Table of contents

Diffuse low-energy electron diffraction (LEED) intensities can be observed in all situations where the surface cannot be represented by a perfect biperiodic array of atoms or molecules. For instance, this occurs when binding sites are either occupied or not by adatoms, when several atoms or molecules are co-adsorbed, when one absorbed species can be located at several possible binding sites, etc. So far, the statistical aspects of this problem have been disregarded. In fact, under certain circumstances, it is possible to extract from diffuse intensities a quantity that only depends on the local atomic arrangement near a binding site. A method based upon this possibility has been systematically used in all diffuse LEED studies and consequently these studies have been totally devoted to the determination of this local atomic arrangement. The authors show in this paper that this method only works if one binding site is occupied or not (binary chemisorption). In the other cases, investigation of diffuse LEED intensities requires knowledge of the statistical distribution of occupied sites and thus the statistical aspects of this problem can no longer be bypassed. For this reason and also because this problem is intrinsically interesting, the authors particularly focus on it here. Diffuse LEED intensity can be approximated by a sesquilinear form of the form factors for each adsorbed species. The coefficients of this form are the Fourier transforms on the two-dimensional surface lattice of the site-occupancy pair correlation functions. A self-consistent molecular-field approximation of these correlation functions is given in this paper. Particularly, the validity of this approximation is discussed in detail for the binary chemisorption case. Diffuse LEED intensities are provided in some other cases: (i) two kinds of atom are distributed at the surface of a binary metallic alloy; (ii) two adsorbed species coexist at a single crystal surface. The authors arrive at the conclusion that investigation of diffuse LEED intensities generally requires the direct comparison of measured and calculated intensities.

Ni/Zr multilayer films with single-layer thicknesses of a few hundred nanometres were prepared by electron beam evaporation and by DC magnetron sputtering. The films were subsequently annealed under high-vacuum conditions at temperatures between 570 and 640 K where the progress of the solid state amorphisation reaction was monitored by in situ measurements of the electrical conductivity. From these measurements the interdiffusion coefficient D was derived. The samples were characterised by X-ray diffraction with Cu K alpha ₁ radiation in Seemann-Bohlin geometry and Rutherford backscattering with 25 MeV ¹⁶O ions. Analysis of X-ray diffraction patterns shows that the Zr layers of evaporated samples are under tensile stress, whereas those of sputtered samples are nearly unstrained or under a small compressive stress. In sputtered films, D is about an order of magnitude smaller than in comparable evaporated films which is interpreted as a result of the different stress states. These results agree with the interpretation of the author's recent irradiation experiment. A correlation between the Zr crystallite size and D is found such that D decreases with increasing crystallite size. X-ray patterns also indicate that the nucleation of the amorphous phase depends on the crystallographic orientation in the Zr layers.

In the inelastic neutron scattering spectra of Ni-Ni-Bz(d₆) obtained at 1.8 K the authors observed transitions differing only slightly, but systematically, from those between uniaxial free rotator levels of ammonia groups. They interpreted the deviations as the influence of a weak hindering potential. A twelve-fold cosine potential with a barrier height V₁₂ of 16+or-3 meV together with a rotational constant of 0.71+or-0.02 meV for NH₃ gave a satisfactory description of the experimental results. Spectra obtained at higher temperatures (up to 51 K) showed qualitatively the transition from quantum to classical behaviour.

V₆Ni₁₆Si₇ has a very unusual reciprocal lattice, closely related to that of gamma -brass but requiring a minimum of six basis vectors in order to index all observed reflections. It can be described formally as a three-dimensional, incommensurately modulated beta -brass structure. Characteristic satellite extinction conditions observed at (001) zone axes imply the presence of certain six-dimensional super-space group symmetry operations which can be used to develop a modulated structure approach to the structural description of V₆Si₁₆Si₇ and to provide a structural comparison with gamma -brass.

Vacancy trapping at tin atoms during stage III and stage IV recovery as well as tin precipitation were studied in a fast-quenched Al-0.014 at.% Sn alloy by parallel positron lifetime and Mossbauer spectroscopy. The results show that the substitutional ¹¹⁹Sn isotope with a trapped vacancy has 2.275 mm s^-1 isomer shift. Trapping of more vacancies may cause significant line shift and leads to the appearance of a line at 2.86 mm s^-1 in the Mossbauer spectra. The origin of this line position was interpreted as the relaxation of substitutional tin atoms to tetrahedral and octahedral interstitial positions in the centre of the vacancy clusters nucleated at the tin atoms. The defect recovery taking place in the alloy was found to be dominated by the strong tin-vacancy interaction. In stage III annealing, both vacancy annihilation and the formation of vacancy clusters independent of tin atoms are considerably suppressed by the strong trapping of vacancies at tin atoms. In stage IV recovery, the dissolution of dislocation loops, the migration of tin-multivacancy complexes and the formation of tin-atom clusters were observed. Precipitation of metallic tin exhibits several substages. The positron trapping observed upon precipitation was ascribed to the mismatch free volumes at the incoherent particle-matrix interface of the precipitates.

In the present paper, an equation of state is proposed which has been found to be applicable at high pressures and high temperatures. The proposed equation of state has been used to compute the pressure-volume behaviour, the isothermal bulk modulus and its pressure derivative, the coefficient of thermal expansion and the Anderson-Gruneisen parameter of NaCl and CsCl at high pressures (up to 400 kbar) and high temperatures (up to 1073 K). The results obtained for V(P,T)/V(0,T) are found to be in very good agreement with Decker's data in the whole ranges of pressure and temperature. Further, the other calculated results are also found to be in good agreement with the experimental data. The present calculations definitely lend support to the proposed equation of state and to the method used to extend the applicability at higher temperatures.

The authors have performed detailed high-pressure studies up to 110 kbar to investigate the nature of the structural phase of YBa₂Cu₃O₇ using energy-dispersive diffraction techniques. They observed a reversible orthorhombic to tetragonal phase transition at between 70 and 80 kbar. During compression it is found that the c axis has the highest compressibility; thus c/a decreases. A possible explanation for the relationship between the variation of structure with pressure and the transition temperature is given.

This paper concerns the similar gamma incommensurate phases found in bis (n-propylammonium)tetrachloro-manganate and -cuprate. The modulation is driven by the propyl chains. A simple physical picture explains the microscopic origin of these modulated structures in terms of interatomic interactions. The ends of the propyl chains from successive layers interpenetrate at the 'peaks' of the modulation, while at the 'troughs' of the wave the adjacent chains from the same layer are brought closer together. Both effects serve to lower the Van der Waals energy between chains. Some theory and some computer simulation round off the picture and give quantitative support.

A comparison between experimental and theoretical densities of states (DOS) is performed for CoSi₂. Partial DOS of CoSi₂ are studied by soft X-ray spectroscopy (SXS) through Si K beta (3p to 1s) and Co L alpha (3d to 2p₃2/) emission bands. In addition, X-ray photoemission spectroscopy (XPS) measurements of the valence band are presented. On the theoretical side, a scalar relativistic self-consistent augmented plane wave calculation is performed from which the partial DOS are obtained. The theoretical DOS, with a Lorentzian broadening, are compared with the experimental results and found to be in good agreement.

A simple pseudopotential scheme, which incorporates compositional disorder as an effective potential, is proposed for calculation of the band structure of ternary compound semiconductors. It is shown that the present theory, which is free from any additional parameter, satisfactorily produces the band-gap bowings of ternary compounds when the lattice mismatch is small.

Applicability of the mean-field approximation is examined, when the slave boson technique is used for SU(N_d) Anderson lattice model. The fluctuation component of the slave boson field is explicitly introduced and its interactions with electrons are studied. The self-energy parts are calculated up to the second order of the interactions without relying upon the usual 1/N_d-expansion rule. Using perturbed propagators, the author obtains modified self-consistency equations for the mean-field parameters. They are solved at the transition temperature and also at zero temperature. The transition temperature decreases by a certain factor when the fluctuation is incorporated. This factor mainly depends on the location of the atomic f level and the degeneracy N_d. The author newly adopts this quantity as the quantitative measure of the fluctuation effects. It is confirmed that the mean-field theory is less affected by the fluctuation if the system is in the Kondo limit and the degeneracy is large. It is found that the fluctuation becomes more effective as the location of the atomic f level is closer to the Fermi level or the degeneracy of the orbital is smaller. The consequences for application of the mean-field theory to the real heavy-fermion compounds are discussed.

The electronic states responsible for transport constitute a wildly fluctuating minority of the total number of states in a disordered system. These fluctuations are the essence of the problem and demand a rich and powerful formulation if they are to be described accurately. In this paper the author introduces a new formalism for treating transport in three-dimensional systems based on the transfer matrix. The method gains its power from the interplay of theory of the symmetric group, which enables irrelevant parts of the mathematics to be thrown out, leaving the essentials intact but simplified. Analogies with replica theory are drawn, and replica symmetry breaking discussed.

For pt.I see ibid., vol.2, p.3273 (1990). In the previous paper a formalism was established for the calculation of intensities reflected from a disordered system, and hence for the transmitted intensities via the unitarity relationship. Here the author shows how the method can be applied to calculation of the conductivity of a three-dimensional (3D) system in the limiting case of weak disorder. First the one-dimensional (1D) situation is discussed and canonical models developed for the classical diffusive case and for the quantum localised case. The author's 3D theory can then be mapped onto the classical 1D case in the limit of weak disorder, and onto the quantum 1D case in the limit that all lateral hopping is eliminated and the author has a collection of independent 1D systems. Thus the theory has the power and flexibility to describe both localised and delocalised systems, a unique advantage in discussing more complex effects.

Non-dispersive transient currents in non-homogeneous films with a single discrete trap level have been calculated, the non-homogeneity consisting in spatial variations of trap depths or the presence of constant built-in electric fields in the limit cases of shallow and deep traps. Both kinds of spatial non-homogeneity may result in identical shapes of transients. Methods for determination of the spatial variations of trap depth and internal built-in electric field on the basis of experimental data are discussed.

In the present work the authors have investigated the influence of spatial non-homogeneity of trap distribution on thermally stimulated currents (TSC) due to transport of carriers in insulating layers. It has been assumed that at zero time carriers are generated only in a thin layer near one of the contacts. They have proved that for both non-dispersive and dispersive transport the initial increase of TSC is strongly dependent on the spatial distribution of traps. On the other hand, the dependence of TSC maximum on electric field strength, layer thickness and heating rate, and in the case of dispersive transport also on the final current decay, is almost the same as for a homogeneous trap distribution. The analytical results have been found to agree satisfactorily with the results of Monte Carlo simulation of non-isothermal carrier transport for all the model spatial and energetic trap distributions considered. The possibility of determining spatial trap distribution on the basis of the measured TSC has been discussed.

This is a report of measurements of the electrical resistivity and of the Hall coefficient of CoGe and CoGe₂ films between 80 and 520 K. Both germanides behave as metallic conductors with resistivities increasing linearly with increasing T over the whole temperature range. There is a slight deviation from linearity for CoGe₂ which is attributed to the fact that the mean free path of the carriers becomes comparable with the lattice dimensions at high temperatures. The Hall coefficient is negative for CoGe, giving an apparent carrier concentration of (0.86+or-0.01)*10²⁸ m^-3 and a Hall mobility of (11.0+or-0.1)*10^-4 m²V^-1s^-1 at room temperature. On the other hand, p-type conduction is revealed for CoGe₂ with apparent carrier concentration and Hall mobility equal to (1.93+or-0.01)*10²⁸ m^-3 and (2.2+or-0.1)*10^-4 m²V^-1s^-1, respectively, at 300 K. For a quantitative explanation of the resistivity and Hall behaviour of germanides, detailed information about the energy band structure near the Fermi level is required. At this information is lacking, the rough approximation of the two-band model is the only way to obtain insight into the electrical properties of germanides.

The internal friction in the high-T_c superconductor YBa₂Cu₃O_{7- delta} arising from the diffusion of oxygen atoms between the O(1) and O(5) sites under the influence of a periodic mechanical strain is studied microscopically in the linear approximation of the relaxation process. It is analysed for various specific cases in both the tetragonal and the orthorhombic phases as well as in the case of polycrystalline samples.

Spectra of paramagnetic spin fluctuations in insulating and metallic ferromagnets, obtained by inelastic neutron scattering, are interpreted in terms of appropriate Heisenberg models. A mode coupling approximation is used which takes full account of the lattice structure and the extended range of the exchange interaction, the form of which is obtained from the interpretation of spin wave spectra measured in the ordered state. The favourable outcome of confronting theoretical spectra with experimental data for EuO, Pd₂MnSn and Fe leads the authors to conclude that the mode coupling theory provides a more than adequate description of paramagnetic spin fluctuations. The dependence of the spectra on an increasing neutron wave vector is a change from a Lorentzian shape to a distinctly squared-up shape, but there is no strong evidence in any of the three systems of a damped collective mode contribution even for zone boundary wavevectors.

The magnetic properties of monocrystalline erbium have been determined with low-field AC susceptibility measurements between 4.2 and 150 K for different frequencies and magnetic fields with the applied field parallel and perpendicular to the c axis. The real component of the susceptibility shows anomalies at the antiferromagnetic transition (T_N=85 K) and at the ferromagnetic transition (T_C=19 K) as well as for the basal plane ordering transition at 52 K and for commensurate transitions at 34 and 28 K. The magnetic effects at these transitions are discussed in relation to the results of previous calorimetric measurements on the same crystal. To account for the observed non-zero value of the imaginary part of the susceptibility at the transitions, different loss mechanisms are suggested.

Magnetic and neutron diffraction studies on the series of mixed (Co_1-xMn_x)₂P phosphides are described. The magnetisation of CoMnP has been measured as a function of field at 4.2 K. The temperature dependence of the magnetisation along the b axis of the compounds with x=0.2, 0.5, 0.63 and 0.65 has been measured. Single-crystal neutron diffraction data at room temperature were used to refine the structure of two compositions with x=0.5 and 0.62. Polarised neutron measurements on CoMnP are interpreted in terms of antiferromagnetic-like domains associated with the orientation of the cobalt moments. The spin wave dispersion relations of CoMnP and (Co_1-xMn_x)₂P, x=0.65, have been determined at 300 K and in a low-temperature helical phase for the latter. A theoretical understanding of these results is possible in terms of the canting of cobalt moments due to a Dzyaloshinski-Moriya interaction.

Neutron scattering experiments were performed in the temperature range from 1.5 to 293 K on polycrystalline samples of the non-metallic 5f³ electron systems UBr₃ and UCl₃. The hexagonal crystal structure P6₃/m was refined at T=293 and 10 K. UBr₃ orders at T_N=5.4(1) K with a second-order transition to a 0+-configuration of magnetic moments of magnitude mu _U^(3.9K)=2.08(2) mu _B oriented parallel to the c axis. At T_t=3.0(5) K occurs a second transition to an ordered antiferromagnetic ground state, corresponding to the lattice constants a_m= square root 3a, b_m=a, c_m=2a. This structure is anticentered in the basal plane. The magnetic moments of magnitude mu _U^(1.6K)=0.82(4) mu _B are oriented parallel to the b axis. In both magnetic phases the antitranslation (0, 0, ¹/₂) suggests predominantly antiferromagnetic exchange within the chains of uranium ions along the c axis. The situation is similar for UCl₃: the first transition at T_N=6.5(1) K with an ordered magnetic moment mu _U^(4.0K) of 1.8(1) mu _B, and the second transition at T_t=3.8(5) K with a magnetic moment mu _U^(1.6K) of 0.8(1) mu _B. By means of inelastic neutron scattering the authors have measured the crystal-field splitting of UBr₃ and UCl₃ and determined the crystal-field parameters. The assumption of anisotropic magnetic interactions reasonably predicts the value of the zero-field magnetisation and provides an explanation for the magnetic phase transitions.

High-field magnetisation- and temperature-dependent susceptibility of Laves phase inter-metallic compounds Y(Co_1-xAl_x)₂ are investigated in the strongly exchange-enhanced paramagnetic region (0<or=x<or=0.11) in pulsed magnetic fields up to 100 T. In the whole concentration range, a sharp metamagnetic transition is observed in the low-temperature magnetisation process, while the low-field susceptibility exhibits a maximum at finite temperatures. The transition field B_c and the temperature of the susceptibility maximum T_max obtained as a function of x indicate a linear relationship, i.e. B_c/T_max=constant. The result clearly suggests that the susceptibility maximum of the nearly ferromagnetic metals is strongly correlated with itinerant electron metamagnetism.

Using Mossbauer effect (ME) spectroscopy and powder X-ray diffraction an uncalcined iron manganese oxide catalytic precursor was examined at room temperature after a magnetic separation of the magnetic and paramagnetic components was made. A previous identification of the paramagnetic component as the gamma -FeMnO₃ defect spinel is confirmed. Its interplanar spacings have now been measured and they show a strong resemblance to the gamma -Fe₂O₃ defect spinel. Both defect spinels give rise to similar mixed indices indicating a non-FCC structure. The previously postulated gamma -Mn_0.16Fe_1.84O₃ magnetic defect spinel could not be confirmed. The magnetic component is now compared with a synthetic magnetite. The ME parameters are strikingly similar and the X-ray spectrum of the magnetic fraction exhibits predominantly magnetite-like rings. Furthermore the quadrupole splitting observed in the Mossbauer spectrum of this phase is very small, confirming normal spinel behaviour. It is now postulated that the magnetic spinel is a Mn-substituted magnetite with the approximate composition Mn_0.6Fe_2.4O₄. Both phases have the same lattice parameter a=0.844 nm calculated on the assumption of cubic structures.

Numerical results are presented for the dynamical properties of a dilute, central force, triangular network with 60 degrees bond bending forces. The rigidity percolation threshold P_R for such a network lies between the connectivity percolation threshold and the central force rigidity percolation threshold. A LU factorisation technique is used to determine the eigenvalues of the dynamical matrix and the results combined with finite size scaling to yield values for P_R and the fracton dimensionality d of 0.40<or=P_R<or=0.405 and 1.25<or=d<or=1.3.

The authors have fabricated a 2D array of 1D narrow channels in a square lattice configuration on a high mobility GaAs/GaAlAs heterostructure. When a gate voltage is applied the underlying two-dimensional electron gas is depleted creating 4000 dots, of size 0.2 mu m, leaving a grid-like conducting path of quasi one-dimensional channels that are linked every 0.5 mu m. For low values of magnetic field the Hall voltage measured on either side of the device is quenched across its entire width of 25 mu m. The longitudinal magneto-resistance reveals Aharonov-Bohm oscillations with a flux period of h/e for magnetic fields up to 0.4 T. The conductance versus gate voltage shows structure that is consistent with Bragg reflections of the electron waves when the one-dimensional subband wavelength at the Fermi energy, in the direction of current flow, is equal to one superlattice period. Dips in the conductance when this criteria is met show a ln T temperature dependence, similar in origin to a 2D quantum interference. A magnetic field quenches these dips when the cyclotron orbit is comparable with the superlattice period.

A general classification of the solutions of the Ginzburg-Landau equation for UPt₃ is presented. The free energy includes an E_1g superconducting order parameter, an antiferromagnetic order parameter and the coupling between them. There are four distinct regions of the parameter space. In one of these the coupling acts as a perturbation and the split transition in UPt₃ is essentially a superconducting transition. The strong coupling region gives a mixed character to the transition. Phenomenologically it appears that the first alternative is most likely to be the correct one.

The authors have carefully re-measured the spin-wave dispersion relationship of holmium along (00l) in its spin-slip b=11 phase using a neutron triple-axis spectrometer. Their results are consistent with the existence of an energy gap in the spin-wave dispersion relationship close to the midpoint of the magnetic Brillouin zone. The measured magnitude of the gap of approximately=0.6 meV is in good agreement with the value predicted for a spin-slip model of holmium by Jensen in 1988.

Directional spin dependent Compton profiles of ferromagnetic nickel metal have been measured with 55 keV circularly polarised synchrotron radiation and the results have been compared with an APW calculation of the electron momentum density. At low momenta where the negatively polarised 4s-p band reduces the spin dependent momentum density, the model predicts accurately the contribution to the (100) lineshape but significantly underestimates the (111) profile.