Table of contents

Simple model theories have been applied to the thermodynamic properties of liquid 3d transition metals. The one-component plasma model approach of Itami and Shimoji (1984) gives a better description of the thermodynamic quantities than the Percus-Yevick phonon model. For a fully quantitative discussion, however, information on the density of states at the Fermi energy and on its density dependence is essential. Similarly, accurate experimental values are indispensable.

Martensitic transitions occur in Li and Na at low temperatures. In Li the structure produced in the 9R ( delta -Sm structure). For K the martensite clearly does not occur in bulk, but there is corroborative evidence from much transport, thermodynamic and structural data which would support an embryonic martensitic presence. There also exists optical microscope evidence that RB can undergo some conversion, but above all there exists neutron diffraction evidence for K that the system shows an appropriate two-phase character up to 90K at least, at the level of about 1 part in 10⁴. The model of a precursor martensitic state appears more consistent with experiment and theory than the CDW model supported by Overhauser and co-workers (1983).

The authors report results of measurements of the anisotropy of the high-field magnetoresistance in single crystals of potassium. They find no evidence for the presence of open orbits. This result casts doubt on the interpretation of the induced torque experiments of Coulter and Datars (1985) in terms of a charge-density wave ground state in potassium.

The authors have studied the photon-energy dependence of photoemission from the Shockley-type surface state on the (111) face of a Cu-6.8% Pd random alloy in the range 20<or=h nu <or=85 eV. Their measurements are in good quantitative agreement with first-principles KKRCPA calculations of the photocurrent. They conclude that such calculations are capable of reproducing most of the observed properties of the surface state.

Values of delta ln sigma / delta P, where sigma is the magnetisation per unit mass, for YFe₂ (ZrFe₂) of -8.4+or-0.4 (-6.3+or-0.3)*10^-4 kbar^-1 were obtained from forced magnetostriction measurements on polycrystalline samples at 4.2K in fields of up to 12 T. These values are shown to be in reasonable agreement with ab initio computer calculations of delta ln sigma / delta P but contradict earlier experimental work. The agreement between experiment and theory for delta ln sigma / delta P provides further evidence that YFe₂ and ZrFe₂ are ferrimagnetic.

For pt.I see ibid., vol.14, L.155 (1984). The time dependence of the magnetisation of a system of fine spherical cobalt particles in dipole interaction is computed by using an appropriate Monte Carlo algorithm. For weak interactions, with volume packing factor p up to about 10%, the time dependence follows an exponential law but with a corresponding particle volume v_eff which exceeds the actual volume v by an amount proportional to p. For strong interactions (p up to about 50%), the time dependence was found to be logarithmic due to the interaction-induced distribution of relaxation times. The corresponding time t_1/2 which elapses before the magnetisation falls to half its original value follows a Vogel-Fulcher law (ln t_1/2)^-1 approximately (T-T₀) with a characteristic temperature T₀, at high temperatures but deviates from this law at low temperatures so that the system never experiences any phase transitions. The temperature T₀ depends linearly on the volume packing factor, thus confirming that the Vogel-Fulcher law arises from short-range cooperative phenomena.

¹¹⁹Sb emission Mossbauer spectra are measured for very dilute, quenched Au-Sb alloys. Two distinct satellite lines observed after quenching and subsequent annealings were assigned to vacancy-solute pairs and vacancy-solute clusters, respectively. Their electronic state and behaviour are investigated during annealing. The effect of oxidation was also examined and separated from that of vacancy-solute interactions in the analysis.

The authors demonstrate, by calculation and discussion of experimental data, that a change in nearest-neighbour spacing is not sufficient of itself to account for changes in physical properties when metallic glasses are annealed. The authors examine the consequences of this result for several current models of structural relaxation.

The atomic structures of the FCC/FCC metallic superlattices Au/Ni and Cu/Ni were simulated by the molecular dynamics method, which Morse potentials were used as the interatomic potentials. In contrast to the other previous studies, the authors have treated here incoherent realistic structural models. The simulated structures were used to calculate the Young's modulus along the FCC (111) direction, Y_Y(111), and the biaxial modulus, Y_b(111). The biaxial moduli for Au/Ni and Cu/Ni were compared with the experimental results and it is shown that the enhancement of elastic moduli can be well reproduced qualitatively by the authors' incoherent models. The enhancement of elastic moduli can be ascribed to the existence of periodic strain in the multi-layered films.

Cd_1-xHg_x alloys (x=0.15 and 0.35) have been studied up to 8 GPa by the four-probe resistance and X-ray diffraction techniques. Cd_0.65Hg_0.35 shows a BCT to HCP transformation at 2 GPa while Cd_0.85Hg_0.15 does not show any transition up to 8 GPa. This is consistent with the structural trends expected from the pseudopotential treatment of Hafner and Heine (1983). The possible mechanism of BCT to HCP transformation is discussed.

The electronic structure of a simple model for (La, Ce, Th, U)Be₁₃ compounds is determined self-consistently within the density functional theory. The rare earth or actinide atom is considered as a spherically symmetric impurity embedded in a jellium. The width and position of f-localised states result naturally from the scattering of conduction electrons by the impurity potential. The results of the author's calculation for electronic specific heat, position and width of f-resonant states are compared with available theoretical results and experimental data.

Measurements of the resistivity rho of liquid Rb-Pb and Cs-Pb alloys as a function of composition and temperature are presented. Around the equiatomic composition, the resistivity reaches non-metallic values in both systems; equiatomic Cs-Pb ( rho =7000 mu Omega cm at T=620 degrees C) may be classified as a liquid semiconductor. This is explained in terms of the liquid structure of these alloys, in which the Pb ions presumably join in covalently bound clusters. With rising temperatures, these clusters disintegrate. As a consequence, the resistivity decreases rapidly. For the Cs-Pb system, a DSC (differential scanning calorimetry) survey of the liquidus temperature as a function of composition has been performed. A tentative Cs-Pb phase diagram is presented.

Ultrasonic quantum oscillations in pure Bi and semimetallic Bi_1-xSb_x alloys (x=0.012, 0.019, 0.033) have been investigated. Experiments were carried out in the temperature range 1.2-4.2K using longitudinal ultrasonic waves of frequencies up to 90 MHz propagated along the trigonal axis. Magnetic fields between 0.05 and 2.3 T were applied in the yz and xz planes. The angular dependence of the peak positions was studied. The Landau and spin quantum numbers of the oscillation peaks were obtained from the experimentally measured periods and peak positions of the oscillations. The experimental data indicate that, for most orientations of the magnetic field in the yz plane, the magnetic energy levels of electrons in the alloys studied can be described successfully by the ellipsoidal non-parabolic (ENP) model. The magnetic field dependence of the Fermi energy and the carrier density were calculated in the framework of the ENP model for electrons and the parabolic band for holes, using the previously determined band parameters. The magnetic field values at which the electrons in each pocket enter the quantum limit were deduced from the magnetic field dependence of the carrier density. The quantum limit fields obtained from calculations agree with the experimentally predicted values. It is concluded that within the magnetic field range studied the variation of the Fermi energy with magnetic field has a negligible effect on the period of ultrasonic quantum oscillations from light electrons.

A new method is introduced for treating the electronic structure of a solid surface by means of the linear muffin-tin orbital (LMTO) method with the most localised orbitals. This method is applicable to both the thin-film geometry and the semi-infinite crystal geometry. As a test of the method, calculations are performed in a simple surface dipole field model for the Cu (001) surface of a film of 15 atomic layers. The localised and resonance states are discussed in detail in terms of the bond orbitals and the canonical description is proposed.

The critical concentration required for the appearance of magnetic moments on nickel atoms deposited on a palladium substrate and covered with a thin palladium overlayer is calculated in the Ginzburg-Landau model. The effect of the palladium surface is characterised in this model by a surface extrapolation length lambda . The extrapolation length lambda is determined by comparing the calculated and observed critical concentrations for a Pd-Ni-Pd sandwich with 1% of Ni atoms covered by a palladium overlayer of four atomic layers. The negative value of lambda obtained by this method implies that the exchange enhancement of the palladium surface is stronger than in the bulk.

The non-linear part of the magnetisation of pure Pd and dilute PdNi alloys (up to 0.56 at.% Ni) was measured in semi-continuous fields up to 35 T. The Ni-doped Pd shows a negative curvature at high fields whilst for pure Pd a small positive curvature was found as in previous measurements. In the alloys the curvature becomes more pronounced for higher Ni concentrations. These effects and their relation to recent DHVA measurements are discussed in this article.

The authors have measured the magnetisation of Gd_68.9Y_31.1 single-crystal samples in the temperature range 230-170K. The material appears to order initially with a c-axis ferromagnetic moment at 213K while the basal plane is disordered. This persists down to 206K, whereupon the c-axis moment collapses into a helical antiferromagnetic structure. The antiferromagnetism is still present at 182K when the structure is thought to become complex before a simple canted ferromagnet is established at lower temperatures.

The temperature dependence of the AC susceptibility Fe-rich Fe-Zr amorphous alloys having mictomagnetic or spin-glasslike behaviour has been measured in a superposed DC field. The Fe-Zr amorphous alloys containing more than 8.6 at.% Zr clearly show sharp peaks at the Curie temperature T_C and the spin cluster freezing temperature T_sf. These two temperatures seem to coincide at a critical concentration for the Fe-7.6 at.% Zr alloy, which is considered to be the percolation limit for the long-range ferromagnetic interactions.

By magnetisation measurements the authors have studied the temperature dependence (1K<T<50K) of the susceptibility of La₃Al and La₃Ga alloys containing a low concentration of magnetic praseodymium ions. These alloys were prepared in both a crystalline state and an amorphous state by a rapid quenching method. The results are strongly different for La₆₈Pr₇Al₂₅ and La₆₉Pr₆Ga₂₅ crystalline alloys, the Van Vleck susceptibility being ten times larger for La₆₉Pr₆Ga₂₅. The different crystal field symmetries (cubic and hexagonal) explain this difference. Results are nearly identical in the two corresponding amorphous alloys indicating that the short range structural order of the crystal is poorly preserved around Pr in the amorphous state. A quadratic crystal field model (the Fert and Campbell model) gives a good fit with the amorphous susceptibility data.

The technique of radio-frequency collapse is used to determine the quadrupole splitting distributions of ferromagnetic amorphous alloys, Fe_xNi_75-xSi₁₀B₁₅ and Fe_xSi_90-xB₁₀. In both cases the concentration dependence of the average quadrupole splitting QS, determined by Mossbauer spectroscopy, suggests a change in short-range order as predicted by structural models.

3d to 4f core electron energy loss spectra (CEELS) are presented for Ce intermetallic compounds. A comparison with XAS data shows that the final state mixing is different in XAS and CEELS. A more complete relaxation towards the energetically more favourable f² final state is observed in CEELS. This increased relaxation may be attributed to non-dipole effects in CEELS and it appears to correlate with the hybridisation strength between f and conduction states.

A second-principles theory of the angle-resolved photoemission spectra from the bulk states of substitutionally disordered alloys is developed within the single-site approximation (SSA). The initial valence states are described by a new version of the tight-binding-based alloy theory which allows treatment of different constituent and band shapes exactly within the SSA. The optical matrix elements are governed by atomic-like dipole selection rules. The theory is applied to the evaluation of the normal unpolarised spectra from the (100) and (111) faces of single crystals of pure Cu and Cu-25% Ni and Cu-15% Pd random substitutional alloys.