006.
15.
K.S. Hur, Y.S. Song, J.B. Dale and N. E. Izatt, “Commercial MRT Bismuth Removal Plant at LS-Nikko Copper Refinery”, EMC-2005 European Metallurgical Conference, Dresden, Germany, September 19-23, 2005.
16.
R.L. Bruening, J.B. Dale, N.E. Izatt, and S. R. Izatt, “Selective Extraction and Recovery of Gold, Copper, and Other Base Metals from Mine leach Cyanide Solutions Using Molecular Recognition Technology (MRT),” The International Conference on the Science Technology, and Industrial Applications of Gold 2003, Vancouver, British Columbia, Canada, September 28- October 1, 2003.
17.
R.L. Bruening, J.B. Dale, N.E. Izatt, and W. Young, “The Application of Molecular Recognition Technology (MRT) for the Recovery of Gold and Cyanide at Primary Mining Operations,” Hidden Wealth Conference, Johannesburg, South Africa, South African Institute of Mining and Metallurgy, 1996, 143-149.
18.
N. Ezawa, S.R. Izatt, R.L. Bruening, N.E. Izatt, M.L. Bruening, and J.B. Dale, “Extraction and Recovery of Precious Metals from Plating Solutions Using Molecular Recognition Technology,” Trans. IMF. 2000, 238-242.
19.
G. Gutekunst, S.R. Izatt, N.E. Izatt, J.B. Dale, and R.L. Bruening, “The Commercial Application of SuperLig® Molecular Recognition Technology (MRT) Products to Recycling of Potassium Gold Cyanide from Spent Gold Plating Solutions,” International Precious nd Metals Institude 32 Annual Conference, Phoenix, AZ, June 7-10, 2008.
20.
C. Smith, “Facing the E-Waste Crisis,” Recycling Product News, September, 2009, 6.
13 ALTA Free Library www.altamet.com.au
21.
S. Kulandaisamy, J.P. Rethinaraj, P. Adaikkalam, G.N. Srinivasan, and M. Raghavan, “The Aqueous Recovery of Gold from Electronic Scrap,” JOM, Vol. 55, August 2003, 3541.
22.
E. Gies, “Leading Africans to Responsible Recycling,” New York Times Global Edition, Business of Green Special Report, New York, NY, USA, May 25, 2009, III.
23.
L. June, “Dell Bans Export of E-Waste to Developing Countries,” (May 13, 2009). Retrieved April 30, 2010, from engadget website: http://www.engadget.com/2009/05/13/dell-bans-export-of-e-waste-to-developing countries/
24.
M. Darsana, “E-Waste: New Challenges,” Kerala Calling, Vol. 28, No. 4, 2008, 41-42.
25.
E. Ogg, “Samsung Will Take Back Used Electronics for Free,” (September 8, 2008). Retrieved April 30, 2010, from Cnet website: http://news.cnet.com/8301-17938_105-10035277-1.html?tag=mncol;title
26.
E. Gies, “Bring Out Your Dead TVs,” New York Times Global Edition, Business of Green Special Report, New York, NY, USA, May 25, 2009, IV.
27.
“Statistics on the Management of Used and End-of-Life Electronics”, (updated to 2007). Retrieved April 30, 2010, from U.S. Environmental Protection Agency website: http://www.epa.gov/osw/conserve/materials/ecycling/manage.htm
14 ALTA Free Library www.altamet.com.au
TRACE ELEMENTS DEPORTMENT IN GOLD PROCESS SOLUTIONS
By
Jim Kyle, Vera Gella and Peter May Parker Cooperative Research Centre for Hydrometallurgical Solutions Murdoch University, Australia.
Presented by
Jim Kyle J.Kyle@murdoch.edu.au
CONTENTS ABSTRACT
2
INTRODUCTION
2
EXPERIMENTAL
4
RESULTS AND DISCUSSION
5
CONCLUSIONS
8
REFERENCES
20
ALTA Free Library www.altamet.com.au
ABSTRACT The deportment and speciation of trace elements during mining and minerals processing is becoming an area of increasing concern with regards to potential health and environment risks within the immediate vicinity and surrounding area of mining and minerals processing operations. This concern also extends to the transport of concentrates and other products from site to other facilities for further processing. In many cases the hazards associated with trace elements in the ore are exacerbated through their concentration in the processing plant. In Australia, the National Pollutant Inventory (NPI, 2010) lists antimony, arsenic, beryllium cadmium, chromium (III&VI), cobalt, copper, cyanide, lead, manganese, mercury, nickel, selenium and zinc as monitoring targets. From NPI data, the main trace metal emissions to the environment from metal processing operations are generally lead, arsenic, antimony and cadmium. Other trace elements of possible concern are mercury, bismuth, selenium and tellurium. However, little is known about the deportment and speciation of these trace elements in gold processing solutions, and how they vary in the leaching, carbon adsorption and tailings disposal facilities. The use of chemical equilibrium studies can enhance our basic understanding of the deportment and speciation of trace metals under conditions present in gold processing solutions, and assist in laboratory and field investigations. This paper presents some initial results from an initial two-year study into the deportment and speciation of trace metals in gold processing solutions. Although equilibrium models can be very useful in determining the deportment and fate of trace metals in process solutions, it must be remembered that these models have a number of limitations. The accuracy and precision of the modelling is very dependent on the thermodynamic database from which the models are generated, and the range of conditions for which the data is available. In addition, models based on thermodynamic data may in some instances be unrealistic for processing solutions which often have a short residence time, and for which the speciation of metals may be governed more by kinetic than thermodynamic factors. For these reasons, care must be used in the interpretation of the information generated by the models and should, where possible, be confirmed by plant data or by laboratory tests.
1. INTRODUCTION The chemistry and deportment of copper and zinc during cyanidation of gold ores are well known (Marsden and House, 2006). However, little is known of the behaviour of the minor trace elements and in particular lead, cadmium and mercury; arsenic, antimony and bismuth; and selenium and tellurium during the cyanidation process for gold recovery. In Australia, the trace elements antimony, arsenic, beryllium, cadmium, chromium (III&VI), cobalt, copper, lead, manganese, mercury, nickel, selenium and zinc are listed by the Australian National Pollutant Inventory (NPI) as monitoring targets that need to be measured and reported annually for all mine and mineral processing facility outputs. The NPI is a public Internet database that displays information about diffuse sources and emissions of 90 different substances to air, land and water reported by industrial facilities. The most recent reports can be viewed on their website (http://www.npi.gov.au). The latest reported Australian emissions data for the trace elements of interest to this review are reported in Table 1. Bismuth and tellurium are not included as they are not required to be reported.
2 ALTA Free Library www.altamet.com.au
Table 1: Estimated NPI Reported Emissions (2007-08) from metal ore mining and basic nonferrous metal manufacturing in Australia (tonnes/year)
Metal ore mining
Pb
Cd
Hg
As
Sb
Se
Air
220
5.9
1.4
55
2.5
2.9
Water
13
1.4
<0.1
3.1
1.5
1.0
Total
240
7.4
1.4
60
5
4
Basic non-ferrous metal manufacture
Pb
Cd
Hg
As
Sb
Se
Air
340
15
8.7
72
21
0.8
Water
2.9
0.6
<0.1
1
0.6
0.5
Total
340
16
8.8
74
22
1.4
The data, that include emissions from all aspects of the mining and mineral processing operations including the burning of fossil fuels on site, indicate that the main trace metal emissions to the environment from metal processing operations are lead, arsenic, antimony and cadmium. In general, emissions to air dominate the total emissions, although for certain trace elements (e.g. antimony, selenium) water emissions can also be significant. The main emissions from gold mining operations are arsenic (~17 out of a total of 60 tonnes/year) and mercury (~6 out of a total of 8.8 tonnes/year). As such, gold mining is the main contributor to mercury air emissions in Australia. Currently there is only limited understanding of the behaviour during cyanidation of the various minerals that contain trace elements or the deportment of these elements. Lead has been used as an additive to catalyse the gold leaching reaction and assist in the control of soluble sulfides during gold processing (Marsden and House, 2006). Lead toxicity is now well established but little is known of its deportment in the gold processing solutions. Similarly, cadmium is now widely classified as a human carcinogen but again its deportment in gold processing solutions has not been widely investigated. Mercury toxicity and the hazards from its use in gold amalgamation are well known, but mercury as a trace element of concern in ore roasting and gold leaching are more recent issues. Arsenic, another known carcinogen, has been widely studied in gold processing, particularly the precipitation and disposal of solubilised arsenical wastes from the roasting, pressure leaching or bacterial leaching of gold ores and concentrates. But its deportment during gold cyanidation is less well known. Antimony and bismuth have not been well studied due to their generally lower concentrations in gold ores, and while bismuth has a lower toxicity, antimony can be compared to arsenic in its toxic effects on humans. Selenium is an essential nutrient in trace amounts but is very toxic in high concentrations; it is reported to be the third most toxic trace element after mercury and lead (Peters et al., 1997). Tellurium is thought to be less toxic but still to be handled with care. Little information is currently available regarding the deportment of selenium or tellurium in mineral processing plants. However, the toxicity of antimony and selenium, and to a less extent bismuth and tellurium compounds, requires that we know more about their chemical behaviour during gold processing.
3 ALTA Free Library www.altamet.com.au
2. EXPERIMENTAL GOLD PROCESSING SOLUTIONS The data from a recent survey of the chemical composition of gold processing solutions from 25 processing plants throughout Australia were used to simulate two “typical” gold leaching solutions – representing low and high salinity. The chemical constituents in these process solutions are given below in Table 2. These process solutions formed the “background” process solutions to which the trace elements were “added” during our chemical modelling simulations. Each solution contained varying levels of sodium, magnesium, calcium, chloride, sulphate, carbonate and cyanide species. The compositions of the two solutions chosen to represent a solutions typically found in gold processing operations are presented in Table 2.
Table 2: Compositions of Simulated Gold Process Solutions Total Dissolved Solids
Na+
Mg2+
Ca2+
Cl-
SO42-
CO32-
CN-
High Salinity
64.00
5.00
0.33
98.42
19.76
0.49
0.20
0.50
0.07
0.05
0.50
0.28
0.07
0.20
Concentration (g/L)
(~190 g/L) Low Salinity (~1.5 g/L) To each of these solutions, cadmium, lead, mercury, arsenic, bismuth, selenium, antimony and tellurium 4 were “added”, in turn, at a concentrations of 1x10 M (7.5 – 20 mg/L). Using JESS (Joint Expert Speciation System) for chemical equilibria (Murray and May, 2008), speciation diagrams for each of the metals in solution were constructed. JESS is a large body of computer software concerned primarily with the modelling of chemical phenomena in solution. It places a strong emphasis on chemical speciation in complex solutions (i.e. the identity and relative abundance of the chemical entities in solution). The database contains in excess of 72,000 reactions and 215,000 constants, of which all have been critically reviewed and given a weighting factor. JESS includes ALL available equilibrium constants in the literature rather than, as is the case with most software, single values for the thermodynamic constants o o o (G , H and S ). A weighted average is used in its calculations with obviously incorrect data being given a weighting of zero. JESS is much more comprehensive than many other modelling packages, and more suited to this type of modelling, because all types of chemical equilibria can be modelled, including protonation, complex formation, redox, solubility and adsorption. Variations in solution temperature, pH and ionic strength are all taken into account. JESS, like all software systems, is constantly evolving, changing and improving. The diagrams were constructed between pH 7 and pH 13 and the Eh was varied from 810 mV at pH 7 to 450 mV at pH 13 in order to mimic the redox potential of an air-saturated solution at ambient temperature and pressure (Breuer, 2010). The speciation in the solutions containing arsenic, antimony, selenium and tellurium was also examined at a redox potential of zero mV (or 45 mV) due to their multiple oxidation states. The aim was to examine how the oxidation states of these species may change in oxygen deficient solutions that could exist in tailings storage facilities. The presence of sulfide minerals and/or soluble sulfide ions was not considered at this stage of the modelling exercise. The production of chemical speciation models not only depends on entering a number of chemical parameters into a program and pressing the “SOLVE” button. The inputs and outputs from the modelling software must be carefully examined to ascertain how the information was obtained and whether it is appropriate for the situation at hand. Process solutions are often not “at standard conditions” of 4 ALTA Free Library www.altamet.com.au
temperature, pressure and ionic strength where most thermodynamic data is obtained. The thermodynamic equilibrium constants used in models can be very dependent on these variables. As such, the use of the data obtained from modelling must be used with caution, or adapted to the appropriate conditions of interest.
3. RESULTS AND DISCUSSION TRACE ELEMENTS SPECIATION Lead (Pb) Lead is not often present in gold ores but is sometimes used as an additive, usually in the form of lead nitrate, to assist in the gold dissolution process. Although it is not certain exactly how it works, the lead is thought to reduce the passivation of the gold by deposition of a metallic monolayer on the gold surface thereby disrupting the AuCN monolayer causing the passivation (Marsden and House, 2006). In gold process solutions, lead is known to form insoluble lead hydroxide or lead sulfide, if sulfide ions are – present (Sandenburgh and Mahlangu, 2007). The lead hydroxo-complex Pb(OH)3 may be present at very low concentrations and at high pH while lead cyanide complexes are very weak (Perera, 2001) and probably do not form. Lead Speciation (Figure 1): At low salinity, the speciation diagrams generally are in line with what is known, with lead forming an insoluble precipitate up to pH 10.5. At low salinity lead carbonate was the predominant species below pH 7.5 and hydrocerussite (also called basic lead carbonate) above pH 7.5. Above pH 10.5, soluble lead hydroxo complexes become significant, with a range of complexes forming as the pH increases. What has not previously been recognised is that at high salinity, the modelling indicates that lead remains soluble up to about pH 8, with numerous lead complexes with chloride and sulphate being identified (Figure 1). The implication of these results is that lead can be solubilised in saline process waters below pH 8 as may be prevalent in tailings storage facilities. This has previously not been considered as an issue in tailings ponds and is worthy of further investigation. Experiments on synthetic solutions are currently underway to test this model. From pH 8 to pH 11, insoluble hydrocerussite predominates with red PbO(s) the dominant species above pH 11. The soluble hydroxo complexes do not become significant until about pH 13. Cadmium (Cd) Cadmium is known to form stable cyanide complexes and as such Cd minerals may dissolve during gold cyanidation. The complexes dissociate readily at low pH such that the complexed cyanide reports as weak acid dissociable (WAD) cyanide. At higher pH, under gold cyanidation conditions (pH ~10) at high 2 2 cyanide concentrations (10 M), the dominant cadmium species is Cd(CN)4 , whereas at lower cyanide -4 (10 M) the dominant cadmium species are Cd(OH)2 (s) and the Cd(CN)3 complex (Flynn and Haslem, 1995). Cadmium Speciation (Figure 2): The majority of the cadmium formed cyanide complexes above pH 7 (low salinity) or above pH 8 (high salinity). At high salinity below pH 8, the mixed chloro and sulfatocomplexes of cadmium predominated (as most cyanide is present as HCN). 2
At high salinity and above pH 8, both the tetracyanide (Cd(CN)4 ) and tricyanide (Cd(CN)3) complexes formed with the tetracyanide species dominating above pH 9. At both salinities, cadmium hydroxide formed just below pH 13. Minor Cd(CN)2 (aq) was present at low pH and low salinity.
5 ALTA Free Library www.altamet.com.au
In order to remove cadmium from solution, the cyanide concentration must be lowered sufficiently to precipitate the cadmium as hydroxide. This generally occurs above pH 8 at very low cyanide concentrations in a tailings storage facility or during cyanide detoxification (not shown). Mercury (Hg) During the cyanidation process, generally only 10-40 % of mercury in the ore is extracted along with the gold and silver, although the extent of mercury extraction is highly dependent on the concentration of cyanide in solution and the mineralogy of the mercury (Zaraté, 1985). Dissolution of mercury minerals leads to the formation of a number of species including the neutral mercuric cyanide (Hg(CN)2), which predominates at low free cyanide concentrations (10−4M), and the cyano-complexes Hg(CN)3− and 2− -2 Hg(CN)4 , the latter predominating at high free cyanide (10 M) (Flynn and Haslem, 1995). Mercury Speciation (Figure 3): At low salinity, the speciation modelling indicated that the neutral species (Hg(CN)2) predominated at pH 7 (low free cyanide concentrations), the complex Hg(CN)3− at pH 2− 8, and Hg(CN)4 above about pH 9 (at high free cyanide). -
Unexpectedly, increased salinity favoured the formation of the mercury(II) tricyanide complex (Hg(CN)3 ) 2 over the tetracyanide complex (Hg(CN)4 . From pH 10−13, the tricyanide to tetracyanide ratio remained constant at ~40:60 (high salinity) compared to only ~15:85 at low salinity. This preliminary result needs to be confirmed experimentally. Mercury complexes, being more stable than their cadmium counterparts, are harder to break down to form hydroxide or carbonate precipitates. In general, mercury is removed by precipitation during cyanide detoxification, possibly by precipitation as a heavy metal cyano-complex. Arsenic (As) Under alkaline conditions found in gold leaching by cyanidation, arsenic sulfides in the presence of oxygen are decomposed to arsenite (AsO2−) and arsenate (AsO3−), with the proportion of each dependent on solution composition, oxidation potential and pH. AsS + 4OH− AsS + 7OH−
= =
AsO2− + S + 2H2O + 3e− 2−
HAsO4
+ S + 3H2O + 5e−
Arsenate(V) forms sparingly soluble salts with calcium while calcium arsenites(III) are generally more soluble, meaning that arsenate(V) will often exist in solution at relatively low concentrations while arsenite(III) can be present at much higher concentrations (Magalhães, 2002). Neither arsenite(III) nor arsenate(V) form stable complexes with cyanide. Arsenic Speciation (Figure 4): Arsenic speciation is difficult due to the large number of species that may be formed, and the fact that the species actually formed are the result of kinetic than by thermodynamic factors. As such, the results of this and other arsenic speciation diagrams should be interpreted with caution. Our very preliminary results are based on no interconversion between the As(III) and As(V) species, and no sulfides being present in solution. The general conclusions are that, at low salinity in the presence of sufficient calcium, As(V) will mostly exist as the solid species Ca5(OH) (AsO4)3 (arsenate apatite or johnbaumite) with variable but minor 2 soluble As(V) in solution. For As(V) at high salinity, the solubility range of the arsenate species (HAsO4 ) is increased up to above pH 8 before the solid species johnbaumite is formed. As(III) will be soluble 3 2 across the pH spectrum as a either HAsO3 or AsO3 . Salinity makes little difference to As(III) speciation.
6 ALTA Free Library www.altamet.com.au
Antimony (Sb) Antimony has a very similar chemistry to arsenic in solution. It does not form stable complexes with cyanide but does form soluble antimonites(III) on dissolution of its sulfides, affecting cyanidation by consuming cyanide and oxygen and retarding gold dissolution (Marsden and House, 2006). These can be further oxidised to antimony(V) compounds in oxidising solutions. 2
3
2SbS3 + 6CN + 3O2 = 6CNS + 2SbO3
3
2SbO3 + O2 + 8H2O = 2Sb(OH)6aq) + 4OH
Antimony Speciation (Figure 5): In aerated solutions at equilibrium all antimony is present in the Sb(V) oxidation state. At low salinity, up to about pH 7.2 (and 7.5 in high salinity solutions) the predominant species is the insoluble solid Sb2O5, which dissolves at higher pHs to form the soluble complex Sb(OH)6. At the lower Eh of zero (Figure 6), the main solid species formed at low pH is Sb(IV) oxide (Sb2O4). Between pH 9 and 10, this species oxidises and dissolves to form the soluble Sb(V) complex Sb(OH)6. There is very little difference between the speciation at low and high salinity. The modelling indicates that antimony is relatively soluble in the normal pH range of cyanidation liquors, and will exist in solution as the hydroxo-complex Sb(OH)6. The long-term solubility of antimony in alkaline solutions has been confirmed in studies on 40-year-old weakly alkaline tailings pore waters which showed high concentrations of soluble antimony (>30 mg/L) in solution (Lazareva et al., 1999). Bismuth (Bi) What is known is that, under cyanidation conditions, bismuth minerals may be oxidised to bismuth(III) which will precipitate as bismuth oxyhydroxide. Whether minor concentrations of bismuth remain in solution as hydroxo- or other complexes is uncertain. Bismuth Speciation (Figure 7): The speciation diagrams indicate that bismuth exists mainly as solid bismuth hydroxychloride (BiCl(OH)2) or bismuth oxyhydroxide (BiOOH). Some soluble bismuth hydroxy complex (Bi(OH)4) may form at very high pH (above pH 12) and at high salinity. Although the solid form of bismuth needs to be confirmed, the modelling confirms that bismuth will tend to be insoluble except at very high pH. Selenium (Se) 2-
In alkaline solutions in the absence of cyanide, selenate(VI) (SeO4 ) generally predominates under highly 2oxidising conditions, while under moderately oxidising conditions, the selenite(IV) (SeO3 ) species are dominant. Metal selenite salts can precipitate under these conditions. However, the most stable form of selenium under normal conditions is elemental selenium, Se(0), which is relatively stable at all pH values in waters that are free of oxidising and reducing agents. At low redox potentials, which may occur below the sediment-water interface in tailings storage facilities, Se(IV) and (VI) can be reduced to the elemental selenium. Selenium Speciation (Figure 8): Under aerated conditions, all selenium is present in the Se(VI) oxidation 2 state, almost exclusively as the selenate ion (SeO4 ), with very minor quantities of the neutral but
soluble species CaSeO4 (aq) and MgSeO4 (aq) at low salinity, and only MgSeO4 (aq) at high salinity (due to the much higher concentration of Mg ions in solution). 0
At Eh = 45 mV redox potential (Figure 9), Se is stable up to about pH 9 (low salinity) or pH 8.5 (high 2 salinity) with the selenite(IV) species (HSeO3 and SeO3 ) becoming predominant at higher pHs. The selenate(VI) ion begins to form above pH 12. This means that in oxidising conditions (i.e. above Eh = 0 mV), these ions may build up in solution, but at low redox potentials, which may occur below the 7 ALTA Free Library www.altamet.com.au
sediment-water interface in tailings storage facilities, Se(IV) and (VI) can be reduced to the elemental selenium. At this stage, the models do not include other metals in solution. However, if dissolved heavy metals are also present under these conditions, quite insoluble metal selenides can precipitate. Because of this, selenium concentrations are generally less than 0.01 mg/L in water that is in equilibrium with solids which contain heavy metals (Frankenberger and Benson, 1994). Tellurium (Te) Tellurium is found mainly as gold telluride minerals that are refractory during cyanidation giving low gold dissolution (Henley et al., 1995). Jayasekera et al. (1991) have studied the dissolution of calaverite (AuTe2) in both acid and alkaline cyanide media and have attributed the refractoriness to passivation of the mineral surface by tellurous acid, a reaction product. The tellurous acid is soluble at low and at high 2 pH (>12) forming the tellurite ion (TeO3 ).
AuTe2(s) + 2CN + 8OH = Au(CN)2 + 2H2TeO3(s) + 2H2O + 9e The tellurite(IV) ion can be oxidised to tellurate(VI) ion under oxidising conditions: 2
2
2TeO3 + O2 = 2TeO4
Tellurium Speciation (Figure 10): At low salinity, the preliminary speciation indicates that at Eh values of 2 810-450 mV, (aerated solutions) only tellurate(VI) species (H2TeO4HTeO4 and TeO4 with minor H6TeO6) exist at equilibrium, whereas at lower potentials (Eh = 0, Figure 11) only the tellurite(IV) species 2 (H2TeO3HTeO3 and TeO3 ) are in solution, with the degree of protonation being governed by the pH. The models at high salinity are similar, except that the solid species TeO2 predominates below pH 7.5 at Eh = 0. Although the selenium and tellurium oxyanions are seen to be soluble over most of the pH range studied, we have not considered metals present in solution that may precipitate the tellurium species as metal tellurates and tellurites. With heavy metals in solution, solid selenate and/or tellurate salts may form in the solid phase.
4. CONCLUSIONS Modelling can be a valuable tool in determining the speciation and deportment of trace metals in gold process solutions. However, care must be taken when using the models to investigate solutions at temperatures and ionic strengths where thermodynamic data is not generally available. Initial investigations have shown that the models can simulate the general deportment of the trace metals under varying conditions of pH and salinity. However, these models need to be confirmed by experimental data and measurements in actual process solutions. More reliable thermodynamic data is required under process conditions to improve the accuracy of the models.
8 ALTA Free Library www.altamet.com.au
-4
Figure 1 Lead Species Distribution. Low and High salinity. [Pb] = 10 M 100
Low Salinity Pb(OH)3 90 Pb(OH)2 Pb(OH)4 -2
80
) % ( 70 n o it u 60 ib rt si 50 D s ie ce 40 p S d 30 ae L
Pb3(OH)2(CO3)2 (s) PbCO3 (s) PbO(OH) -
20
10
0 7
8
9
10
11
12
13
pH 100
High Salinity 90
80
) % ( 70 n o it u 60 b ir ts i 50 D se ic 40 e p S d a 30 e L
PbCl6 4PbCl5 3PbCl3 Pb(SO4)4 6Pb(OH)3 Pb3(OH)2(CO3)2 (s) PbO(red,s) PbO(OH) -
20
10
0 7
8
9
10
11
pH
9 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 2 Cadmium Species Distribution. Low and High salinity. [Cd] = 10 M
100
Low Salinity 90
80
Cadmium Species Distribution (%)
70
60
Cd(CN)2 Cd(CN)3 -
50 Cd(CN)4 240
Cd(OH)2 (s)
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity 90
80 Cd(Cl)SO4 -
Cadmium Species Distribution (%)
70 Cd(Cl)2SO4 Cd(CN)2
60
Cd(CN)3 50
Cd(CN)4 2Cd(OH)2 (s)
40
30
20
10
0 7
8
9
10
11
pH
10 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 3 Mercury Species Distribution. Low and High salinity. [Hg] = 10 M 100
Low Salinity 90
80
Mercury Species Distribution (%)
70
60 Hg(CN)4 2Hg(CN)3 -
50
Hg(CN)2 40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity
Hg(l)
90
HgCN4-2 HgCN380
HgCN2 HgClCN3
Mercury Species Distribution (%)
70
HgClCN2-
60
50
40
30
20
10
0 7
8
9
10 pH
11 ALTA Free Library www.altamet.com.au
11
12
13
-4
Figure 4 Arsenic Species Distribution. Low and High salinity. [As] = 10 M 10% As(V) and 90% As(III) 100
Low Salinity 90
80
Arsenic Species Distribution (%)
70
60 AsO3 3-
50
Ca5(OH)(AsO4)3 (s) HAsO3 2-
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity 90
80
Arsenic Species Distribution (%)
70
60 HAsO4 2HAsO3 2-
50
AsO3 3Ca(OH)(AsO4)3 (s)
40
30
20
10
0 7
8
9
10 pH
12 ALTA Free Library www.altamet.com.au
11
12
13
-4
Figure 5 Antimony Species Distribution. Low and High salinity. [Sb] = 10 M 100
Low Salinity 90
80
Antimony Species Distribution (%)
70
60
Sb2O5 (s)
50
Sb(OH)6 -
40
30
20
10
0 7
8
9
10
11
12
13
pH
100.0
High Salinity
90.0
80.0
Antimony Species Distribution (%)
70.0
60.0 Sb2O5 50.0
Sb(OH)6 -
40.0
30.0
20.0
10.0
0.0 7
8
9
10
11
pH
13 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 6 Antimony Species Distribution. Low and High salinity. [Sb] = 10 M, Eh=0 100
Low Salinity Eh=0
90
80
60 Sb2O4 (s) 50 Sb(OH)6 -
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity Eh=0
90
80
70
Antimony Species Distribution (%)
Antimony Species Distribution (%)
70
60
50
Sb2O4 (s)
40
Sb(OH)6 -
30
20
10
0 7
8
9
10
11
pH
14 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 7 Bismuth Species Distribution. Low and High salinity. [Bi] = 10 M 100
Low Salinity
90
80
Bismuth Species Distribution (%)
70
60
50
BiOOH(s) BiCl(OH)2 (s)
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity
90
80
Bismuth Species Distribution (%)
70
60 Bi(OH)4 BiOOH(s)
50
BiCl(OH)2 (s) 40
30
20
10
0 7
8
9
10 pH
11
15 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 8 Selenium Species Distribution. Low and High salinity. [Se] = 10 M 100
Low Salinity
90
80
Selenium Species Distribution (%)
70
60 SeO4 250
MgSeO4 (aq) CaSeO4 (aq)
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity
90
80
Selenium Species Distribution (%)
70
60
SeO4 2-
50
MgSeO4 (aq) 40
30
20
10
0 7
8
9
10
11
12
13
pH
-4
Figure 9 Selenium Species Distribution. High and Low salinity. [Se] = 10 M, Eh = 45 mV 16 ALTA Free Library www.altamet.com.au
100
Low Salinity Eh=0
90
80
Selenium Species Distribution (%)
70
60 Se(0) 50
SeO3 2SeO4 2HSeO3 -
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity Eh=0
90
80
Selenium Species Distribution (%)
70
60 Se(0) 50
SeO3 2SeO4 2HSeO3 -
40
30
20
10
0 7
8
9
10
11
pH
17 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 10 Tellurium Species Distribution. Low and High Salinity. [Te] = 10 M 100
Low Salinity
90
80
Tellurium Species Distribution (%)
70
60
TeO4 2-
50
HTeO4 H2TeO4 40
H6TeO6
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity 90
80
Tellurium Species Distribution (%)
70
60 TeO4 2HTeO4 50
H2TeO4 H6TeO6
40
30
20
10
0 7
8
9
10
11
pH
18 ALTA Free Library www.altamet.com.au
12
13
-4
Figure 11 Tellurium Species Distribution. Low and High Salinity. [Te] = 10 M. Eh =0 100
Low Salinity Eh=0
90
80
Tellurium Species Distribution (%)
70
60 TeO3 250
HTeO3 H2TeO3
40
30
20
10
0 7
8
9
10
11
12
13
pH
100
High Salinity Eh=0
90
80
Tellurium Species Distribution (%)
70
60 TeO3 2HTeO3 50 H2TeO3 TeO2 (s) 40
30
20
10
0 7
8
9
10
11
pH
19 ALTA Free Library www.altamet.com.au
12
13
5. REFERENCES Breuer, P.L., 2010. Personal communication. Flynn, C.M. and Haslem, S.M., 1995. Cyanide Chemistry – Precious Metals Processing and Waste Treatment. Information Circular 9429. U.S. Bureau of Mines. Frankenberger, W.T., Benson, S. 1994. Selenium in the Environment, Marcel Dekker, Inc., New York. Jayasekera, S., Ritchie, I.M., Avraamides, J., 1991. Prospects for the direct leaching of gold tellurides – recent developments. In Proceedings World Gold ’91, Cairns, April 21-25, 1991. AusIMM, Parkville, Victoria, 181-183. Lazareva, E.V., Bortnikova, S.B., Kolmogorov, U.P., Kireev, A.G., Tsimbalist, V.G., 1999. Metal redistribution within the sulfide tailing body. In Ármannsson, H.(Ed.) Geochemistry of the Earth’s Surface, Proceedings 5th International Symposium Geochemistry of the Earth’s Surface, Reykjavik, Iceland, 16-20 August, 1999, 195-198. Magalhães, M.C.F., 2002. Arsenic. An environmental problem limited by solubility. Pure and Applied Chemistry, 74, 1843–1850. Marsden, J. O. and House, C. I., 2006. The Chemistry of Gold Extraction, 2nd ed. Society for Mining, Metallurgy, and Exploration, Littleton, Colorado. Chapter 6. Murray, K., May, P.M., 2008. Joint Expert Speciation System (JESS) Primer. Manual to accompany JESS software. National Pollutant Inventory (NPI), 2010. Accessed at www.npi.gov.au, December 2009. Perera, W.N., 2001. Hydrolysis and cyanide speciation of some heavy metals relevant to the fate of cyanide in the environment. Ph.D. Thesis, Murdoch University, Western Australia. Peters, G. M., Maher, W. A., Barford, J. P., Gomes, V.G., 1997. Selenium associations in estuarine sediments: redox effects. Water Air and Soil Pollution, 99, 275-282. Sandenburgh, R.F., Mahlangu, T., 2007. The use of equilibrium diagrams to better understand the influence of lead additions on gold leaching in aqueous cyanide. In World Gold 2007, Cairns, Australia, 22-24 October, 2007. AusIMM , Parkville, Victoria, 265-268. Zaraté, G.C., 1985. Copper and mercury behaviour in cyanidation of gold and silver minerals. In Oliver, A. J. (Ed.), 15th Annual Hydrometallurgical Meeting. Vancouver, Canada, August 18-22. CIM Metallurgical Society, Paper No. 26.
20 ALTA Free Library www.altamet.com.au
CARBON ACTIVITY MEASUREMENTS. RATE LIMITING MECHANISMS
By
R. Pleysier, P Austin and C. Wingate Parker CRC for Integrated Hydrometallurgy Solutions CSIRO Minerals Down Under National Research Flagship CSIRO Process Science and Engineering, Australia
Presented by
Ron Pleysier ron.pleysier@csiro.au
CONTENTS
1. 2. 3. 4. 5.
INTRODUCTION EXPERIMENTAL RESULTS AND DISCUSSION CONCLUSION REFERENCES
ALTA Free Library www.altamet.com.au
2 2 3 10 10
1. INTRODUCTION The adsorption rate of gold cyanide onto activated carbon (commonly referred to as activity) controls the solution tails in a CIP circuit. The activity determines the amount of carbon required, contact time (tank size), gold inventory (lockup), and overall efficiency of the plant (Fleming 1982). Maintaining this adsorption rate is a critical parameter in the CIP process. The activity of carbon is typically determined by measuring changes in gold concentration of a known solution upon the addition of a test carbon. The rate of change in the gold concentration is plotted against time and a mathematical relationship determined. The rate constant (k) in the rate equation provides a measure of the speed of a reaction independent of the initial concentrations and thus provides a convenient measure of activity for comparison. The measured activity of plant carbons are generally compared to new carbon and used as a guide for carbon regeneration. Activated carbon has a large surface area that consists of macro, meso, and micro pores. It is widely accepted that there are three major mechanisms controlling the rate of gold adsorption onto activated carbon: 1) film diffusion: mass transport of gold cyanide from the bulk solution phase through a hypothetical hydrodynamic boundary layer or film surrounding the carbon particle; 2) pore diffusion: mass transport of the dissolved species in the solution filling the pores; and 3) surface diffusion: migration of the adsorbed molecules along the internal pore walls (Jones et al., 1989; Le Roux et al., 1991; Jones et al., 1988; Nicol et al., 1984; Muir, 1982; Demopoulos and Cheng, 2004). It is generally believed that the initial rate of gold adsorption is controlled by film diffusion (Nicol et al., 1987), whilst pore or surface diffusion or both in parallel become the dominant mechanism at high gold loadings (Ahmed et al., 1991). As different mechanisms dominate with time, rate equations for the adsorption of gold onto carbon often represent the process only over a narrow range of conditions (Le Roux et al, 1991). A number of rate equations that describe the various rate limiting mechanisms have been proposed (Le Roux et al., 1991), and found to be adequate for loading periods up to 5 hours and gold concentrations of 100 mg L-1. The mathematical complexity required, however, is often beyond that required for routine carbon activity tests. A common alternative approach for rate constant determinations is an empirical one where a mathematical model is produced to fit the observed data without a mechanistic foundation. This approach has been applied by a number of researchers (La Brooy et al., 1986) with varying degrees of success. Most are modifications of the first order rate equation and attempt to extend the conditions under which the equation accurately describes the process. Ideally rate plots would be determined under conditions similar to that in the plant. However, CIP is a continuous process with carbon in contact with leach liquors for 2-4 weeks in a process where carbon is moved in batches counter current to the slurry. It is not practical to simulate this in the laboratory and experiments only need to provide operators with consistent data for addition, regeneration and control of carbon movement to minimise gold losses. Problems arise when the experimental conditions for activity determination are not consistent. As activity procedures are not standard across the industry, adsorption rates are not directly comparable, and in some cases the rate mechanisms being measured are not representative of the carbon's activity. Meaningful comparisons of the rate constant thus require an appreciation of the initial conditions and the range for which the right mechanism is valid.
2. EXPERIMENTAL The mechanisms of gold adsorption onto activated carbon were studied using x-ray computed microtomography and standard wet chemical techniques to determine gold adsorption rates. 2.1 MICROTOMOGRAPHY X-ray computed microtomography uses penetrating X-ray radiation to image an opaque object and determine its internal features (Kyle and Ketcham, 2003; Lin and Miller, 2000; Lin and Miller, 1996; Miller and Lin, 2004). Contrast in a microtomographic image is generated by differences in X-ray attenuation that arise principally from density variations within the object. Each image corresponds to a slice through the object and by stacking a series of slices, a continuous three-dimensional map of the density variations in the object can be constructed. The effectiveness of CT scanning
2
ALTA Free Library www.altamet.com.au
technique was demonstrated by Kyle and Ketcham (2003) in identifying gold grains due to their extremely high density in contrast to other gangue minerals in gold ores. In this study, activated carbon grains loaded with gold over 24 hours were analysed by microtomography using a Skyscan 1172. The carbon samples were rotated during the x-ray imaging and the data produced a three dimensional density map of the sample. The data files allow sectional imaging (slices) through the sample in each plane. As activated carbon has a low background density (low z) the images were produced using a 60 kV source without filters. 2.2 CARBON ACTIVITY TEST The rate of gold adsorption onto activated carbon (activity) was measured by the change in gold solution concentration with time. New carbon was screened (+2.36, -1.70 mm removed), washed o and dried at 60 C prior to use. Gold solutions were prepared from standard stock potassium gold cyanide solution (1000 mg L-1) and diluted to the desired concentration with a sodium cyanide solution (500 mg L-1 NaCN). The carbon activity tests were completed for a range of initial gold and carbon concentrations. In each case the carbon (0.5 - 15 g) was added to 500 mL of gold solution -1 (0.1 - 100 mg L ) in a 600 mL beaker which was agitated using an overhead stirrer at 250 rpm. Solution samples (5 mL) were taken at various time intervals for gold assay by ICP.
3. RESULTS AND DISCUSSION
3.1 FILM DIFFUSION AT LOW GOLD LOADINGS -1
Figure 1 shows the microtomography images of an activated carbon grain loaded with 2500 mg kg gold, a typical loading observed in a gold plant. The image pair represents a coronal (front) view of the sample with the trans-axial (plan) view (lower) through the slice indicated by the dotted line. The attached trace shows the density variations through the sample (solid line), with the large pore in the centre of the carbon clearly visible in the trace. The increase in density at the edge of the -1 trace indicates the presence of gold. It is clear that at 2500 mg kg gold loading, the gold is adsorbed predominantly at the external surface of the carbon, where the adsorption sites are readily accessible. This observation agrees with the findings of La Brooy et al. (1986) who analysed the gold distribution on a loaded pellet of Norit R2515 peat carbon from a typical CIP circuit using electron microprobe and found that most of the adsorbed gold was contained in the outer surface layer of carbon. Under such conditions, it is not hard to envision that the rate of adsorption would be limited by the diffusion of gold cyanide through the hydrodynamic boundary layer surrounding the carbon particle (film diffusion) to the external surface. This is also consistent with the protruding surfaces loading in preference to the lower areas, since the boundary layer (i.e. stagnant solution phase) around the latter is thicker. It is also worth noting in the density plot that there is a density gradient penetrating into the depth of the carbon, indicating -1 that at a loading of 2500 mg kg , the gold starts to diffuse into the internal part of the carbon. This suggests the commencement of a different adsorption mechanism. The surface active sites on the carbon and to a depth of approximately 300 - 500 microns are readily accessible to the gold solution and provide the gold loading capacity at the film diffusion rate. Beyond 500 microns the rate decreases and the adsorption kinetics become surface diffusion controlled.
3
ALTA Free Library www.altamet.com.au
0 mm
2 mm
4 mm
Grayscale
2 mm
4 mm Pixel
-1
Figure 1: Microtomography slice of a carbon loaded with 2500 mg kg gold. Coronal view (top left), trans-axial view (bottom left), and greyscale map (right: 1 pixel = 5.15 um)
3.2 SURFACE DIFFUSION AT HIGH GOLD LOADINGS The microtomography images of a carbon grain loaded with 25,000 mg kg-1 gold are shown in Figure 2. The existence of large macro pores is evident in both the coronal and trans-axial images. However, surprisingly, the gold loading around these macro pores is no more significant than the bulk of the carbon interior (this was also evident in Figure 1). This indicates that contrary to popular belief, the so-called pore diffusion, which involves the mass transfer of gold through the stagnant solution within the large carbon pores, is not a major factor in the adsorption of gold cyanide onto activated carbon. It is also worth noting in the density plot that the gold loading around the perimeter of the carbon displays a plateau, indicating a loading saturation at the external surface. This plateau (reflected as a golden-coloured band in the trans-axial image) is extended to a depth of ~500 microns, indicating that the gold is readily transported into the interior of the carbon. The observation that this penetration is uniform for both macro and micro pores (the latter are not visible in the microtomography images) is consistent with the diffusion of the adsorbed gold cyanide along the internal pore surfaces (i.e. surface diffusion) being rate limiting. Similar conclusions were reached by Vegter and Sandenbergh (1996), who compared the modelling results of various researchers and found that in most instances the estimated pore diffusivities were orders of magnitude higher than the molecular diffusivity of dilute gold cyanide in aqueous solution after they had been corrected for porosity and tortuosity. Thus, there must be another mechanism other than pore diffusion responsible for the transport of gold into the interior of the carbon, namely surface diffusion. Based on microtomography data, two major rate limiting mechanisms can be used to describe the adsorption of gold cyanide onto activated carbon. Initially at low gold loadings, due to the high availability of external surface active sites, the adsorption is rapid and controlled by film-diffusion through a hypothetical hydrodynamic boundary layer. As the adsorption at the external surface approaches saturation, however, further adsorption requires the slower transport of adsorbed gold along the internal carbon surfaces into the depth of the carbon, and thus surface diffusion becomes rate limiting. In the CIP process only the initial fast film diffusion mechanism is utilised to recover
4
ALTA Free Library www.altamet.com.au
gold. When using activity tests it is only required to determine the loss of activity due to fouling of the surface sites. Test data which includes gold loadings dominated by surface diffusion reduce the calculated rate constant and provide an incorrect and low activity values.
0 mm
2 mm
4 mm
Grayscale
2 mm
4 mm Pixel
-1
Figure 2: Microtomography slice of a carbon loaded with 25,000 mg kg gold. Coronal view (top left), trans-axial view (bottom left), and greyscale map (right: 1 pixel = 5.15 um)
3.3 CARBON ACTIVITY TESTS The rate of gold adsorption onto activated carbon (activity) is typically calculated using the slope of the first order plot (ln([Au]0 / [Au]t versus time) (Nicol et al., 1984) with the rate constant, k, corrected for unit carbon concentration. However, as pointed out by (La Brooy et al., 1986), these plots are rarely linear over an extended time. A common problem with many laboratory tests is that the carbon to gold mass ratio utilised is generally small (typically less than 400). Under these -1 conditions the carbon is loaded to a greater degree than would occur in the plant (> 2500 mg kg ). Under these high loading conditions, the readily accessible external surface sites are consumed rapidly, and the gold loading becomes limited by surface diffusion into the interior of the carbon (as discussed above). Under these experimental conditions the data spans two rate controlling mechanisms and thus does not remain linear with time. -1
To illustrate, an activity test was performed with an ultimate gold loading of 20,000 mg kg (Figure 3). It is clear that the data is not linear, with the rate constant decreasing with time. Such a result is consistent with a shift in the gold adsorption rate limiting mechanism from film diffusion to surface diffusion. It is also expected that the rate of adsorption from surface diffusion would not be first order, as the distance required to transport gold to available sites increases with loading. The overall effect is to produce a plot which curves indicating a decreasing rate constant over the course of the experiment. Determination of the initial gradient (rate constant) is thus too low and also difficult to measure accurately.
In contrast, under typical plant conditions where the gold loading on carbon is lower, the gold adsorption is limited by film diffusion, and the adsorption reaction follows a pseudo 1st order rate equation (Nicol et al, 1984). This is shown in Figure 3 for carbon with an ultimate gold loading of -1 500 mg kg , where the data was linear during the first 30 minutes of the experiment. After this time,
5
ALTA Free Library www.altamet.com.au
the data deviates from first order as the gold adsorption approaches equilibrium (gold in solution becomes depleted).
6
500 mg/kg 20,000 mg/kg
ln([Au]0/[Au]t)
5
4
3
2
1
0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
Time (hours) Figure 3: Activity test presented as a first order log plot for gold loadings of 500 and 20,000 mg kg-1 An alternative, and recommended, mathematical approach to determining the rate constant from the slope of the data shown in Figure 3 is to plot the data on a linear graph and fit a first order + equilibrium relationship. The rate constant is then obtained by a least squares regression. The first order + equilibrium equation is derived as follows: rate = k([Au] - [Au]equil) Integration (for a batch reactor) yields: ln (([Au]t - [Au]equil) / ([Au]0 - [Au]equil)) = -kt -kt
[Au]t = ( [Au]0 - [Au]equil ) e
+ [Au]equil
(1)
Where [Au]t is the solution gold concentration at time t, [Au]0 is the initial solution gold concentration, and [Au]equil is equilibrium solution concentration of gold.
Figure 4 shows Equation 1 (a first order + equilibrium model), fits the data for the low loaded carbon significantly better than the high loaded carbon. The main advantage of this method for the activity calculation is that the regression is weighted towards the initial adsorption, where the concentration is the highest. So when the equilibrium gold solution concentration is low (as it is when the loading is low), a simple first order fit will yield comparable results. For instance, when the ultimate loading of gold was 500 mg kg-1, the corrected rate constant (unit carbon concentration) from the first order regression was 0.604, compared to 0.613 for the first order + equilibrium model.
6
ALTA Free Library www.altamet.com.au
1.2
500 mg/kg 20,000 mg/kg
1
[Au]t/[Au]0
0.8
0.6
0.4
0.2
0 0.0
0.5
1.0
1.5
2.0
2.5
3.0
Time (hours)
Figure 4: Activity data presented as a first order + equilibrium plot for gold loadings of 500 and 20,000 mg kg-1 respectively (data as per Figure 3.) Carbon activity tests can produce data from the two rate controlling mechanisms depending on the initial test conditions employed. Only the film diffusion mechanism, however, correctly represents the activity of carbon in most gold operations. Also, the mechanism can change within the time frame of the test, (ie at intermediate loadings), where the rate limiting step may vary from film diffusion to surface diffusion as the readily accessible external surface sites are depleted. This may lead to an error in the rate constant (k) determination. To illustrate the shift from film diffusion to surface diffusion, the activity (corrected rate constant) of new carbon was determined for a range of initial carbon and gold solution concentrations. The data spans four orders of magnitude, and is presented in Figure 5 on a log scale for clarity. To facilitate the comparison of results for different initial test conditions, the range of reagent and carbon concentrations is presented as the ultimate gold loading, which is calculated assuming that all the gold is loaded from solution (i.e. 1 g of -1 -1 carbon in 500 mL of 1 mg L Au gives an ultimate gold loading of 500 mg kg ). -1
It is clear that within an ultimate gold loading of 100 â&#x20AC;&#x201C; 3000 mg kg , the carbon activity tests return a constant value for k of approximately 0.6. These initial test conditions provide the fastest adsorption mechanism and reflect the external surface sites available for gold adsorption. However, when activity test are performed with low carbon to gold ratios (ultimate loading above -1 3000 mg kg ), the surface active sites are rapidly saturated, and hence the gold needs to diffuse deeper into the internal structure of the carbon. These experimental conditions are less appropriate for activity determinations, as the internal sites are not usually occupied during loading in most gold plants. When the activity tests are performed with higher carbon to gold ratios (ultimate loadings -1 below 100 mg kg ), loss of gold from the solution was rapid, resulting in very low gold concentrations remaining in solution. The process approaches equilibrium within the first three samples taken and the calculation errors increase. The determination of carbon activity under these conditions is not recommended as significant variability is observed in the data (Figure 5).
7
ALTA Free Library www.altamet.com.au
0.7 0.6
Activity (k)
0.5 0.4 0.3 0.2 0.1 0.0 10
100
1000
10000
100000
-1
Ultimate gold loading (mg kg )
Figure 5: Rate constant (carbon activity) as a function of ultimate gold loading.
3.4 CIP PLANTS Commercial CIP plants are a continuous process and a gold adsorption rate constant can be obtained from plant data. Plants regularly monitor the gold solution concentrations in their adsorption tank and these values represent a snapshot of the loading kinetics in the tanks. This can provide an indication of mixing efficiency, short circuiting, carbon leakage and a comparison of the overall plant efficiency (Fleming, 1982). The value of the rate constant (with respect to solution change), can be obtained from the rate equation for a continuous process; rate = k[Au] â&#x2C6;&#x2020;Aus/t = k[Au] (Auin - Auout) = k t [Auout]
(2)
Where Auin = solution gold concentration entering the tank Auout = solution gold exiting tank t
= residence time of the slurry in the tank
k
= the rate constant for the tank/system
The gold solution values and residence (slurry) time of tanks are known and hence the value of k (rate constant for the tank) can be easily calculated. Using an example set of plant data shown in -1 -1 Table 1, a (tank) rate constant of 0.5296 hr is obtained. At 12 g L carbon the value of k becomes -1 0.044 hr per unit carbon (Table 2). This is significantly lower than the 0.6 hr-1 obtained in laboratory batch tests. This inefficiency of CIP adsorption tanks is well known with CIP adsorption rates typically only 3 - 10 % of those obtained in laboratory tests.
8
ALTA Free Library www.altamet.com.au
The reason for this is that the rate of gold adsorption in standard CIP plants is controlled by the mixing efficiency of the adsorption tank rather than by the carbon activity. In addition, the CIP circuits typically operate closer to equilibrium carbon loadings than those which are utilised in the activity measurements. Adsorption of gold onto carbon is fast (non fouled carbon in clean circuits) and the rate controlling mechanism is usually the transport of the gold to the surface of the carbon. A recent process design takes advantage of the potential for faster adsorption rates by increasing the mixing efficiency of the adsorption vessels. These "pump cells" have much faster slurry turnover and result in higher adsorption rates than conventional tanks. Some comparative data is given in Tables 1 and 2.
Table 1: Plant data from CIP operations in WA Variable
Ads tank
Pump cell
2.33
2.51
0.90
1.03
3
2.5
170
411
-1
12
7
(t)
8.16
9.45
Solution grade in
Auin
(mg L-1)
Solution grade out
Auout
(mg L )
Retention time
t
Volume flow
V
-1
(hrs) (kL hr-1)
Carbon concentration [C] Carbon
(g L )
Ct
Table 2: Calculated rate constants from data in Table 1.
k (wrt â&#x2C6;&#x2020;Aus)
-1
Ads tank
Pump cell
0.044
0.082
7.3
13.7
-1
(hr g )
% of lab rate
Laboratory data tends to provide a rate constant with respect to the change in solution grades. For plant data rate constants are more commonly determined via changes in carbon loadings. To allow comparison of both plant and laboratory data common units can be achieved using: ks * Vt / Ct = kc (wrt â&#x2C6;&#x2020;Auc)
(4)
Where ks = rate constant wrt changes in solution gold kc = rate constant wrt changes in carbon gold loadings Vt = tank volume (kL) -1
Auc = gold loading on carbon (g t ) Ct = carbon in tank (t)
9
ALTA Free Library www.altamet.com.au
4. CONCLUSION Microtomography was used to investigate the activated carbon grains that had been loaded with different concentrations of gold. The microtomographic images obtained clearly showed that the gold adsorption was not enhanced around the macro pores. It was thus believed that the diffusion of gold cyanide in the aqueous phase in macro pores is poor. As a result, the contribution of pore diffusion to the total loading of gold is low and could thus be excluded from the rate limiting mechanisms. The microtomographic images also showed that at low gold loadings, the gold is predominantly adsorbed at the external surface with film diffusion being the rate limiting mechanism; whilst at high gold loadings, the external surface sites are saturated and further adsorption requires the transport of the adsorbed gold into the interior of the carbon via a surface diffusion mechanism. The initial carbon activity test conditions, and hence ultimate gold loading, have a significant impact on the measurement of carbon activity. At ultimate gold loadings below 100 mg kg-1 or above -1 3000 mg kg , the gold adsorption is either approaching equilibrium too rapidly and prone to analytical errors, or limited by the surface diffusion mechanism respectively, and hence the results obtained are not reliable or representative of the gold adsorption under typical plant operating conditions. Therefore, it is recommended that activity tests be carried out under initial conditions -1 that give an ultimate gold loading of between 100 and 3000 mg kg . The units and variables used to determine the rate constant is not consistent between either laboratory work or from plant data. It is recommended that the outlined first order method to determine the rate constant with respect to solution changes be use to determine carbon activities. Comparisons with plant data can be achieved by a simple rearrangement of the equations to obtain rate constants from either carbon loading changes or solution profiles.
5. ACKNOWLEDGEMENTS The support of the CSIRO Minerals Down Under National Research Flagship and Parker CRC for Integrated Hydrometallurgy Solutions (established and supported under the Australian Governmentâ&#x20AC;&#x2122;s Cooperative Research Centres Program) is gratefully acknowledged.
6. REFERENCES Ahmed, F.E., Young, B.D., and Bryson, A.W., Comparison and modelling of the adsorption kinetics of gold cyanide onto activated carbon and resin in a silica slurry. Hydrometallurgy, 1992, 30, 257275. Demopoulos, G.P. and Cheng, T.C., A case study of CIP tails slurry treatment: comparison of cyanide recovery to cyanide destruction. European Journal of Mineral Processing and Environmental Protection, 2004, 4(1), pp. 1-9. Fleming, C.A., Some aspects of the chemistry of carbon-in-pulp and resin-in-pulp processes. In Carbon-in-Pulp Technology for the Extraction of Gold, The Australasian Institute of Mining and Metallurgy, Melbourne, 1982, pp. 415-440. Jones, W. G., Klauber, C., and Linge, H. G., The adsorption of gold cyanide onto activated carbon. In Randol Gold Forum â&#x20AC;&#x2122;88 Perth, Randol International, Golden, Co., 1988, pp. 243-248. Jones, W. G., Klauber, C., and Linge, H. G., Fundamental aspects of gold cyanide adsorption on activated carbon. In World Gold '89. Society for Mining, Metallurgy and Exploration, Littleton, Co., 1989, pp. 278-281. Kyle, J.R. and Ketcham, R.A., In situ distribution of gold in ores using high-resolution X-ray computed tomography. Economic Geology, 2003, 98(8), 1697-1701. La Brooy, S. R., Bax, A. R., Muir, D. M., Hosking, J. W., Hughes, H. C., and Parentich, A., Fouling of activated carbon by circiut organics. In Proceedings of the International Conference on Gold. Volume 2: Extractive Metallurgy of Gold. South African Institute of Mining and Metallurgy, Johannesburg, 1986, pp. 123-132.
10
ALTA Free Library www.altamet.com.au
Le Roux, J.D., Bryson, A.W. and Young, B.D., A comparison of several kinetic models for the adsorption of gold cyanide onto activated carbon. Journal of the South African Institute of Mining and Metallurgy. Vol. 91, no. 3, pp. 95-103. Mar. 1991, 1991, 91(3), 95-103. Lin, C.L. and Miller, J.D., Cone beam X-ray microtomography for three-dimensional liberation analysis in the 21st century. International Journal of Mineral Processing, 1996, 47(1-2), 61-73. Lin, C.L. and Miller, J.D., Network analysis of filter cake pore structure by high resolution X-ray microtomography. Chemical Engineering Journal, 2000, 77(1-2), 79-86. Miller, J.D. and Lin, C.L., Three-dimensional analysis of particulates in mineral processing systems by cone beam X-ray microtomography. Minerals & Metallurgical Processing, 2004, 21(3), 113-124. Muir, D.M., Recovery of gold from cyanide solutions using activated carbon: a review. Carbon-inPulp Technology for the Extraction of Gold, Murdoch, Australia. 13-15 July 1982. pp. 7-22. Nicol, M.J., Fleming, C.A. and Cromberge, G., The absorption of gold cyanide onto activated carbon. I. The kinetics of absorption from pulps. South African Institute of Mining and Metallurgy. Vol. 84, no. 2, pp. 50-54. Feb. 1984. Nicol, M.J., Fleming, C.A., and Paul, R.L. The chemistry of the extraction of gold. In The Extractive Metallurgy of Gold in South Africa, ed. G.G Stanley. South African Institute of Mining and Metallurgy, Johannesburg, 1987, pp 831-905. Vegter, N.M. and Sandenbergh, R.F., Rate-determining mechanisms for the adsorption of gold dicyanide onto activated carbon. South African Institute of Mining and Metallurgy, 1996, 96(3), 109118.
11
ALTA Free Library www.altamet.com.au
GOLD ION EXCHANGE By Marthie Kotze Mintek: Hydrometallurgy Division MarthieK@mintek.co.za
Carbon vs Resin Large proportion of Au produced in CIS is via RIP/RIL Muruntau: 1.3 million tons of ore/month
Western World: CIP preferred technology Gold-selective resins available outside CIS since 1990â&#x20AC;&#x2122;s Resin for CIS now produced outside (Ukraine plant closed down) Advantages of resin:
Higher loading capacities higher loading rates are less likely to be poisoned by organics do not require thermal regeneration
Disadvantages of resin: Higher cost Has been a smaller product, new resins in order of 1mm
Generally: RIP/RIL more cost effective than CIL/CIP no regeneration required: CAPEX and OPEX lower
ALTA Free Library www.altamet.com.au
1
Carbon vs Resin Why CIP/CIL still preferred technology? Entrenched technology: o in operation for 4 decades o Low risk
CAPEX and OPEX of metallurgical plant often small proportion of overall costs RSA: deep mines, costs primarily in mining, so little incentive to reduce metallurgical plant costs
RIP/RIL cost-effective option for niche applications Improved recovery of Au from pregrobbing ores Minimise energy costs, no regeneration of C Small gold operations in remote areas
Resin Au recovery operations in Western World: Golden Jubilee (RIP) New Caledonia: Barbrook in SA (RIL) Avocet Mining o Penjom Gold Mine, Malaysia (RIL) o Indonesia (RIS)
Anglo Asian Mining o Gedabek Cu/Au operation, Azerbaijan (RIS)
Commercial resins Strong-base
Medium base
Minix AM2B Puro A100
Guanidine = Aurix
Weak base Inefficient at high pH
ALTA Free Library www.altamet.com.au
2
Adsorption and stripping Strong-base resin Adsorption: R+(SO4)½ + Au(CN)2- →
R+Au(CN)2- + ½(SO4)2-
H2O
Elution: + 2TU + 2H2SO4 → R+HSO4 + Au(TU)2+HSO4 +2HCN
Weak/medium-base resin Protonation: R”NR2 + H+ → R”NR2H+ Adsorption: + Au(CN)2- → R”NR2H Au(CN)2 Elution: + NaOH → R”NR2 + NaAu(CN)2 + H2O
Minix and AM2B-type CH2 CH2
CH2 Functional Group
CH2
CH2
CH2
CH2 Functional Group
CH2 CH2 CH2
CH2 Functional Group
CH2
CH2
CH2 Functional Group
TBA
Polystyrene matrix Macroporous structure Large beads: >700 µm
ALTA Free Library www.altamet.com.au
3
Strong-base resins Conventional SB resin: trimethylamine 4 eq/kg More selective for multivalent species: Ni(CN)42-, Cu(CN)32-
Minix: tributylamine (TBA) More hydrophobic groups Lower functional group content Unreacted groups interfered with performance: removed About 1 eq/kg total SB capacity
AM2B resin Mixture: trimethylamine + dimethylamine Capacity: 1 eq/kg SB; remainder (~3 eq/kg) WB Only SB active at pH of Au adsorption However, WB groups ‘absorb’ acid in stripping Neutralisation required prior recycling to adsorption
Gold eluted (%)
Minix Elution
0
2
4
6
8
10
12
Time (hours) 1M Thiourea, 0.5M H2SO4 (60C) 0.5M Zn(CN)4 (50C)
2M NH4SCN (25C)
Au strongly extracted: need destruction of Au(CN)2 complex Elution 6 hours using thiourea + H2SO4 at 60 ̊C Residual Au < 50g/t No regeneration required
ALTA Free Library www.altamet.com.au
4
Stripping: AM2B type resin stripping Fe and Cu with sodium cyanide (40 g/L+1.5 g/L NaOH) stripping Zn and Ni with dilute sulfuric acid (50 g/L) optional stripping of silver with thiourea solution (large quantities of silver) 2-stage stripping of gold with thiourea solution (80 g/L containing 25 g/L H2SO4) regeneration of resinwith NaOH converting it back to the OH- form for quaternary groups and the tertiary groups are deprotonated (3 mol/L resin) Osmotic shock on resin Elution 24-48 hours
Relative performance Adsorbent
Metal on resin, mg/kg
Solution, mg/kg
Au
Ag
Zn
Ni
Co
Cu
Fe
5
0.5
2
5
1
10
10
Au/M
Strong-base (water)
13300 <250
9000
17800
4600
15800 21300
0.16
Activated C
25200 <200
<200
460
<200
<200
1500
0.93
36300
10000
7300
<330
1200
<330
0.65
Minix
300
ALTA Free Library www.altamet.com.au
5
Relative costs: RIP vs CIP Cost Saving for RIP over CIP (%) 10 25 150 300 20 33 22 13 52 50 39 37
kt/month CAPEX OPEX
RIS vs CIS Gold mine in SA Mill water from milling-in-cyanide Water used for repulping dump material, some gold lost due to pregrobbing Problematic due to low CN content: Cu species Cu(CN)2-1 Feed: flowrate Feed: gold concentration Adsorbent loading Resin elution Carbon elution/regen Configuration Column: diameter Column: height Adsorbent: bed height Adsorbent: volume
(m3/h) (mg/L) (mg/L) (h) (h) (m) (m) (m) (m3/column)
150 1.5 3500 8 24 lead-lag-lag 2.1 4 3.1 11
ALTA Free Library www.altamet.com.au
6
RIS vs CIS: CAPEX Relative costs
1.0 0.8 0.6 0.4 0.2 0.0 CIS Adsorption Regeneration
RIS Stripping Electrowinning and smelting
CAPEX ~50% higher for CIS
RIS vs CIS:OPEX Relative costs
1.0 0.8 0.6 0.4 0.2 0.0 CIS Adsorbent Labour
RIS Reagents Maintenance
Power/Diesel
OPEX >50% higher for CIS (reagents and power)
ALTA Free Library www.altamet.com.au
7
Relative costs
CIP vs RIP: relative CAPEX 2000 1800 1600 1400 1200 1000 800 600 400 200 0 CIP Adsorption
RIP Elution/regen
Inventory
CIP vs RIP: relative OPEX R elative co sts
140 120 100 80 60 40 20 0 CIP Pow er
MINRIP Reagents
Adsorbent
ALTA Free Library www.altamet.com.au
8
Pregrobbing: RIL vs CIL 100 90 80
Au recovery (%)
70 60 50 40 30 20 10 0 ROM-A Aufeed = 5.2 g/t Corg = 0.59%
CN
CIL
ROM-B
FC-A
Aufeed = 4 g/t Corg = 0.42%
CIL+BA
Aufeed = 32 g/t Corg = 0.42%
RIL
RIL+BA
Relative CAPEX: RIL vs CIL R elative co sts
5 4 3 2 1 0 CIL Adsorption
RIL Elution/regen
Inventory
ALTA Free Library www.altamet.com.au
9
R elative co sts
Relative OPEX: RIL vs CIL 12 10 8 6 4 2 0 CIL Pow er
RIL Reagents
Adsorbent
Minix relative durability Cumulative loss, %
60
A161 RIP GT2 MP1 MP2 GT1 Minix
50 40 30 20 10 0 1
2
3
4
5
Cycles
Minix 3.5 g/t loss at Penjom (according to GL at conference) A161 RIP used at Golden Jubilee: 10 g/t loss Still most durable resin when compared to uranium products
ALTA Free Library www.altamet.com.au
10
Penjom Gold Mine
Avocet Mining: Penjom Gold Mine: Malaysia Converted CIL to RIL during 1999 Improved recoveries as resin more effective against pregrobbing Diesel used as blinding agent, C blinding significant, difficult to generate Minix resin loss: 3.5 g/t (GLewis)
North Lahut, Indonesia
North Lahut Gold Mine, Indonesia Heap leaching, RIS using Minix resin Avocet Mining Heap leach recovery low, investigated other options
ALTA Free Library www.altamet.com.au
11
New Caledonia: Barbrook Refractory and pregrobbing ore Converted CIL to RIL to optimise recoveries Employed Kemix pump cells for adsorption circuit Used kerosene to blind pregrobber Closed down
Anglo Asian Mining Gedabek, Azerbaijan: opened May 2009 25 000 oz in 9 months (March 2010) 300 000 oz over 6 yrs Heap leach + RIS using Minix Cu: Au ratio in cyanide PLS = 1000 with Au =1.5 mg/L Cu too low to make acid pre-leach economical Loading on resin Au:Cu ratio = 1 Currently open pit, might expand to underground Fixed beds for Au recovery
ALTA Free Library www.altamet.com.au
12
Gedabek
Thiourea recovery & destruction Thiourea entrained in resin after elution washed out by displacement Thiourea bleed stream and entrained in resin recycled to adsorption destroyed via CN destruction
INCO CN destruction applied to CN solution containing TU 300
Concentration, mg/L
250 200 150 100 50 0 0
10
20
30 Time, min TU CN
ALTA Free Library www.altamet.com.au
40
50
13
Gold price vs enquiries
Conclusions Number of commercial resins for gold available Resins with superior size and durability being developed for uranium and base metal RIP can compliment Au resin improvement Numerous other IX metallurgical applications successful could also promote use of IX for Au Previous poor acceptance due to perceived risk has been demonstrated by various operations RIP/RIL/RIS effective for niche applications Pregrobbing: RIL can improve recoveries significantly Smaller/medium operations: CAPEX for regeneration high Remote areas where diesel is used for power Scavenging of Au from low grade solutions
ALTA Free Library www.altamet.com.au
14
Mintek Site: Randburg, RSA
www.mintek.co.za
ALTA Free Library www.altamet.com.au
15
NEW SELECTIVE STRONG BASE ANION EXCHANGE RESINS WITH PROMISE FOR COMMERCIAL GOLD CYANIDATION By CR Marston and DJ Gisch Dow Water & Process Solutions The Dow Chemical Company
Presented by
Charles Marston cmarston@dow.com
CONTENTS
1. 2. 3. 4. 5. 6. 7.
INTRODUCTION GOLD ION EXCHANGE GOLD SELECTIVITY ON SBA RESINS NEW HYPER-SELECTIVE SBA GOLD RESIN EXPERIMENTAL CONCLUSION REFERENCES
ALTA Free Library www.altamet.com.au
2 2 3 6 10 11 11
1. INTRODUCTION
A commercial gold selective strong base anion exchange resin, patented by MINTEK of South Africa and manufactured by Dow Water and Process Solutions, XZ-91419 (“DOWEX™ MINIX”), has proven to be a stellar performer in several commercial gold mining operations. Its rapid kinetic response has made it an ideal absorbent for gold from carbonacious cyanidation processes. In addition, Anglo Asian’s Gedabek site takes advantage of the “DOWEX™ MINIX” resin’s exceptional gold/copper selectivity. This great performing resin exhibits a gold-to-copper selectivity in cyanidation PLS of about 600:1, which can be upgraded to 1000:1 by employing simple elution techniques. The new class of hyper-selective strong base anion exchange resin (Dow’s HSGR) introduced in the paper has even higher gold capacity than “DOWEX™ MINIX”, and has a gold-tocopper selectivity coefficient of about 12,000:1 . 2. GOLD ION EXCHANGE
2.1 ION EXCHANGE FOR GOLD RECOVERY IN CYANIDATION Historically, ion exchange resins have been employed for recovery of gold from commercial cyanidation leach solutions only in the former Soviet Union (FSU). In a 1998 publication by Lukey et. al.1 he asked the question: “Is Ion-Exchange Technology for Gold Extraction Ready for Commercialization?”. At that juncture Lukey’s answer was “Not Yet”. Lukey asserted that many questions need to be answered before ion exchange technology was truly ready for the commercial challenge. The author insisted that no commercially available gold resin possessed the following required attributes: -high equilibrium loading for gold cyanide
-selectivity for gold cyanide over the base metal complexes
-fast adsorption and elution kinetics
-especially copper cyanide
-simple and non-fouling elution characteristics
-high mechanical strength and toughness of resin particles
Today, hydrometallurgical engineers are discovering that the answer is now a resounding “Yes”. Gold selective ion exchange resins are considerably more versatile than activated carbon. Recent developments in ion exchange resin design make it possible to economically synthesize specialized resins for optimal gold selectivity, physical stability, kinetic performance, and resistance to organic fouling. Unlike activated carbon, resins can be eluted and returned to use without high temperature re-activation. The higher cost of resin can be easily absorbed through lower overall Capital 2 Expense (CAPEX) and Operating Expense (OPEX) . In this paper we reveal the latest advance in gold selective resin design. A few years after Lukey’s publication, Dow introduced XZ-91419 (“DOWEX™ MINIX”) gold selective 3 resin, manufactured under license to MINTEK of South Africa , and Cognis introduced AURIX gold selective resin to the marketplace. Today, several commercial mines outside of the former Soviet Union utilize ion exchange resins for recovery of gold from cyanide solutions. Avocet Gold uses Dow XZ-91419 (“DOWEX™ MINIX”) gold selective strong base anion exchange 4 resin at Penjom in Malaysia and North Linut in Indonesia. The first in a Resin-in-Leach (RIL) process and the second in a Resin–in-Solution (RIS) process. At both locations, resins are employed to overcome the limitations of activated carbon in the processing of carbonacious ores. 5 More recently, Anglo Asian installed the same resin at the new Gedabek mine in Azerbaijan . This time, “DOWEX™ MINIX” gold selective resin was applied to overcome problems with high copper content in the PLS. The elution circuit is as simple as practically possible. The eluant, consisting of 0.2M sulfuric acid and 1M thiourea at 50°C, is pumped down-flow through the column, exiting to the electrowinning cell and pumped back into the eluant tank. Gold is recovered by removing the cathodes from the electrowinning cell and washed with water hoses in a small tank. The gold and copper content of the PLS is typically 1-1.5 g/t and 1,100 g/t respectively, giving a copper-to-gold ratio of approximately 1,000:1. Despite the very high copper content, the typical loadings of gold and copper on the MINIX ion-exchange resin are 2,000 g/m3 and 2,600 g/m3 of dry resin, respectively. These figures illustrate the high selectivity of the resin for gold over copper.
DOWEX™ Trademark of The Dow Chemical Company 2 AMBERLITE™ Trademark of Rhom & Haas
ALTA Free Library www.altamet.com.au
2.2 STRONG BASE VS WEAK BASE AND INTERMEDIATE BASE RESINS Most of the ion exchange resins used in commercial Former Soviet Union (FSU) gold recovery are strong base anion (SBA) exchange resins but more recently a hand-full of gold selective “intermediate weak base resins” (IBRs) have been introduced to the gold mining community. Dow patented a series of these resins in the late 1980’s, based upon commercially available amine 6 chemistry , but has opted not to pursue the technology commercially. The most successful gold 7 selective IBR was jointly developed by MINTEK and Henkel (now Cognis), AuRIX® 100, who’s 8 chemistry is based on a guanidine functionality . These mildly basic gold selective resins are of particular interest because the adsorbed gold cyanide could be eluted with caustic, while SBA resins generally require acidification and somewhat exotic elution chemistries. Weak base anion exchange (WBA) resins and intermediate base anion exchange resins must be protonated in order to become positively charged and bind Au(CN)2 – ions (Figure 2). R
R
Polymer N: HX
R
R Polymer N: H X Polymer N: H Au(CN)2 + + MAu(CN)2 + MX R R
Free Base
Conjugate Salt
Gold Complex
Figure 1: Gold loading on weak base anion exchange resins Standard commercial WBA resins convert from their conjugate acid forms to their free-base forms at pH above 7 (Figure 2), which is well below the operating pH of common cyanidation leach 9 processes. In contrast, IBRs such as AuRIX and Dow’s diaminomethylpentane (DAMP) resin , remain at least partially protonated in the lower to middle pH range for commercial gold cyanidation. This property allows IBRs to be employed for adsorption of gold at pH 9-11 and can be efficiently stripped of gold with stronger alkaline solutions.
50 45 Au (mg/g) resin)
40
Intermediate Wk Base Resin
35 30
Std. Wk Base Resin
25 20 15 10 5 0 4
6
8 pH
10
12
Figure 2: Effect of pH on protonation of weak base anion exchange resins Since the density of charged quaternary groups within the IBRs backbone depends heavily upon the pH of the pregnant leach solution (PLS), so too does gold selectivity. Hence, careful control of the PLS pH is essential for optimal performance. This makes it considerably more challenging to develop a robust commercial process around IBRs than around SBA resins. In contrast, SBA resins retain their positive charge in highly alkaline solution so they do not release their Au(CN)2 – ions even at pH 14. 3. GOLD SELECTIVITY ON SBA RESINS
3.1 PREVIOUS STATE OF THE ART –
Standard commercial grade SBA resins bind Au(CN)2 ions strongly, but are largely non-selective, capturing a wide range of other anions as well. It has long been known that low charge density SBA resins, and WBA resins having a low volume density of strong base anion exchange sites, 10 11 show enhanced gold selectivity in cyanidation leach liquors .
3
ALTA Free Library www.altamet.com.au
*DOWEX and AMBERLITE are Trademarks of The DOW Chemical Company Feed TEC SSC Capacity (g/L) RESIN Ref. (eq/L) (eq/L) gold iron zinc nickel copper AM-2B (1) 7.96 0.23 6.43 0.55 1.30 1.47 1.97 AM-p (1) 1.22 0.73 3.65 6.08 1.94 2.48 11.52 AP-3x8p (1) 2.01 0.47 6.27 1.57 1.57 2.16 9.36 AP-2x12p (1) 2.17 0.46 5.59 1.34 1.43 1.47 6.64 DOWEX MWA-1 (2) 1.00 0.17 3.90 0.10 2.20 1.50 0.60 AMBERLITE IRA-93 (2) 1.20 0.12 2.20 0.60 1.70 0.80 0.60 AMBERLITE IRA-400 (2) 1.40 1.29 3.00 6.80 9.24 13.50 10.00 DOWEX MSA-1 (2) 1.00 1.01 2.20 6.80 7.90 9.90 6.80 AMBERLITE IRA-458 (2) 1.20 1.11 0.30 8.60 10.00 5.90 8.60 Dow XZ-91419.00 (3) 0.30 0.30 4.62 2.32 16.15 3.09 1.42
Select. Au/all 1.215 0.166 0.428 0.514 0.886 0.595 0.076 0.070 0.009 0.201
Select Select Au/Cu Au/Cu+Fe 3.26 2.55 0.32 0.21 0.67 0.57 0.84 0.70 6.50 5.57 3.67 1.83 0.30 0.18 0.32 0.16 0.03 0.02 3.26 1.24
(1) B.N. Laskorin et.al. J. of Appl Chemistry of U.S.S.R.,1977, Vol. 47 (8). (2) P.A. Riveros, 1993 Hydrometallurgy, 33: 43-58 (3) C. Marston, D. Gisch, THE DOW CHEMICAL COMPANY
REF. # 1 2 3
STOCK SOLUTION CONCENTRATIONS, (ppm) pH CN Free Gold Iron Zinc Nickel Copper 11.0 11.0 11.0
200 200 200
0.60 4.90 5.10
1.10 9.50 8.60
0.60 8.50 8.50
1.20 14.30 14.20
1.50 18.80 18.60
Silver NA NA NA
Table 1: Historical gold selective resin performance While Table 1 provides a convenient way to compare the “apparent” metal over metal loading selectivities of one resin versus another, it does not fully express the “true” selectivity of the resins, because it does not consider the metal concentrations of the original challenge solution. The “apparent selectivity” expressed in Table 1 is actually an expression of the metal loading ratio, rather than of the “true selectivity”. Until now, gold selective resins generally load all anions – reasonably well, but Au(CN)2 “squeezes” or “bumps” most of them off the resin later in the loading stage due to its higher affinity for fro the resin. This is “apparent selectivity”. “True selectivity” is seen in resins that do not load other ions readily but instead loads only Au(CN)2 –. Later in this paper the authors will discuss the “true selectivity” demonstrated by two apparently selective resins and one truly gold selective resin. The internal polymer matrix of strong base anion exchange resins with low volume charge density (Salt Splitting Capacity (SSC)), such as Dow XZ-91419, are less hydrophilic than standard grade, high SSC resins. The cyanide complexes of gold, particularly Au(CN)2 , are well known to be sparsely hydrated, and therefore, less hydrophilic than the multi-valent cyanide salts typically present in mining solutions (e.g. Cu(CN)43-, Cu(CN)32-, Ni(CN)42- , Zn (CN)42- and Fe (CN)64-). In addition to the lower hydrophilicity, SBA resins with low SSC will generally have their charged sites somewhat dispersed within the resin back bone. This separation of charges makes low SSC resins less “multivalent-selective” and more “mono-valent selective”. 3.2 A ROADMAP TO IMPROVED GOLD SELECTIVITY The chart of Figure 2, derived from the data in Table 1, readily reveals the relationship between gold loading capacity and gold-to-copper loading selectivity vs. SSC when the resins are challenged with a typical cyanidation pregnant leach solution (PLS). Interestingly, the gold loading capacity changes little with large changes in SSC, counter intuitively showing a slight tendency to load more gold as SSC gets smaller. The Chart also shows that gold selectivity increases with decreasing SSC, ultimately revealing an exponential increase as SSC approaches zero.
4
ALTA Free Library www.altamet.com.au
7.00 Selectivity
6.00
6.00
Gold Capacity
5.00
5.00
DOW XZ-91419 "DOWEX MINIX"
4.00
4.00
3.00
3.00
2.00
2.00
1.00
1.00
0.00 0.00
0.20
0.40
0.60
0.80
1.00
1.20
Gold Loading (g/L)
Gold Selectivity (Au vs Cu)
7.00
0.00 1.40
SSC
Figure 3: SBA Resin- Exponential relationship between SSC and Au:Cu selectivity The data in Table 1 as expressed in Figure 3 makes it clear that resins with very low SSC can load gold to surprisingly high levels, to the exclusion of the multivalent anions. 3.3 ANATOMY OF ANION EXCHANGE RESINS Most commercial anion exchange resins are manufactured employing a cross linked polystyrene matrix (Figure 4) which is activated via a chloromethylation process and then aminated with a secondary amine to form a WBA resin or with a tertiary amine to produce a SBA resin. + NR3
R3N +
Cl
NR3
Cl Cl
+ NR3
Cl
Cl
Cl
Cl Cl
+ NR3 + R3N
+ NR3
R3N +
Cl
R3N + Cl
+ R3N + NR3
Cl Cl
Cl
+ NR3 R3N +
R3N +
R3N +
SBA Cl
HNR2
R2N
NR2
NR2
Chloro-methylated Polymer NR2
R2N
R2N
+ R2 N
R3N +
N R2 +
NR2 R2N
WBA Figure 4: Synthesis of polystyrene based anion exchange resins Standard SBA resins are high in SSC and are relatively unselective as gold recovery resins (see AMBERLITE™ IRA-400, AMBERLITE™ IRA-458, and DOWEX™ MSA-1 SBAs in Table 1). Standard WBA resins typically have a small number of artifactual salt splitting sights due to amine bridging or tertiary amine contamination of the secondary amine. Weak base resins without salt splitting capacity are not selective for gold; however, some WBA resins have up to 20% of their amines functionality as quaternary salt splitting sights. This extra salt splitting capacity in WBA resins is often incorporated intentionally. Such is the case with DOWEX™ MWA-1 WBA resin. Serendipitously, this latent salt splitting capacity (~15%) makes DOWEX™ MWA-1 a reasonably good gold selective resin. The Russian gold selective resin AM-2B, employed in several gold mines of the FSU, has about 20% of its amine sites as quaternary salt splitting sights.
5
ALTA Free Library www.altamet.com.au
Without extraordinarily careful control of the chloromethylation process, it is virtually impossible to homogeneously distribute the chloromethyl groups on the polymer backbone. Distribution which is necessary for good quaternary group sight separation. Separation that is important in the deselection of multivalent ions. Dow applies proprietary functional group distribution chemistries in the manufacturing process for “DOWEX™ MINIX”. In addition to the nearly ideal sight separation in “DOWEX™ MINIX” resin, the functional amine is tri-n-butyl versus the more common tri-methyl amine. The aliphatic character of the tri-n-butyl amine helps make the resin even more 12 hydrophobic, further enhancing gold selectivity .
4. NEW HYPER-SELECTIVE SBA GOLD RESIN
4.1 LEARNING FROM “DOWEX MINIX” For the past 20 years, resin manufacturers and mining technology houses have tried to synthesize a more selective gold resin based on the principle of hydrophobicity and charge separation within the resin polymer backbone. “DOWEX™ MINIX” (Dow’s XZ-91419) was developed with these properties as the basis for their design and has successfully taken its place as an economical alternative to activated carbon, especially in the case of preg-robbing carbonacious ore4 and high 5 copper ores . “DOWEX™ MINIX” gold selective resin is fully commercialized and has been adopted as the primary mode of gold adsorption at several commercial mines. It is revered as the most gold selective commercially available strong base anion exchange resin. This resin has gold selectivity over copper that allows it to be used in very high copper PLS. At Anglo Asian’s Gedabek mine, the gold and copper content of their PLS is typically 1-1.5 g/t and 1,100 g/t respectively. Even with this high copper to gold ratio, their simple elution process gives a phenomenal copper-to-gold ratio of approximately 1,000:1 with moderate gold loadings and 98 plus percent gold recovery, with only 10 minutes of adsorption contact time. While “DOWEX™ MINIX” has proven to be highly – 32selective, especially for Au(CN)2 over Cu(CN)4 and Cu(CN)3 , a resin with even greater selectivity could enable significant reduction in overall mine site operation expense. 4.2 SYNTHETIC STRATEGY As remarkable as is the performance exhibited by “DOWEX™ MINIX” gold selective resin, the authors have taken the technology to a new level. The exponential trends evident in the chart of Figure 2 are an interesting teaser to an ion exchange chemist and support the hypothesis that says “lower SSC and more efficient site separation, will result in SBA resins with extremely high gold over copper selectivity without sacrificing gold loading”. According to this hypothesis, a revolutionary gold selective SBA resin should be achieved by lightly reducing the SSC to a level slightly lower then the lowest on the chart, and/or by more efficiently spacing the salt splitting quaternary groups in the resin backbone. Development of a synthetic pathway to achieve such a resin, pushes the limits of the art, nevertheless, the authors have successfully tested these limits. While the technology applied by the authors to this task is progressive, it is also elegantly simple. The authors started with a standard chloromethylated copolymer with a great number of chloromethyl groups that are far too close together to achieve well separated charge density within the polymeric backbone. Cl
Cl Cl
Cl
Cl
Cl
Cl
Cl Cl
Cl
Cl
Cl
Catalyst
Cl
Cl
Cl
Cl
Figure 5: Methylene bridging for separation of functional sites The authors applied Dow’s proprietary “methylene bridging” technology to catalytically cross link most of the chloromethyl groups, eliminating a great number of the chloromethyl groups in the process. The resulting resin has extraordinarily well dispersed chloromethyl functionality.
6
ALTA Free Library www.altamet.com.au
4.3 A HYPER-SELECTIVE GOLD RESIN Upon amination of the methylene bridged chlorometylated copolymer with tri-methylamine, the authors had prepared a candidate resin with the potential to exemplify a new class of highly selective strong base anion exchange resin. CH3 CH3 N CH3 +
CH3 CH3 N CH3 + CH3 CH
3
+ N CH3 CH3
CH3 CH 3
+ N H3C
+
CH3
N CH3
Figure 6: Dow Hyper-Selective Gold Resin (HSGR) Samples of the new resin (Dow’s HSGR) were challenged with mock cyanidation PLS in the same fashion as was employed for the collection of data for Table 1, with the following results: *DOWEX and AMBERLITE are Trademarks of The DOW Chemical Company Feed TEC SSC Capacity (g/L) RESIN Ref. (eq/L) (eq/L) gold iron zinc nickel Dow XZ-91419.00 (3) 0.30 0.30 4.62 2.32 16.15 3.09 Dow HSGR (4) 0.09 0.09 5.72 0.36 0.62 0.77 Dow HSGR (5) 0.09 0.09 6.03 0.09 0.34 1.04
copper
Select. Au/all
Select Au/Cu
Select Au/Cu+Fe
1.42 0.03 0.01
0.201 3.200 4.087
3.26 168.09 602.90
1.24 14.51 60.29
(3), (4), and (5) C. Marston, D. Gisch, THE DOW CHEMICAL COMPANY
REF. # 3 4 5
pH 11.0 11.0 11.0
STOCK SOLUTION CONCENTRATIONS, (ppm) CN Free Gold Iron Zinc Nickel Copper 200 56 144
5.10 2.25 2.33
8.60 10.60 10.50
8.50 18.90 19.30
14.20 4.89 4.93
Silver
18.60 20.30 19.93
NA 1.06 1.08
Table 2: hyper-selective gold resin performance vs. “DOWEX MINIX” (XZ-91419.00)
700.00
7.00
600.00
6.00 Selectivity
500.00
5.00
HSGR High Free Cyanide
Gold Capacity
400.00
4.00
300.00
3.00
200.00
2.00
Gold Loading (g/L)
Gold Selec. Log (Au/Cu)
From Table 2, it is clear that the synthetic strategy chosen by the authors has resulted in dramatically improved gold selectivity with an increase in gold loading capacity even though the SSC is a mere 30% of the SSC in “DOWEX™ MINIX”.
HSGR Low Free Cyanide 100.00
0.00 0.00
1.00
0.20
0.40
0.60
0.80
1.00
1.20
0.00 1.40
SSC
Figure 7: Exponential relationship between SSC and Au/Cu selectivity verified The chart of Figure 7, shows that the trends revealed when the data from Table 2, is added to the data from Table 1. The exponential trend from the chart of Figure 2 is indeed verified. The Dow’s 7
ALTA Free Library www.altamet.com.au
HSGR resin gives unprecedented performance for gold selectivity from mock cyanidation solutions. Interestingly, gold selectivity is strongly enhanced in solutions having free cyanide values in the 150 ppm range. This improvement in selectivity is related to the enhanced equilibrium distribution of tri32 valent Cu(CN)4 over the lower valent copper cyanide species, Cu(CN)3 and Cu(CN)2 -, in high free cyanide background concentrations. The gold selective anion exchange resins of this invention (Patent Pending) exhibit gold to copper loading selectivityâ&#x20AC;&#x2122;s almost 600 times greater than XZ-91419 resin, without sacrificing gold loading capacity. The authors benchmarked two industrially important gold selective resins against the new Dowâ&#x20AC;&#x2122;s Hyper-Selective Gold Resin at high and low free cyanide, at 200:1 solution to resin volume ratio and at 1000:1 solution to resin volume ratio. Equilibrium Resin Loading Capacity mg/L Low Free Cyanide 56 ppm 200:1 fluid to resin vol ratio Au Ag Cu 1202 527 5600 AURIX 1273 536 7818 XZ-91419 HSGR 1524 524 1172
Feed Reference (5) Table 2 Fe Ni Co 1098 2059 187 461 2593 444 414 959 <200
Low Free Cyanide 56 ppm 1000:1 fluid to resin vol ratio Au Ag Cu AURIX 4130 778 4559 4400 600 4286 XZ-91419 HSGR 5749 643 4396 High Free Cyanide 144 ppm 200:1 fluid to resin vol ratio Au Ag Cu 1241 522 4337 AURIX XZ-91419 1277 535 6526 1541 512 <20 HSGR
All Metals 18447 23530 4625
Zn 9655 14286 4396
All Metals 24217 28427 19242
Fe 604 502 337
Feed Reference (5) Table 2 Ni Co 2977 375 3143 <200 1792 237 Feed Reference (4) Table 2 Ni Co 1993 390 2415 508 693 74
Zn 7741 10375 3096
All Metals 16828 22138 6253
Fe 2757 2324 <500
Feed Reference (4) Table 2 Ni Co 3108 <200 3089 <200 1038 <200
Zn 7297 16154 335
All Metals 20189 28340 8340
Fe 1743 1712 2029
High Free Cyanide 144 ppm 1000:1 fluid to resin vol ratio Au Ag Cu AURIX 4378 757 1892 4619 737 1417 XZ-91419 6029 938 <20 HSGR
Zn 7774 10405 32.41
Table 3: Hyper-selective gold resin performance in high and low free cyanide solutions
4.4 GOLD LOADING Hyper-selective gold resin provides rapid loading and elution kinetics. Similar to commercial XZ91419 gold selective resin, HSGR adsorbs gold from cyanide solutions at a rate many times that of commercial activated carbons. Gold Adsorption on HSGR 10 cc re sin, 2L, 100ppm AuCN , 150 ppm Free Cyanide, RT
100
Gold ppm
80 60 40 20 0 0
50
100
150
Minutes
Figure 8: HSGR Kinetic Response - Gold Adsorption
8
ALTA Free Library www.altamet.com.au
In Figure 8 we observe the loading of HSGR with a gold cyanide solution in a stirred vessel. Ten mL of HSGR was stirred in 2L of 100ppm gold cyanide at pH 10 and with 150 ppm free cyanide. The ultimate loading capacity at equilibrium with the final gold concentration of 6 ppm was 18.8 g/L. This rapid kinetic response in the adsorption of gold cyanide is typical of gold selective anion exchange resins and is a property which makes them ideal for use in carbonatious preg-robbing ores. The resin beads capture gold at a much faster rate than does the freshly exposed native graphite in newly ground ore, driving the adsorption to the resin vs. the ores own latent carbon. In contrast, if using an activated carbon adsorbent, there is a strong competition between the preg robbing native carbon and the activated carbon. As a result of the slower adsorption of gold by activated carbon, than by resin, more gold is lost to the preg robber. 4.5 GOLD ELUTION Several commercial installations of the commercial XZ-91419 (MINIX) resin have proven the simplicity and elegance of the acidic thiourea elution of gold from gold selective anion exchange resin. While it is known that standard grade strong base anion exchange resin elute gold sluggishly with acidic thiourea, and there are even some reports of non-selective strong base resins catalysing the decomposition of thiourea, users of XZ-91419 report no significant degradation and excellent 13 elution characteristics . Some report that they are achieving complete elution of gold in 3-4 bed volumes, eluting with 0.5 Molar Thiourea in 1 Molar Sulphuric Acid at 60°C. Johns et. al. of 13 MINTEK describes in detail, the commercial thiourea elution/electrowinning process . The HSGR of this paper also elutes cleanly under these conditions. Elution of Gold from HSGR 4 BV/hr, 1BV=10cc, 0.5 M Thiourea in 1 M Sulfuric Acid, 60 deg C
6000
Au Conc. (ppm)
5000
>99.5% Recovery
4000 3000 2000 1000 0 0
2
4
6
8
10
12
14
Bed Volumes
Figure 9: HSGR Kinetic Response - Gold Elution Figure 9 shows the elution curve for a case where HSGR was loaded as in Figure 8. Essentially all of the gold (18.8g/L) elutes within a span of 6 bed volumes. These results were achieved using 10 cc of loaded resin in a 1cm diameter glass column. The authors expect that this performance can be significantly optimized when performed at a larger scale.
4.6 COMMERCIAL IMPORTANCE In a continuous counter current resin-in-pulp (RIP) plant, the resin in the first stage will load up to high levels of metal, while the resin in the last stage will have very little metal on it. As a result, the 1000:l solution-to-resin volume ratio simulates the first RIP stage, while the 200:l ratio mimics the last RIP stage. If a resin loads all metals when there is lots of resin capacity, and then relies on high gold affinity to achieve selectivity by the action of gold “squeezing off” the other metals (apparent selectivity) then, in continuous operation, base metals will load onto the resin in the last stages and be “bumped” or “squeezed” off in the earlier stages. Consequently, the base metal concentration will accumulate in solution to progressively higher levels going from the first to the last stages, which ultimately results in a reduction of the ultimate gold loading capacity. Therefore, it is much preferred if the resin is “truly selective” for gold, and has little affinity for the base metals, even when overall metals concentration is low. Unlike other “gold selective resins”, in cyanidation solutions, Dow’s HSGR appears to be “Truly Selective” for gold over base metals. Selectivity coefficients were calculated for the three resins, AuRIX, “DOWEX™ MINIX” and Dow’s Hyper Selective Gold Resin. The results are given in Table 4. In high free cyanide solutions, Dow’s 9
ALTA Free Library www.altamet.com.au
HSGR shows selectivity for gold over all metals in solution that is greater that the other two resins by a factor of ten. In addition, the selectivity of Dow’s HSGR for gold over copper in high free cyanide solution is 11,375:1, which is more than 1000 times greater than that of the other two resins. Selectivity Coefficient ka = {[Au]R/[Au]S} x {[other metal]R / [other metal]S} R: resin phase (units g/t) S: solution phase (units mg/L Ag: included in "All Metals" Selectivity Coefficient Low Free Cyanide 56 ppm Feed Reference (5) Table 2 All Metals Au/Cu Au/Fe Au/Ni Au/Co 15 24 33 7 31 AURIX 13 27 36 7 66 XZ-91419 HSGR 42 45 52 25 91
Au/Zn 9 6 42
High Free Cyanide 144 ppm All Metals AURIX 20 14 XZ-91419 HSGR 217
Au/Zn 14 6 652
Au/Cu 63 90 11375
Feed Reference (4) Table 2 Au/Fe Au/Ni 21 7 27 8 243 51
Au/Co 69 76 133
Table 4: Selectivity coefficients for hyper-selective gold resin vs. free cyanide conc. The amount of free cyanide in solution is an operating parameter that varies from plant to plant. Some plants require a high free cyanide level in solution to drive the gold leaching kinetics. Other plants can achieve the target gold extraction with much lower free cyanide. The “gold selective resins” are often far more selective for gold at high free cyanide levels than at low levels. This is true, in particular, for gold selectivity over copper, a metal that is commonly found in gold plant 2leach solutions. This phenomenon is attributed to the fact that at high free cyanide, the Cu(CN)3 complex dominates, whereas at low free cyanide, Cu(CN)2 dominates. With the hydrophobic resin matrix and the charged site separation of gold selective resins, single charged anions are selected over double or triple charged anions. 5. EXPERIMENTAL
5.1 PREPARATION OF FEED SOLUTIONS The resin loading test work was conducted using two synthetic gold cyanide leach solutions. Both solutions contained the same concentrations of gold, silver and base metals (zinc, nickel, cobalt, iron and copper). The concentrations of un-complexed, free cyanide were 20 mg/L for Solution 1 [Feed (5) Table 2] and 110 mg/L for Solution 2 [Feed (4) Table 2]. The solutions were prepared by dissolving the required amounts of the metal salts in de-ionized water, and making up the volume to 80 liters. The solutions were then adjusted to approximately pH 11 with sodium hydroxide before use for testing. The target and actual concentrations of metals and free cyanide of the two feed solutions are listed in Table 5.
Table 5: Synthetic gold cyanidation solutions for resin testing 10
ALTA Free Library www.altamet.com.au
5.2 RESIN ADSORPTION TESTWORK Resin adsorption was carried out by contacting 7mL of each resin sample (in the CN- form) with both feed solutions. At solution-to-resin volume ratios of 200-to-1 and 1000-to-1, for 24 hours. The loading tests at solution-to-resin volume ratios of 200 (1.4 L solution) were conducted in a rolling bottle, while those at the 1000-to-1 solution-to-resin volume ratio (7 L solution) were carried out in an agitated pail. Following loading, the loaded resin was recovered, dried and prepared for analysis of Au, Ag, Cu, Fe, Ni, Co and Zn. A sample of the barren solution was submitted for the same analyses, plus a cyanide titration using silver nitrate. The results are summarized in Tables 2 and 3. 6. CONCLUSION The hypothesis that states “lower SSC and more efficient site separation, will result in SBA resins with extremely high gold over copper selectivity without sacrificing gold loading” is supported by the work reported in this paper. The elegant synthetic technique, employing catalytic methylene bridging of the chloromethylated copolymer precursor, imparted near ideal site separation of the quaternary ammonium groups in the backbone of Dow’s HSGR. Dow’s new developmental hyperselective gold resin is perhaps the most gold-selective resin ever reported. It’s true gold selectivity could make it, or other resins manufactured on the same principles, the resins of choice for commercial gold recovery in resin-in-pulp and resin-in-solution applications. Research continues on Dow’s HSGRs as we are investigating their kinetic performance, elution chemistries, physical strength, sensitivity to varied amine chemistries, and resistance to fouling. Dow’s HSGR is not just a laboratory curiosity since Dow has demonstrated the capability to economically produce this type of resin on a full commercial basis. Hence, in the near future, Dow is planning to introduce a new, truly selective, commercial gold selective resin. 7. REFERENCES 1. Lukey et.al., 1998 “Is Ion-Exchange Technology for Gold Extraction Ready for Commercialization?”, G. C. Lukey, J. S. J. van Deventer, D. C. Shallcross, 19-23 April 1998, Proceedings AusIMM ’98 – The Mining Cycle, pgs. 349-354. 2. Johns et.al., 1993 “A Technical and Economic Comparison Between the Carbon-In-Pulp Process and MINRIP Resin-In-Pulp process” M.W.Johns, D March, 1993, Randol Conference, Beaver Creek, pgs. 293-299. 3. Green et.al., 1992 “Gold Selective Ion Exchange Resins”, Green et al., Jul. 28, 1992, US PATENT 5,134,169, MINTEK, South Africa. 4. Lewis et.al., 2000 “Resin-In-Leach: An Effective Operation for Gold Recovery from Carbonacious Ores”, G.O. Lewis and W. Bouwer, Randol Gold Forum 2000, Vancover. 5. Paul, 2010 “Resin-in-Solution Approach Solves Gold-Copper Selectivity Problem”, R. Paul, March 10, 2010, 11:19, Engineering and Mining Journal online. 6. Harris et.al., 1988 “Reactive Resins Useful for Precious Metal Recovery”, W.I. Harris, J.R. Stahlbusch, July 19, 1988, US Patent # 4,758,413. 7. Green et.al., 1986 “Recent Developments in Resins for the Extraction of Gold”, B.R. Green, A.H. Schwellus, and M.H. Kotze, 1986, Gold 100, Proceedings of the International Conference of Gold, Vol2: Extractive Metallurgy of Gold, C.E. Fivaz and R.P. King, eds., The South African Institute of Mining and Metallurgy, Johannesburg, South Africa, pgs.321 -333. 8. Kolarz et.al., 1999 “Influence of the Structure of Chelating Resins with Guanidyl Groups on Gold Sorption”, B.N. Kolarz, D. Jermakowicz-Bartkowiak, A.W. Trochimczuk, and W. Apostoluk, 1999, Reactive and Functional Polymers, 42, pgs. 213-222. 9. Harris et.al., 1992 “The Extraction of Gold from Cyanide Solutions Using Moderate Base Polyamine Ion Exchange Resins”, W.I. Harris, J.R. Stahlbusch, W.C. Pike and R.R. Stevens, 1992, Reactive Polymers, 17, pgs. 21-27. 11
ALTA Free Library www.altamet.com.au
10. Riveros, 1993 “Selectivity Aspects of the Extraction of Gold from Cyanide Solutions with Ion Exchange Resins”, P.A. Riveros, June 1993, Hydrometallurgy, Volume 33, Issues 1-2, Pages 43-58. 11. Green et.al., 1989 “ A Dedicated Resin for Gold – the Stimulus Needed for Universal Acceptance of Resin-In-Pulp”, B.R. Green, K.G. Ashurst, T.E. Chantson, 1989, In: R.B. Bhappu and R.J. Harden (Editors), World Gold 89 (Reno NV), pgs. 339-346. 12. Lukey et.al., 2000 “The Speciation of Gold and Copper Cyanide Complexes on IonExchange Resins Containing Different Functional Groups”, G C. Lukey, J. S. J. Van Deventer, R L. Chowdhury, D C. Shallcross, S T. Huntington, C J. Morton, 2000, Reactive and Functional Polymers, Volume 44, Issue 2, Pages 121-143. 13. Johns et.al., 1995 “Elution and Electrowinning of Gold from Gold-Selective Strong-base Resins” M.W.Johns, D, PJ Conradie, RJ Fowles, 1995, Hydrometallurgy, Volume 37, Pages 349-366.
12
ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
BIO-OXIDATION
ALTA Free Library www.altamet.com.au
BIOLEACHING â&#x20AC;&#x201C; PROBLEMS FOR THE MINING STATE
By
Carla Zammit1, Francoise Sandow 1, Lesley Mutch1, Helen Watling2 and Elizabeth Watkin1 1Parker Cooperative Research Centre for Integrated Hydrometallurgy Solutions, Curtin University of Technology, School of Biomedical Sciences, Australia 2
Parker Cooperative Research Centre for Integrated Hydrometallurgy Solutions, CSIRO Minerals Down Under Flagship, Australia
Presented by
Carla Zammit carla.zammit@postgrad.curtin.edu.au
CONTENTS 1.
INTRODUCTION
2
2.
HUNTING FOR MICROBES
4
3.
CONCLUSION
6
4.
OUR GROUP
6
5.
ACKNOWLEDGEMENTS
6
6.
REFERENCES
6
ALTA Free Library www.altamet.com.au
1. INTRODUCTION Bioleaching is the process by which microorganisms catalyse the breakdown of mineral sufides; it is an ancient technique with the first reports of crude processing systems dating as far back as AD 166 on the island of Cyprus. However, not until the 1940’s was it realised that microorganisms were involved in the process. In 1959 the first commercial bioleaching operation began at the Bingham Canyon Mine, in Utah, USA. Since that time bioleaching has been an expanding industry used globally as an economic way of processing low-grade and difficult to process ores from their mineral sulphides [1, 2]. 1.1
ADVANTAGES OF BIOLEACHING
As the world’s reserves of assessable high-grade ore diminish it is essential to develop systems for the economical extraction of metals from low-grade ores. This is the major driving force for the use of bioleaching; as it entails relatively low set-up and maintenance costs. As bioleaching microorganisms accelerate the leaching of sulfide minerals, ore which may have been categorised as low quality or too difficult to process by traditional methods may be economically extracted by bioleaching. This holds particularly true for high sulfide containing ores, but also means that secondary ores, waste ores and residue ores from old mine sites can also be viably processed. From an environmental point of view bioleaching is advantageous as gaseous emissions associated with smelting operations are avoided. Additionally, as the microorganisms involved in bioleaching fix carbon dioxide from the atmosphere, they may potentially help to reduce nett carbon emissions. 1.2
BIOLEACHING PROBLEMS – WESTERN AUSTRALIA
Bioleaching has been used in all corners of the globe to leach a large range of materials, including; gold, silver, copper, cobalt, zinc, lead, nickel and uranium. Most notable; it is estimated that 20% of the worlds copper production comes from bioleaching [1]. However, in Western Australia only two out of the significant 1032 [3] operational mines use bioleaching; Whim creek (copper) and Wiluna (gold). In 2007-08 the Western Australian mining industry was worth $58.66 billion, accounting for 83% of the state’s merchandise exports and 40% of the country’s merchandise exports [3]. So the question stands as to why Western Australia has not been receptive to bioleaching technologies.
The answer to this question lies with the microorganisms that are used for bioleaching. Bioleaching microorganisms are ubiquitous, found naturally in the environment; an exposed ore body will often support its own population of microorganisms. Bioleaching microorganisms have been found in environments that representing extreme environmental systems with high heat, high levels of metal and low pH, such as; acid mine/rock drainage site, volcanic sediments and hot springs. Although these microorganisms demonstrate resilience to the extremes of life, in none of these environments are the microorganisms exposed to high levels of salt, hence these microorganisms can be inhibited -1 by levels well below that of seawater (35 gL NaCl).
This low tolerance to salt has greatly affected the use of bioleaching in Western Australia and Chile. Salt can be introduced to bioleaching systems via three independent instances: i.
during bioleaching there is a build-up of salt due to the dissolution of gangue; this salt accumulates within the system, inhibiting the growth of bioleaching microorganisms;
ii.
large volumes of fresh water required for bioleaching may not be accessible and as such desalinization plants are installed or saline water is used (a common problem in Chile and Western Australia);
iii.
chloride ions can be added to some forms of acid leaching (non-biological) in order to speed up the leaching process [4, 5].
As Western Australia suffers from both gangue and difficult to access fresh water, the inability of traditional bioleaching microorganisms to withstand salt is a major problem for the use of bioleaching in Western Australia. The aim of this study was to find some microorganisms which would be capable of bioleaching under high levels of salt, for possible use in the Western Australian and Chilean mining industries.
2 ALTA Free Library www.altamet.com.au
1.3
SOLUTION
Fortunately, Western Australia is dotted with environments that are low in pH, and contain high levels of metals and high levels of salt. In many agricultural areas saline ground waters are rising due to the clearing of deep-rooted vegetation [6]. When these waters approach the surface a condition called â&#x20AC;&#x2DC;dry land salinityâ&#x20AC;&#x2122; occurs, thus causing the salinisation of the soil and above ground water supplies (figure 1). One solution used to address dry land salinity is to drain the groundwater. Drains are drilled to bring the groundwater to the surface in a controlled manner. The water is then drained into rivers or lakes or evaporated off in man made channels [7]. In the south of Western Australian the saline groundwater typically has a pH 3 to 4.5, high levels of heavy metals analogous to acid mine drainage systems and the perfect conditions to support the growth of salt loving bioleaching microorganisms.
Figure One: Arial view of salt lakes in Western Australia The acidic saline lakes in Western Australia are very unique. Unlike acidic saline lakes around the world, these lakes are not formed by volcanic or hydrothermal waters. The majority of saline lakes around the world are neutral to alkaline, in Western Australia the pH can be as low as pH 1.5. The process by which these lakes are formed is not well understood; alkaline and acidic lakes can be separated by only a couple of meters.
3 ALTA Free Library www.altamet.com.au
2. HUNTING FOR MICROBES Sixteen different sites where chosen for the investigation (figure 2). These were chosen based on their geochemical characteristics; low in pH, high levels of metal (in particular iron and sulfur) and high levels of salt. These environments were sampled and investigated for the presence of bioleaching microorganisms.
B
A
Figure Two: A. Run-off from an acidic saline drain located in the south of Western Australia. The run-off was an orange yellow colour caused by ferric rich iron sediments and is similar to what is seen in acid mine drainage systems where it is referred to as “yellow boy”. B. An acidic saline drain where the samples investigates in this study were isolated from.
In the laboratory samples were enriched to encourage the growth of microorganisms that could be used in bioleaching. From the original sixteen samples three contained microorganisms that were able to oxidize iron (II), a key trait when looking for bioleaching microorganisms. From these three samples one was chosen to study in further detail based on its superior ability to oxidize iron.
Figure Three: Growth of the environmental mixed culture in media with added NaCl: ▲ 0, 7, 12.5, ∆ 20, 30 and ≥40 gL-1. Adapted from [8]
4 ALTA Free Library www.altamet.com.au
The selected sample was tested for growth and iron (II) oxidation under different levels of NaCl (figure 3); the sample was found to grow optimally at 12.5 gL-1 NaCl, and able to tolerate levels of salt up to 40 gL-1 NaCl. This was the first report of a bioleaching microorganism which grows in such acidic conditions preferentially growing under high levels of salt. Additionally, it gives promise that other microorganisms may be present in our environment that can be used for bioleaching under the high levels of salt experienced in Western Australia.
Figure Four: Phylogenetic tree of the four different microorganisms isolated from an acidic saline drain in Western Australia.
Group I
Group II
The microorganisms identified in the acidic saline drain sample fell into two groups. One of the groups of microorganisms identified was similar to Thiobacillus prosperus (figure 4), that has only ever been isolated from a volcano in Italy [9]. The base of this volcano, where this microorganisms was isolated, was exposed to geothermally heated seawater, had a pH of 6.5 and was rich in minerals. This ‘Italian’ microorganism is able to oxidize -1 iron in the presence of up to 35 gL NaCl, but preferentially grows in the absence of NaCl.
Percentage number of microorganisms
Molecular techniques were used to study how the population changed under different levels of NaCl. The two groups of microorganisms identified in the sample were present in varying proportions under different levels of NaCl (figure 5). What this means, in terms of application, is that this consortium of microorganisms is pliable, and will naturally change to suit the level of salt within the bioleaching operation.
Figure Five: Molecular analysis of the two groups of microorganisms isolated from the acidic saline drain. The graph shows the percentage of each group of microorganisms under different levels of NaCl. Group I, Group II.
-1
Concentration of NaCl (gL )
Research is now underway to individual isolate each of the strains in our mixture. The use of laboratory scale bioleaching columns under high levels of salt are currently being used to investigate if these microorganisms can be used for commercial bioleaching operations. Additionally, the mechanisms of salt tolerance of these microorganisms will also be of interest. Looking at the differences between the Thiobacillus prosperus from Italy and the ‘Thiobacillus prosperus’ like strain from Western Australia may help to shed some light on how these microorganisms tolerate such high levels of NaCl.
5 ALTA Free Library www.altamet.com.au
3. CONCLUSION To date, this study has made a significant contribution to the development of bioleaching systems to be used in Western Australia. Additional research will assist the application of this research on an industrial scale. Although the bioleaching process has a long history, it is still not fully understood. Research into the microorganisms within bioleaching communities, their interaction with each other and their environment is necessary if the optimisation of the bioleaching process is to be realised.
The most interesting part of this research project is the approach; to look at an industrial problem and then look to the environment for a solution. We now have the technology to upscale what the environment does and to apply it to our industry, encompassing all of the 4.5 billion years of evolution into developing simple solutions for seemingly complex problems. 4. OUR GROUP Our Environmental Microbiology Group is located at Curtin University of Technology, Perth. The research that our group undertakes involves bioleaching, acid sulfate soils and root nodule bacteria. We work closely with the Parker CRC for Integrated Hydrometallurgy Solutions and CSIRO Minerals. If you are interested in our group and require further information please contact Dr. Elizabeth Watkin (E.Watkin@curtin.edu.au) or visit our website: http://www.biomed.curtin.edu.au/EnvironMicro/. 5. ACKNOWLEDGEMENTS CZ would like to thank the Australian Government for the Australian Postgraduate Award, the Minerals and Energy Research Institute of Western Australia and Curtin University of Technology. The support of the Parker CRC for Integrated Hydrometallurgy Solutions (established and supported under the Australian Governmentâ&#x20AC;&#x2122;s Cooperative Research Centres Program) is gratefully acknowledged. Brad Degens for help with the locations to sample and providing unpublished reference data. 6.
REFERENCES
1. Watling, H., R, The bioleaching of sulphide minerals with 7. Stewart, B., K. Strehlow, and J. Davis, Impacts of deep emphasis on copper sulphides - A review. open drains on water quality and biodiversity of Hydrometallurgy, 2006. 84(1-2): p. 81-108. receiving waterways in the wheatbelt of Western Australia. Hydrobiologia, 2009. 619: p. 103-118. 2. Watling, H.R., The bioleaching of nickel-copper sulfides. Hydrometallurgy, 2008. 91(1-4): p. 70-88. 8. Zammit, C.M., et al., The Characterization of Salt 3. Western Australian Mineral and Petroleum Statistics Tolerance in Biomining Microorganisms and the Search Digest 2007-08, S.o.W.A. Department of Industry and for Novel Salt Tolerant Strains. Advanced Materials Resources, Editor. 2008. Research, 2009. 71-73: p. 283-286. 4. Lu, Z.Y., M.I. Jeffrey, and F. Lawson, The effect of 9. Huber, H. and K.O. Stetter, Thiobacillus prosperus sp. chloride ions on the dissolution of chalcopyrite in acidic nov., represents a new group of halotolerant metalsolutions. Hydrometallurgy, 2000. 56(2): p. 189-202. mobilizing bacteria isolated from a marine geothermal 5. Kinnunen, P.H.-M. and A.P. Puhakka, Chloride-promoted field. Archives of Microbiology, 1989. 151(6): p. 479leaching of chalcopyrite concentrate by biologically485. produced ferric sulfate. Journal of Chemical Technology & Biotechnology, 2004. 79(8): p. 830-34. 6. Dickson, B. and A. Giblin, Features and effects of some acid-saline groundwaters of southern Australia. Chinese Journal of Geochemistry, 2006. 25(0): p. 227-227.
6 ALTA Free Library www.altamet.com.au
Tank Bio-Oxidation Agitation Systems
By B. Gigas & R. Kehn
SPX Flow Technology â&#x20AC;&#x201C; LIGHTNN Rochester, New York USA
Presented by
Bernie Gigas bernie.gigas@spx.com
CONTENTS
1. 2. 3. 4. 5. 6. 7.
ABSTRACT INTRODUCTION GENERAL SOLID SUSPENSION GASSED SOLID SUSPENSION SPECIAL MODIFICATIONS FOR BIO-OXIDATION CONCLUSIONS REFERENCES
ALTA Free Library www.altamet.com.au
Page 2 Page 3 Page 3 Page 9 Page 12 Page 13 Page 13
1. ABSTRACT The requirement for solid suspension in mining refinery applications is very common and agitation design is reasonably well understood by many agitator vendors, engineering contractors and mine refinery operators. This paper presents a summary of the general principles used for both agitator and vessel feature design for solid suspension duty. In addition, this paper presents basic design features required for successful slurry suspension in gassed operations such as aerated leach or sulfur dioxide reduction. Finally, the paper discusses a novel agitator design for aerated leach. This design was specifically developed to reduce both the compressed air and agitator power requirement for bio-oxidation applications in gold leach bio-oxidation. Development and characterization of this approach was carried out in 500 liter test tanks water trials followed by in-process pilot testing in both water and 3 slurry at 21 m [1]. Results show that the proposed LIGHTNIN A315-A340 combination is superior to other agitator designs with respect to both required agitator power and compressed air feed requirements.
Page 2 ALTA Free Library www.altamet.com.au
2. INTRODUCTION Bio-oxidative leach of gold ore bodies is an industrially applied extraction process. It is used prior to cyanide leach as an alternative to pressure oxidation autoclaves (POX). A typical bio-oxidation plant requires several in-series stages of flat bottom, cylindrical slurry processing tanks that are both aerated and mechanically agitated. Aside from the typical leach tank solid suspension requirement, the bio-oxidation process is based on the use of generally three bacteria â&#x20AC;&#x201C; acidithiobacillus ferrooxidans, acidithiobacillus thiooxidans, and leptospirillum ferrooxidans. These bacteria require oxygen and the gas feed requirements for bio-oxidation leach tanks are not trivial and significantly complicate the agitator process design. We therefore have two primary process requirements that need to be addressed by the agitator designer: 1. Uniform solid suspension to keep solids from accumulating and to provide predictable residence time distribution. 2. Gas dispersion and mass transfer of the oxygen containing feed stream. There are additional design parameters such as blending, etc. â&#x20AC;&#x201C; these are relatively minor and are not discussed in this paper. For gassed slurry applications, the requirements for solid suspension and gas handling are interlinked and an agitator design that focuses primarily on the gas dispersion requirements will possibly fail to address the diminished solid suspension performance characteristics of the impeller(s) under gas load. Successful bio-oxidation reactors use the LIGHTNIN A315 impeller, Figure 1:
Figure 1: LIGHTNIN A315 Impeller LIGHTNIN released the A315 impeller in 1987 for processes requiring high efficiency axial flow in addition to strong gas handling capability. A key component to a successful impeller design for the bio-oxidation process is the ability to maintain axial flow and therefore solid suspension capability even under relatively high gas loads. While the use of a single A315 impeller is still the standard for bio-oxidation leach tanks, recent developments have shown that the addition of an upper LIGHTNIN A340 impeller (Figure 2) can improve the overall efficiency of the agitated leach tank in terms of both compressed gas consumption and total agitator power [1]. This finding is similar to the use of the A340 impeller in copper pressure oxidation (POX) described in 2006 [2].
Page 3 ALTA Free Library www.altamet.com.au
Figure 2: LIGHTNIN A340 Impeller While the focus of this paper is on gold ore bio-oxidation agitator design, the concepts discussed apply to any gassed and agitated slurry tank in gold, copper, or nickel ore processing.
3. GENERAL SOLID SUSPENSION General solid suspension has been, and continues to be a deeply studied area of agitator design. Probably the best known and most widely used correlation for agitator solid suspension design was developed and first published by Zwietering in 1958 [3] and is presented in Equation 1. Note that this correlation determines NJS, the agitator speed required to just suspend all solids off the vessel bottom. It does not predict the power required to ‘uniformly’ suspend the solids in the vessel.
N JS S
0.1
g s l c l
0.45
X 0.13 d p0.2 D 0.85
Eq 1
Where, NJS S gc s l X dp D
= Just Suspended Agitation Speed [RPS] = Dimensionless Constant = Kinematic Viscosity = Gravitational Acceleration = Solid Density = Liquid Density = Solids Mass Fraction = Mass Mean Particle Diameter [m] = Impeller Diameter [m]
The challenges faced in slurry agitation designs using this correlation are summarized as follows:
Page 4 ALTA Free Library www.altamet.com.au
3.1 Particle Size Distribution Zwietering’s work was done with monodisperse (sand) slurry at relatively low solids concentration (2% or so). Real world applications are polydisperse and will require a judgment call on the design particle size. Choose P100 to be safe and risk overdesign, choose P80 and have potential solids accumulation in case of a wide particle size distribution (PSD).
Figure 3: Examples of Solids Particle Size Distribution The examples in Figure 3 illustrate this point. The PSD on the left has a P80 of approximately 90 micron and a P100 of 250 micron. In this case, designing to P80 uniform and P100 off bottom is a reasonable approach since an agitator designed for uniform suspension of 90 micron particles is likely to be able to suspend 250 micron particles off the tank floor. In the case of the PSD on the right hand side of Figure 3, P80 is approximately 100 micron and P100 is approximately 500 micron. Here, a design using P80 may lead to solids accumulation on the tank floor. In mining slurries, it is not uncommon to find a P80 less than 300 micron with a P100 in excess of 1 mm and correct selection of the design PSD is of more than academic interest. In order to avoid oversized agitators, a compromise solution may require the intermittent milking of large solids from the tank bottom.
3.2 Settling Behavior, Viscosity & Scale Up Zwietering’s work is based on free settling solids, i.e., applications where the slurry is not dominated by particle-to-particle interactions. Many mining slurries are hindered (dominated by particle-toparticle interaction) and may even be viscous. There are few reliable references published for hindered slurry agitation design and vendor experience plus test work is recommended to develop a suitable mixer selection [6]. Scale up from test scale to full scale is generally done via a power (P) per volume (V) relationship per Equation 2:
V PFullScale PSmallScale FullScale VSmallScale
x
Eq 2
Here, the exponent x can vary from 0.80 to 0.95 depending on the exact nature of the application. Once again, vendor experience is critical.
Page 5 ALTA Free Library www.altamet.com.au
3.3 Suspension Uniformity As implied above, uniform solid suspension requires an adjustment on NJS. While most agitator vendors will have experience in design for uniform solid suspension, the definition of the word ‘uniform’ requires some discussion. Oldshue [4] defines uniform solid suspension as the point at which the addition of more mixer power does not appreciably change the solids suspension, meaning neither the height of the suspension nor the particle size distribution can be further improved by the use of higher agitation power. While this definition may not be very satisfactory from a user’s perspective, it does acknowledge the physical limitations found in tall tanks and processes that have a wide particle size distribution and/or a high density difference between the liquid and solid phase. In practical terms, Oldshue’s definition of uniformity recognizes that there is a vertical classification of particle size distribution and solids concentration particularly in full scale mining applications. Full scale field tests (Figure 4) in 14 m diameter by 30 m tall alumina precipitators with multiple flights of axial flow impellers (non draft tube) have shown top-to-bottom that solids concentration uniformity better than 5% is achievable.
Figure 4: Full Scale Solids Uniformity Tests No particle size distribution data was available from this work but small scale tests in a 2.5 m diameter tank have confirmed that particle size classification does occur. In large scale slurry tanks this means that there is virtually always a thin clear layer at the top of the fluid that contains few solids. The use of simple overflows at the top of the tank is therefore not recommended as this requires a specific agitator design to avoid eventual sanding in of the lowest impeller. In order to assure removal of all possible size fractions either bottom draw off or riser pipes are recommended as the simplest solution. Riser pipes should be located as close as possible to the bottom but above the ultimate settled solids bed depth in case of power failure. Refer to Figure 5.
Tank Inlet Δh
Settled Bed Depth Figure 5: Riser Pipe General Layout
Page 6 ALTA Free Library www.altamet.com.au
Tank Outlet
There are a number of design criteria that must be considered in riser pipe design: 1. The design velocity should be between 6-10 times the settling velocity of the largest particle to be transported. 2. The pipe diameter should be such that the pipe Reynolds number is at least 2,000 at the lowest flow condition. This assures full turbulence and eliminates the possibility of a reverse flow (downward) inside the riser pipe near the wall. 3. The pressure drop, Δh (Figure 5) of the riser pipe should be calculated at the highest design flow rate to assure sufficient head is available. All necessary equations can either be simply derived or taken from Perry’s Chemical Engineer’s Handbook [5]. Assure that the total tankto-tank head requirement includes the launder drop as well. 4. The addition of air into the riser pipe as a means of transport assist in cases of low flow and/or low head capacity is not recommended as this practice changes the density of the slurry in the riser pipe. This reduces the ability of the system to transport the larger solids fraction out of the tank.
3.4 Tank – Mixer Geometry One of the parameters in Zwietering’s equation that receives the most scrutiny is the ‘S’ value. While this is often referred to as an impeller constant, it is neither constant nor a function of the impeller alone. The system geometry affects solid suspension design (and the S-value) via the following major contributors: 1. Impeller geometry. High efficiency axial flow impellers are best suited for solid suspension applications. LIGHTNIN released the A310 (Figure 6) in 1981 and other vendors have followed suit with similar designs.
Figure 6: LIGHTNIN A310 Impeller These impellers have specific blade shapes, bends and hub attachment angles to produce truly axial flow at minimum invested power. As impellers get bigger and thicker, it can be difficult to fabricate the exact blade shape and the hub angle is sometimes steepened to produce a higher power number impeller. Both of these changes can have a significant impact on the performance of an otherwise well characterized standard impeller design. Specific validation may be required on a case by case basis. 2. Impeller diameter to tank diameter ratio (D/T). Ideally, this should be small to allow for the selection of a lower torque and hence lower cost agitator. In practice, typical slurry applications are designed at a lower limit of 0.3 D/T for free settling solids and 0.35 for hindered slurries. Smaller impellers tend to struggle with generating sufficient up flow near the wall to produce reliable solid suspension performance.
Page 7 ALTA Free Library www.altamet.com.au
3. Impeller off bottom to impeller diameter ratio (C/D). Figure 7 shows a plot of the relative power required for solid suspension as a function of impeller off bottom. Note Region I where there is relatively little effect on required power as the impeller is moved off bottom. There is however a distinct inflection point where the required agitator power is a strong function of impeller off bottom distance. This is referred to as Region II.
Region II
Region I
P/V
C/D Figure 7: C/D Impact on Required Mixer Power Since mixer hardware design and cost for mining tanks is often driven by shaft design; there is an incentive to make the shaft as short as possible in order to reduce capital cost. This must however be done in conjunction with a careful analysis of the solid suspension Region to assure successful field operation. Incorrect off bottom dimensioning of the lower impeller is one of the more common application errors in slurry tanks, there can be significant differences in solid suspension capacity of an impeller between 0.5, 1.0 and 1.5 impeller diameter off bottom clearance.
Page 8 ALTA Free Library www.altamet.com.au
4. GASSED SOLID SUSPENSION The preceding section covered solid suspension in typical slurry operation. These applications may have complex particle size distribution and rheology but impeller performance is generally well understood, testing protocols are established and there are a large number of installed tanks to reference for design purposes. All of these parameters are still important for gassed slurry applications but the design must now also consider the impact of the added gas flow to the overall system capacity. All impellers have limited gas handling capabilities. As gas is applied to an impeller, the local effective fluid density decreases and the impeller’s ability to impart momentum to the surrounding fluid is diminished. This phenomenon is easily observed via the reduced power draw of an impeller in gassed versus ungassed operation and is expressed as the k-Factor, or ratio of gassed power draw to ungassed power draw per Equation 3:
k
Pgassed Pungassed
Eq 3
The k-Factor of an impeller is primarily a function of its geometry and the aeration number, NAE or ratio of applied gas rate and impeller primary pumping capacity per Equation 4:
N AE
Qgas Qimpeller
Eq 4
Every impeller type has an upper limit of aeration, at which point the impeller is flooded, per Figure 8. The picture on the left shows a well dispersed condition, the picture on the right shows a flooded impeller:
Figure 8: Well Dispersed and Flooded Gas Dispersion This flood point aeration number is a function of impeller geometry, applied power per volume and sparge device used. For high efficiency type impellers like the A310 (Figure 6), the aeration number limit can be as low as 2%, making it less than ideal for gassed slurry applications. Higher solidity axial flow impellers like the A315 (Figure 1) have a gas handling capacity in excess of 8% and are better suited to this type of application, as demonstrated in bio-oxidation.
Page 9 ALTA Free Library www.altamet.com.au
From a design perspective, an agitator analysis for gassed slurry operation based solely on k-Factor and aeration number is too simplistic. It is critically important to review an impellerâ&#x20AC;&#x2122;s flow performance under gas loading that is below the flooding limit. Axial flow impellers are well suited to solid suspension applications because their downward flow profile sweeps the tank bottom to pick up and suspend solids. Figure 9 is an idealized discharge pattern off an A315 impeller in ungassed operation:
Figure 9: Idealized A315 Flow Discharge This discharge pattern changes in gassed operation. The upward momentum of the gas causes a radial deflection of the discharge from the impeller per Figure 10 until the impeller discharge becomes fully radial at the flood point.
Fluid
Gas Figure 10: Discharge Angle Alteration Under Gas Load This radial deflection of the impeller discharge begins to occur as soon as gas is added to the process, even at low aeration numbers well below the flood point. As a result, the impellerâ&#x20AC;&#x2122;s ability to suspend solids is reduced and the transition point from Region I to Region II is shifted to a lower off bottom distance, C/D. If the impeller location is poorly chosen, the power draw, gas dispersion and mass transfer capacity may still exist but solids may accumulate on the tank bottom as shown in Figure 11:
Page 10 ALTA Free Library www.altamet.com.au
Figure 11: Settled Solids in Gassed Suspension
In summary, gassed slurry operation requires the following considerations: 1. Total gas flow for hot slurries must consider the increased gas flow rate from water vapor. 2. The choice of impeller type must consider the gas load relative to the impellerâ&#x20AC;&#x2122;s ability to disperse the gas and still maintain suspension. 3. k-Factor and aeration number analysis must be augmented by a correct impeller off bottom location to assure suspension. Off bottom distances of 0.5 to 0.75 impeller diameters are common in gassed slurry service.
Page 11 ALTA Free Library www.altamet.com.au
5. SPECIAL MODIFICATIONS FOR BIO-OXIDATION BIOX速 belongs in a special category of aerated and mechanically agitated slurry applications. It is characterized by a 20% by weight slurry with high superficial gas velocity to effect the necessary mass transfer and keep the ferrooxidans and thiooxidans bacteria alive. BIOX速 is operated at approximately 40 oC. This process to treat refractory gold ores prior to conventional cyanide leach was developed by Goldfields in South Africa and was placed into commercial operation some 20 years ago. Virtually all BIOX速 installations worldwide use the LIGHTNIN A315 impeller due to its gas handling and solid suspension capabilities. Recent efforts have focused on improving the overall efficiency of the BIOX速 plant, both from a capital and operating expense point of view. Since the current agitator arrangement with single lower A315 impeller requires a significant investment of agitator power for gas handling and solid suspension, a method was sought to reduce the fed gas requirement of the process and thereby also reduce the agitator power required. The results of laboratory and 21 m3 pilot plant studies were first reported by Van Deventer and Gigas in 2009 [1]. The test sequence used the typical lower A315 impeller in conjunction with an upper A340 impeller near the surface to draw gas in from the head space, Figure 12:
Figure 12: A315 (left) & A340 (right) Impellers The addition of the upper A340 impeller improved mass transfer capability of the reactor significantly on both the laboratory and pilot plant scales: 1.8
1.6
A315 A315 - A340
1.4
Relative kLa
1.2
1
0.8
0.6
0.4
0.2
0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
Relative Gas Rate
Figure 13: Relative Mass Transfer Capability A315 versus A315-A340 Note that Figure 13 data is from testwork in the 21 m3 pilot scale tank.
Page 12 ALTA Free Library www.altamet.com.au
The above data is for water trials. Later process tests using slurry were also completed and showed excellent results with positive effect on mass transfer and no effect on solid suspension. The conclusion of the test program is quoted directly from Van Deventer and Gigas’ paper: “From a BIOX® process view point the combination LIGHTNIN impeller configuration is superior to any other mixing system tested in the BIOX® process, achieving a reduction power per unit volume as well as reducing the volume of air required to maintain the same sulphide sulphur oxidations without compromising any bacterial activity or gold recovery.” [1] The next step in the development of this promising alternative design is to seek a suitable full scale beta test site.
6. CONCLUSIONS Solid suspension is one of the most common applications for industrial mixers. Impeller configurations and their performance limits are reasonably well understood and applied in typical mining slurry applications. We have presented general guidelines for geometry, impeller choice and analysis criteria for ungassed slurry design. The addition of sparged gas adds additional complexity to the design requirements, particularly for impeller choice and off bottom distance. The special case of agitator design for bio-oxidation was presented. Here, the lower impeller is designed primarily for gas dispersion and solid suspension while a second up pumping impeller near the surface is used to draw in air from the head space. This configuration has shown the ability 3 to reduce both the sparged gas rate and the applied mixer power in 21 m pilot studies.
7. REFERENCES [1]
R. Van Deventer and B. Gigas, “New Biox® agitator concept test work”, Advanced Materials Research, Vols. 71-73 (2009), pp 461-464
[2]
B. Gigas and J. Wilmot, “Computational Analysis and Commercial Demonstration of Improved Pressure Leach Vessel Agitator Design”, ALTA Conference Proceedings, Perth, Australia, 2006
[3]
T. Zwietering, “Suspending of solid particles in liquid by agitators”, Chem. Eng. Sci., 8, 244; 1958
[4]
J. Oldshue, “Fluid Mixing Technology”, Chemical Engineering McGraw-Hill Publications, New York, 1983
[5]
R. Perry and D. Green, editors, “Chemical Engineers’ Handbook”, McGraw-Hill Publications, New York
[6]
R. Kehn et al., ”Agitator Design For Large Hindered Settling Slurry Tanks”, ALTA Conference Proceedings, Perth, Australia, 2009
Page 13 ALTA Free Library www.altamet.com.au
ALTA 2010 GOLD ORE PROCESSING SYMPOSIUM
PRESSURE OXIDATION
ALTA Free Library www.altamet.com.au
ACID BRICK LINING OPTIMIZATION FOR THE TWIN CREEKS MINE
By
Kevin Brooks Koch Knight LLC, USA
Presented by
Kevin Brooks kevinbrooks@kochknight.com
CONTENTS
1. 2. 3. 4. 5.
INTRODUCTION LINING DESIGN MATERIALS SELECTION MAINTENANCE PROCEDURES CONCLUSION
ALTA Free Library www.altamet.com.au
2 2 5 6 7
1. INTRODUCTION 1.1 HISTORY The Twin Creeks Mine is located near Golconda, Nevada, USA, in the Carlin Trend gold production region. The processing capability at the mine includes a pressure oxidation plant to process the gold sulfide ores. The pressure oxidation plant utilizes two pressure oxidation trains. The acid brick lined vessels in each train include an autoclave, low pressure flash vessel, high pressure flash vessel, low temperature heater and high temperature heater. The two trains share one slurry cooler vessel. The Twin Creeksâ&#x20AC;&#x2122; design was a scale up of a similar process in use at the nearby Lone Tree Mine. At the time, these were the largest acid brick lined vessels for pressure oxidation in service. The autoclaves are a four (4) compartment configuration with a 5.1 m process diameter and a length of 20.1 m. The support vessels are of a similar scale with low pressure flash vessel dimensions of 5.1 m diameter by 4.9 m tangent to tangent and high pressure flash vessel dimensions of 6.1 m diameter by 4.3 m tangent to tangent. The plant construction was executed in two phases. The linings for the first phase were installed in 1996 with start-up in 1997. The second phase was completed in 1997 with start-up in early 1998. 1.2 IMPROVEMENTS During start-up and the early days of operation several unanticipated issues were encountered with the scale up of the lining systems. Most of these issues were in the autoclave. With input from operations personnel, consultants and the brick lining manufacturer, the lining material, lining configuration and maintenance practices were improved to provide reliable and predictable service. At this time the brick lining is no longer a constraint on operations and is maintained on regularly scheduled shutdowns for mechanical preventative maintenance.
2. LINING DESIGN 2.1 LINING CONFIGURATION The autoclave linings consist of a membrane to protect the steel shell from corrosion overlaid with a brick and mortar system to provide thermal and physical protection for the membrane system. The membrane in Phase I is a 6 mm thick lead lining. In Phase II a PYROFLEXÂŽ DUPLY membrane with an overall thickness of 12 mm was selected to reduce the both the project cost and schedule duration. The brick linings consist of one brick layer with a 114 mm thickness and two layers at 76 mm thick. In Phase I the 114 mm thick bricks were installed in the first course against the membrane. In Phase II the 114 mm bricks were installed in the mid course. The dimensions at the back of the 114 mm thick brick are 228 x 114 mm in comparison to 228 x 152 for the 76 mm thick brick. The larger surface area at the back of the 76 mm thick brick made it easier to prop the brick tightly to the membrane until the mortar cured.
Figure 1: Phase I and Phase II Autoclaves
2
ALTA Free Library www.altamet.com.au
2.2 COMPARTMENT WALLS AND BAFFLES The autoclaves were initially constructed with brick compartment walls and brick baffles at the sidewalls. Titanium baffles were fastened to the center of the compartment walls. The configuration of the compartment walls with all of the brick in compression proved to be a reliable design. The autoclaves currently still operate with brick compartment walls. Over time it was determined that the brick baffles were not the best method of construction. The compartment walls were constructed 343 mm thick with alternating header and stretcher brick courses (Figure 2). The walls were inset into a chase at the vessel walls with an expansion joint built into the chase. The top of the wall incorporated a reverse arch to impart a compressive load on the top of the wall. Desanding holes were built into the bottom of the walls to allow for some underflow.
Figure 2: Brick Compartment Wall The baffles attached to the compartment walls were titanium anchored with rods that extended through the brick compartment walls. The original brick baffles were carefully designed to interlock the courses of brick and tie the brick back into the sidewall lining (Figure 3). These efforts were unable to compensate for the tendency of the compressive forces at the face of the sidewall to pinch the base of the brick baffle and shear them off after a few cycles in operation.
Figure 3: Original Brick Baffle Configuration The brick baffles at the sidewalls were replaced with titanium baffles with anchors imbedded in the brick lining (Figure 4). There were concerns that the higher thermal expansion of the titanium would cause fracturing of the brick at the anchors. This risk was managed by minimizing the size of the anchors imbedded in the brick. The titanium baffles have been in service for several years and have required very little maintenance. Confidence in the titanium baffle design has allowed for an increase in the baffle size and improved mixing.
3
ALTA Free Library www.altamet.com.au
Figure 4: Titanium Baffle Configuration 2.3 NOZZLE LINING The nozzle linings were constructed with a weld overlay, acid brick lining and titanium inserts (Figure 5). The acid brick linings were designed to minimize the amount of heat conducted to the membrane lined shell. Nozzle brick thicknesses varied by nozzle size from 140 to 190 mm.
Figure 5: Typical Nozzle Detail The scale up of the autoclaves included a requirement for larger nozzles. The largest nozzles were the agitator nozzles with a diameter of 2000 mm. These were over 30% larger than previous autoclave nozzles. The vessel configuration also included several large â&#x20AC;&#x153;hillsideâ&#x20AC;? nozzles. As with the steel vessel, brick lining stresses are concentrated at the nozzles. The larger nozzles magnified the higher stresses from the selection of more brittle bricks and mortars. The standard procedure of grouting the inserts in place also contributed to higher lining stresses. The elliptical shape of the compression rings created further issues due to the irregular distribution of the stresses. These conditions lead to spalling and additional maintenance requirements at the nozzle compression rings. Maintenance at the nozzle openings was reduced by the change in lining materials and configuration to reduce the lining stresses. Smaller pieces of brick were used in order to incorporate a greater number of mortar joints into the compression rings. This reduced the stiffness of the compression rings further reducing the stress concentration at the nozzles. For the largest nozzles, two compression rings were installed. The titanium inserts were not grouted in place so that the thermal expansion of the titanium did not contribute to the lining stress.
4
ALTA Free Library www.altamet.com.au
3. MATERIALS SELECTION 3.1 MEMBRANE The Phase I autoclave was lined with a panel lead lining system. The lead panels installed were approximately 300 x 600 mm. A 3 mm thick ceramic paper was applied over the lead lining before the brick lining installation. Portions of the lead lining were exposed during early repairs to the brick lining. It was observed that the center of the exposed lead panels would separate from the shell by up to 6 mm. This was addressed by drilling a hole at the center of the panel and using a vacuum to pull the panel back to the steel shell. The Phase II autoclave was lined with a PYROFLEX® membrane. This is an asphaltic rubber polymer formulated for use as a membrane behind acid brick linings in higher temperature acid service. The PYROFLEX® membrane was overlaid with a secondary DUPLY™ membrane of glass reinforced resin. During operation small amounts of the membrane were extruded at the nozzle insert flanges due to the thermal expansion and plastic properties of the membrane. The displaced material allowed for the irreversible chemical expansion of the brick lining reducing the overall lining stresses. With the autoclaves operating at 227 °C, the membranes are exposed to shell temperatures of up to 106 °C during operation in the summer. Both membrane systems are in service as originally installed and have required no maintenance. The vapor zone nozzles were lined with weld overlay due to higher shell temperatures. The higher shell temperatures are due to heat conducted from the nozzle blinds which limit the insulation provided by the brick linings in the nozzles. The weld overlay has only required minor repairs, primarily at the nozzle shell intersection. 3.2 BRICK The initial brick selection was focused upon the acid resistance of the brick. The acid solubility and porosity were considered the key properties at the time. In order to provide the lowest acid solubility and porosity, the SiO2 content of the brick was increased. The increase in SiO2 created more glass in the brick body during the firing process. The mechanical properties of the brick were considered of secondary importance. The stresses created by the scale up of the vessel and increased nozzle sizes proved that the mechanical properties of the brick and brick/mortar composite were of equal importance. The initial brick linings installed had significant spalling. Studies of these issues showed that bricks with a lower modulus of elasticity (less brittle) and lower chemical expansion were needed to minimize the compressive stresses in the lining. Since the initial installation the brick formulations have undergone several generations of continuous improvement. Current practice is to install brick with properties that balance chemical resistance with the desirable mechanical properties to provide for long service life. Another area for improvement was the brick selection immediately below the agitators. This area tended to erode by abrasion from solids swirling underneath the agitators. A denser more abrasion resistant acid brick was incorporated into the lining in this area. This was accomplished while maintaining the needed insulating properties of the brick. 3.3 MORTAR The mortar installed in the original autoclave linings was a lower porosity silica filled potassium silicate. This mortar had been used extensively in prior autoclave installations. In the Twin Creeks autoclaves this mortar tended to soften and erode in the vapor zone area. This required pointing of the receded joints during regular outage intervals. A recently developed mortar was available at the time that used an alternate silicate based solution. This solution provided a dense mortar that had improved resistance to the steam in the vapor zone. As with the brick, the trade off for the improved resistance to the operating environment was undesirable mechanical properties. The new mortar created a very brittle brick/mortar composite causing considerable spalling of the brick lining in the vapor zone. The remaining choice for the vapor zone mortar was a lead oxide based mortar. Lead oxide mortars were known to perform well in higher pressure steam service but required additional special equipment and procedures to manage the health and environmental risks. Once the lead oxide mortar was installed the maintenance intervals for the vapor zone brick lining were greatly reduced.
5
ALTA Free Library www.altamet.com.au
The present lining configuration utilizes a potassium silicate mortar in the liquid zone and a lead oxide mortar in the service course of the vapor zone. Both materials perform satisfactorily.
4. MAINTENANCE PROCEDURES 4.1 THERMOGRAPHY During operation it is desirable to monitor the condition of the brick lining in order to predict maintenance requirements. Over time the lining can erode or be damaged by operational changes. The brick lining can tolerate losses in thickness of up to 25 mm of the 76 mm thickness of the brick service course. Further reduction in thickness risks the unkeying of the brick lining. With experience the maintenance personnel at Twin Creeks have found the best method to determine the condition of the brick lining during operation is thermography. Thinner spots in the brick lining show up as “hot spots” on the thermal image. The temperatures at these hot spots can be compared to thermal gradient calculations to determine the lining thickness as shown in Table 1. Through regular monitoring of the shell temperature, the rate of change in the lining thickness can be determined. Also significant events such as choke valve failures can be identified allowing for shutdown and repair before additional damage is caused. Table 1: Predicted Shell Temperatures Lining Thickness
Ambient Summer Temperature 38 °C
Ambient Winter Temperature 21 °C
Complete Lining 76+114+76
90 °C
80 °C
Lining Eroded by 38 mm 76+114+38
95 °C
85 °C
Lining Eroded by 76 mm 76+114+0
100 °C
90 °C
Since the initial construction of the pressure oxidation plant in the late 1990’s the resolution of thermographic equipment has improved significantly and high end units have become more affordable. The Twin Creeks operation currently uses a ThermaCAM® P60 by Flir Systems for infrared inspections. The lined vessels are surveyed weekly and the results compared to previous surveys. More detailed studies of areas of interest are done as needed. The thermographic equipment also is used to monitor other mechanical and electrical equipment throughout the operation. It has been used to predict replacement of bearings and transformers before unplanned failures occur. 4.2 MAINTENANCE SHUTDOWNS Maintenance shutdowns of the autoclave trains are scheduled at regular intervals to allow for inspection and replacement of wear parts such seals and agitator blades. To avoid lost production, the lining maintenance must be completed within the time interval allocated for these other activities. The large size of the Twin Creeks autoclaves can require longer periods of time to perform the lining maintenance. The preparations for entry and scaffolding requirements take up substantial portions of the time available to perform the work. The service course brick in the vapor zone require replacement approximately every 3 to 5 years. A complete vapor zone at Twin Creeks reline requires an extended shutdown beyond the time needed for normal maintenance. A strategy of partial replacement was developed to keep the lining maintenance within the standard shutdown time frame. When it was time to replace the vapor zone brick, the linings in compartments 1 and 3 were replaced on one planned shutdown. The remaining compartments 2 and 4 were replaced on the subsequent shutdown. This allowed for performance of the work without additional lost production. The added benefit was that part of the chemical expansion from the reline of the first two compartments could be relieved when the second portion of the vapor zone was replaced.
6
ALTA Free Library www.altamet.com.au
5. CONCLUSION Scale up of existing equipment requires careful study of the implications of the larger dimensions on the materials of construction. Minor maintenance issues on existing equipment will be magnified during scale up. Material selection requires careful consideration of all properties that could influence their performance. Well thought out maintenance strategies by experienced personnel become more important for larger equipment. Tasks that normally only take a few hours or days can take substantially more time. Special equipment and procedures may be needed to reduce maintenance schedules to acceptable intervals. The lessons learned over the past twelve years of operation of the pressure oxidation autoclaves at Twin Creeks have made important contributions to the knowledge required to construct, operate and maintain large acid brick lined autoclaves. The combination of proper materials selection, good designs and well thought out maintenance strategies will provide reliable service for the Twin Creeks operation and future projects. This experience can be applied to both the continued Twin Creeks operations and future pressure oxidation plants of a similar scale.
7
ALTA Free Library www.altamet.com.au
GOLD PRESSURE OXIDATION â&#x20AC;&#x201C; AN UPDATE By Karel Osten AMEC Minproc, Australia Karel.Osten@amec.com
Agenda
Gold pressure oxidation Historical cost influences An example - Macraes POX Developments Conclusion Questions
2
ALTA Free Library www.altamet.com.au
1
Gold pressure oxidation
Gold pressure oxidation Commercialised Gold POX
Typically used to liberate refractory Au from sulphide host minerals often with As Has been used on “double” refractory concentrates (Newmont process) Aqueous process at elevated pressure and temperature, low chloride Process for concentrates and whole ore Typically 190°C - 230°C, uses heat of reaction from sulphide oxidation
Less than 6%sulphur typically requires heat recovery ,“flash and splash”
Well above the melting point for sulphur (115°C)
Increased kinetics at higher temperatures
Lower solubility of precipitates at higher temperatures
Rejection of Fe as Fe2O3 and m.Fe3(SO4)2.(OH)6 and As as Fe.As.O4.2H2O
Mostly acidic but some alkaline Oxidation extent typically greater than 90% Tonnage oxygen used as oxidant, typically cryogenic
4
ALTA Free Library www.altamet.com.au
2
Gold pressure oxidation Uncommercialised Gold POX Other POX technologies uncommercialised for Au
Activox
CESL
UBC / AARL
Platsol
Nitric Acid Processes
Low temperature, fine grind, partial oxidation
Medium temperature, chloride assisted, partial oxidation
Medium temperature, fine grind, surfactant
High temperature, chloride assisted
NSC, Arseno, Redox, Nitrox
5
Gold pressure oxidation Abbreviated autoclave chemistry Oxidation of pyrite and arsenopyrite
2FeS2(s) + 7O2(g) + 2H2O = 2FeSO4(aq) + 2H2SO4(aq) 4FeAsS(s) + 11O2(g) + 2H2O = 4HAsO2(aq) + 4FeSO4(aq)
Further oxidation to ferric and arsenate
4FeSO4(aq) + 2H2SO4(aq) + O2(g) = 2Fe2(SO4)3(aq) + 2H2O 2HAsO2(aq) + O2(aq) + 2H2O = 2H3AsO4(aq)
Hydrolysis reactions
Fe2(SO4)3(aq) + 3H2O = Fe2O3(s) + 3H2SO4(aq) 3Fe2(SO4)3(aq) + 14H2O = 2H3O.Fe3(SO4)2(OH)6(s) + 5H2SO4(aq) (hydronium jarosite) 2H3AsO4(aq) + Fe2(SO4)3(aq) +4H2O = 2FeAsO4.2H2O(s) + 3H2SO4(aq)
Reactions consuming CaCO3 and MgCO3
CaCO3(s) + H2SO4(aq) = CaSO4(s) + CO2(g) + H2O MgCO3(s) + H2SO4(aq)) = MgSO4(aq) + CO2(g) + H2O 6
ALTA Free Library www.altamet.com.au
3
Gold pressure oxidation Typical flowsheet FEED PREPARATION Vent Quench Water Oxygen
Flocc Wash Water
Limestone / Lime, Air (Float Tails)
PRESSURE OXIDATION
NEUT’N
Lime, Air, Cyanide, Carbon
Lime
RSF
SO2, Air, Lime
NEUT’N
CIL
DETOX
Air Caustic, Cyanide, Fluxes
Au RECOVERY
$$$ 7
Gold pressure oxidation Installations Project
Client
Year
Temperature
McLaughlin
Homestake
1985
195°C
Sao Bento
Gold Fields
1986
190°C
Mercur
Barrick
1988
195°C
Getchell Gold
FirstMiss
1989
210°C
Goldstrike
Barrick
1990 – 1993
218°C
Campbell Red Lake
Placer
1991
195°C
Porgera
Placer
1992
190°C
Nerco Con
Nerco Minerals
1992
210°C
Lone Tree
Newmont
1994
180°C
Twin Creeks
Santa Fe
1997
225°C
Lihir
Rio Tinto
1997
210°C
Macraes
GRD
1999
225°C
Hillgrove
NEAM
1999
220°C
Kittila
Agnico-Eagle
2008
190°C
Amursk
Polymetal
2011?
200°C
Pueblo Viejo
Barrick
2011?
230°C
Lihir Expansion
Newcrest?
2012?
210°C
8
ALTA Free Library www.altamet.com.au
4
An example - Macraes POX
An example - Macraes POX Key parameters
Double refractory, Newmont technology Pyrite, arsenopyrite Low and high preg robbing ores 22 t/h @ 10.3 %S equivalent to 54.4 t S/d 3.5 m dia. x 12.6 m = 57 m3 working volume Horizontal 3 compartment vessel, 4 agitators 225°C and 3,140 kPa.g Koch Pyroflex membrane and AP302 bricks 1.5 MW Boiler 2 Stage CCD 24 h Feed storage US $14 M in 1999 BOC 180 t/d cryogenic oxygen plant 10
ALTA Free Library www.altamet.com.au
5
An example - Macraes POX Simplified Schematic
BLAST SPOOL XM-359 SAF2205 Ceramic Lined SAF 2205 SCRUBBER SR-350 SAF2205 Titanium Gr2 SAF 2205 TO CCD
SAF 2507 Titanium Gr2
OXYGEN 316S/S STEAM 316S/S QUENCH WATER 316S/S
ACID TANK - TK 358 Carbon Steel
TO CCD FLASH VESSEL PV-352 Carbon Steel Bric k Lined
SULPHURIC ACID 316S/S
316S/S
CS
GEHO PP-367 Poly
POX FEED PP-358/359
1A
1B
2
AUTOCLAVE Carbon Steel Brick Lined
11
An example - Macraes POX
12
ALTA Free Library www.altamet.com.au
6
An example - Macraes POX
13
Historical cost influences
ALTA Free Library www.altamet.com.au
7
Historical cost influences Gold price
15
Historical cost influences Titanium sponge price
16
ALTA Free Library www.altamet.com.au
8
Historical cost influences Crude oil price
17
Developments
ALTA Free Library www.altamet.com.au
9
Developments Use of FEA in design
FEA for vessel design
ASME VIII Div. 2 Design by analysis
ASME VIII Div. 1 Design by rule
FEA for lining system design
Interactions with shell
Effect of nozzle size and placement
19
Developments Lining systems 1 Membrane and brick vs Ti clad
Safety concerns due to potential ignition of titanium CESL demo plant Improvements in vessel fabrication due to HPAL experience
Image source: WE Smith Engineering
20
ALTA Free Library www.altamet.com.au
10
Developments Lining systems 2 Membrane choices
Panel bonded lead Homogenously bonded lead Weld overlay Organic membranes
Mortar Choices
Potassium silicate for slurry phase Litharge based for vapour phase Alternatives to litharge
Masonary Choices
ASTM C279 type II or type I Vendors have developed their own formulations
21
Developments Compartment wall construction 1
22
ALTA Free Library www.altamet.com.au
11
Developments Compartment wall construction 2
23
Developments Oxygen safety
BLAST SPOOL XM-359 SAF2205 Ceramic Lined SAF 2205 SCRUBBER SR-350 SAF2205 Titanium Gr2 SAF 2205 TO CCD
SAF 2507 Titanium Gr2
OXYGEN 316S/S STEAM 316S/S QUENCH WATER 316S/S
ACID TANK - TK 358 Carbon Steel
TO CCD FLASH VESSEL PV-352 Carbon Steel Bric k Lined
SULPHURIC ACID 316S/S
316S/S
CS
GEHO PP-367 Poly
POX FEED PP-358/359
1A
1B
2
AUTOCLAVE Carbon Steel Brick Lined
24
ALTA Free Library www.altamet.com.au
12
Developments Oxygen safety - tight shut off
Requirement for vent line isolation
Previously provided by severe service ball valves
Earlier ceramic severe service control valves could not achieve tight shut-off
New pressure control valve design has special trim
Issues of material compatibility in oxygen service Potential for excessive velocities downstream of isolation valve
Hydraulic actuator available for improved control 25 in POX vent service from 3â&#x20AC;? to 6â&#x20AC;?
Image source: Caldera Engineering
25
Developments Oxygen safety - automation
26
ALTA Free Library www.altamet.com.au
13
Developments Oxygen / Quench injection
Lammers (2004) reported the following benefits of combined oxygen and quench water injection
Cooling of oxygen lines
Reduction in vessel penetrations
Improved oxygen utilisation, survey results 80-90% vs 60-70% for separate addition
In addition provides means of draining autoclave
27
Developments Oxygen / Quench injection
28
ALTA Free Library www.altamet.com.au
14
Developments Alternate cooling strategies
Sulphide oxidation is exothermic requiring temperature control Quench water injection - default technology
Simple, flexible
Dilutes process (less of an issue for gold than for Cu, Zn etc)
Increases slurry volume, increasing autoclave size
Can be used for flushing and cooling of injection lines
Can be used to improve oxygen mass transfer
Flash cooling (atmospheric and vacuum) Internal cooling coils (slurry and vapour) External shell and tube heat exchanger External Klarex style self cleaning heat exchanger Internal steam recycle using gassing impellers
29
Developments Agitation 1
Oxygen mass transfer requires power input and gas dispersion Wetted parts are titanium
High cost
Subject to erosion and scaling
Limited access for impeller blades
30
ALTA Free Library www.altamet.com.au
15
Developments Agitation 2
Impeller Options
Rushton turbine
Smith Turbines
Lightnin A340 (up-pumping)
Ekato gassing impeller
Coatings to improve blade life Improvements in mechanical seal design Improvements in seal water supply VSDs for autoclave agitation
Significant drop in VSD cost
For gassed - ungassed conditions
Allow optimisation of mass transfer, wear rate, scale growth
Image source: Post Mixing 31
Developments Other influences
Influence of HPAL
Higher temp and pressure
Larger capacity feed pumps, lines, isolation and control valves
Generational improvements in valve coatings
Use of hydraulics for valve actuation
Improvements in titanium lined vessel fabrication
Use of drain nozzles
Development of â&#x20AC;&#x153;fleatersâ&#x20AC;?
R&D into high temperature chemistry
Use of CFD for design
Use of bus technology for process control Increased number of specialist vendors and service suppliers
32
ALTA Free Library www.altamet.com.au
16
Conclusions
Conclusions
Expect more POX autoclave circuits
World class operations such as Pueblo Viejo
Satellite treatment plants such as Amursk
Small operations such as Macraes
Au, Au + As, Au + Cu + Co
Larger autoclaves 5m diameter, 8 compartments, twin discharge Use of POX to stabilise As in AMD solutions Total oxidation not chloride assisted Horizontal multi compartment rather than multiple vertical units Increasing use of hydraulic actuation Increasing use of tight shut off valves Improved productivity, availability and safety
34
ALTA Free Library www.altamet.com.au
17