Table of contents

Photo- and Auger electron spectra have been measured simultaneously from vapour and solid metal for both zinc and cadmium by using Al K alpha X-ray excitation. The differences in the shifts of Auger and photoelectron spectra have been determined with essentially improved experimental accuracy from directly observed vapour-metal shifts.

Transition metal impurities act as deep trapping centres in III-V and II-VI compounds and give rise to near-infrared absorption and luminescence. In many cases the charge state of the ion participating in these optical transitions is either unknown or the subject of some controversy. Luminescence studies of GaP:Ni are described. Measurements of polarised spectra under application of uniaxial stress support the assignment of the charge state Ni⁺ (d⁹).

The first measurements of total light decay of photoluminescence (PL) of a chalcogenide glass in the time regime >100 ns are reported. The time decay is described by t^-m exp(- nu ₂t) where m=1- delta +T/T₀, at temperature T with nu ₂^-1=1.4 ms, delta =0.1, T₀=300K. These results allow a microscopic description of PL in a-As₂S₃ and indicate that PL may arise from charge-transfer de-excitation of defect pairs.

Presents an account of recent studies on the structure of amorphous semiconductors, oxide and chalcogenide glasses and amorphous metals. Emphasis is placed on simple solids which are representative of each class of amorphous material and also on so-called direct structural techniques-for example, X-ray and neutron scattering EXAFS and electron microscopy. For most glasses the arrangement of near-neighbours is similar to that found in at least one crystalline form of the material. Local structure in the so-called transition metal-metalloid glasses can be described in similar terms or by models with random atomic packing and the arguments for an against each hypothesis are presented. Definition of the medium range (8-15 AA) structure is necessarily more vague.

A theory for the memory function of the velocity autocorrelation function for a monatomic liquid is presented, based on a general kinetic approach. Strong emphasis is put on the coupling of the motion of a single particle to the collective motions in its surrounding, splitting the memory function into an essentially binary collision part and a more collective tail, including recollisions to all orders. Numerical comparisons with molecular dynamics data on liquid rubidium have shown quite remarkable agreement.

The M-centre statistical production process in NaCl has been studied as a function of sample temperature, dose rate and impurity content. For a given sample the process is found to be independent of both temperature and dose rate over the temperature range 15 to 95 degrees C and for dose rates of the order of 10⁴ Rh^-1. However, a clear inverse dependence on sample impurity content is observed. The results obtained provide further evidence that F-centre production during the first colouring stage is localised within the crystal lattice around impurities.

Spin relaxation is considered in superfluid ³He-B at low temperature. The relaxation-time result is shown to be the exact solution of the linearised Boltzmann equation. The spin relaxation time is calculated and evaluated with the s- and p-wave approximation for the scattering amplitude. The agreement with experiment is satisfactory.

A momentum-space formalism for calculating the total energy of solids is derived. This formalism is designed particularly for application with the self-consistent pseudopotential method. In the present formalism, the total energy is obtained through band-structure calculations without involving additional integrations. The Hellman-Feynman theorem is derived, as is a modified virial relation for the pseudopotential Hamiltonian which provides an alternative way of calculating forces and total energies.

A semiempirical method of crystalline simulation, the effective supersystem method (ESS), is applied for the first time to the covalent crystals, e.g. diamond and silicon. The geometrical parameters and the bulk moduli are found with reasonable accuracy and the experimentally determined positions of the impurity levels in the band gap agree well with theory.

A simple one-dimensional model of a diatomic chain is used to explain the unusually large values of the transverse effective charge (Z_T*) in cubic IV-VI compounds. By using a pseudopotential formalism, the authors have related Z_T* to the chemical electronegativity and found good agreement with experiment in both the magnitude of Z_T* and in its trend along the series.

For pt.I see ibid., vol.12, no.21, p.4431, 1979. In order to explain the unusually large values of the transverse effective charge (Z_T*) in cubic IV-VI compounds, Z_T* was calculated using pseudopotential parameters derived from detailed three-dimensional band structure calculations; the results agree with experimental values to within 20%. This new technique is extremely simple and is not confined to IV-VI compounds. It requires only a knowledge of the pseudopotential form factors, without the calculation of wavefunctions, matrix elements, or band energies, which are cancelled out by the f sum rule. The transfer for IV-VI compounds validate the conclusions of the earlier simplified one-dimensional model.

The electronic part of the dielectric constant epsilon _infinity is calculated using a simple analytical model. The technique involves the use of the f sum rule to evaluate most of the detailed band structure information, and is closely related to the Penn model of an isotropic semiconductor in the use of this method, although the present technique is quite valid for the strongly anisotropic IV-VI materials. Local fields are included through the off-diagonal elements of the dielectric matrix epsilon (G, G') and are shown to be important. The results reproduce the experimental trends of epsilon _infinity down the series and are in good agreement with experiment.

The effective static dielectric constant of an electron plasma in the simultaneous presence of an electromagnetic wave and a uniform DC magnetic field is discussed. It is found that as the radiation field frequency approaches the plasma frequency a breakdown in screening occurs. As the radiation frequency approaches the electron cyclotron frequency, however, the potential of a static charge felt by the electrons becomes vanishingly small.

Experimental results of the thermal quenching of Co²⁺ intraimpurity ⁴T₂ to ⁴A₂ infrared luminescence in wide-gap II-VI semiconductors are presented. Several theoretical models are discussed and some arguments are given that the observed phenomenon is due to the position of the Co²⁺ ground-state level in the forbidden gap of a II-VI semiconductor. Values of ionisation energies of Co²⁺ ions are determined.

It is shown that the cubic linear pseudo-Jahn-Teller model (s(+)p)(X)( alpha _1g(+) tau _1u (+) epsilon _g(+) tau _2g) is invariant under SO(4) if all modes are equally coupled and are of equal frequency and if the two electronic states are degenerate. On relaxing these restrictions, except that the epsilon _g and tau _2g modes must be equivalent, SO(3) invariance is found. The possibility of applying the models to F-centres in the alkali halides and to the CaO:F⁺ centre is discussed.

The SO(4)-invariant pseudo-Jahn-Teller system (s(+)p)(X)( alpha _1g(+) tau _1u(+) epsilon _g(+) tau _2g) in equal coupling and with electronic degeneracy is examined. Phonon states with definite SO(4) symmetry are defined in terms of the group chain U(9) contains/implies SO(9) contained in/implied by SO(4) and a method for obtaining matrix elements of phonon-creation operators is derived. It is shown how a quantitative calculation might be performed and a relation between ground-state reduction factors is found. The possibility of applying the model to F-centres in the alkali halides and to the CaO:F⁺ centre is discussed.

A perturbed angular correlation study of the nuclear quadrupole interaction at Ta in the commensurate charge density wave (CDW) state of 1T-TaS₂ yields two inequivalent lattice sites with a population ratio of 0.86(9):1. The electric-field gradients (EFG's) for the low- and high-frequency site differ by 11.6(7)% and show different asymmetry parameters of eta =0.12(3) and eta =0.19(2), respectively. Both the EFG difference and the asymmetry allow the identification of sites within the star shaped cluster model of Brouwer and Jellinek (1976). In the quasicommensurate CDW state both the asymmetry parameters and the EFG difference remain constant whereas the population ratio and the EFG distribution vary, especially at high temperature. The disintegration of star-shaped clusters could be responsible for these observations. In the incommensurate CDW state a unique site with a 4.8(3)% EFG distribution is observed.

The ground states of the Ising spin glass with random +or-J bonds on a square lattice have been studied numerically. The distribution of effective fields is obtained as a function of the concentration of negative bonds, together with the proportion of 'living' bonds (i.e. bonds which are frustrated in some ground states only). The comparison of different ground states shows the existence of large packets of solitary spins which keep the same relative orientation in all cases. For the (18*18) samples, the largest packet contains on average more than half the spins, even for maximum bond disorder. The results reveal a striking rigidity of the system at zero temperature, which is very different from the behavior of simple paramagnets and fully frustrated models.

Bond- and site-diluted spin-1/2 Heisenberg and Ising ferro- and antiferromagnets in one, two and three dimensions are treated by a real-space scaling method. Critical parameters (pure transition temperature, where non-zero, and percolation concentration) and critical exponents are evaluated. The critical curves of the simple cubic spin-¹/₂ diluted Heisenberg and Ising systems are also obtained and agree very well with experimental results for KMn_pMg_1-pF₃ and Co_pZn_1-pCs₃Cl₅.

A systematic cluster expansion is applied to dilute ferromagnets with quenched site disorder. The influence of both a spatially homogeneous and a spatially inhomogeneous effective field on the static properties of a ferromagnet are studied. Furthermore, these properties are compared in the one-, two- and three-site approximations in order to test the convergence of this expansion. Numerical calculations are presented for a simple cubic lattice with spin s=¹/₂.

The dependence of c₆₆ in ferromagnetic Tb on a magnetic field applied along a hard axis in the basal plane has been studied experimentally. It has been predicted that c₆₆ should vanish at the second-order transition from a phase where the magnetisation is parallel to the field to a phase which has a finite moment perpendicular to the field. Although it has not been possible to detect the ultrasonic waves in the close neighbourhood of the transition, c₆₆ is found to be substantially reduced at fields slightly higher than the critical value. The observed behavior of c₆₆ can be described quite accurately, at all temperatures below the Curie point, by a random-phase theory without the use of any free parameters.

The hyperfine structure parameters for the trigonal ²³⁵U⁴⁺ centres in fluorite are determined. For centre I, g/sub ///^app=4.03+or-0.01 and A^app=(107+or-0.5)*10^-4 cm^-1 and for centre II, g/sub ///^app=5.60+or-0.02 and A^app=(234.5+or-1)*10^-4 cm^-1. On the basis of the superhyperfine structure attributable to the F^- ligands, a model with six F^- ions and two O^2- ions on the same body diagonal is proposed for centre II.

The niobium spin-lattice relaxation time T₁ and Knight shift K in Ti_1-xNb_xH_1.94 (0.05<or=x<or=0.65) were measured as a function of temperature and niobium concentration. For analysis, the customary scheme of electronic band structure for transition metals in a tight-binding model was assumed. In this approximation the presented data together with values of the magnetic susceptibility have been partitioned into spin (s and d) and orbital components. The latter was found to be strongly dependent upon the composition. It decreases with the increase of niobium concentration and becomes near zero for x=0.65. The d-band density of states decreases also with the increase of x, however, with moderate rate. The importance of these results for proton NMR and low-temperature specific heat data in the studied system is examined.

The series of compounds CuCr_2-xFe_xO₄ with x=0.02, 0.04, 0.10 and 0.20 has been studied by Mossbauer spectroscopy between 77 and 873K. In all samples, Fe²⁺ and Fe³⁺ ions are observed. The occurrence of Fe²⁺ ions, the variation in the quadrupole splitting and the isomer shift with temperature are discussed in terms of a band model for CuCr₂O₄ with exchange of an electron occurring between two Fe³⁺ ions through the conduction band and a relaxation of the Jahn-Teller distortion between the three space directions. The superposition of these two processes implies phonon-electron interactions. The Neel temperature, deduced from the Mossbauer spectra does not vary significantly with the iron content and a value of 130K is obtained, in agreement with the value of 135K deduced from magnetic measurements on CuCr₂O₄. As expected, the hyperfine field of the Fe³⁺ ions follows the Brillouin curve for S=⁵/₂.

Light scattering from the magnetic excitations in the mixed antiferromagnet KMn_1-xNi_xF₃ has been studied experimentally. Measurements have been made on four samples with X=0.84, 0.73, 0.54 and 0.15. At low temperatures the spectra show three two-magnon bands corresponding to the creation of excitations on neighbouring pairs of Ni ions, neighbouring Ni and Mn ions and neighbouring Mn ions. The frequencies of these bands are in excellent agreement with the predictions of an average Ising cluster model, provided that allowance is made for the change in the exchange constants with lattice parameter. The resulting exchange constants are then in agreement with those deduced from other measurements. The intensity of the Mn-Mn band is much weaker than that of the Ni-Mn band than would be expected on the basis of simple theory. It is suggested that this arises because the incident light is nearly resonance with some electric transitions of the Ni ions so that the scattering occurs very largely in the neighbourhood of the Ni ions. The temperature dependence of the magnetic bands was also measured.

A new study has been made of the defect-induced one-phonon infrared absorption in a selected type-Ia diamond. Particular attention has been paid to the wavelength calibration of the spectrometer, and the frequency of a band near the one-phonon cut-off is put at 1331.8+or-0.5 cm^-1 which is the same as the frequency ( omega _R) of the first-order Raman line (1332.5+or-0.5 cm^-1) within experimental error. Other related measurements on diamond and silicon are reviewed, and it is concluded that a peak occurs in the one-phonon density of states of diamond at, or very near, omega _R, supporting the proposition that the maximum lattice frequency is greater than omega _R. This is in agreement with previous speculations and calculations. Such a peak could explain the 'anomalous' band in the second-order Raman spectrum of diamond, although the fact that the frequency of the latter is slightly greater than 2 omega _R remains to accounted for.

Raman spectra of KBr(F) found by excitation of the low-energy side of the F band can be well interpreted by the off-resonance approximation when the '1s-2p'-like transition of the F electron is taken into consideration. A 58% softening of the force constant between the F-centre and its nearest neighbours affects the positions of the two Gamma ₁⁺ resonant modes at 110 and 121 cm^-1. The relative values of the electron-phonon coupling parameters have been determined in a phenomenological manner by considering the vibrations of the first five shells of ions neighbouring the F-centre. When the laser line is tuned from the F to the K band, an enhancement of the LO modes is observed in the Raman spectra relative to the vibrations of Gamma ₁⁺ symmetry. The F-centre Raman scattering theory developed by Robbins and Page (1977, 1978) accounts for the change in lineshapes of the first-order Raman spectra when the interference effect between the F and K band contributions to the Raman scattering is taken into consideration, but it cannot justify the relative values of the electron-phonon coupling parameters.

Calculations are performed on several aspects of the luminescence of pure CsI and CsI:Na. These include electronic-structure calculations by both pseudopotential and semi-empirical molecular-orbital methods, as well as lattice-configuration studies. The results suggest that the main observed emission in CsI:Na at 2.95 eV involves the recombination of a self-trapped exciton immediately adjacent to the substitutional Na impurity.

A technique involving the laser-induced narrowing of a fluorescence line is used to obtain high-resolution spectra of Pr³⁺:LaCl₃. The second-order effects caused by Zeeman and nuclear interactions are systematically investigated in the ³H₄E₂ to or from ¹D₂A₁ transition. A non-zero perpendicular Zeeman splitting of the ground level, proportional to the square of the magnetic field, is measured and interpreted as being caused by the perturbation by the four upper Stark levels belonging to the ³H₄ multiplet. The variation of the number and of the relative intensities of the hyperfine lines is demonstrated to be closely related to the angle between the external magnetic field direction and the crystal axis. It is unambiguously attributed to an angular dependence, with regard to the magnetic field direction, of the hyperfine states of the excited level. The importance of the effect, caused by a very distant perturbing level, is shown to be a result of the quasidegeneracy of the mu =0 excited level.

Analysis of the infrared spectra has led to the determination of the mechanism of fixation of the furan molecules adsorbed on NaCl and their orientation on the crystal surface. No modification of the infrared spectra has been found at the temperatures of the liquid-solid or solid I-solid II phase transitions. By increasing the pressure of furan vapour in the cell, the formation of a new condensed phase (NPC) showing the same structure as the adsorbed phase is observed. When the adsorbed and solid II phases are simultaneously obtained, the latter is spontaneously transformed into the NPC phase. It is deduced that the interaction energy between the molecules of the NPC phase is more important than the lattice energy of solid II.

The strong interaction of a charged particle with the surface excitations in a semi-infinite medium is discussed. Only the example of a particle outside the material and trapped near the surface by its image potential is considered. By adapting a simple unitary transformation formalism, presented for the Feynman polaron model to the present problem, the ground-state energy of a charged particle-surface excitation system is determined. The coupling of a particle to the bulk excitation modes is not treated here.