An ab initio calculation of electronic excitation energies based on the GW approximation is performed by using the all-electron mixed-basis approach. The generalized plasmon-pole (GPP) model is used to determine the electron self-energyand the results are compared with our previous results using the full frequency integration (S. Ishii, et al., Phys. Rev. B 63 155104 (2001)). It is found that the two methods with and without the GPP model agrees well in the case of sodium clusters.

View full abstract

A norm-conserving pseudopotential database (NCPS95) and updated version NCPS97 have been available since 1995. NCPS97 has been updated to NCPS2K and will be distributed by CD-R media. In the updated version (NCPS2K), new numerical pseudopotential data for the rare gas atoms (He, Ne, Ar, Kr, Xe, Rn) and Ba and At have been prepared. In addition, some pseudopotentials in NCPS97 have been updated and retained. In this database, lattice properties and the absence of ghost bands in the electronic band structures are checked.

View full abstract

We Study the low-temperature thermodynamical and statistical properties of the ASYNNNI model, used to describe oxygen ordering in YBa₂Cu₃O_6+2c. Analyzing the excited states we show that this model undergoes a second order phase transition at absolute zero temperature. From the divergence of the relative energy fluctuations, in the limit T→0, we deduce that for T≈0 the deviation of the chemical potential μ from its ground state value is linearly related to the partial derivative of the fraction of threefold coordinated Cu(1) ions n with respect to oxygen concentlation c. We use the cluster variation method (CVM) to determine the numerical value of the coefficient of this relation and find that it is equal to −1⁄2, independently of the values of the interaction parameters. This establishes a full equivalence between the ASYNNNI model and the one-dimensional Ising model, where the role of the nearest-neighbor interaction of the latter is played by the V₂ interaction in the former. We comment on the general applicability of low levels of the CVM approximation to systems equivalent to the one-dimensional Ising model.

View full abstract

A numerical method based on the Method of Fundamental Solutions coupling with the Fourier transformation is presented. The advantage of the new method is demonstrated by solving the nonlinear partial differential equations, which are derived for the simulation of the pulse discharged sintering process to obtain the electric potential and temperature profiles in the system. In this problem, the delta function is revealed at the interface between the different component materials, graphite mold/punch and powder sample. The properties change there as a step function. The efficiency of the method is indicated with regard to solving the problems including the discontinuous functions as a whole without dividing the system to two different material parts. The calculation results showed that the temperature in the system is controlled by the heat conductance from the punch, where a large Joule heat evolution is observed without depending on the sample properties. The calculated temperature in the sample is relatively uniform.

View full abstract

Complementary to the experimental studies on the magic Ar₆Fe⁺ cluster [Bililign et al., J. Chem. Phys. 108 (1998) 6312], the geometry, magnetic properties, dynamical stability, and magnetic stability are studied in detail by using ab initio molecular orbital calculations. The geometry expected by experimentalists is confirmed. When Ar₆ cluster is doped with Fe⁺, due to the charge transfer, the electronic shells in the Ar atoms are changed from closed to open, resulting in stronger interactions and producing a more compact structure. Because of orbital hybridizations, the Fe atom carries magnetic moment of only 3μ_B.

View full abstract

Ab-initio calculations are performed to study interaction of hydrogen on aluminum clusters. The results show large binding energies for a single hydrogen atom on Al₇ and Al₁₃ and two hydrogen atoms on Al₆. The highest occupied-lowest unoccupied molecular orbital (HOMO-LUMO) gaps of Al₇ and Al₁₃ become dramatically large by absorbing hydrogen. These large binding energies and HOMO-LUMO gaps make Al₇H and Al₁₃H clusters magic(s). The atomic structures and binding energies of Al₆H₂ and Al₈H₂ are used to explain the reason why hydrogen is likely to be dissociated only on Al₆. These results are discussed in the light of available experimental data and are in very good agreement with them.

View full abstract

Density functional theory has been used to investigate the adsorption of small molecules on Fe-based multi-component clusters. The energetic, electronic and structural properties of CO and H₂ on the catalyst surfaces are calculated. It has been found that Cr, Mn, Mo, Zr and Re could be used as additional elements in the Fe-based catalysts, since the activation of CO and H₂ molecules is increased as compared with the unsupported Fe catalyst. The obtained results are in agreement with available experimental data that confirmed the validity of selected calculation method and models.

View full abstract

Clustering behaviour of the constituent alkali metals, in bimetallic clusters has been investigated for M₈K₆ clusters (M=Li, K, Rb, Cs, Mg and Al) using a recently developed orbital free molecular dynamics method. For Li₈K₆ and Na₈K₆, the lithium or sodium atoms form a distorted inner cube whose six faces are capped by K atoms. In contrast, for Rb₈K₆ and Cs₈K₆, the K atoms form an inner octahedron whose faces are capped by Rb or Cs atoms. The homoatomic K₁₄ and Al₁₄ clusters form an icosahedral configuration with one of its face being capped by the fourteenth atom. For Mg₈K₆ and Al₈K₆ clusters, the Mg or Al atoms form a distorted cubic inner core of the cluster. From these results, it is inferred that for such bimetallic clusters the element with smaller atomic radius, higher electronegativity, and larger number of valence electrons, forms the core of the cluster.

View full abstract

A large-scale ab initio molecular dynamics simulation of the insertion of nitrogen atom into C₆₀ is carried out by using the all-electron mixed-basis approach. In this approach, a one-electron wave function is expressed by superposing truncated numerical atomic orbitals and plane waves. The LDA (local density approximation) is used to calculate electronic states. 4169 plane waves having a cut-off energy of 7Ry (1Ry=13.6 eV), and 1s, 2s and 2p atomic orbitals for carbon and nitrogen atoms are used in this calculation. Consequently, we found that with proper incident kinetic energy around 80 eV (1 eV=1.602×10⁻¹⁹ J) the nitrogen-encapsulated C₆₀, N@C₆₀ is finally realized by a collision of a nitrogen atom against C₆₀.

View full abstract

The behavior of clathrate hydrate of structure I with and without enclathrated methane and xenon guests has been investigated at high pressure and wide temperature range using the quasiharmonic lattice dynamics approach up to absolute stability limits. The mechanical stability boundaries of the clathrate hydrate of structure I have been determined from the Born stability criteria via the calculations of the elastic constants at positive and negative pressures. For comparison the spinodals of the clathrate hydrates are calculated. It is found that in spite of weakness of interaction the enclathrated guests with host lattice the guest molecules contribute greatly to the stability of the hydrate framework structure.

View full abstract

Bondlengths of Si–Si, Si–Ge, and Ge–Ge pairs in Silicon-Germanium alloys were determined as a function of hydrostatic pressure using ab initio electronic structure calculations. A series of ordered structures was selected to represent the various atomic environments in actual alloys. Enthalpies of formation computed for these structures were used to model the phase stability under hydrostatic pressure.

View full abstract

We present systematic first-principles calculations for vacancy formation energies E_V^F in most of elemental metals (Li∼Au) of fcc and bcc structures, as well as bulk properties such as lattice parameters a and bulk moduli B. The calculations are based on the generalized-gradient approximation in density-functional formalism, proposed by Perdew and Wang in 1991 (PW91-GGA), and apply the full-potential Korringa-Kohn-Rostoker Green’s function method for perfect crystals and point defect systems, developed by the Jülich group. First we show that the calculated bulk properties for all elements studied are in excellent agreement with the experimental results: the PW91-GGA corrects the deficiencies of the local spin density approximation for metals, i.e., the underestimation of a and the overestimation of B and the theoretical errors in a and B are reduced within ∼ 1% and ∼ 10% of experimental results, respectively. Second we show that E_V^F for most of fcc metals are reproduced within the experimental errors, while E_V^F for most of bcc metals are overestimated by 10%–20% of the experimental results. It is noted that the comparison with the experimental results needs the inclusion of the thermal lattice expansion effect in the first-principles calculations because most of the experimental results were derived from positron annihilation measurements at high temperatures. The remaining discrepancies between theory and experiment are also discussed.

View full abstract

We present systematic first-principles calculations for impurity-impurity interaction energies (E_int) of 4d elements in 4d bcc metals Nb and Mo, and 4d fcc metals Pd and Ag. The calculations are based on the generalized-gradient approximation in density-functional formalism, proposed by Perdew and Wang in 1991 (PW91-GGA), and apply the full-potential Korringa-Kohn-Rostoker (FPKKR) Green’s function method for point defects, developed by the Jülich group. First we examine the distance dependence, from 1st to 8th neighbors, of E_int and show that for most cases, the 1st nearest-neighboring impurity-impurity interaction energies (E_int¹) are dominant. Second it is shown that most of the types of phase diagrams of binary alloys of impurity and host elements, such as segregation, solid solution, and order, known experimentally, may be very well discriminated by use of the sign and magnitude of E_int¹. Third we show that the temperature dependence of solid solubility limit of Rh in Pd, which is segregated at low temperatures and becomes disordered at high temperatures, are reproduced fairly well by the free-energy calculations based on the cluster variation method with the present results for E_int (up to the 8th neighbor). It is also shown that the inclusion of the impurity-cluster interaction energies up to the four-body (a tetrahedron of 1st-nearest neighbors), being also determined by the present first principles calculations, leads to the complete agreement with the experimental result. We also show that the chemical trend for E_int¹ is understood by use of the Friedel’s band filling mechanism for d-states, if the dependence of the band width on element is taken into account in the model.

View full abstract

In statistical thermodynamic analysis for condensed phase (either liquid or solid), atomistic (or microscopic) information of the condensed phase is derived from experimentally determined macroscopic equilibrium P-T-C (pressure-temperature-composition) relationships. This analysis procedure is especially adequate for non-stoichiometric compound containing at least one interstitial constituent. Atomistic information derived by statistical thermodynamics includes clustering pattern of metal atoms around an interstitial atom as well as pair-wise interaction among constituent atoms. Desired calculation in the statistical thermodynamic analysis is not very demanding and thence, compared with some other branches of computational materials science and engineering research, statistical thermodynamics requires only modest computational capacity of currently available advanced high-performance PC (personal computer). In this article, I would like to review aspects of atom clustering evaluation by statistical thermodynamics and also potentiality of statistical thermodynamics for predicting optimised material manufacturing condition.

View full abstract

The bonding state has been studied by the discrete variational (DV)-Xα method in the vicinity of the core of the Peierls-Nabarro edge dislocation in the BCC(Fe) crystal the slip systems of which are {112}⟨111⟩ and {110}⟨111⟩. Bond order and net charge distribution were evaluated in the perfect crystal model cluster and edge dislocation core model cluster with 32 atoms. In the {112}⟨111⟩ slip system, the total bond order of the nearest neighbor atom pairs in the dislocation model was found to be relatively larger than that in the perfect crystal model. On the other hand, the total bond order of the second neighbor atom pairs in the dislocation model was a little smaller than that in the perfect crystal model. It was also found that there was asymmetry in the bond order between the (110) plane on the left side and the one on the right side of the (110) plane which contains the core atom of dislocation. Net charge and its effect in the dislocation model were larger and wider than those in the perfect crystal model. Almost the same results were also attained for the {110}⟨111⟩ slip system.

View full abstract

The electronic structure of Al grain boundary with segregated Si and S impurity atoms, respectively, is calculated by first principles pseudopotential method. It is found that the segregated Si atom bonds the neighboring Al atoms, forming the metallic-covalent character mixing bonds. These bonds will prevent the rearrangement of atoms, such as sliding under stress. Therefore the embrittlement promoted by Si segregation should be classified into ‘bond mobility model’. However, the segregated S atom bonds only one of the neighboring Al atoms, forming a metallic-covalent character mixing bond. Other Al–S bonds in the boundary may become weaker than the former Al–Al bonds. Therefore it can’t be decided that the mechanism of the embrittlement promoted by S segregation is classified into ‘bond mobility model’ or ‘decohesion model’.

View full abstract

We have performed molecular dynamics (MD) simulations for vitreous P₂O₅ using isotropic pair potentials composed only of coulombic and repulsive interactions. The P–O pair distribution function obtained had the two distinguishable peaks expected from the results of neutron diffraction experiments, in the nearest-neighbor P–O correlation. The neutron-weighted real-space correlation function was also in semi-quantitative agreement with that obtained from the experimental results. The distribution of the coordination number for O around P and P around O showed that most P atoms form tetrahedral PO₄ units in the glass, and that three-fifths of O atoms are bridging oxygen atoms, O_B, and the remaining are terminal oxygen atoms, O_T. The pair distribution functions for P–O_B and P–O_T show that the PO₄ units have three long P–O_B bonds and one short P–O_T bond. We have concluded that the short-range structure obtained for vitreous P₂O₅ agrees well with the picture derived from many experiments. This fact indicates that the short-range structure of vitreous P₂O₅ can be described mainly by both charge ordering and packing based on the differences in ionic charge and size between cation and anion.

View full abstract

We have investigated in detail the dependence of scanning tunneling microscopy (STM) images of the Si(001)-p(2×2) surface on bias and tip-sample distance, based on first-principles molecular-dynamics simulations. STM images on the terrace of the Si(001)-p(2×2) surface are found to be similar to those at the S_A-step edges of the Si(001) surface. The present theoretical calculations predict that the STM images strongly depend on the sample bias and tip-sample separations, and that the π and π^∗ surface states contribute to the changes of the STM corrugation images.

View full abstract

First-principles calculations of electron-transport properties for a single-row atomic wire of Na under the application of a finite bias voltage are presented. Calculations are carried out by a density functional Green-function approach based on the Lippmann-Schwinger equation and the Landauer-Büttiker formula to evaluate the conductance. The model consists of a linear wire, a pair of jellium electrodes and two pyramidal clusters as the interface between the linear wire and the jellium electrode. As a result of the calculations, we found that the voltage drop is generated neither in the pyramidal clusters nor in the jellium electrodes, but in the linear wire. The conductance of Na atomic wire evaluated from the electron transmission is about 84% of the quantized unit 2e²⁄h, and this low conductance is caused by partial reductions of the transmission in some parts of the incident energies. The main reason for this reduction is that the spatial distributions of some states responsible for electron transport become discontinuous around the linear wire in the case of applying a finite bias voltage.

View full abstract

We investigate the relation between the geometrical structure and electrical conduction of an infinite single-row gold wire in the process of its elongation using first-principles molecular-dynamics simulations based on the real-space finite-difference method. This relation has not been explicitly explained by experiments yet. Our theoretical study predicts that the single-row gold wire ruptures when the average interatomic distance increases to more than 0.30 nm, and that the wire is conductive before breaking but changes to insulator at the rupturing point.

View full abstract

Tight-binding molecular dynamics (TBMD) simulations are performed to investigate atomic and electronic structures during neck formation processes of nanocrystalline silicon carbide at high temperature. For calculating the electronic energy and forces we use a linear-scaling method (the Fermi-operator expansion method) with a scalable parallel algorithm. The TBMD simulations of collision of SiC nanospheres show that processes of neck formation depend strongly on contact angles between the two grains. Electronic populations at grain boundaries are rather uniform, even in a disordered structure of the grain boundary between misaligned nanospheres. Atomic diffusions at elevated temperature are, on the other hand, quite different in the necks formed with different orientations of particles.

View full abstract

The grain growth phenomenon in a polycrystalline metal was studied by the molecular dynamics method. The time evolutions of the grain boundary network in the nano-grained polycrystals composed of 64 grains were analyzed by incorporating the boundary curvature and the grain boundary energy composed by the interatomic potentials. The grain growth phenomenon accompanied by the T1 and T2 processes was successfully observed. The rate of the grain growth was fairly close to the parabolic law. The grain boundaries were dominantly composed by random boundaries, but a number of coincidence boundaries were also observed.

View full abstract

Green function and first principle theories are used to investigate electron transport through the polyphenylene-based molecular rectifying diode switch, which is attached to gold electrodes at both ends. The coupling at the interface between the molecular device and the electrodes is treated carefully. The Green function of the electrode is built by the standard parameterized tight-binding method. The effect of the molecular orbitals on the conductance is shown clearly in the conductance curve. The conductance calculated from the approximated Green function method, which is composed of spd/sp/s model, is presented. The difference between our Green function and the current approximated Green function is discussed. The result suggests that the Green function made from sp orbital model is a good approximation. Besides, the molecular structure of the diode-diode-type AND gate, composed of the polyphenylene-based molecular wires, is energetically optimized. Some of its orbitals, near the Fermi level, are presented. Our calculation method can be extended straightforwardly to any organic molecular system, which is connected by the electrodes.

View full abstract

Density functional study has been carried out using B3PW91/6-311G(d,p) method for bi-pyridinium molecule connected with poly-methylene chains (upto n=9), reported experimentally to behave like molecular switch. Fully optimized geometries, electronic structures, HOMO-LUMO gaps, Muliken point charge distribution and orbital orientation have been analyzed to understand the electronic behavior in this molecule. The electronic transport across the molecule has been explained assuming that the incoming electron passes through the lowest unoccupied molecular orbitals. The conduction barrier is determined from the energy levels of HOMO and LUMO energy states. From the results it is predicted that a bias voltage of 1.66 V is required to transfer one electron from metal electrode to the LUMO energy state, quite good in agreement with the experimentally reported results.

View full abstract

Self-assembled islands are spontaneously formed during the heteroepitaxial growth of InAs on GaAs substrate. The island formation creates strain fields in the substrate and this leads to elastic interaction between the islands. We calculate the strain energy of the array of InAs islands on the GaAs substrate by the finite element method, and extract the elastic interaction between the islands from the variation of the energy with the distance between the islands. The interaction decays with the third power of the inter-island distance, in agreement with our previous work. The effect of the interaction on the correlation of the island positions is examined by comparing the coefficient of the interaction with the phase transition criteria of the two-dimensional dipole system. When the island size is large enough, depending on the island density, an array of islands can be in either a liquid phase where the islands are randomly distributed or in a solid phase where the islands form a two dimensional lattice. Only the liquid phase exists when the island size is small.

View full abstract

A molecular dynamics simulation by the embedded atom method was conducted to investigate hydrogen embrittlement of a nickel single crystal, which is composed of 163311 nickel atoms on the nanometer scale and has a [011]-oriented notch under uniaxial tension along the [100] direction at room temperature. The hydrogen-free specimen showed good ductility associated with pronounced blunting of the crack tip. Hydrogen influence was most serious in the specimen that had been hydrogen-charged in the notched (100) planes ahead of the crack tip. In the specimen that had been hydrogen-charged in the notched area, a hydrogen-assisted fracture occurred macroscopically on the (100) plane perpendicular to the tensile direction and the elongation at failure decreased with increasing hydrogen content. A low hydrogen content caused strain localization only, while a high hydrogen content caused microvoid formation in the notched area as well. The specimen containing a thin layer of hydride fractured and exhibited brittleness due to significant microvoid formation and subsequent microvoid growth and linkage at the early stage of deformation. The simulation results show good agreement with the published experimental observations.

View full abstract

Machining mechanism of EEM (Elastic Emission Machining) is investigated at the atomic level by employing first-principles molecular-dynamics simulations. Based on experimental results it has been suggested that the removal of surface atom occurs through a surface chemical reaction between the work surface and the powder surface. This reaction has been so explained that powder particle chemisorbs on the work surface, and is succeedingly separated, removing a work-surface atom. In this paper, we investigate the machining mechanism by the analyses of the reaction process and the binding energy between work and powder surface based on the molecular-dynamics method. First, dependency of machinability on the powder materials is evaluated. It is recognized that the bond between the machined surface atom and the work surface is weakened, and the dependency on powder material is consistent with experimental results. Second, the molecular-dynamics analysis shows that the removal actually occurs in the unit of atom without any potential barrier and without introducing any defect around the machined atom.

View full abstract

A model of cluster growth process based on Direct Simulation Monte Carlo (DSMC) and first principle calculation is introduced to explore the effects of experimental conditions on the clustering process and size distribution of the clusters gathered at the nozzle. The behavior of material and inert gas atoms, which fly through the cluster formation equipment, is simulated under different experimental conditions including flight path length, concentration of the material gas to the inert gas, total gas pressure and the initial temperature of the inert gas. Size distributions of generated clusters under various conditions are obtained. The results of the present simulation suggest that the size distribution strongly depends on the experimental conditions.

View full abstract

The homogeneous nucleation process induced by supercooling liquid Pt has been studied by means of molecular dynamics employing the embedded-atom method for the potential energy. The process was simulated by cooling an equilibrated liquid to a low temperature, while structure analysis was performed during the subsequent time evolution of the system under constant temperature and pressure conditions. In order to investigate the effects of degree of supercooling and cooling rate on crystallization, cooling temperature was varied from 900 to 1300 K and three processes at different cooling rates were studied. The results proved that certain degree of supercooling is necessary for homogeneous nucleation of crystals due to the existence of a free energy barrier in forming critical nuclei from supercooled liquid, as described by the classical theory of homogeneous nucleation. The scale of the degree of supercooling has much effect on the incubation period of homogeneous nucleation and nucleation rate. The crystallization towards fcc and hcp phases takes place in all homogeneous nucleation processes from liquid in our simulations. The progression of crystallization is sensitive to cooling rate. A very high cooling rate has been found to prolong incubation period, decrease nucleation rate, thus suppress crystallization. This may be associated with the phenomenon that high cooling rate prohibits the rearrangement of atoms towards forming a crystal lattice at high temperatures during the early stage of cooling.

View full abstract

The modified Monte Carlo simulation of grain growth has been performed in the model applying actual orientation data by SEM/EBSP analysis in a pure aluminum to an initial microstructure. Changes in microstructural features and textures have been compared between experimental and simulated results. The SEM/EBSP analysis reveals that the pure aluminum has a residual rolling texture after annealing. The texture becomes sharper as the annealing time increases. The simulations are carried out under various conditions of grain-boundary energy and mobility. Microstructural and textural developments in the simulation taking the dependence of grain-boundary energy on misorientation into consideration were in good agreement with those in the experiment. The present model is expected to make it possible to predict the microstructural evolutions precisely.

View full abstract

Microstructure modeling of aluminum alloy casting has been carried out by using a modified cellular automaton (CA) method coupled with macro heat transfer calculation. Continuous nucleation model is applied to describe heterogeneous nucleation in liquid metal. Dendrite tip growth kinetics and preferential ⟨100⟩ crystallographic orientation are taken into account. Therefore, the stochastic nature of nucleation process as well as the deterministic of dendrite growth is considered to simulate the crystal growth. The actual dendritic shape is substituted for the square envelope in the ordinary cellular automaton model. In addition, the coordinate transformation method is employed to describe the grain growth and to maintain it onto the original orientation of nucleus. Simulation results show that grain size is different at various positions of a step-shaped sample casting. The grain size is smaller when cooling rate is faster and vice versa, which is in good agreement with the experimental results and the solidification mechanism as well.

View full abstract

The mixed buoyancy-thermocapillary flow in a half-zone liquid bridge with the dynamic Bond number Bd≈o(1) is investigated numerically by the finite volume method. The similarities of temperature and velocity structures are realized under conditions of fixed Marangoni and Rayleigh numbers in steady two-dimensional regime for a suitable Prandtl number range.

View full abstract

The present study evaluates the superplastic material characteristics of Ti-alloy “Ti–4.5Al–3V–2Fe–2Mo”, SP-700, and Al-alloy “A5083”, which have been obtained by utilizing a multi-dome forming test. First, the flow stress vs. strain rate relationship was established for certain strain rate range, from a single circular sheet containing four different domes, sized 35, 30, 25, and 20 mm diameter, respectively. The multi-dome forming experiments were conducted under constant pressure at a superplastic temperature of 800°C for Ti-alloy and of 530°C for Al-alloy. The bulge forming of domes was also simulated by finite element method (FEM) to verify the experimentally-obtained material characteristics. Then the material characteristics obtained through multi-dome forming test were compared to those obtained by tensile step strain rate tests. In contrast to the predictions of the tensile test, the multi-dome forming test’s predicted time of deformation was generally in reasonable agreement with the results of the FEM simulation. There was no great difference, however, between the multi-dome forming test and the tensile testing with regard to the predicted final thickness distribution.

View full abstract

In order to clarify the effects of casting microstructures on the fatigue properties of Al–9%Si–1.6%Cu–0.55%Mg alloy castings, non-degassed, degassed and HIP treated castings with different sizes of microstructures were examined. The fracture elongation and fatigue properties markedly increased with decreasing the amount of gas in the melt. Gas porosities in the castings greatly affect the crack growth as well as the crack initiation. In the castings with no porosities, Fe intermetallic compounds with a needle shape act as a crack initiation site. In the castings composed of fine microstructures with smaller aspect ratios, the existence of the compounds scatters the values of fatigue strength. In the castings composed of coarsened microstructures with larger aspect ratios, on the other hand, the compounds provide not only crack initiation sites but also growth paths, resulting in the significantly decreased fatigue strength.

View full abstract

Pure magnesium and AZ31 magnesium alloy plates 4 mm in thickness were butt welded without addition of filler wire using a high voltage electron beam welding machine. Mechanical properties and microstructures of the welded joints were investigated. Regardless of the materials, the welded joints were almost free from welding defects and showed good bead appearance under appropriate welding conditions. The arcing phenomena tend to appear at low welding speed. Optimum beam current and welding speed were smaller than those for electron beam welded joints of aluminum alloys. Hardness in the fusion zone of the joints are nearly equal to those of the base metals. Microstructure on the fusion zone of pure magnesium joints was remarkably coarse, although the weld interface could not unambiguously detected. The fine crystal grains observed on the fusion zone of AZ31 alloy joints. Regardless of the welding conditions, both tensile strength and ductility of the joints show same value to those of the base metals, but the elongation of the joints are inferior to those of the base metals. From the tension test and impact test, pure magnesium joints fractured at the center of the fusion zone, but in case of AZ31 alloy joints, crack occurred at weld interface and it propagated through the fusion zone.

View full abstract

The structural characteristics of several modulations of Al–Co–Ni decagonal quasicrystals and crystalline approximants, all of which are found in and around an Al–Co–Ni decagonal phase with a wide compositional range, are discussed systematically. The structures of a Ni-rich basic structure and the other modulations are characterized as quasiperiodic arrangements of columnar clusters of atoms with decagonal sections of 3.2 and 2.0 nm in diameter, respectively. Various types of electron diffraction patterns including so-called superlattice reflections are interpreted by tiling features of the atomic clusters as well as ordered arrangements of the atomic clusters with pentagonal symmetry in the tilings. From the structure of a crystalline approximant (W-(AlCoNi) phase), precise models of the atomic clusters are proposed, and also the origin of diffuse reflections often observed in the Al–Co–Ni decagonal quasicrystals may be interpreted.

View full abstract

Mechanical alloying (MA) was performed, with Fe–N alloy powders as nitrogen source, on stainless steels of Fe–Cr–Ni–N (Cr–Ni type) and Fe–Cr–Mn–Mo–N (Cr–Mn type) systems in an Ar atmosphere using a planetary ball mill. In the MA processing, high nitrogen nanocrystalline stainless steel powders with 0.45–0.90 mass%N were readily manufactured after about 500–700 ks of processing. Nitrogen highly enhanced austenitizing of the powder products. The onset temperature (M_d) of strain-induced martensite formation in nitrogen-free 19Cr–11Ni stainless steel MA powders was estimated to be about 100 K or more higher than those obtained under conventional plastic deformation processes. Compaction of the MA powders at 1173 K using a spark-plasma sintering (SPS) process still retained their nanostructure. When 5 vol% of dispersion particle agents AlN or NbN were added to the MA powders, grain growth in the compacts was greatly suppressed during SPS processing. On SPS processing, nitrogen in the Cr–Mn type system was completely retained but about 10–15 mass% of nitrogen in the Cr–Ni type powder samples was observed to escape, depending on the composition. However, when Mn, Nb or Ta was added to the samples or Cr content was increased, the nitrogen loss was greatly decreased. Furthermore, by simultaneous addition of these elements, such nitrogen loss was almost completely prevented. The nitrogen retention depending upon the chemical composition during SPS processing and the marked raising of M_d temperature due to the MA processing are thermodynamically explained.

View full abstract

An equation expressing the transformation curve for crystallization of metallic glasses has been proposed by using two parameters, viscosity and melting temperature (T_m). The Vogel-Fulcher-Tammann (VFT) equation, η=η₀exp(B⁄(T−T₀)) where η₀, B and T₀ (ideal glass transition temperature) are empirical parameters, was used for expressing the viscosity. A Time-reduced Temperature-Transformation (T-T_r-T) diagram was calculated using the following five quantities: reduced temperature (T_r=T⁄T_m), three reduced viscosity parameters (η_0r=η₀⁄T_m, B_r=B⁄T_m and T_0r=T₀⁄T_m), and reduced critical cooling rate (R_cr=R_c⁄T_m) for formation of the glassy phase. The R_cr in the T-T_r-T diagram was calculated for Ni, metallic glasses and SiO₂. The glass-forming ability (GFA) was estimated from T_0r−R_cr(R_cr=R_c⁄T_m) diagrams corresponding to T_g⁄T_m−R_c diagrams obtained experimentally. The metallic glasses with T_0r of 0.55 are calculated to have R_cr ranging from 10⁻⁵ to 10⁰ s⁻¹, which agrees with the experimental data that metallic glasses with T_g⁄T_m of 0.6 or more have R_c of less than 10³ K/s. The following three points are clarified: (1) the higher GFA of metallic glass is achieved because of higher viscosity, (2) higher viscosity causes both the homogeneous nucleation frequency (I_v^hom) and the ratio (I_N⁄I_max), at reduced nose temperature against the maximum of I_v^hom, to decrease, and (3) the R_cr is numerically derived from reduced viscosity parameters.

View full abstract

Development of microstructure and crystal orientation distribution in melt-spun and subsequently annealed Ni–36 at%Al and Ni–38 at%Al alloy ribbons was investigated. In Ni–36 at%Al ribbon, the martensitic transformation from β to β^′ phase occurred during rapid cooling showing strong ⟨110⟩_β^′ fiber texture, while β phase with strong ⟨100⟩_β texture was frozen in Ni–38 at%Al ribbon. During annealing at 1073 K, a peculiar fiber texture of strong ⟨100⟩_β and ⟨110⟩_γ^′ texture was developed in Ni–36 at%Al ribbon. Formation of the texture was discussed focusing on the initial microstructure and the orientation relationship between phases during phase transformation.

View full abstract

6082 Al alloys are commercial and medium strength alloys, widely used as materials for welded structures. The purpose of this study is to investigate the effects of Mn addition on toughness of welded Al–Mg–Si alloys. To evaluate microstructural effects quantitatively, in-situ SEM observations of crack initiation and propagation behaviors through weldment are carried out. For the consideration of in-situ observation of fracture toughness test, stress field at crack-tip is analyzed using elasto-plastic finite element method (Hereinafter, FEM.) assuming that a crack is near a boundary between a weld metal and heat affected zone (Hereinafter, HAZ.). When small amount of Mn is added, recrystallization is completely suppressed as compared to specimens to which no Mn is added, thereby fibrous grains are kept. On the other hand, recrystallization of HAZ causes drastic decrease in fracture toughness in the case of no Mn addition. With the extension of a main crack, many microcracks are formed at grain boundaries ahead of a crack-tip despite the fact that the stress is relatively low. Such microcracking is not attributed to so-called liquation cracks, but the degradation is caused by the formation of film like Al–Mg intermetallic compounds at grain boundaries. The microcracks are aligned ahead of the crack-tip at an angle of 60 degrees from an initial notch direction. This is attributable to the experimentally-observed direction of the intermetallic compound film, which is also confirmed by the numerical analysis.

View full abstract

Calcium-deficient hydroxyapatite (DAp) having a rod-like or thin-blade-like shape was produced by hydrolysis of α-tricalcium phosphate (α-TCP) at different pH values. Mixed powder composed of 40 mass%DAp and 60 mass% stoichiometric hydroxyapatite (HAp) was suspended in an aqueous solution containing a small amount of dispersant by colloidal process. After dehydration, HAp-40 mass%DAp consolidated samples were sintered at 1473 K for 2 h. Rheological behaviour of the slurry, microstructure and bending strength of HAp-40 mass%DAp prepared by this process depended strongly on quantity of the dispersant. Good dispersion of HAp and DAp powders at an optimum quantityof the dispersant resulted in high relative density of HAp/DAp composites with homogeneous microstructure. DAp starting powder readily decomposed to α-TCP and HAp at 1473 K, while the formation of α-TCP phase took place less intensively in HAp-40 mass%DAp composites sintered at that temperature. In contrast to HAp/α-TCP composites, an interdiffusion of ions between HAp and DAp during sintering may result in the increase in the thermal stability of DAp. HAp-40 mass%DAp composites prepared by a suitable colloidal process exhibited higher bending strength than HAp-40 mass%α-TCP composites and stoichiometric HAp alone.

View full abstract

The phase transformations in a Ti–10 mass%Zr alloy during continuous cooling and isothermal holding have been investigated mainly by means of hot stage optical microscopy. During cooling from β phase at 100°C/s, α^′ plates formed martensitically at temperatures around 180°C together with sharp and coarse surface reliefs. The Widmanstätten α laths mostly nucleated directly on a β phase grain boundary and grew into one of the grains during isothermal holding at temperatures between 650 and 850°C. At β grain boundaries, α allotriomorphs often nucleated and Widmanstätten α laths grew from them during isothermal holding at 800°C. Widmanstätten α plates exhibited sharp surface reliefs. The precipitation of α phase exhibited the C-curve with the nose at around 775°C. The Ms temperature so far reported to be ∼ 805°C is not the martensite start temperature but the nose temperature for Widmanstätten α laths.

View full abstract

It was found that the Ti_45−xZr_xNi_45−zCu_z (0≤x≤10, 0≤z≤10) metallic glasses crystallize polymorphously and the crystal phase forms a solid solution of Zr and Cu in the TiNi phase. Differential scanning calorimetry (DSC) was used to study the crystallization kinetics of Ti–Zr–Ni–Cu metallic glasses near glass transition temperature. The fraction transformed y, as a function of time are of sigmoidal type at each temperature. For 0.10≤y≤0.85 the plots satisfy the Johnson-Mehl-Avrami equation and approximate to straight lines with exponent n varying from 3.1 to 3.9 as the annealing temperature increases. The activation energy for the crystallization of Ti₄₅Zr₅Ni₄₅Cu₅ metallic glass estimated from the temperature corresponding to 50% transformation is close to 380 kJ/mol.

View full abstract

The Ostwald ripening phenomena is studied using a phase-field model. The phase-field parameters are obtained at a thin interface limit proposed by Kim et al. It is shown that small solid particles preferably melt out, and large particles grow and agglomerate due to the curvature undercooling. Calculation results are consistent with the LSW rule. The rate constant of average particle radius is estimated as 0.57 at 0.2 solid fraction. The change in shape of the invariant particle size distribution due to finite volume fraction agrees with experiment and a statistical model. The difference of the calculation results by Diepers et al. and by the present work is discussed. The change in the shape of the invariant size distribution depending on solid fraction is also shown. The shape becomes more wide and symmetry with increase of the solid fraction.

View full abstract

The hydriding/dehydriding (H/D) rates and pressure-composition isotherms (PCTs) of Mg–Pd, Mg–Nd and Mg–Pd–Nd alloys with ∼ 90 at% Mg have been measured, and the characteristics of these hydrogen-storage properties have been discussed in relation with the phase structures and microstructures of the alloys studied by X-ray diffraction (XRD) and scanning electron microscopy (SEM). All these alloys exhibit a hydrogen absorbency of ∼ 5 mass% at 573 K . The PCTs of the alloys containing Pd exhibit three plateau-like regions. A complementary examination of the phase structures, microstructures and PCT characteristics at 523–623 K shows that the H/D kinetics is strongly influenced by the following three factors: (1) A catalytic effect of NdH_2.5 and NdH₃ which assist hydriding and dehydriding of Mg matrix; (2) the disproportionation reaction of Mg₆Pd to form Mg₅Pd₂ and MgH₂, which retards the overall reaction kinetics; and (3) the refinement of the microstructure which enhances the kinetics of all the H/D reactions. A melt-spun and crystallized Mg–Pd–Nd alloy which shows a fine-grained (∼ 1 \\micron) microstructure exhibits the best H/D kinetics, especially the desorption kinetics, among the three alloys investigated, and it can be a promising candidate for an efficient hydrogen storage material in the future. The formation of Mg₅Pd₂H_∼5 hydride is suggested from the PCTs of the alloys containing Pd, although it has not been evidenced by XRD and SEM.

View full abstract

This paper describes a new measuring apparatus for the maximum bubble pressure method to confirm the pressure change pattern during bubble formation for wetting and nonwetting systems. Generally speaking, a sawtooth pressure pattern has been widely accepted by many researchers in the maximum bubble pressure method. Nevertheless, when the accumulated gas volume is very small, the pressure change pattern shows a chopping wave pattern as theoretically estimated for both systems. These phenomena were confirmed by the observation of the bubbles during the measurement in an aqueous system, and also confirmed by the identical pressure pattern in a mercury system.

View full abstract

To analyze the arsenic behavior in the copper smelting process, the phase relations and the arsenic activity in the miscibility gap of Cu–Fe–S–As system with arsenic as a minor element have been determined at 1473 K by the quenching method and the double Knudsen cell-mass spectrometric method, respectively. The experiments have been conducted for the charges in the miscibility gap with mass%Fe/mass%Cu ratios of 0, 0.023, 0.059, 0.098 and 0.131 with varying arsenic content. The activity measurements indicate that the arsenic activities in both the metal and matte phases show extremely negative deviation from the ideal behavior. The Raoultian activity coefficients at infinite dilution in the metal and matte phases are found to be almost constant against the charges in the miscibility gap and they are 4.3×10⁻³ and 5.5×10⁻² for the metal and matte phases, respectively. The vapor pressures of predominant As, As₂ and AsS species in the gas phase equilibrated with the immiscible solutions have been calculated on the basis of the determined activity coefficients and they are very small at less than 1 Pa even when the arsenic content in the metal phase is increased up to 10 mass%.

View full abstract

The equilibrium between Ga–As melt, containing arsenic up to 5 mass%, and B₂O₃ flux was investigated from 1273 to 1523 K in a silica ampoule. The effect of arsenic on equilibrium contents of oxygen, boron and silicon in the melt was investigated. The results were analyzed using interaction parameters. The equilibrium distribution ratio of oxygen between Ga–As melt and B₂O₃ flux decreased with the increase of temperature, Ga₂O₃ content in the flux and arsenic content in the melt. The activity of Ga₂O₃ in the B₂O₃ flux was determined in the temperature range from 1273 to 1523 K, which is relevant to the practical process of crystal growth of GaAs.

View full abstract

The dynamic restoration process of 99.99 mass% purity aluminum has been investigated using a ⟨111⟩ oriented single crystal. The deformation was carried out by constant compression speed testing (at an initial strain rate of 1.67×10⁻³ s⁻¹) at temperatures ranging from 373 to 853 K . True stress-strain curves showed gradual hardening typical of 99.99 mass% grade aluminum. The microstructure after the testing contained newly generated high angle boundaries. Continuous dynamic recrystallization was found to take place even at 373 K . High angle grain boundaries appeared at a strain of 0.28 at 453 K, while at 0.20 at 613 K . However, grains where all the surrounding boundaries have high angle boundaries were very few.

View full abstract

Changes of phase transformation temperatures of a NiTi alloy have been investigated under non-isothermal conditions by using a differential scanning calorimeter (DSC). It has been found that the duration for phase transformation declines exponentially with increasing the heating rate. The specific heat capacity of the NiTi alloy at a constant pressure (C_P) was also determined in the experiment. Differences in the variation tendency of the C_P for the martensite and austenite phases have been observed. The temperature distribution of NiTi wire, when heated by electric current, was calculated based on C_P measured, discussions addressing the calculation results are given.

View full abstract