Liquids are known to characterized by the structural homogeneity as well as the fluidity. This is true in various simple metals and alloys in the liquid states. However, some of liquid alloys, particularly systems including the semiconducting elements such as Te and Se, indicated a peculiar behavior in their electronic and thermodynamic properties at a certain concentration near the melting point and then the structural inhomogeneity is quite feasible in such liquid alloys. An attempt has been made in this paper to provide current information on the evidence of structural inhomogeneity in liquid alloys using the X-ray diffraction results of liquid Ga-Te alloys as an example.
The theory of diffusion in melts is outlined. The inter-diffusion (or mutual diffusion) coefficient D_<12> in a binary molten mixture can be divided into two contributions: one factor arises from concentration fluctuations related to stability of the mixture and the other one can be written in terms of self-diffusion coefficients of the respective components. The concept of the velocity autocorrelation function and the memory function is briefly described, and comparisons between theory and experiment are shown for diffusion properties of liquid metals. It is emphasized that microgravity experiments are very important in the studies of diffusion in melts, since they are free from the effects of convection and gravity inevitable in the experiments made on the earth.
Computer simulation of molecular systems represented by Molecular Mechanics, Monte Carlo (MC) and Molecular Dynamics (MD) simulations have become more and more powerful tools thanks to the development of fast computers. The MC and MD methods, which can be employed for liquid systems, are described in comparison with each other. The Ewald method essentially needed for calculating the Coulombic potentials and forces is explained. For MD simulation of polyatomic rigid molecules, Euler's equation has to be used and a special technique called the quanternion method is usually invoked. For these simulations, the validity of pair potentials holds the key to reproducibility of the real systems.
The physical properties of oxide melts such as silicate, aluminate, and borate are of importance not only to the extractin metallurgist as slags but also glass technologist. Recently, theese properties have been important in the production dof solid ionics materials from oxide melts. Numerous physical properties of oxide melts such as density, viscosity, electric conductivity, diffusivity and thermal conductivity have bean reported on the oxide melts, mainly those of silicate melts. However the values obtained from different studies are not in agreement, the reason being mainly the difliculty of measureing at high temperature. Some models have been constracted of the structure of oxide melts such as silicates and borates with reference to the physrcal propertres of oxide melts and X ray diffractron measurements, infrared spectroscopic studies, and Raman spectroscopic studies for quenched glassy samples.
Homogeneous crystal growth is expected in space, because buoyancy driven convection does not occur under microgravity conditions. In some cases, however, growth melt may possibly be controlled with Marangoni convection owing to the strong effect of surface tension on a free surface, and ideal crystal growth can't be obtained. So, may studies have been made in order to investigate the fundamental characteristics of Marangoni convection. In Japan, we have a study of flow visualization on Marangoni convection already planned in the program of FMPT, and the preliminary studies are being made. In this paper, the effect of surface tension on flow behavior, and examples of space experiments (SL-1, D-1) concerning thermocapillary flow are reviewed. Moreover, the result of the aircraft experiment under low-gravity environment for FMPT, preliminary study of Marangoni convection in FMPT, and theoretical study of a control of Marangoni convection are presented.
A series of computer simulation was carried out in order to understand the different kinds of behavior between the transverse MCZ and the axial MCZ on the oxygen contents in Si crystals; oxygen contents in silicon crystals are usually very low in the transverse MCZ growth, but frequently quite high in the axial MCZ growth. It has been shown that the oxygen contents in silicon crystals are mostly affected by melt convection. High oxygen contents in silicon crystals are obtained both by weak melt convection and strong melt convection, and low oxygen contents in silicon crystals are obtained by intermediate melt convection. High oxygen contents in the axial MCZ growth are due to excessive suppression of melt convection. The reasons of the dependence of oxygen contents on melt convection are discussed.
Experimental and analytical studies on the oxygen-incorporation mechanism in Czochralski (CZ) silicon crystal growth are summarized, with special emphasis on the influence of convections on transportation and mixing in the melt. Included dre experiments on oxygen concentration distribution in silicon melt during crystal growth, silica dissolution and SiO evaporation rates, and a theoretical analysis of the conservation of oxygen atoms in the melt. Oxygen concentration distribution is uniform through the bulk of the melt. This is mainly due to thermal convective mixing. A steep concentration gradient exists at the melt-silica crucible interface. The rate of oxygen transfer through this interface by the crucible dissolution is determined by diffusion in a melt layer with thickness δ_x, which is mainly controlled by forced convection due to crucible rotation and thermal convection adjacent to this interface.
Experimental and numerical approach to convection of molten silicon was surveyed. Natural and Forced convection of molten silicon during Czochralski single crystal growth was directly observed using X-ray radiography. Flow pattem of the melt was monitored using a solid tracer method. The tracer, whose density and wettability were adjusted to that of molten silicon, was newly developed. Observed flow of natural convection of molten silicon in a crucible was not only steady but also transient, and not axisymmetric. This asymmetry is attributed to asymmetric temperature distribution within the crucible. The flow velocity of molten silicon in a 75 mm-diameter crucible was 10 to 20 mm/sec. The experimental results were compared with previous numerical calculations.
Growth of crystals from solutions were investigated by means of in situ observation methods, so that the formation of defects or growth striations can be investigated in relation to the degree of supersaturations and the mode of convection or flows. It was confirmed that unstable convection plumes start to develop at a critical supersaturation, 〜0.5%, around a growing Barium Nitrate crystal in an aqueous solution, which results in the large growth rate dispersion, >40%. It was also confirmed that due to the higher supersaturation at the surface very adjacent, <0.1 mm to the edge of a crystal, growth rate of a spiral hillock is accelerated nearly twice by generating the steps with double step density, even in a running solution. The acceleration was found to cease wheh the spiral centre relatively moves, being away, >0.1 mm, from the edge of the crystal as the crystal grows larger.
Morphological studies on CdS crystals produced using "Conically converging shock-wave technique" have been carried out by scanning and transmission electron microscopes. Reflecting the special growth condition, many growth features have been observed. They were classified into polyhedral crystal based on hexagonal prism, sphere, matchstick-like, hexagonal hollow, ribbon-like, needle-like, elephant-nose-like, flag-like, arrowhead-bulb-like crystals. They were the wurtzite structure. The crystallographic orientation relationship of above characteristic shapes have been elucidated.
Phase equilibria studies of the binary system Al_2O_3-SiO_2 have a very extended history for more than 80 years, yet, the melting nature of mullite, the only one compound in this system, is still ambiguous and controversial. In this paper, the published investigations on this problem are critically reviewed from the standpoint of crystal growth, and the conflicting problems involved in ordinal phase research methods are pointed out. It is suggested that the control of the solidification from this silicate liquid is the most important factor to avoid the misinterpretation of the results. Using the slow cooling float zone method, the phase diagram of this system is revised.