Search Documents

By Hani Henein

"Heat transfer between droplets/particles and a gas phase plays an important role in the transport of numerous materials processing operations. These include rapid solidification operations such as gas atomization and spray forming, as well as chemical systems such flash furnaces. Chemical reaction rates and solidification are dependent on the rate of gas-particle or gasdroplet heat transport. Using a heat transport model validated using single fluid atomization of molten droplets; the effect of process variables on heat losses from droplets was examined. In this work, the effect of type of gas, droplet size, gas temperature, gas-droplet relative velocity on the heat transport from AA606 l droplets was examined. The most critical of these process variables to heat transfer is identified and will be presented.IntroductionIn materials processing, droplets are often encountered and used in processing a wide range of liquids and slurries. The small size of the droplet enables high reaction rates or high solidification rates in many systems. In order to capitalize on these high rates of transport processes, sprays are· also developed to maximize throughput. The design of these sprays, however, does not account for the transport steps and rate limiting mechanisms at the droplet surface. Rather, spray design focuses on maximizing the break-up of a stream or slurry into small size droplets or controlling the spread of the spray. Spray systems in industrial use typically generate droplets of wide size distribution leading to a wide range of reaction times for each of the droplets. Often the generation of a spray involves a number of coupled phenomena such as stream break-up, droplet-gas transport, droplet size distribution and spray geometry. Using gas generated sprays, it is not possible to identify how individual process variables contribute and control each of these coupled phenomena. While many mathematical models have been developed of sprays and they attempt to simulate coupled droplet transport, they do not provide insight into the contributions of different mechanisms in heat transfer. This is primarily because the basic fundamental heat transport between droplet and gas cannot be isolated from other gas spray phenomena experimentally such as two-way thermal coupling [l]. Thus, experimental validation of heart transfer of large and complex industrial spray models is not carried out. Hence, the validity of spray models remains in question and the approach required to optimize spray processes cannot be clearly identified"

Jan 1, 2004

By Jyri Talja, Shaolong Chen, Hannu Mansikkaviita

"For optimum smelting results, proper treatment of the feed is needed to ensure favourable conditions for ignition characteristics, heat balance, mass balance with tolerable impurity levels, and overall process economics. Feed preparation consists of physical treatments before drying (blending), chemical treatments before drying, (roasting, if required), drying itself and physical treatments after drying (fractioning, if required).Effects of fluctuation on mineralogy and moisture should be minimized by blending the concentrates in a bedding plant prior to drying. As an alternative to drying, additional treatments such as roasting for impurity removal may be needed for part of the concentrate.Sustainability, energy efficiency, utilization of waste energy, low gas and dust emissions, product quality, availability and occupational exposure reduction are decisive criteria in drying today. These criteria are best met by the Kumera steam dryer, which has the best energy efficiency, zero CO2 emission, lowest operational and maintenance expenses, and superior availability and product quality.IntroductionFeed preparation plays a crucial role in having a smooth smelting operation and also the overall smelting economics in metal production. Therefore, it is of significant importance for all copper, nickel and lead smelter operations in which sulphide concentrates are fed to the primary smelting furnace. Generally speaking, the higher the metal content in the matte after primary smelting, the more important the feed preparation. This is of particular importance in the direct blister (DB) process for copper production and direct oxygen nickel (DON) for nickel production.Bedding is the first stage in feed preparation. This is very important for the sake of smoothing the mineralogical fluctuation in the feed for better smelting control. Its importance increases as the number of concentrates available for the feed mix decreases."

Jan 1, 2012

By Chikabumi Yamauchi, Mitsuru Tanahashi, Tsuyoshi Taniguchi, Nobuo Furuta, Toshiharu Fujisawa

"It is essential to clarify the equilibrium relations among nonmetallic inclusions of the Mn0-SiO2-CrOx system and molten stainless steel in order to control non-metallic inclusions formed by the deoxidation of stainless steel with silicon. In the present study, the phase equilibria of the MnO-SiO2-CrOx system were measured by conducting the following experiments under controlled oxygen partial pressure, Po2, at 1873 K:1) Phase equilibria of the MnO-Cr20 3 binary system were measured for Po2 = 2 x 10-6 ~2 x 102 Pa.2) Phase equilibria of the MnO-SiO2-CrOx ternary system were measured for Po2 = 2 X 10-6 Pa.Based on the results, the phase diagram of this ternary system was established under the condition of the Si-deoxidation process of stainless steel (Po2 = 2 x 10-6 Pa and T = 1873 K).IntroductionAt the deoxidation stage in steelmaking process, generally multiple deoxidizers participate in deoxidation reactions, and the composition of deoxidation products ""non-metallic inclusions"" greatly varies depending on deoxidizing efoments and their contents. In order to control non-metallic inclusions formed by the refining process of steel, it is essential to clarify the equilibrium relations among non-metallic inclusions and molten steel.Especially, during the refining process of stainless steel, nonmetallic inclusions in molten steel seem to- change as follows. After the oxidation refining, Mn-chromite is formed in molten steel as non-metallic inclusion. By the deoxidation with silicon addition (mainly less than 1 mass%), Mn-chromite inclusions are changed their compositions to Mn-silicate. Segawa et al.[l] have reported the effect of amount of silicon addition on composition and mass ratio· of non-metallic inclusions during the Sideoxidation process of 18Cr-8Ni stainless steel at 1873 K, as shown in Figuri; L At higher silicon content than 0.5 mass%, only Mn-silicate was identified as non-metallic inclusion. On the other hand, at lower silicon content, both of Mn-chromite (MnO Cr2O 3) and Mn-silicate were identified. However, the thermodynamic and the kinetic analyses of the above-mentioned Sicdeoxidation process have not been made due to lack in the thermodynamic data for the non-metallic inclusions· of the MnO-SiO2-CrOx system, such as the phase diagrams, the activities of the components, and the free energy of intermediate compounds. In relation to the phase diagram of this system, those of the MnO-SiO2 [2] and the SiO2-Cr2O 3 [3] binary systems have been only reported"

Jan 1, 2000

By Charles Washburn

The authors review mercury sources, and contaminant streams, with particular emphasis on mercury that is extracted from precious metal ores. The design and operation of both large and small mercury distillation facilities are detailed. The relevant properties and characteristics of mercury and mercury vapor are presented. Facility engineering with respect to industrial hygiene, area ventilation, and material handling is presented.

Jan 1, 2003

By Alexander I. Raitchenko

"An electroconductive viscous liquid may flow in laminar or turbulent regime under Lorentz-force generated by the intercrossed electric current and magnetic field. When there is a starting flat layer between hard and free boundaries with surface tension the spatial Lorentz-force can create a specific stationary regime of flow: the plane free surface will transform into the wavy one, and the movement in the depth will be not laminar and not turbulent, but this regime will be characterized by organized forms with indications of ""rolling of long cylindrical drops"". Wave-length is determined by the viscosity, the density, the surface tension, the average thickness of layer, and Lorentz-force. ·The limit of wavy regime is determined by most long of wave which is directly proportional to the critical value of specific flow and inversely proportional to the kinetic viscosity. The proper critical value of Reynolds number is determined by the dimensionless combination containing surface tension, density, viscosity, and Lorentz-force. The mentioned dimensio.nless combination in our case is some criterion substituting the known Kapitsa number for case of the thin viscous liquid layers under gravity. Spreading of admixture by mechanism of wavy regime is more rapid than by common diffusion or by laminar flow. The suggested conception can be a contribution into the theoretical basis for elaboration of some specific liquid metals technologies."

Jan 1, 1999

By Zhang Chuanfu

The fibrous nickel oxide powder was prepared by the method of chemical precipitation and pyrolysis, that is, the NiO precursor was firstly precipitated from nickel ions solution with addition of oxalate salt and then the final NiO powder was obtained by pyrolysis of the prucursor. The effects of the pH, reaction temperature, reactant concentration, addition of surfactant on the morphology of the precursor particles were investigated. The influences of the various conditions, including the temperature and time, the air flow and the speed of the temperature increase on the morphology, average size and specific surface area of the NiO particles were also addressed. The fibrous NiO particles with IOO~ 120 in axis/diameter ratio and 5.5 m2/g of the specific surface area may be produced under certain operation conditions by this process.

Jan 1, 2003

By D. Tahija

The Getchell Mine produces gold from a geologically complex ore bearing high concentrations of arsenic and carbonate minerals. The mill treats the ore using high temperature aqueous oxidation in autoclaves, followed by conventional carbon-in-leach gold recovery. We laboratory tested a variety of Getchell ores to determine the effects on gold recovery of conventional factors such as grind size and retention time, as well as the influence of factors such as ore variability and preliminary carbonate destruction. Because the mill is currently operating, we were able to directly compare certain laboratory predictions with operational results.

Jan 1, 1999

By Michael L. Free

Bacterial oxidation of an arsenopyritic gold-ore .concentrate was carried out with continuous-flow .operation in 14-liter, mechanically agitated reactors. Samples from the slurry reactors were separated into solid and liquid fractions and analyzed separately to determine the electrochemical rates of sulfide conversion by the bacteria and the rates of direct bacterial oxidation of sulfide. The rates of the individual mechanisms are compared to the overall rates of biooxidation in the continuous-flow slurry reactors. The results show that the commonly hypothesize4 direct and indirect mechanisms of biooxidation both play a significant role in the bacterial oxidation of sulfide ores.

Jan 1, 1991

By Kenneth L. Reid

The Mineral Resources Research Center at the University of Minnesota has recently installed a QEM*SEM system (Quantitative E.valuation of Materials using Scanning Electron Microscopy) to aid in detailed mineralogical evaluation of potential mineral resource developments in Minnesota and to provide services to the North American mineral industry. A primary application for the unit is to serve the Minnesota taconite industry by providing mineralogical analysis of ores and concentrator products. Minnesota taconite ores contain fine-grained magnetite in a quartz matrix. The ores undergo several size-reduction/magnetic-separation steps to produce a final concentrate with approximately 75% passing 325 mesh and containing about 5% gangue. The paper describes a typical taconite concentrator flowsheet and reports some preliminary results of the applications of QEM*SEM particle image maps and cumulative yield liberation data to samples from selected points in the concentrator circuit.

Jan 1, 1988

By L. Klinger, T. Werber, LA Levin, A Katsman

"The process of catastrophic carburization known as ""metal dusting"", whereby a solid metallic material is converted into fine particles in a carburizing atmosphere, was analyzed. It was shown that this conversion is the result of plastic deformation and subsequent fracture accompanying diffusional phase formation in the near-surface layer. The process is controlled by internal stresses arising during phase transformation. Competition between stress generation and relaxation may result in the attaining the ultimate strength in the near-surface layer and its fracture. Such a mechanism of metal disintegration may occur in a certain temperature interval depending on the kinetic and geometrical parameters of the system - the diffusivities of the alloy's components, the ratio between specific volumes of the new and the old phases, and the ultimate plastic strain. The temperature interval for disintegration of carbon steel, the rate of disintegration, and the period of the disintegration cycle, were evaluated and compared with experimental data. IntroductionCatastrophic carburization, known as ""metal dusting"", consists in conversion of a solid metallic material into fine particles. The process occurs in a carburizing atmosphere at elevated temperatures. The corrosion product, usually named ""coke"", consists of metallic particles dispersed in a loose carbon matrix. Under industrial conditions metal dusting takes place, for example, in a gas mixture of CO and H2 or in hydrocarbon environments. Steels, nickel- and cobalt alloys are susceptible to this kind of wastage in the temperature range 400-800°.A mechanism of metal dusting in steels was proposed by Hochman [1] and later developed by Grabke and coworkers [2-5]. According to the proposed model, the process consists of the following reaction steps:"

Jan 1, 1996

By John ?gren, Martin Schwind, Henrik Larsson

"A new model for internal oxidation, as well as similar processes, has been implemented on computer. It is based upon a combination of random walk and cellular automata and is capable of treating 3-dimensional cases. To model diffusion, it utilises quantised units of matter. All steps of the process are included: nucleation, growth and coarsening. The effect of grain boundaries on diffusion as well as nucleation is taken into account. Further possible development is briefly discussed; especially the treatment of nucleation needs further consideration. IntroductionComplex diffusion-controlled processes may now be simulated on a routine basis when concentration varies along one co-ordinate only. One-dimensional simulations capture much of the information of interest but not all. For the case at hand, especially the spatial and the size distribution of precipitates, as well as the impact of grain boundaries, would be interesting information not accessible from 1-D simulationsSolving the full Fick-Onsager law in three dimensions does not seem feasible with the computers of today. In the model described in this article, random walk by quantised units of matter are considered. The model should be useful for a number of cases involving dispersed systems. With further development, it would indeed seem possible to simulate most types of phase transformations.Previous workWe have previously published two articles [1,2] on the new approach discussed in this work. In those articles we described the basic idea and showed, among other things, that very good agreement with an experimental concentration profile can be obtained, in a single-phase system, using independently assessed thermodynamic and kinetic data. Also, Schwind used the same random walk based concept to model coarsening [3]."

Jan 1, 2001

By Andreas H. Vassiliou

Organic materials are major constituents of the ore in several uranium deposits. Some of these deposits have been described as "urano-organic" because, the organic materials are believed to have provided the dominant control for the accumulation and deposition of uranium and are directly and intimately intermingled with the uranium. The nature of these urano-organic deposits has been problematic in terms of (a) the origin of the associated organic materials and (b) the specific form of' occurrence of uranium. The organic materials in urano-organic deposits have been genetically related to both coal and petroleum or their progenitors or deriva-tives. Analytical data presented here (infrared, chemical. D.T.A.) suggest that coaliferous materials or related plant derivatives represent the main source of the organic fractions in practically all major urano-organic ores analyzed from the Colorado Plateau and other areas. Recognition of the specific form of occurrence of uranium in urano-organic deposits has been problematic mainly because the uranium is usually in an extremely fine-grained and disseminated state in the organic matrix. Electron microprobe and X-ray diffraction analyses presented here show that fine-grained uranium minerals, specifically uraninite or coffinite. are disseminated in the organic matrix as the predominant form of uranium in most major deposits; however. the presence of true organo-uranium compounds is also suggested for some deposits.

Jan 1, 1981

By E. Kuzeci, R. Kammel

"Lately various voltametric methods have gained increased attention as an analytic tool in hydrometallurgy. Voltammograms with mercury drop electrodes or glassy carbon electrodes provide information about the electrocrystallization of cathodic deposits and allow calculation of kinetic data by means of characteristic parameters like the peak potential and peak current and their variation with the voltage scanning rate.Cyclic voltammograms are also suitable for identification of reaction sequences. Voltammograms with carbon paste electroactive electrodes containing metal sulphide additions reveal information about the dissolution behavior. The stripping voltammetry can be applied to analyze metal ions and organic additives (e. g., thiourea) in extreme low concentration. Typical application of the various methods studied with microcomputer controlled measurement system are explained by examples selected from copper hydrometallurgy.IntroductionThe use of electroanalytical methods in metallurgical process control can be dated back to the starting period of industrial copper refining, when the deposition of the metal ions was followed electro-gravimetrically by weighing the electrode with deposits (1). The principles of metal deposition and dissolution have not changed but the methods to follow these phenomena were subjected to noticeable developments and especially the introduction of polarographic investigations by Heyrowsky (2) has greatly contributed to the progress in electro-analysis. A review about the fundamental work on the electrochemical kinetics in earlier time is compiled in a monograph of Vetter (2). With the steady improvements in the instrumental techniques, it is nowadays possible to apply several current and potential modulations for measurements with improved selectivity and sensibility. There is however, still a lack in theoretical interpretation of the data obtained. More theoretical information are already available for linear voltametric methods for systems where reduced as well as oxidized species are dissolved in the electrolyte (4, 5). Some recent investigations have also shown the reliability of these techniques to study electro-crystallization (6, 7). Main fields of the application in the hydro - and electro-metallurgy are:"

Jan 1, 1989

By Peter K. Eriksson

Nanofiltration is very efficient in separating heavy metals from water solutions, even if they contain strong acids. Desalination System, Inc., has many years of experience in this area, and this paper deals with three different such cases. The first case deals with concentrating the heavy metals in the pregnant leach solution from a copper dump leaching site with a surplus water problem. The increased copper concentration in the pregnant leach solution increases the capacity of the following copper extraction stage. The nanofIltration permeate is polished with a biomass adsorption column to remove the small amounts of remaining heavy metals, before being discharged. This method eliminates sludge handling and disposal, which are quite costly with the alternative precipitation process. In the second case, a complicated precipitation method was used to remove heavy metals from the waste stream of a specialty metals manufacturer. Nanofiltration was installed to increase the metals concentration four times, which greatly simplified the precipitation process. The chemical consumption and sludge production decreased by more than 80 percent, and the labor cost decreased by 66 percent. In the third case, nanofiltration was used to purify waste sulfuric acid solutions containing heavy metals, to meet the specifications for technical grade acid. Sulfuric acid concentrations above 30 weight percent were successfully purified in a multiple pass membrane system.

Jan 1, 1996

By Yutaka Yasuda

The particle dispersion phenomena in a turbulent gas jet of the flash smelting process has been investigated. The particle number density in a non- reacting jet was measured experimentally by taking photographs of the particles in flight and counting them using the image analysis. Numerical computations were also performed using a mathematical model for a turbulent particle laden gas jet The experimental particle number density data were then compared with the predicted results of the model. Good agreement was obtained between the experimental and predicted data. The effects of the parameters such as the injector design, particle loading and air flow rate are discussed. Optimization of the concentrate burner design can be achieved with the help of experimental and theoretical simulation as performed in this work.

Jan 1, 1994

By F. Chabchoub

A mathematical model has been developed to study the influence of process variables on the horizontal Ohno continuous casting process. The model simulates the heat transfer and fluid flow in a wide plate ingot of pure tin. A two dimensional numerical analysis based on a control-volume finite difference approach, the Semi-Implicit Method for Pressure-Linked Equations algorithm (SIMPLER), and an Enthalpy Transforming Model was used. The shape and position of the solid-liquid interface were calculated at different casting speeds, ingot thicknesses, and cooling device characteristics, as well as various mould temperatures. Results of the mathematical model were compared with

Jan 1, 1992

By Yasuhiro Konishi, Kazunari Mizoe

"This paper describes a new processing of ultrafine, magnetic ferrites from sulfuric acid pickling solution in steel industry. In the ferrite processing, the iron (II) in the pickling solution was first oxidized to iron (III) by using the acidophilic thennophile Acidianus brierleyi. The bio-oxidized pickling solution was treated with solvent extraction to separate the iron (III) into the organic solution of tertiary aliphatic monocarboxylic acid. After that, the carboxylate solutions of iron (III) and other metal (cobalt) were hydrolyzed and precipitated ultrafine particles of crystalline ferrites at l90°C and l. 3 MPa for 20 min.IntroductionHot-rolled steel sheet has always a layer of oxides on its surface. The oxide layer arises as scale in hot rolling processes in steel industry, and the scale consists of the three iron oxides such as hematite Fez03, magnetite Fep4, and wustite FeO. Pickling removes chemically the scale layer that adheres to the steel surface. For normal steels, pickling is carried out in 15-20% hydrochloric acid at 60-70°C or in 20-25% sulfuric acid at 95-100°C. When steel is pickled with sulfuric acid, the scale oxides dissolve to give iron (II) sulfate.Spent pickling solutions are hazardous wastes containing a large amount of dissolved iron (II) and sulfuric acid, e.g., about 40 kg/m3 of dissolved iron (II) at pH 0. In typical commercial processes, spent pickling solution is usually cooled evaporatively to 5°C, and then to 0°C to crystallize iron (II) sulfate as its heptahydrate. After the crystallization operation, the sulfuric acid can be reused in the pickling operation.In this paper we describes a new processing of ultrafine, magnetic ferrite from sulfuric acid pickling solution, in order to recover dissolved iron (II) in the form of a valuable product. In the ferrite processing, the iron (III) was first supplied by the microbial oxidation of iron (II) in the pickling solution, and then the biooxidized pickling solution was treated with solvent extraction to separate the iron (III) into the organic carboxylate solution. After that, the carboxylate solutions of iron (III) and other metal (cobalt) were precipitated crystalline ferrites under hydrothermal conditions."

Jan 1, 2000

By Wei-ling Wang, Ryosuke O. Suzuki, Xiang-ning Meng, Miao-yong Zhu

"A finite-element thermal-stress model of continuous casting mold is conducted to predict deformation of copper plates and its change with different cooling structure. The results show that deformation behavior of copper plates is mainly governed by cooling structure and thermal- mechanical conditions, deformation amount is related to structure geometry, and a small deformation mutation occurs in cooper-nickel boundary due to different properties. The maximum deformation of hot surface centricities of wide face locate at 100 mm below meniscus and that of narrow face locate at meniscus and terminal of water slots and significant curvature fluctuations on both sides of copper-nickel boundary. The maximum deformation of centricities is increased up to 0.05 mm with thickness increment 5 mm of copper plates, and maximum deformations are only depressed 0.01 mm and 0.02 mm with increments of 1 mm nickel layer thickness and 2 mm water slot depth respectively.IntroductionIn conventional continuous casting of steel, a large amount of sensible and latent heat of molten steel dissipate in primary cooling zone and make continuous casting mold under enormous heating load giving rise to thermal stress, deformation, high-temperature fatigue and creep not only depress mold life-span, but reduce casting stability [1] Many studies focused more on heat transfer, solidification and shrinkage of solidifying shell to provide inverse proposals for optimizing cooling system and even designing new taper of mold [2,3], but it is generally useless to those of having been manufactured or being put into production. Therefore, mold-related forward problems show necessity to be discussed deeply, and stress and deformation distributions based on special thermal conditions in mold should be analyzed in order to explore suitable cooling structure of mold directly. Samarasekera et al. [4] examined heat flow and distortion of two-dimensional longitudinal sections and investigated effect of cooling water velocity, copper plates thickness and carbon content on formation of rhomboidity and longitudinal comer cracks. O'Connor et al. [5] developed a three-dimensional mathematical model to compute thermo-mechanical state in thin-slab mold and calculated temperature and deformation of mold wall to assess roles of various process parameters on impacting mold life. Thomas et al. [6] predicted temperature, distortion and residual stress in mold and indicated increased distortion requires increased remachining, which reduces mold life and increased residual stress increases the probability of a catastrophic crack failure. Park et al. [7l determined heat-flux profiles in funnel-shaped and parallel thin-slab molds based on an inverse heat conduction model and thermocouple temperature measured in-plant and put them into three- dimensional finite-element thermal-stress model to predict temperature, distortion and residual stress of mold wall. Korie et al. [SJ focused on prediction of temperature, distortion and residual stress in mold and in particular to clarify that mold life is significantly affected by temperature and its rigidity, but little effect by mechanical including clamping force, bolt prestress, friction and ferrostatic pressure. Meng et al. [9l predicted temperature and thermal stress in a slab mold and analyzed effect by cooling system geometry and casting speed quantitatively. In present study, a three-dimensional finite-element thermo-mechanical model with type-switched element is transformed from previous heat-transfer finite-element model [9l, and mold deformation is simulated in order to shed light on thermo-mechanical behavior with special thermal conditions and different mold geometry at high casting speed."

Jan 1, 2013

By Hideyuki Itou

Municipal wastes are mainly incinerated in Japan. Both bottom ash and fly ash from the incineration treatment are smelted recently to decrease the volume. Direct smelting of municipal wastes is also becoming popular. The smelting treatment generates fly ash which contains concentrated heavy metals such as Zn, Pb and Cu, and the harmfulness is causing concern. The smelter fly ash is mainly landfilled after stabilization. An effective treatment process of the smelter fly ash is required to recover the metals. For the treatment process, an application of chloridizing volatilization technique is considered. Chloridizing volatilization experiments were carried out to determine the effects of various factors on the volatilization behaviors in this study. A pretreated ash was mixed with a main raw material which mainly consists of hematite. Calcium chloride water solution was added into the mixture. The mixture was pelletized, dried and used as samples for the volatilization experiments. The sample was placed in the reaction tube heated in an electric resistance furnace. 5%02-95%N2 gas with controlled humidity was introduced into the reaction tube. Effects of humidity in the carrier gas, addition of calcium chloride and mixed ratio of the ash to the main raw material were investigated on the volatilization of Zn, Cu and Pb. To determine the volatilization behavior, samples were taken from the furnace at specified temperatures on the heating stage. In addition, experiments at specified holding temperatures for 15 minutes with shortened heating time were conducted. The two series of experiments revealed the effect of heating temperature on the volatilization behavior of Zn, Cu, Pb and Cl.

Jan 1, 2006

By C. D. Chapman

Current regulations regarding the landfill disposal of asbestos containing wastes are inadequate to ensure there are no ground water or airborne emissions. Accordingly, there is growing recognition that asbestos wastes can only be safely disposed of by using techniques that destroy the hazardous fibres. An innovative high-temperature plasma arc technology (The MAGRAM Process) is being developed for the recovery of magnesium metal from asbestos together with supplementary magnesia containing feedstock. The furnace will operate in. open bath mode with aluminium and/or ferrosilicon being employed as reductant. An important advantage of the process is that magnesium vapour within the furnace will be generated close to atmospheric pressure rather than at reduced pressure under which existing processes operate. Thermodynamic modelling work has been undertaken to define the range of slag composition and furnace temperature under which magnesium production at atmospheric pressure is feasible. The effect of slag chemistry, reductant addition, and temperature on product quality and yield have also been investigated and key aspects have been tested on bench scale. A 300 kW pilot plasma furnace unit has been employed for the treatment of calcined asbestos wastes. Aspects of design and operation of the unit are detailed. Acknowledgement: This work is substantially financially supported by the European Communities under the Industrial and Materials Technologies (Brite EuRam H) Research and Development programme.

Jan 1, 1995