Search Documents

By B. G. Thomas

"Mathematical process models can be applied in several different ways to serve industry by inducing beneficial changes to process operation. This paper attempts to summarize the different uses, types, and attributes of these models, and to offer some ideas to consider when developing, using, and reporting on them. Process models range from empirical to mechanistic in nature and vary in complexity from simple analytical solutions to coupled, 3-D transient numerical simulations. They are classified here according to the way in which they are implemented in practice, viz. fully-online models, semi-online models, off-line models, and literature models. The purpose of the model should dictate how choices are made during its development. Ways to validate and compare the model with experiments are suggested. Examples are taken in the context of the author's experience in modeling the continuous casting of steel.IntroductionThe objective of this paper is to provide a partial introduction for this symposium by discussing some basic aspects of the application of mathematical process models in industry. Essentially, the aim is to explore how process models are used to benefit industry, and to offer some general comments and suggestions, based on the experience of the author, on how models can best meet this goal.This work is concerned with process models, which consist of systems of mathematical equations and constants that are solved on a computer to make quantitative predictions about some aspect(s) of the process. The specific variables required as input data and generated as output data are important features of the model. The equations often derive from a numerical solution to one or more differential equations and their boundary conditions. The model also includes the constants, which represent material properties, empirical relationships, and other knowledge about the process, and usually require considerable effort to obtain. Thus, general purpose commercial software packages, (ie. finite-element or finite-difference based codes), are not models in this context. They are useful tools, however, serving as frameworks for the development of process models."

Jan 1, 1994

By B. J. Vorster

AngloGold South Africa region currently consists of 12 Gold Plants. These plants use a combined total of $20 million of cyanide per annum. Of this, the .major portion (60%) is consumed at two Ergo dump retreatment plants. Historically the primary motivation for cyanide control at Ergo has been one of leachlcost optimisation. However, more recently, with the increased public awareness of cyanide in the environment, a secondary but increasingly important motivation for control has been to ensure that only the minimum amount of cyanide is added to the process whilst not compromising leach performance. Following a brief overview of the Ergo process, the methodology in detennining the amount of cyanide to be added is described. The paper then traces the developments in cyanide control from very basic? manual systems to the current automated control system. In line with the increasingly sophisticated control systems, developments also took place in regards to the method of cyanide analysis. The development of these analysers is also discussed. Whilst cyanide is one of the major drivers, if not the major driver of gold dissolution, it cannot be viewed in. isolation particularly with respect to the relationship between cyanide and oxygen as ascertained from the Elsner's well known equation. Consequently, in order to control cyanide addition, knowledge of the relative cyanide/oxygen profiles is necessary. Various means investigated at Ergo for pulp oxygenation have therefore been included for the sake of completeness. Finally a comparison is made of the control achieved from the current system compared to the? original manual system. The knowledge and experience gained at Ergo is now being used to draw up guidelines for the installation and optimisation work at the other AngloGold plants. The net effect of this will be reduced cyanide consumption for the region, which also, apart from the obvious economic benefit, translates into a reduced environmental; risk.

Jan 1, 1998

By Judy Che, Steven Ceccio, Gretar Tryggvason

"A detailed understanding of drop deformation and solidification is essential for quantitative predictions of geometry and properties of parts made by ballistic particle manufacturing processes. Axisymmetric results of towers built by the deposition of multiple drops are presented. In these computations, fluid flow, heat transfer, and phase change are fully coupled throughout the drop deformation and solidification process. The solution method is based on a single set of conservation equations written for all phases simultaneously. The phase boundaries are treated as imbedded interfaces by adding the appropriate source terms to the conservation equations. A parameter study demonstrates the effect of Weber number, Peclet number, and drop deposition frequency on the final shape of the structure. Numerical results show that most successful method of building uniform towers is when two conditions are met. First, each drop solidifies at the same time in the deformation cycle. Second, the cooling rate is slow enough to allow for good adhesion between drops. The ability to predict the final shape of deposited drops will greatly reduce the amount of experimentation necessary to select the desired set parameters for any artifact.IntroductionMaterial deposition techniques based on layering small molten metal drops holds the promise to build small, coherent three-dimensional structures. In this process, numerous 20-100 µm diameter drops are deposited onto a substrate with a low impact velocity (on the order of 5 mis). Both the momentum and trajectory of each drop are controlled precisely to build up complex parts. This material layering technique differs from processes such as spray forming and splat quenching which involve high impact velocities (on the order of 100 mis) and high frequency deposition with little control over individual drop placement. This process of layering small drops can produce parts with near net shape accuracy, while spray forming produces slabs of material which must then be worked to make a usable part. The ability to computationally predict final drop shape and cooling rate has the advantage of allowing the user to select parameters to produce parts with desired shape and mechanical properties. However, the development of such predictive methods requires a thorough understanding of both drop deformation and solidification"

Jan 1, 1999

By Arthur S. Klein, Hugh R. Lowry

"The Levitation Casting Process is a new continuous casting method that uses an electromagnetic levitation field, instead of a mold, to support and contain a column of solidifying metal. By counteracting gravitational forces and hydrostatic pressure on the metal column, this process eliminates friction and adhesion at the interface between the solidifying metal and the cooled walls of the casting chamber. The process inherently provides high casting speeds, excellent dimensional control and smooth, continuous emergence of solidified product from the top of the casting chamber.Of major importance is that the electromagnetic field produces an intense stirring of the liquid metal both before and during solidification. As a result of this stirring, the cast product has a homogeneous, equiaxed, fine grain structure generally suitable for immediate drawing or other forming operations without the need for hot rolling or other expensive additional processing.This paper reviews the fundamentals of the Levitation Casting Process, results obtained in casting various pure metals and alloys, and the production sized system at Lowell, MA, that uses this process. IntroductionThe Levitation Casting process was specifically developed to meet several requirements including:• Economic at low casting rates• Usable with pure metals and alloys• High, uniform product quality suitable for fine wire drawing or cold rolling into flat strip and tape• Easy cast product shape and size changes• Process inherently stable and easily automatedIn meeting these criteria it was also desired to retain the advantages of General Electric's well known Dip Forming continuous casting process; namely the inherent ""fail-safe"" properties of upward casting, solidifying metal by means other than pressure contact with a cooled mold, and use of a sealed furnace, casting chamber and cooling tower, filled with a protective atmosphere, to prevent oxidation of the molten metal or cast product."

Jan 1, 1988

By Jerry L. Smith

Industry on a whole has not been very receptive to government prodding to implement waste minimization programs. Waste minimization has the appearance of substantial benefits, however, there are underlying difficulties and suspicions that need to be addressed before industry will support pollution prevention initiatives. They can be summarized as follows: - Human behavior resists change, uncertainty, and unknowns. - New waste minimization techniques are difficult to find and research. - Calculating the true cost of present waste practices as well as new techniques being considered is difficult and time consuming. - Getting a consensus opinion of all the individuals affected on the course or technique to be employed in a pollution prevention program. New software products have addressed many of the problems associated with implementing pollution prevention programs: They contain research on a large number of pollution prevention techniques and case studies; they generally are accessed by EPA ID number, chemical name, or synonym; and some programs have the option to print a comprehensive written user-specific pollution prevention program.

Jan 1, 1994

By W. R. Opie

Fundamentals of blast furnace smelting of low grade copper and precious metal scrap and residues are reviewed with emphasis on emission control, by-products produced, slag cleaning for high recovery of metal values and the evolution of an off-gas system to eliminate hydrocarbons that result from decomposition of plastics in the blast furnace- charge. Problems with tin, antimony and nickel in the smelter-refining cycle are discussed.

Jan 1, 1985

By G. A. Irons, D. Guo

"The average spout height and width produced by bottom gas injection in a water model of a steel ladle were measured with an image processing technique. It was revealed that the spout height could be described by a Gaussian curve. A combined SIMPLE -VOF model was developed to simulate the liquid surface and flow. The results showed that, though the model produced reasonable velocity distributions and free surface positions, the spout height due to the dynamic head of the rising liquid was substantially lower than observed, indicating that the bubble dynamics at the bath surface play an important role in spout height.IntroductionDue to the enormous importance of gas injection and stirring in metallurgical processes, there have been numerous publications on water and numerical modeling of such vessels. The free surface of the bath has received considerably less attention, despite its importance. At the free surface there are often important chemical phenomena occurring that require a good understanding of the fluid flow; these would include reoxidation of steel, metal splash and fume formation. Furthermore, to understand the details of slag-metal interactions such as slag entrainment, metal losses to slag and the transport phenomena associated with refining reactions, it is necessary to understand the plume and spout interactions first.There have been few experimental studies of the free surface of liquid in gas-injected ladles [1- 3], and even fewer on numerical modeling. Brimacombe and co-workers [1] studied the gas void fraction in the spout region, using a double-tip electrical resistance probe. This sensor cannot distinguish between gas in bubbles about to break and gas in the freeboard. Thus, it is a matter of judgement as to the position of the free surface. Schwerdtfeger and co-workers [2] used a single probe configuration with a wire and a counter electrode in the liquid, so that the height of the continuous liquid phase was unambiguously determined. Irons and co-workers [3] measured the spout height by ultrasonic and electrical capacitance techniques. These techniques were as reliable as the resistance techniques, but were restricted to low flow rates."

Jan 1, 2001

By Willy Persson

A common feature of the kinds and amounts of impurities present most pyrometallurgical, processes is injection of oxygen to, remove various in the feed. The degree of success of a specific process step relies critically on control of the instantaneous oxygen potential/oxidation stage, throughout the step as well as upon termination of the step. This will, for instance, determine the distribution of impurities between metal and slag. As long as the operation is perturbation free -detailed analysis of all additions and application of advanced models often yield satisfactory control means However, a pyrometallurgical process is seldom free from disturbances, which reduces the accuracy of the static models and leads to inconsistencies in scheduled route might be uncontrollable changes in quality and quantity of feed, operator interventions and shift changes, timing of cranes, lining wear and accretions etc. The implementation of dynamic, control to correct for disturbances and, more generally for process optimization, requires continuous on-line information on, the actual status of the process. The implementation of sensors for continuous process monitoring, e.g., as regards the instantaneously prevailing oxygen potential or oxidation stage, has been hampered by the highly aggressive environment in smelters; This is. where remote-sensing techniques based on optics and spectroscopy, have proven to form invaluable, tools towards the realization of dynamic production control. Since about 15 years Semtch has developed and industrialized a remote-sensing technology, based on ,optics and spectroscopy, for continuous on-line process monitoring and production control of various pyrometallurgical processes. The presentation will discuss applicability aspects with examples and experiences from installations and tests in ferrous and non-ferrous industry. Keywords: Applied spectroscopy, Metallurgy, Process monitoring, Production. Control

Jan 1, 2003

By Jennifer Murphy, Brad S. Schake, Karol Holmes, John R. FitzPatrick, Mike H. West

"The separation and purification of uranium from either nitric acid or hydrochloric acid media can be accomplished by using either solvent extraction or ion-exchange. Over the past two years at Los Alamos, emerging programs are focused on recapturing the expertise required to do limited, small-quantity processing of enriched uranium. During this period of time, we have been investigating ion-exchange as a method for extraction/purification of uranium from acidic media. In addition, waste stream polishing is associated with this effort in order to achieve more complete removal of the uranium prior to recycle of the acid. Extraction chromatography has been demonstrated to further polish the uranium from both nitric and hydrochloric acid media thus allowing for a more complete recovery of the actinide material and creation of less waste during the processing steps.We will discuss contact experiments for evaluation of eleven different resins for uranium removal from nitric acid, the evaluation (again by contact experiments) of uranium elution from the ion-exchange media by various chemicals, and small-scale experiments for the removal of 235U from nitric acid. Larger-scale (1 DO-gram) flow experiments, using depleted uranium, were done after the small-scale evaluation of the selected resin. A discussion is also included concerning the shortcomings of the contact experiments and how the flow experiments were modified in order to overcome these problems. We will also discuss the application of ion-exchange for some specific process needs at Los Alamos.Contact experiments were used to evaluate some commercially available solvent-impregnated resins in addition to evaluation of several resins of this type synthesized in our laboratory. These experiments involved both nitric and hydrochloric acid media. The data from these experiments will be used to determine the column size and amount of resin required to polish the waste stream produced by the ion-exchange process described above. Small-scale flow experiments will be described and data given on the mass balances achieved using this two step process.Future work for evaluation of new ion-exchange and solvent-impregnated resins as well as the search for new eluants will be discussed. Other future experiments for both processes, including large-scale experiments, will be discussed in detail."

Jan 1, 1996

By Jukka Tuominen

The history of making blister copper in an Outokumpu Flash Smelting Furnace dates back to the late 1960?s, when Outokumpu first piloted the Outokumpu Direct Blister (ODB) process. The first commercial scale applications have been in use at Glogow, Poland, since 1978 and at Olympic Dam, Australia, since 1988. These applications utilize special concentrates, which are low in iron, and thus the formation of slag is also low. In 1995 the same principle was applied to smelting copper matte into blister copper in industrial scale at Kennecott Utah Copper. This application, the Kennecott-Outokumpu Flash Converting process, will celebrate its 10th anniversary of commercial scale operation this year. This paper outlines the development in efficiency and flexibility of Blister Flash Smelting processes compared to other smelting and converting options. The focus is on the economical aspects i.e. on the total feasibility of both greenfield investment projects and revamping of existing smelters.

Jan 1, 2005

By M. A. Barakat

Large amounts of zinc ash and flue dust containing more than 80 % zinc are accumulated during galvanization processes at the surface of molten zinc bath and in the chimney respectively. Pyrometallurgical recovery of zinc from both of the ash and dust samples has been carried out. The ash was sieved with the help of an industrial sieve shaker provided with mesh sieve having a pore size 0.9 mm. The oversize fraction cut (+ 0.9 mm) was melted at 550 °C under ambient conditions. The produced ingot (1) was analyzed whereas the coarse fraction of the produced slag (+ 0.9 mm) was treated as the coarse ash or added to it. The dust was enriched with suitable fluxing salt prior to heat treatment at 600 °C for 1 h. The produced ingot (2) was analyzed whereas the resulting slag was mixed with the fines of the ash and treated by hydrometallurgical processing to prepare valuable salts of zinc. Parameters affecting recovery processes such as time, temperature and flux percentage were studied. Results obtained revealed that the two prepared ingots (1&2) were successfully recovered with recovery efficiency amounts to =85 %. The obtained samples meet the standard specifications with relatively low price. Also in the light of the strict environmental regulations of pollution control, reuse of these wastes materials will help in keeping the environment clean.

Jan 1, 2000

By Robert M. Leary

The Kennecott Utah Copper smelter operates two Outokumpu flash furnaces: a smelting furnace for production of high-grade matte and a converting furnace for production of blister copper. Flash furnaces typically possess several characteristic features that must be considered in process control development such as long lag times associated with settler volume, non-linear input and output responses, large process gains, and unmeasured disturbances. Additionally, the Kennecott smelter possesses certain unique characteristics which must be considered such as single mine concentrate source, limited concentrate blending capacity, and feed preparation facilities that inevitably cause significant lags between furnace control parameters and manipulated set points. In order to improve furnace control and on-line time, a process control system was developed to address these issues. The feed forward control module provides predictive control and operates in response to furnace feed input changes. The feed back control module provides responsive control and is based on furnace output deviations. To provide regular feed back information to the control room supervisors, new sampling methods were developed which provide the operators with molten metal and bulk material samples at regular intervals to provide information for control. The control system was implemented on the Flash Converting Furnace in May 1999 and on the Flash Smelting Furnace in August 1999.

Jan 1, 2000

By B. Q. Li, H. C. de Groh, X. Lu

"A fully 3-D, transient finite element model is developed to represent the magnetic damping effects on complex fluid flow, heat transfer and electromagnetic field distributions in a metallic melt. The model development is based on our in-house finite element code for the fluid flow, heat transfer and electromagnetic field calculations. The numerical model is tested against numerical and experimental results for water as reported in literature. Various numerical simulations are carried out for the melt convection and temperature distribution with and without the presence of a transverse magnetic field. Numerical results show that magnetic damping can be effectively applied to stabilize melt flow, reduce turbulence and flow levels in the melt and over a certain threshold value a higher magnetic field resulted in a greater reduction in velocity. Also, for the study of melt flow instability, a long enough running time is needed to ensure the final fluid flow recirculation pattern. Moreover, numerical results suggest that there seems to exist a threshold value of applied magnetic field, above which magnetic damping becomes possible and below which the convection in the melt is actually enhanced.1. INTRODUCTIONThe imposition of a DC magnetic field on a moving, electrically conducting fluid, such as molten metal or semiconductor melt, produces an opposing Lorentz force that results in a reduction of flow velocity in the liquid. This effect has been explored in both the semiconductor and metals industries to control natural convection and/or surface tension induced Marangoni convection as well as thermally induced flow instability during solidification processing of these melts [1-6]. A mathematical model can be of great use in developing both fundamental understanding of the fluid-magnetic field interactions in these magnetically-assisted processing systems and basic guidelines for process design, control and optimization.This paper presents a 3-D mathematical model for magnetically damped fluid flow and heat transfer in a melt cavity, which is useful for interpreting results for unidirectional solidification processes at slow solidification rate, and some preliminary results concerning the magnetic damping effects on the flow instability in the melt. The model development is based on the finite element solution of the Navier-Stokes equations, energy equations and equations for induced electric field. The model is tested against reported results and is then applied to study the fluid flow and thermal distribution in the metallic melt with and without an external magnetic field."

Jan 1, 1999

By R. Boussaa, Y. Fautrelle, K. Zaidat, O. Budenkova, B. Saadi, X-D. Wang, L. Hachani, H. Ben Hadid

"Comparison of experimental data obtained in a benchmark experiment of solidification of a Sn-3%wt.Pb ingot and numerical results of two-dimensional and three-dimensional simulations of the process is presented. The benchmark experiment consisted in solidifying a rectangular ingot of Sn-3%wt.Pb alloy using two heat exchangers. Conditions of solidification were thoroughly controlled due to six thermocouples embedded into each heat exchanger. Due to another array of thermocouples the evolution of the temperature field at a lateral wall of the crucible was registered during the whole experiment. A treatment of the temperature data provided an estimation of the velocity field in the experiment. Post-mortem analyses of the sample provided the pattern of macrostructure of the sample and qualitative segregations pattern. 2D and 3D numerical solidification models are developed to give the distribution of the solute in a solidified sample.IntroductionConvection, either natural or forced, plays a very important role during the solidification. Its intensity and flow pattern affect heat and mass transfer and, consequently, macrostructure and segregation in a solidified ingot [1-4]. The effect of buoyancy convection during solidification has been studied in many laboratory-scale experiments [5-6]. Due to these experiments a general comprehension of formation of various defects in the ingot has been achieved. However, mentioned studies were not consistent since alloys, solidification conditions and configurations were every time chosen arbitrary. That is why the role of each parameter characterizing solidification process remains unclear. Moreover, the question of reproducibility of defects under the same solidification conditions still exists. In addition, most of the experiments suffer from the insufficient knowledge of control parameters that makes difficult their use for validation of models for solidification. These parameters serve as boundary conditions for numerical modeling and their incomplete description leads unavoidably to the errors in simulations. This situation motivated us to develop a setup for benchmark solidification experiments and to perform a set of the solidification experiments which we reported in [7]. Here we presented a comparison between results obtained in a benchmark experiment of solidification of Sn-3%wtPb ingot and its numerical modeling, which was performed in two-dimensional and three-dimensional configurations."

Jan 1, 2012

By Entov Vladimir M., Gilmer R. Burgos, Andreas N. Alexandrou

"Processing of metal alloys in their mushy state represents a new trend in metal processing. As part of this process, known also as thixocasting, a specially prepared slurry at a temperature in the mushy region is injected into a die. Parts produced using this technology offer better mechanical properties than those produced by casting, and comparable to those of forged alloys.In thixocasting, semisolid slurries are highly concentrated suspensions of rounded rosette-like crystals in eutectic liquid. The mechanical behavior of the slurry is determined by the structure and properties of the skeleton formed by the alpha phase particles. The skeleton structure is almost never at equilibrium, and its evolution is, governed by a number of kinetic phenomena of different c11'aracteristic time-scales. Therefore, the skeleton structure depends on the mechanical and thermal history of the material. As a result of these kinetic processes, the rheological properties of the material, such as effective viscosity and yield stress, decrease with structure breakdown, and increase with its development. In the present investigation, the behavior of semisolid slurries during processing is modeled using the conservation equations, and a Herschel-Bulkley fluid model. The rheological parameters are assumed to be functions of the solid volume fraction, and of a structural parameter that changes with processing history. The evolution of the structural parameter is described by a first order kinetic differential equation that relates the rate of build-up and break-down of the solid"

Jan 1, 2000

By Xiang-dong XING, Zhen-yang wang, Xing-le LIU, Jian-Liang Zhang

Experiments were carried out by adding CaF2 as catalyst in a Ar atmosphere to study the isothermal reduction kinetics of vanadium titano-magnetite carbon composite pellets under high temperature in the range from1473 to 1673K. By analyzing reduction mechanism, it was found that the rate controlling step was gas diffusion, and the activation energy was 178.39kJ•mol-1 without adding any catalysts. Adding CaF2 of 3% to vanadium titano-magnetite carbon composite pellets can decrease the apparent activation energy of reduction, and the decrease extent was 14.95kJ•mol-1. In addition, temperature is an important factor influencing on reaction rate.

Jan 1, 2015

By Zhijing Zhang

Long-term thermal stability of ionic liquids is essential for reliable performance of ionic liquids in any engineering applications. TGA studies were carried out on a series of ionic liquids with different cations and anions such as 1-butyl-3-methylimidazolium chloride ([C4mim]Cl), 1-hexyl-3-methylimidazolium chloride ([C6mim]Cl), 1-hexyl-3¬methylimidazolium tetrafluoborate ([C6mim]BF4), 1-hexyl-3-methylimidazolium hexafluorophosphate ([C6mim]PF6), 1-octyl-3-methylimidazolium hexafluorophosphate ([Csmim]PF6), and 1-butyl-3-methylimidazolium bis(trifluromethanesulflonyl)imide ([C4mim][Tf2N]). The short-term thermal stability of ionic liquids was evaluated by measuring the onset temperature of decomposition of ionic liquids, and the long-term thermal stability of ionic liquids was also evaluated by weight loss of ionic liquids over a long period of time at various temperatures. The experimental results showed that the long-term thermal stability of ionic liquids decreased with increasing temperature and was significantly lower than the onset temperature of decomposition of ionic liquids.

Jan 1, 2002

By Adam J. Gesing

The eddy current separator (ECS) is designed to separate nonmagnetic (nonferrous) electrically conducting (metallic) particles from nonconductive residue. Nonferrous metallic particles are repelled out of the material stream passing over the eddy current rotor (ECR). The rotor is housed in a fiberglass composite outer shell that serves as a conveyor head pulley. This results in the desired metaVnonmeta1 separation. The behavior of ferromagnetic particles, mainly iron and iron oxide (i.e. rust), is quite different. These particles are strongly attracted to the magnets and the smaller particles tend to align their long axis in the direction of the magnetic flux. Since the direction of the flux changes with the passage of each magnet pole, the small particles spin at very high rates causing wear of the belt and of the composite outer shell housing the ECR. Large particles, when trapped under the ECS belt, cannot spin, but can be heated by the eddy current effects and subsequently can burn through the composite outer shell. We discuss the practical implication of this behavior of iron and rust on the design, performance and maintenance of eddy current separators.

Jan 1, 1999

By George Kachaniwsky

Until mid-1982 the Mines Gaspe copper smelter consisted of a fluid bed roaster, a reverberatory furnace, two Peirce-Smith converters of which only one was blowing at any one time, anode casting facilities, and a single contact sulphuric acid plant treating both roaster and converter gases. Reduced concentrate supplies have led to the shutdown of the roaster and, as a result, it was necessary to adapt the converter and acid plant operation to the new operating mode: green concentrate smelting with resulting lower matte grades and acid plant treatment of off-gases from only one blowing converter. Therefore, modifications were carried out to existing equipment and methods of operation and control. This paper describes the result of work done to obtain steady, maximum strength, converter off-gas flow and an efficient acid plant operation. Topics covered include: optimization of converter operation, tightness of hood and off-gas ducting, improvements in acid plant/converter communications and modifications of acid plant operation and control to cope with the fluctuations in converter gas flowrates and strengths.

Jan 1, 1983

By Zhongliang Tian, Yun Cheng, Yexiang Liu, Ming Jia, Yanqing Lai

"A fundamental electrochemical study of Si ions in Na3AlF6-LiF melts and electrolysis using liquid cathodes was conducted. Cyclic voltammograms (CVs) clearly showed the depolarization of silicon deposition on liquid cathode by forming chemical complexes with the liquid solvent metals. Besides, co-deposition of Al and Si is possible when Si ions are insufficient in the molten salt. Based on the electrochemical results, silicon diffusion into liquid cathode is modelled and evolution of Si activity with concentration is deduced by a mathematic approach. Electrochemical study using liquid metals as working electrodes, can be used as an alternative approach for the acquisition of basic data on alloy behavior. Galvanostatic electrolysis showed that liquid cathodes are appropriate to collect the deposited solid silicon and the highest collection efficiency is almost 85%.IntroductionRecent years have witnessed an increasing demand for solar-grade silicon (Si) by the photovoltaic industry [1-4]. The preparation of high-purity metals through electrolysis and electrorefining offer many advantages such as low cost, low-energy consumption, and environment-friendly process, and thus has been widely used worldwide. From the perspective of electrochemistry [5-7], by utilizing the electronegativity difference between the impurities and Si and controlling the relevant process parameters, it is possible in principle to electrodeposit monatomic Si from a Si4+-containing compound melt. Moreover, because the carbon content of the high-purity Si obtained through electrolysis can be effectively controlled and used as a very ideal raw material in the preparation of electronic-grade and solar-grade Si, the preparation of high-purity Si through electrolysis has received widespread attention [8,9]. Monnier et al. [9] used a graphite electrode to obtain high-purity Si (99.9 %-99.99 %) from Si02/Na3AlF6 system through electrolysis; however, because the Si was precipitated in a solid form and showed poor conductivity, the deposition rate was too slow. Olsen et al. [10] obtained high-purity Si films via electrodeposition in K2SiF6/KF-LiF system by selecting electrode materials with high surface quality (such as Ag and graphite). However, in order to ensure the film quality, the deposition current density must be relatively low (10-50 mA/cm2), resulting in a low current efficiency and low deposition rate, which are not conducive to reducing energy consumption and applying to practical use"

Jan 1, 2014