Table of contents

An expression is obtained for the low-velocity stopping power of an electron plasma in thermal equilibrium, which applies for all degrees of plasma degeneracy. In the limit of high densities the result agrees with the Fermi-Teller approximation for a cold plasma, and with a previous approximation of Skupsky (1977) for all degeneracies. It gives better agreement with the Lindhard-Winther result (1964) at lower (metallic) densities for cold plasmas, and it also agrees with a previous result of Arista and Brandt (1981) in the opposite limit of dilute nondegenerate plasmas.

The effective transmitted intensity of LEED is calculated to show a strong damping of the incident beam inside a crystal under surface state resonance conditions, and is compared with recently reported current image diffraction (CID) patterns. Good agreement between the transmitted intensities and CID patterns have been obtained, which shows that CID patterns originate mainly from surface state resonances. This result implies that CID patterns should be sensitive to surface states of a crystal.

Electrical resistivity, thermopower and specific heat have been measured on single crystals of monoclinic eta -Mo₄O₁₁ grown by vapour phase transport. The anisotropy of the electrical resistivity establishes that eta -Mo₄O₁₁ is a quasi-two-dimensional metal. All the physical properties show that two phase transitions take place at T_c1=109K and T_c2=30K. The authors propose that these transitions are due to charge density wave instabilities.

The authors report analytic results of omega -dependent dielectric function with local-field corrections, using the Penn model within the RPA. The calculations, which are valid for the entire energy range, yield good agreement with detailed numerical calculations as well as experimental data.

For original paper see ibid., vol.15, p.3967 (1982). Recently, P. Wurfel presented a paradox, where an engine converts heat into work with an efficiency of 25%, although it functions between two heat reservoirs at identical temperatures. It is shown in the present paper that the paradox results from neglecting the entropy creation at surfaces of bodies that emit radiation without at the same time absorbing an equal amount of radiation. By correctly taking into account the entropy creation, it is shown that Wurfel's engine has an efficiency of 0%, i.e. cannot produce work in a stationary (or in a cyclic) process. The opportunity is taken to extend the model to a wider field of applications, resulting in valuable formulae, e.g. the maximum efficiency of the conversion of solar energy into work by use of a single solar collector.

Free-energy barriers separating metastable incommensurate states are studied theoretically for a simple model of modulated systems. Analogies with superfluid ⁴He and superconducting wires are exploited. Two types of free-energy barriers associated with the creation of a vortex ring and the unwinding process are identified and are discussed in relation to the decay processes of metastable states.

Measurements have been made on fluid Hg in the temperature range 300-1800K and at pressures up to 200 MPa. The authors have measured the vapour pressure, density of saturated liquid and liquid phase isochores, and the electrical conductivity of saturated liquid and vapour. The critical temperature, pressure and density were estimated as 1750K, 172 MPa and 5.8 g cm^-3, respectively. It was argued that the low-density limit of the supposed semiconducting region for Hg occurred at the critical density.

Using the T-matrix technique the local density of states of a hydrogen atom for hydrogenated a-germanium is calculated. It is shown that a resonant mode is obtained in the acoustic range and that this mode is similar to that of a simple mass defect.

The density of states of a disordered system whose interaction matrix consists of 2m+1 diagonals of non-zero elements (a band matrix of width 2m+1), is obtained by a direct application of the Gauss elimination procedure, without having recourse to a reduction process to a tridiagonal basis. The method is extended to the evaluation of spatial amplitude correlations and related quantities. The method is illustrated by computation of the density of vibrational states of the simple square and cubic lattice, with nearest-neighbour interaction.

By equating the electronic structure yielded by a tight-binding model Hamiltonian of the covalent solids with only nearest-neighbour interactions to the corresponding free-electron energy bands, Froyen and Harrison (1979) have shown that the LCAO parameters should vary essentially as 1/d². This simple relationship has since been employed in studies of the effects of disorder by Tanaka and Tsu (1981) and in calculations of surface relaxation and reconstruction by Chadi. In this paper the authors investigate the accuracy of this assumed 1/d² dependence by matching the energy spectra obtained from the LCAO model Hamiltonian to the pressure-dependent, self-consistent, OPW band-structure calculations on Si by Stukel and Euwema (1970). They conclude that the 1/d² variation significantly underestimates the changes in electronic structure with lattice constant. Finally, they consider the implications of this result in relation to the determination of surface geometries.

On the basis of augmented-plane-wave band-structure calculations for the insulating compound MgO the authors have calculated the momentum density. Compton profiles, and B(z) functions. In order to derive physical insight from the results they have put much effort into understanding the momentum density in terms of real-space wavefunctions and nearest-neighbour interactions. To this end they performed a backtransformation of the momentum densities into real space, thus obtaining Wannier-like functions. The Compton profile and the B(z) function along (100) are distinctly different from the ones along the other two main directions, and this is also revealed by the experimental results. However, there is a significant discrepancy between the results and experiment for the B(z) function along (100). A similar discrepancy is also observed by another calculation using a different model. The source of this discrepancy is not understood.

Numerical calculations have been made of the rates of scattering and momentum relaxation of electrons caused by unscreened polar optical phonons in quasi-two-dimensional layers, using a simple infinite-depth square-well model. The results are compared with approximate analytic formulae, and with the situation in the bulk. Some of the major differences from scattering in the bulk are (i) an abrupt threshold for phonon emission appears, (ii) rates just above threshold are enhanced and (iii) the scattering just above the threshold of emission is more isotropic. These differences are intimately related to the energy dependence of the density of states and to the polar interaction, and are expected to be largely independent of the model. The form of the polar interaction strongly favours intra-sub-band scattering over inter-sub-band scattering and this effect has consequences for the luminescent spectrum expected from hot injected carriers. The large increase of the momentum relaxation rate at the emission threshold has consequences for transport phenomena.

A theory is given of the free energy, the mass shift and the linewidth of the cyclotron resonance of electrons coupled to ripplons (ripplonic polarons) on the liquid He surface for the entire range of the magnetic field. The path-integral approach developed previously is utilised. The theory is constructed in such a way that, while it has the correct limit at weak magnetic fields, the full quantum character due to the Landau quantisation is incorporated at high magnetic fields. The linewidth at high fields is shown to have the form that is expected from the self-consistent Born approximation. The results are critically compared with experiments.

The free energy, the effective mass and the DC mobility of electrons strongly coupled to ripplons (ripplonic polarons) on a thin film of liquid He are calculated. The path-integral technique is utilised. It is found that the temperature dependence of the mobility is of exponential type with the characteristic energy proportional to the coupling constant, and the prefactor is independent of the coupling constant. The conditions for strong coupling are also examined with reference to experiments.

Hall effect and deep-level transient spectroscopy (DLTS) are employed to investigate the electrical properties of Co-doped InP. Co doping is found to give rise predominantly to only one deep level at E_v+0.32 eV in contrast to earlier conclusions. This leads to p-type material with a resistivity in the range 2*10³ to 2*10⁵ Omega cm depending on the Co concentration. Resistivities above 10⁴ Omega cm only arise for closely compensated material. The level at E_c-0.5 eV reported in the authors' earlier investigation is shown to be present in high concentration only near to the surface of material which has undergone 350 degrees C heat treatment during ohmic contact alloying. Similar deep-level production by low-temperature heat treatment is also reported for InP:Fe.

A recoupling procedure is used to show that the spin-correlated crystal field (SCCF) is included in the general two-electron-correlation crystal-field (CCF) parametrisation. A parallel result is that the spin-correlated Coulomb interaction (SCCI) is included in Judd's parametrisation of three-electron correlation in free ions. Experimental results and ab initio calculations show that the SCCI is important but does not dominate three-electron correlation. This suggests that contributions other than the SCCF may also be important in the CCF.

The thermodynamic properties of the two-dimensional Ising model with random (+or-J) nearest-neighbour interactions are determined using a numerical transfer matrix method. The calculations are performed on a long strip of width L(3<or=L<or=11) with periodic boundary conditions and finite-size scaling arguments are used to extrapolate to L= infinity . The central result is that the spin-glass correlation length diverges as T^{- nu} at low temperature T with nu =2.59+or-0.13. The heat capacity is similar to the Schottky anomaly. At zero temperature the entropy per spin is 0.0701+or-0.0005 and the internal energy per spin is -(1.4024+or-0.0012)J.

The equilibrium properties of the two-dimensional Ising model with nearest-neighbour interactions uniformly distributed between -J and J have been determined using a numerical transfer matrix method. The calculations are performed on a long strip of width L(3<or=L<or=11) with periodic boundary conditions and finite-size scaling arguments are used to extrapolate to L= infinity . The spin-glass correlation length diverges as T^{- nu} at low temperature with nu =2.96+or-0.22. The correlation length scaling function differs from that for the random Ising model with interactions of magnitude J and random sign showing that the two models with different distribution functions belong to different universality classes. The heat capacity is linear at low temperature and can be interpreted using a continuous distribution of two-level states which is finite at zero energy and increases approximately linearly with energy. At zero temperature the entropy is zero and the internal energy extrapolated to L= infinity is -(0.7943+or-0.0010)J.

Neutron spectroscopy measurements have confirmed an earlier intimation of an 'anomalous' anisotropy in the dispersion of 'acoustic magnon like' excitations in yttrium iron garnet Y₃Fe₅O₁₂. The present experiments have shown this anisotropy to be large and intricate, with energies at a fixed excitation wavelength 25 Å, varying from 5 to 20 meV. A contention that such anisotropy might arise from strong magnon-phonon coupling effects is not supported by experimental investigation of (i) a spin-reorientation transition in Tb_0.3Y_2.7Fe₅O₁₂ or (ii) applying variously directed magnetic fields to YIG itself. The experimental dispersion curves are replicated instead by linear spin-wave theory, with fitted values for all exchange interactions J_1ad, J_2ad, J_3ad, J_1aa, Δ_1aa, J_2aa, J_3aa, J_1dd, J_2dd, J_3dd, J_4dd, J_5dd closer than a/2∼6.2 Å. In this new model, the experimentally observed 'acoustic magnon' anisotropy is attributed to a complex coexistence of acoustic and low-energy optic magnon branches: ensuing computations suggest that possible 'soft modes' of such 'pseudo-acoustic' branches may relate to previously reported 'cantings' of the ferrimagnetic structure in various substituted garnets. The new model also helps in understanding some reported anomalies in existing magnetisation, specific heat and microwave measurements.

Thermally detected EPR has been used to obtain the spectra of cobaltous ions at concentrations of up to 5% in cadmium chloride, cadmium bromide and magnesium chloride in the range of frequencies 8-40 GHz. These have been interpreted in terms of single ions and four types of pair. The anisotropic exchange parameters in the spin Hamiltonian for these pairs have been determined, but no spectrum attributable to nearest-neighbour pairs has been observed.

The authors have investigated radiative recombination in a-P at low temperatures using optically detected magnetic resonance (ODMR) and standard luminescence techniques. The ODMR results show that two distinct regions can be identified in the photoluminescence (PL) spectrum, corresponding to excitonic and pair emissions. The exciton emission has a maximum near 1.1 eV, and the pair emission band peaks near 1.4 eV in bulk a-P material. The results are compared with the existing PL and time-resolved spectroscopy (TRS) date on a-P, and a consistent model can be formed involving exciton and pair recombination at intrinsic centres. From arguments relating to the exciton-pair energy separation, it is inferred that intimate valence alternation pairs (IVAPS) form suitable candidates for the recombination centres.

Using synchrotron radiation, the intrinsic luminescence excitation spectra of RbI and RbBr were measured between 40 and 130 eV, at low temperatures, and the impurity luminescence at room temperature. Some surprising features correlated with competing decay routes after photoexcitation, which increase in complexity as the exciting energies enlarge. Luminescence can be used to determine core-edge thresholds and core-exciton binding energies.