Search Documents

By Klaus-Markus Peters, Robert M. Cravens, James G. Hemrick

"Degradation of refractories in molten aluminum applications leads to energy inefficiencies, both in terms of increased energy consumption during use as well as due to frequent and premature production shutdowns. Therefore, the ability to enhance and extend the performance of refractory systems will improve the energy efficiency through out the service life. TCON_ ceramic composite materials are being produced via a collaboration between Fireline TCON, Inc. and Rex Materials Group. These materials were found to be extremely resistant to erosion and corrosion by molten aluminum alloys during an evaluation funded by the U.S. Department of Energy and it was concluded that they positively impact the performance of refractory systems. These findings were subsequently verified by field tests. Data will be presented on how TCON shapes are used to significantly improve the thermal management of molten aluminum contact applications and extend the performance of such refractory systems.IntroductionThe production of molten aluminum creates aggressive environments for processing equipment, thereby requiring refractory materials to contain the aluminum during melting, transfer, treatment, and casting operations. Degradation of these refractories not only causes issues with reduced product quality and production yields, but also increases heat losses as the insulative properties of the refractories are compromised. Furthermore, in addition to disrupting production output, refractory failures lead to further impacts upon energy efficiency as large amounts of energy are lost during the cooling and subsequent reheating of the equipment as the refractory linings are repaired or replaced.The need for improved refractory materials is recognized throughout the industry. A joint U.S. Department of Energy and industrial workshop held in 2000 on applications for advanced ceramics in aluminum production [1] noted that refractory materials used in melt processing equipment have many limitations. The workshop report noted that these issues may be addressed by using advanced ceramics, but new materials must be cost effective when compared to current refractory materials. Also, in 2005 the Canadian aluminum industry produced a technology roadmap [2] having the goal of identifying and bridging gaps between current technological resources in the industry and future requirements. One of the areas identified as being crucial to the future of the industry was the need for improved molten aluminum-resistant materials."

Jan 1, 2009

By Paykan Safe

Computational Fluid Dynamic (CFD) modeling provides a powerful tool to assist with the design of ventilation and fume control systems in smelters and other high temperature metallurgical facilities. For this paper, this tool has been used to analyze the off-gas flow patterns exiting the mouth of a Peirce-Smith converter into a water-cooled hood and drop out box. The effects of various process and physical plant design parameters on process gas and fume capture and potential build up on the converter hoods was examined, and the optimum design and operating parameters were determined.

Jan 1, 2000

By Dong Liang, Qijie Zhai, Zhichen Zhang, Qixin Li, Zhishuai Xu

"The formation, distribution and survival of pure Al nuclei under surface pulsed magneto oscillation (SPMO) treatment were investigated with thecombination of experiments and numerical simulation. The results indicated that the nuclei originated from the whole top surface of melt owing to the oscillation of top surface induced by SPMO. The temperature field significantly affected the survival of the nuclei The electromagnetic force and flow field of the melt with SPMO treatment were simulated by finite element software ANSYS. The electromagnetic force mainly distributed below the top surface near the mould wall. The electromagnetic farce not only made the melt oscillate but also increased the convection of meh below the top surface vigorously. The oscillation near the top surface of the melt promoted the nucleation largely, and the convection increased the uniform of the temperature distribution in the whole meh which favored the survival of nucleiIntroductionAs we all know, fine equiaxed grains are beneficial fur the improvement of the mechanical property, process ability and the uniformity of metallic materials [1-2]. With the researches of recent decades, people discovered that the electromagnetic field is an efficient method to refine the solidified structure of metals [3-7]. Pulsed magnetic field technology obtains more and more attentions in recent years due to its smaller power consumption, bigger instantaneous magnetic field intensity, non-contacting with the meh and simple convenient equipment manufacturing [8-9]. However, the melts would splash fiercely when pulsed magnetic field is applied with a slightly higher power. Thus, based on the understanding of the solidification refinement mechanism of electric current pulse (ECP) [10], Gong et al [11] developed a novel technology called pulse magneto-oscillation (PMO) to refine the solidification structure. The technology can be easily applied in the solidification structure refinement of continuous casting slab. But fur aluminum continuous casting, the sump is very shallow, which greatly limits the application of pulse magneto-oscillation technology."

Jan 1, 2014

By D. E. Larson

The U. S. Bureau of Mines has developed a image analysis methodology to correlate the energy consumed during the grinding of an ore to the mineral liberation achieved. Energy consumption values are obtained using a laboratory-scale ball mill equipped with a torque sensor, and an automated image analyzer is used to determine the liberation present in ground products. By combining liberation information with energy consumption data an appraisal can be made of the energy required to liberate the mineral(s) of interest. Results of tests performed on a complex sulfide material determined the effects of grind duration on the liberation achieved and the energy efficiency.

Jan 1, 1995

By Xin Zhang, Mengqi Li, Hongxu Li, Xinrong Zhang, Hao Bai, Ning Li

"Organic Rankin Cycle (ORC) is an effective technology to transform low-grade energy resource into electricity. In this paper, based on thermodynamic principle, considering the waste heat resource lower than 120°C, thermodynamic properties of ORC with working fluids ofR245fa, mixed R21/R245fa and mixed Rl52a/R245fa were assessed. The results show that high evaporation temperature contributes to improved system performance. The Rankine Cycle that uses R21/R245fa can increase net cycle output, thermal efficiency and exergy efficiency by 3.02%, 2.35%, and 4.44%,respectively, as compared to that with pure R245fa. When using Rl52a/R245fa as working fluid, the net cycle output can increase by 3.83% at most, while the thermal efficiency decreases by 8.7%. These results indicate that R21/R245fa has a good advantage in improving the cycle thermodynamic performance.IntroductionOrganic Rankine cycle is usually used for low temperature power generation, and is applied to the conversion of geothermal energy, low-temperature solar energy, ocean thermal energy and industrial low-temperature waste heat etc. Traditional Rankine cycle is quite mature already, however, organic working fluids would affect significantly the general efficiency of the entire thermodynamic system. Thus, selection of proper working fluids and development of new highperformance cycle is becoming important."

Jan 1, 2013

By Min Chen, Lei Xu, Weijun Huang, Shuai Zhang

For the purpose of extending the service life of MgO-C bricks used as linings of vanadium-extraction converters, the effect of Fe addition on properties of MgO-C bricks was investigated. The results showed the addition of Fe could effectively improve the sinterability of the oxidized layer of the MgO-C refractory at 1400°C due to the effect of FeO (formed by oxidation of Fe) on formation of MgO-FeOss and modification of the microstructure, and excellent corrosion resistance against vanadium containing slag was also obtained due to the increase of compactness of the oxidized layer and concentration of FeO in the oxidized layer. Consequently, it is considered that MgO-C brick by introducing the new phase of Fe is a favorable substitute of MgO-C refractory to be used as linings of vanadium-extraction converters.

Jan 1, 2015

By Wenxian Zhu, Xuema Li, Patrick R. Taylor

"Experiments were performed in a non-transferred arc, thermal plasma flow reactor to investigate the synthesis of ultra fine boron nitride (BN) powder from the reaction system of boron oxide (Bz03), methane (CH4), ammonia (NH3) and hydrogen (H2). Argon (Ar) and nitrogen (N2) were used as plasma gases. The thermodynamics and kinetics of the synthesis reactions were examined and a chemical reaction mechanism was proposed. The powders produced under certain operating conditions had high purity and good conversions. All the products were extremely fine powders (< 30nm) and were uniform in size distribution. The powder products were analyzed by chemical and physical characterization. The effects of the reaction variables on the conversions from B203 to BN and the purity of the products were investigated.IntroductionIn the past few decades, boron nitride (BN) has attracted a great deal of interest because of its excellent thermal, chemical and electrical properties (1-4) It has a high thermal stability and thermal conductivity. Chemically, it is very stable and is not wetted by most molten metals and slag. It is also an excellent low loss insulator. Since hexagonal boron nitride (h-BN) is light in weight, strong and easily fabricated, it has applications in a variety of specialized high temperature areas, such as a high temperature lubricant, crucibles for molten metals, thermocouple protection tubes, break rings for horizontal continuous casting of steel, and friction reducing coatings. In cubic form, boron nitride is one of the hardest materials known, second only to diamond, and it is an excellent material for tools in which the use of carbon based materials is limited by oxidation or reaction with metals."

Jan 1, 1998

By Musa Karakus

Reflected light microscopic (RLM) examination of dust particles from lead and copper smelters can effectively characterize their mineralogy, texture, and origin. Scanning electron microscopic-energy dispersive spectroscopic (SEM-EDS) analysis provides confirmation of the optical identifications and chemical data on the phases. Studied lead smelter dust samples were from the: I) blast furnace, 2) sinter plant scrubber, 3) sinter plant crusher, 4) dross furnace, and 5) refinery. Phases identified included galena fragments, melted PbS, metallic Pb and Cu, Cu-Ni, Cu-Zn, Zn-Ag, Zn-Cu-Ag, Ni-Cu-Sb, and Ni-Cu-Zn-Ag metal alloys, Ni-sulfide, Cu-sulfide, and Cu-Pb-sulfide dross .phases; slag and sinter particles, coke fragments, unreacted original feed phases: pyrite, chalcopyrite, sphalerite, siegenite, bravoite; and oxidized particles: iss, bornite, chalcocite, covellite, magnetite, hematite, and goethite. Dust from a secondary copper refinery were found by RLM and SEM to consist of tenorite, cuprite, cassiterite, two silicate phases. Magnetite, hematite, Fe-Al-Cr-Mg spinel, periclase, forsterite, and wollastonite were derived from the furnace linings during initial cleaning.

Jan 1, 1992

By Kazuki Morita

A new route to produce solar grade silicon (SOG-Si 6N) from metallurgical grade (MG-Si 98%) has been developed in Japan, which may satisfy the increasing demand for solar cells in the near future. Since some elements such as P and B have very large segregation coefficients, solidification refining is not effective for their removal and costly processes, such as electron beam vacuum melting and oxidation plasma refining, are required. According to our previous results which showed the effectiveness of leaching procedure with Ca addition for Fe and Ti removal, possibility of P and B removal by acid leaching was focused expecting that their affinities for Ca in molten silicon might be considerable. Hence, the effects of Ca addition on the thennodynamic properties of P and B in molten silicon alloys have been investigated for overall control and optimization of the SOG-Si production process in the present study. The interaction parameters between Ca and P, and Ca and B were determined by chemical equilibration techniques at 1723 K. Although the interaction parameter between Ca and B was slightly negative, that between Ca and P showed a large negative value. Also, the results for the leaching test with various amount of Ca addition revealed that the removal fractions of both elements, especially P, drastically increased with increase in the Ca content, which showed very good agreement with the prediction from the interaction parameters derived in the present work. The present results suggest that the preliminary leaching procedure can reduce the processing time and energy for SOG-Si refining. Keywords: Silicon, Solar Cell, Refining, Acid Leaching, Thennodynamics. Interaction, Phosphorus, Boron, Calcium

Jan 1, 2003

By Keiji Okumura, Jiayi Shan, Jian Yang, Masamichi Sano

"A new method of deoxidation and desulfurization of molten iron with magnesium vapor has been developed. An immersion tube, which was filled with pellets of the mixture of magnesia and carbon powders, was put into the molten iron. Magnesium vapor produced by carbothermic reduction of magnesia inside the tube was injected directly into the molten iron with argon carrier gas. The effects of temperature and carrier gas flow rate on deoxidation and desulfurization were investigated. A mathematical model of the reactions was developed. The rates of deoxidation and desulfurization were controlled by mass transfers of magnesium vapor in the gas-phase and oxygen and sulfur in the liquid-phase. This method may offer high efficiency for magnesium deoxidation and desulfurization of molten iron.IntroductionMagnesium is a strong refining agent in steel-making processes1 • At steelmaking temperatures, magnesium has a very low solubility in liquid iron, and a very high vapour pressure2 • Deoxidation and desulfurization with magnesium take place essentially through the interaction of magnesium vapor with oxygen and sulfur dissolved in molten iron, and magnesium can change chemical composition of nonmetallic inclusions3 • However, addition of magnesium to the steel bath is dangerous and does not give high efficiency in the refining process4 • Furthermore, as a refining agent, the price of magnesium is higher than aluminium and CaO fluxes. In the present study, a new method of the deoxidation and desulfurization of molten iron was developed using magnesium vapour in-situ produced by carbothermic reduction of magnesium oxide. Pellets made of mixture of magnesia and carbon powders were charged into an immersion tube, which was made of alumina or graphite. The magnesium vapor was injected directly into the melt. The effects of pellet mass, experimental temperature and carrier gas flow rate were studied on the refining rate. A mathematical model of the refining reaction was proposed on the basis of the experimental results."

Jan 1, 2000

By I. M. Richardson, M. Janssen, W. Geerstma, I. J. Bennett, V. A. Popovich

Avoiding wafer breakage is a big challenge in the photovoltaic silicon industry, limiting production yield and further price reduction. Special fracture strength tests suitable for thin silicon solar wafers and solar cells, to be used in combination with Weibull statistics, finiteelement (FE) modelling and digital image correlation have been developed in order to study the mechanical stability of solar wafers. The results show that removal of the saw damage significantly increases the strength of crystalline silicon wafers. Furthermore, it was found that silicon crystallinity and the location where the wafer is extracted from the cast Si ingot have a significant effect on the strength, namely samples taken from the bottom of the ingot are 30% stronger than those taken from the top. The study also showed that there is a decrease in fracture strength when an anti-reflective SiNx coating is applied, which is caused by high thermal stresses.

Jan 1, 2015

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 R. Kola

Since more than 15 years the Waelz process at "Berzelius" in Duisburg has proved suitable for reprocessing the zinc and lead containing dust and .sludge which inevitably occur in making steel in EAF-plants. The produced Waelz-oxide is agglomerated in a hot-briquetting plant and charged to an Imperial Smelting Shaft Furnace to recover the valuable elements in form of metallic zinc and lead. In view of the actual importance of the dumping problems of the EAF wastes a description of the "Berzelius" treatment route is given but also the situation in Western Europe is illustrated. Dust from stainless steel production is successfully treated in the Scan Dust Plasma Process in order to recycle Chromium and Nickel, while the resulting Zinc-Dust is treated via the Waelz-ISF-Route. For treatment of low Zinc-containing dust, sludge and fines from integrated steelworks, pilot scale tests have been carried out successfully in the Circulating Fluid Bed in order to recover metalized iron and .zinc/lead fume. A 30,000 tpy Demonstration Plant is planned?.

Jan 1, 1990

By Ashish D. Patel, Ramesh S. Minisandram

"Vacuum Arc Remelting (VAR) is the final and, hence, critical melting step for the production of a sound, segregation-free ingot for the manufacture of alloys used in rotating components in aircraft engines. A mathematical model for this process has been developed by the Specialty Metal Processing Consortium (SMPC). The model predicts temperature distribution in the solidifying ingot, flow in the molten metal pool, and the electromagnetic field in the ingot. These predictions may be used to estimate the effect of melting parameters on the pool profile in the ingot. In this paper, typical pool shapes and flow fields during VAR of Alloy 718 and Ti-6Al-4V are presented. The differences in the pool profile are primarily due to vastly different melting conditions. The model predictions are compared to the observed pool shapes from industrial-size VAR ingots.1. IntroductionVacuum Arc Remelting (VAR) is a process in which an electrode is melted in vacuum by the heat from a DC electric arc. Metal droplets from the electrode fall into the molten metal pool, which solidifies in a water-cooled copper crucible to form the ingot, as shown in Figure I. The ingot obtained from VAR is sound, and has superior cleanliness and chemical homogeneity. Hence, the process is the final melting process for most nickel-base superalloys and titanium alloys used in the manufacture of critical rotating components (discs & blades) of aircraft engines. Modeling of this process provides an insight into the effect of melting parameters on the molten metal pool in the solidifying ingot, which is a key first step in process optimization and defect prevention. In this paper, typical pool shapes, both measured and predicted, for VAR of Alloy 718 and Ti-6Al-4Vare presented. Alloy 718 is the most common VAR-processed nickel-base superalloy; whereas Ti-6Al-4V is the ''workhorse"" alloy for titanium-based alloys."

Jan 1, 2004

By Leijun Li, M. M. Mahapatra

"A three-dimensional finite element model with moving heat source was developed for pulsed-laser powder deposition (PLPD) process. Experiments were conducted to study the repair of superalloy components using the PLPD process. During the experiments, IN 625 powder was deposited on a superalloy substrate similar to IN 738. The process was modeled using three-dimensional finite element analysis. Element birth and death technique was used in the finite element model for simulating the deposition of powder layer on the substrate. A moving distributed heat source was used in the thermal model for realistic simulation of the process. Temperature dependent thermal material properties including enthalpy were used for the powder and the substrate in the transient thermal analyses. The fusion zones and peak temperature distributions of the deposits were predicted from the transient thermal analyses.IntroductionPoor weldability is a critical issue of superalloy repairing. The microstructural directionality of the superalloy components are disrupted when repaired by the fusion welding processes. Currently there is no widely accepted method to economically and reliably repair directionally solidified superalloys components. Pulsed laser powder deposition process (PLPD) is a promising repairing technique as the substrate is not severely affected by the heat source.In this study PLPD technology is used as the tool for superalloy repairing. In most of the land based turbine engines the directionally solidified nickel based superalloy similar to IN 738 [1] is used as the bucket material. The turbine engine blade tips are subjected to a combination of extreme operating conditions and corrosive environment. The tip of the nickel based turbine engine blades can be repaired using IN 625 material as suggested in the literature [1-2]. IN 625 is considered to be ductile and strong enough for the purpose of Ni based turbine engine blade tip repair[1-2]. During the present PLPD investigation, IN 625 powder was used as the deposit material on an experimental super alloy substrate similar to IN 738 superalloy used as the cast superalloy for the turbine engine blades."

Jan 1, 2012

By Shizhe Liu, Bing Li, Heng Wang, Zuxin Zhao

In this paper, cyclic voltammetry and chronoamperometry have been applied to investigate the effects of Nd2O3 concentration on the anode effects, critical potential for the anode effects and Nd2O3 reduction in LiF-NdF3-Nd2O3 melts at 950℃. The results show that the oxidation peaks in the potential of about 2.7V are attributed to the oxygen ions oxidation and the peak current density increases with Nd2O3 concentration increase from 0 to 0.5wt%. The critical potential caused the anode effects shifts to 0.3V more positive value as Nd2O3 content in the melts decreases from 0.01wt% to 0.5wt%. The cyclic votammograms on W electrodes in the LiF-NdF3-Nd2O3 melts present two pairs of reduction and oxidation peaks, indicating that the reduction of Nd ions in the LiF-NdF3-Nd2O3 melts in two steps rather than one step.

Jan 1, 2014

By Oleg Ostrovski

Industrial pyrometallurgical processes in ferrous metallurgy are based on carbothermal reduction of metal oxides. Carbothermal reduction of stable oxides requires high temperatures. Lowtemperature reduction can be implemented by using methane-containing gas with high carbon activity, or by carbothermal reduction under reduced CO partial pressure. Under standard conditions, methane is thermodynamically unstable above 550 °C and decomposes to solid carbon and hydrogen. At appropriate CH4/H2 ratio and temperature, carbon activity in the methane-containing gas phase can be well above unity relative to graphite, which provides favorable thermodynamic conditions for reduction. To maintain these conditions, the rate of reduction/carburisation should be higher than the rate of solid carbon deposition. The paper discusses reduction of pure manganese and chromium oxides at relatively low temperatures, and constraints in reduction of manganese and chromium ores. Reduction of metal oxides by carbon in hydrogen as an alternative use of natural gas is also discussed.

Jan 1, 2014

By William Brinkman, Mark Jolly, Nasim Soleimanian, Karl Dearn, Oliver Brinkman

"Glass manufacturing is a heat intensive process. There is a direct correlation between the batch distribution techniques and the furnace energy consumption, productivity, and quality of the glass manufactured. All four major segments (float, container, fibre, and specialty glasses) would benefit from using an optimised batch distribution technique where possible. Oscillating batch chargers (OBC) have been in use since the early 70s, despite their superior batch shape, coverage, and in turn positive effects on the energy consumption and productivity of the furnace they are almost exclusively used in container glass manufacturing. The OBC’s main difference compared with other charging methods is its ability to directly influence the batch initial velocity. This paper reports on results achieved in CFD models (GFM) used to study effect of the machine on the overall energy consumption in the doghouse and the melt space.IntroductionGlass manufacturing can be divided to four major segments float glass, container glass, fiberglass, and specialty glasses. The financial cost of manufacture, its impact on the environment, lack of standardisation and mounting pressure imposed by the authorities for reduction in emissions means that the industry and its suppliers are utilising CFD modelling in order to tackle these challenges. The need for high temperature to create and maintain the viscous flow, homogenisation of the glass melt, and maintenance of the process to allow continuous output of glass of the right quality are the contributors to the energy intensive activities that are involved in continuous glass manufacturing. This in turn has resulted in design of large container glass furnaces which have high energy consumption, low specific performance, and high CO2 emissions where the most efficient furnaces in the container glass sector have a specific primary energy consumption of 3.8 GJ /tonne of glass at a level of 50% cullet in the batch making even very small improvements in efficiency by operation optimisation a value adding activity [1-14].The theoretical calculated energy required to melt glass is 2 to 3 times less than the energy actually used to melt glass as a result of glass melt passing a zone 5 to 8 times before leaving the tank.[8]. This causes contamination as the result of mixing between the completely molten glass free from bubbles returning to the batch area where there is freshly molten non-homogeneous glass, still containing very large numbers of bubbles and un-molten batch. At high temperature some components are dissolved rather than melted. Given that an un-molten batch is a good thermal insulator, it can be transported a long way into the furnace, increasing the need for homogenization which also increases the glass residence time of glass melt. This is why large unbroken strings of batch should be avoided, as should an overall blanket without gaps."

Jan 1, 2014

By E. N. Mounsey

The paper addresses the design of the Zhong Tiao Shan Company (ZTS) Houma Ausmelt Technology copper smelter in People&apos;s Republic of China. The smelter represents the third copper smelter development using Ausmelt Technology, the first large scale (35,000 tpy blister copper) and the first smelter to use the Ausmelt Technology smelting and converting process steps in separate TSL vessels in a continuously operated process. The project will be realised through the combined efforts of ZTS, Ausmelt and ENFI, the Beijing Non-Ferrous Engineering Institute who provided partial detail engineering services in China. Ausmelt will provide the technology licence, basic and detailed furnace engineering packages and additional technical services including cold commissioning support and hot commissioning supervision.

Jan 1, 1999

By Takeshi Yoshikawa, Kazuki Morita, Kunitoshi Matsunaga

"In metallurgical grade silicon (MG-Si), most impurities mainly exist at the grain boundaries and distribute in silicon grains at extremely small contents. Grain boundary phases in MG-Si transform into liquid above the eutectic temperature, and migrate towards high temperature regions when heated under a temperature gradient. With the migration of all liquid phases in MG-Si to the surface, the impurity contents will be decreased less than their solubilities in principle.In the present work, the fundamental research was conducted to discuss the possibility for purification using the liquid phase migration technique. The purification limit was estimated by thermodynamic calculations, and metallic impurities were estimated to be effectively removed except for Fe and Al. The migration behavior of the liquid phase in the synthesized silicon substrate was investigated by unidirectional heating under infrared radiation at 1523-1548 K, and the proceeding of migration was confirmed. The migration velocity was 1.4--2.0xl0-7 mis, and found to be controlled by the liquid phase diffusion.IntroductionPhotovoltaic energy is one of promising energy resources in a sustainable society because of its inexhaustibility and cleanliness. Although the production and installation of solar cells have increased significantly during the last decade, the cost of solar cell installation is still high. To enable the widespread use of photovoltaics and for the photovoltaic electricity price to reach grid parity, a low-cost and abundant feedstock of solar grade silicon (SOG-Si) is essential. Recently, a modified Siemens process and a fluidized bed reactor process were developed to address the cost and energy issues associated with the conventional Siemens process. However, their cost reduction potential is limited because of the large energy consumption needed to convert silicon into gaseous compounds. Purification of silicon through metallurgical routes offers a promising way for producing SOG-Si from metallurgical grade silicon (MG-Si), because it enhances productivity and can also lower the process costs. Although several metallurgical processes such as directional solidification [l ], vacuum refining [2-4] and plasma oxidation [5-7], have been developed, processing times and refining efficiencies still have to be optimized."

Jan 1, 2013