Search Documents

By Robert E. McIvor

"The design/optimization of the pump and cyclones to maximize grinding circuit efficiency first requires a concise understanding of, and clearly defined specifications for, desired grinding circuit performance. Subsequently, the pump and cyclone performances that provide the desired grinding circuit performance can be defined. Once the performance requirements of each process unit are determined, selection of physical equipment parameters that will deliver those performances can be carried out.The Functional Performance Equation shows that there are two distinct efficiencies in closed circuit ball milling. One of these, “Circuit Classification System Efficiency” (CSE), is the fraction of “coarse” material inside the ball mill upon which grinding energy is deliberately expended, versus the remaining fraction of “fines” or finished size material inside the mill, upon which grinding energy is wasted (on over-grinding). The role of the pump and cyclones is to maximize CSE, which is the desired grinding circuit performance. The obtainable target CSE for any given circuit can be estimated, for example, by referring to a plant data base of CSE’s. Design/optimization guidelines then allow us to specify the pump and cyclone performances that will achieve the target CSE.This paper outlines the step-by-step, systematic method of pump and cyclone selection which maximizes CSE. Numerous examples of its successful application are provided.THE FUNCTIONAL PERFORMANCE EQUATIONObservation of the feed and product size distributions going into and discharging from the mill in grinding circuits with low versus high circulating load ratios initially led to the definition of “Circuit Classification System Efficiency” (or CSE, McIvor, 1988a), and subsequently the discovery of the Functional Performance Equation (McIvor, 1988b; McIvor et al, 1992). The CSE is the fraction or percentage of “coarse” (usually in reference to the plus P80) size material inside the mill, versus “fines” or product size material finer than the circuit P80. It can be readily calculated by taking the average of the percentage of said coarse material in the mill feed and mill discharge streams. As this also represents the percentage of the mill energy expended on targeted, coarse size material, it is directly proportional to overall grinding circuit efficiency and production rate. This is all clearly expressed in the Functional Performance Equation for ball milling circuits, for which the derivation has often been presented (e.g. McIvor, 2006)."

Jan 1, 2011

By F. D. Ford, Z. Dai

"Vale’s Ontario Operations re-introduced Cu/Ni separation at Clarabelle Mill in 2006. The mill processes ores from a number of mines in the Sudbury basin. A saleable Cu concentrate is produced at 30-60% Cu recovery from mill feed, which has to meet the quality specification. In particular, Ni content in Cu concentrate must be less than 0.6%, otherwise, penalties will apply and diminish the value of Cu concentrate, in addition to the lost Ni value. Generally, a Ni grade of less than 0.4% is obtained in Cu concentrate. More recently, while processing Cu-rich ore, Cu concentrate contained more than 0.6% Ni which is abnormally high. Detailed mineralogical investigation shows that Cu-rich ore contains more millerite in comparison to typical Sudbury Ni sulphide ore. Millerite is found to be more floatable than pentlandite and is much more difficult to depress using high-lime aeration conditioning, which is the standard practice in Cu/Ni separation to float chalcopyrite while depressing pentlandite. The current operation uses sodium cyanide at 20 g/t to depress pentlandite and pyrrhotite. However, it is found that a high dosage is required to depress millerite. It is also shown that starch and dextrin are promising in millerite depression.INTRODUCTIONPrior to bulk smelting in 1991, Cu/Ni separation was carried out at Vale’s Copper Cliff Mill, Ontario Operation. The typical Cu concentrate grade was around 30% Cu and 1% Ni. This is very high in comparison to today’s value of 0.4% Ni and would not be acceptable for external processing. However, the Cu concentrate was processed internally at the Copper Cliff smelter/refinery, so the Ni contained in Cu concentrate was not lost and no actual penalties applied. During the SO2 abatement initiative, bulk smelting was chosen because of the perceived simplicity and consequently, Cu/Ni separation was eliminated (Agar, 1991) and the Copper Cliff mill was subsequently demolished. In 2005 and after more than one decade of bulk smelting, business changes required the removal of some Cu concentrate from Clarabelle Mill bulk concentrate to provide smelter capacity for Voisey’s Bay concentrate in the Copper Cliff smelter. In essence, this strategy calls for Cu removal at Clarabelle Mill (Lawson, 2006). In this case, Cu concentrate is sold externally and it must meet the quality requirement of less than 0.6% Ni."

Jan 1, 2011

By Klaus Tietz, Hamish Adam, John Selby, Ian Glazier, Brian Sinfield

"Room-temperature tunable diode laser (TDL) analyzers are gaining in acceptance for highly critical safety monitoring applications in a variety of industries. This paper describes the testing, evaluation and deployment of a laser-based hydrogen cyanide monitoring system at the Campbell Complex of Goldcorp’s Red Lake Gold Mines in Balmertown, Ontario, Canada.A trial, single channel installation was followed by a staged multi-channel installation. Unexpected technical issues with the multi-channel installation were addressed and resolved.INTRODUCTIONLaser-based gas detection has several advantages over existing gas detection techniques (Ballangrud, 1993; Silver, 1992; Bomse, 1995; Table 1). For practical purposes, lasers generate light at a single wavelength. Room temperature tunable diode lasers (TDL) emit light in the near infrared (NIR) part of the electro-magnetic spectrum. Many gases of interest, including HCN, NH3 and others, absorb energy in the near infrared (NIR). Each gas has a unique absorption signature, or spectrum, made up of a large number of individual “absorption lines”. A TDL can be tuned to select a single absorption line of a target gas, which does not overlap with absorption lines from any other gases. Therefore, laser gas sensing is selective and does not suffer from interference from other gases. The single-wavelength nature of laser light also means that laser gas analysers have a linear response over a wide dynamic range.Tunable Diode Lasers (TDL) generate only a few mW of power. However, all this power is concentrated at one wavelength, the wavelength where gas absorption occurs, resulting in high signal-to-noise ratios and short response times, typically of about 1 or 2 seconds (Ballangrud, 1993). This also means that long path lengths can be used, enabling accurate measurements of very low concentrations, even with weakly absorbing gases. Tunable Diode Lasers provide all the advantages of other semiconductor devices: they are small, solid-state devices, operate at room temperature, and have long-term reliability (over 15 years). Currently, they can be obtained for any wavelength throughout the NIR (700nm to 2000nm). Because many gases absorb in the NIR, the technique is widely applicable."

Jan 1, 2011

By M. R. Khalesi, S. Bellec, C. Bazin, D. Hodouin

This paper proposes a classification model to predict the split coefficient of gold ore particles as a function of the particle size and gold content. The model is calibrated and validated using data obtained from the sampling of an industrial grinding circuit. Model calibration is based on a complete simulation of the circuit and involves parameters related to the gold grain distribution in the ore as well as a grinding and liberation model. Since cyanide is added into the studied grinding circuit, assumptions related to gold dissolution in the mills are made to allow simulation results which reflect industrial behaviour patterns.

Jan 1, 2011

By J. L. Viljoen

t has long been the drive of the industry and all operations to create an environment that is free of injury and illness. We have come a long way; the leaders, executives, managers, supervisors and workers are to be commended for relentlessly driving to create, what I call a ?White Flag Day Every Day?. Yet I am also aware of the huge shift we still need to create in the industry, on the operations and in the workplaces to have them free of injury and illness. Keywords People, values, care, trust, belief.

Jan 1, 2011

By F. Sorensen

ABSTRACT Electricity production by the Electricity Supply Commission of South Africa (Eskom) is the most important beneficiation process for South African coal. As one of the world’s lowest cost producers, energy-intensive industries such as ferroalloys and aluminium were introduced. However the poor were neglected. The white paper on energy policy (Department of Minerals and Energy, 1998) sought to redress this. Also, chasing economic growth to generate jobs resulted in high electricity demand and interruptions. Lifespan estimates of South African coal reserves based upon peak theory vary widely. Approaching peak, coal production in the future will be more expensive and, thus, contribute to more expensive electricity, which could inhibit industrial development unless South Africans work smarter.

Jan 1, 2011

By Carl Bertrand, Patrice Nadeau, Claude Bazin

"Rio Tinto Fer et Titane processes the ore extracted from the Lac Tio orebody to concentrate the hemo-ilmenite (FeO.TiO2-Fe2O3) mineral that is subsequently smelted to produce pig iron and titanium oxide slag. The ore is crushed, heat treated (roasted) and concentrated using dry magnetic separators. This paper describes a model to predict the magnetic susceptibility of the roasted ore and its behaviour in the magnetic separation plant. The model uses a classification of the ore as a function of the size and magnetic susceptibility of the particles. A subsequent division into mineral species allows the simulation of the magnetic separation process. The model is used to assess various operating options when process disturbances occur.INTRODUCTIONOre processing operations consist of ore size reduction followed by the recovery of valuable species in a concentrate that is subsequently sold for its minerals or metal content. The concentration processes are based on the physical and/or chemical properties of the minerals contained in the ore. Flotation is typically used to recover minerals of base metals such as copper, zinc, lead or nickel. Iron ore and other dense minerals are typically concentrated by gravity separation processes. Magnetic separation uses the magnetic properties of the minerals to concentrate magnetite and other minerals, to clean coal concentrate and to recycle ferrosilicon for dense separation processes (Brophy, 1985).Magnetic separators can be divided into low and high intensity separators and high gradient magnetic separators (Gillet, 2008). Most of the low gradient magnetic separators are built as rotary drums that are operated in wet or dry conditions. Dry magnetic drum separators are used by Rio Tinto Fer et Titane in Sorel-Tracy, Canada to concentrate hemo-ilmenite from the ore extracted at the Lac Tio Lake mine near Havre St-Pierre, Canada. The analysis of production data suggests that process improvement could be achieved by identifying key variables for the control of the magnetic separation circuit and/or by modification to the circuit flowsheet. The use of a process model was considered to be an effective approach to conduct a preliminary analysis of the circuit with a view toward process optimization. However, the literature contains few detailed studies dealing with the modelling of magnetic separation processes."

Jan 1, 2011

By Maarten van de Vijfeijken, Marco Rufli

"By using the latest generation of medium voltage frequency converters, new possibilities have arisen to improve the grinding process. Beside all the direct and indirect advantages, several dedicated functions have now been implemented for ring geared grinding mills in the minerals industry. The advanced operational functions ensure a smooth, safe and reliable operation with minimum stress for the mechanical equipment and therefore highest availability of the mill. Several papers have already been presented concerning the benefits of using this state of the art technology for mill drives (Ravani von Ow and Bomvisinho, 2010; Ravani von Ow and Gerhard, 2010). However this paper will not only show and explain the operational benefit of these functions, but will focus on the experiences gained with this kind of sophisticated ring geared mill drive system. Several measurements of different operating conditions (starting, frozen charge remover, controlled roll back, backlash, etc.) made on an installed and commissioned 2 x 5 MW dual pinion ring geared mill drive system will be shown and discussed. Also the accuracy of the load sharing between the two 5 MW motors during starting, normal operation and stopping will be demonstrated using real measurements from an operating system. Finally the inherent features of these drive systems are briefly discussed.INTRODUCTIONLooking at the demands of grinding mills, we can split them into operational, maintenance and protective demands. For smooth and safe operation, it is important that critical situations are avoided as much as possible. Maintenance functionality must be quick and easy to execute, while protection of the system is important during all operating conditions. ABB has considered all these demands during development of their dedicated mill functions. By adding an additional controller (mill controller), not only a lot of application related functionality and protection is included, but also the interface between the mill drive system and the customer’s distributed control system (DCS) is simplified."

Jan 1, 2011

By Tassos Grammatikopoulos, Sarah Prout, William Mercer, C. Gunning

"The Nechalacho rare metal deposit is located in NWT, Canada. Avalon Rare Metals Inc. is now evaluating processing techniques for recovery of REE to be used in the industry. QEMSCANTM analysis coupled with whole rock, REE, Zr and Nb, and electron microprobe analyses were utilized to identify and quantify the mineralogy of the deposit, and define the REE distribution among minerals. QEMSCANTM mineralogy was performed on fifty intact drill samples using EXPLOMINTM that provided textural information for the mineralization; crushed samples were used to determine the variability of the ore and define REE domains, and composite samples to determine the liberation and association of the REE minerals. The sample matrix predominantly consists of biotite, feldspars, quartz, muscovite/clays, minor chlorite and amphibole, magnetite and hematite and carbonates. REE include in decreasing order of abundance allanite (3.6%), monazite (1.5%), synchysite (0.9%), columbite (0.9%), fergusonite (0.6%), bastnaesite (0.4%), and zircon 11.0%. The majority of the HREE are carried by fergusonite and zircon, and the LREE by monazite, allanite, synchysite and bastnaesite. QEMSCANTM mineralogical data is extremely valuable in assessing the mineralogy of the ore. The data have been implemented in defining the mineralogy of the ore body, support metallurgical test work, predict recoveries and grades, and finally calculate reserves of the deposit.INTRODUCTIONAvalon is in the unique position of owning and developing one of the world’s largest rare earth deposits in the Northwest Territories of Canada.The term Rare Earth Elements (REE) is used to describe the 15 lanthanide elements or metals and, more recently, the metals yttrium and scandium, which display similar physical and chemical properties. REE are not particularly rare and have a similar abundance in the earth’s crust to nickel, tin or tungsten. Unlike these better known metals however, the rare earths have little tendency to become concentrated into commercial ore deposits. When they do become concentrated, deposits of REEs usually display enrichment in one of two groups. The most commonly occurring deposits of REEs contain a sub-group of elements referred to as the “light rare earth elements” (LREEs), which include lanthanum, cerium, praseodymium, neodymium, and samarium. Less commonly, deposits of REEs contain the much more valuable sub-group of elements referred to as the “heavy rare earth elements” (HREEs), which include europium, gadolinium, terbium, dysprosium, holmium, erbium, thulium, ytterbium and lutetium. Yttrium and scandium most commonly occur in association with this second HREE sub-group. Generally, LREEs are produced from the mineral bastnaesite, and the HREEs have been produced from the minerals xenotime and monazite. LREEs continue to be produced from bastnaesite, however in the 1980s, China started production of HREEs from unique laterite clay deposits. China now produces approximately 95% of the world’s rare earths."

Jan 1, 2011

By Qionghua Zhou, Wei Liu

"Nanocrystalline silver bulk material with average grain size of 20.3 nm was prepared by warmvacuum- compaction method. The as-prepared nanocrystalline silver was characterized by X-ray diffraction (XRD), differential scanning calarmeutry analysis (DSC), thermogravimetric analysis (TG), and 6157 type Electrometer test, respectivily. The experimental results show that the average grain size and microstrain of the nanocrystalline silver are 20.3 nm and 0.088%. The melting point and melting enthalpy are 1228.7 Kand 82.34 Jig, which are lower than that of coarse-grained silver by 6 Kand 22.6 Jig, respectively. The electrical resistivity of nanocrystlline silver increases with higher temperature when the range of temperature is between 233K and 293K. The electrical resistivity at 293 K is 1.475xlO -7 Q·m, which is higher than that of coarsegrained silver by a factor of 9.IntroductionN anocrystaline ( nc) materials or nano structured ( ns) materials are three-dimensional (3 D) bulk materials with average grain size typically smaller than 100 nm. Ne materials have been of great interest for their unique properties, such as superior tensile strength [1-2], high microhardness [3- 4], superplasticity [5-6], magnetically quasi-isotropic [7-8] , and so on, making them one of the most promising nano-scale materials with wide applications. Ne materials with average grain size of 5-10 nm were firstly fabricated by Gleiter et al. [9] in 1980s, using inert gas condensation (IGC) method. Unfortunately, the large surface to volume ratio creates a high surface energy component in the total free energy of the system and results in a metastable state in the materials. Therefore, due to the signfica nt disordered grain boundary regions, nanomaterials are thermally unstable and are subjected to a strong driving force for grain growth. The grain growth even occurred at room temperature [10].The conflict or deviation of measured date, e.g. strength, microhardness, thermal stability, partly due to the grain boundary ratio or the distribution of grain size, and partly derived from imperfections such as voids and microcracks while earlier processing techniques are known to result in incomplete dens:fication [11-12]. Therefore, the higher quality samples produced in subsequently years still show such a general trend."

Jan 1, 2011

By Rodolfo Espinosa

"Industrias Peñoles, S.A. de C.V. in Mexico, has nine mining operations, six managed by Peñoles (base metal mines) and three managed by Fresnillo plc (precious metal mines). Seven of the operations use flotation and the other two are heap leach operations, one for gold and the other one for copper.This paper gives a brief introduction of Peñoles, shows mine locations and describes the main process characteristics of each concentrator. It also presents four of the new operating practices that were implemented over the last four years to improve the economics of the mining operations. These include (1) the use of centrifugal gravity concentrators to increase the recovery of gravity recoverable gold (GRG) and silver (GRS), (2) the use of expert systems to reduce process variability while increasing grades and recoveries, (3) the application of the Six Sigma methodology to key projects to ensure success and personnel development and (4) the use of water chemistry fundamentals to identify chemistry problems and design adequate solutions for improving grades and recoveries.INTRODUCTIONPeñoles was founded in 1887. It is a mining group with integrated smelting, refining and chemicals producing operations. Peñoles is the world largest producer of refined silver and metallic bismuth, the leading Latin American producer of refined gold and lead. It is amongst the world leading producers of refined zinc and sodium sulfate.Peñoles operates and controls six underground mines in Mexico and is the largest producer of zinc and lead of the country. It also produces copper concentrates and cathodes. Through its 77.1% equity stake in Fresnillo plc (a London Stock Exchange listed company); Peñoles is also a significant producer of gold and silver. Fresnillo plc operates three precious metal mines in Mexico. While Fresnillo is the world largest primary silver mine, Ciénega and Herradura are two leading gold mines of Mexico. Figure 1 shows a map of Mexico indicating the locations of the Peñoles mines. The main process characteristics of each concentrator are listed in Table 1."

Jan 1, 2011

By S. S. Dimov, B. R. Hare

This paper describes two micro-beam analytical techniques(Dynamic SIMS and TOF-SIMS) which have become an intricate part of a comprehensive analytical approach for precious metal (PM) ore characterization and process mineralogy or deportment studies. The value of the D-SIMS technique is its capacity for accurate analysis in the sub ppm range along with the ability to produce high resolution elemental maps. The TOF-SIMS is significant for speciating and quantifying preg-robbed Au on the surface of carbonaceous material. In combination with traditional assays, the two techniques can provide invaluable information regarding the distribution of PMs within an ore along with providing insight into potential sources of PM loss in processing or where losses have occurred in process streams.

Jan 1, 2011

By C. Fleming, J. R. Goode

ABSTRACT: Data were produced from two uranium project studies that were conducted in 2008, both involving uranium recovery from leached slurry by resin-in-pulp (RIP) with strong base resins. In both studies, excellent metallurgical results were produced. The loaded resins were eluted with strong sulphuric acid and recycled through RIP circuits in an integrated process. Resin losses by either abrasion or mechanical breakage were immeasurably low. Metallurgical results from these two campaigns are presented, along with a case study economic analysis of RIP versus the alternative processes of counter-current-decantation or filtration and solvent extraction.

Jan 1, 2011

By Deyou Yin, Kongfang Feng, Bo Wang, Jieyu Zhang, Xin Hong, Shaobo Zheng, Jinyin Xie

"The flow patterns and mixing characteristics in a water model of C-H2 smelting reduction furnace are experimentally studied in this paper. Water modeling is employed to investigate the features of gas jet stirring and the fluid flow patterns in the bath of a smelting reduction furnace during the top, bottom, and side combined blowing process. The influence of the gas flow rate for top, bottom and side nozzles, angle, submerged depth of each side nozzles and the height of the top nozzle has been examined. It showed that the gas flow rate of side nozzles and bottom nozzle had a key role on the fluid flow and mixing characteristics. The gas jet from the top nozzle could change the agitation and fluid flow pattern caused by the gas jets from the side nozzles in the bath.IntroductionFacing the current natural resources decrease and the environmental degradation, smelting reduction iron-making technique is believed to be the most representative at present. Mid last century, the concept of producing liquid iron or semi-steel directly from ore fines and inexpensive fuel was the subject of considerable pilot plant development activity. Among these projects, the direct smelting process, which is the reduction and melting of iron-bearing materials by carbon and other hydrocarbons in the vessel containing a molten bath, is the most popular one. A number of direct smelting processes have been developed around the world, e.g. COREX, DIOS, HIsmelt, AISI, ROMELT, etc.These studies put forward a new technology, which through bottom nozzle blowing rich hydrogen, top nozzle blowing rich oxygen and provide preheating ore and solvent, side nozzles blowing oxygen-coal to realize smelting reduction iron-making process. Owing to the metallurgy reactions always in high temperature, which is above 1600 °C, it is difficult to visualize the mixing behavior and measure the desired phenomena in the reduction vessel. With water modeling, the primary advantages are that it is relatively easy to observe and measure the flow phenomena in the bath."

Jan 1, 2011

By S. Aktas

The present work describes an investigation of a hydrometallurgical process for the recovery of mercury from spent silver oxide button cells. The cell paste was leached in nitric acid to solubilize silver, mercury and zinc. The optimum leaching conditions, in which mercury is entirely converted to Hg2+ ions and, therefore, does not co-precipitate with silver chloride during treatment with potassium chloride solution, were 1.0 M HNO3, 45° C, 1 hour and 200 rpm. After the silver was removed from the solution, mercury was cemented out with zinc powder with an efficiency greater than 99%. In the cementation experiments, the amount of zinc powder, shaking rate, temperature and solution pH were investigated in detail. The mercury recovery (%) increased as the amounts of zinc powder and the temperature were increased. Zinc hydroxide was precipitated from the waste solution by the addition of sodium hydroxide. The Zn(OH)2 precipitate was then converted to zinc oxide. The proposed process is proven to be simple, straightforward and environmentally friendly.

Jan 1, 2011

By P. Nelson, B. Salt, J. A. Scott, H. Shang

ABSTRACT In a nickel smelting process, large amounts of off-gas emissions are generated, which can cause serious environmental and plant hygiene problems if not properly treated. With increasingly stringent environmental regulations, an enhanced understanding of well-controlled, off-gas cleaning systems is vital. In this paper, dynamic models for an industrial smelter, off-gas system tackling hazardous sulphur dioxide, carbon monoxide, and carbon dioxide are developed using mass continuity, momentum, and energy conservation laws. The effect of air in-leakage is explored under different operating conditions. The developed models provide an enhanced process understanding and a necessary basis for highperformance control analysis of smelter off-gas systems.

Jan 1, 2011

By E. T. Tuzcu

Microscale information about single particle fracture can be incorporated into the macroscale model of the whole size reduction process in the mill. The ultrafast load cell (UFLC) device provides a substantial amount of information about fracture characteristics of a particle under impact loading. UFLC captures high-resolution compressive force signals during particle breakage. The measurement of the compressive force at the instant of fracture of the particle allows computation of a number of comminution-related particle properties, as well as the energy absorbed during fracture. In the current study, limestone particles 50 mm in size were fractured to obtain specific fracture energies. Then, based on the experimental data, two models were proposed to quantify the fracture process. The specific fracture energy distribution of the particles in the range of 6 to 50 mm was defined by the well-known Weibull distribution. The other model is a probabilistic fracture model. This model is a link between energy events occurring in a grinding mill and the breakage rates of particles. The impact energy spectra or energy events occurring in a grinding mill can be obtained with a discrete element method (DEM) simulation. One can model the breakage rates of the mill material by combining the energy spectra with the probabilistic fracture model. This novel approach for the modeling of breakage rates via DEM and UFLC is shown here.

Jan 1, 2011

By M. Mai Manga, M. Renaud, A. Faucher, C. Bazin

ABSTRACT Grinding circuits are operated in the presence of disturbances resulting from ore changes and equipment malfunctions, such as hydrocyclone overloading, worn or partly plugged apexes, surging, and internal liner detachment that may cause a loss of classification performance. Metal recovery losses in downstream leaching or flotation circuits justify the development of methods for rapid detection of hydrocyclone malfunction. Two methods are presented in this paper. The first method uses measurements from sensors available in a grinding circuit and the second uses vibration sensors installed on the individual hydrocyclones of a nest.

Jan 1, 2011

By Asbjørn Solheim

The service life for aluminium reduction cells is often limited by the cathode carbon, due to preferential wear along the periphery of the cell leading to the so-called "W" wear pattern. The reason for this preferential wear is not well understood. Two mechanisms are suggested in the present paper; 1) non-stationary dissolution of aluminium carbide into a layer of sludge that periodically forms and disappears, and 2) electrochemical formation and dissolution of aluminium carbide within an aluminium carbide layer containing bath-filled pores. The latter mechanism may be related to the high local current density at the periphery of the cathode.

Jan 1, 2011

By J. E. Dutrizac, Alba Sunyer

The factors affecting the precipitation of arsenic-free alunite (KAl3(SO4)2(OH)6) were investigated in a series of laboratory experiments. Although the amount of alunite precipitated increased with increasing retention time, temperature, solution pH and the concentrations of Al(SO4)1.5 or K2SO4, the composition of the products remained nearly constant. The preferred conditions identified for the precipitation of arsenic-free alunite were then used to assess the behaviour As(III) and As(V) during alunite precipitation at 98°C. Increasing additions of As(III), as NaAsO2 or As2O3, resulted in a slight increase in the amount of precipitate formed at pH 2.80. However, X-ray diffraction analysis showed all the precipitates to be alunite, and <0.2% As(III) was detected in the products. In contrast, increasing additions of As(V), as KH2AsO4 or As2O5, resulted in significant increases in the amount of precipitate formed at pH 2.80 or 3.00. Furthermore, the arsenate (AsO43-) contents of the precipitates increased up to 15%, and the sulphate contents decreased, as the As(V) concentration of the synthesis solution increased from 0 to 9 g/L. X-ray diffraction analysis of the products detected only alunite, but there was a progressive expansion of the alunite cell with increasing arsenate content. The implication is that arsenate substitutes for sulphate in the alunite structure. Preliminary observations on the environmental stability of the arsenate-bearing alunite using the TCLP protocol gave solubilities <0.4 mg/L from a precipitate containing 8.5% AsO4 and 7.8 mg/L from a product containing 19.5% AsO4. Thus, it appears that alunites containing less than -10% AsO4 would be an acceptable vehicle for the disposal of arsenic from hydrometallurgical processing solutions.

Jan 1, 2011