Table of contents

The critical dynamics of a random uniaxial dipolar ferromagnet is studied for the range of temperatures t_x<<t<<1, where t_x is related to a degeneracy of the renormalisation-group scaling relations at leading order. The behaviour of the characteristic frequency is described by an effective dynamic critical exponent which is parametrised by the concentration of impurities.

Samples of n-type GaP doped with either (a) sulphur or (b) silicon to concentrations of about 10¹⁸ atoms/cm³ have been compared using EPR and IR measurements following a series of small doses of 2 MeV electron irradiation at room temperature. The EPR spectrum due to the (P-P₃) antisite defect is produced up to a similar strength in both materials. The measurements suggest that it is unlikely that the defect involves an impurity. The authors speculate that the (P-P₃) centre may be a gallium vacancy distorted to give (V_p-P_Ga) or a trigonal (V_P-P_Ga-Ga_i) complex produced by a knock-on process.

Using a ZnO transducer, two peaks at 1.250 and 1.212 eV in the photoacoustic spectra of n-InSe were obtained at 300K. The energy and its temperature variation of the 1.250 eV peak are well explained in terms of a free exciton annihilation. Another peak at 1.212 eV is interpreted by a surface related transition. The results suggest the usefulness of the low-temperature photoacoustic studies on the recombination processes in semiconductors.

The lifetime of sticking of the spin-polarised atomic hydrogen interacting with the third sound in the superfluid helium is found to have new exponential dependence on temperature.

For original paper see ibid., vol.17, p.L419 (1984), for comment see ibid., vol.18, p.867 (1984). In spite of N. d'Ambrumenil's objection, it is maintained that R.B. Laughlin's suggestions are not the whole answer to the fractional quantum Hall effect, which has more physics behind it.

For highly anisotropic layered material the influence of weak disorder due to randomly distributed impurities is studied in diagrammatic self-consistent approximation. It is shown that, in contrast to a quasi-one-dimensional case, extended states in an upper part of an energy band immediately appear if a weak interlayer tunnelling is switched on. Intensity of this tunnelling is characterised by the value of interplane exchange integral omega . The borderline energy between extended and localised states is equal to epsilon _c=(h(cross)/ pi tau )ln( square root 2h(cross)/ omega tau ), where tau is the intralayer relaxation time. For epsilon _F< epsilon _c system behaves as a dielectric and for epsilon _F> epsilon _c it behaves as a metal. Here epsilon _F is the Fermi energy. AC conductivity sigma ( omega ) is found for both cases. The localisation length is calculated for epsilon _F<or= epsilon _c.

The authors have investigated the low-temperature characteristics of GaAs FETs manufactured by different companies. All devices showed conductance oscillations and had common features. The minima could be fitted to a law based on periodicity in electron separation.

The influence of quantum fluctuations on the magnon spectrum of helimagnets is investigated. It is shown that quantum effects result in the shift of the helix wavevector value, while the general structure of the spectrum, i.e. the existence of the two Goldstone modes, remains unchanged.

The author studies from general considerations, the critical behaviour of systems which condense, at their critical temperature T_c, into a weakly coherent critical state with an infinitely large number of weakly coherent critical state with an infinitely large number of internal degrees of freedom. If this number (per volume of size xi ^d where xi is the correlation length) varies as xi ^rho with 0< rho <d, the reduced dimension d=d- rho is naturally associated with the extremely large fluctuations at the transition. Assuming that xi approximately (T-T_c)^{- nu} with nu = nu (d) of some regular d-dimensional model, the negative value of the specific-heat exponents alpha = alpha (d)- rho nu follows. He proposes a relation between rho and other critical exponents and compute self-consistently the dimensional reduction. The short-range Ising spin glass below an intermediate critical dimensionality (probably d_icd=4) may provide a realisation of such an entropy-dominated transition. If all frustration effects are accounted for by the effective shift in dimensionality, the following consequences are anticipated: (i) The lower critical dimension is d_lcd=2. (ii) The 3d exponents are nu _SG=1 and alpha _SG=-1. These conjectures agree surprisingly well with several recent numerical simulations.

The recently published Knight shift measurements in metallic Ge:As under saturation (111) stress close to the metal-nonmetal transition are here supplemented by measurements of T₁ and T₂. Significant changes in these parameters are caused by the application of stress. The T₁ value becomes longer by about a factor of four, and the T₂ value, previously recognised at zero stress as anomalously small, becomes longer as the stress is applied. The relaxation time data exhibits behaviour that indicates that, under stress, the electron gas becomes more inhomogeneous. The relevance of these results to the question of whether or not an impurity band separates out under stress is discussed.

The process of the polarisation reversal within the commensurate ferroelectric structure is discussed. The nucleation theory is used to calculate the switching time and the coercive field epsilon _c. It is shown that epsilon _c decreases critically on approaching the transition to the incommensurate phase, which represents one of the first examples for precursor phenomena within the commensurate phase.

The spectral profiles of Pd M_4,5N_4,5N_4,5 transitions of Ag₈₀Pd₂₀ and Cu₉₅Pd₅ are calculated using the Cini-Sawatzky model for a filled band system the multiplet structure obtained from atomic structure calculations and single electron densities of states obtained from the results of SCF-KKR-CPA calculations. Reasonable agreement with experiment is obtained for Ag₈₀Pd₂₀ but not for Cu₉₅Pd₅. Agreement with experiment is improved if the calculated intensities of states at the bottom of the band, associated with bonding states, is arbitrarily reduced and the M₅N_4,5N_4,5:M₄N_4,5N_4,5 intensity ratio is modified to take account of the influence of M₄M₅N_4,5 Coster-Kronig transitions.

The object of this study was to separate the different quasi-elastic components related to the tumbling motion of the C₃ molecular axis and to the uniaxial rotation around the C₃ molecular axis. In agreement with symmetry considerations, quasi-elastic components related to the uniaxial rotation have been observed for Q along the (110) and (111) crystallographic directions. The relaxation time tau ₃ derived from the widths of these components agrees with the incoherent neutron scattering analysis.

The orientationally disordered phase of sulphur hexafluoride at 150K has been studied here using the molecular dynamics simulation technique, with particular emphasis placed on the origin and nature of the orientational disorder. A model involving the frustration of two competing intermolecular interactions that had been proposed previously has been shown to be a good description of the origin of the disorder through calculations that have given information concerning the correlations between the orientations of neighbouring molecules. It has also been shown that the disorder is dynamic rather than static in nature, and features of the molecular motion that are related to this dynamic disorder have been identified. It is suggested that some of the features of orientationally disordered crystals that have been studied might be observed experimentally in a coherent neutron scattering study, and some preliminary calculations of S(Q) are reported.

A theoretical model which accounts for the photoacoustic amplitude and phase generated at first- and second-order phase transitions in solids by using a piezoelectric detection system is proposed. This model is based on a previously published approach to the study of phase transitions by using microphonic detection systems. Measurements of the photoacoustic signal on CsCuCl₃ and (N(CH₃)₄)₂CoCl₄ are presented. The results obtained are discussed in the framework of the theory.

A model of low-frequency relaxation in condensed matter is provided. The model involves time-dependent transition rates which are obtained from time-independent transition rates. The time dependence arises because of an environment that provides a time-dependent entropy contribution to the free energy that controls the transitions. The model has two related consequences. Firstly, relaxations at low frequencies conform in general to a fractional exponential decay function, phi = phi ₀ exp(-(t/ tau _p)^1-n), 0<or=n<1, independent of both the nature of primary relaxation species and the type of condensed matter. This fractional exponential decay property of relaxation has been widely verified. The general occurrence of this behaviour allows a confident determination of the key parameter, n, in the fractional exponential decay function, to be made. A second consequence of the model is the concept that the relaxation time tau _p observed in a particular experiment is related to a direct relaxation time tau ₀ that would occur in a time-independent environment by means of a relation dependent on the same key parameter n determined from the relaxation measurement. The latter relation provides various renormalisation relations appropriate to particular situations.

The authors have performed a molecular dynamics simulation of the superionic conductor Li₃N at 300K and 400K in order to characterise its structure and transport mechanisms. The distinct Li(1) and Li(2) sublattices are observed experimentally to give complicated temperature-dependent anisotropic mobilities. These have been completely reproduced by the simulation at a qualitative level, although simulated transport coefficients are too high. The defect structure is found to consist of low levels of Frenkel disorder on the Li(2) sublattice and at the higher temperature interstitials are identified in the interplanar region. In accord with observation, Li(2) ions are found to have a remarkable degree of site-localisation.

For pt.I see ibid., vol.17, p.6623 (1984). The authors analyse the trajectories of lithium ions obtained by molecular dynamics simulation of the superionic conductor lithium nitride. Intermediate scattering functions are calculated for scattering vector parallel and perpendicular to the c axis. Detailed transport mechanisms are obtained by direct analysis of trajectories. Highly correlated migrations of a type unknown in normal ionic conductors are observed in the Li₂N plane. At high temperature, the interplanar interstitials play a dominant role in conduction both parallel and perpendicular to the c axis.

In order to carry out calculations of electronic properties efficiently using localised basis functions it is desirable to have a strictly orthonormal basis system, at least to the extent of the range considered. At the same time these functions should mutually transform under a given crystal symmetry. Furthermore they should contain as few expansion terms as possible. The authors propose an orthonormalisation procedure which meets all these requirements. Its usefulness has recently been demonstrated in the calculation of the density response and the phonon dispersion of Si (Falter et al., 1984).

A theoretical approach is developed for the local-environment effects on the electronic spectra of an impurity in weakly disordered alloys (WDA). It is obtained from an embedded-cluster technique using suitable approximations for the WDA. This approach is illustrated using a simple density of states shape for the host alloy and a description of the impurity centre using one state per atom. The simplicity of the formalism involved in this approach makes it a good candidate for applications to real alloys, as shown by the application to the determination of nitrogen-bound electron levels in Ga_1-xIn_xP:N.

A new variational procedure is proposed for calculating the ground-state energy of mixed-valence (MV) semiconductors such as SmS and SmB₆. This procedure uses the localised basis involving one- and two-electron states of a dielectric Anderson lattice with electron-hole attraction. The mixed-valence state can be obtained only beyond the Hartree-Fock approximation. The non-orthogonality catastrophe arising due to the overlap of the exciton wavefunctions is solved by means of a modified Lowdin-Carr method. Some physical consequences of the model accepted are discussed; it is shown in particular that the model resolves the controversy between the semiconducting behaviour of the static dielectric permittivity and its metallic high-frequency properties responsible for the golden colour of SmS in the high-pressure phase.

The self-energy of an electron liquid is studied by including the infinite series of particle-hole and particle-particle ladder interactions beyond the RPA. The effects of these two local-field corrections on the quasiparticle dispersion, the chemical potential mu , the quasiparticle renormalisation factor z_F and the effective-mass ratio m*/m have been numerically examined in the metallic density region.

Deep-level defects related to silver and iron in n- and p-type silicon have been investigated using deep-level transient capacitance spectroscopy. In n-type silicon, a silver-related electron trapping state at E_c-0.54 eV was observed, whilst in p-type silicon a hole-trapping state at E_v+0.29 eV was dominant, in agreement with values previously assigned to acceptor and donor states respectively. In iron-diffused silicon, a hole-trapping level at E_v+0.32 eV was observed in the p-type samples. All of these defects were passivated by reaction with atomic hydrogen, and concentration profiles for the two silver-related centres were measured as functions of the temperature of the exposure to the hydrogen plasma. Gamma irradiation produced an additional silver-related hole-trapping level at E_v+0.48 eV, and an iron-related hole-trapping level at E_v+0.39 eV in p-type samples. Both of these centres have been observed in doped samples quenched from 1175 degrees C. Finally, the room-temperature motion of the silver-related hole-trapping (donor) centre (E_v+0.29 eV) under the influence of the electric field in reverse-biased junction diodes is reported. An estimate for the mobility of 2*10^-14 cm² V^-1 s^-1 at 25 degrees C was obtained for the centres associated with this level.

Wave propagation in 1D systems is extremely susceptible to disorder. A well known result is the localisation of waves which leads to an exponential dependence of the transmission coefficient (or resistance in electrical terms) on the length. In this paper the authors show that the disorder also restrict the bandwidth available when the system is used for signalling purposes. The strength of the effect depends on the group velocity and can be expected to be small for light waves, sometimes important for electrical conductors in fast switching operations and crucial when acoustic waves are involved.

In the impurity system V²⁺:MgO, a d³ ion replaces a Mg²⁺ ion at an octahedral site. The resulting Zeeman spectra have been a puzzle for many years even with techniques employing high magnetic fields and low temperatures. In this paper, the system is treated as a Jahn-Teller ⁴T_2g*( epsilon ₁+ epsilon ₂) system. In this treatment the triply degenerate electronic state designated by the group theoretical irreducible representation T couples to a spin state of J=³/₂ to form spin-orbit states which couple through the electron-phonon interaction to doubly degenerate distortions of the lattice denoted epsilon which break the electronic degeneracy. While the electronic states can couple to the entire phonon spectrum of epsilon -distortions, the cluster approximation is usually employed which approximates by coupling to one suitably chosen frequency mode. In ⁴T_2g*( epsilon ₁+ epsilon ₂) coupling is taken to two different modes epsilon ₁ and epsilon ₂ whose frequencies and coupling strengths are varied to probe the phonon density of states. This multimode treatment is found to yield better agreement with experimental results due to the sensitivity of the intensity quenching to a weakly coupled low-frequency mode.

A detailed experimental study in the near infrared has been made of the low-lying excited states of Dy³⁺ in the ethyl sulphate lattice. The results, which were obtained using different forms of polarised light and the technique of magnetic circular dichroism, show significant differences from the earlier work on this system. Theoretical calculations of energy levels and g-values that include the effects of intermediate coupling and J-mixing have been made. A very satisfactory agreement between theory and experiment is found for the lower ⁶H_11/2 manifold when using published crystal-field parameters derived to fit the experimental data on the ground state. Making some changes in these parameters yields an even closer fit.

The hot-electron kinetic equations recently introduced by Herbert and Till (1981, 1982, 1983) are analysed for silicon using relaxation approximations. The results yield a threshold field for strong electron heating and under the approximation of neglecting optic phonon absorption allow a simple solution for the monolithic hot-electron transistor structure described by Shannon (1981, 1983). The results show how the effective scattering rates for quasi-ballistic electrons at high energy can be reduced from the normal zero-field values, thus allowing large transport factors. The transport factors are approximately independent of the detailed field distribution in the base and collector, yielding a new overshoot effect.

A quantitative treatment of the non-ohmic current response due to delocalisation of shallow localised electrons in a model of a disordered solid is given. Phonon-assisted tunnelling in an electric field from shallow localised to extended states is confirmed as a mechanism leading to the dependence which was experimentally found for chalcogenide glasses.

In n-Si dependence of the current density j and the transverse electric field E_{perpendicular to} on the applied electric field strength E_x is investigated theoretically and experimentally in the presence of a magnetic induction B perpendicular to the current, which can lead to a multivalued electron distribution between the valleys (MED) because the total field E (not only E_x) heats the carriers (this is related to the fact that partial transverse currents of electrons in different valleys exist regardless of the absence of the transverse current). For a homogeneous distribution of j and E numerical calculations deliver S-type j-B characteristics in the region of MED, which in turn, on account of the instability of a state with negative dj/DB, leads to an inhomogeneous current distribution. A layer of high current density, in which the electrons are repopulated into the valleys with high mobility along E_x, is built up near one side of the sample. With increasing B the wall parallel to the current between the layers with high and low current density quickly moves to the opposite surface. This is accompanied by a jump-like change of the transverse field and a strong increase of the current.

The first experimental results of photoconductivity measurements in a-Si:H using in-phase and quadrature frequency-resolved spectroscopy (FRS) are reported. The main feature of the response curves of quadrature FRS is a double peak structure. The relative intensities of the peaks are strongly dependent on external parameters such as temperature and generation rate and on the Staebler-Wronski state, indicating the existence of two recombination mechanisms with different characteristic times.

The DC electrical conductance G has been measured in the approximate range 2-300K for crystalline samples of V_nO_2n-1 with 4<or=n<or=8. A novel feature of the experimental results is the effect of surface treatment, which can profoundly affect G(T) in the range below the metal-insulator transition. The authors show that G(T) below a transition can be described by the small-polaron model but with a vibrating potential barrier.

The reaction-field approximation is applied to the d-dimensional q-state quantum Potts model in an external transverse field. With the exception of d<or approximately=3, the tricritical value q_c(d), below which the system has a second-order phase transition, decreases with increasing external field strength. From the ground state of the quantum d-dimensional model, an improved value of q_c is obtained for the classical model in d+1 dimensions. A direct explanation is also obtained for the expected tricritical point in finite-thickness slabs of cubic ferromagnets (like DyAl₂) with a field in the (111) direction.

The self-energy of spin waves in classical ferromagnetic chains in the presence of a magnetic field is investigated by a non-perturbative Green function approach which includes the leading effects of magnon bound states. The latter are found to be essential for the elimination of unphysical divergences of the usual perturbative result for the spin wave damping. The qualitative differences with respect to the quantum case introduced by the assumption of classical statistics are discussed. The authors present qualitative results for the damping at several fields and temperatures and indicate how these predictions can be tested by computer simulation experiments.

The specific Hall resistivity rho _H of MnAs_1-xP_x crystals (x=0.1, 0.123) has been measured at stationary conditions for temperatures between 50 and 250K using the double-AC method with B_peak<or=0.4 T. The combination of experimentally determined magnetisation values with the isothermal curves rho _H(B) allowed for the separation of the normal and anomalous Hall constants R₀ and R_S, which show different anomalies at the phase transition temperatures T_N, T_l and T_k. The dependences on R₀(T) are interpreted in terms of a two-band model leading to the concept of charge-carrier localisation at T_k i.e. at the low-temperature magnetic phase transition beta to or from gamma .

Ferromagnetic resonance of quasi-two-dimensional K₂CuF₄ Has been studied for various frequencies between 10 and 100 GHz (external magnetic fields between 0.5 and 3.5T) at various temperatures up to 30K. For 10 and 35 GHz, the measurements were performed with conventional microwave techniques and the spectra exhibited structures which are analysed as originating from magnetostatic modes. For higher frequencies, a quasi-optical method has been used and here the structures in the absorption are explained as propagation effects and ferromagnetic polaritons. From the ferromagnetic resonance data, the authors obtain the temperature and field dependence of the anisotropy field H_A and of the saturation magnetisation M₀ or of the combination A=H_A+M₀, respectively. The temperature dependence of the latter follows a A(T)=A(0)(1-CT³2/) law with decreasing values of C for increasing external field, i.e. increasing resonance frequency. The temperature dependence of the magnetisation and the anisotropy field alone has been derived mainly on the basis of theoretical considerations.

Samples of the quasi-two-dimensional ferromagnet K₂CuF₄ with a thickness comparable with the radiation wavelength have been studied by magnetic resonance in the millimetre wave range. The observed transmission and reflection spectra can be explained only on the basis of electromagnetic propagation and magnetic polariton effects and thus provide experimental evidence for these quasiparticles of mixed magnon photon character. In particular, the transmission spectra clearly show the 'polariton gap' between the resonance and the anti-resonance field while the reflection spectra exhibit the 'restsrahlen band' well known in the case of phonon-polaritons. The analysis of the experimental data is performed not only in a qualitative way but also on in a quantitative way, namely the authors calculate theoretically the power transmission or power reflectivity of the sample and achieve good agreement with the experimental data. In this calculation, they have to use the Fresnel coefficient for normal incidence in a form modified for the case where the permeability and not the dielectric constant causes the spectral features and where the magnetic field of the electromagnetic wave (and not the electric field) is considered the basic quantity.

The relaxation behaviour of single-domain magnetic particles is investigated. The first non-vanishing eigenvalue of an appropriate Fokker-Planck equation, characterising the dynamical response as well as the decay of the metastable states of the system, is calculated using a variational approach. The lower bound to the eigenvalue is also computed. The present treatment yields estimates for the eigenvalues that are better than those obtained by the Kramers method.

An investigation of the origin of quasielastic light scattering from silicon surfaces, first observed by Sandercock (1972, 1978) is described. Measurements have been made of the dependence of the quasielastic spectrum upon temperature, wavevector, doping level, and polarisation and intensity of incident light. The results rule out the possibilities of scattering by free electrons or by entropy fluctuations in a collision-dominated phonon regime. A simplified calculation suggests that two-phonon difference scattering may be the origin of the observed spectrum with the main contribution coming from pairs of optic phonons on the same dispersion branch.

The authors present the results of a spectroscopic investigation of the luminescence at 0.74 eV from GaAs:V. They are very well interpreted in the framework of a ³T₂ to ³A₂ transition at a V³⁺(3d²) ion. This type of model was introduced for the first time by Skolnick and co-workers (1983) in their study of InP:V.

The authors have measured the effect of temperature and hydrostatic pressure on the far-infrared plasma edge of an epitaxial film of indium-doped Pb_0.79Sn_0.21Te. At temperatures below 15K at all pressures the sample displayed the large increase in carrier plasma frequency characteristic of persistent photoconductivity. Application of pressure to 8 kbar at 80K, however, is not equivalent to lowering the temperature to change the band gap by the same amount. Instead, the application of pressure causes a relatively small shift in the plasma frequency at both high and low temperature. This demonstrates that the persistent photoconductivity is unrelated to the indium level entering the band. An alternative picture involving a configurational potential barrier between the impurity and band states explains the qualitative features of the results.

LEED intensity measurements are reported for pure Ni(100), and for one monolayer, four and nine layers of Cu on Ni(100). The intensities of the (00), (10), (11), (20) and (22) sets of spots were measured using a spot photometer. The largest variation in the intensity pattern with coverage was observed for the (00) spot. Dynamical calculations were performed for the specular and non-specular beams. It is found that the clean Ni(100) surface has the same interlayer distance as bulk nickel, while one monolayer of Cu on Ni(100) is found to be expanded by about 1.7%.