R A Moore

A computer and analytic study of the metallic liquid–glass transition. II. Structure and mean square displacements

A study, using both computational and analytic techniques, of the extent to which the cage concept of liquids applies to the metallic liquid–glass transition is presented. Monte Carlo calculations on a Ca0.7Mg0.3 system yield the angular distribution functions required to determine the cage structure and the atomic motion. The analysis parallels that for normal liquids. The results indicate that the cages are icosahedral, that they exist from the melting temperature down into the glass state, and that the cage concept applies quantitatively to the static properties of metallic binary alloys. Finally, the cage concept itself does not appear to describe fully atomic diffusion in metallic systems.

A computer and analytic study of the metallic liquid–glass transition

The present work examines two aspects of metallic liquid–glass transitions. First, a computer experiment, consisting of a temperature‐quench Monte Carlo calculation for the liquid–glass transition of Ca0.7Mg0.3, is carried out. A comparison with such calculations for monatomic metals indicates that there is a significant difference in the Wendt and Abraham criterion for monatomic metals and metallic alloys. In addition, a comparison with laboratory experiments indicates that the discrepancies between the Wendt–Abraham parameters obtained in these laboratory experiments and in the computer experiments can be ascribed to different cooling rates. Second, an analytic study of the pair distribution function near the transition temperature is presented and reproduces the linear temperature dependence of the Wendt–Abraham parameter. This gives a physical picture for this parameter and indicates that the liquid–glass transition is more likely to be a first‐order transition than a second‐order one.

On the use of the one-component-plasma model for variational thermodynamic calculations

Abstract It is shown that the plasma parameter obtained by fitting the presently available one-component-plasma structure factors to the real structure factors of the alkali liquids cannot be reasonably reproduced by existing variational thermodynamic calculations, even using a very accurate pseudopotential theory. The implications of this discrepancy are discussed.

Evidence of sp-mixing effects on thermodynamic and lattice-dynamic properties of Li and Be

It is shown that the generalised non-local model pseudopotential theory, developed recently, can be applied to the calculation of the thermodynamic properties of liquid metallic Li and Be in a fashion similar to its application to the transition metals. The corrections to the effective ionic interaction attributed to one-body plus two-body terms are much less important than those from the three-body plus higher-body terms. In analogy to sd-mixing in the transition metals, the latter are interpreted as arising from sp-mixing effects. These corrections have relatively similar magnitudes in the calculation of the phonon spectra of Li and Be and the latter corrections are required to yield reasonable agreement with observation. Finally, the Be results vindicate the physical interpretation.

On variational thermodynamic calculations for liquid metals

Abstract It is demonstrated that a recently described ab initio variational thermodynamic calculation for pure liquid metals can be significantly improved by using the hard-sphere Yukawa model instead of the usual hard-sphere model as the reference system.

Molecular dynamic simulations for crystallization of metallic liquids under different pressures

Molecular‐dynamics simulations, using 500 particles, have been performed in order to study the crystallization of supercooled liquid Na under 1 and 2000 atm of pressure. The pseudopotential method is used in order to explicitly include conduction electron contributions, which are found to require careful evaluation by the method of a two‐dimensional interpolation. The liquid and crystal structures are analyzed using a pair analysis technique. Under the above increase in pressure and with the same cooling rate, the liquid‐bcc phase transition point shifts upward by ∼25 K and the transition zone narrows.

Theoretical evaluation of the g shift in the alkali metals

A numerical evaluation of a very recent reformulation of the theoretical expression for the conduction electron g shift in metals is carried out. The calculations use the spherical cellular approximation in which, in each case, the cellular potential in the one-electron Schrodinger equation is approximated by a single spherically-symmetric ion-core potential plus the Hartree field of a uniform charge distribution equal to the average conduction electron charge density. The additional terms are not so important for Li, Na and K and hence previous favourable agreement with observation is maintained. However, these terms are important for Rb and Cs and tend to improve the theoretical values, although significant discrepancies still remain.

On the use of the one-component-plasma model for the structure factors of liquid metals

Abstract It is demonstrated that the perturbative correction, appearing in the structure factor of the simple liquid metals in the one-component-plasma model, is important and may have to be calculated more rigorously than by the usual semi-non-local or local pseudopotential theory.

A relativistic model pseudopotential: Applied to Cs and Pb metals

The authors generalise a recently published energy-dependent model pseudopotential to derive a model pseudopotential including spin-orbit interaction for use in the relativistic calculation of the electronic properties of nontransition metals. In this work the derived model pseudopotential is used to calculate (i) the electrical resistivity of Cs and Pb metals in the liquid phase and (ii) the phonon spectra of Pb metal. It appears that (i) the spin-orbit interaction enhances the liquids electrical resistivity by about 1% in Cs and about 7% in Pb and (ii) the spin-orbit effect on the phonon spectra of Pb is not at all as significant as was anticipated in the previously published work, being about 3%.

A theoretical determination of the phonon frequencies in metallic lithium

The repulsive and electronic contributions to the phonon frequencies have been calculated for metallic lithium in the symmetry directions. It is shown that the former are truly negligible as is usually assumed. For the latter the energy wavenumber characteristic is calculated using the electron-ion matrix elements for energy conserving transitions on the Fermi surface and then extended, by fitting the elastic constants, to simulate, in turn, local and nonlocal approximations. It is demonstrated that the nonlocal approach is essential in a complete theory of transport properties and phonon frequencies. Also, the crossword in the (100) branches can be explained as a reflection of the electron-ion matrix element characteristic of lithium. There are indications that anisotropic effects in the electron-ion matrix element may be important and that perturbation theory may not be adequate in lithium metal.