Table of contents

Based on the recent bond-order-bond-length-bond-strength correlation mechanism (Sun C Q, Chen T P, Tay B K, Li S, Huang H, Zhang Y B, Pan L K, Lau S P and Sun X W 2001 J. Phys. D: Appl. Phys.34 3470) and the criterion of thermal-vibration-exchange-interaction energy equilibrium, an atomistic model has been developed for the Curie temperature (T_C) suppression of ferromagnetic nanosolids. At T_C, the atomic thermal vibration energy (E_V) overcomes the atomic cohesive energy (E_coh), which triggers the order-disorder transition of the spin-spin exchange interaction. Besides, the coordination-number (CN) imperfection at a surface enhances the strength of the bonds of the surface atoms. The CN reduction and bond-strength enhancement modifies the surface atomic E_coh from that of an atom inside the bulk. As such, the critical E_V for an atom at a free surface will be different from the bulk value and, hence, the T_C of a nanosolid will change with the portion of surface atoms. Matching between predictions and experimental observations on the T_C suppression of Fe, Ni and Co nanofilms evidences the validity of the current premise, in which no assumptions or freely adjustable parameters are involved.

The nanometre-scale materials and devices that are under ever more intense study at present often depend for their unique properties on buried interfaces between two phases. Yet the number of experimental techniques that can specifically probe such interfaces, particularly with magnetic sensitivity, is limited. We here report a novel type of non-destructive method for spectroscopically studying buried nanometre-scale interfaces and other nanostructures with soft x-ray standing waves. Strong standing waves with a period of 4.0 nm and approximately 3:1 contrast ratios are created via Bragg reflection from a synthetic multilayer of form [B₄C/W]₄₀. By growing a wedge-shaped Fe/Cr bilayer on top of this multilayer, the mechanical translation of the bilayer exposed to a fixed and finely focused synchrotron radiation beam is converted into a translation of the standing wave through the interface. Analysing various core photoelectron intensities as a function of angle and beam position permits derivation of layer thicknesses and interface mixing/roughness scales. Magnetic circular dichroism in photoemission from the 2p and 3p levels of Fe and Cr further permits derivation of the positions and widths of regions with decreased (increased) ferromagnetic alignment for Fe (Cr), showing that normally antiferromagnetic Cr becomes ferromagnetic just below the centre of the interface but with antiparallel alignment with respect to Fe, and that the equal-concentration region in the centre of the interface strongly inhibits magnetic alignment for both species along the direction of net magnetizations that is probed. The magnetically altered regions in both metals are only 1-2 atomic layers in thickness. 3s spectra from Fe and Cr further indicate that the local spin moments on both atoms do not change on crossing the interface. This standing-wave-plus-wedge method should have a range of applications for the characterization of nanostructures and their interfaces.

Electron correlation effects in Fe are analysed using a first-principles linear combination of atomic orbitals scheme. In our approach, we first use a local orbital density functional theory local density approximation solution to introduce a Hubbard Hamiltonian without fitting parameters. In a second step, we introduce a many-body solution to this Hamiltonian using a dynamical mean-field approximation. Our analysis shows that magnetism in Fe is an effect associated with the first atomic Hund's rule. Moreover, we also find important correlation effects in the Fe spin-polarized density of states. The photoemission spectra can be explained using a value of U^eff as large as 4 eV, provided that the satellite peaks appearing around 3–5 eV below the Fermi energy are interpreted appropriately.

X-diffraction experiments on liquids are most commonly analysed within the independent atom approximation (IAA), despite the fact that compounds of light elements may show deviations from this simple approximation. An analysis of high energy x-ray diffraction data on liquid water in terms of the electron density auto-correlation function is given. The analysis shows that some but not all deviations from the IAA can be accounted for by a charged atom model. This corroborates findings from Compton scattering experiments on solid water that the nature of hydrogen bonding is at least partly covalent. The electron densities in H₂O and D₂O are found to be indistinguishable within the limits of the present experiments. This is in agreement with the interpretation of the differences in the intermolecular structure of H₂O and D₂O as being quantum mechanical in origin, and not the effect of a differing interaction potential.

The disordered nature of glass-forming melts gives rise to non-Arrhenius and non-exponential behaviour of their dynamics. With respect to the microscopic details involved in the structural relaxation, these materials have remained an unsolved puzzle for over a century. The observation of spatial heterogeneity regarding the dynamics provides an important step towards understanding the relation between the macroscopic properties of soft condensed matter and the molecular mechanisms involved. On the other hand, dynamic heterogeneity is the source of several new questions: What is the length scale and persistence time associated with such clusters of relaxation time? What is the signature of heterogeneity at high temperatures and in the glassy state? How do these features depend on the particular material and on the correlation function used for probing these heterogeneities? This work attempts to review the various approaches to heterogeneous dynamics and the generally accepted results, as well as some controversial issues. Undoubtedly, heterogeneity has provoked a number of novel experimental techniques targeted at studying glass-forming liquids at the molecular level. It will be emphasized that the picture of heterogeneity is a requirement for rationalizing an increasing number of experimental observations rather than just an alternative model for the dynamics of molecules.

In this work, a study of photon avalanche at 476 nm (¹G₄→³H₆) in Tm-doped fluoroindogallate glasses is reported. Transient Tm³⁺ emissions at 476, 450 nm (¹D₂→³F₄) and transmission at the pumping wavelength (645 nm) were monitored as a function of excitation power and Tm³⁺ concentration. We propose a theoretical treatment based on a set of five non-linear coupled rate equations, which include sequential two-photon absorption (TPA), two cross-relaxations among Tm³⁺ ions, and energy migration at the ³F₄ metastable state. It is demonstrated that energy migration in highly doped samples plays a fundamental role in the photon avalanche process, and may even suppress it. Additionally, sequential TPA must be considered in order to explain the transient behaviour of the 450 nm emission in the photon avalanche regime.

Using a piston-cylinder displacement technique, the pressure-volume (P-V) relation of Zr₄₁Ti₁₄Cu_12.5Ni₁₀Be_22.5 bulk metallic glass (BMG) has been measured up to a pressure of 4.5 GPa at room temperature. In the lower part of the high-pressure region the change of volume with pressure has remarkable non-linear characteristics indicating the existence of a large amount of free volume. A simple equation of state (EOS) with a cubic polynomial can be used to fit the experimental P-V data very well. From this, a potential function describing the EOS of the BMG has also been determined through combining experimental and theoretical analysis.

The behaviour of metastable phases involved in phase transformations of a simple fluid confined to a nanoscopic slit-pore is investigated within the mean-field lattice-gas approximation. For small displacements δρ_i (where i labels lattice sites) of local densities about their values ρ_i⁰ at the (metastable) minima of the grand potential functional Ω[ρ], the change in Ω is quadratic in δρ_i. The `force constants' associated with the δρ_i are the elements of the Hessian evaluated at the minimum (i.e., (∂²Ω/∂ρ_i ∂ρ_j)₀). If the walls of the pore are homogeneous, the Hessian reduces, in reciprocal space, to a single n_z×n_z matrix, where n_z is the number of lattice planes parallel with the walls. The complete phase diagram, including spinodals, is determined for n_z = 3, in which case the Hessian can be diagonalized explicitly. As the spinodal of a given (metastable) phase is approached, a characteristic eigenvalue of the Hessian tends to zero. The components of the corresponding eigenvector, which are proportional to homogeneous density fluctuations in the lattice planes, can indicate the nascent phase to which the metastable phase is transforming. The new phase is not necessarily the globally stable one. This suggests that sorption could take place via stepwise transformations involving phases that are metastable on the timescale of the observation.

We use molecular dynamics (MD) simulations to investigate the modification of the dynamic and static properties of liquid toluene when confined in cylindrical mesopores a few molecular diameters across. Due to the strong influence of the substrate on the dynamics of the confined liquid, we choose a procedure where no additional thermal interactions between the wall and the liquid are taken into account. We observe the characteristic oscillations of molecular density profiles (layering) when temperature and pore size are changed. Mean square displacements and intermediate incoherent scattering functions of the centre of mass of the molecules are calculated as functions of different distances from the wall along the principal axis of the pore z and along the perpendicular x- and y-directions. At 200 K the relaxations of the two correlation functions slow down by one order of magnitude as compared to the bulk, with a slightly more pronounced slowing down in the x-direction. This slowing down increases strongly when the wall is approached. However, we do not observe any layer-specific dependence of the dynamics, but instead a continuous change. When the molecules are arrested near the wall in the time window (1 ns) of the simulations, we find hopping processes.

High-temperature measurements of the electrical conductivity σ(T), thermo-emf S(T) (in a range from the vicinity of the melting temperature, T_m = 1371 to 1600 K), and viscosity η(T) (from T_m to 1420 K) were performed for equiatomic liquid CdTe alloys and with admixtures of In, Ge, Sn. It was shown that all the compositions kept their semiconducting properties during transition from a solid to a liquid state. A gradual melt metallization has been observed in the course of further heating. The results obtained are interpreted in the framework of the two-structure model suggesting coexistence of two areas: a densely packed metallic one and an area with the crystalline CdTe clusters. Transition to metal conductivity is considered as a gradual increasing of the densely packed metallic fraction. The influence of admixture on the electrical conductivity and thermo-emf is explained by the s-p hybridization phenomenon.

We study periodic chain clusters comprising equal numbers of two-dimensional bubbles of two different areas, with at most four bubbles per period. The cluster energies are calculated as a function of the ratio λ of bubble areas and of the imposed strain . We identify the clusters of lowest energy for each (λ,), and obtain the Young's moduli of all clusters in their unstrained state. An approximately linear correlation has been found between the Young's moduli and the cluster binding energy per unit area, similar to that which holds for crystalline solids (e.g. metals).

The crystal structures and electronic properties of MV₃O₇ (M = Cd, Ca and Sr) and their mixed systems Cd_1-xCa_xV₃O₇ and Ca_1-xSr_xV₃O₇ have been explored by x-ray four-circle diffraction and through measurements of magnetization and electron paramagnetic resonance. The structure changes systematically according to the ionic radii of M. The analysis of susceptibility in terms of high-temperature series expansion of up to the eighth order indicates that the V-V exchange couplings sharing oxygen corners are antiferromagnetic and they are one order of magnitude larger than those sharing oxygen edges: one of the latter couplings is ferromagnetic and another coupling changes from ferromagnetic to antiferromagnetic depending on M. The transitions to the stripe phases for M = Ca and Sr are understood qualitatively. The magnitudes of the exchange couplings estimated here account for the spin-flop phenomena and their composition dependences are consistent with those of the V-V distances and V-O-V angles.

The structural phase transition in CaMn₇O₁₂ has been investigated by using high-resolution synchrotron and neutron powder diffraction. Both measurements show a phase coexistence phenomenon: between 409 and 448 K two different crystallographic phases coexist in the material. The first one is trigonal and it has a charge ordering (CO) of the Mn³⁺ and Mn⁴⁺ ions, while the second one is cubic and charge delocalized (CD). The volume fraction of the CD phase increases with temperature from zero at 400 K up to 100% about 460 K. Both phases have domains of at least 150 nm at each temperature in the PS region. A percolation scenario assuming a growth of the volume of the highly conducting CD regions at the expense of the volume of the insulating CO matrix is discussed and it is found to be in agreement with literature data of the CaMn₇O₁₂ resistivity.

In this paper we study the electronic structure, electron density distribution and bonding mechanism in transition-metal (TM) di-aluminides Al₂TM formed by metals of group VIII (TM = Fe,Ru,Os) and crystal structures of TM di-silicides C11_b (MoSi₂), C40 (CrSi₂) and C54 (TiSi₂). A peculiar feature of the electronic structure of these TM di-aluminides is the existence of a semiconducting gap at the Fermi level. A substitution of a 3d TM by 4d or 5d metal enhances the width of the gap. From the analysis of the charge-density distribution and the crystal-orbital overlap population we conclude that the bonding between atoms has strong covalent character. This is confirmed not only from the enhanced charge density halfway between atoms, but also by a clear bonding-antibonding splitting of the electronic states. Groups of bonding and antibonding states corresponding to a particular bonding configuration of atoms are separated by a gap. As such a gap is observed in all bonding configurations among atoms in the unit cell it results in a gap in the total density of states. The bandgap exists at a certain electron per atom ratio e/A≈4.67 and also occurs in TM di-aluminides of groups VII and IX. For group VIII TM di-aluminides the Fermi level falls just in the gap.

The thermal expansion and low-temperature specific heat measurements for Y₆(Mn_1-xFe_x)₂₃ compounds have been carried out in order to discuss the relationship between the thermal expansion anomaly and the spin fluctuations. The thermal expansion coefficient α significantly increases above the Curie temperature T_C. This behaviour, the so-called anti-invar characteristic, is attributed to a remarkable increase of the amplitude of spin fluctuations in paramagnetic temperature regions. The anti-invar characteristic becomes weaker with increasing x. The temperature dependence of the thermal expansion coefficient of the magnetic term α_mag is convex upward in the paramagnetic temperature range, which is explained within the framework of spin fluctuations.

The atomic exchange-correlation (xc) potential with the correct -1/r asymptotic behaviour constructed by Parr and Ghosh (Parr R G and Ghosh S K 1995 Phys. Rev. A 51 3564) is adapted here to study, within time density functional theory, the linear response to external fields of (i) neutral and charged sodium clusters, and (ii) doped clusters of the type Na_nPb (n = 4, 6, 16). The resulting photoabsorption cross sections are compared to experimental results, when available, and to results from previous calculations using local and non-local xc functionals. The calculated static polarizabilities and plasmon frequencies are closer to the experimental values than previous results.

Hole band structures of p-doped semiconductor heterostructures are presented. The full six-band Luttinger-Kohn Hamiltonian generalized to treat different materials is solved in conjunction with the Poisson equation in a plane-wave representation. Self-consistent solutions of the multiband effective-mass-Poisson equations are obtained for unstrained and biaxially strained zinc-blende GaN/In_xGa_1-xN and GaAs/In_xGa_1-xAs quantum wells and superlattices (SLs), in which the acceptor doping concentration and its profile, the SL period, and the alloy content x are varied. The particular features observed in the valence subband structure of GaN/InGaN systems are stressed in a comparison with other selected In-derived III-V heterostructures, such as GaAs/InGaAs SLs.

Extensive bulk magnetization and ac susceptibility measurements have been performed over a wide temperature range on well characterized polycrystalline Ni₇₅Al₂₅ samples `prepared' in different states of site disorder. A detailed data analysis unambiguously establishes (i) the existence of multiplicative logarithmic corrections to the mean-field (MF) power laws in the asymptotic critical region near the ferromagnetic-paramagnetic phase transition and (ii) a gradual crossover to the Gaussian fixed point at temperatures outside the critical regime, irrespective of the degree of site disorder present. The latter crossover is followed by yet another crossover from Gaussian to pure MF regime in all the samples. Accurate determination of the universal amplitude ratio R_χ = DB^δ-1Γ, the asymptotic critical exponents β, γ and δ and the logarithmic correction exponents x^-, x⁺ and x⁰ for spontaneous magnetization, initial susceptibility and the magnetization versus field isotherm at the Curie temperature T_C, coupled with the observations made on the same system previously, not only rules out completely the possibility of isotropic short-range Heisenberg or isotropic long-range dipolar or uniaxial dipolar asymptotic critical behaviour in Ni₇₅Al₂₅ but also indicates strongly that, in the asymptotic critical region, the weak itinerant-electron ferromagnet Ni₇₅Al₂₅ behaves as an isotropicd = 3, n = 3 ferromagnet in which the attractive interactions between magnetic moments decay with intermoment distance (r) as J(r)~1/r^(3/2)d, and that site disorder is irrelevant in the renormalization group sense.

A high-pressure Raman study of microcrystalline tungsten oxide was performed in the 0.1 MPa-30 GPa pressure range under hydrostatic and non-hydrostatic conditions. Two phase transitions are evidenced; they take place below 0.1 GPa and at about 22 GPa and are of first order. Two spectral anomalies are observed at about 3 and 10 GPa; they may be related to diffuse weak structural transitions. The number of observed Raman bands remains practically unchanged in the 0.1-30 GPa range and thus the symmetry changes are likely to be small. Surprisingly, the non-hydrostatic conditions do not induce inhomogeneous band broadening and do not modify the transition sequence observed in hydrostatic conditions. The compressibilities of the different observed phases are estimated from spectral data and discussed within Hazen's polyhedral approach.