Search Documents

By Takafumi Kondo, Satoshi Yamashita, Noboru Masuko

"Aqueous oxidation of phosphite ions to phosphate ions with dissolved oxygen was examined at 80° C and pH 4.5 mainly from the stichiometric point of view. Dissolved oxygen alone can not oxidize sodilim phosphite in aqueous solution, but when sulfur dioxide is coexistent, cooperative oxidation for both sulfite and phosphite can occur in aqueous solution. Sulfur dioxide itself is well known as a reducing agent, but it also reacts as an oxidation-promoting agent for other species under a certain condition. So sulfur dioxide is· a promising cheap reagent for waste phosphite processing.IntroductionIn a liquid-phase oxidation reaction using the dissolved oxygen gas, the eoexistence of a copper salt is required generally. Recently, it was reported that the oxidation reaction progressed by using mixed gas of sulfur dioxide and oxygen without the copper catalysis, and the attention has been collected1·3). The present authors also noticed the phenomenon since the early time and have been studying on the reaction mechanism of sulfur dioxide in liquid phase oxidation 4l. It was found that; by using the SO2-O2 mixed gas, the oxidation of phosphite ions to phosphate ions proceeded efficiently in sulfuric acid solutions. In the present report, the applicability of SO2-O2 mixed gas as a cheap oxidizing agent to the industrial processing of the waste solution of electroless nickel· plating is investigated. It is well known that the sulfur oxide dissolved in aqueous solution is oxidized ·by the oxygen in the air to sulfuric acid according to the following equation"

Jan 1, 2000

By John R. Goode

Barrick Gold Corporation operates gold mines in the United States, Chile, 'Peru, Canada, and in the near future, in l1anzania. Barrick meets environmental requirements and limits the potentially adverse effect of cyanide discharge to the environment using various cyanide destruction techniques. Barrick operations use, or have used, natural degradation, ferrous sulphate, hydrogen peroxide, Caro's acid, alkaline chlorination, and the SO/Air process for cyanide destruction. This paper presents a general discussion of the options for controlling cyanide in plant effluent and provides details of Barrick's operating experience including general notes on costs.

Jan 1, 1998

By Bing Hu, Zhu-Cheng Huang, Guang-Bin Xia, Hu Peng, Hua Wang

"The temperature rising characteristics and reaction of the reduction of iron oxide pellets with anthracite fines were studied using a MW-L0316V microwave oven and Leica-DM-RXP polarizing microscope. The results show that iron oxide pellets and anthracite fines have good microwave absorbing properties, and high temperature and reducing atmosphere in a relatively short time for reduction of iron oxide pellets can be achieved. The reduction reaction, which follows the unreacted core model, is rapid and the metallization increases from 13.41% to 56.57% as the reduction end point temperature increases from 850oC to 950oC. The iron is formed, fine grain crystals of iron are transferred and the contraction of metal phase occurs firstly at the surface of pellets, which hinders the reduction reaction at the center of pellets. The rate of reduction reaction increases slowly as the reduction end point temperature increases from 950oC to 1050oC.IntroductionIron ore direct reduction can be divided into gas-based reduction and coal-based reduction. Owing to the effect of the resource and the related economic benefits [1, 2, 3, and 4], the development of gas-based direct reduction technology is restricted, but the development of coal-based direct reduction supported by the relevant industrial policy has very good prospects for application.As a source of clean energy, in China microwaves have been applied in many fields [5, 6, and 7]. Chen Jin studied the reduction of carbon pellets under microwaves, and has made some progress on reduction kinetics and reaction mechanisms. Other countries also have related research in this field [8 - 15], for example, Kotaro Ishizaki and Kazuhiro Nagata have done research on the reduction reaction of carbon magnetite heated by microwaves. The research shows that Fe3O4 is deoxidized to FeO at about 800oC, and FeO is deoxidized to Fe at 1000oC-1025oC. Due to the very good heat absorbing properties of iron oxide pellets and anthracite fines, the selective and rapid heating of microwaves can rapidly raise the temperature of pellets and anthracite, and the iron oxide pellets can be rapidly reduced. In order to provide enough theoretical basis for direct reduction by microwave heating, this paper studies the temperature rising characteristics, behavior of the reduction reaction and the changes of the iron oxide pellets and anthracite in the microwave field."

Jan 1, 2011

By Stavros A. Argyropoulos, Henry H. Hu

"The assimilation of exothermic additions in liquid metals exhibit an array of unique coupled heat, mass and momentum transport phenomena. These phenomena are further complicated by the presence of a moving boundary.This paper describes a) A mathematical model which simulates the complex phenomena, and b) An extensive verification of the mathematical model. The SIMPLER algorithm was employed to solve numerically the pertinent partial differential equations. The computational results indicated that the exothermic heat of mixing led to a rapid increase in temperature around the moving boundary, which produced an enhanced convective flow in the liquid phase. The intensification of fluid flow around the moving boundary resulted in an acceleration of the melting process.Verification of the model was carried out in two contexts. The first was, in a low temperature physical model consisting of ice immersion in different sulfuric acid solutions. In this physical model, both temperature and velocity measurements were carried out. The model results were compared with experimental measurements and they were found to be in excellent agreement. The second context employed high temperatures, involving the assimilation of silicon in high carbon liquid iron. The model was also applied to predict fluid flow, heat and mass transfer for these high temperature experiments and a good agreement was obtained.In addition new dimensionless heat transfer correlations that quantify these complex phenomena will be presented."

Jan 1, 2001

By Blas Melissari, Stavros A. Argyropoulos

"In this paper a computational approach was employed to estimate the melting of solid spheres immersed in different liquid metals and under different convective conditions. The spheres are made from the same material as the liquid metal. The SIMPLER algorithm was implemented using a three-dimensional system in Cartesian co-ordinates. Experimental validation of the dimensionless correlation was carried out in two fluids with vastly different Prandtl numbers. The dimensionless heat transfer coefficient (Nusselt number) is obtained as a function. of the Reynolds and Prandtl number.IntroductionThere is a paucity of dimensionless convective heat transfer correlations applicable to fluids like liquid metals. The existing correlations cover single Prandtl number metals [l] or at most two different Prandtl number metals [2]. A knowledge of heat transfer rates from particles at high flux levels has become increasingly important to the design of energy transfer systems. Melting of solid particles has been studied extensively as it relates to many engineering and environmental applications, such as materials processing, space applications, water resource and energy conversion. When objects are immersed in a fluid with a temperature higher than their melting point, they melt and produce molten material that then mixes with the surrounding fluid. These objects experience not only size variation but also changes in their shape. This observation is more pronounced when the melting occurs in a forced convective flow.Yuge performed pioneering experimental work on the heat transfer from a sphere to air under mixed convection [3]. He suggested procedures for predicting the Nusselt number. Hieber et al [4] studied the spherical system analytically, but their study was limited to flow with small Reynolds numbers. Many researchers studied the melting dynamics of ice spheres in water at different convective regimes. Aziz et al [5] and Hao et al [6] performed measurements of the heat transfer coefficient in the water system by measuring the melting time of ice spheres in forced convection. Hao et al [7] performed visualization studies of ice spheres melting in water under mixed convection regime. All of this groundbreaking work, however, did not involve the use of liquid metals"

Jan 1, 2004

By K. W. Cathrine Solem

Filtration of liquid aluminium is widely used in the industry for the removal of inclusions, and Ceramic Foam Filters (CFFs) are often the filtration media of choice. It is known that PH3 (phosphine) can be released from used phosphate bonded CFFs when in contact with water. Additionally, there is a need of an improved understanding of the thermal stability and the chemical reactivity of these types of filters, as there is limited information available in the public domain. In the present preliminary study, three CFFs, i.e. Substrate 1–3, with varying AlPO4 (aluminium phosphate) content were studied under controlled conditions. Samples of the substrates, as produced and in contact with 5N pure aluminium, were heat-treated in a vacuum induction furnace at 850 °C and 1300 °C, as well as thermally studied at 850 °C using a Differential Scanning Colorimeter connected to a Thermogravimetric Analyser (DSC-TG). All tests were performed under an inert atmosphere of argon (Ar). Mass changes of ∼0.001 % were registered for the pure substrates, and 0.003-0.10 % when in contact with aluminium. In the latter case, diffusion of P (phosphorous) from the bulk of the substrate to the interface was established to have taken place. A colour change, from white to orange/brown, was also observed at the interface, which is a clear sign that a chemical reaction has taken place. As a result, the thermal stability of the substrates can be questioned under present conditions.

By John R. Howell

"SUMMARYThe process of producing a new cast component requires significant design and development by both the design engineer and the foundry engineer. This effort strives to produce a high quality component at minimum cost. To assist in this development, an analysis system, consisting of several computer programs, has been developed to simulate the metal casting process. This system incorporates analytical geometric solid modeling, finite element analysis, results display, and an innovative technique to model the casting mold with surface boundary conditions.The simulation has produced excellent thermal results and work is in progress to develop techniques to enable use of this data in additional analyses to predict porosity and residual stress. Research is required to produce working parametric formulations of material properties. Software development is required to develop good fluid flow analyses for large. 3D models to simulate filling of and convection in molds.Integrating this system of programs is difficult due to proprietary database structures used by software developers. However, an acceptable means has been accomplished' to move data from one program to another. Further software development is required to facilitate this task. INTRODUCTIONThe process of producing a new cast component requires significant design and development by both the design and foundry engineer. This effort strives to produce a high quality, reliable component at minimum cost in a timely manner. The design engineer needs to know the physical and mechanical properites which can be expected in the actual part at critical areas. This will enable him to design a part with minimum material while fulfilling the performance required. The foundry engineer needs to produce a sound, quality casting with a minimal amount of material In risers, the feeding system and machining stock. This must be done with a minimum· investment in tooling and development time and maximum productivity of labor and capital employed. This task Is continually being made more difficult because of increased competition in the marketplace and the use of castings in more severe service. This has strongly influenced the need for high quality castings.The task of developing a new casting is difficult for both the design and foundry engineer. To assist in this development, Deere & Company has been engaged in the development of an analysis system to simulate material processes. A major portion of this program involves simulation of the metal casting process. This analysis system has evolved into a package of computer programs which simulates casting solidification and subsequent cooling through the finite element method"

Jan 1, 1986

By Takao Yasugi

The majority of copper in the raw material migrates to the de-copperized dross that forms when the crude lead produced in the blast furnace is treated in the de-copperizing process. The de-copperized dross is treated in the electric furnace producing matte, speiss, and crude lead. The de-copperized dross also contains the silver. Therefore, it is important to improve the rate of silver migration to the crude lead. For this reason, the behavior of silver between the matte, speiss and crude lead phases was examined. At the same time, several tests to improve the percentage of silver migrating to crude lead were conducted.

Jan 1, 2003

By Derek Fray

Molten salts have been used for the extraction of metals since the 19th century but, recently, many interesting processes have been investigated using mixtures of molten salts and different cathodic materials. In several existing processes, the cation is simply discharged from melt but, in some cases, complications arise when the electrode potentials of the cations are relatively close. For example, the co-deposition of sodium with aluminium in the Hall-Heroult process is discussed in terms of electrode potentials of the components in the molten electrolyte. Similar considerations will also be given to the anodic reactions with respect to chlorine, oxygen, carbon dioxide or monoxide evolution from chloride-oxide melts. In some cases, the cathodic product interacts with the cathode to produce new and exciting products such as nanotubes from carbon or titanates and reduced titanium oxides from titanium dioxide. Depending on the electrode potentials, it is also possible to completely change the cathodic reaction from discharge of the cation in the salt to ionisation of the anions in the cathode. Examples are given of the reduction of metal oxides to produce metals, high technology alloys and nanomaterials and the cathodic refining of liquid metals.

Jan 1, 2006

By J. -S. Park, H. -S. Jeon

This study has been carried out to develop a relatively simple hydrometallurgical process for recovering nickel as nickel sulfate from the spent nickel-cadmium battery residue in which cadmium was removed by vacuum distillation. This residue is composed of 36 % nickel, 20 % iron, and other metals such as cobalt. First, crushing and selective classification were performed to separate iron to produce a nickel-rich product (over 40 % nickel). This product was dissolved in sulfuric acid to obtain a nickel sulfate solution. During the dissolution, nitric acid and air were added to the system to enhance the iron oxidation reaction rate. The free acid remaining in the dissolved solution was neutralized by adding nickel carbonate. Iron was precipitated as iron hydroxide and removed from the solution. The nickel sulfate solution was then fed to a vacuum evaporator to concentrate the solution by evaporation of water and crystallized at around 45°C to obtain nickel sulfate hexahydrate. A recrystallization process could remove other impurities such as zinc and manganese.

Jan 1, 2000

By Lanre M. Oshinowo, André Bakker

"An understanding of the parameters that govern the just-suspended impeller speed, Njs, and the distribution of solids, is critical to the efficient operation of hydrometallurgical and other processes involving solid-liquid suspensions. In this paper, the distribution of solids in stirred tanks under a range of solids loadings (0.5 to 50 vol%) was predicted using CFD and validated against experimental data obtained from the literature. The multiphase flow is modeled using the Eulerian Granular Multiphase model. This paper will also review the established design parameter Njs in the context of scale-up and compare it to the quality of solids dispersion as a means of assessing correct scale-up in suspension tank design. The results of this study will describe a straightforward procedure to obtaining comprehensive information about reactor behavior with complex CFD models.The performance of hydrofoil impellers and a 45° pitched-blade turbine at suspending solids under different agitation speeds was studied. Both single and dual impeller operation have been evaluated. The settled solids fraction for speeds below N js, and the cloud height for impeller speeds above Njs were predicted. The CFD predictions are in good agreement with experimental literature data on velocity distribution and cloud height."

Jan 1, 2001

By J. R. Hartin

Traditionally, many of the advances in casting technology have been made by expensive and time consuming trial-and-error methodology. Modeling of the complex solidification process promises production of castings which are made "right the first time." Even though several models are available for simulation of conventional casting processes, none of them address this relatively new technology of investment cast titanium aluminides. In this paper, modification and application of a transient, three-dimensional, fully coupled fluid flow and heat transfer solidification modeling computer program to the analysis of centrifugally cast titanium aluminides is discussed. The predictive results of the program are compared to qualities of actual cast parts for validation.

Jan 1, 1992

By Jun Sugiyama, Hiroyuki Sano, Toshiharu Fujisawa, Hideki Takeuchi, Chikabiuni Yamauchi

"The solid .state electrotransport (SSE) is one of the most promising methods for the elimination of gaseous impurities from chemically active metals, such as rare earth metals. . In the present work, the SSE process was applied for the oxygen removal from praseodymium. The SSE factors, such as effective valence, were measured under the various temperatures, electric field intensities and initial oxygen contents, and the effects of those conditions on the SSE process were discussed. Furthermore, the solubility of oxygen in .solid praseodyillium was also measured·, by using the SSE method and metal-oxide equilibrium method.IntroductionWhen a direct current is passed through a solid alloy, not only electrons but also solute atoms are transported as charge carriers toward the anode or cathode. This phenomenon is referred to as an electrotransport or electromigration. The solid state electrotransport (SSE) is the application technique, based on this principle, to the pUrification of the solid metallic materials.Rare earth metals are used for various functional materials, such as magnetic, optical or electronic materials, therefore their high purity metals are required [l ,2]. However, rare earth metals are chemically active and have a strong affinity with gaseous elements (e.g. oxygen, nitrogen, hydrogen and carbon). In many cases, eliminating operations of gaseous elements have not been always made commercially, because these impurity elements could not be eliminated by usual treatments applied to the purification of ·common metals. As an alternative method, authors have proposed the· application of the SSE process to the elimination of gaseous impurities from rare earth metals. Investigations have been carried out on the yttrium-oxygen and neodymium-oxygen systems and the results were already reported[3,4]."

Jan 1, 2000

By John E. Litz

The Denver Process is a chemical/metallurgical process for treating radioactive or mixed wastes by extraction and concentration of the radioactive and/or hazardous component(s). The process was developed at bench scale, evaluated at pilot scale, and will be implemented at full scale to remediate a low-level radioactive sludge. Six unit operations comprise the process: feed handling, leaching, solvent extraction, electrodeposition, precipitation, and tailing preparation. Unit operation design may be modified to tailor the process for varying feed stock composition. Radioactive or mixed waste is reacted with aggressive reagents in an aqueous phase, solubilizing targeted components. These components are then selectively extracted and concentrated in subsequent steps. Waste solids and solution exit the process as non-hazardous non¬radioactive waste. Dehydration of aqueous waste by submerged combustion eliminates aqueous discharge.

Jan 1, 1994

By Tapio Ahokainen

Commercial modeling software and increasingly efficient computers enable today computational fluid dynamics (CFD) to be used as a tool for a more comprehensive understanding of the flash smelting and flash converting processes. CFD modeling can be used for design purposes in many technological areas, but some relevant physical and chemical phenomena are still missing in the commercial CFD solver packages. Therefore, Outokumpu has developed several subroutines for solvers to describe analytically the phenomena in the flash smelting environment. CFD sub-models of the copper concentrate and matte combustion, the particle-particle-wall interactions and for more accurate radiation properties of the gas-particle suspension are needed to ensure reliability of numerical simulations on the flash smelting and converting vessels. Several basic combustion phenomena can be modeled with a number of commercial solver packages, but own models for combustion must be included in the case of sulfides as the feed. Mathematical models for copper concentrate and solid matte combustion have been developed during the last decade. They calculate compositions of the particles and temperature, based on extensive experimental research work in laboratory scale furnaces with validated kinetic parameters and temperature measurements of reacting particles. Inside and near the concentrate burner, the standard particle tracking methods do not include the particle-particle interactions and the particle-wall interaction is too much simplified to describe properly the particle collisions with the distribution cone. Therefore, additional models for the particle-particle as well as particle-wall interactions have been written and implemented. For thermal radiation, models for the emissivity of a gas mixture and for calculating the particle absorption and scattering coefficients have been developed and tested. In addition, the advanced radiation model can be used i.e. in the enhanced simulation of off-gas cooling in the heat recovery boiler area of the flash smelter.

Jan 1, 2006

By Mingchun Lu, Yunjin Yan

"This paper describes a new type of bell-less top installed on two relined blast furnaces at Tianjin Iron Works and another new-built at Echeng Steel Works in China. A compact installation was presented for existing furnace top innovation with the least reformation of existing top structure. A water cooling and sealing gearbox used for driving the chute discards completely the traditional idea about necessity of continuous supply of cooling gas for the gearbox. A multiring spreader and its dozens of selectable charging patterns enables an operator to achieve optimal burden distribution easily under the condition of a minimum investment of control equipment. Some useful conclusions have been made through a small-size model and a full-scale model tests. The excellent characteristics and the superiority of the new top have been fully proved at both works.IntroductionIt is known that there are some advantages with the bell-less top of a blast furnace: suitable for furnace to operate at high top pressure, capable of realizing optimal burden distribution, less investment and maintenance costs, etc. But why do most of blast furnaces in the world still remain using conventional bell top. One of the concerns is that current bell-less tops still have some shortcomings:1) The height of a current bell-less top might be lower than three or four-bell top or two-bell plus valve seal top, but still much higher than an ordinary two-bell top.2) The gearbox for driving the rotating chute is positioned onto the lip of the furnace. It is subject to attack by crude hot bf gas. Continuously pumping cooling gas (pressurized nitrogen or clean bf gas) into the gearbox is indispensable to withstand such a harsh working environment. Therefore, a plant furnished with blower8 and its ancillaries has to be built, It involves a large sum of investment and much expense in its annual running costs.3) Although the current bell-less charging device as a means of controlling burden distribution possesses a wide range of freedom, it is difficult to realize optimal burden distribution. There are some large blast furnaces realized better burden distribution, but sophisticated control means is indispensable so that the investment of control means is much higher."

Jan 1, 2000

At its Copper Cliff Smelter, Inco Limited processes a bulk copper-nickel concentrate in two Inco oxygen flash furnaces, The furnace matte is converted to low iron matte for further processing, and the molten converter slag is recycled to the flash furnaces. The flash furnace slag is discarded. Natural gas or coke supplies a minor proportion of the process heat requirements, thus limiting the flash furnace matte grade to the target 45% CuNiCo. Coke addition to the furnace offers a potential means of establishing a reducing barrier on the surface of the molten bath. Such a practice would be conducive to producing sulfur deficient (partially metallized) matte. At the same iron content, the iron activity of sulfur deficient matte is higher than that of regular matte. A supernatant coke barrier could, therefore, be expected to lead to higher nickel and cobalt recoveries, reduction or even elimination of magnetite furnace bottom build-up and increased sulfur elimination as S02 in the flash furnace. Preliminary experiments were conducted in a miniplant flash furnace to investigate the effect on process metallurgy of the suggested modified scheme. This paper presents the experimental results and discusses key aspects of the possible operation of the Inco flash furnace with a coke barrier separating the freeboard from the molten bath.

Jan 1, 2003

By J. D. Budai, D. H. Lowndes, D. P. Norton, D. K. Christen, B. C. Chakoumakos, J. R. Thompson, B. C. Sales

"Artificially-layered structures, consisting of (Sr,Ba,Ca)CuO2 layers in the tetragonal, ""infinite layer"" crystal structure, have been grown by pulsed-laser deposition. Superlattice chemical modulation is observed for structures with SrCuO2 and (Sr,Ca)CuO2 layers as thin as a single unit cell (~3.4 Å). In addition, novel thin film superconductors were formed by using the constraint of epitaxy to stabilize SrCuO2/BaCuO2 superlattices in the infinite layer structure. Using this approach, two new structural families, Ba2Srn-1CUn+l O2n+2+ d) and Ba4Srn-1CUn+3O2n+6+ d, have been synthesized which superconduct at temperatures as high as 70 K. These results represent not only the synthesis of new structural families of superconductors, but also demonstrate that pulsed-laser deposition and epitaxial stabilization can be effectively used to engineer artificially-layered thin-film materials.Advances in the understanding of epitaxial thin-film growth of the cuprates have heightened interest in creating artificially layered superconductors.(1-3) The layered structure of the copper oxide superconductors opens the possibility of forming novel materials by carefully controlling the epitaxial growth of sub-units of the crystal structures. Thin-film growth methods offer unique opportunities for the atomic engineering of new high-temperature superconducting (HTSc) materials through the ability to form artificially layered crystal structures.(4,S) Moreover, the surfaces of single-crystal substrates provide an ""atomic template"" that can be used to stabilize epitaxial films in metastable crystal structures.An important development in forming new HTSc compounds by artificially-layered epitaxy has been the epitaxial growth of (Sr,Ca)CuO2 films in the so-called infinite layer structure.(6-17) This material, which can be considered the parent compound of all of the copper oxide superconductors, has a relatively simple, layered structure consisting of CUO2 planes separated by planes of alkaline earth elements.(18) Bulk synthesis of (Sr,Ca)CuO2 with the tetragonal, infinite layer structure generally requires the use of high pressure and high temperature bulk processing techniques.(19-24) However, recent experiments show that this metastable compound can be epitaxially stabilized at less than atmospheric pressure, resulting in the growth of tetragonal (Sr,Ca)CuO2 single crystal thin films of the infinite layer defect perovskite structure by pulsed-laser deposition (PLD) over a wide range of growth conditions.(6-17)"

Jan 1, 1994

By Guo Xian Jian

In this paper, the gas composition calculations on the FeS-H20 system at elevated temperature were performed. The results show that the gas constituents are H20, S02, H2S, Sj, H2, and O2 and the order of the compositions partial pressure is PH2o> PH2 >PSO2 >PSj >PH2S >PO2 when FeS is oxidized into Fe304 and FeO by steam in the temperature of 700°C -960°C, the Claus process occurs with the decrease of gas temperature in the off-gas and the S02 and H2S in the gas are transformed into Sj , the critical temperature for the transformation is 600°C, and conversion percent of Sj is 97% at 330°C,

Jan 1, 1996

By H. Itou, J. M. Toguri, T. Takasu

"Marangoni convection has a large effect on materials processing since it remarkably enhances heat and mass transport near a free interface. To have a better understanding of the behavior of Marangoni convection in molten salts, observation and calculation of the thermal convection in a column held between a pair of Pt disks (2mm in diameter) fixed to hot thermocouples, were carried out.Calculated flow patterns and absolute values of fluid velocity agreed well with the observed results. When the temperatures of both disks were set equal and the height of liquid column was 2.4mm, four vortexes were generated on the longitudinal plane with a typical velocity of 15mm/s . From the calculated isothermal lines, the behavior of the enhanced heat transport was obvious; fluid heated at the disks was conveyed to the free surface and the cooled fluid at the surface was conveyed to the inner region. Increase in the height of column and/or increase in the absolute value of applied temperature difference lead to increase in fluid velocity lineally in this calculation range.IntroductionIn the processing of many materials, fluid flow of melts has a large effect on refining capabilities and on the characteristics of the produced material. Marangoni convection is generated when a spatial gradient of interfacial tension exists on the free interface due to differences in temperature and/or concentration. This promotes heat and mass transfer locally near the interface.For example, Marangoni convection which are created following the dissolution of interfacially active materials, increases the local corrosion rate of refractories (1-3). In absorption and discharge of oxygen and/or nitrogen in a molten metal, the effective reaction rate of the gas components through the gas-liquid interface increases (4-7). Also in crystal growth, the distribution of defects and impurities in the crystal is affected and this can alter its quality (8-11 ). The temperature distribution is also affected and can greatly change the productivity. For these reasons, the effectiveness of Marangoni convection in refining should be quantitatively clarified and its control should be better understood."

Jan 1, 1997