Table of contents

Electron capture measurements have been made using a double pulse method to give evidence that the two acceptor levels E_v+0.82 eV and E_c-0.24 eV originate from the same impurity centre, being a double acceptor. By comparing the photo cross-sections sigma _p,1⁰ and sigma _n,1⁰ of the first level E_v+0.82 eV with data in the literature the author concludes that the impurity is an isolated Fe atom on a Ga site.

A self-consistent theory of localisation is used to determine the mobility edges in random binary alloys. The model takes into account both diagonal and off-diagonal disorder and avoids many-cluster calculations. Results are presented in the limit of low concentration and compared to other theories. In particular, this theory predicts the existence of a multitude of mobility edges.

The low-frequency ( approximately 1 MHz) breakdown of the integer quantum Hall effect can be understood if semiclassical orbits are present in the sample. These orbits be very long and thus absorb very low-frequency radiation. The diagonal conductivity is not zero, precluding the observation of the resistance plateaux. Using results from percolation theory, one can predict a frequency-dependent conductance sigma _xx( omega ) approximately omega ^{-4 rho} exp(- omega ₀/ omega ), where rho approximately=²/₃ and omega ₀ can be of the order of 1 MHz or less. This form is universal in the range omega < omega ₀. In this range sigma _yx( omega ) does not deviate appreciably from its quantised value. For omega >or= omega ₀ there is a cross-over to a non-universal regime where sigma _xx( omega ) and sigma _yx( omega )- sigma _yx(0) are of comparable magnitude.

A non-local mean field approximation for amorphous Heisenberg ferromagnets is derived. It is shown that the so-called homogeneous mean field approximation is an inconsistent approximation and that their results have to be modified.

Dielectric measurements on semi-insulating gallium arsenide reveal a discontinuous change of the spectral shape of the response at around 250K. This is interpreted as an 'electronic phase transition' arising from a discontinuous change in the interactions between localised charge carriers and the lattice.

The authors report here on X-ray diffraction, magnetic hysteresis, resistivity and thermoelectric power measurements carried out at different temperatures for the system Ge_xFe_1-xNiCrO₄(x=0.0, 0.2, 0.4, 0.6, 0.8 and 1.0), which crystallises with a spinel structure. On the basis of the site-preference energy values of the individual cations (as given by Miller, and Smit and Wijn) the cation distribution is expected to be given by (Ga_x³⁺Fe_1-x³⁺) (Ni²⁺Cr³⁺)-O₄, the Ga³⁺ and Fe³⁺ ions being on the A sites and the Ni²⁺ and Cr³⁺ ions on the B sites, thus providing a selective A-site magnetic dilution with increasing x. The saturation magnetisation measurements indicate a canting of the B-site moments with an average canting angle that increases with x. Whereas the lattice is found to contract with x, the activation energy deduced from the resistivity plots and the Seebeck coefficient is found to increase with x. For a given x the resistivity decreases with temperature (T) and for a given T it increases with x. The Seebeck coefficient is found to be independent of the temperature. These results can be explained by assuming a small-polaron hopping mechanism to be responsible for the conductivity and thermoelectric effect observed in this system.

The Peierls transition in the quasi-one-dimensional conductor K_0.3MoO₃ has been studied to clarify its dynamical nature and the electron correlation problem. A giant Kohn anomaly and the central peak in S(q, omega ) were observed. The amplitude below the transition temperature T_c was also measured. An asymmetry of the elastic diffuse scattering above T_c around the reciprocal point which corresponds to the superlattice point can be explained by taking into account the structure factor of the charge-density wave. A detailed type of modulation of Mo atoms is presented by the X-ray study. No evidence for the correlation effect was observed and the transition of K_0.3MoO₃ seems to be explained by a simple band picture.

It is shown that colloids can be formed much more easily in microcrystalline powders than single crystals. The colloids can be formed at lower and lower temperatures with decreasing particle size of the powders. The ease of colloid formation is attributed to the dislocation-aided motion of F centres.

In type-IIb semiconducting diamond an electron paramagnetic resonance spectrum, which is labelled C-NL1, is observed. For detection of the resonance low sample temperatures (T<4K) and the application of a large uniaxial stress (P approximately=1 GPa) are required. The spectrum can be analysed with an effective spin J=³/₂ and a spin Hamiltonian including terms depending on the magnetic field and on the strain. From the angular dependence of the magnetic field at resonance the spectroscopic splitting factors are determined, g₁'=1.10 (+or-0.05), g'₂=+0.01(+or-0.02), and the ratio of the deformation potentials d'/b'=1.55 (+or-0.05). By analogy to the corresponding characteristics of shallow acceptors in silicon, the resonance is associated with the neutral acceptor (boron) in p-type diamond. The width of the resonance can be explained as due to broadening by random internal strains.

The experimental free-energy data for (N(CH))ZnCl and (N(CH))CoCl are obtained from C_p measurements in the temperature range 60-350K. The experimental results have been used to determine the values of the phenomenological coefficients in a free-energy expansion proposed by Mashiyama (1980) which describes the phase transition sequence. The calculated coefficients determine the experimental temperature dependence of the incommensurate wavevector as well as the experimental pressure-temperature phase diagram.

This work approaches the problem of incoherent diffusion or transport in disordered media as seen through transient grating experiments which allow a direct observation of the diffusion parameters. Using both the concept of the fractal space and the percolation theory in the critical region the authors predict the behaviour of the transient grating experiments in the two extreme cases when the interfringe is much larger or smaller than the correlation length of the disorder. They calculate the characteristic decay time of the transient grating as a function of the interfringe distance and of the critical exponents of the percolating system. In particular they show that when performed in the small-interfringe limit, the results of these experiments will directly reflect the properties of the fractal space of the diffusion through an anomalous interfringe dependence of the decay time of the transient grating.

The 1/N corrections to the large-N limit of the U= infinity Anderson model are evaluated as gaussian fluctuations about the mean-field theory of previous work, in terms of the complex sigma -field (Coleman boson). The results at T=0 for the magnetic susceptibility chi ₀ and the f occupation n_f are in agreement with those of Rasul and Hewson (1983-4). The result for the linear coefficient of the specific heat gamma is in agreement with an exact relation due to Yoshimori to O(1/N). Fluctuations in the phase of the order parameter Delta produce infrared divergences which force it to vanish as, Delta / Delta ₀ varies as Lambda ^alpha , as the infrared cut off Lambda to 0, with alpha =n_f²/N, restoring the symmetry which was broken in n_f+b⁺b, consistent with this interpretation. The same results for n_f, chi ₀ and gamma are also obtained in polar coordinates for the sigma -field and in the gauge used in the previous work.

Measurements are reported of frequency- and temperature-dependent cyclotron resonance in the ternary alloy Ga_xIn_1-xAs, close to its lattice match with InP. The band-edge effective mass is found to be 0.041 m₀ for x=0.47, in agreement with earlier measurements, but it is found that the non-parabolicity of the conduction band is approximately double that expected on the basis of k.p perturbation theory. This is thought to be due to alloy disorder. Magnetophonon resonance was also observed, giving a further measurement of the effective mass, together with the first observation of two-phonon-mode behaviour in this material.

A unified treatment of the arbitrary frequency moments I_qq⁽ⁿ⁾ of the longitudinal dielectric matrix epsilon _qq'( omega ) and of the moments F_qq⁽ⁿ⁾ of epsilon _qq'^-1( omega ) is given for the inhomogeneous electron gas. The I_qq(ⁿ⁾ are in the RPA, whereas the F_qq⁽ⁿ⁾ are exact in the non-relativistic limit excluding phonons. Furthermore, the author derives formulae allowing the F_qq⁽ⁿ⁾ to be calculated from I_qq⁽ⁿ⁾ and vice versa for a given approximation, which obeys the Kramers-Kronig relations. Using inequalities between the moments the author derives an inequality relating the static dielectric function of a non-conductor and the electron density at the nuclear sites.

A Gauss-function-like model including three parameters is established for the loss function. By fitting the moments of order one, three and five, these parameters are expressed by the charge density, potential energy and a tensor, whose trace is the kinetic energy density of the electrons. From this model a formula for the static macroscopic dielectric constant of non-conductors is derived, which is tested for diamond and which is used to discuss the Philips and van Vechten theory of electronegativity.

Unusual magnetic properties in CeSb and CeBi are interpreted in terms of the mixing mechanism between the 4f state and the valence bands. To explain the magnetic properties quantitatively the effective 4f level should be considered, which is a virtual concept introduced to interpret the higher-order effects of the p-f mixing. The calculated effective 4f level becomes shallower than the level observed by the XPS measurements.

For pt.I see ibid., vol.18, p.2697 (1985). The crystal-field splittings in CeP, PrP and NdP are calculated by considering the point-charge Coulomb interaction, the intra-atomic d-f Coulomb interaction, and the p-f and d-f mixings. The p-f mixing mechanisms, not only between the occupied 4f states and the conduction bands, but also between the unoccupied 4f states and the valence bands make an important contribution to the crystal-field splitting. The fact that the crystal-field potential in CeP is smaller than those in PrP and NdP is due to the occupied 4f level in CeP being shallower. The values of the Slater-Koster integrals (pf sigma ) and (pf pi ), are determined uniquely from the crystal-field fitting for PrP and NdP.

For pt.II see ibid., vol.18, p.2709 (1985). An anomalously small crystal-field splitting in the paramagnetic region in CeSb and CeBi compared with those in CeP and CeAs is explained by the mixing mechanism between the occupied 4f states and the valence band holes. In the paramagnetic region, the above p-f mixing gives the effective crystal-field splitting, which is estimated to nearly cancel the normal splittings extrapolated from CeP and CeAs in good agreement with experiment. The formula for the second-order transition temperature at which the population ratio of the 4f Gamma ₈ states begins to be unbalanced, is also derived. In CeBi the second-order transitions may occur, while in CeSb the first-order transition should occur in agreement with experiments.

For pt.III see ibid., vol.19, p.2721 (1985). Extremely strong magnetic anisotropy in the ordered phases in CeSb and CeBi is explained by the anisotropic mixing mechanism between the 4f states and the valence bands. The character of the ferromagnetic ground state is investigated in detail. The anisotropy energy calculated by using realistic valence bands is consistent with the strong anisotropy observed experimentally.

For pt.IV see ibid., vol.18, p.2731 (1985). To analyse the magnetic phase diagram of CeSb and CeBi, the energies of some magnetic ordered states are calculated considering the p-f mixing, the d-f exchange interaction and the crystal field by the point-charge model. Both in CeSb and CeBi a type-IA antiferromagnetic structure becomes most stable at T=0K in zero applied field in agreement with experiments. The non-magnetic (001) plane can coexist with the ferromagnetic plane in CeSb but not in CeBi in agreement with experimental results. The calculated energies of various ordered states are very close and this tendency is consistent with the observed complicated magnetic phase diagrams.

For pt.V see ibid., vol.18, p.2745 (1985). The indirect f-f interaction derived from fourth-order perturbation theory with respect to the p-f mixing is calculated using the valence bands obtained by the APW bands calculations. The type of the f-f interactions is described as the coupling through the symmetry exchange under the cubic crystal field, which cannot be written as a simple bilinear type of 4f spin operator. It is necessary to consider the short-range-ordering effect as well as the nonlinear effect of the p-f mixing to explain the fact that a type-I antiferromagnetic ordering is established by the second-order transition in CeBi.

The time-dependent shift and band width of luminescence bands due to radiative transitions between electrons trapped in tail states and recombination centres are calculated assuming two different competitive non-radiative channels: (a) non-radiative tunnelling into a second set of centres or (b) escape over tail states in the low-temperature case. It is shown that in case (a) the shift of the band maximum is logarithmic with time whereas in case (b) a much slower double-logarithmic dependence occurs. The decay of the total intensity is approximated by power law dependences in both cases.

Within a real-space renormalisation-group framework, the authors study the phase diagram of a semi-infinite cubic lattice q-state Potts ferromagnet, in which the free-surface coupling constant J_s=(1+ Delta )J_B might be different from the bulk one J_B. The authors calculate the starting value Delta _c(q) above which surface order is possible even if bulk order is absent. Their results can be alternatively seen as approximate for the simple cubic lattice (as a matter of fact, the Ising value Delta _c(2) the authors obtain approaches reasonably well the series and Monte Carlo results; consequently Delta _c(q) is expected to be quite satisfactory even for q not=2) or as exact for a well defined diamond-like hierarchical lattice. In the q to 0 limit, Delta _c diverges as q^-12/.