Table of contents

The structural phase transitions occurring in the chain compounds (CH₃)₄NMnCl₃ (TMMC) and (CH₃)₄NCdCl₃ (TMCC) are studied through X-ray diffraction measurements. In both compounds, the existence of a new high-temperature orientationally disordered phase is established (phase I') with space group P6₃/mmc and Z=2 (parent phase). The room-temperature phase I of TMMC and TMCC (space group P6₃/m and Z=2) derives from I' by a rotation of the octahedra chains about the c axis; orientational disorder of the (CH₃)₄N⁺ group (TMA) in phase I is described by a complex Frenkel model involving at least five different orientations of the TMA. Phase II of TMMC (space group P2₁/b with Z=4) is derived from phase I by antiphase translational displacements of the octahedra chains along the c axis; an important amount of residual disorder of the TMA is observed just below the I to or from II transition temperature. In TMCC, the space group P2₁/m (Z=2) of phase III is confirmed, whereas phase IV exhibits a complex structure (space group P2₁/b with Z=12) corresponding to a trebling of the lattice constant along c and a doubling along b. X-ray diffuse scattering experiments on TMMC and TMCC are also reported. In phase I, the diffuse scattering patterns are interpreted in terms of linear translational disorder of the octahedra chains along the c axis. In phase II of TMMC, a residual translational disorder of the chains is evidenced, just below the I to or from II transition temperature. These results show that orientational disorder of the TMA group is coupled to translational disorder of the octahedra chain sublattice.

For pt.I see ibid., vol.2, p.8209 (1990). The structural phase transitions occurring in the chain compounds (CH₃)₄NMnCl₃ (TMMC) and (CH₃)₄NCdCl₃ (TMCC) are studied by means of Raman scattering and ultrasonic measurements. The Raman spectra recorded in the different structural modifications are fully understood on the basis of the structural data previously established in paper I. In addition, these data provide useful information on the transition mechanisms, which are of complex nature. Generally, these mechanisms involve order-disorder processes due to the reorientations of the (CH₃)₄N⁺ groups (TMA) coupled with a displacive contribution coming from the MCl₃ octahedra chains (rotatory and translatory soft modes). The ultrasonic measurements clearly show that the I to or from III phase transition (P6₃/m(Z=2) to or from P2₁/m(Z=2)) is pseudo-proper ferroelastic, as well as the I to or from II (P6₃/m(Z=2) to or from P2₁/b(Z=4)) transition; this latter should be of the trigger type, in order to account for the unit-cell doubling.

For pt.II see ibid., vol.2, p.8229 (1990). A phenomenological model is developed in the framework of Landau theory, in order to account for the structural phase transitions occurring in TMMC and TMCC. This model includes pseudo-spin coordinates attached to the orientation of tetramethylammonium groups that describe order-disorder processes, coupling terms with rotatory and translatory modes of the octahedra chains leading to displacive contributions and coupling terms with the strain components that account for the ferroelastic behaviour. Thus, the phase transitions between phases I', I, II and III are described in a satisfactory way, when compared with the experimental data. It is shown that the complex structure of phase IV must be seen as the result of a lock-in transition from a hypothetical incommensurate phase.

A model of solitary-wave excitation for B-DNA has been proposed by taking into account the stretching of the hydrogen bonds. Using the continuum approximation the author has derived a nonlinear Klein-Gordon equation which has the solitary-wave solution. The stretch amplitude of the solitary wave is found to be 0.37 AA. The energy and width of the solitary wave are also estimated. It is shown by a linear stability analysis that the solitary wave is unstable. The numerical simulation of the nonlinear differential-difference equation (i.e. the equation without using the continuum approximation) shows that the solitary wave is pinned by the lattice. An unstable pinned solitary wave is used to explain the opening mechanism of the bases in a DNA chain.

A re-investigation of the room temperature phase of RbFeF₄ by single crystal and powder X-ray diffraction is presented. Using a Rietveld method for refinements, the space group is shown to be orthorhombic D_2h¹¹-Pmab instead of D₂³-P2₁2₁2 as proposed in the most recent work. As a result, the space group of the room temperature phase (phase IV) is not a subgroup of the space groups of the high temperature phases III (D_2h¹³-Pmmn) and II (D_4h⁵-P4/mbm). However, it obviously remains a subgroup of the highest temperature phase I (D_4h⁴-P4/mmm). A discondensation is deduced.

For pt.I see ibid., vol.2, p.8269 (1990). The layer compound RbFeF₄, which undergoes structural phase transitions at 923 K, 416 K (second order) and 380 K (first order), has been studied by polarized Raman scattering (under the microscope) in the temperature range 473 K to 143 K. The Raman spectra were analysed with the help of a calculation of the phonon spectrum in the higher symmetry phase (phase I, P4/mmm) and the determination of symmetry compatibility with subgroups. The spectra collected in phase II (923-416 K) are consistent with the P4/mbm, Z=2, symmetry resulting from the condensation of the M₃ mode. Very little changes are observed in the phase III Raman spectra, which can be explained with the Pmmn, Z=4 space group induced by the additional condensation of X₃ modes Below 380 K (phase IV) the Raman spectra exhibit drastic changes which are consistent with the Pmab, Z=4, symmetry. The transition has to be imputed to the condensation of one component of the twice degenerate soft mode M₉, maintaining the condensation of one of the components of the already condensed X₃ mode, while M₃ is no longer condensed. Pmab is not a Pmmn subgroup, which corroborates the first-order character of the III-IV transition. The sequence of phases in RbFeF₄ can be schematized as a⁰a⁰c⁰ to a⁰a⁰c⁺ to a_p⁺b_p⁺c⁺ to a_p⁺b_p^-c⁰. It is also shown that the FeF₆ sheets can be characterize by several internal mode frequencies.

The effect of the nitrogen vacancy on the bonding in the VN_x crystal is investigated by means of the CNDO/2 cluster calculations. It is shown that the existence of the N vacancy leads to local lattice relaxation and rearrangement of the interatomic bonds in its neighbourhood. One of the most interesting results is the appearance of strong V-V bonds through the vacancy site. The presence of the vacancy leads to the lowering of the total energy of the cluster and stabilization of the defect phase.

A scheme for a fast decomposition of the total density of states evaluated by the recursion method is presented. It is suggested that all the projected densities can be obtained in a single recursion calculation instead of applying the complete recursion procedure to each projection separately. The proposed scheme appears to be exact for the correct choice of the initial vector of the recursion transformation. It is shown that when the random initial vector approach to the total density of states is implemented, good results can be obtained, in particular for calculating the partial densities of states. The applicability and efficiency of the proposed method are demonstrated on two model systems. A simple criterion of reliability is formulated. Possible ways to eliminate spurious effects caused by the random initial vector approach are discussed.

A method of studying strongly coupled orbital triplet ions forming part of a trigonal tetrahedral complex is presented. It is based on the unitary transformation and energy minimization method described previously by Dunn and Bates (1989) for regular tetrahedral clusters, and it is assumed that the ion is coupled to the e-modes and to one set of the t₂-modes of the trigonal cluster. The coordinates of the 13 minima in the potential surface are found and they are related to the tetragonal, trigonal and orthorhombic minima of the regular cluster. Results are presented for both the radial and transverse t₂-modes. The ground states located in each of the minima are also deduced. These calculations form the basis for the evaluation of Jahn-Teller reduction factors for orbital triplet ions in such environments.

The nonlinear conductivity in orthorhombic and monoclinic TaS₃ has been measured as a function of temperature from the Peierls transition temperature down to 4.2 K in the electric field range 10^-2-500 V cm^-1. At liquid-helium temperature a sharp increase in conductivity is measured in a very limited variation range of voltages. A general description based on phase-slip processes is given to explain these nonlinear properties in the whole temperature range.

The energy level structure of a quasi-two-dimensional electron system in a parabolic quantum well with the magnetic field tilted with respect to the sample plane was investigated. Based on this model and on the linear response theory the analytical expressions for the magnetoconductivity tensor were derived. Owing to the anisotropy induced by the tilted field, two diagonal components of conductivity are no longer equivalent. In high mobility samples, precisely quantized Hall plateau develop in the off-diagonal component of conductivity regardless of field orientation.

The temperature and magnetic field dependence of the width intensity and position in magnetic field of the derivative of the nonresonant microwave absorption has been measured in a granular sample of YBa₂Cu₃O_{7- delta} . Measurements were made on a single sample in order to keep constant the effect of sample microstructure. They were made with appropriate control of the many parameters known to influence the behaviour of the absorption. The measurements reveal new behaviour. The hysteresis is shown to display a time dependence after the removal of a magnetic field attributed to the decay of the magnetization. The theory of Xia and Stroud (1989), for the microwave response of a superconducting loop in a magnetic field formed by Josephson links, is extended to the case of a granular composite where there is a distribution of loop areas and orientations with respect to the magnetic field. Assuming that the distribution of loop areas is determined by a Boltzmann function of the energy of the current loop in a magnetic field, it is possible to completely account for the line shape and the temperature dependence of the intensity. The temperature and magnetic field dependence of the field position and width of the line are concluded to be a result of the temperature and field dependence of the flux trapping properties of the superconductor. It is shown that the microwave absorption can be used to measure the flux trapping behaviour of the material, such as the decay of the magnetization.

Soliton excitations in a one-dimensional Heisenberg ferromagnet are studied by means of the Holstein-Primakoff representation. Writing the Hamiltonian and the equation of motion into dimensionless forms, the authors show that the relative ratio of epsilon to eta is important for the determination of the modified terms of the nonlinear Schrodinger equation; epsilon =1/ square root S (S is the spin length) is the small dimensionless parameter used in the semiclassical approximation and eta =a/ lambda ₀ (a is the lattice space and lambda ₀ is the characteristic wavelength of the excitations) is another small dimensionless parameter used in the long-wave approximation. The soliton solutions are given in three cases ( eta =O( epsilon ), eta =O( epsilon ³2/) and eta =O( epsilon ²)) which correspond to the different physical conditions of the system. The results obtained by Pushkarov and Pushkarov (1977), de Azevedo et al. (1982), and Skrinjar et al. (1989) are included in this approach.

A new formalism for the static paramagnetic spin susceptibility of metals is given which includes a directional dependence on the magnetic response. This approach leads to an anisotropic generalized Stoner condition. The authors applied this theory to HCP and FCC cobalt. The starting point was the fully relativistic band structures and densities of states of both phases. The calculated susceptibilities show that an anisotropy appears in the HCP phase, but is absent for the FCC phase within this linear response theory. They also find a temperature dependence of this anisotropy. At low temperatures (<2500 K) the easy axis is along the c axis whereas at higher temperatures it lies in the ab plane.

The magnetic structure of the weak ferromagnet ludlamite, Fe₃(PO₄)₂.4H₂O, has been re-examined by both polarized and unpolarized neutron diffraction from single crystal samples of the mineral. The crystal structure has been refined from three dimensional data collected at 20 K, the unit cell being monoclinic, a=10.542(8), b=4.650(2), c=9.278(7) AA, beta =100.75(7) degrees at this temperature. At 4.2 K the contributions to the weak b-axis ferromagnetic moment produced by an external field of 1.5 T parallel to (010), determined from (h0l) polarized beam flipping ratios, are 1.73(3) and 0.62(2) mu _B for the Fe²⁺ ions at the centre of symmetry and in a general position, respectively. Limited three-dimensional unpolarized data collected at 4.0 K were used to refine the complete magnetic structure. The moments on the closely linked triads of ions are predominantly ferromagnetically coupled, with the inter-moment angle of 26(2) degrees . The total moment at 4 K and in zero field is best fitted by the values: Fe1, 4.51(6) mu _B at angles Theta =30(2) and phi =48(2) degrees ; Fe2, 4.00(5) mu _B at Theta =6(2), phi =90(7) degrees , where Theta is measured from the c axis and phi from the a-c plane. The orientations and magnitudes of the ionic moments are related to the competition between exchange and spin-orbit coupling.

The authors report low-field (0-1200 Oe) magnetic properties of the CoCl₂-graphite intercalation compounds (GIC). Using high-quality stage-1, CoCl₂-GIC samples, they determine that there is only one low-temperature field-induced transition at H_t(0)=380 Oe in these compounds. They estimate that the interplanar exchange interaction J'=0.025 K, and hence the interlayer-to-intralayer exchange coupling ratio is mod J'/J mod approximately 10^-3 for the stage-1 CoCl₂-GIC. The H_t-T phase diagram for the stage-1 CoCl₂-GIC has been mapped out using both in-plane susceptibility chi ' and c-axis resistivity rho _c measurements. No field-induced transition in the susceptibility that could be ascribed to the presence of a sixfold in-plane anisotropy field H₆ was observed. Zero-field susceptibility chi = chi '-i chi " measurements on CoCl₂ intercalation compounds with larger c-axis repeat distances than the stage-1 compounds showed one maximum in the real component chi '(T) at T_max, and two maxima in the imaginary component chi "(T) at T_cl and T_cu, where T_cl<T_max<T_cu. Additionally, the magnitude of the chi '(T) maximum at T_max is approximately 10-15 times greater than the magnitude of the maxima in chi "(T). For the stage-2 compounds they obtain mod J'/J mod approximately 10^-4, representing an order of magnitude decrease with respect to the stage-1 compound.

Magneto-volume effects of weakly ferromagnetic metals are discussed, based on the assumption that the local spin fluctuation amplitude is almost constant. The author's result on the thermal expansion agrees with the result of Moriya and Usami (1980). He derives the universal relation of the ratio of thermal expansion above the Curie temperature and the spontaneous magnetostriction. Based on a simple thermodynamic argument, he also derives a relation connecting the magneto-volume coupling coefficient with the pressure dependence of the Curie temperature, which is slightly different from the Stoner-Wohlfarth theory.

Luminescence decay curves have been measured for the Eu³⁺ emission and Tb³⁺ emission in the cubic crystals Cs₂NaTb_1-xEu_xCl₆ at 80 and 293 K using pulsed dye laser excitation into the ⁵D₄ state of Tb³⁺ and the ⁵D₁ state of Eu³⁺. For excitation of Tb³⁺, efficient energy transfer occurs to the ⁵D₀ state (but not the ⁵D₁ state) of Eu³⁺ and there is fast energy migration within the ⁵D₄ state. When excitation is into the ⁵D₁ state of Eu³⁺, phonon-assisted coupling to the Tb³⁺ ions results in enhanced relaxation to the ⁵D₀ state but the lifetime of ⁵D₀ is not affected by the presence of Tb³⁺. In all cases the non-resonant energy transfer involves creation or annihilation of phonons which have odd parity with respect to the metal centers and cannot be explained within the usual Born-Oppenheimer treatment of phonon-assisted energy transfer. The mechanism of the energy transfer processes is considered in detail.

Neutron diffraction experiments over the momentum transfer range 0<Q<40 AA^-1, have been used to investigate the properties of liquid sulphur from 131 degrees C to 230 degrees C, i.e., at conditions around the lambda -transition (T_lambda =159 degrees C). The data have been transformed to give an accurate pair correlation function from which the absolute numbers of atoms in given molecular configurations may be determined. These data seem to be inconsistent with the picture of 'crown'-shaped S₈ molecules below T_lambda . Also it has been found that the diffraction pattern may be analysed as though the liquid was an assembly of roughly spherical molecular units. In this case each unit would contain about six atoms, and the lambda -transition could be related to the percolation limit for the units. It is concluded that a fundamental diffraction analysis of the states of sulphur has thrown new light on long-standing problems.

The structure of molten MOD (M=⁷Li, Na and K) at 773 K has been investigated by neutron diffraction measurement using a nickel metal cell. The intraionic O-D bond lengths were 97, 98 and 98 pm and the most probable interionic M-O distances were 194, 242 and 279 pm for ⁷LiOD, NaOD and KOD, respectively. The difference in the intraionic O-D distance among these melts was within the experimental error. The M-O distance depends linearly on the ionic radius of the cations; the OD^- ion possesses an effective ionic radius of 139 pm in the melts. Long ranged oscillatory character was observed in the Gⁿ(r) for ⁷LiOD but not for NaOD or KOD.

Observation of silica aerogels by scanning electron microscopy and transmission electron microscopy are presented. They are supplemented by information on pore volume and pore size distribution determined by thermoporometry. Comparison with small-angle neutron scattering allows precise characterization of the fractal domains. The authors identify three classes of structure. Primary chains, 30-50 AA in diameter, form the substructures of grains whose size varies from 100 to 500 AA depending on density. The grains form large-scale aggregates, leading to pores on scales exceeding 1000 AA. Since the system is demonstrated to be fractal between the chain and grain dimensions, they re-examine the phonon-fracton crossover and the fracton dimensionality determined by Brillouin scattering and Raman scattering.

A simple approach based on a statistical mechanical model is used to discuss the nature of atomic ordering in Cd-based (Cd-Bi, Cd-Mg and Cd-Ga) liquid alloys in the framework of the chemical short-range order parameter, concentration fluctuation in the the long-wavelength limit and the excess free energy of mixing. It has been observed that the nature of atomic order in binary liquid alloys could be understood with the help of their melting curves.

The possibility of modelling the partial structure factors of the Ni₃₃Y₆₇ amorphous alloy with simple purely repulsive potentials is investigated by extending and refining an approach previously proposed for the Ni_xTi_1-x alloys. Non-additive hard- and soft-sphere models are carefully studied by using integral equations originally derived in the theory of liquids. The quality of the approximations used and the consequences of softening the repulsion have been examined. The overall agreement between model and experimental data allows a meaningful discussion of the role of different regions of the interaction potential in the real system.

The free energy and the static elastic constants of a hard-sphere glass are calculated using density functional theory. More specifically, Bennett's glass model is used where the packing fraction of random closed packing, eta _RCP, is taken as an adjustable parameter. As regards the density functional, the modified weighted density approximation (MWDA) is employed. The results are compared with the FCC hard-sphere crystal. It is found that the free energy of the glass is always higher than the crystal free energy; however, for high densities and high eta _RCP, it is lower than the free energy of a fluid resulting in a first-order liquid-to-glass phase transition. For the same temperature and densities, the bulk modulus in the glass is higher than that in the FCC crystal. For high eta _RCP and high densities, the shear modulus is smaller than that of a polycrystalline sample.

The scaling equations which determine the master functions for the beta -relaxation process within the mode coupling theory of supercooled liquids are reformulated such that a direct numerical solution becomes possible. Tables are presented that allow an accurate derivation of the scaling function and susceptibilities by elementary manipulations. As an example of their application the beta -peak scenario as predicted by the theory is demonstrated.

The viscoelastic properties of monolayers of n-pentadecanoic acid at the air-water interface have been studied using surface light scattering. The monolayers displayed first-order liquid-condensed to liquid-expanded transitions: the surface properties reported are those of well defined, reproducible monolayer states. Two viscoelastic moduli were determined: for shear transverse to the surface and for uniaxial dilation in the surface. For experiments both above and below the triple point the surface viscoelasticity was found to depend upon the monolayer phase in a complex fashion. Both moduli displayed viscoelastic relaxation, the strength and timescales being different for the two moduli, as well as depending upon the monolayer state. At one point in the phase diagram it was shown that the high-frequency dilational viscosity measured by light scattering could be identified with the conventional surface shear viscosity. The results confirm the complexity of the viscoelastic behaviour of molecular films which is not adequately represented by a single surface viscosity.

The equiatomic alkali-group-IV (Si,Ge,Sn,Pb) compounds usually crystallize in the NaPb structure. This structure contains negatively charged tetrahedra of group-IV ions, which are separated by alkali ions. The Li compounds, however, have either a CsCl (LiSn,LiPb) structure or a structure with a three-dimensional network of group-IV atoms, e.g., LiGe. A similar difference between Li compounds on the one hand, and Na, K, Rb, Cs compounds on the other hand, has been found in the corresponding liquid alloys. The author explains this difference in structures in terms of two factors: the existence of a gap in the density of states at the Fermi level in the phase with tetrahedra, and the radius of the alkali atom. When the alkali ion is too small, like Li⁺, to separate the tetrahedra, then either the CsCl structure is the most stable, or a threefold-coordinated anion network will be the most stable structure. The latter structure occurs when the group-IV anions have a strong covalent interaction like for Ge and Si.

A set of integral equations for determining the liquid and electronic ion structures in a liquid metal have been derived by treating a liquid metal as a nucleus-electron mixture. By these integral equations with the atomic number as the only input, an electronic excited state of an ion in a liquid metal can be determined in a self-consistent way with the ion and valence-electron density distributions around it. It is shown that the authors' integral equations lead to a liquid-state version of the spherical-solid model proposed by Almbladh and von Barth to treat the spectroscopic problem in a solid. The integral equations combined with Slater's transition state method are applied to evaluate the K-edge position of a liquid metallic lithium: the value 51.36 eV is obtained at 470 K, which is compared with the experimental one of 51.26 eV. Thus it is ascertained that their integral equations can determine precisely the electronic structure of an ion as well as the liquid structure without the use of any information other than the atomic number.

The authors consider the dissipation arising from the activated motion of a flux line lattice in the presence of dilute pinning centres. By regarding this process as an effective motion of the pins relative to the flux lattice, they map the problem into the diffusion process of a particle interacting with the flux line lattice. They solve this problem analytically using a novel microscopic diffusion theory. In particular, they show that anomalous behaviour of the linear resistivity is expected at the flux lattice melting, and may show up as a dip in its temperature dependence.

For the first time both R=0 and diamagnetic transitions have been observed together for the ErBaCuO system for T>200 K, giving credence to earlier global reports of extra-high T_c. The problems of unstable aspects and rare reproducibility are believed to be linked with the small relative volume fraction of the extra-high-T_c entity estimated in the present study. The material considerations for achieving extra-high T_c are also indicated.

A calculation of the crystal field parameters for Ho³⁺ and Er³⁺ impurities in metallic copper is reported. The results can be understood in terms of the hybridization of d-s band states around the Fermi energy of the metal host. The possibility that the model can be used for the lanthanide series in noble metals, including those for which no experimental information is available, is discussed.

Near T_c=846 K barium sodium niobate exhibits a strong thermal focusing of laser light polarized along its optic c axis (focal length F=+2 to +3 cm for 1 W in a 0.45 mm beam waist). The beam convergence theta (T) is proportional to K^-1(T)dn_c(T)/dT, where K is the thermal conductivity and n_c is the c-axis index of refraction. For zero applied field, theta (T) far from T_c is dominated by the isotropic dn/dT term, which varies as t^{2 beta -1}, where t is the reduced temperature and beta is the coexistence curve exponent (approximately ¹/₄ in this tricritical system). However, within a degree or two of T_c, theta (T) is dominated by K(T), which becomes highly anisotropic and has a cusp-like dip for thermal diffusion along the polar c axis. In the present experiment the authors apply a DC electric field along c and show that a few kV cm^-1 is sufficient to suppress the cusp.