Table of contents

The intensity of X-ray diffuse scattering was measured at room temperature for FCC Au-Cr alloys whose compositions are 20.2, 12.6 and 8.4 at.% Cr. Diffuse maxima due to atomic short-range order (ASRO) were observed at 1¹/₂O and its equivalent positions for the three specimens, while their intensities decrease with decreasing Cr content. From computer-simulated arrangements of Cr atoms in a FCC lattice, based on the observed ASRO parameters, the Cr atoms were found to have a preference for lining up along the (100) direction for the Au-20.2 at.% Cr alloy. On the other hand, the Cr atoms tend to have random positions for the Au-8.4 at.% Cr alloy and have an intermediate arrangement for the Au-12.6 at.% Cr alloy. The relationship between the ASRO state and the magnetic property in FCC Au-Cr alloys is discussed.

The properties of the ferroelectric IV-VI semiconductor GeTe have been studied as a function of pressure up to 25 GPa at room temperature, by powder X-ray diffraction in a diamond anvil cell. In a truly hydrostatic medium, the transition from the rhombohedral to cubic NaCl-type structure was not observed up to 8 GPa while it occurred below 5 GPa in a solid pressure transmitting medium. In both cases, no discontinuity could be observed in the lattice parameter. However, the transformation probably remains of the first order, as at normal pressure, but with a very small volume discontinuity. The presence of small anisotropic stress components has a large effect on the equilibrium pressure of the transition and possibly on its mechanism. In addition, a transition to an orthorhombic structure takes place in the same pressure range; no cubic CsCl-type structure was evident below 25 GPa.

The cubic-rhombohedral phase transition at 450 degrees C of AlF₃ is studied by DSC, X-ray powder diffraction and Raman scattering. It is demonstrated that the transition is of first order with a hysteresis of about 6 degrees. It is established by X-ray powder diffraction patterns that the room temperature space group is R3c. A temperature study of the Raman scattering spectra (that confirms the above conclusion) evidences the presence of two soft modes. It is shown from group theory that the transition can be imputed to the condensation of the R₅ mode of the cubic Brillouin zone and the attribution of the Raman lines is deduced on the basis of the compatibility diagram between the cubic and rhombohedral symmetries. The frequencies of the Raman lines are used to adjust the parameters of a rigid ion model and to calculate the phonon spectrum in the cubic phase. The calculated phonon density of states appears to be strongly dependent on the soft phonon frequency.

The effect of pressure on incommensurate phases is studied on the basis of thermodynamic theory. It is shown that there are two types of pressure-temperature phase diagrams and the influence of the pressure on different properties of incommensurate crystals is clarified. The character of the diagram and of this influence depends on the type of the incommensurate phase (existence or absence of the Lifshitz invariant, and the order of the anisotropy invariant). The results are compared with experiments and show good agreement.

A method assigned to solve exactly the Schrodinger equation with non-muffin-tin crystal potential is numerically tested. The approach is based on the Green function technique. It differs from the conventional multiple-scattering methods in that the wave field psi _k is sought at some points within a cell rather than at the boundaries. The empty lattice and three-dimensional Mathieu problems are studied with the emphasis on convergence properties. Generally, the convergence is governed by three independent parameters, resulting as a consequence of the truncation of some infinite series, namely, the expansions of the Green function, potential and wavefunction psi _k sought. It is numerically shown that, by increasing the three parameters mentioned, the calculated energy eigenvalues approach the exact ones.

Polarised absorption and laser-induced fluorescence spectra of LiYF₄:Eu³⁺ were recorded at liquid-nitrogen temperature in the visible spectral range. Transitions with pure electric dipole, pure magnetic dipole and mixed dipolar character have been observed. Spectroscopic assignments to the energy levels are made on the basis of the observed energies, polarisation characteristics of the spectral lines and electric and magnetic dipole selection rules relevant to the S₄ point group. 37 Stark components belonging to ten multiplets of Eu³⁺ have been established. Free-ion and crystal-field parameters have been obtained. Simulation of the crystal-field energy levels has been attempted with and without J-J mixing effects using intermediate-coupling wavefunctions.

The authors have measured the thermopower S, from 2 to 300 K, of several ternary compounds of cerium and uranium: CeTX and UTX with T identical to transition metals Ni, Pd or Pt and with X identical to In or Sn. The behaviour of S for these cerium compounds seems to be capable of interpretation as due to the strength of the hybridisation of the Ce-4f electron with conduction electrons. On the other hand, the behaviour of S for the uranium compounds changes enormously for each compound. Prominently sharp peaks of S as a function of temperature were observed both for UNiSn near its semiconductor-to-metal transition temperature and for CeNiSn near its metal-to-semiconductor transition temperature.

The asymptotic transfer method was used to calculate the electron energy spectra, electron envelope functions and the dispersion relations of type-II semiconductor coupled quantum wells when an electric field was applied along the growth axis. It turned out that the energy spectra were nearly evenly spaced. The electron envelope functions displayed strong localisation and the dispersion relations showed a non-parabolic shape.

Based on the real-space correlated-basis-functions theory and the collective oscillation behaviour of the electron gas with effective Coulomb interaction, the many-body wavefunction is obtained for the quasi-two-dimensional electron system in the semiconductor inversion layer. The pair-correlation function and the correlation energy of the system have been calculated by the integro-differential method. A comparison with other previous theoretical results is also made. The new theoretical approach and its numerical results show that the pair-correlation functions are definitely positive and satisfy the normalisation condition.

The author numerically computes the two lowest eigenvalues of finite length spin-1 chains with the Hamiltonian H= Sigma _i(S_i.S_i+1- beta (S_i.S_i+1)²) and open boundary conditions. For a range of beta , including the value 0, he finds that the difference of the two eigenvalues decays exponentially with the length of the chain. This exponential decay provides further evidence that these spin chains are in a massive phase as first predicted by Haldane (1982). The correlation length xi of the chain can be estimated using this exponential decay. He finds estimates of xi for the Heisenberg chain ( beta =0) that range from 6.7 to 7.8 depending on how one extrapolates to infinite length.

The method of anisotropy measurement of gamma -ray emission from spin-polarised nuclei has been applied to the FCC Fe₅₄Mn₄₆ alloy to deduce the ground-state magnetic structure. Measurements were made at temperatures down to 240 mu K, and the observed anisotropy of the emission from ⁵⁴Mn isotopes embedded in the single-crystal sample gives definitive evidence for the triple-Q spin-density wave in this material.

The linear spin-wave theory of a single magnetic impurity in a Heisenberg ferromagnet with planar single-ion anisotropy is presented. The impurity-host exchange is assumed to be ferromagnetic or antiferromagnetic and the ground state of the host taken to be the classical aligned state. The calculation consists of deriving the double-time thermodynamic Green function from the corresponding equation of motion. From the poles of the Green functions the criteria for the impurity modes are obtained and calculations of their energies as functions of various perturbation parameters and for various values of the anisotropy given.

Using the mean-field approximation, the spontaneous magnetisation and paramagnetic susceptibility of localised spin ferromagnetic-superlattices, in which the antiferromagnetic films possess an odd number of the atomic planes which are parallel to the interfaces of the superlattices, are calculated numerically. A study of the spontaneous magnetisation as a function of temperature shows that the superlattices display five kinds of magnetism, and the compensation point appears for some values of the parameters J_ab, J_a and J_b. The calculation for the paramagnetic susceptibility shows that this system displays a paramagnetism that is similar to ferromagnetic, antiferromagnetic or ferromagnetic paramagnetism, depending upon J_ab, J_a, J_b and the period of the system.

The authors report that the electron paramagnetic resonance field H_res measured at a fixed microwave frequency in KCuF₃, a compound having one-dimensional Heisenberg antiferromagnetic properties in spite of its pseudocubic crystal structure, is temperature independent for the external field H parallel to the chain axis, while H_res for H perpendicular to the chain axis shows a very weak decreasing tendency with decreasing temperature to T_N=39 K over the wide region of short-range order. Considering the recent discovery that this compound has a Dzyaloshinsky-Moriya (DM) antisymmetric exchange interaction the authors suggest that H_res is governed by the DM interaction and the antisymmetric spin-correlation functions (S_n^alpha S_n^beta ₊₁-S_n^beta S_n^alpha ₊₁) with alpha not= beta ( alpha , beta =x, y, z), resulting in the temperature independence of H_res. They also point out that the two inequivalent g-tensors in this compound can be the origin of the very weak decreasing tendency of H_res observed for H perpendicular to the chain axis.

Anomalous (non-classical) temperature behaviour of birefringence at the second-order normal-incommensurate phase transition in Rb₂ZnBr₄ crystals is studied experimentally. An attempt is undertaken to discuss it using the existing theoretical approach. A critical review of some theoretical results concerning the temperature dependence of the local mean square of an order parameter is given both for the region of small corrections to the results of the Landau theory and for the scaling region. It is shown that, in the region of 'small corrections' (at mod tau mod approximately=10^-1), qualitative agreement of the experiment with the results of fluctuation theory can be recognised. Near the critical point, at mod tau mod <10^-3, the deviation from the critical behaviour of the XY model is found; the nature of this deviation is not clear.

The solvus phase boundaries of the Co-Cu system were recalculated by taking into account the effect of copper on the magnetic free energy of alpha -Co. It was confirmed that the solubility of copper in alpha -Co is significantly lowered owing to the magnetic ordering energy below the Curie temperature T_c and there is an anomaly in the Arrhenius plot of solubility. The precision lattice parameter determination by X-ray diffractometry and the intrinsic coercivity H_ci measurement by SQUID was performed and the calculated results were justified. Two different thermodynamic treatments of magnetic free energy were also compared in this study.

In the early 1970s, Basinski et al. (1972) observed that, if two solid solutions of different solutes in a given solvent (copper or silver) had the same initial flow stress at a given temperature in the range 4-380 K, then they had the same activation volume. Moreover, two alloys based on the same solvent metal, which had the same initial flow stress and activation volume at a given temperature, would show the same temperature dependence of initial flow stress and activation volume throughout that range. These observations, which are known as the 'stress equivalence of solution hardening', have been shown to be readily accounted for in terms of the kink-pair formation model of solid-solution hardening.

FeTi is considered a good material for solid state hydrogen storage; however, it must undergo an initial activation by either a vacuum or hydrogen annealing procedure after preparation or exposure to air. Previous studies indicate that the initial activation in vacuum causes a reduction of surface iron oxides by a solid state reaction with FeTi to produce TiO₂. Palladium has also been shown to be an effective barrier for eliminating O₂ or H₂ impurities in typical H₂ charging gases from reaching active FeTi. These impurities cause FeTi decomposition to iron and TiO₂. In this study, experiments have been performed to investigate interfacial reactions that occur between Fe₂O₃ and FeTi layers located below protective palladium overlayers. Vacuum annealing procedures were found to crack the palladium overlayer while forming both iron and titanium oxides below the surface. Hydrogen annealing procedures allowed reduction of the subsurface Fe₂O₃ by FeTi without evidence of cracking within the palladium overlayer. Both procedures produced Fe-Pd alloys, the latter procedure also forming Fe-Pd hydrides at >or=623 K. Results indicate that the application of palladium overlayers on FeTi may be successfully used provide the initial activation is performed in hydrogen environments.

The dielectric relaxation mechanism of 2,2'- and 3,3'-dibromodiphenyl has been examined from measurements at different frequencies and temperatures. All dielectric measurement have been done on diluted solutions in p-xylene. A simple dielectric absorption band was observed in both cases. It was found that the Goulon-Rivail model provides a good description of the dielectric relaxation mechanism.

A nickel metal cell has been used in the neutron diffraction measurements of highly corrosive molten NaOD. The scattering intensities originating from the nickel metal cell could be satisfactorily subtracted from those from the cell and the melt. With respect to the intraionic O-D correlation, the result obtained from real space is consistent with that from reciprocal space. The correlations for O-D and Na-O have been observed at 98 pm and 242 pm, respectively, while the Na-D correlation has not been observed as an isolated peak.

Quantum path integral Monte Carlo calculations have been used to study the properties of the alkali atoms Li, Na and Cs immersed in liquid ammonia. The solvent has been treated using a pairwise additive intermolecular potential fitted to experimental data. The alkali-atom solvent potential consists of two parts; an ion-core-solvent interaction fitted to quantum chemical calculations and a valence-electron-solvent pseudopotential taken from the solid-state literature. Two distinct forms of pseudopotential have been employed, one having an attractive, and the other a repulsive core. In the latter case, the equilibrium structure of Na and Cs is found to be an ionised state consisting of a fully solvated ion core, plus a well-separated, compact solvated valence electron. In the case of Li, the equilibrium structure for both models appears to be a dipolar or excitonic atom. The relative merits of the two models are discussed and, where possible, contact is made with data on metal ammonia solutions.

Theoretical work has shown that the timescale separation required for the application of the Smoluchowski, Fokker-Planck and Langevin equations to interacting Brownian particles in concentrated liquid solutions may not exist. In particular, it was suggested that the current autocorrelation function of the Brownian particles may not decay sufficiently quickly in comparison with the intermediate scattering function at appropriate wave vectors. Here, molecular dynamics (MD) calculations are performed for hard-soft-sphere liquid solutions containing Brownian-type particles of large mass and volume. The mass and volume ratio are varied systematically. The MD results indicate the following: (i) for solutions of equally sized particles, the current autocorrelation function of the massive particles decreases sufficiently quickly compared with the intermediate scattering function, thus ensuring timescale separation; (ii) massive particles of large volume diluted among light and small particles show a time decay of the current autocorrelation function, which approaches that of the intermediate scattering function. However, the similarity of the two correlation functions reduces with increasing mass ratio. Hence, for asymptotic conditions timescale separation can be expected.

The long-range behaviour of the orientational fluctuations of a pair distribution function is studied. The existence of the instability of an isotropic liquid against formation of the state with bond orientational order is shown. The estimated for the instability points for a Lennard-Jones liquid are obtained, and a comparison with computer simulation is carried out.

A neutron diffraction study has been carried out on 1 mol kg^-1 solutions of nickel chloride in deuterated methanol. Isotopic substitution was applied to the nickel ions and the hydroxyl hydrogen of the methanol. The Ni²⁺ first-order difference radial distribution functions show that the Ni²⁺ . . . methanol conformation is characterised by the mean distances r_NiO=2.07(2) AA, r_NiH(1)=2.62(2) AA and r_NiC=3.16(2) AA (H(1) refers to the hydroxyl hydrogen/deuterium atom). The pair radial distribution function, g_NiH(1)(r) was calculated, and combined with the first-order difference results to demonstrate that Ni²⁺ is coordinated to 3.7(3) methanol molecules and at least 0.8(3) Cl^-ions.

A theory for the calculation of static structures of a liquid binary alloy, using neutron scattering data on the concentration-concentration structure factor, is proposed together with its validity criterion. On the basis of this criterion, the result of the modified mean spherical model is critically discussed.

For pt.I see ibid., vol.21, p.3165 (1988). The density-functional theory of freezing is extended to the case of binary mixtures of hard discs viewed as a simple model for monolayers adsorbed on a smooth substrate. The theory is applied to the fluid-disordered crystal transition. The results are similar to those already obtained for the three-dimensional case: when the ratio of diameters is lowered, the phase diagram changes rapidly from a spindle, over a lower azeotrope to an eutectic diagram. Evidence is also found for the existence of a two-dimensional analogue of the empirical Hume-Rothery rule.

The authors report exhaustive calculations of the density and temperature dependences of the time evolution of the velocity, transverse stress and energy current density correlation functions and the corresponding transport coefficients for Lennard-Jones fluids over wide ranges of densities and temperatures. The calculated results are compared with the recent molecular dynamics data. Overall, good agreement has been achieved.

The authors present a self-diffusion study of long cylindrical micelles of lecithin in an organic solvent. The results are compared with those obtained for similar micelles of CTAB (cethyltrimethylammonium bromide) in brine. Qualitatively they observe a similar behaviour for both systems. The measured self-diffusion coefficient D_s has a power law dependence on the surfactant concentration. However, the exponent, which is salinity-dependent in the CTAB system, is constant and in good agreement with a theory for living polymers in the lecithin system. In the case of very elongated lecithin and CTAB micelles, D_s becomes independent of surfactant concentration.

The authors have performed self-consistent calculations for molecules adsorbed on a metal substrate in the presence of a scanning tunnelling microscope (STM) tip. These calculations of comprise (i) estimates of the inhomogeneous electric field, (ii) the self-consistent calculation of both electronic structure and molecular geometry using an extended semi-empirical molecular orbital method, and (iii) estimates of the tunnelling current using the resulting wave-functions and geometries. The STM tip causes changes in atomic positions which can complicate the link between current and geometry. In particular, induced molecular reorientation may well explain the lack of observation by STM of adsorbed CO on metals.

Calculations of the electronic structure, equilibrium volume Omega _e and elastic constants C_ij for the disordered Al_1-xLi_x and Al_1-xMg_x alloys have been performed. The authors use the virtual crystal approximation, first-principles pseudopotentials and the previously discussed method of calculating C_ij, based on the density functional approach. The results describe fairly well all the peculiar features of the C_ij(x) and Omega _e(x) dependences in these alloys and show that the anomalies in these dependences are due to the band structure effects.

For Cs on Cu(111) the work function is found to depend on the Cs thickness for coverages less than three atomic layers. The magnitude of the work function variation is as predicted for a free-electron-like quantum well. The photo-yield observed for a photon energy slightly above the threshold for photoemission can be used as a measure of the adsorbed amount. Via the yield, the thermal desorption of the first three atomic layers can be followed layer-by-atomic-layer. The desorption energy differs by around 35 meV between the second and third layer.