Search

The effects of different amounts of segregated Ga (substitutional) on an Al grain boundary have been investigated by using a first-principles pseudopotential method. The segregated Ga is found to draw charge from the surrounding Al due to the electronegativity difference between Ga and Al, leading to a charge density reduction between Ga and Al as well as along the Al grain boundary. Such an effect can be enhanced by increasing the Ga segregation amount. With further Ga segregated, in addition to the charge-drawing effect that occurs in the Al–Ga interface, a heterogeneous α-Ga-like phase can form in the grain boundary, which greatly alters the boundary structure. These effects are suggested to be responsible for Ga-induced Al intergranular embrittlement.

We have developed an efficient parallelized program for large-scale calculations in the framework of the first-principles pseudopotential method. This program uses the residual minimization method–direct inversion in the iterative subspace method which was originally proposed by Pulay and recently implemented by Kresse and Furthmuller, to efficiently obtain the electronic ground state. This method is more suitable for parallel computations with respect to each band than other algorithms such as Car–Parrinello or conjugate-gradient schemes. We have tested the message passing interface (MPI)-parallel computations of the program applied to relatively large systems using two kinds of parallel computer, the multinode supercomputer and the personal computer cluster. We have found that the developed program can attain a very high efficiency of parallelization for the two kinds of computers.

We calculate the electronic structure of an Al Σ9(21)/[110] tilt grain boundary with substitutional sodium or silicon impurity atoms using the first-principles pseudopotential method. For the grain boundary with sodium impurities and that with silicon impurities, there exists obvious differences in the relaxed configuration, valence charge density and local density of states (LDOS), which indicates a different embrittlement-promotion mechanism for these two cases.

We first review our recent simulation work on dynamic heterogeneity and supercooled liquid rheology. We then treat a supercooled polymer melt to study the stress relaxation function, transient stress evolution, shear thinning, and elongation of chains.

We numerically examine nonlinear rheology of a highly supercooled two-di men sional fluid in shear flow via molecular-dynamics simulation. The viscosity exhibits marked shear-thinning behavior when the shear rate exceeds the inverse of the α relaxation time τα. Bond breakage events among particle pairs are found to occur collectively in clusters. The characteristic size ξ of such clusters grows strongly with lowering the temperature and decreases rapidly with increasing as ξ ~ τb()1/4. Here 1/τb() is the bond breakage rate tending to 1/τα for τα 1 and growing as for τα 1. The viscosity is of order τb() in glassy states.

Ionic conduction in glass is investigated using normal-mode analysis (NMA). The origin of two distinct timescales observed in an earlier nuclear magnetic resonance (NMR) study (Y Kawamoto and J Fujiwara 1990 Phys. Chem. Glasses 31 117) is considered. First, it is shown that the present NMA gives a result that is satisfactorily consistent with molecular dynamics (MD) simulations. This indicates that the NMA method could be successfully applied to the present topic, as well as providing added support to the MD result. The two methods give a consistent explanation for the origin of the NMR observation; the fact that there are two timescales is attributable to the difference in mobility between two classes of fluoride ions, i.e., bridging fluoride (BF) and non-bridging fluoride (NBF). The present NMA also provides information additional to that from the MD results. The difference in mobility between BF and NBF becomes significant only below the glass transition temperature . Further analysis has been made to characterize the conduction mechanism more clearly, by the use of normal-mode excitations. The main finding is that an activation process becomes important for BF conduction below , whereas no such process is observed for NBF.

First-principles calculations of the magnetic anisotropy energies of Pd/Co(111) and Pd/Co(001) metallic multilayers have been performed to investigate the orientational dependence of the magnetic anisotropy. The calculated results indicate perpendicular magnetic anisotropy for both systems in accordance with experiments, and the anisotropy energy of the (111) oriented multilayer is more than four times larger than that of the (001) multilayer. The perpendicular magnetic anisotropies of these systems are explained by the same electronic origin, a large LDOS of |m| =2 character of Co minority spin near the Fermi energy. The small anisotropy of the (001) multilayer can be attributed to the weaker hybridization at the interface as compared to the case of the (111) system.

We prepared multilayers on glass substrates by RF magnetron sputtering at room temperature. multilayers exhibit a magnetic easy axis perpendicular to the film plane up to a larger magnetic layer thickness than do Co/Pd multilayers. In particular, multilayers became perpendicular at 33 A. The magnetic anisotropy can be explained by the strain effect. The role of the magnetoelastic energy due to the in-plane strain in the system is pointed out and discussed. We also investigated the magneto-optical polar Kerr effect of these multilayers. The polar Kerr rotation angle increases when the Au concentration is 4.5 at.% compared to Co/Pd multilayers but decreases with increasing Au content.

Surfactant epitaxy has been expected to be a powerful method to improve thin-film growth from 3D island to layered growth, but the atomic level process of surfactant epitaxy has not been revealed yet. The purpose of this paper is to elucidate the phenomenon of surfactant epitaxy by molecular dynamic simulations. It is found that a smaller surface energy is necessary for the surfactant to be stable on the surface and that the difference of the atomic radii between the surfactant and the adatom plays an important role in the process of quick exchange of their positions.

Molecular dynamics simulations are performed for a ZrF4-BaF2 glass system, which is expected to be a good ion-conducting material, in order to investigate the fundamental conduction dynamics of fluoride ions in the glass. The simulation results show that the ionic conduction is governed by the hopping motion of fluoride ions, with an average hopping distance of 2.5 AA, in the glass. This is clearly different from the diffusive motion observed in the melt. It is also found that the mobility is clearly different between bridging (Zr-F-Zr) and non-bridging (Zr-F-(Ba), (Ba)-F-(Ba)) fluoride ions in both glass and melt. The present result gives a satisfactory explanation of two distinct time scales observed in a previous NMR measurement (Y. Kawamoto and J. Fujiwara, 1990 Phys. Chem. Glasses 31 117). The first-passage-time approach is applied to the melt to determine the self-diffusion coefficients of the bridging and the non-bridging fluoride ions separately.

Molecular dynamics simulations are performed for ZrF4-BaF2-EuF3 glass systems in which 0, 5 and 10% F- ions are substituted by Cl- ions. Changes in local structures of the glasses around each cation are investigated in terms of increasing amount of substitutions of Cl- for F-. The average coordination number of F- around Zr4+ is found to be approximately 7.7 and is independent of the amount of Cl- substitution. The power spectra of the velocity autocorrelation functions are also calculated. The Zr-F stretching mode observed at around 800 cm-1 shows no shift with Cl- substitution. These results show that the local structure around Zr4+ is not greatly affected by the small amount of Cl- substitution. The F- coordination number around Eu3+ at a radius of 3.0 AA and that around Ba2+ at a radius of 3.5 AA are found to decrease from 8.0 to 5.6 and from 9.2 to 5.1, respectively, as the Cl-fraction increases. This means that Cl- ions substituted for F- exist almost around the Eu3+ or Ba2+ cation. Thus, even the small amount of Cl- substitution greatly changes the local structures about Eu3+ and Ba2+.

Coherent and incoherent multilayer models are constructed by structural relaxation using variable-cell molecular dynamics under constant pressure. The models are restricted to two dimensions for computational efficiency. A fully relaxed multilayer does not show the supermodulus effect and its elastic constant does not vary even when repeat length changes. In contrast, the elastic constant of the multilayer models in which only one of the constituent layers is strained under the influence of the substrate shows a drastic increase by 50% of that of the bulk. This suggests that the residual stress in the multilayer can cause the supermodulus effect. Several experiments have reported much larger enhancements (several hundred %); therefore, the electronic effect associated with superlattice periodicity is considered to be essential for full interpretation of the supermodulus effect.

The magnetocrystalline anisotropy energies of X/Fe (X=Pd,Pt,Au,Ag) multilayers have been calculated from first principles within the local-spin-density approximation using the linear muffin-tin orbital (LMTO) method, including spin-orbit coupling. Although the calculated magnetostatic energies favour in-plane magnetization directions, the easy axes of Pd(2ML)/Fe(1ML), Au(2ML)/Fe(1ML) and Ag(2ML)/Fe(1ML) (ML: monolayer) multilayers are perpendicular to the film plane, because of the large electronic contribution to the perpendicular anisotropy. The calculated anisotropy energies of Pd/Fe multilayers decrease with increasing Fe layer thickness, which is in good agreement with experiment.

The results of in situ reflection high-energy electron diffraction, X-ray diffraction, transmission electron microscopy and magnetization measurements of Ag/Cr metallic superlattices (with a few monolayers of Cr) are presented. The superlattice structure is confirmed even in the superlattice with one monolayer of Cr. It is found from simulations of the observed X-ray diffraction profiles that dramatic changes in the interplanar distances of Ag and Cr take place as a function of the number of Cr monolayers in a period. The interplanar distance of Ag contracts, whilst, on the contrary, the interplanar distance of Cr is stretched; for the superlattice which includes two monolayers of Cr it is stretched by up to about 10%. Magnetization measurements show that the magnetization per Cr atom at 5 K with an applied field of 5 T is at a maximum for two monolayers of Cr.

The authors present the temperature dependence of positron lifetime parameters measured on a series of specimens of YBa2(Cu1-xMx)3O7-y (M=Fe, Ni, Zn) in the temperature range 20-300 K. In a non-doped specimen, the defect-related component tau 2 showed a step-like increase in the temperature range 40-100 K. Above 180 K, the mean lifetime tau mean decreased with increasing temperature. These variations in tau 2 or tau mean are explained in terms of a thermally activated escape of positrons from the two kinds of trapping centres. The bulk lifetime tau bulk showed a slight decrease in the region 100-300 K. The iron doping moved the shift of the transition region of tau 2 towards a higher temperature and removed the decrease of tau mean out of the temperature range in which experiments were performed. These phenomena are ascribed to the enhancement of the positron-defect binding energies, which is theoretically predicted. On the other hand, nickel and zinc did not much alter the behaviour of tau 2, but affected that of tau bulk. A direct correlation between the superconducting critical temperature (Tc) and the lifetime-temperature curve was not found.

The polar and non-polar interfaces of grain boundaries in the zinc blende structure can be defined by the stoichiometry in the interface region, and it is possible to construct two polar interfaces and one non-polar interface for a symmetrical-tilt grain boundary for which the interface comprises the polar surfaces.

The atomic and electronic structure of the (122) Sigma =9 grain boundary in cubic SiC has been calculated for the first time using the self-consistent tight-binding (SCTB) method. An atomic model consisting of a zig-zag arrangement of five-membered and seven-membered rings similar to that in the same boundary in Si or Ge has been constructed from a high-resolution electron microscope image, although Si-Si and C-C wrong bonds are repeated alternately at the interface in this model. The authors have also performed calculations of the same boundary in Si using the SCTB method for comparison, and have obtained results similar to those previously obtained by other theoretical schemes. The calculated boundary energy in SiC has shown that the present atomic model can exist stably compared with the two surfaces, and the calculated boundary electronic structure in SiC has no deep states in the gap as well as that in Si. These results indicate the possibility that stable boundary structures can be constructed by arranging structural units in SiC as well as in covalent semiconductors. However, it has been found that the presence of the wrong bonds greatly influences the boundary energy and the boundary electronic structure.

The self-consistent tight-binding (SCTB) model proposed by Majewski and Vogl (1987) has been extended to be applicable for calculations of lattice defects in solids or disordered systems with both ionic and covalent characters that cannot be treated using other types of tight-binding theories. The precise formulation of electronic structure, total energy and atomic forces in the supercell technique has been presented. In order to apply this method to lattice defects in SiC, the parameters and functional forms have been examined so as to reproduce the basic properties of Si, SiC and C. The nature of the bonding and the phase stability in Si and SiC have been analysed by the present SCTB method.

Positron lifetime measurements were performed on YBa2(Cu1-xFex)3O7-y with x=0, 0.01, 0.02, 0.03, 0.06, 0.09 and 0.12 as a function of temperature between 20 K and room temperature. The positron mean lifetime and its thermal behaviour strongly depend on the iron concentration. A plausible model is proposed, in which the observed results are explained in terms of positron weak localisation due to copper-oxygen chain disordering.

We have investigated the thermal stability of GaAs-oxides grown by atomic force microscope (AFM)-assisted anodic oxidation to identify the conditions suitable for fabricating oxide nanomasks for molecular beam epitaxy (MBE). The oxides grown at bias voltages, V ox, less than 30 V were desorbed after standard thermal cleaning in MBE, while the oxide patterns fabricated at V ox ≥40 V survived on the GaAs surfaces. From X-ray photoemission spectroscopy, we have found that the better thermal stability of AFM-oxides grown at V ox > 40 V can be attributed to the formation of Ga2O3 and that Ga2O3 can be used as nanomasks for site-controlled MBE growth.

The behavior of a supercooled polymer melt composed of short chains with ten beads near an oscillating plate are simulated by using a hybrid simulation of molecular dynamics (MD) and computational fluid dynamics (CFD). In the method, the macroscopic dynamics are solved by using CFD, but, instead of using any constitutive equations, a local stress is calculated by using a non-equilibrium MD simulation associated at each lattice node in the CFD calculation. It is seen that the local rheology of the melt varies considerably in a thin viscous diffusion layer that arises near an oscillating plate. It is also found that the local rheology of the melt is divided into the three different regimes, i.e., the viscous fluid, viscoelastic liquid, and viscoelastic solid regimes, according to the local Deborah number $De$, which is defined with the Rouse or $\alpha$ relaxation time, $\tau_R$ or $\tau_\alpha$, and the angular frequency of the plate $\omega$ as $De^R$=$\omega\tau_R$ or $De^\alpha$=$\omega\tau_\alpha$. The melt behaves as a viscous fluid when $De^R\lesssim 1$, and the crossover between the liquid-like and solid-like regime takes place around $De^\alpha\simeq 1$.

We have performed a first-principles computational tensile test (FPCTT) on an Na-segregated Al grain boundary (GB) with an Si additive. We show that the Si additive in the GB greatly increases both the tensile strength and the toughness of the Na-segregated Al GB. We demonstrate that the final GB fracture is dominated by the breaking of interfacial stronger Al–Si bonds according to the bond evolution with increasing strain. Based on the Na-induced Al intergranular embrittlement mechanism explored before and the present calculation results, we propose a GB-strengthening mechanism by adding a strengthening element such as Si for Al alloy to suppress the intergranular embrittlement by an Na impurity. Such an intergranular embrittlement suppression mechanism can explain the experimental observations.

Surface acoustic waves of transition-metal nitride TiN/CrN superlattices prepared on MgO (001) and glass substrates have been examined by Brillouin scattering. The superlattices were capped by a 15-nm-thick Al layer to enhance the Brillouin scattering efficiency. Scattering from the Rayleigh wave and the longitudinal guided wave have been observed. The Rayleigh wave velocity is independent of the superlattice period in a range of 30 to 0.6 nm. On the other hand, the longitudinal guided mode velocities exhibit a softening of about 10% and show a minimum at around 5 nm and 3 nm of the superlattice periods for the glass and MgO substrate superlattices, respectively.

Reproduction capacitance signal of a ferroelectric-semiconductor disk memory was studied by numerical simulation under an extremely high-density recording condition. Information is recorded by controlling the direction of the ferroelectric polarization which induces the change of depletion-capacitance of the semiconductor substrate. When the spatial wavelength is less than 1 µm, depletion regions connect with each other, which is called the “punch-through” effect. It has been considered that the depletion-capacitance modulation disappears in the “punch-through” state. The present numerical simulation, however, suggested that for the ferroelectric-semiconductor structure, the modulation signal of depletion-capacitance does not disappear even in the “punch-through” state, because there exist island accumulation regions at the semiconductor surface due to the ferroelectric polarization of domains with opposite direction. This result is consistent with previous experimental results. It is suggested that the ferroelectric-semiconductor disk memory has a potential for higher density recording than silicon nitride-silicon disk memories.

The magnetic anisotropy energies of multilayers which include only magnetic elements, Ni/Co and Ni/Fe, are calculated from first principles. Both systems show perpendicular anisotropy but the anisotropy energy is somewhat smaller than that of well-studied systems such as Pd/Co. The LDOS (local density of states) of Ni is as large as that of Co(Fe) at the Fermi level and contributes directly to perpendicular anisotropy, which means that the magnetic moments of Ni are also normal to the film plane. The anisotropy cannot be explained only in terms of the electronic structure of Co(Fe) near the Fermi level but that of Ni must also be considered. The origin of the perpendicular anisotropy is attributed to a large LDOS of |m| =2 character of Co(Fe) and Ni minority spin near the Fermi energy.

Noncontact evaluation of capacitance disk memory with ferroelectric semiconductor (FS) structure was performed in order to avoid wear of the head and medium. Head-medium spacing was actively controlled using differential capacitance between the head and medium as the active feedback signal. Using this system, head-medium spacing could be stably controlled to be as low as 30 nm, which is less than that of a magnetic hard disk. Ferroelectric film of the medium was poled, which corresponded to recording in memory application, while head-medium spacing was kept constant using the active spacing control system. Change in capacitance due to poling direction, which corresponded to the reproduction signal in memory application, could be observed at the spacing below 100 nm. As a result, we confirmed that ferroelectric/semiconductor disk memory could function without head-medium contact. The surface image was also obtained using the capacitance signal while differential capacitance was kept constant.

The atomic structures of the FCC/FCC metallic superlattices Au/Ni and Cu/Ni were simulated by the molecular dynamics method, which Morse potentials were used as the interatomic potentials. In contrast to the other previous studies, the authors have treated here incoherent realistic structural models. The simulated structures were used to calculate the Young's modulus along the FCC (111) direction, YY(111), and the biaxial modulus, Yb(111). The biaxial moduli for Au/Ni and Cu/Ni were compared with the experimental results and it is shown that the enhancement of elastic moduli can be well reproduced qualitatively by the authors' incoherent models. The enhancement of elastic moduli can be ascribed to the existence of periodic strain in the multi-layered films.

All the Bernal holes were found in a realistic structural model of amorphous iron generated by a computer. Tetragonal dodecahedron occurred much more frequently in number than archimedian antiprism capped with two half-octahedra.

The atomic and electronic structure of the {211}/{111} facets in Si have been examined based on the atomic model of Bourret and Bacmann. The optimum atomic structure has been determined by energy-minimisation calculations in the Keating model, and the electronic structure has been calculated using a supercell technique and a tight-binding model. Relatively large distortions exist at one of the two types of facet junctions between the {211} and {111} interfaces. This distorted region considerably influences the electronic structure at the facet junction compared with that at the other regions. However, there are no states introduced inside the minimum gap.

The migration energies of substitutional impurity atoms based on the vacancy mechanism in cubic metals have been formulated in terms of the pseudopotential theory. Calculations have been carried out for fifteen kinds of impurity atoms in aluminum.

The vibrational density of states (DOS) of a realistic model structure of amorphous iron was calculated in two ways, using the equation of motion method and the recursion method, in which the quantitative disorder due to variations of interatomic force constants was found using the empirical potential of Pak and Doyama for alpha -Fe. It is shown that the vibrational DOS of amorphous iron has two peaks, an enhanced broad one on the low-frequency side and a fairly reduced one in the high-frequency side. The Debye parameters which characterise the whole spectrum and can be obtained by Mossbauer spectroscopy, were calculated. They are about 6% smaller than those values for the crystalline state ( alpha -Fe). The 'phonon' dispersion curves in amorphous iron were also computed from the wavenumber-dependent spectra both for longitudinal and transverse excitations. The authors observed a considerable softening of phonons associated with the metallic amorphous state.

The atomic configurations and relaxation energies of vacancies in alkali metals have been computed using long range oscillatory pairwise potentials. The results for Na were compared with those of the published theoretical work. Agreement was fairly good, in particular, with the results of the recent calculations by Flocken and Hardy (1969), and Ho (1971). The electrical resistivity due to vacancies has also been calculated, using the pseudopotential method in which the lattice relaxation has been taken into account. It was shown that the effect of lattice relaxation was very important in the calculation of electrical resistivity.

Calculations of electrical resistivity of five different types of single interstitials in aluminium have been carried out using the pseudopotential method taking account of the lattice distortion near the defects. The order of resistivity increasing is the body centred, 100 split, 110 split, tetrahedral, and 111 split types. The directional dependence of resistivity due to the anisotropy of the defects has not been considered.

Using the pseudopotential approach, the interaction between an impurity atom and a vacancy, in aluminium, is shown to be long range and may possibly be repulsive in some cases.

Vacuum ultraviolet photoemission spectra (UPS) of the blue bronze K0.30MoO3 were measured at 80 and 293 K by using synchrotron radiation. The spectrum at 80 K showed an enhancement of the photoemission intensity at the binding energies (E B) of around 0.1 and 1.0 eV compared with that at 293 K. These results are interpreted as due to the reduced zone effect by the Peierls transition taking place at 180 K. The inverse photoemission spectrum measured at room temperature has shown that the density of the unoccupied conduction band states is rather low near the Fermi level (E F) and has a broad maximum around 2.5 eV above E F.

We simulated the growth of metal films and investigated the relationships between their change of surface step densities and growth conditions. We modeled surface migration by the Arrhenius relation and considered two cases of decomposition of clusters. In the first case, no decomposition of cluster takes place, and in the other case cluster decompositions are taken into account. When the substrate temperature was above 430 K, clear periodic oscillations of the number of steps on the surface were observed in both cases. The amplitude of the oscillation for the no-decomposition model was larger than that in the decomposition case. However, the number of steps was smaller and the growing surface was flatter in the latter case.

The atomic structure of a Σ=11 tilt grain boundary in gold was simulated by the molecular dynamic techninue using the Morse potential. The simulated structure is compared with the lattice image of a Σ=11 boundary in gold observed by high resolution electron microscopy. The local density of states of phonons was calculated by the recursion method for atoms near the boundary plane. It was found that the resonant mode with a large amplitude occurs with atoms at the boundary plane.

Multicomponent zinc oxide varistors were electron irradiated and the changes of electrical properties were measured. The electron beam energies were 800 keV, 1 MeV and 2 MeV. In the pre-breakdown region leakage current increased and the activation energy decreased. This degradation was symmetrical with respect to the applied field. In the breakdown region the nonlinear exponent decreased greatly and the breakdown voltage also decreased slightly. These results indicate that defects created by electron irradiation cause the increase of donor concentration, which lowers the activation energy of the thermionic emission. The degradation by 800 keV electron beam is the greatest and the difference between 1 MeV and 2 MeV is very small.

Electrical and electroluminescent properties in P-diffused Cd1-x Mg x Te (x=0, 0.2, and 0.35) p-n junction diodes were measured at room temperature and are discussed in detail by using the theory for injection-recombination kinetics with emphasis on radiative recombination. The I-V characteristics of these diodes show an exponential behavior of the form I=I 0exp (q V/m k T), where m=2 at low biases and m

Solid solutions of MgTe and CdTe were prepared throughout the range 0∼0.35 mole fraction MgTe. P-n junctions were formed from CdTe, Cd0.8Mg0.2Te and Cd0.65Mg0.35Te by diffusing P under Cd atmosphere for 10 to 72 hours and electroluminescent properties were investigated at room temperature. The diode with x=0.35 emitted visible red light (∼6800 Å) with external quantum efficiency of 2×10-6 for 10 mA at 300°K. The emission spectra of Cd1-x Mg x Te diodes (x\neweq0) are similar to the spectra due to Schoen-Klasens type transition as supposed in CdTe diodes. Capacitance measurements, current-voltage characteristics and transient effect measurements imply that, at high current density (∼100 A/cm2), diffusion current rather than the recombination current in the space-charge region relates to radiative recombination.