Table of contents

The electron localisation criterion for one impurity atom can be exactly derived for a Cayley tree network of atoms using the tight-binding approximation.

An analysis of the shape of photoconductivity spectral distribution curves of n-MoSe₂ in the range of the direct and indirect fundamental optical transitions is presented, taking surface and volume recombination of the photogenerated charge carriers into consideration. The c-axis hole mobility, mu _h,z, is almost as great as that in the layers, mu _h,xy, at room temperature in accordance with band calculations where the top valence band is three-dimensional, and gives allowance for a low effective hole mass anisotropy. A very large anisotropy of mu _h,xy/ mu _h,z of approximately 10²-10³ at 77K can be qualitatively understood in terms of extrinsic hole scattering from ribbons of stacking faults. The surface recombination rate is enhanced by the presence of (1011) faces on the illuminated surface.

A model of spin glasses containing single-ion uniaxial anisotropy is solved within (replica-symmetric) mean-field theory. For small anisotropy strength D, two phase transitions are predicted as the temperature is reduced. The first transition involves the freezing of the spin component(s) corresponding to the easy axis (axes). At the second transition the remaining spin components freeze. For sufficiently large mod D mod , only the first of these transitions occurs.

First results of an investigation of the magnoms in mixed K₂Co_1-xFe_xF₄ (x<or=0.4) by means of Raman scattering and FIR absorption spectroscopy are presented. Four different magnetic excitations are identified: the Fe gap mode, the Co AFMR, the Co-Fe pair mode and the Co two-magnon mode. In addition the gap mode of an Fe spin in the neighbourhood of three Co²⁺ and one Fe²⁺ ions is observed. Using the exchange and anisotropy parameters of the pure compounds and a Co-Fe exchange parameter derived from the Fe gap mode the authors calculate the frequencies of the other magnetic modes as well as the concentration dependence of the Co two- -magnon mode in a simple cluster model.

Irradiation of GaP with 2 MeV electrons leads to the linear growth of infrared absorption lines at 424 and 419 cm^-1 with a strength ratio of 1.45+or-0.15. These lines were previously attributed to the vibrations of a displaced phosphorus atom in a site of axial symmetry but it is now shown that they can be explained by a single mode of vibration ⁶⁹Ga and ⁷¹Ga respectively. Low-temperature irradiation confirm the authors' earlier data (1971) that the lines do not arise from a primary defect and imply the mobility of displaced interstitial atoms. A comparison is made with the observations of interstitial trapping by substitutional boron impurities.

The results of a package of FORTRAN computer programs for modelling defects in ionic crystals and for fitting experimental data are described. The fundamental concept of the defect simulation is similar to HADES except that the minimisation procedure is different since the package is designed to run on small computers. As an example of the use of this package, the relative stabilities of NN and NNN complexes for various rare earths, Ianthanum and yttrium are considered. First, the data-fitting routine was used to analyse relaxation data for NN and NNN complexes in rare-earth-doped strontium fluoride. The experimental results for strontium fluoride were then used in conjunction with the defect-simulation program to determine potentials for all of the rare earths, yttrium and lanthanum. Those rare-earth potentials were then used in the simulation of calcium fluoride and show that the NNN complex should not be observable except for possibly the smallest rare earths. This implies that the B site of Wright and co-workers or the RII relaxation requires another explanation. Also, the potentials were used in the simulation of barium fluoride, showing that the NN complex should be observable only for the largest rare earths or lanthanum. Next, the enthalpy for NN to NN reorientation via the interstitialcy mechanism was calculated for rare earths in calcium and strontium fluoride. In general, the calculated reorientation enthalpies are larger than the experimental values. However, the variation of the enthalpy with the size of the rare earth is in reasonable agreement with experiment. Finally, the variation of the calculated enthalpy with pressure is found to be in excellent agreement with experiment.

For pt.I see ibid., vol.14, p.3863 (1981). Previous calculations are applied to estimate the contribution relatively small concentrations of dislocations of dislocations make to the diffusion coefficients defined by appropriate gradients of the initial part of conventional tracer penetration plots and usually presumed to represent volume diffusion. A first-order expression is derived for the effective, or dislocation perturbed, diffusion coefficient as a function of dislocation density using an empirical analytic representation of the computed numerical values of mean tracer concentration as a function of penetration distance.

Five models (two regular and three distorted structures) of isolated polysulphur nitride chains are studied by a Hartree-Fock ab initio SCF-FSGO technique. The author finds the Peierls ground state of an isolated (SN)_x chain corresponds to a structure where there is a succession of two longer bonds and two shorter bonds, =(S-N-S=N)=. The Peierls state is coherent with the Hartree-Fock instabilities appearing in the regular models.

For pt.I see ibid., vol.15, p.2613 (1982). The electronic band structures of ferromagnetic chromium spinels CuCr₂Se₄ and CuCr₂Te₄ are calculated self-consistently by using the discrete variational X alpha method. The main features of the band structure are almost common among the three compounds CuCr₂X₄ for X=S, Se and Te: relatively narrow valence bands constructed mainly from the anion p orbitals lie in a lower energy region. Very narrow bands having mainly d character range over the intermediate energy region; a few d bands cross the Fermi level for both spin bands of the selenide and telluride while d bands with the majority spin only cross the Fermi level for the sulphide. Wide conduction bands composed of Cu and Cr 4s, 4p orbitals lie in the higher energy region. The valency of each ion obtained from the self-consistent calculation are Cu^1.25+(Cr^1.41+)₂(Se^1.02-)₄ and Cu^0.85+(Cr^1.03+)₂(Te^0.73-)₄, which are compared with Cu^1.30+(Cr^1.55+)₂(S^1.11-)₄. The atomic spin-polarisations of Cr, Cu and X are +3.54, -0.26 and -0.36 respectively for CuCr₂Se₄, and +3.74, -0.19 and -0.48 for CuCr₂Te₄. (They were +3.20, -0.32 and -0.27 for CuCr₂S₄ by way of comparison.).

The density matrix is introduced to describe a wave in a random medium. The spatial and temporal evolution is analysed in terms of eigenmatrices of the disordered system and a transition in the type of behaviour is found, at a critical value of the disorder above which all wavelength density fluctuations increase exponentially in space or time.

Working within the single-electron picture of two-dimensional disordered systems in the crucial weak-disorder regime, the authors calculate the average two-particle Green function by summing the maximally crossed diagrams occurring in the approximation of Neal and Langer (1966) as developed by Abrahams and Ramakrishnan (1980) to show that Anderson's criterion for location is not satisfied, while there is only a weak absence of diffusion in the sense of Ishii (1973). Summation of ladder diagrams appearing in the Langer-Neal scheme gives an identical result in the small-disorder regime. They reach the same conclusion upon using the CPA to perform the required averaging. Recent mathematical results associate such a weak absence of diffusion with a singularly continuous spectrum.

A theory of light absorption in semiconductors doped by transition metal impurities is constructed. The absorption cross section near the edge of the optical band is calculated by means of k.p perturbation theory within the framework of the Kane model supplemented by the d impurity levels. Special attention is given to the problem of gauge invariance in k.p theory. It is shown that in the case of an impurity state with definite parity the Lucovsky formula is valid when modified by the more detailed account of the central cell corrections. In the substitutional positions without the inversion centre the impurity states possess no definite parity which allows the dipole transitions forbidden by the conventional crystal field theory and causes the linear Stark effect for the t₂ levels.

The electronic conductivity ( sigma _c) of copper selenide in the composition range from Cu₂Se to Cu_1.96Se has been investigated in the temperature interval from room temperature to 430K. Two anomalies found in sigma _c against T curves are associated with a long-range order-disorder transition of the cation subsystem and a rhombohedral-cubic transformation of the cage, respectively. On both sides of the order-disorder transition sigma _c decreases linearly with increasing temperature and the only constant that determines how sigma _c varies with T is the interaction energy (k_BT_c1) of the cation subsystem. Therefore by measuring sigma _c(T) cation disordering can be detected. Also, its temperature dependence may give some relevant information about the disordering mechanism if an appropriate microscopic model that correlates increased disorder with reduction in charge carrier concentration is used.

Nanosecond time-resolved measurements of the current and voltage of an electron-hole plasma generated by avalanche in n-type InSb at 77K under z-pinch conditions are reported. A model assuming cylindrical symmetry and taking into account the field dependence of the generation rate and the electron mobility, various kinds of recombination, electron-hole scattering and circuit effects is presented. The observations are in good agreement with the results calculated from the model. It turns out that the generation recombination processes dominate the behaviour of the pinched plasma. The observed oscillations on current and voltage, which were formerly often connected to sausage instabilities, are shown to originate from density domains moving in radial direction. The authors conclude that a description of semiconductor plasmas in terms of the well-known Bennet model is incorrect, because this omits the generation recombination terms. It is argued that because of de-pinching effects it is generally not possible to determine the values of the recombination parameters from a simple measurement of the conductivity decay of an avalanche plasma.

A theory for spin waves in diluted uniaxial antiferromagnets with random interactions in both exchange and single-ion anisotropy has been developed by extending a recent work on the related ferromagnets. As in the paper for the ferromagnets, the kind of coherent potential approximation (CPA) the authors consider here taking into account the dynamical effects of fluctuations in the exchange and local-anisotropy fields on the same level. For a two-sublattice antiferromagnet, however, the effective medium is self-consistently determined in terms of three parameters which, after using the time-reversal symmetry into the usual CPA T matrix condition for a pair of nearest-neighbour sites, may be reduced to only two independent parameters in the whole energy space. The concentration dependence, at zero temperature, of the zero-field antiferromagnetic resonance (or, equivalently, of the antiferromagnetic-'spin-flop' critical field) is analysed as a function of a local-anisotropy parameter and of the relative size of the anisotropy in the pure magnetic system. For systems with random non-competitive anisotropy, it is found that the general behaviour of these quantities with dilution is very much more sensitive to changes in both the random and non-random parts of the anisotropy than predicted by virtual-crystal approximation.

The temperature renormalisation of the energies of the optic modes in ErFe₂ has been determined from room temperature up to close to the Curie temperature (574K). It is found that the two modes, a dispersive transition-metal mode and a localised crystal-field-dominated mode, cross over at about 420K. The experimental results have been interpreted and are well accounted for by a linear spin wave model, where the level scheme of the lowest J multiplet of the Er³⁺ site has been assumed to consist of pure J_z states with an equidistant energy spacing between the levels.

Diffraction investigations on single crystals of the system Mn_1-xCr_xSb (0<or=x<or=1) are reported. The NiAs-type structure is maintained for the whole composition range. Single crystals were grown by the Bridgman-Stockbarger method. The site occupancies were characterised by combined X-ray and neutron diffraction investigations. By means of neutron diffraction experiments as a function of temperature, a detailed magnetic phase diagram has been established. The uniaxial antiferromagnetic structure of CrSb is preserved up to a 20% substitution of Cr by Mn. In the composition range 0.05<or=x<or=0.80 magnetic orderings are observed which may be understood as mixing the ferromagnetic structures of the MnSb-type and antiferromagnetic arrangements of (Cr_0.80Mn_0.20)Sb. The magnetic behaviour of the intermediate region therefore may be discussed in terms of two superimposed magnetic couplings, the directions of the easy axis adjusting to the temperature.

The nuclear spin relaxation rates in a number of alkali halides have been measured at temperatures above the melting point and in LiBr and LiI were found to be approximately proportional to the reciprocal of the temperature. The rates are explained by the interaction of the nuclear electric quadrupole moment with the field gradient of the surrounding ionic charge but, just as in the case of the quadrupole relaxation in solid alkali halides, the point-ion model for the calculation of the field gradients proved to be untenable. Approximate values for the mean-square gradients obtained from previous data taken in the solid state were used in conjunction with the present measurements to calculate a correlation time. This correlation time, which is the average time that the spatial configuration of a group of near neighbours persists in the liquid state, was found to be about 10^-13 s.

The authors present a discussion of the main features of self-consistency for clean and chemisorbed surfaces in the local density functional formalism. The technical details of the procedure are analysed together with the behaviour of the various contributions to the surface potential. They show that the iterative approach to self-consistency can be made efficient and quick when the starting potential is properly chosen and the components of the potential more sensitive to the self-consistent process are singled out.

Photoemission yield spectroscopy, low-energy electron diffraction and Auger electron spectroscopy measurements are performed on a set of UHV cleaved GaAs (110) surfaces from n- and p-type crystals as a function of Ag coverage theta which is varied from about 10^-4 monolayer (ML) to a few ML. The variations upon interface formation of the atomic structure, the work function phi , the ionisation energy Phi and the semiconductor band bending are analysed together with the changes of the density of filled surface states in the gap and in the valence band, within about 1 eV from the top. At the forming interface, the contribution of different electron states is demonstrated and their possible origins are discussed. Ag is shown to form an abrupt interface and to grow epitaxially as crystallites forming a non-uniform layer.