Table of contents

We have applied the fast Fourier transform, which allows one to compute efficiently convolution sums, to solve the set of self-consistent T-matrix equations to get the Green function of the two-dimensional attractive-U Hubbard model below , extending previous calculations by the same authors. Using a constant order parameter , we calculated as a function of the electron density and interaction strength U. These global results deviate from the BCS behaviour remarkably.

We report the first muon-spin relaxation study of an organic spin - Peierls system, the linear-chain compound MEM(TCNQ). Our results show a crossover from a Gaussian relaxation to an exponential relaxation as the temperature is lowered below the spin - Peierls transition. We associate this behaviour with the slowing down of the electronic fluctuations resulting from the opening of a gap in the magnetic excitation spectrum.

A study of the specular and off-specular (diffuse) x-ray scattering of a diblock copolymer is presented. In the ordered state the surface of the diblock copolymer is covered with islands. It is shown that the periodicity in the ordered state in the direction normal to the surface in a specular scan differs from the periodicity in the same direction observed in an off-specular scan. This result is explained by an analytical calculation of the differential cross section. The introduction of the statistical properties of the island distribution allows a complete analytical calculation of the transverse scans yielding the determination of the mean distance between the islands and the average size of an island.

Neutron diffraction experiments and macroscopic magnetization measurements show evidence of the stabilization of the helical magnetic phase over large temperature ranges for Tb and alloy films grown epitaxially on yttrium. In particular, the temperature of the transition between the helical and the ferromagnetic states is shifted from 220 K for bulk terbium to 160 K for a pure terbium film grown epitaxially on yttrium, despite the low stability of the helical phase in the bulk element. The decrease of the Curie temperature is due to the negative c-axis strain induced by the epitaxial growth on yttrium. The epitaxial strains also induce modifications of the Fermi surface, which leads to an increase of the turn angle even at . At low temperature, a long-wavelength-modulated phase, whose origin still remains to be explained, has been observed.

We have directly compared the values of the magnetic moment obtained from x-ray magnetic circular dichroism (XMCD) measurements with those obtained from polarized neutron reflection (PNR) measurements on strained Ni films grown on Cu(001)/Si(001). The PNR measurements show that the absolute magnetic moments differ from that of bulk Ni. We find agreement within experimental errors between the two magnetometry measurements, confirming that the XMCD sum rules are applicable to this strained low-symmetric system.

Chemical reactions carried out under strongly non-equilibrium conditions can result in a variety of interesting effects such as oscillations, chemical waves, kinetic phase transitions, bistability, instabilities and chaos - all of these effects being mathematically due to the strong non-linearities in the kinetic rate equations describing these chemical processes. The present article reviews the various types of non-linear behaviour observed in several NO reduction reactions over Pt-group metals. A large part of this review deals with oscillations and spatiotemporal pattern formation in the reaction over Rh surfaces: the reaction dynamics of this system has been realized on the microscopic, mesoscopic and macroscopic scales using field emission microscopy, photoemission electron microscopy and macroscopic rate measurements respectively. The various mechanisms of this and the other reactions reviewed in this article are also considered.

The incident-energy dependence and angular distribution of a scattered- yield were measured and compared with those of scattered- and scattered- yields. Below the incident energy of 80 eV, both resonance and Auger neutralization processes are considered to contribute to the ion-survival probability in and scattering, from the analysis based on the calculation by Imke et al (Imke U, Snowdon K J and Heiland W 1986 Phys. Rev. B 34 41, 48). Above the incident energy of 80 eV, the electron promotion mechanism contributes additionally to the ion neutralization process, and causes the difference between and the other ions as regards the survival-ion yields. The threshold energy at which dissociated appears for incidence is . The dissociated is considered to arise via reionization of a D atom which is dissociated from once-neutralized via impulsive collision.

The energy-resolved momentum densities of the valence band of a thin polycrystalline aluminium film have been measured using electron momentum spectroscopy (EMS). The spectrometer used for these measurements has estimated energy and momentum resolutions of 0.9 eV and 0.10 atomic units respectively. The valence band of aluminium was clearly resolved, resembling very closely that of a free-electron parabola. The measurement has been compared to linear muffin-tin orbital (LMTO) calculations for spherically averaged crystalline aluminium. A comparison has also been made between the experiment and Monte Carlo simulations which take into account additional elastic and inelastic scattering events not considered in the LMTO calculations. The final agreement obtained between the measurement and theory for the dispersion and relative intensities of the aluminium valence band is excellent when lifetime broadening of the band is allowed for.

Magneto-impedance effects have been studied in amorphous and nanocrystalline ribbons. Large magneto-impedance responses have been obtained in nanocrystalline samples by applying a longitudinal DC field, but not in amorphous samples. It was found that the transverse field dependences of the resistance and the reactance of nanocrystalline samples show very broad peaks at relatively high fields (above 35 Oe), while the longitudinal magnetic responses show very sharp peaks at the field of 2 - 7 Oe, sharing some common features with those reported for Co-based amorphous wire and ribbon samples, and can be understood from the same mechanism. It was also found experimentally that the properties of the longitudinal and transverse field dependences of the effective permeability of nanocrystalline samples correlates with those of the impedance, showing a sharp decrease for the longitudinal applied field and a slow decrease for the transverse applied field at a corresponding field value and, the higher the effective permeability, the larger is the magneto-impedance effect at a fixed frequency. and a sensitivity of 20% at a low field (2 Oe) are obtained in this paper.

The Ge(111) surface has been investigated as the temperature was varied in the 300 K - 1150 K range by valence band (VB) photoemission spectroscopy and x-ray absorption spectroscopy above the Ge 3d edge. VB photoemission shows that the high-temperature surface has a metallic character and that at temperatures above the 1050 K high-temperature phase transition a clear change of the surface electronic structure occurs. Metallicity at high temperature was also detected by absorption spectroscopy which shows the presence of a well-defined Fermi edge at an energy value consistent with earlier core photoemission results. From absorption data and numerical simulations the width of the Ge(111) surface conduction band was found to be 1.1 eV.

An inversion of the x-ray photoelectron diffraction (XPD) experiment has been performed in which incident electrons are diffracted to reach different sites within a crystal with differing relative probabilities reflecting those for escape of photoelectrons from the same sites in the corresponding XPD experiment. The diffracted-electron flux arriving at each site was monitored by using a Si(Li) detector to measure the intensity of the characteristic x-ray emission following core ionization by the incident electrons. This new technique, termed `inverted' x-ray photoelectron diffraction (iXPD), has been initially applied to Si(100) for rotation about , and the expected close parallels between XPD and iXPD data have been observed. Although the intensity of the x-ray signal increased rapidly as the incident-beam energy was increased beyond the threshold value for core ionization, the extent of the angular anisotropy correspondingly decreased, and there was little advantage in terms of signal:noise ratio in the iXPD pattern in increasing the energy to more than 10 - 15 eV above the threshold. The reduction in anisotropy resulted from a progressive degradation of the directional integrity of the electron beam by inelastic energy-loss processes. The angular resolution in iXPD is governed by the incident-beam divergence, which can easily be as small as , much less than the angular acceptance of typical electron energy analysers. High-resolution iXPD data taken above the threshold showed up to 38% angular anisotropy, much more than for conventional XPD at angular resolution, while iXPD data taken as much as 15 - 30 eV above the threshold still revealed splittings of all of the principal forward-scattering maxima of . Similar splittings have been reported for high-resolution XPD of other systems, but not for Si(100). Their origins are briefly discussed.

We re-examine the calculation of the bulk modulus of random networks with a variable mean coordination that ranges from down through the phase transition from rigid to floppy that occurs at around . In contrast to previous workers, we use random-network models, rather than depleted diamond lattices, so our results are more relevant to glasses. We find that the bulk modulus behaves in a very similar way to that found previously for depleted diamond lattices, with the bulk modulus going to zero at around with an exponent of 1.4. In the course of this study we came across many examples of transitions between different local minima, especially in networks with a low mean coordination. We discuss and illustrate the nature of these metastable states and show that although there are differences in the local minimum energy associated with these states, the bulk modulus (curvature around the minimum) is essentially independent of which minima the system is in. We show that these nearly degenerate local minima are associated with different local confirmations of very short polymer chains.

Porous silica dispersed with silver (Ag) nanoparticles (about 3 nm in diameter) within its pores has been prepared by a new method. The microstructures and the size distribution of particles within pores have been examined by transmission electron microscopy and the Brunauer - Emmett - Teller technique. The Ag nanoparticles are uniformly dispersed within the pores of silica; the particle size follows the log-normal distribution function. The thermal effect of this porous composite, exposed to ambient air for different times, has been investigated by differential scanning calorimetry and thermogravimetric analysis. An endothermic peak has been found for all the doped samples measured. At room temperature (298 K), a short exposure (2 d) to dry air leads to a wide endothermic peak at around 400 K; long exposure (1 month) results in a wider endothermic peak at around 450 K; when the sample was exposed to ambient air with relative humidity greater than 60%, a much higher endothermic peak exists at around 460 K. From the experimental results and discussions of adsorption and oxidation, the endothermic peaks for the samples exposed to dry air can be mainly attributed to the desorption of oxygen physisorbed and chemisorbed on the surface of the Ag particles within the pores, for lower- and higher-temperature peaks, respectively, and the corresponding desorption enthalpy values were estimated to be about 0.26 eV and 0.90 eV, respectively. For the sample exposed to humid air, the endothermic peak originates from the decomposition of silver oxide () formed on the surface layer of the Ag particles, and the bond energy of Ag - O in the film was estimated to be about 1.8 eV.

Phase transitions in polytypic substances can display a rich structure. A polytypic material, being formed from stacked layers, each layer having freedom of orientation, has an infinite number of possible structures. Thus a phase transition between two simple structures could occur directly, or via an infinite sequence of intermediate phases. Such a sequence, called a `devil's staircase', can arise from simple and general mathematical models. This paper presents a simple model in which the phonon free energy drives a temperature-induced phase transition, the mechanism which is believed to cause phase transitions in SiC, and . The form of interaction between changes in the stacking orientation caused by the phonon free energy is found to be inversely proportional to the square of the separation of the changes, but of alternating sign. Although no staircase results from this interaction, one intermediate phase does arise, and others are barely unstable.

It is shown that the melting temperature approach is quite successful in representing the pressure dependence of the self-diffusion coefficient in solids at different temperatures. Moreover, the approach is also capable of giving the activation volume and compressibility as functions of both the pressure and the temperature. The calculations are performed in the cases of Na, Pb, Cd and Zn.

An analysis of a general non-perturbative technique for calculating ground-state properties of extensive lattice many-body systems is presented, in order to extract accurate numerical values characterizing the ground-state spectrum. This technique, the plaquette expansion, employs an expansion about the thermodynamic limit of the coefficients that are generated by the Lanczos process. For the ground-state energy this error analysis, using theorems on the error bounds for the Lanczos method and the truncation in the plaquette expansion, allows for an accurate estimate when the approximation is taken to a given order. As an example we analyse the one-dimensional antiferromagnetic Heisenberg model, and find that the best ground-state energy density is within of the exact value, although the systematic error is . We also find, for this model, systematic improvement with each new order included in the expansion and have not observed any asymptotic tendencies. At equivalent orders of truncation we achieve far better results than for the other moment methods, such as the t-expansion or the connected-moment expansion.

We present calculations on all- trans-polyacetylene (t-PA), using localized orbitals for the calculation of total correlation energies in an ab initio framework. We show that due to the localization properties of the localized Wannier functions, especially the virtual ones, simultaneous interactions between three unit cells of a polymer must be included. However, if a larger number of neighbours are taken into account our method is still faster than those using canonical HF orbitals. Our LO approximation is shown to be able to recover about 90% of the correlation energy obtained in the canonical orbital basis in the equilibrium geometry. Furthermore, we present a different approximation which also reproduces potential curves very well, although this variant recovers only about 80% of the total correlation energy per unit cell calculated with a canonical orbital basis. This failure, however, leads only to a shift of the potential parallel to the canonical one, even in a very subtle case like that of the bond-alternation potential in t-PA which depends strongly on the quality of the correlation calculation method used. For the equilibrium bond alternation (projected onto the polymer axis) of t-PA the coupled-cluster doubles method with localized orbitals yields values almost identical to those obtained with canonical Møller - Plesset perturbation theory of fourth order including single, double, triple and quadruple excitations, published previously in the literature (Suhai S 1995 Phys. Rev. B 51 16 553). Furthermore, our results on agree fairly well with experiment, while the results of density functional calculations, also given in the above-mentioned work, are usually too small. Only one of the functionals applied yields comparable values for - surprisingly, the one which contains no correlation part.

The instability of a strong pulse current is observed in a plate of pure tungsten. The instability appears at low temperatures if the plate is placed in a high magnetic field parallel to the current direction. The magnetic field of the alternating current in the sample is detected by making use of a Hall probe. The inherent frequencies of the instability currents are some tens of hertz. Their amplitude ranges up to about hundreds of milliamperes. The primary directions of the instability fields are along the steady magnetic fields, so that the instability currents have a component at right angles to the initial current direction. There is a threshold value of the amplitude of the pulse current at the beginning of instability. It has been found that the threshold value depends on the steady magnetic induction and is reduced significantly when the induction increases.

The amplitude of the instability currents under study exhibits features of the dynamical chaos regime. The conversion into the chaos regime is effected by the formation of turbulent time intervals intermittent with laminar time intervals. It was found experimentally that the occurrence of the current instability is accompanied by ultrasonic oscillations.

The observed instability is attributable to the fact that the magnetic field of the current pulses changes the trajectories of electrons. A set of equations has been proposed in order to explain the oscillatory behaviour of the fields of instability. This set contains two differential equations of parabolic type like the distributed parameter system with diffusion.

We present the results of a study of thermal and quantum nucleation in an overlap Josephson junction with a finite length. There is a critical length that marks the boundary between nucleation with a uniform phase across the junction and nucleation that is concentrated at one of the ends of the junction. In the thermal activation regime, it is shown to be , while at absolute zero, in the quantum tunnelling regime, it is given by . Here, is the effective Josephson penetration depth. The rates for nucleation at the ends of the junction are given for junctions in the thermal activation regime for all lengths, and for junctions undergoing quantum nucleation at absolute zero temperature for lengths less than or equal to the critical length.

Fe-rich alloys in the Ti concentration range have been successfully prepared in the amorphous state for the first time using the DC co-sputtering method. Exhaustive ferromagnetic resonance (FMR) measurements were performed on these alloys at a fixed microwave-field frequency of about 9.23 GHz in the temperature range 77 - 300 K. In this temperature range, a single resonance is observed for the alloys with x = 10 and 11 as against two resonances (primary and secondary), having different properties, in the case of a-. For the primary (single) resonance in a- (the alloys with x = 10 and 11), the `in-plane' uniaxial anisotropy field scales with the saturation magnetization . This result suggests that the anisotropy energy is of dipolar origin. By contrast, the relation breaks down in the case of secondary resonance. The primary and secondary resonances thus characterize two distinctly different amorphous magnetic phases which widely differ in their chemical as well as geometrical short-range orders. For all the alloys in question, thermal demagnetization is mainly due to spin-wave (SW) excitations, and the SW stiffness D decreases with increasing x. SW modes soften at (the so-called re-entrant transition temperature) and D possesses a reduced value for . The existence of a re-entrant state at low temperatures is also vindicated by the rapid increase in the magnitude of the peak-to-peak FMR linewidth below a certain temperature. The temperature at which an upturn occurs in shifts to lower temperatures as the Fe concentration is decreased, indicating that the re-entrant behaviour is progressively suppressed with increasing x.