Table of contents

The elastic magnetic scattering from a series of compounds exhibiting two-dimensional Ising-like antiferromagnetic behaviour, Rb₂Co_cMg_1-cF₄ (c=1.00, 0.89, 0.85 and 0.70), has been examined below their Neel temperatures. The magnetic scattering from Rb₂Co_0.7Mg_0.3F₄ takes the form of almost smooth elastic Bragg rods along the c direction rather than points in reciprocal lattice space, and provides the best approximation yet found to an ideal two-dimensional long-range order. This two-dimensional magnetic long-range order is found to depend strongly on the heat treatment of the crystals and on the concentration of Mg ions, becoming less pronounced as the concentration decreases.

The results of Paskin, Gohar and Dienes (1977) obtained from their molecular dynamical simulation of shock wave propagation in a solid are analysed. Their kinetic 'temperature' profiles T₁ and T₃ behind the shock front are found to be in error, because these are defined with respect to a reference frame that experiences acceleration, instead of one that moves at a uniform velocity with the compressed lattice as a whole. The basic difficulty with their definition is that with respect to non-Galilean transformation the Newtonian equations of motion are not invariant and the kinetic energy is not conserved. The error incurred is precisely equal to the kinetic energy associated with the oscillatory motion of the atomic planes relative to the compressed lattice. This error is very large in the region immediately behind the shock front, and it is not zero even in the tail region of the shock profile, where the interplaner oscillations are expected to disappear.

Stress parameters associated with the splitting of the GR1 and ND1 lines are examined in the light of the recent suggestion that ND1 arises from the negatively charged vacancy. The stress parameters are totally consistent with this interpretation; Jahn-Teller coupling in the excited states being similar for both centres.

Infrared absorption bands due to localised vibrations of ²⁸Si and ³⁰Si at 431 cm^-1 and 422 cm^-1 have been observed in n-type silicon-doped indium phosphide grown by the liquid phase epitaxial method. The modes are ascribed to Si_In donors, but no evidence has been obtained for the existence of Si_p acceptors. Two other lines at 417 cm^-1 and 436 cm^-1 have been detected but their origin has not been determined.

A first direct observation of the '50K' structural transition in beta -phase palladium hydride has been made using neutron scattering. This observation, of a superlattice reflection at (¹/₂10), allows us to conclude that the transition is essentially an order-disorder transition involving a local rearrangement of deuterium atoms and vacancies between the eight indifferent interpenetrating FCC sublattices having twice the original lattice parameter. The low intensity of the reflection corresponds to a small value for the long-range order parameter and its dimensions in reciprocal space indicate that the long-range order only extends to about 25 AA.

It is shown that due to the hybridisation between the one-particle excitation and the two-roton states in superfluid ⁴He the interaction between the rotons is characterised by an affective roton-roton coupling g_4eff instead of the bare roton-roton coupling. The magnitude of this interaction, estimated from the neutron scattering measurements at momentum transfer between 2.4 AA^-1 and 3.5 AA^-1, is g_4eff approximately=-1*10^-38 erg cm³ and therefore the neutron scattering measurements show the existence of the two-roton bound state even at large momentum values.

The observed dependence between the Racah parameter B for Co²⁺ ions in crystals and the Phillips electronegativity of anions are presented.

The Green function approach is used for the evaluation of second-order self-energy Sigma _1k(E) in Heisenberg antiferromagnets. The decoupling scheme is discussed in detail and its step-by-step correspondence with classes of Feynman graphs is shown. The final result is in agreement with previous results obtained by a different method and with the diagrammatic approach. It is, instead, in contrast with an expression for Sigma _1k(E) found very recently by Wenzel and Wagner (1977). Some of the reasons which cause this discrepancy and lead to the incorrectness of their final result are discussed.

Mossbauer absorption is reported of ⁵⁷Fe in monocrystalline KFeS₂ perpendicular to the c axis of the crystal which yields a negative sign for the EFG and suggests that the c axis is the axis of magnetisation.

The F-level (⁴I₁₅2/) of Dy³⁺:CaF₂ is found to fluoresce for temperatures greater than 300K. The decay times indicate an overlap of the tetragonal and the cubic centres of Dy³⁺ and approximate G-level positions are obtained for both the centres.

It has been reported previously that chromium in gallium arsenide gives rise to two luminescence bands, one centred around 0.80 eV with a no-phonon line at 0.839 eV, the other centred around 0.56 eV. High-resolution cathodoluminescence measurements are reported which show that the line at 0.839 eV consists of nine components, and that there are three no-phonon components at approximately 0.574 eV associated with the lower energy band. Temperature dependence measurements show that there is no shift in the transition energies in the temperature range 4-25K and indicate that splitting occurs in both the initial and final states of the transitions.

Results of a study of pulse height defects in surface barrier detectors bombarded with low energy (E<500 keV) heavy ions (Z<or=18) are presented. A 'residual' energy defect is obtained after corrections for energy lost by the incident particles in the gold contact layer and to nuclear stopping. Since the behaviour of this residual energy defect does not follow that expected for a simple energy absorbing layer (supplementary entrance window) and since it shows changes correlated with depth in the detector, it appears reasonable that the residual energy defect is due mostly to carrier recombination in the sensitive zone presenting an electric field perturbation near its surface.

Nearly complete transitional phonon dispersion curves have been measured from the orientationally disordered (plastic) phase of crystalline adamantane. The translational modes were observed above a broad diffuse scattering component which is believed to arise from damped librational modes. No discrete librational modes were observed. The transitional modes are fitted by the solution of the dynamical matrix for a force field assumed to arise from a superposition of inter-atomic Buckingham potentials from all pairs of atoms between nearest and next-nearest-neighbour molecules. This model has been used to calculate the phonon dispersion curves for the low temperature, tetragonal phase of adamantane. The elastic and inelastic diffuse scattering have also been measured.

The crystal structure of s-triazine at 150K has been determined by X-ray diffraction. The unit cell is monoclinic with a=0.6884 nm, b=0.9569 nm, c=0.7093 nm, beta =126.61 degrees . The low-temperature structure can be generated from the high-temperature hexagonal form by shearing the crystal in the a*c plane and rotating the molecules about the b axis. The shear and rotation angles have been measured at a number of temperatures between 110K and the transition temperature of 198.07K. They are found to be proportional to each other and their variation with temperature is in good agreement with that expected from the order parameter in the case of a quasi-continuous transition.

For pt.I see ibid., vol.11, p.1761 (1978). The specific heat of s-triazine has been measured using an adiabatic calorimeter in the temperature range 100-335K. A rapid rise in the specific heat, followed by a small latent heat discontinuity was observed at around 198K and this is identified with the quasi-continuous structural phase transition observed in previous X-ray diffraction studies. A mean-field model is shown to account for the main part of the anomaly, but significant features remain which are attributed to fluctuation effects. A further specific heat anomaly was observed around 268K and this is briefly discussed.

For pt.II see ibid., vol.11, p.1771 (1978). Previous calculations of the lattice energy of s-triazine have been extended to investigate the nature of the structural phase change. When the energy is plotted in a two dimensional space defined by the shear angle of the unit cell and the rotation angle of the molecules, the points corresponding to the high- and low-temperature structures lie along the foot of a valley. The inclusion of electrostatic interactions represented by a quadrupole-quadrupole potential results in the structure at absolute zero of temperature corresponding with a minimum in the lattice energy and the stabilisation energy of the low-temperature structure is found to be close to that predicted by other methods. A 'soft mode' associated with the phase transition is identified.

The role of boundary scattering of phonons in reducing the lattice thermal conductivity is investigated in silicon, germanium and silicon-germanium alloy. The theoretical model gives a satisfactory explanation to the reduction in thermal conductivity observed in thin undoped films of silicon compared to bulk material. The magnitude of the reduction in heavily-doped hot-pressed Si-Ge alloys is somewhat less than previously reported and discrepancies still remain between theory and experimental data.

The stability of the equilibrium orientation of l parallel or antiparallel to a counterflow is studied both dynamically and statically. In contrast to the recent work of Hall and Hook (1977)-who, it is shown, use incorrect dynamical equations-the equilibrium is found to be stable in the Ginzburg-Landau regime. At lower temperatures an instability should develop.

A form of the cellular method is developed in which the non-spherical terms in the potential can be treated exactly. Examples of its use are given for copper metal and graphite crystal. In the latter, eigenvalues change in most cases by 0.2-0.3 Ryd as a result of introduction of the non-spherical potential and it is also found that layer-layer interaction cannot be considered small.

A scheme which allows one to calculate the effect of the lattice on the density response function of an electron gas is considered. This is based upon summing an infinite subset of gradient terms in the density functional theory of the exchange and correlation energy of the electrons. The scheme contains both the local density approximation and the uniform gas limit as special cases. The effect of a single gradient correction term to the local density approximation is investigated, and it is shown that this effect can be significant.

The diffusion thermoelectric power of bismuth in a magnetic field is calculated assuming a two-band model with parabolic dispersion for the holes and either parabolic or non-parabolic dispersion for the electrons. The scattering is described by a relaxation time which is assumed to be either dependent on energy as for intravalley acoustic scattering or independent of energy. Below 80K the density of charge carriers is only weakly temperature dependent and the Bethe-Sommerfeld expansion is used to derive expressions for the weak-field thermomagnetic coefficients. Above 80K a re-evaluation of earlier mobility data is performed to take into account the thermal smearing of the Fermi surfaces, and numerical integrations of the transport integrals are used. The results give a good agreement with experimental data, but this agreement is not compatible with the existence of an energy gap in the electron spectrum, such as is found in magneto-reflection experiments.

The trapping by F-centres of conduction electrons in KBr, KCl, KI, RbCl and RbI is studied via their photoconductivity at low temperature and in high magnetic fields. In zero magnetic field the carrier lifetime before final trapping is found to approach the scattering time at 4K when the density of F centres exceeds 10¹⁷ cm^-3. The trapping process is then important reducing the carrier mobility. In KBr the triplet state trapping cross-section is less than 3% of the singlet state trapping cross-section.

Boat-grown gallium arsenide doped with both silicon and chromium has been subjected to Hall effect measurements, mass spectrometry, optical and electron microscopy, and localised vibrational mode (LVM) absorption measurements. The net silicon donor concentrations ((Si_Ga)-(Si_As)) determined by LVM exceeded the carrier concentrations measured electrically by (1.5*10¹⁷)-10¹⁸ cm^-3 depending upon the silicon doping level. These discrepancies are attributed to acceptor levels associated with the chromium. A semi-insulating crystal has been prepared with a silicon concentration of 2*10¹⁷ cm^-3.

A simple method for calculating the electronic structure of metal surfaces is presented and applied to Na (001). The ionic lattice is treated as a pseudopotential perturbation on a jellium slab. The self-consistent linear response to the pseudolattice perturbation is treated with a model dielectric function of an electron gas bounded by an infinite potential barrier. Exchange and correlation effects are included in the local density approximation. The results are in semiquantitative agreement with previous non-linear calculations for Na (001).

If one is interested in the asymptotic q- kappa - omega dependence of frequency-dependent spin correlation functions it is shown that the intricacies of the corresponding mode-coupling equations are greatly reduced by assuming the validity of certain general properties of the correlation functions. A simplified version of the mode-coupling equations results for the frequency-dependent self-energy. The asymptotic solutions of these equations in the entire critical q- kappa - omega region of three-dimensional anisotropic ferromagnets above T_c are presented. The modifications in the omega -dependence brought about by the anisotropy are especially emphasised. The solutions satisfy the intrinsic assumptions self-consistently. The frequency dependence of the self-energy leads to deviations of the shape functions from a Lorentzian line shape. These anomalies are studied for some special cases. Finally the self-energy of isotropic ferro- and antiferromagnets for general dimensions d>2 are derived in order to compare the author's results with recent renormalisation group calculations valid near d=6 and d=4 respectively. The asymptotic q- kappa - omega dependences turn out to be identical in both procedures.

The magnetic anisotropy and the magnetostrictive properties of Tb in the metallic host GdAl₂ have been studied using single crystals of Tb_xGd_1-xAl₂ where x=0.02 and 0.05. The anisotropy and magnetostriction constants are found to be proportional to x and extrapolation to x=1 yields an anisotropy constant consistent with measurements on TbAl₂. The magnetostriction results are in clear conflict both with existing measurements using X-rays and with crystal field calculations; no satisfactory explanation for the discrepancies can be found.

The relationship is investigated between the equilibrium orientations of the sublattice magnetisations in a cubic antiferromagnet with linear magnetostriction and the presence of the externally applied mechanical stress. The necessary and sufficient conditions for equilibrium are examined and it is shown that variation of the stress can cause the sublattice magnetisations to rotate through a right angle.

Field modulation can be used in optically-detected magnetic resonance experiments to enhance inhomogeneously broadened signals by factors that can be as large as twenty or more. The conditions under which such enhancements are expected are discussed and illustrated with results obtained for ZhSe:F and ZnS:Al. The use of chopped field modulation to eliminate unwanted non-resonant background signals is also described.

The Debye-Waller factor is calculated classically for the Mossbauer effect of an atom which is subject to a very anharmonic force from its neighbours in an otherwise harmonic lattice. The required integration over configuration space can be performed analytically over the coordinates of all atoms in the crystal but the active atom and its neighbours. The remaining integration, which, for a face-centered cubic lattice involves still 39 coordinates, is done by a Monte Carlo method. The numerical work assumes an FCC lattice with nearest-neighbour interactions, and a hard-core repulsion between the active atom and its neighbours, the core radius being one-half the cube edge. The Debye-Waller exponent, W, varies with wavevector k and temperature T approximately as k²(A+BT), A, B, being constants, and A by no means small. For fixed k, this resembles the behaviour predicted from simpler models, but no Einstein model with a spherically symmetric potential can reproduce the dependence on both variables, thus showing that it is essential to take into account the many-body aspects of the problem.

The magnetic structure of NiS_2-xSe_x has been investigated by Mossbauer effect of ⁵⁷Fe doped NiS_2-xSe_x. Mossbauer patterns of the non-collinear first-kind ordering of FCC have been observed both in metallic and semiconductive antiferromagnetic phases. The local environment effect on the magnetic moments is weak in this system; the magnetic moments decrease uniformly at all nickel sites by selenium substitution. With increasing x the magnetic moments decrease without any discontinuous change at the metal-semiconductor phase boundary. The magnetic moment of nickel rotates towards the (111) direction with increasing temperature in NiS_1.6Se_0.4.

All the existing qualitative rules and the empirical relations in the light of the available data on the X-ray absorption edge-shifts have been tested. Although the qualitative rules are still valid, the empirical relations are found to be inadequate in interpreting the data on edge-shifts. New empirical relations based on the concept of the effective ionic charge are proposed and tested. The success of these new relations in the case of each of the 3d¹⁰, 4p⁶, 5s², 6p⁶, 4d¹⁰ and 4f¹⁴ isoelectronic series of compounds is established. The present results suggest that the edge-shifts are mainly governed by ionic charge and these depend linearly on the product of effective ionic charge and the electronegativity of the anion.

High-resolution absorption and photoluminescence measurements have been carried out on the GR1 zero phonon lines, and absorption measurements on the GR2-8 lines. These transitions are all believed to occur at the GR defect which is created by radiation damage. Carefully selected and annealed natural semiconducting diamonds have been used to minimise line broadening effects due to impurities or lattice strain. The line widths observed at 10K are believed to be close to the intrinsic values. The 1.665 eV line which is due to an A to T transition at the GR1 centre is almost three times as wide as the line at 1.673 eV associated with the E to T transition. This finding is consistent with theoretical models which propose that the A state is a vibronic level derived from the E electronic state. There is, however, no evidence for higher vibronic levels. The linewidths of the GR2-7 transitions are not inconsistent with the recently proposed two-electron model of the GR defect, but this scheme will need to be modified to account for the GR8 lines. There is some evidence, however, which suggests that the lines labelled GR8 may not be associated with the GR defect.

The L₂-L₃M ₄₅ and L₁-L₂₃M₄₅ Coster-Kronig energies for solid state systems near threshold have been calculated within the intermediate coupling framework with a semiempirical method. The importance of these processes on L-MM Auger spectral profiles is discussed. For several systems the changes in Coster-Kronig energies on oxidation of the elemental material and for the solid to vapour phase transition are considered. A qualitative explanation for the intensity variation of the low-energy satellite on the L₂-MM Auger profile for elements and their oxides near the zinc system is provided. Changes in the L₂₃-M₄₅M₄₅ Auger spectra profiles zirconium to molybdenum are predicted due to the closing of L₁-L₂M₄₅ channels. From detailed calculations L₁-L₃M₄₅ Coster-Kronig processes predicted to be allowed energetically up to metallic tin.