Search Documents

By Junli Xu, Xianwei Hu, Kaiyu Xie, Liangxing Li, Zhongning Shi, Zhaowen Wang, Bingliang Gao, Aimin Liu

Al-Ti master alloys were prepared by aluminothermic reduction between the dissolved titanium dioxide and aluminum in cryolite melts at 960oC. The kinetic analysis by differential scanning calorimetry indicated that the apparent activation energy of the reaction of reducing titanium dioxide by aluminium is 22.3 kJ/mol, and the reaction order is 0.5. The products were analyzed by means of X-ray diffraction, X-ray fluorescence, scanning electron microscopy and energy dispersive spectrometer. Results from X-ray diffraction showed that the phase compositions of produced alloys are Al and Al3Ti. In addition, Al-Ti master alloys containing 2-6 mass% Ti were formed at different reduction time of 2-5h in aluminothermic reduction experiment.

Jan 1, 2015

By S Ivanaj

Oxidation of chalcopyrite from a copper concentrate in an electrochemical cell having separate-anodic and cathodic compartments was studied by potentiodynamic, potentiostatic and galvanostatic experiments. Oxidation of chalcopyrite particles was mainly realized by in-situ electrogenerated FeIII and Cull ions. The use of an ion exchange membrane, in this experimental set, avoids the reduction of the ferric and cupric ions. The effects of particle size, current feeder potential, current intensity and pulp density were studied to optimize the oxidation parameters. Selective complete oxidation of chalcopyrite was achieved. Oxidation of pyrite was negligible. Mom than 85 % of sulfur was obtained in its elemental form. This may allow selective extraction- of copper from the chalcopyrite concentrate without sulfur dioxide emissions.

Jan 1, 1995

By Xianzhong Hu, Qingbo Yu, Qin Qin

"Oxy-fuel combustion is one promising option among different CO2 capture technologies. The object of this work is to show the effect of CO2 dilution on the laminar flame speed, flame structure and reaction pathway for methane oxidation at O2/CO2 atmosphere. In this paper, kinetic simulations were made using CHEMKIN with GRI-Mech 3.0 mechanism, which was validated by experimental data of laminar flame speeds. The reaction pathway of CH4/O2/CO2 mixture was derived from reaction flow analysis. Results showed that CO2 decreased the flame speeds, lowed concentration of major intermediates and increased flame thickness. The major chemical pathways of CH4 to CO2 have not changed at O2/CO2 atmosphere comparing with O2/N2 atmosphere, but contributions to major intermediates formation from each related elementary reaction have changed a lot. In addition, the secondary chemical pathways of CH3 oxidation became different at O2/CO2 atmosphere.IntroductionNowadays, as the global economy is developing fast and the pollutant levels (especially global warming gases) are increasing, a lot of researches are focusing on new technologies which can cut greenhouse gas emissions [1]. New technologies have been developed, such as high Mild combustion, flameless combustion and Integrated Gasification Combined Cycle (IGCC) technology. Oxy-fuel combustion using carbon dioxide as dilution of oxygen is considered as a promising approach for reducing CO2 emission. The properties of CO2 in flue gas are different from N2 in air. On one side, the transport, thermal and radiative properties of CO2 are different from N2 in air. On the other side, CO2 is not inert but directly participates in chemical reactions and recent researches [2-5] also present the chemical effects of CO2 on oxy-combustion. For these differences above, the characteristics of oxy-combustion are not alike with traditional combustion. It is necessary to develop the investigation of oxy-combustion.In all combustion processes, laminar flame speed is an important parameter, which links to combustion behavior. Analysis of this parameter can provide a better understanding of the effect of CO2 dilution in the process of oxy-combustion. Several investigations of laminar flame speeds of oxy-fuel flame have been improved, including experimental measurements and numerical computations. A.A.Konnov et al. measured the flame speeds of oxy-methane and oxy-ethane mixture using the heat flux method with a flat burner [6-8]. V. Ratna Kishore et al. used heat flux method to measure the adiabatic laminar flame speeds of hydrogen [9] and H2/CO mixture [10]. Adiabatic laminar flame speeds of hydrogen in O2/CO2 atmosphere are much smaller than in O2/N2 or in O2/Ar. The burning velocity of H2/CO mixture decreased with increase in CO2 content in mixture. Jeongseog Oh et al. [11] used Bunsen flame to measure the laminar flame speeds of premixed oxy-methane mixture with angle methods in atmospheric condition (300k, 1atm). The results were SL=29.1cm/s and SL=29.5cm/s respectively derived from oxy-methane from CH* images and Schlieren photos."

Jan 1, 2014

By Jilai Xue, Jun Zhu, Germain Kouma Ndong, Qiaochu Liu, Luxing Feng

"High-temperature oxidation and corrosion behaviors of Ni-Fe-Cr alloy as inert anodes for aluminum electrolysis have been studied in oxygen and molten electrolyte. The oxidation and corrosion scales on the anodes tested were analyzed using XRD and SEM-EDS. The oxidation rate is found to increase with increasing temperature from 700 °C to 950 °C, which can be approximately described by an inverse power rate function. The oxidation scales at 750 °C, 920 °C and 950 °C contain Cr-rich phase along with FeCr204 and (Feo.6Cro4)203. The corrosion extent of Ni-Fe-Cr anodes in electrolyte is dominated by temperature, which can make the scales thickness double from 700 °C to 750 °C or from 920 °C to 950 °C. Cr and Fe in the scales on the anode in electrolysis corrode preferentially into the molten electrolyte, while the nickel oxides could better sustain the corrosive environment in electrolysis. The results can be useful for developing inert anode material for potential application in aluminum electrolysis.IntroductionThe development of an inert anode for aluminum electrolysis has been a long term goal for aluminum industry. The present process for primary aluminum production involves the reduction of alumina with carbon anodes, forming carbon dioxide:"

Jan 1, 2013

By Linfeng Zhou, Yuanbo Zhang, Rong Sun, Guanghui Li, Zhiwei Peng

The volatilization behavior of rhenium in the process of oxidative roasting of a molybdenite concentrate with 43.55 wt.% Mo and 321 g/t Re is explored based on identifying the relationship between volatilization of sulfur and rhenium. The results show that rhenium volatizes after sulfur and precise control of operating variables plays an important role in separating rhenium from molybdenite concentrate. When the roasting parameters are optimized as follows: roasting temperature 600 oC, roasting time 2 h, roasting atmosphere air, the volatilization rate of rhenium reaches 89.26%. In the process of solidification roasting, the volatilization rate of rhenium decreased almost to zero when the ratio of slaked lime/molybdenum concentrate was 10% excess to the stoichiometric requirement at 600 oC for 2 h.

Jan 1, 2016

By Signe Kjelstrup, Leiv Kolbeinsen, Marit Takla, Odne Burheim

"This paper presents a Seebeck-type solid state thermoelectric power generator unit, designed to investigate the possible recovery of radiant waste heat at a silicon plant in Norway where roughly 703 of the total input energy leaves the process as waste heat. The unit is 0.5 m x 0.5 m and is built of 36 commercially available thermoelectric devices based on bismuth telluride p - and n - type semiconductors. Thermoelectric devices directly convert fractions of a heat flux into an electric current (and vice versa) and may therefore be convenient for recovery of waste heat when the heat source is discontinuously available. The purpose of the thermoelectric unit is to give a proof of principle of direct thermoelectric energy conversion on a large scale. We present the initial test results for the generator. At an average temperature difference between the heat reservoir and the heat sink of 98°C, we measured an open ciruit potential of 23.44 ± 0.06 V and the matched load power output was 39.4 W. At a temperature difference of 220°C, we predict the generator open circuit potential to be 40.8 V, which corresponds to an efficient Seebeck coefficient of 242 µV /K, and matched load power output to be 120 W.IntroductionThe world's power demand increases, driven by population growth and increase in living standards and simultaneously global warming is recognized as one of our most serious environmental problems. According to the International Energy Agency[l], the industrial sector accounted for approximately one third of the total global power consumption in 2005. Increasing the energy efficiency of industrial processes thus represents a benign and substantial route for addressing environmental concerns and energy security.In accordance with the second law of thermodynamics, all real processes are associated with lost work which is energy being dissipated as heat to the surroundings[2]. Improving the energy efficiency of industrial processes is about finding the process path associated with a minimum of lost work. The more evenly the losses are distributed, the smaller are the losses [3, 4]. In addition, exploiting the unavoidable waste heat will contribute to improve the energy efficiency. The waste heat may be utilized for heating purposes and/or it can be upgraded to work for instance via some sort of heat engine."

Jan 1, 2012

By Milton Manrique

The thermal decomposition of zircon concentrates in a non-transferred arc thermal plasma flow reactor was investigated. The plasma torch was operated using argon-helium and argon-nitrogen mixtures as plasma gases. The effect of power, feed rate and zircon particle size was studied. Products were characterized by XRD, SEM, TEM and chemical analytical techniques. The average size of the product was in the submicron range. Free energy minimization plots were used to predict the behavior of the system under different operating conditions. A kinetic model to describe the different stages of the thermal decomposition process is proposed. The results of this investigation showed a significant segregation process where the zirconia-rich product is collected in the reactor while the silica-rich product is collected in the filter. TEM analysis indicated that silica vapor is condensed on the surface of the zirconia particles.

Jan 1, 1996

By Jan Dabrovska

Solidification and cooling of a continuously cast slab and simultaneous heating of a mould is a very complicated problem of transient heat and mass transfer. Nowadays, the solving of such a problem is impossible without numerical models of the temperature field not only of the slab itself, while it is being processed through the whole concasting machine (Caster), but of the mould as well. Two original numerical models have been developed and used in the investigation of a continuously cast steel slab. The first (one of two) 3D model of the temperature field of a concasting is capable of simulating the temperature field of a caster. Experimental research and data acquisition have to be conducted simultaneously with the numerical computation-not only to confront it with the actual numerical model, but also to make it more accurate throughout the process. After computation, it is possible to obtain the temperatures at each node of the network, and at each time of the process. The utilization of the numerical model of solidification and cooling of a concasting plays an indispensable role in practice. The potential change of technology-on the basis of computation-is constantly guided by the effort to optimize, i.e. to maximize the quality of the process. The user can therefore choose any appropriate longitudinal or cross-section of a slab and display or print the temperature field in a 3D or 2D graph whenever necessary. The second numerical model of dendritic segregation of elements assesses critical points of slabs from the viewpoint of their increased susceptibility to crack and fissure. In order to apply this model, it is necessary to analyze the heterogeneity of samples of the constituent elements (Mu, Si and others) and impurities (P, S and others) in characteristic places of the solidifying slab. The numerical model, based on measurement results obtained by an electron micro-probe, generates distribution curves showing the dendritic segregation of the analyzed element, together with the distribution coefficients of the elements between the liquid and solid states. The combination of both models enables the prediction of cracks and fissures in critical points of the continuously cast carbon-steel slab. Both models had been applied in the industrial investigation of a cast low-carbon-steel slab.

Jan 1, 2003

By Akira Yazawa

The continuous converting has attractive advantages, but further experimental studies on the slag are required, when the process is considered thermodynamically. The summarized experimental results, obtained with iron silicate slag and with ferrite slag and results relating to actual operation are described and discussed in this paper. At first, equilibrium relations at controlled oxygen potentials among three melts of silica-saturated iron silicate slag, white metal and copper are investigated. The results obtained for slag agreed with the thermo-dynamical predictions. Namely, to suppress a magnetite separation, the system should be at high temperature; in the case of the S02 gas of 10 kPa, for example, magnetite did not separate from the slag at 1300°C, but at 1250 and 1200°C the slag was saturated with magnetite even at lower SO2 pressure. Copper losses caused by dissolution of its oxide increased with increasing 5O2 pressure, and were slightly lower than those reported in previous papers without the matte phase. When the continuous converting slag does not contain any appreciable amount of silica - in contrast to the conventional smelting slag ferrite slag would form, but there was no published fundamental work on-ferrite slags. Hence, the melting behavior and activities of the system CaO-Cu2O-Fe2O3 have been investigated. The liquidus diagram in nitrogen atmosphere was, constructed using both thermal analysis and quenching method. The low liquidus temperature region without magnetite separation was found in the part of low content of Cu2O in the section Cu2O-CaO-Fe2O3 and Cu2O?Fe2O3¬CaO?Fe2O3. The activity of Cu2O in the system was also measured. Slags, composed-of CaO, Cu2O and Fe2O3, in platinum crucibles were equilibrated at 1250°C and controlled oxygen pressures. After quenching, the-composition of slag and platinum crucible were determined qualitatively. On the thermo-dynamic analysis of the results, iso-activity lines were drawn, and it was shown that the slag in the converting furnace of the Mitsubishi process exists in an adequate region of the composition triangle.

Jan 1, 1976

By Xiang (Frank) Chen, Mikhail A. Sokolov, Randy K. Nanstad

"Nickel-base Alloy 617 is being considered as a structural material for application in the secondary heat exchanger of the New Generation Nuclear Plant, a very high temperature gascooled reactor. Thermal aging of Alloy 617 plate and welds is being performed with tensile, Charpy impact, and fracture toughness tests conducted at temperatures to 950°C. Results of testing for thermal aging to 5,300 h have been obtained and are presented; varying effects of thermal aging temperature and time on fracture toughness are observed. The 9Cr-1MoV (Grade 91) ferritic steel is a candidate for structural applications in the sodium fast reactor. Fracture toughness testing of unaged Grade 91 steel has been performed to evaluate specimen size effects in preparation for future testing of the material in the thermally aged condition. Results for material in the mill-annealed and heat treated conditions are presented and show that this heat of Grade 91 steel does not indicate a small specimen bias on the fracture toughness Master Curve reference temperature.IntroductionThe behavior of advanced engineering materials will be a critical factor in the performance of Generation IV reactors. In the case of materials utilized for reactor internals and pressure vessels, the effects of irradiation and long time thermal aging are major issues. The environmental conditions for some of the Gen IV reactors are generally beyond present day reactor technology, requiring new experimental data as well as analytical predictions of expected behavior of candidate materials at conditions for which there are only sparse or no experimental data. The radiation environments and several key materials issues for some Gen IV systems are summarized in Section 5 of [l] as a starting point for modeling and microstructural analysis of Generation IV materials. More detailed descriptions of materials research and development needs and recommended activities for the SCWR, VHTR, and GFR are provided in references [2, 3, 4], respectively. Mansur, et. al. provide an interesting comparison of materials needs for Generation IV reactors to corresponding needs for fusion reactors and spallation neutron sources in [5]. The two primary candidate reactor systems currently undergoing significant research and development by the U.S. Department of Energy (DOE) are the New Generation Nuclear Plant"

Jan 1, 2012

By Ray D. Peterson

The hydrothermal enrichment of chalcopyrite by CuSO, occurs in the absence of oxidizing or reducing reagants. Kinetics of the enrichment process were determined under autoclave conditions in the temperature range of 398°K (125°C) to 473°K (200°C). Covellite (CuS) formed first, followed by digenite (Cu,,,S), with rejection of iron to the solution. The kinetics have been explained by a model which includes diffusion during the first stage formation of covellite followed by surface reaction rate control when digenite be. The effects of particle size, temperature, pH, and cupric and ferrous ion concentrations on the leaching kinetics were examined. Additions of ferrous ion to solution markedly increased the rate of reaction.

Jan 1, 1994

By S. J. Macnaughton

The Olympic Dam deposit in South Australia contains 2% copper and 0.06% uranium, which are present as bomite, chalcocite, chalcopyrite, uraninite, coffinite and brannerite. The copper is principally recovered using flotation. The flotation tailings, which contain the bulk of the uranium, and typically 0.3% copper, are leached using sulfuric acid and an oxidant at atmospheric pressure. This paper describes the results from a detailed investigation of the leaching behaviour of this ore. The influence of acid strength, Fe3+ concentration, redox potential and temperature on the dissolution of uranium and copper minerals was studied. Experimental data were obtained from extensive batch leach tests. Mineralogical studies, including SEM analysis of individual mineral grains, were also carried out. The maximum extraction of uranium determined experimentally was in good agreement with that predicted from SEM determination of the proportion of refractory brannerite in the ore. The leaching rate of copper was found to be independent of acidity over the range 20 g L-' to pH 2.0, but was strongly dependent on temperature and Fe3+ concentration. Uranium extraction showed a weak de endence on acidity, Arrhenius temperature dependence and a complex co-dependence on F>+ concentration and redox potential. A kinetic model has been developed, which predicts copper and uranium dissolution. This model will be incorporated into a more comprehensive model that can be used to predict leach performance and reagent consumption for specific ore samples.

Jan 1, 1999

By Jana Dobrovska

Solidification and cooling of a continuously cast slab and simultaneous heating of a mould is a very complicated problem of transient heat and mass transfer. Nowadays, the solving of such a problem is impossible without numerical models of the temperature field not only of the slab itself, while it is being processed through the whole concasting machine (Caster), but of the mould as well. Two original numerical models have been developed and used in the investigation of a continuously cast steel slab. The first (one of two) 3D model of the temperature field of a concasting is capable of simulating the temperature field of a caster. Experimental research and data acquisition have to be conducted simultaneously with the numerical computation-not only to confront it with the actual numerical model, but also to make it more accurate throughout the process. After computation, it is possible to obtain the temperatures al each node of the network, and at each time of the process. The utilization of the numerical model of solidification and cooling of a concasting plays an indispensable role in practice. The potential change of technology-on the basis of computation-is constantly guided by the effort to optimize, i.e. to maximize the quality of the process. The user can therefore choose any appropriate longitudinal or cross-section of a slab and display or print the temperature field in a 3D or 20 graph whenever necessary. The second numerical model of dendritic segregation of elements assesses critical points of slabs from the viewpoint of their increased susceptibility to crack and fissure. In order to apply this model, it is necessary to analyze the heterogeneity of samples of the constituent elements (Mn, Si and others) and impurities (P, S and others) in characteristic places of the solidifying slab. The numerical model, based on measurement results obtained by an electron micro-probe, generates distribution curves showing the dendritic segregation of the analyzed element, together with the distribution coefficients of the elements between the liquid and solid states. The combination of both models enables the prediction of cracks and fissures in critical points of the continuously cast carbon-steel slab. Both models had been applied in the industrial investigation of a cast low-carbon-steel slab.

Jan 1, 2003

By C. P. Broadbent

The process metallurgy of many nickel laterite smelting operations is controlled largely by the physical chemistry of the slag system. Consequently, accurate phase diagram information and physicochemical property data are essential to interpret and predict the behaviour and optimisation of the overall process. The system (SiO2 - MgO - FeO) first described by Bowen et al. between 1914 and 1935 and the system (SiO2 - MgO - FeO - Fe2O3) determined by Muan and Osborne in 1956 are of particular relevance to the study of nickel laterite smelting slags as these systems form a reasonable approximation of many industrial slags. This paper will review the literature for these (and other relevant) phase systems. In addition data available from the ternary and (quasi)ternary phase diagrams will be compared with data obtained from industrial slags. The physico-chemical properties under consideration will include; melting and crystallisation behaviour (solidus and liquidus temperatures), viscosity - temperature relationships, heat flux considerations, electrical conductivity, surface and interfacial tension.

Jan 1, 1993

By Keiji Okumura, Hong Lan, Masamichi Sano

"Magnesium has been well known as one of the excellent agent of desulfurization, deoxidation, etc. in iron- and steelmaking because of its high reactivity. Since its production is energy nitensive, the price is very high. Thus magnesium has not widely been used as an agent in metal refining ·processes. In the present work, reduction of magnesia by carbon and aluminum is examined. Because the reaction is very sensitive to temperature, non-isothermal technique is applied to investigate the reaction kinetics. The effects of heating rate .and reaction interface area on the reaction kinetics .are studied. .A kinetic model is proposed based on the experimental results.IntroductionFor the metal refining processes, the production of high purity metal has been one of the most important research subject. In order to produce the high purity metal, the amount of the refining agent and the consumed energy is increasing. Therefore, if the future earth environment is taken into consideration, the further efficient and energy saving processes are required. Generally, Cao slag has mainly been used for desulfurization of molten iron as a desulfurization agent. However, environmental problems, such as recycling of disposal slags, are still unsolved. On the other hand, the deoxidation of molten steel is mainly performed by addition of aluminum. Therefore, the alumina inclusions· remain in. molten steel and the great efforts have been made to the removal of the inclusions. The affinity of magnesium with sulfur and oxygen is strong, and magnesium is well known as a powerful desulfurization and deoxidation agent. However, magnesium has been seldom used for the steel' refining process. The following reasons are considered. Generally, magnesium is expensive. And at iron- and steelmaking temperatures, the refining efficiency becomes low due to its abruptly .evaporation."

Jan 1, 2000

By Martín Reyes

Iron concentration by ion flotation techniques, from sulfate solutions in presence of zinc was studied in a laboratory flotation column by continuous mode using a synthesized sand shell plate sparger, an anionic collector, promoters, and surfactant propylenglicol 400. Effects of experimental parameters, such as the concentration of reagent: potassium amyl xanthate and dithiophosphate, superficial gas velocity, and superficial liquid velocity were studied in terms of the recovery and enrichment of Fe (III). The results show that the iron elimination from the solution in the presence and absence of zinc (II) are of 59 % and 72 % respectively. The flotation efficiencies decrease with an increase of the concentration of xanthate and air flow rate. The best recovery was achieved, at Jg 0.1 cm/s, Jl 0.72 cm/s and a mixture of xanthate and dithiophosphate as promoter. It was also demonstrated that it is possible to selectively separate and concentrate iron (III) from solutions containing zinc. The behavior of iron ion flotation in a flotation column is reported in this research as a gas hold up, superficial gas velocity, bubble size and superficial bubbles area flux function. The results show that the extraction of aqueous ions of metal iron by ion flotation is strongly affected by: [C], Jg (cm/s), eg (%), Db (mm), Jl (cm/s), y Sb (s-1).

Jan 1, 2009

By Kent D. Peaslee

When supersonic oxygen jetting is used in electric furnace steelmaking, the lance jet axis is typically inclined at larger angles to the vertical than is used in basic oxygen steelmaking. The supersonic jet is important not only for refining the steel, but for sustaining the foamy slag in the electric furnace. Physical modeling was carried out using a room temperature, two-dimensional slice apparatus in which mercury represented the metal and water-glycerol the slag. High speed video photography with image analysis was used to study the effects of the angle of inclination on the slag/metal bath. Mathematical modeling using FLUENT provided a method of predicting flow patterns and the trajectories of drops injected at various points in the flowfield. A one-half scale water model of the system, in which lance heights, angles, and flowrates simulated those encountered in steelmaking operations, was used to provide a better understanding of some aspects of the oxygen blowing process in electric furnace steelmaking.

Jan 1, 1994

This report presents overview of the authors' results of thermodynamical investigation of some important metallurgical processes and phenomena. Among them: carbo thermal reduction of silica, and boron oxide, silicon carbide obtaining from different precursors, investigations on copper and manganese ores processing, investigations on pollution from different sources. Special attention is paid-to thermodynamical approach to investigation and estimation of multi component multi element compositions stability, developed by the authors. Also are discussed some auxiliary but' important for the thermodynamical investigations issues.

Jan 1, 1997

By Xie Zhang, Jiang Diao, Xiaojun Li, Bing Xie

"Vanadium is widely utilized as an important element to improve the quality of alloyed steels in iron and steel industry. Moreover, the crystallization process of V-containing spinels plays a significant role due to the aim of achieving a fairly larger size of spinel crystals in vanadium extraction. Thus, a comprehensive exploration and discussion on the competitive precipitation and growth of vanadium, chromium and titanium spinel crystals in vanadium slag during a cooling process were carried out in the present study on the basis of a crystallization kinetic model. To be more specific, the abilities of forming crystal, precipitation order and suitable temperature range for crystallization of the three kinds of spinels were discussed with both theoretical and experimental methods.IntroductionVanadium is an important element which is used almost exclusively in ferrous and non-ferrous alloys due to its excellent physical properties such as high tensile strength, hardness, and fatigue resistance [1]. Various kinds of processes of vanadium extraction from V-containing hot metal are available at present. However, only four processes which are the Highveld process (South Africa), the Pan-steel process (China), the NTMK duplex process (Russia) and the New Zealand steel process (New Zealand) take the dominant positions [2, 3]. Although the type of reactor and the initial vanadium content in hot metal are different in these techniques, they share the same principle that vanadium is oxidized by oxygen into vanadium slag."

Jan 1, 2012

By Manabu lshimaru

Solid-liquid interfaces as well as defect formation processes during crystallization of melted silicon (Si) have been examined by using molecular-dynamics (MD) simulations. The ordinary Langevin equation employing the Tersoff interatomic potential was used for the calculation of the atomic trajectories. A solid-melt interface has a tendency to form {]] I} microfacets during the crystal growth in the [001] direction, while it is essentially flat during the (Ill) crystallization. Because of the difference of the interfacial structures, the structures of grown-in defects and defect evolution depend strongly on the growth direction. We discuss defect structures and their formation processes within the context of our results as well as previous experimental observations.

Jan 1, 2001