Table of contents

An Ising model for a layered two-sublattice metamagnet, with ferromagnetic interactions strong compared to the antiferromagnetic ones, is randomly diluted. Such a model has been used to describe experiments on dilute FeCl₂. The author points out a novel percolative effect which may occur above the percolation threshold of the layers. Over a range of concentrations, finite intralayer clusters can be coupled by antiferromagnetic bonds to form an infinite network in the bulk material; the infinite intralayer clusters, if present, are excluded from the network. This network may support an antiferromagnetic phase transition in which the finite intralayer clusters do not participate because of hysteresis.

An inhomogeneous structure model is applied to the Se-Te system to account for the observed relationship between the isothermal compressibility and the thermal expansion coefficient as functions of composition. The phenomena can be consistently explained in terms of changes with composition, temperature and pressure, in the fractions of two kinds of domains present: one metallic, the other non-metallic. The overall consistency obtained for the thermodynamic and electrical properties suggests that groups of 20-30 atoms simultaneously take part in the structural changes. It is concluded that heterogeneous electrical transport occurs and that the non-metal-metal transition occurs at a percolation threshold.

The authors show that for an uncompensated semiconductor such as Si:P the metal-insulator transition occurs for -K_Fl>1 and the decrease of the conductivity sigma near the transition can be accounted for by perturbation theory. A universal dependence of sigma as a function of electron density n is given. sigma decreases with decreasing n due to formation of wavefunctions decaying with distance as a power law, causing a decrease of the diffusion constant. Electron correlation has only a small effect on sigma far above the metal-insulator transition. However, as n decreases and tends to n_c, electron correlations cause a sharper decrease of sigma . For uncompensated samples transport is in a conduction band, the density of states deviates only slightly from a free-electron-like behaviour, whereas sigma drops below sigma _B, the Boltzmann value of the conductivity, due to a reduction of the diffusion constant. A discontinuous transition to an impurity band occurs when the conductivity in the conduction band is about 0.03 sigma _B and thus somewhat below Mott's value sigma _min=0.03 e²/h(cross)a, which is correct for compensated samples. Conductivities much below sigma _min for any sample must be due to long-range fluctuations or inhomogeneities. For uncompensated Si:P the authors argue for a minimum metallic conductivity of about ¹/₃ sigma _min.

The authors generalise the Mooij rule (1973) for disordered metals. They find that the temperature dependence of the conductivity sigma of a disordered metal as a function of temperature must change slope due to diffusion effects, and if interaction effects are included, sigma changes its slope three times. The crossover temperature (if it occurs at high temperatures) from positive to negative d sigma /dT due to diffusion effects varies as C⁵, where C is the average concentration of impurities or scattering centres. Another crossover temperature which separates electron correlation effects from diffusion effects is predicted. This explains the temperature dependence observed for Ge_1-xAu_x. It is also shown that the non-metallic behaviour of the AC conductivity is accounted for by diffusion effects and there is no need to invoke the concept of a pseudogap due to electron interaction. The negative magnetoresistance follows from the theory for diffusion effects in agreement with the perturbation theory of Kawabata (1980).

On the basis of linear response theory, the Hall conductivity is expressed as a sum of two contributions: one corresponding to the classical Drude-Zener formula, and a second which has no classical analogy. The developed theory is applied to the Hall effect, thermopower and thermal conductivity in two-dimensional systems. The periodic potential is taken into account.

ODMR investigations of as-grown and electron-irradiated GaP confirm that the P_Ga antisite defect behaves, in its singly ionised state P_Ga⁴⁺, as an electron trap in direct competition with band edge recombination processes. A broad photoluminescence band peaking at 1.1 eV shows triplet (S=1) ODMR which is associated with the two-electron P_Ga³⁺ state of the antisite. The triplet excited state has (111) symmetry and is formed by spin-dependent electron transfer from shallow donors to the P_Ga⁴⁺ antisite.

The experimental determination of partial structure factors for liquids which contain more than one chemical species is described. The results for a variety of liquids are reviewed and are compared with theoretical models in which specific chemical effects are included.

The structure of liquid NiTe₂ and NiTe has been investigated by applying the technique of neutron diffraction to isotopically enriched samples. The three partial structure factors relating to Ni-Ni, Te-Te and Ni-Te correlations have been successfully extracted from the experimental data. It is concluded that tellurium exists in these liquids as a dense assembly of Te^2- ions. The average coordination of Ni²⁺ is characterised by approximately three Ni²⁺ ions at 2.54 AA with four Te^2- ions at 2.56 AA for NiTe and approximately two Ni²⁺ ions at 2.81 AA with four Te^2- ions at 2.56 AA for NiTe₂. Comparison is made with the results of a new analysis of earlier data taken on liquid CuTe.

For pt.I see ibid., vol.15, p.4627 (1982). The resistivity and thermoelectric power of liquid Ni-Te alloys have been measured as a function of temperature over a wide range of compositions. It is shown that the resistivity of liquid Te can be halved by the addition of less than 10 at.%Ni. Comparison with other liquid alloys containing Te is made and a unified model for the atomic and electronic structure is proposed.

For pt.II see ibid., vol.15, p.4635 (1982). The total structure factor of liquid Tl₂Te has been measured by neutron diffraction and compared with the prediction in k-space of the mean spherical model of an ionic liquid consisting of Tl⁺ and Te^2- ions. Good overall agreement is achieved. The total radial distribution function in real space has been investigated and it is concluded that models of the melt based on discrete molecules should be abandoned.

Ionic models are commonly used in defect studies in sixfold-coordinated I-VII and II-VI compounds. The authors discuss the use of models of this type for defects in ZnSe, notably for the cation vacancy centre. They show that methods based on the shell model plus empirical interatomic potentials provide a powerful and important tool, complementary to the commoner studies of defect electronic structure. The new method is particularly effective quantitatively in calculations of (i) energies of closed-shell interstitials, (ii) optical charge-transfer energies, (iii) Stokes shifts, e.g. band gap excesses, (iv) thermodynamic energies, such as internal energies and entropies (through entropies are not calculated in the present paper), and (v) the distortion and polarisation fields near defects needed as a preliminary to fuller studies of electronic structure. Their specific application to the zinc vacancy centre in ZnSe predicts successfully the nature of the ground state, the mean optical charge-transfer energy, and other properties. They also calculate the energies of Frenkel and Schottky disorder.

The temperature dependences of the mean square fluctuation of the displacement of the lattice ions and of the local spectral density at the lattice ions sites in KH₂PO₄ have been evaluated as a function of the ratio between the correlation length xi and the range of the proton-lattice ion interactions. A comparison with recent neutron scattering and magnetic resonance measurements showed that the range of the proton-phosphorus interactions is short and about the same as the range of the proton-proton interactions whereas the range of the proton-potassium interactions lambda is so large that xi becomes of the order of lambda only in the immediate neighbourhood of T_c. The soft mode in KH₂PO₄ type crystals thus seems to be better described by a model where the whole H₂PO₄ group represents the pseudospin ¹/₂ which couples to the K motion than by the conventional Kobayashi model.

The effects of thermal treatment on CuInSe₂ have been studied by comparing the as-grown and post-anneal states. Since iron appeared to be a major contaminant, material purposely doped with this impurity has also been studied. At least four different EPR signals, labelled I, II, III, IV, have been observed: III and IV are ascribed to Fe²⁺ and Fe³⁺, II to a hole trapped on the chalcogen and I to shallow donors. The relative strengths of the signals depend strongly on the heat treatments.

The authors report measurements of quasi-elastic light scattering associated with ammonium reorientation at the order-disorder phase transitions in NH₄Cl and NH₄Br as a function of temperature and hydrostatic pressure. NH₄Cl and NH₄Br show respectively 'ferro' and 'antiferro' ammonium order at low temperatures and zero pressure. The scattering shows critical divergence in NH₄Br where the soft mode is associated with the M points of the Brillouin zone but not in NH₄Cl where the soft mode is at the Gamma point. A theory of the quasi-elastic scattering cross section is developed in which the main scattering mechanism is a second-order process involving two ammonium reorientational excitations of equal and opposite wavevectors and the dominant contributions are from M points of the zone. This is an extension of theories that successfully predict the disorder-induced Raman spectra and it gives a good description of the main features of the quasi-elastic scattering.

The authors report measurements of the elastic and photoelectric constants of TCNE cubic phase which is the phase stable at low temperature, whereas the high-temperature phase is monoclinic. The elastic constants have been studied between 100K and the transition temperature (292K). Their behaviour does not show any precursor sign of this transition. Room-temperature elastic constants are compared with the values obtained from atom-atom potential calculations and are found to be in good agreement, whereas the point dipole model fails in the quantitative description of the photoelastic properties of TCNE.

A theoretical analysis of the electron density of states in the delta -phase Au₃Si compound is given and the main aspects of the hybridisation between Si p and gold d orbitals are discussed. Differences in the electronic behaviour of Au-Si compounds with respect to those produced by nearly noble metals are pointed out. The theoretical results are compared with new photoemission data, which have been taken with synchrotron radiation using photon energy at the Cooper minimum, in order to distinguish the orbital contribution to the main structures. The general features of the experimental results agree with the theoretical model. A description of the main aspects of the chemical bond at the interface between Si and noble metals is provided.

A simple one-parameter electron density profile and the variational principle for the energy is used to calculate analytically surface energies and work functions of simple metals. The effect of the lattice of ions as well as the influence of the different corrections for the exchange and correlation energy to the local-density approximation is considered. The results agree with experiment and more sophisticated calculations.

Experiment and theory are shown to contradict each other about the image-induced binding energy B of positrons (e⁺) at metal surfaces, with (i) experimental values ( approximately 0.1 au) inferred from thermally activated positronium (Ps) emission, assuming that the barrier between the e⁺ and Ps channels is no higher than dictated by energy conservation; and (ii) calculated values (<or approximately=0.05 au) from the hydrodynamic model of the metallic plasma (with cut-off), embodying long-range correlations and collective effects. Higher theoretical values ( approximately 0.08 au) previously claimed for the same model can result only from a cut-off procedure which, through widespread, is physically unreasonable for a half-space (as opposed to an unbounded medium), truncating sums over normal modes but not affecting the amplitudes and frequencies. The author shows that this is not a true cut-off in Fourier space; it fails to limit wavenumbers, or to describe the physically expected delocalising effects of a cut-off. The quantum theory of the model is developed compatibility with a true wavenumber cut-off. By using the e⁺ workfunction phi ₊ as input, he by-passes the difficulty of calculating ab initio the important repulsive effects from the metal interior and from the surface dipole layer.

Surfaces of KCl and NaCl have been subjected to electron bombardment and the emitted species have been monitored by mass spectrometry as functions of substrate temperature and electron beam intensity. In addition to neutrals, electron-induced sputtering of positive ions and clusters is reported for the first time. Experimental evidence shows that they are formed in the surface layer during the electron bombardment. An attempt is made to explain the energy transfer mechanism by the Auger decay model. The possibility of secondary sputtering by ions or neutrals is discussed and this effect is proposed to explain the particular sputtering pattern encountered in the studied alkali halides.