The progress in fcc-based phase diagram investigation has been made based on the Cluster Variation Method. We first performed the prototype calculation in which pair interaction energies are given as constant parameters, and the importance of atomic correlations played in the free energy formula has been pointed out. Although the resulting phase diagrams demonstrated overwhelming advantages over the ones obtained by the Bragg-Williams approximation, serious drawback is the completely symmetrical feature of the phase boundary around 50at%. We, then, proceeded on the phenomenological model of which parameters are fitted to the thermodynamic data. By the introduction of concentration dependency to the pair interaction energies, the asymmetry of Cu-Au phase diagram has been actualized. Finally, together with the electronic structure calculation via ASW method and statistical numerical calculation via Cluster Variation Method, the first-principles calculation was attempted to draw the phase diagrams of three kinds of noble-metal alloys; Cu-Au, Cu-Ag and Ag-Au. The physical origin of the distinctly different alloying behavior of these systems are partly elucidated by microscopic analysis based on the electronic structure calculations. The discrepancies between the calculated phase diagrams and experimental ones are pronounced for the system which has a large difference in atomic size. The importance of the local lattice relaxation is pointed out.

View full abstract

General forms of Fourier expressions for the one- and two-dimensional non-linear dif fusion equations containing the gradient energy term are proposed and computer-simulations in some supersaturated solid solutions are derived on the basis of the Fourier expressions. The phase-decompositions proceed to form periodic zone-arrangements in the high solute alloys, while in the low solute alloys a few composition peaks sparsely distributed appear. In the calculation for the phase decomposition of an elastically anisotropic solid solution, the <100> modulated structure is formed, which explains well the morphology in actual spinodally decomposed alloys having an elastic anisotropy factor greater than unity. The computer simulations obtained based on the non-linear diffusion equations are very useful and powerful for the basic understanding of phase decomposition.

View full abstract

In order to design superplastic titanium alloys having an improved strength-density ratio, it is necessary to control the compositional and microstructural parameters such as volume fraction and degree of solid-solution strengthening for the primary a phase and electron-atom ratio for the prior β phase in titanium alloys. For that purpose, a method for calculation of α-β phase equilibria in multi-component titanium alloys, based on the analyses of a and β phase compositions, was developed by using multicomponent 20 titanium alloys. The reliability of the method was confirmed to be satisfactory by excellent agreements between the observed value and the value calculated by the method for the volume fraction of a phase and the partition ratio of each element between a and β phases in not only 20 alloys which were used to construct the method but also other 8 alloys. Furthermore, thermodynamic calculation was performed on multi-component 20 titanium alloys. The good agreements between the calculated value and the observed value for the volume fraction of α phase and the partition ratio of each element between α and β phases demonstrate that an adequately selected set of parameters can describe the thermodynamics of the α and β phases in multi-component titanium alloys.

View full abstract

Structure change during reversion of phase-decomposed alloys has been investigated by using the kinetic Ising model (Kawasaki dynamics) on fcc lattice, and the results have been compared with the experimental results obtained by means of synchrotron radiation small-angle scattering (SR-SAXS) study of Al-Zn binary alloys. From the real-space configuration of solute atoms obtained by the simulation, the structure of the system was investigated. The structure function was also calculated by Fourier transform, in order that direct comparison with SAXS experiments was possible. It was made clear that the time evolution of the structure parameters during phase decomposition agreed qualitatively with those obtained by Lebowitz et al, for a simple cubic lattice. The structure change during reversion was found to be divided into two distinct stages: The earlier one is characterized by the rapid reduction of the composition of precipitates, with almost constant volume fraction. The later one is characterized as the stage where precipitates shrink to dissolve with almost constant composition. For reversion below the coexistence curve, the coarsening process at the reversion temperature followed the reversion process. It was found that the structure change during reversion in Al-Zn binary alloys was very similar to the one observed in the present simulation, and the further discussion based on the two phase analysis with the small-angle X-ray scattering results agreed well with the present results.

View full abstract

Based on the statistical theory of texture controlled grain growth by the present authors comprehensive computer simulations were carried out. For a system consisting of two textural components the influence of the most important parameters, such as initial volume fractions of the components, initial mean radii and grain growth diffusivities were investigated. It was found that during grain growth no quasistationary texture is achieved. Instead complete transitions of the texture occur, leading to an intermediate broadening of the grain size distribution, which for certain values of the involved parameters appears as pronounced secondary recrystallization. The grain growth kinetics does not follow the R-t^1/2 law, but in a double logarithmic plot (log R vs. log t) shows a stepwise curve with alternating flat and steep ranges.

View full abstract

Experimental data in the literature on the partition of alloying element during the growth of Proeutectoid ferrite allotriomorphs in 9 Fe-C-X systems (X=Al, Si, V, Cr, Mn, Co, Ni, Cu and Mo) are compared with the theory of difusional growth which assumes that all component species are in local equilibrium (LE) at the advancing interface and also with the theory which assumes that only local equilibrium with respect to carbon is maintained at the interface (paraequilibrium (PE)). The temperature ranges in which the formation of ferrite occurs with or without alloy partition are explained in a qualitative manner by the former theory. The partition behavior in each alloy is discussed in terms of thermodynamic properties of alloy element. A major deviation of partition from the LE theory occurs at lower reaction temperatures in alloys with higher Mn or Ni content. A practical means of estimating the temperature of the cassation of partition in such alloys is considered. The growth kinetics of alloy element partitioned and no partitioned ferrite allotriomorphs are briefly reviewed. In alloys with X which has a strong affinity with carbon in austenite, such as Cr, Mn, Mo, etc., the growth kinetics differ remarkably from both the LE and the PE theories, the reasons for which are yet to be explained adequately.

View full abstract

The structure and energy of the ∑=5 (130) symmetrical tilt grain boundary in Si were calculated theoretically. The valence force fields, the bond orbital model, and the semiempirical tight-binding method with the use of the supercell technique were used and the computed results were compared with each other. All the computations have shown that the structural model by Bacmann et al. is more energetically favored than that by Hornstra and that the optimum rigid-body translation of the former model is characterized by a translation of about 1/8a₀ along the <001> direction. These results are in good agreement with the experimental observation of a germanium bicrystal. However, there are diferences in the absolute values of calculated boundary energies. The energy values obtained by the valence force potentials with the parameters fitted to the elastic constants and by the bond orbital model are overestimated, although the energy values by the valence force potential with the parameters fitted to the finite wave vector modes of lattice distortions are comparable with those by the semiempirical tight-binding method.

View full abstract

Using the computer simulation technique based on the LCAO (linear combination of atomic orbitals) electronic theory, we study the grain boundary properties of the metal systems (Ni₃Al crystal with L1₂ structure) as well as of covalent semiconductor (Si) crystals. Specifically, we calculate the atomic configuration, segregation energy of solute atoms and cleavage strength of the symmetrical tilt grain boundaries. We assume that the total energy of the system is given by a sum of the band structure energy and the pairwise repulsive potential energy. For the calculation of band structure energy, we use the recursion method of s, p and d-basis orbitals and local charge neutrality condition. For the repulsive potential energies, specific spatial dependence is derived by either Poudolocky-Pettifor formulation (for the intermetallic compound) or the extended Hiückel theory (for the covalent semiconductor), depending on the nature of atomic bonding of the material. We will show that the present analysis of the LCAO scheme is quite useful and suggestive for the interpretation of various experimental results on grain boundary properties of metal systems and of covalent semiconductor crystals.

View full abstract

A molecular dynamics study has been carried out of Li₄SiO₄ melt and glass. The computer-generated structures of Li₄SiO₄ melt and glass seemed to be realistic in comparison with our Raman and X-ray diffraction data. The simulated correlation functions were in satisfactory agreement with the one obtained from the X-ray diffraction analysis. Furthermore, the results obtained indicated that the Li₄SiO₄ glass consisted of a few kind of isolated silicate units.

View full abstract

Two different numerical methods were used to simulate the solidification of eutectic alloys. The first one was the classical approach based on an implicit, one dimensional, axisymmetric finite difference method (EUCAST program), while the second one involved a two dimensional control volume implicit scheme with a boundary condition describing the heat flux at the casting-mold interface (BAMACAST program). To account for the heat evolution term in the conductivity equation, an original latent heat approach was introduced, consisting of the calculation of the latent heat evolved by the fraction of solid formed as a function of time. In turn, the fraction of solid has been calculated based on nucleation and growth kinetics. The heat transfer coefficient used for simulation was determined based on an analytical approach for hollow cylinders.
Gray cast iron and aluminum-silicon alloys of eutectic composition were used for the validation experiments. Cylindrical bars of different diameters were poured and the cooling curves were recorded at various locations in the metal and the mold. Experimental and computed data were compared, with excellent results for solidification time, undercooling, width of the mushy zone and heat transfer coefficient.

View full abstract

Molecular dynamics is used to study the homogeneous nucleation and growth processes for the sodium system of 864 atoms with constant volume. The equilibrium liquid at 381 K has been cooled directly to 181K. The processes have been monitored by studying mean-square displacement of atoms, potential energy, and pair-distribution functions. The supercooled liquid is found to crystallize into bcc structure. The number of atoms in the critical nucleus is at least greater than 43 atoms which is almost coincident with prediction from the classical nucleation theory. We have obtained the solid-liquid interfacial energies from the local pressures of each atom. The energies increase with increasing cluster size. The nucleus is found to advance at a rate of 187m s^-1, which is extremely larger than the rate from the kinetic theory for growth.

View full abstract

Concerning a heat source model for estimating thermal cycles near the fusion line in an arc welding from the heat conduction theory, the details of the frame model formed by line heat segments which correspond to heat transfers due to the plasma stream and molten metal flow have been proposed and can be determined using a personal computer system designed for the following steps:
1) Rough selection of line segments on the basis of an expert knowledge of the weld pool and simplification for practical use,
2) Evaluation of an effect of each segment on the thermal cycle by the multiple regression analysis, and
3) Checking adaptability of the segments from the distribution of the residuals and the time at a maximum temperature-elevation due toeach segment.
As a result, the model for a given welding condition can be obtained on the basis of the data of the weld penetration measured by many engineers for a long time.
Moreover, the system to predict hardness and microstructural constituents in the heat affected zone and to determine optimum welding conditions has been developed by using an estimated thermal cycle and the database of CCT diagram for welding.

View full abstract

The interaction between a screw dislocation and self-interstitial atoms (SIA's) in iron and vanadium lattices under applied shear stress has been investigated by computer simulation using Johnson-Wilson potentials for iron and vanadium. Two structures of SIA, i.e., a <110> dumbbell and a <111> crowdion have been taken into account. Although the former is stable in the matrix, the <111> crowdion is more stable than the <110> dumbbell in a screw dislocation core. On the encounter of a gliding screw dislocation with a SIA (dumbbell type), conversion from the dumbbell to the crowdion occurred in a screw dislocation core, resulting in the formation of three kink configuration. These kinks propagated along a dislocation line under shear stress. This greatly contributes to the understanding of irradiation softening observed in iron single crystals after low temperature irradiation. The interaction between a screw dislocation and di-interstitial atoms was simulated in the case of vanadium and also contributed to the irradiation softening.

View full abstract