Table of contents

The pressure dependence of the effective mass m* of the negative ion in He II has been investigated for the first time. On this assumption of a free-particle-like ionic dispersion relation, the values of m* deduced from the measurements range from 108 m₄ at 11 bar to 77 m₄ at 25 bar, where m₄ is the ⁴He atomic mass. There are indications in the raw data, however, that there may be a small quartic term in the dispersion relation. If the data are re-analysed on the assumption that such an extra term does exist, the derived values of m* are reduced by about 8%, which then brings them into excellent agreement with the predictions of the bubble model of the negative ion.

Behaviour of electrons on a sinusoidally corrugated surface is investigated as a function of height U₀ and wavevector K, using molecular dynamics. The internal energy, interaction energy and diffusion constant are found to be periodic functions of K with absolute minimum values in the lock-in phase, K=2, when U₀=k_BT; the system is nearly isotropic at K=1. Results for other values of U₀ and K are also presented.

A disordered and frustrated Potts model is described which exhibits a number of curious features. For the m-state model the authors find a spin glass transition which within replica symmetric theory is continuous for m<or=6 but discontinuous for m>6. Replica symmetry breaking effects are much stronger than for conventional systems and the conventional Parisi ansatz fails for m>or=4. At low temperatures an infinite-range model is always finally ferromagnetic.

The K beta ^VII,VIII X-ray satellite band spectrum is measured in metallic Mg with high resolution and interpreted in terms of KL² to L²V transitions from the initial multiplets of the KL² configuration.

The (1+1)-dimensional Hamiltonian Potts model is studied for q>or=4 using finite lattice extrapolation techniques. The ground-state energy and its first derivative give information about the free energy and the latent heat of the classical two-dimensional Potts model, while the gap in the excitation energy corresponds to the inverse of the correlation length. It is shown that the finite latent heat for q>4, when the transition is of first order, comes from the crossing of levels; nevertheless there are no excitations for which the gap would vanish in the thermodynamic limit i.e. the correlation length is also finite.

Measurements of sound velocity and thermal expansion for some hexa-ammine metal (II) salts are presented over temperature regions spanning their phase transitions. Some evidence is found that the combination of elastic constants (C₁₁+2C₁₂+4C₄₄) decreases as the phase transition in Ni(NH₃)₆Cl₂ is approached from above. All phase transitions investigated are shown to be discontinuous with an associated volume change. Model calculations for the entropy increment associated with the phase transitions in the halide salts are also reported. The calculations are based on a removal of pseudo-isotropic reorientations of H₃ groups at T_c, and give results that are in reasonable agreement with experimental data.

The thermal conductivity, lambda , and the specific heat capacity at constant pressure, c_p, have been measured for solid sodium chlorate (NaClO₃) in the temperature range 100-300K. Differential scanning calorimetry (DSC) was used to measure c_p at atmospheric pressure, whilst the transient hot-wire method was used to measure simultaneously both lambda and the heat capacity per unit volume rho c_p up to a pressure of 2 GPa. The authors found that measured DSC values of c_p exceeded those previously predicted from a lattice dynamical model, consistent with a previous prediction of disorder in Cl atom positions at high temperatures. However, such disorder, if it exists, did not affect the transport of heat, since they found that lambda (T) under isochoric conditions appeared to be determined entirely by three-phonon umklapp interactions. They found no evidence for a phase transition in NaClO₃ up to 2 GPa at 300K, contrary to some previous results. However, they did observe an explosive chemical instability to occur at about 1 GPa and 130K.

Specific heat measurements on monoclinic Gd₂O₃ show a cooperative peak at 3.8K. The entropy associated with this peak is the full 2R ln 8 expected for the Gd³⁺ ions. Measurements on GdAg are affected by the presence of small inclusions of Gd₂O₃, but it proves possible to allow for these and to deduce values for the electronic specific heat constant gamma which is 11.4+or-1.7 mJ mol.^-1 K^-2. For LaAg, gamma =9.9+or-0.2 mJ mol.^-1 K^-2. The lattice heat capacities are analysed in terms of a Debye temperature which is found to vary strongly with temperature; the low-temperature limiting values are 208K and 150K for GdAg and LaAg respectively.

Results of measurements of the thermal conductivity of crystals of ³He and ⁴He containing up to 0.1% Ne which have already been reported by Berman and Livesley (1981) are considered in more detail. The way in which the various changes associated with a point defect should be combined to obtain the net phonon scattering strength is considered and a qualitative argument is used to show how the observed results may be explained by a compensation of the effect of mass changes by force constant changes. A form for the equilibrium temperature-concentration phase diagram is suggested and is shown to be consistent with most of the observations, but some results obtained at high concentrations remain puzzling.

Mullin's method (1968) is used to calculate phase separation temperatures for helium at molar volumes between 18.2 and 21.5 cm³ mol.^-1, containing 0.01% to 1.0% Ne. Although direct comparison of the results of the calculations with experiments made at appreciably smaller molar volumes is not possible, the interpretation of those observations in terms of an equilibrium phase diagram seems likely to be correct.

A perturbation method is presented for calculating the mean value of the overlap of the wavefunction amplitude in off-diagonal disordered alloys. Tight-binding Hamiltonians for some lattices (one-dimensional chain, Cavley tree, square lattice and d-dimensional cubic lattice) are considered, with nearest-neighbour interaction only and with random off-diagonal matrix elements, where the random transfer energy assumes one of two possible values. The nature of the zero-energy states (localised or extended states) is examined by calculating the mean square value of the wavefunction amplitude in the dilute limit of the concentration c. The mean value of the overlap of the wavefunction amplitude is, to first order in the fraction c, given by the summation of the squares of the lattice Green functions in the perfect lattice over all bonds of the lattice. It is shown that in real d-dimensional lattice the zero-energy state is localised for d<or=2 and the necessary condition for delocalisation is satisfied for d>2. It is also found that in the Cayley tree of coordination number z the necessary condition for delocalisation of the zero-energy state is satisfied for z>2.

An exact formula for the conductance of a finite sample in which interacting electrons move in one dimension in the presence of random impurity potentials is derived. It has a form similar to the single-channel Landauer formula (1970) for non-interacting particles. The transmission coefficient is given by a temperature-dependent 2k_f density response function of the interacting system.

The authors study the bulk conductivity of a random network of diodes and insulators as a function of the concentration of diodes p near the percolation threshold. The network considered is a square lattice and the diodes are oriented along the positive x or y axes. The results of the analogue and digital simulations show that the conductivity in the easy direction varies as (p-p_c)^t, where p_c is the critical probability for directed percolation, and t=0.73+or-0.10.

The coherent potential approximation and a k.p alloy Hamiltonian are used to provide a detailed analysis of the influence of disorder scattering on the zone-centre band structure and the electron mobility in II-VI semiconductor alloys. The weak-scattering approximation to the electron mobility is shown to apply to the common II-VI alloys. The effect of alloy disorder on the zone-centre band structure of a zero-gap alloy is examined. The total electron mobility in the mixed-crystal system Hg_1-xCd_xTe is evaluated, and the relative strength of disorder scattering in comparison with other scattering mechanisms is assessed, for a range of impurity concentration and alloy composition.

The phase diagram of the square-lattice Ising model with first- and second-neighbour interactions in an external magnetic field has been calculated by the interface method. The first-neighbour interactions are assumed to be antiferromagnetic while the interaction ratio, R, assumes arbitrary values. Closed-form second-order phase boundaries are obtained for three ordered phases (here called antiferromagnetic, layered and degenerate phases). This is not in agreement with Monte Carlo (MC) computations which proposed two ordered phases (called antiferro- and superantiferromagnetic phase). A proof is given that the layered and degenerate phase are two separate phases. For R>or=0 all transitions are second-order order-disorder transitions. For R<0 the transition from the ordered (antiferromagnetic) to the disordered phase is first order at low temperatures, and a tricritical-point estimate, which is in good agreement with MC and renormalisation-group calculations, is found. An estimate for the critical chemical potential of the hard-squares lattice gas with first- and second-neighbour exclusion is found. This is about 70% higher than the currently accepted value.

Using the recently formulated method of Plefka (1971), the authors derive the TAP free-energy functional for the infinite-range vector spin glass model. The stability condition, in the case of zero magnetic field, has a simple form: the average of the squares of the single-site susceptibilities is less than or equal to the inverse variance of the random interaction. For an applied magnetic field, stability conditions are also obtained.

As well as ClO₃ paramagnetic defect centres of ClO₂ and O₃^- are identified in Ba(ClO₄)₂.3H₂O and Ba(ClO₄)₂.3D₂O single crystals gamma -irradiated at 300K O₃^- radicals are found to be trapped at two chemically inequivalent sites. One of them is supposed to be at an interstitial site, the other being at the anion site. The O₃^- radical trapped at the anion site is perturbed by a chlorine-containing radiation damage fragment which has been identified as ClO^- from the nature of the chlorine hyperfine interaction and this radical can best be described as O₃^-...ClO^-. The O₃^- radical trapped at the interstitial site exhibits an isotropic spectrum and is supposed to be undergoing molecular rotation. The ClO₂ radicals are trapped at the anion site and exhibit proton superhyperfine interaction. The orientation, the effect of thermal and UV light bleaching and radiation damage mechanisms under different conditions of irradiation are discussed.

Measurements have been made of the quasi-elastic light scattering from molten Na, K and Cs chlorides and bromides. The salts were measured at about 35 degrees above the melting point in the frequency shift ( nu ) spectral range of 4-400 cm^-1 (0.5-50 meV). In the quasi-elastic region the depolarisation ratios ( rho _d) are large and decrease rapidly with nu up to 50 cm^-1 whereas in the inelastic range there are regions where rho _d is constant. It is noteworthy that both Cs salts show depolarisation greater than the relatively small values for Na and K salts. The reduced-intensity spectra show broad bands whose peak positions vary in a similar manner to those of the optical vibrational frequencies in the solids, i.e. with the inter-ionic force constants.

A simple scheme is described whereby X-ray absorption spectra, in clouding the very strong, rapidly varying structure near an absorption edge, may be calculated using only a short program and small amounts of computer time. Comparison with experiment shows good agreement right across the periodic table for K and L edges, with rather poorer agreement for M contributions. The method is of use in situations where a knowledge of the fine detail of the absorption spectrum is not needed, but where an accurate reproduction of the absorption over a wide energy range, and of the major structure near the edge, is. It is also of use in calculating matrix elements for use in EXAFS calculations.

A LEED intensity analysis of the Cu(001)c(2*2)-Cl structure finds the Cl atoms in the fourfold symmetrical hollow sites with a Cl-Cu interlayer spacing of 1.60+or-0.03 AA and a slightly expanded Cu-Cu first interlayer spacing of 1.85+or-0.03 AA (bulk value 1.807 AA). The calculated Cu-Cl bond length is 2.41+or-0.02 AA, in fair agreement with the value 2.37+or-0.02 AA determined with SEXAFS by Citrin and co-workers (1978, 1982). The effect of the Cl scattering potential on the LEED intensities, and hence on the structural parameters, is quantitatively analysed with r-factor calculations. The results obtained in this work speak in favour of the same overlayer model found by an earlier LEED analysis of the closely related Ag(001)c(2*2)-Cl structure but contradicted by photoemission studies, and thus help resolve the discrepancy between LEED and photoemission about the atomic geometry of this system in favour of the LEED structure.

The authors have measured the low-temperature (0.02<T<10K) electronic conductivity, magnetoresistance and Hall constant in thin (20-200 nm) bismuth films. The aim was to study weak localisation, spin-orbit and electron-electron interaction mechanisms in this well suited material. The authors compared their data with theories of Hikami et al. (1980) and Altshuler et al. (1980, 1981). It was found that their data must be interpreted in terms of two independent mechanisms, i.e. electron-electron interaction and spin-orbit scattering. Weak localisation is completely suppressed by the strong spin-orbit scattering in this material.