Table of contents

The author presents results for the integrated spectral density and the localisation length of one-dimensional tight-binding models with weak diagonal and off-diagonal disorder. The scaling law near the band edge is satisfied for all kinds of disorder and its simple analytical derivation is given in terms of 'annealed' or 'effective-medium' approximations.

The authors present angle-integrated photo-emission spectra of the Ta 4f levels of pure 2H-TaS₂ and of Ag_xTaS₂, obtained at high resolution using synchrotron radiation, and of the valence-band regions. The data throw light on the band picture of the intercalated samples, and on the extent of the charge transfer from the Ag⁺ intercalant to the host lattice. The samples are shown to be inhomogeneous, with some regions including no Ag⁺ ions while in others the d(z²) conduction band of the TaS₂ host appears to be virtually filled due to charge transfer from the intercalant.

The authors have studied, by numerical simulations, packings of equal spheres with radius R, built on the surface of a central sphere with radius R'=kR. The packing fraction of those 2D packings does not vary with k. From that result, the authors derive the mean coordinance and the distribution of values of the coordinance of a sphere with radius R' in a 3D packing of equal spheres with radius R.

The authors show that information may be obtained about a linear combination of the short-range forces in molten salts-the short-range ordering potential nu ^s(r)-by using the ordering-potential analysis which was originally introduced to study liquid semiconductors. This procedure requires a full knowledge of the charge-charge partial structure factor, S_QQ(k). In order to fit the small-k behaviour of the experimental neutron data of S_QQ(k) in molten CuCl it becomes necessary to introduce partial ionicity in the Coulomb interaction between the ions in the melt and the authors discuss the possible reasons as to why this may be so. They also present the results of calculations of the partial structure factors, S_{alpha beta} (k), of molten CuCl. These are carried out within the mean-spherical approximation using a model of charged spheres, with charge Z=0.16 and a radii ratio of 0.36. The results reproduce the main qualitative features as compared with the neutron diffraction data. Finally they suggest that the apparently unusual features of the S_{alpha beta} (k)s in molten CuCl are likely to be typical of those systems which melt from a superionic phase, such as the copper halides and silver iodide.

A two-dimensional Euclidean plane can be identified with the Gauss-Argan plane, so that a two-dimensional triangular lattice can be identified, apart from a scaling factor, with a set of 'triangular integers' of the complex quadratic field associated with a complex cubic root of 1. It is shown that a two-dimensional modulated structure on the lattice can be characterised by a 'rational triangular number' of the quadratic field. Possible modulated structures are shown to be determined by the number-theoretical properties of the 'triangular integers'. The continued-fraction-expansion algorithm for ordinary rational numbers is extended to the case of 'rational triangular numbers'. Commensurate structures on the triangular lattice are classified by the orders of the continued-fraction expansions of the relevant 'rational triangular numbers'. The atomic configuration in a commensurate structure of a system of atoms adsorbed on a substrate with a triangular lattice is shown to have a hierarchical structure corresponding to the hierarchy among the convergents to the relevant 'rational triangular numbers'.

The density functional theory of freezing is applied to the calculation of the liquid-solid phase diagram of the restricted primitive model of a molten salt. As the temperature is lowered, the structure of the solid changes from a disordered, face-centred cubic (FCC) structure to a caesium chloride (CsCl) structure, with a triple point at about 2500 K for ions of diameter 3 AA.

The previous results on the domain volume distribution in the tetragonal phase of KMnF₃ with the space group I4/mcm have been re-examined by X-ray and neutron single-crystal diffraction in a temperature range down to 110 K; the transition temperature between the cubic and tetragonal phases is 186.6 K. For a smaller specimen examined with X-rays, the distribution is fairly constant in a temperature region close to the transition, whereas it shows a gradual change at lower temperatures. For a larger specimen examined with neutrons, the distribution is nearly constant over the whole temperature range. The difference in the behaviour can be attributed to the size or shape effect and/or to the different distributions of defects. All the previous data can be explained on the basis of the present results.

The thermal conductivities of 17 alkali halides with the sodium chloride structure have been calculated in the medium- and high-temperature ranges. Dispersion relations throughout the whole Brillouin zone have been obtained from a deformation dipole model and low-temperature input data. The linearised Boltzmann equation is solved by a variational method with two different trial functions proportional to the wave-vector and the group velocity, respectively. The results differ significantly and optic branches are found to be very important.

Raman spectra of liquid ³He have been measured with vertically and horizontally polarised incident beams, with the direction of the easy axis of the analysing polariser vertical, under a pressure of 2.05 kg cm^-2 at 1.21 K. The depolarisation property of the Raman spectrum of liquid ³He has been found to be similar to that of normal liquid ⁴He.

The analytic quadratic method of integrating singular integrals such as the density of states over the Brillouin zone has been extended to include the effect of k-dependent matrix elements. The new general algorithm is not substantially slower than the version for the density of states while retaining the advantage over methods based on expansions to linear order only.

The archetype model for phonon-assisted tunnelling is considered. Two alternative approaches are presented that mutually supplement each other. After parity-ordering (Fulton-Gouterman) a reflective ansatz for the ground state is invoked, which establishes mode-mixing and prevents strict mathematical localisation even in the 'ohmic dissipation' coupling situation. This variational argument is supplemented by a rigorous hierarchical calculation, which confirms the statement about localisation. On the other hand, a Green function technique with Hartree-Fock factorisation is employed, which traces the quantum-diffusion problem back to the solution of the oscillatory Fano problem and thus displays all mode-mixing features of the latter. The HF factorisation is justified by a coincidence argument, which interconnects the two approaches. Numerical results are given for the time evolution and an analytical long-time law ( approximately t^-1(lnt)^-2) is found for T=0. Furthermore, in the strong-coupling limit analytical estimates are given for the transmission time at T=0 and at low temperatures that deviate from the conventional results in several aspects.

An extended s-f model is used in order to study the stability of collective magnetic order in 4f systems which have antiferromagnetic ground states in the normal-valence phase and fluctuate between a magnetic and a non-magnetic configuration in the intermediate-valence phase. The local magnetic 4f moment m_f is discussed in terms of the hybridisation V and f-level position E_f, which simulates the influence of external pressure or of alloying with proper impurities. Increasing V and reducing the gap by shifting E_f towards the conduction band intensity electronic fluctuations leading to a partial or even total quenching of the local f moment in the intermediate-valence phase. Corresponding phase diagrams are derived and discussed. The authors observe magnetic first-order transitions from a highly polarised antiferromagnetic state into th non-magnetic state or into a less polarised antiferromagnetic state. In the latter case the final transition into the non-magnetic state is then of second order. The magnetic behaviour is a direct consequence of the sensitive reactions of the quasiparticle densities of states to the variations in E_f and V. This is illustrated by some typical examples.

The magnitude of the spin wave stiffness parameter. D and its dependence on temperature in YIG, Fe₃O₄ and Li-Zn ferrite has been examined and compared with theory. It is shown that values of D obtained using the exchange constants that satisfactorily explain the equilibrium magnetic properties of these systems yield results in agreement with experiment for YIG and Fe₃O₄ but are inappropriate for the Li-Zn ferrite. The temperature dependence of D in YIG and Fe₃O₄ is satisfactorily explained on the basis of the interaction of acoustic magnons with optic magnons.

Results of investigations of polarisation by injected carriers in p-terphenyl crystals are presented. A model of dielectric polarisation of p-terphenyl is discussed. The low-frequency peak of losses found in p-terphenyl crystals is interpreted by relaxation of charge carriers among deep traps.

In a heavily doped semiconductor the optical band gap increases due to the Pauli principle. In this paper the properties of Hg_0.8Cd_0.2Te are discussed using Thomas-Fermi theory which is shown to be appropriate. An expression for the Burstein-Moss shift is derived and related to the position of the transmission edge. The width of the transmission edge is shown to be a function of temperature, doping and variations in the local x-composition. Experimental results are presented to support the results of these calculations.

Starting from the continuous energy spectrum of a quantum well subjected to a perpendicular electric field, electro-absorption is analysed within the effective-mass approximation. The electric field considered is assumed to be strong enough that the exciton effects may be neglected. The influence of the difference between the well and bulk effective masses on the electron and hole wavefunctions, as well as on absorption, is thoroughly discussed. The absolute values of the well absorption are determined for a couple of examples.

The luminescence spectra and decay curves of Sb³⁺-doped Cs₂NaMBr₆ (M=Sc,Y,La) are presented. In this lattice the Sb³⁺ ion (5s²) is octahedrally surrounded by six bromide ions. The emission consists of one band. The Stokes shift increases in going from Sc to La. This is connected to the larger space available for the Sb³⁺ ion in the same sequence. The excitation spectra can be interpreted by assuming a dynamical Jahn-Teller effect to be operative in the excited state of the Sb³⁺ ion. The results are compared with those for Sb³⁺ in Cs₂NaMCl₆ and for monovalent s² ions in alkali bromides.

Energy transfer from lead to manganese ions has been investigated in NaCl and KCl as a function of lead concentration. In both crystals the intensity of the orange manganese emission was found to increase as lead concentration increases up to approximately 45 p.p.m. Beyond this lead concentration, the orange emission decreases probably due to a self-quenching effect taking place in the mixed lead and manganese aggregates. From the fluorescence intensity measurements performed in the crystals, it was concluded that approximately 70% of the lead ions undergo energy transfer in the system NaCl:Pb:Mn and that the efficiency of the Pb to Mn energy transfer process is approximately 75%. In contrast with this result, the number of lead ions undergoing transfer in the system KCl:Pb:Mn was found to be less than 30%. These results are in very good agreement with the ionic radii criteria suggested by Rubio and co-workers (1985). These criteria appear to be quite useful in order to select, for a given alkali halide host crystal, sensitiser and activator impurity ions between which energy transfer is wanted to be produced.