Table of contents

The authors have studied spin-freezing phenomena along the magnetic easy axis of the insulating spin glass Fe₂TiO₅ by magnetisation, AC susceptibility and neutron scattering experiments. The characteristic measurement time for these techniques varies over more than fourteen orders of magnitude. The results, which show a decrease of the freezing temperature with decreasing characteristic frequency, are analysed in terms of three models.

Antiphase boundaries associated with the charge density wave superlattice in 1T-TiSe₂ have been observed by satellite dark-field transmission electron microscopy. By analogy with 2H-TaSe₂, the antiphase boundaries in TiSe₂ are interpreted as antiphase discommensurations associated with a commensurability of order two.

It is shown numerically that electron tunnelling probability out of a one-dimensional delta potential in an electric field is much larger than standard tunnelling theories predict.

Strong frequency dependence is shown in the quantised magnetoconductance in high-mobility silicon MOSFETs, with the loss of integer quantised plateaux and enhancement of fractional quantisation. Fractional plateaux observed are quantised to within 5%, which is within experimental accuracy. The frequency at which the onset of these phenomena is observed is shown to be dependent upon sample length and Landau level filling factor. The results are similar to those previously obtained using low-mobility GaAs heterostructures, though in the highest mobility cases no effect is seen up to 50 MHz. The roles of sample length and disorder are discussed.

As there has been considerable interest (Jerphagnon et al. 1978, Walpole 1981) in the description of physical properties of condensed matter by invariants and since also the acoustic activity is perhaps the first example of a physical property characterised by a fifth-rank tensor, form-invariant expressions for the acoustic gyrotropic tensor leading to invariant acoustic gyrotropic coefficients for all the non-centrosymmetric crystal classes are reported in this paper.

n-type and undoped GaAs samples have been irradiated with 2 MeV electrons in stages at room temperature up to high electron doses of approximately=3*10¹⁹ electron/cm². At each stage the concentration of paramagnetic defects was monitored by EPR. In the n-type sample a signal previously ascribed to uncomplexed As_Ga antisite defects grew rapidly, reached a maximum, subsequently decreased in strength and then grew again almost linearly. In the undoped samples the signal grew almost linearly over the whole dose range. The authors explain these results by postulating the existence of at least two defects which give rise to indistinguishable EPR signals. Annealing and re-irradiation studies on n-type samples provide evidence that the initial growth of the EPR signal is due to a change in the charge state of the defects which are probably complexes involving As_Ga centres such as (As_Ga-V_Ga) or (V_As-As_Ga-Ga_i) rather than isolated As_Ga defects.

The origins of the dielectric anomalies associated with the Jahn-Teller phase transitions in DyVO₄ and DyAsO₄ are examined in detail. The lowest, nearly degenerate, Dy electronic levels couple strongly to distortions that carry an electric dipole moment parallel to the tetragonal axis. The electric susceptibility as a function of temperature, chi _c, accordingly shows Curie-like behaviour above T_D and a prominent cusp at the phase transition. In the perpendicular direction the electric dipole moments are negligible, and chi _a, chi _b have a temperature dependence similar to that of the order parameter. In sensitive experiments chi _a and chi _b are observed to differ in magnitude and temperature dependence, suggesting a non-linear dependence on order parameter. Analysis shows that the principal contribution to chi _a, chi _b comes from matrix elements of the electric dipole operator between the ground doublets and excited crystal-field levels. These Van Vleck contributions lead to a temperature dependence both linear and quadratic in the order parameter in good agreement with experimental results.

The quadratic Pade method-a new method for calculating the local density of states in various physical systems-is introduced and discussed. The method is based upon the use of Hermite-Pade polynomials and it makes the calculation of densities of states a straightforward and relatively simple matter. Its advantages over other methods with similar generality and complexity are outlined and numerical results for various systems, which illustrate the virtues of the new method, are presented and discussed.

The authors analyse the connection between the mathematical properties of iterated maps and the physical properties of electronic systems. Specifically, they treat a linear disordered chain of one-orbital atoms governed by a tight-binding Hamiltonian. The physical properties of the quantum states are connected to the mathematical properties of the orbits of an iterated map.

Optical joint density of states (OJDOS) functions have been obtained from Kramers-Kronig analysis of reflectivity measurements for the layered-type materials TiS₂, TiSe₂, ZrS₂, ZrSe₂, HfS₂ and HfSe₂. The reflectivity measurements were made at near-normal incidence over the photon energy range 0.6-14 eV at 77K. Comparison of the OJDOS functions shows that there are many similarities in the band shapes which can be explained in terms of the amount of trigonal distortion present in the crystal lattice and the differences in binding energy of electron levels in the atoms.

For pt.I see ibid., vol.16, p.6491 (1983). The electrical conductivity sigma (T,x) of amorphous Si_1-xCr_x films is studied for T=0.1-300K and x=0.09-0.26. The existence of a finite minimum metallic conductivity sigma _mm is shown by two independent methods, which yield 270+or-50 Omega ^-1 cm^-1 and 250+or-30 Omega ^-1 cm^-1 respectively. On both sides of the metal-semiconductor transition (with respect to the conductivity scale) only one parameter depending on x and preparation conditions is needed to describe sigma (T) for low T. In particular, on the semiconducting side a scaling law, sigma = sigma (T/T₀(x, . . .)), is valid. On the metallic side a contribution beta T^1/2 is observed, where beta is completely determined by sigma (0K). The function beta ( sigma (0K)) exhibits a square-root singularity at sigma _mm. The critical Cr concentration x_c is weakly influenced by the preparation conditions; it amounts to about 0.14. At high temperatures a second conduction mechanism becomes important. This mechanism is present on both sides of the transition. A quantitative description of the related conductivity contribution is given for both the metallic and semiconducting regions. Annealing of semiconducting samples at temperatures of about 300 degrees C or higher destroys the scaling behaviour, presumably due to the formation of metallic clusters. In particular, it gives rise to a decrease of the pre-exponential factor of sigma (T).

Geometric resonances in the two-dimensional magnetoplasmon excitation are calculated in the framework of classical kinetic theory for arbitrary values of k nu _F/ omega _c and finite values of the scattering time tau (k=plasmon wavevector, nu _F=Fermi velocity omega _c=(e/cm) B=cyclotron frequency). The geometric resonances lead to an oscillatory behaviour of the magnetoplasmon resonance amplitude and dispersion correlated with subharmonic fields B_c/n (n=2, 3, . . .). Calculations are compared with magnetoplasmon resonance experiments in electron inversion layers of Si and the possibility of geometric resonances in single quantum well systems of GaAs-AlGaAs heterostructures is discussed.

The authors present a systematic formulation of the conduction band Hamiltonian in the presence of a magnetic field by an invariant expansion and by expressing the coefficients of this expansion in terms of the band parameters of an extended Kane model. This Hamiltonian can be used to describe non-parabolicity effects such as dipole-allowed spin-flip and higher-harmonic cyclotron resonance transitions and their angular dependence, and the magnetic-field dependence of the cyclotron mass and g-factor.

The luminescence of Cr³⁺ in borate glasses was investigated by conventional and laser spectroscopic methods. In these glasses emission is observed from two levels (²E and ⁴T₂). From spectral analysis of excitation and absorption the most inefficiently luminescent Cr³⁺ ions are inferred to be primarily low-crustal-field ions. Furthermore, decay plots indicate energy transfer between these levels, i.e. from the ²E to the ⁴T₂ level, which decreases the total luminescence efficiency even further.

The effective Debye temperature Theta _eff for (001) and (111) nickel faces, for directions normal to these faces and for those making an angle of 50-65 degrees with this normal is measured by LEED for different azimuths. The electron energies vary between 70 and 100 eV and 130 and 200 eV for (001) and (111) respectively, and the incidence angle for the primary beam is 50-70 degrees C. Under these conditions the electrons are diffracted mainly in the first layer of atoms. For the specular beam, Theta _eff is averaged over the azimuthal angles in order to eliminate the multiple scattering effects and Theta _{perpendicular to} equal to 228.5 and 222K is obtained for the (001) and (111) faces, respectively. The minimum values of Theta /sub /// calculated from the data for backscattered beams are taken as the closest to the surface Debye temperature for the direction parallel to the surface. These values are equal to 348, 344 and 357K for the (110), (001) and (111) faces, respectively. Rotational LEED diagrams for the specular beam, for an incidence angle close to 60 degrees and an electron energy between 70 and 105 eV, are presented for both the faces investigated.

The perturbative variational method is applied to the surface impurity states in semiconductors. It is found that this method can be used to calculate any state, no matter how high it may be, and the surface impurity levels are obtained from those of the hydrogen atom simply by a redefinition of the charge. As a concrete example of the application numerical results are obtained for surface impurity states in GaAs using this method.