Table of contents

The acoustic polaron via the deformation potential under the influence of a Coulombic impurity is investigated within the modified variational scheme of the Lee-Low-Pines theory which is valid for the whole range of the electron-phonon coupling. The phase diagram obtained shows how the polaron changes its state according to the values of the electron-phonon coupling constant and of the Coulombic impurity potential.

Polystyrene spheres (diameter approximately 3000 AA) trapped at the air-water interface and interacting via dipole-dipole repulsion have been seen to form a two-dimensional triangular lattice. The authors report the first molecular dynamics study of such an interfacial colloidal crystal. It is shown that the system undergoes a first-order melting transition in the region Gamma =59-65, where Gamma is the ratio of the average potential to kinetic energy. The transition entropy per particle is 0.3 k_B.

The magnetic susceptibility of YIG near the critical temperature T_c has been examined as a function of frequency and temperature. A diagrammatic technique has been used to construct a master curve of response as a function of frequency which reveals the existence of dynamic scaling in the temperature range available. The scaling is of the form appropriate for a one-component system in which there are no conserved quantities.

An expression for the integrated density of states of a liquid simple metal is presented in the Green function theory in the self-consistent (i.e. renormalised) second-order perturbation theory (SCSOPT) by taking the energy dependence of the pseudopotentials into account. It is much more advantageous than the previous expression. It is shown also that the expression, when considered to be a functional of the Green function, is stationary at the point where the Green function is given by the SCSOPT.

It is shown that the scaling theory of Abrahams et al. (1980) while giving correctly the dependence of the conductivity on specimen size or inelastic diffusion length, cannot be applied in three-dimensional systems to its behaviour when the Fermi energy lies near a mobility edge. The existence or otherwise of a minimum metallic conductivity has therefore to be examined by other methods, and a summary is given of our view of the present position. The scaling theory in two dimensions is also discussed; at zero temperature the conductivity in the limit of low field goes to zero with increasing specimen size, and in addition the electric field itself produces a cut-off length.

Calculations are reported of the static and dynamic properties of an effective Heisenberg model of Fe with exchange interactions extending to fifth-nearest neighbours. Exchange parameters chosen to fit the spin wave dispersion curve at room temperature lead to little short-range order above T_c and to a neutron scattering function S(q, omega ) which is compatible with Lynn's 'constant omega ' plots. The dispersion curve obtained by plotting the positions of peaks in 'constant omega ' plots is in excellent agreement with the observed one but it does not correspond to propagating spin waves above T_c. The authors' results do not, however, reproduce the observed rapid drop of intensity at low frequency in 'constant q' scans.

A generalisation of the continuous time random walk (CTRW) approach for a system consisting of inequivalent states is presented. The new CTRW equation is found to bear the same structure as that of the CTRW equation for equivalent states, but the parameters have modified forms and meanings. In the earlier approaches the hopping time distribution function was taken to be same for all states, whereas in our generalisation this function is different for different states. Because of the nature of the distribution function, the authors have introduced the probability that the carrier can remain unmoved on a state due to the particular nature of that state. They have generalised the CTRW approach in the situations: (a) where the hopping time distribution is a function of both space and time; and (b) in which the space and time parts of the hopping time distribution function are uncorrelated, i.e. the decoupled case. In the latter situation, the time part of the function is not rigorously space independent in the sense that it depends upon the nature of the state on which the carrier is situated at an instant.

The structure of molten ZnCl₂ has been investigated by applying the technique of neutron diffraction to isotopically enriched samples. The three partial structure factors relating to Zn-Zn, Cl-Cl and Zn-Cl correlations have been successfully extracted from the experimental data. It is concluded that the correct structural description of the melt is in terms of Zn²⁺ and Cl^- ions rather than the range of complex ionic species which have earlier been proposed. It is further shown that specific chemical bonding effects associated with the Zn-Cl interaction make a very minor contribution to the structure.

The authors present results of calculations of the low-q behaviour of the liquid structure factor S(q) of the alkali metals for a wide range of thermodynamic states. Assuming that the interionic forces can be modelled by the effective pairwise potential arising in second-order pseudopotential theory, they have calculated S(q) for small q using two different random-phase approximations. The first employs a hard-sphere reference system and the second uses the one-component plasma as the reference system. The second scheme enables them to evaluate S(0) and the coefficient of q² in S(q) analytically.

Solid solutions alpha -Er(H,D)_x, with 0<or=x<or=0.03, have been investigated by measuring the resistivity evolution between 4.2K and room temperature. Similarly to the earlier studied phases alpha -LuH_x and alpha -Tm(H,D)_x, a resistivity anomaly is observed at 160-180K, which is a function of isotope mass and concentration. Hydrogen defects were introduced by electron irradiation at 15K and/or by quenching from room temperature. These defects interact strongly with the magnetic structure of erbium and anneal out in the region of the anomaly. The phenomena are interpreted in the framework of the pair model, assuming a stable H-H (D-D) configuration below the anomaly region, the defects being isolated H-atoms which result from the pair dissociation through electron bombardment or are quenched in from the high-temperature alpha -phase.

Specific heat measurements performed on the ferroelectric (N(CH₃)₄)₂ZnCl₄ reveal that the phase transition occurring in this compound can be classified into two groups. The commensurate to incommensurate to ferroelectric to paraelectric sequence belongs to the first group, whereas in the second group, for which a new anomaly was detected, the authors have three phase transitions associated with the organic group movements. The critical behaviour for the commensurate-incommensurate phase transition is discussed.

Using the neutron scattering technique, the wavevector of the modulation has been carefully measured as a function of temperature in the incommensurate phase of d-TMATC-Zn. The modulation curve is qualitatively compared with a constant- mu Devil's staircase of the one-dimensional model. A great difference appears between the behaviour of h and d compounds, which is attributed to the isotope effect. Integrated intensities of first-order satellites follow a linear law over a 10 degrees C temperature range below the transition temperature T_i=24.7 degrees C. Odd-order satellites have been observed up to the seventh order.

In bootstrap percolation, a fraction p of the sites of a lattice is first randomly occupied, and then all occupied sites with fewer than a given integral number m of occupied neighbours are successively culled (rendered unoccupied) until either (a) a configuration of occupied sites, each with m or more occupied neighbours, ensues or (b) all sites of the lattice are vacated. The authors treat this problem as a function of m on square, triangular and cubic lattices by computer simulation. Various types of percolation transitions are discovered, including some which are discontinuous and some which are continuous but characterised by non-universal critical exponents.

The usual continuous time random walk (CTRW) method for analysing the motion of particles between discrete localised states in disordered systems is fairly simple to apply but involves an averaging procedure whose accuracy is unknown. The Laplace transform of the rate equations for such a system is expressed in matrix form, and it is shown that the usual CTRW approximation corresponds to a certain decoupling of their configurational average. This is the first of a hierarchy of CTRW(n) approximations, which treat exactly transitions involving n steps, and conditions are established for the convergence of this series of approximations.

For pt.I see ibid., vol.14, p.3195 (1981). In order to test the accuracy of the usual continuous time random walk (CTRW) method, calculations were performed using it and its extension, the CTRW(n) method, for two types of one-dimensional disordered system. For the first of these, which corresponds to direct hopping between localised states of similar energy, the usual CTRW method was found to be qualitatively correct, but quantitatively unreliable because of its neglect of correlation between successive transitions. For the second type of system, which corresponds to band transport with multiple trapping, the usual CTRW method was found to be quite accurate.

Electrical properties of single crystals of Ti_1-xHf_xSe₂ (0<or=x<or=0.5) are studied at temperatures between 20 and 290K. The nature of the conduction is metallic for the range of x investigated except for a resistivity anomaly. The resistivity anomaly associated with a phase transition appears when 0<or=x<0.4. The sign of the Hall coefficient at room temperature is positive for 0<or=x<0.4, but negative for x>or=0.4 where the transition is completely suppressed. The results show that the presence of holes is essential to the occurrence of the phase transition. Furthermore, a characteristic change is found in the temperature dependence of the Hall coefficient as x is increased. This feature is similar to the contrasting temperature dependence of Hall coefficient between a pure and impure system which is predicted for an excitonic insulator.

Within the classical and quantum S=¹/₂ versions of the one-dimensional Ising model in a transverse field non-linear domain-wall-like excitations are studied. For the quantum case a domain wall (DW) wavefunction is constructed from the first (improper) excitation band. Various one-DW expectation values are then calculated. In particular, it is shown that the proposed DW wavefunction indeed yields the expected stationary-wall-like profile for the (S_l^z) expectation value. Results are compared with the ones obtained for the classical model and used to analyse the quantum-classical correspondence of domain walls in both models. The extension to non-stationary quantum DW is briefly discussed.

When crystals of Cs₂CdCl₄ (K₂NiF₄-type structure) are doped with Cr³⁺ a number of different impurity centres are formed which are readily distinguishable by EPR spectroscopy. The doped crystals contain both compensated and uncompensated Cr³⁺ centres. The resonances from the uncompensated Cr³⁺ exhibit axial symmetry about the crystallographic c axis. In the compensated centres the charge discrepancy of the Cr³⁺ appears to be balanced by a Cs⁺ vacancy. A centre with axial symmetry and a centre with rhombic symmetry can be attributed to Cr³⁺ ions compensated by Cs⁺ vacancies. If Li⁺ or Na⁺ are doped into the Cs₂CdCl₄ crystals along with Cr³⁺, a new rhombic centre is observed where the charge discrepancy is balanced by one of the small monovalent ions. The assignment and analysis of the EPR spectra from the doped crystals are discussed in detail.

The absorption spectra of ZnSe:Co and ZnS:Co contain sharp weak lines in the vicinity of 2.5 eV (20000 cm^-1) in addition to the quartet-quartet bands of Co d⁷. Photocapacitance measurements on ZnSe:Co reveal a deep level in the lower half of the gap approximately 2.2 eV (17700 cm^-1) below the conduction band edge. When the level is filled, cobalt is in the neutral d⁷ configuration. Both the electron and the hole photoionisation cross section spectra have been measured for this level. On the basis of the results, a model for the new absorption lines is proposed in which some quartet-doublet transitions gain intensity from the degeneracy of the final states with the conduction band while a phonon barrier decreases the auto-ionisation probability.

For a certain composition (close to GaSb), GaAlAsSb has a band gap (0.8 eV) equal to the spin-orbit splitting at the top of the valence band, and Auger recombination involving the spin-split-off band can proceed with zero threshold energy. An analytic expression for the zero-temperature recombination rate due to this process has been derived assuming a large heavy hole mass. The recombination rate is found to depend on the injected carrier density to the power 7/3 in contrast to the cubic dependence for the non-degenerate case. The room-temperature lifetime against this Auger recombination at electron and hole densities of 10²⁴ m^-3 typically found at lasing thresholds is estimated to be 3*10^-10 s, and order of magnitude smaller than the lifetime against spontaneous emission. This shows that efficient lasing is not possible with this alloy.

A number of reliable photoionisation spectra for the excitation of holes from deep level impurities to the valence band of silicon have recently been published. Experimental spectra for the impurities S, Se, Zn, Au, Pt and a neutron-radiation induced defect are compared with calculated spectra with respect to both spectral distribution and absolute value. The calculation is made with the binding energy as a parameter and using a delta-function impurity potential, plane-wave final states and a non-parabolic valence band. The fits obtained show that the spectral distribution of the levels studied is not significantly influenced by the chemical identity of the impurity. Similarly, the charge state is of little importance. Discrepancies in the absolute values clearly demonstrate important differences between Au and Pt, and between S and Se.

For pt.I see ibid., vol.13, p.1567 (1980). Many low-energy satellites in the beta -region of K spectra of elements ¹³Al to ⁴⁴ Ru have been assigned to K to M₁M_2,3, K to M_2,3M_2,3 and K to M_2,3M_4,5 radiative Auger transitions. The effects of configuration interaction on level splittings of the state M_2,3M_2,3 are discussed.

A large experimental data set of I(V) curves was collected over an energy range 50-400 eV at three diffraction geometries from the copper (111) surface. The experimental curves were compared with I(V) curves calculated from model structures. The result obtained for the relaxation of the topmost layer is -0.3%+or-1% (+or-0.02 AA) of the bulk interlayer spacing. The precision of the result comes from the large experimental data base and the application of statistics to an R-factor analysis. R-factors are presented for the formalisms both of Zanazzi and Jona and of Pendry. The two R-factor analyses give similar results for the relaxation of the copper (111) surface. However, the Pendry R-factor is found to be more sensitive to model parameters and faster to compute than the Zanazzi and Jona R-factor.

Photoemission yield spectroscopy measurements, associated with LEED and Auger characterisation, are performed on a set of silicon (111) surfaces with different bulk dopings as a function of silver coverage theta , from zero to a few monolayers, obtained by UHV evaporation on a sample kept at room temperature then annealed at 500 degrees C. When the square root 3 × square root 3 structure which is completed theta _c approximately=²/₃ monolayer is substituted for the 7 × 7 LEED pattern the work function of the system does not vary significantly, while the ionisation energy decreases by about 0.5 eV; it corresponds to a band-bending variation in the Si space-charge region that is in agreement with the change of surface-state density in the gap. The persistence of a Si dangling-bond peak after completion of the first silver layer favours models where a two-dimensional ordered layer is formed allowing Si atoms to be at the surface.