Table of contents

The enhancements of the vapour pressures of solid H₂ and solid Ne produced by the action of high-pressure He have been measured at low temperatures and at pressures up to 80 MPa (800 atm) using a static method. Values deduced for the mixed virial coefficient, B₁₂, for the He-Ne system agree well with those derived recently from the more complex flow method, while for the He-H₂ system there is fair agreement with an extrapolation of values derived from P, V, T measurements.

The energy level splitting due to tunnelling motion of the methyl groups in sodium acetate trihydrate, Na(CH₃COO)3H₂O, has been measured by inelastic neutron scattering at 4K for hydrostatic pressures up to 5 kbar. The splitting falls from 5.7 mu eV at atmospheric pressure to 2.2 mu eV at 5 kbar with an exponential pressure dependence indicating a linear dependence of the height of the hindering potential on pressure.

Absorption bands at 393 cm^-1 and 464 cm^-1 in GaAs are due to the localised vibrational modes (LVMs) of nearest-neighbour (Si_Ga-Si_As) defects. A third line at 367 cm^-1 has been attributed to this defect but there have been doubts about the assignment. Samples have now been subjected to prolonged 2 meV electron irradiations and subsequent anneals, and it is shown that the strength of the absorption at 367 cm^-1 is not correlated with that of the line at 393 cm^-1. Thus the defect responsible for the 367 cm^-1 line is not the pair centre and the defect is designated (Si-Y). Another grown-in centre (Si-X) gives an LVM line at 369 cm^-1 and it is now shown that other silicon impurities do not give rise to detectable infrared absorption.

The dielectric response of solid electrolytes is obtained using a model which relates chi " to the charge-carrier number fluctuations. chi was found to vary with frequency as omega " where the exponent falls in the range -1<or=n<or=+1, depending on the dominant process in which energy diffusion occurs and on the dimensionality of the system.

The far-infrared absorption spectra of the 3-D uniaxial antiferromagnet FeF₂, substitutionally doped with Co²⁺, exhibits a defect-localised magnon mode at 85.5 cm^-1 just about 6 cm^-1 above the magnon band of the host. In an external magnetic field parallel to the easy axis, the energy separation between the local mode and the top of the magnon band, and consequently the localisation of the defect excitation, can be changed considerably, with the consequence that the absorption coefficient and the effective g factor will show a magnetic field dependence.

An argument by Edwards et al. (1977) regarding four as the critical dimension in the Anderson-Mott transition is reviewed. It is pointed out that convergence of the Green's function series for some random systems is not determined by its expectation value alone. A simple model also demands convergence of the series for the variances and this series, which is identical to that for the two-particle Green's function for this model, has eight as the critical dimension.

The critical temperature of any two-dimensional Ising model can in principle be determined exactly, providing there is a finite unit cell. This property is used to investigate the dependence of the critical temperature on the size of the unit cell where the exchange coupling between spins in the unit cell are assigned from a Gaussian distribution. It is found that as the number of spins in the unit cell is increased the critical temperature falls. Extrapolation to infinite unit cell, which corresponds to a spin glass, indicates that the critical temperature vanishes.

The excitation spectrum of the R line in MgO:Cr³⁺ at the foot of the ⁴A₂ to ⁴T₂ band is obtained. The coupling between ⁴T₂ and ²T₁ by spin-orbit interaction and the energy transfer operated by the Jahn-Teller effect are shown to be important in explaining the spectrum. The role of odd vibrations is discussed. The linear dichroism under stress of the ⁴A₂ to ⁴T₂ transition is determined for Cr³⁺ and V²⁺. These results are compatible with the absence of a strong Jahn-Teller effect.

An analysis of the orientation correlation between the neighbouring molecules of nitrogen and oxygen is given. The correlated form factor f₂(Q) is formed giving proper weighting to the parallel and perpendicular orientations and taking into account that fact that the centre structure factor in the region of the first peak should be well produced by a hard-sphere potential. These orientation correlations give about 62% and 64% parallel orientations in nitrogen and oxygen respectively using the proper choice of hard-sphere diameters and bond lengths from the measurements of Furumoto and Shaw. The very good agreement between experimental values of S_m(Q) and those calculated on the basis of this simple model shows that the independent-atom approximation is valid even for linear molecules like nitrogen and oxygen.

The monolayer is represented here by a two-dimensional lattice model in which a molecule in the singly bound state is a 'monomer' occupying one lattice site while a molecule in the doubly bound or 'extended' state is a 'dimer' occupying two neighbouring sites. There is an energy and an entropy of transition (both negative) from the former to the latter state. A series treatment of a lattice incompletely occupied by dimers, modified to allow for a monomer as well as a 'hole' or vacant-site species, is used. The three types of thermal anomaly can be represented by sets of theoretical surface-pressure/area isotherms in three reduced temperature ranges. A Flory-Huggins approximation is used to incorporate nearest-neighbour energies of interaction between the monolayer molecules and the three behaviour types are again reproduced.

A temperature-dependent (preaveraged) potential independent of orientation, which is thermodynamically equivalent to an orientation-dependent two- and three-body potential, is developed. The influence on the preaveraged potential of angle-dependent pair interactions (which arise from permanent electric-multipole moments, induced dipole moments, anisotropic dispersion and overlap forces), of three-body non-additive interactions (which arise from dispersion and classical electric-induction interactions between asymmetric molecules) is considered. This preaveraged potential is used to devise a simple perturbation scheme for the calculation of the thermodynamic properties in terms of the properties of a hard-sphere fluid. The calculated thermodynamic properties (configurational energy, entropy and pressure) for nitrogen and methane are in very good agreement with the experimental data.

A structural phase transition has been observed in CsFeF₄ at 250K by dielectric constant measurements, optical studies and X-ray diffraction. The structure below the transition, which appears to be of second order, has orthorhombic symmetry (space group D₂³-P2₁2₁2,Z=4).

The authors consider the question of the observability of the plasmon mode in molten salts by the study of frequency moments of the dynamical structure factor S_{mu nu} (k, omega ). They establish approximate relationships between the fourth moments and phonon frequencies in the solid which are confirmed by detailed calculations in KI. Various approximation methods for obtaining the neutron scattering function S(k, omega ) from the moments give consistent results indicating that the plasmon mode should be observable.

Single crystals of sodium chloride doped with bivalent anions (S^2-) have been prepared. Electric conductivity measurement of these crystals yields an enthalpy of migration ( Delta h_a) of anion vacancies in sodium chloride of 1.24+or-0.05 eV. This result is considered to be more reliable than previously published values.

Electronic densities of states of tight-binding s-bands for simple cubic and body-centred cubic lattices were determined using the modified-moments method. The weight functions close to the density-of-states functions for SC and BCC lattices were found. These functions define the orthogonal polynomials belonging to the class of polynomials considered by Bernstein and Szego. The analytical expressions for modified moments were derived and the densities of states in the form of continued fractions which reproduce 76 moments for the SC and 78 moments for the BCC lattice were obtained and compared with exact results. The convergence of the modified-moments method as a function of the level at which the continued fraction is terminated in an approximate way has been examined.

The crystal-field splitting of the ²F_5/2 ground state of Ce³⁺ in CeD_2.12 has been studied by inelastic neutron scattering. A single crystal-field transition, consistent with a cubic environment for the Ce ion, is observed at an energy of 20 meV. Low-temperature measurements of the quasielastic scattering in a magnetic field, however, cannot be explained in terms of crystal-field effects alone.

Experimental results for the dark current as a function of orientation, potential difference and temperature are reported for millimetre-sized single crystals of the solid-state-polymerised conjugated-polymer bis(p-toluene sulphonate) ester of 2,4-hexadiyne-1, 6-diol (PDA-TS). The results show that the carrier mobility along the polymer chains is 1000 times that perpendicular to the chains. Both ohmic and space-charge-limited conduction are dominated by a localised level 0.80+or-0.02 eV from the band edge and of concentration corresponding to <or=1 centre per metre of chain length. The ohmic conduction is due to impurities donating one excess carrier per 30 metres of chain.

Anisotropic scaling allows the transformation of an isotropic system into an anisotropic one. When the critical point of the isotropic system is known, the transformation gives the critical line of the anisotropic system. The author illustrates the method on the s-state Potts model using a partial Migdal-Kadanoff transformation.

The author presents calculations, using a transfer matrix method, for the one-dimensional classical planar model in an applied magnetic field. In particular, the correlation lengths for spin fluctuations parallel and transverse to the applied field are calculated. For low temperatures and antiferromagnetic coupling, these show behaviour characterising a crossover from planar to Ising-like behaviour, corresponding to a lowering of the effective spin dimensionality by the applied field. For low temperatures it is shown that the crossover can be characterised by universal scaling functions. The results are shown to be in good agreement with experimental values, obtained by quasi-elastic neutron scattering from TMMC which, for low temperatures, is expected to be well described by the planar model.

The magnetic structure factors in paramagnetic V₅S₈ have been measured using polarised neutrons. The measurements were made on a single crystal in which the moment was induced at 35K by an external field at 4.8 T. The availability of only small crystals limited measurements to the stronger nuclear reflections which precluded interpretation using Fourier techniques. The results have been analysed in terms of a single, spherically symmetric form factor to give the induced magnetic moments on the three crystallographically inequivalent vanadium sites. With the limited data it was not possible to determine accurately the division of moment between the V(2) and V(3) sites, but it is clear that the results are inconsistent with the model based on NMR and susceptibility data which assigns essentially all the moment to the V(1) site. The best assignment of induced moments is 0.053(8) mu _B for V(1) and 0.120(11) mu _B for the sum of the V(2) and V(3) moments.

The authors present a detailed study of the evolution of the magnetisation in mixed CoS_2-xSe_x compounds with various selenium concentrations from pure ferromagnetic CoS₂ to pure paramagnetic CoSe₂. They show that at low temperature these compounds are of three types: (i) below x=0.28, the compounds exhibit normal ferromagnetic behaviour with a magnetic moment of about 0.9 mu _B cobalt atom; (ii) 0.28<x<0.60, the 'metamagnetic' transition first observed by Adachi and co-workers takes place: in this range the samples which are weakly magnetic at low fields, recover a moment of 0.9 mu _B per cobalt atom under an applied magnetic field; (iii) above x=0.60 the samples are paramagnetic. However, in the intermediate range (0.28<x<0.60) the samples still have a low ferromagnetic contribution to their magnetisation and for ferromagnetic samples (x<0.28) when the temperature is increased a transition similar to that observed for intermediate concentrations is seen above the Curie temperature.

For pt.I see ibid., vol.12, p.4269 (1979). From neutron diffraction and NMR experiments it is shown that the magnetism of chemically homogeneous CoS_2-xSe_x compounds is characterised by inhomogeneously distributed moments on the Co atoms. Some of the cobalt atoms carry the same moment, 0.9 mu _B, as in pure CoS₂ whereas others are non-magnetic. The concentration and field dependence of the magnetisation are almost entirely due to the corresponding variation in the number of magnetic atoms, while the moment carried by these atoms remains nearly constant. The temperature dependence of the magnetisation is also attributable, for the main part, to the same mechanism, whereas thermal variation of the moment carried by these atoms remains nearly constant. The temperature dependence of the magnetisation is also attributable, for the main part, to the same mechanism, whereas thermal variation of the moment is only responsible for a small part of the loss of magnetisation when T increases. Furthermore, the thermal variation is almost concentration-independent. It is suggested that the cobalt atoms in CoS_2-xSe_x compounds have two magnetic states depending on the local environment, temperature and magnetic field: a non-magnetic and a magnetic state, which always has the same moment, 0.9 mu _B. In this type of material field-induced transition results from cobalt atoms switching from the non-magnetic to the magnetic state; cooperative phenomena acting through the coupling between cobalt atoms can then explain the hysteresis observed for this transition.

The magnetic phase diagrams of pure and diluted uniaxial Heisenberg antiferromagnets with crystal-field anisotropy of the form D(S^z)² are considered. Treating the anisotropy exactly and only the exchange interaction in the mean-field approximation, it is shown that for S=1, the phase transition between the paramagnetic and the spin-flop phase is first order for values of the anisotropy parameter D>D_c and is second order only for D<D_c. The tricritical point, at which the order of the transition changes, appears when the ratio of the anisotropy to the exchange energy is equal to ²/₃ and, in the MFA, is independent of temperature. For S=³/₂ the tricritical point coincides with the bicritical point at T=0K; for S>³/₂ no tricritical point is observed in the mean-field approximation. The antiferromagnetic to spin-flop phase stability limit never exceeds the antiferromagnetic to paramagnetic critical field.

Internal dilution with simultaneous maintenance of the concentration of magnetic ions at a higher level at the surface is proposed as a source of surface magnetism. Calculation shows that the most favourable conditions for this effect occur for a concentration of internal magnetic ions slightly in excess of the percolation concentration.

The rare-earth compounds exhibiting the mixed valence or the interconfiguration fluctuation (ICF) behaviour have been investigated by a number of techniques. However, no NMR work has yet been reported in the literature which is aimed at studying the effects associated with the fluctuating electric fields which themselves arise from the fluctuating charges on the rare-earth atoms in these systems. The authors report the first magnetic resonance studies of these effects. The systems chosen for these investigations are the members of the series RECu₂Si₂ and RECu₂Ge₂ (RE=rare earths) which belong to the ThCr₂Si₂ structure in which copper atoms occupy sites with non-cubic symmetry. These studies show that in favourable circumstances the quadrupole interaction measurements provide an attractive and sensitive alternative to the Mossbauer measurements in those cases where the latter cannot be performed. In conclusion, it has been pointed out that through the spin-lattice relaxation measurements of copper isotopes (which are quadrupolar nuclei) and silicon nuclei (I=¹/₂), it may be possible to investigate the charge and the spin fluctuation times.

The dependence of the adatom vibration frequency on the change in the local density of states between sites is examined. Adsorption on a surface of the same species with a half-filled band is assumed. The adatom bond length R is less than the bulk spacing and it decreases with decreasing coordination z. The short-range repulsion and hopping integral between the adatom and near neighbours are assumed to vary with R as exp(-pR) and exp(-qR) respectively. For p/q=2, the adsorption energy E_ad⁰ is independent of z, and the force constant for vibration normal to the surface is k=pq cos² theta E_ad⁰, where theta is the angle between the bond and the surface normal. If E_ad⁰ varies little between sites, a much used approximation is to take k dependent on theta but not on z. This is exact in the present model for p/q=2. This value is typical of simple molecules, but rather larger values are appropriate for transition metals. For all p/q the dependence on z of k is less than the linear variation obtained with an additive site-independent pair potential.