Table of contents

Suggestions are made to help resolve the discrepancies between LEED and SEXAFS results for the Al-O distance in the Al(111) 1*1-O structure. The need for proper characterisation of a surface structure and for an adequate data base in intensity analyses is emphasised. It is shown that with the LEED data available it is not possible to discriminate between the two proposed interatomic distances.

Phonon dispersion curves of Rb_xWO₃ have been measured by the neutron inelastic scattering technique. Nearly dispersionless phonon branches are found at the energy of about 6 meV. They can be considered as the local modes of Rb atoms which are loosely bound in the WO₃ skeleton.

The charge densities calculated with the self-consistent symmetrised orthogonalised plane wave (SCSOPW) method (X_a: alpha =1235 SOPW at each of the symmetry points Lambda , A, K, H, K', H', L and M) for trigonal Se are compared triplet by triplet with the self-consistent pseudopotential (SCPSP) densities of Starkloff and Joannopoulos (1978). The SCSOPW charge density maxima are typically twice as high as the SCPSP maxima. The SCSOPW charge density of the lowest triplet shows clear intrachain pile-ups which are missing from the SCPSP density. The reasons for the differences are discussed.

The low-frequency hopping conductivity of disordered systems is calculated in a cluster approximation at low temperatures where the temperature dependence of the DC conductivity is described by the Mott law. Calculations show that in the high-frequency region sigma ( omega , T) approximately omega ^s(T) exp(-p(T₀/T)^p), where s(T) increases, tending to unity as the temperature T decreases, p=¹/₄ for a finite density of electronic states at the Fermi level. The results obtained agree with the available experimental data on frequency and temperature dependence of the impurity conductivity in doped semiconductors.

Inelastic neutron scattering experiments were performed on the dimeric cluster system CU₂(CD₃COO)₄.2D₂O. The intensity of the 37 meV singlet to triplet transition was found to depend on the relative orientation of Q and the Cu-Cu axis as expected from theory. No energy dispersion of the magnetic excitation was observed along a* and c* of the Brillouin zone, in accordance with the localised molecular nature of the transition.

Using the ultrasonic pulse-overlap technique and multipass Brillouin spectroscopy the complete elastic stiffness tensor was determined in the temperature range from 300 to 30K. The transition to the incommensurate phase (at T₂=127K) is reflected in the elastic stiffness functions c₁₁, c₂₂, c₃₃, c₄₄, c₅₅, c₁₃ and c₂₂. The lock-in transition into the commensurate phase (at T₃=93K) has no significant effect on all the measured stiffness functions apart from c₅₅, which experiences large variations around T₃. The observed effects are interpreted in terms of coupling to energy-density fluctuations in the case of c₃₃, T>T₂. In the incommensurate phase the stiffnesses c₂₂, c₃₃ and c₄₄ are augmented by a contribution from quartic anharmonic interaction with the onset of incommensurate ordering. Another type of quartic anharmonicity is responsible for the decrease in c₅₅ in the commensurate phase. The gradual transition at the lower end of the incommensurate phase is an indication of soliton-like discommensurations.

A group-theoretical analysis of the possible continuous phase transitions in crystals belonging to the D_2d class with space centred Bravais lattice of the type Lambda _q^v is performed. The general assumption is that if the representation inducing the transition is comprised in the vibrational representation of the crystal lattice, the frequency of one of the vibrational modes of the lattice related with that representation tends to zero when approaching the transition point from a high-temperature and is the source of crystal instability during the transition. An example of results predicted by this analysis being confirmed by experiment is provided by KDP-type crystals.

A calculation of the excitations in a model of the rare gas solids using the dielectric matrix is presented. The model has only four free parameters: the effective mass of the conduction band electron, the energy gap, the size of the hole orbital and the lattice parameter. It is shown that a hydrogenic model can be used. All matrix element effects are included approximately. A broad peak in the loss spectra is predicted and is interpreted as being due to only single-particle excitations. Agreement with experiment is reasonable considering the crude nature of the model. Brief consideration is given to the effect of the finite width of the valence band and triplet-singlet splitting.

DLTS and ODLTS measurements in p-type and n-type LPE GaAs have been used to characterise the deep Cr acceptor level. Both its electron and hole emission rates have been determined separately. Together with previous determination of electron and hole capture cross sections, these measurements lead to a coherent description of the deep Cr level, which is shown to correspond to the Cr³⁺ or Cr²⁺ states according to its electronic filling. The Fermi function of the level has been deduced from these experimental results without any approximation. Furthermore, it is shown that the concentration of the Cr level can be determined not only in p-type GaAs, but also in n-type material from ODLTS measurements using only a Schottky diode. Comparison with results of chemical analysis allows us to conclude that most, if not all, the Cr atoms are electrically active in GaAs.

The spin dynamics of an impure Heisenberg antiferromagnet with site diagonal disorder only, at frequencies near the single impurity local magnon frequency, are re-described by an effective Hamiltonian in which the fundamental dynamical entities are the local modes associated with each impurity. At T=0 and at the low impurity concentrations (<1%) for which the model is expected to be useful, the q=0 response, broadened by the spatial disorder, is calculated and compared successfully with magnetic resonance experiments in FeF₂:Mn. The theory merges smoothly with results of a CPA calculation, expected to be valid at higher concentrations.

Extending to easy-plane ferromagnets the approach developed in a previous paper (Balucani et al., 1979), the kinematic inconsistencies of the Hartree-Fock temperature-dependent magnon energies are removed. To restore this consistency at the lowest order in the anisotropy a second-order theory is worked out and the leading terms in the self-energy are taken into account.

The magnetic phase diagram (H-T diagram) of the nearly two-dimensional magnetic system CoCl₂ is investigated. The spin-flop-paramagnetic phase boundary was determined by heat capacity measurements under various static magnetic fields. A self-consistent molecular-field calculation of the phase boundary based on the energy levels of the Co²⁺ ion in CoCl₂ is presented. Good agreement is found between the calculated and observed phase lines for normalised critical temperature and fields T/T_N and H_c/H_c(0).

The critical properties of the spin-¹/₂ Ising ferromagnet on a diamond lattice are calculated using the Monte Carlo technique. The heat capacity, the susceptibility, the magnetisation, and various pair and multispin correlation functions are calculated as functions of the temperature. The critical temperature, and the critical parameters pertaining to the magnetisation, are determined. The results are consistent with series analyses. Using the critical values of the correlation functions, it is shown that the basic assumption in the theory of Frank and Mitran (1978) is not exact for the diamond lattice.

The 9.4 GHz electron paramagnetic resonance (EPR) of Gd³⁺ doped in YbCl₃.6H₂O single crystals has been studied at 90-297K. The g values and linewidths of the Gd³⁺ spectra in the Kramers' host Yb³⁺ ions have been compared with those for the isostructural diamagnetic Y³⁺ lattice. The observation of resolved Gd³⁺ spectra in the presence of paramagnetic Yb³⁺ ions has been interpreted in terms of a random modulation of the interaction between the Gd³⁺ impurity and Yb³⁺ host ions by rapid spin-lattice relaxation of Yb³⁺ ions using the generalised theory of magnetic resonance by Kubo and Tomita (1954). It is found that the effective host spin-lattice relaxation time tau ₁ varies as T^-n where n=6.10(76) (B mod mod Z axis) and n=5.6(1.1) (B mod mod X axis) if 90<T<297K. Good agreement is observed between values of tau ₁ estimated from EPR linewidths of Gd³⁺ in YbCl₃.6H₂O at 90K, and those obtained by extrapolation from the direct measurement of tau ₁ for Yb³⁺ below 77K.

Spin relaxation of conduction electrons in InSb at high magnetic fields is considered theoretically and compared with existing experiments. A three-level model of band structure is used, which takes into account the main peculiarities of InSb: narrow energy gap and strong spin-orbit interaction. Spin mixing in the electron wavefunctions allows for spin-flip transitions due to electric-type perturbations (Elliot mechanism, 1954). Times of spin relaxation between the two lowest spin sub-bands are calculated for scattering by ionised impurities, acoustic phonons, optic phonons and for scattering by localised magnetic moments. It is shown that the spin relaxation in existing experimental conditions is mainly due to the scattering by ionised impurities and acoustic phonons (deformation potential interaction). The theory well describes some experiments performed with a spin-flop Raman laser, while in others the observed spin relaxation times are shorter than the calculated ones. Possible sources of these discrepancies are considered.

¹¹B NMR studies of SmB₆ support the intermediate valent character of the Sm ion: the Knight shift and linewidth data are constant over the temperature range 100-480K in agreement with its susceptibility behaviour. Furthermore, attempts have been made to explain qualitatively the electrical resistivity data.

Mossbauer spectra of ⁵⁷Fe in the chromium-substituted Zn ferrite system are reported in the temperature range 295-823K. These ferrites have been obtained by replacing a part of Fe³⁺ ions in ZnFe₂O₄ by Cr³⁺ ions. Well resolved quadrupole doublets are obtained for the two compositions studied. The value of quadrupole splitting increases with increasing concentration of chromium. This is attributed to the change in the oxygen parameter of the spinel structure with the addition of chromium. No detectable variation in the value of quadrupole splitting with temperature is observed and hence it is evident that the quadrupole splitting is mainly of lattice origin. The Debye temperatures of both the compositions have been calculated from the temperature dependence of the second-order Doppler shift.

The variation of the dielectric constant ( epsilon ') and the dielectric loss ( epsilon ") of CeO₂, Pr₆O₁₁, Tb₄O₇ and RE₂O₃ (where RE=Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm and Yb) are studied at 1.542 kHz in the temperature range 4.2-300K. A peak in epsilon " around 216K for CeO₂ and 84K for Pr₆O₁₁ is observed and identified as being due to dipoles produced by defects. An unusual peak in epsilon ' around 42K is observed for Sm₂O₃. All other sesquioxides (with the exception of Tb₂O₃) behave as low-loss ionic solids showing a systematic increase in both omega ' and epsilon " with temperature. Tb₂O₃ and Tb₄O₇ become exceptionally lossy above 150K and both epsilon ' and epsilon " increase rapidly above this temperature.

Resonant Brillouin scattering from longitudinal and transverse phonons has been investigated in alpha -ZnP₂ by tuning thermally the absorption edge through the 632.8 nm line of a He-Ne laser. The observed cancellation of scattering cross section has been discussed by a destructive interference between the resonant and the non-resonant transitions near the indirect band edge.

An explanation is given of the shifts by chemical combination of the K and L₃ absorption limits of Cu, Ag and Au given by previous researchers. Studies are made of the effect of chemical combination on the M₅ absorption limits of Pb and Bi and the effect of chemical combination on the L₁, L₂ and L₃ absorption limits of Mo. It is shown that the shifts of the absorption limits by chemical combination are explained mainly by previous proposals by the authors (Kawaka and Maeda, 1973, 1977). Two discontinuities of about comparable magnitude, L₁-I and L₁-II, have been observed in the L₁ absorption spectrum of Mo in MoO₃, and the mechanism of their appearance is discussed.

A previous LEED theory/experiment comparison for the NiO (100) surface (Kinniburgh and Walker, 1977) has been extended by the inclusion of a new I(V) data set recorded at 120 degrees C for primary energies down to 60 eV. The possibility of a small degree of rumpling of the ions in the surface layer was explicitly tested in the course of the analysis using the CAVLEED suite of programs. The results are consistent with the earlier findings and confirm that the surface is not rumpled on a scale detectable by LEED. The necessity of including exchange effects in the theoretical calculations at low energies is also demonstrated.

The problem of the reliability of the experimental data in the structural investigations by LEED is studied by artificial enhancement of some of the major causes of error. In particular, the effect on the intensity spectra of the surface imperfections and of the instrument response function is investigated. The results indicate that in the energy range from 30 to 200 eV the spectra are only slightly affected, so that a high degree of reliability of the experimental data can be accepted.