Table of contents

The structure of the liquid-vapour interface of Na-Cs alloys has been determined from Monte-Carlo simulations.

The electronic structure of amorphous Cu_xZr_1-x (x=0.65, 0.50 and 0.35) alloys is calculated for realistic, structurally relaxed models by using the linear muffin-tin orbital method in the atomic sphere approximation and the recursion method. The orbitals of the Cu 3d and 4s and the Zr 4d and 5s electrons are included in the calculation and the electronic charge density is determined self-consistently. The hybridised mixing between the Cu 3d and Zr 4d states can be seen and the calculated density of states agrees well with the observed XPS and UPS data. The observed change of the Zr-Zr distance depending on the Cu contents is attributed to charge transfer into the Zr 4d band.

The electronic Gruneisen constant gamma _e in Nb is calculated on the basis of a Slater-Koster interpolation of augmented-plane-wave X alpha bands. Dependence of gamma _e on the electron concentration is estimated in the rigid band model and the results are compared with the experimental results in Zr-Nb-Mo-Re alloys.

A new mechanism is proposed for the quenching of phonon drag in the electrical resistivity of the alkali metals. The mechanism is based on the anisotropy of electron-dislocation scattering, and is thus appropriate to strained samples. The theory is applied to the low-temperature resistivity data for strained samples of potassium.

The de Haas-van Alphen (DHVA) effect has been investigated in Pd and Pd-Ni alloys in magnetic fields extending to 40 T. With increasing Ni content, the incipient ferromagnetism gives rise to an unusual behaviour which is dominated by a field dependence of the Stoner factor.

Giant AC susceptibility anomalies from 30 to 35K, and 'shoulder' structures at lower temperatures are observed in amorphous Zr_100-xCu_x (x<or approximately=35) alloys diluted with only 4 at% Gd. Similar behaviour is also seen in Zr-RhGd alloys. On the other hand, amorphous Mg-ZnGd (antiferromagnetic) and Y-AlGd (ferromagnetic) alloys containing up to 8 at% Gd are spin glasses or cluster glasses typically found in other amorphous Gd alloys. To explain these results, the role of an indirect exchange interaction via the d band is emphasised.

The trapping of He in bubbles produced by implantation of Al films and the dependence of the pressure-shifted He 1¹S₀-2¹P₁ transition on the He concentration C_He have been studied by electron energy-loss spectroscopy, transmission electron microscopy, and proton-enhanced scattering. Application of an equation of state for super-dense He allows the evaluation of the dependence of the bubble concentration C_B on C_He. The authors find that bubble nucleation takes place predominantly for C_He<or approximately=6 at%. In excess of 6 at% C_B decreases rapidly with increasing C_He indicating the onset of bubble migration and coalescence.

A structure study of the liquid Pd_0.825Si_0.175 alloy has been performed by X-ray and neutron diffraction. As the contribution of the Si pairs to the scattered intensity is very small, the partial interference functions have been evaluated for the Pd-Pd and Pd-Si atomic pairs and compared with those obtained from a hard sphere model. The partial interference function J_Pd-Pd is very similar to that deduced from the Percus-Yevick calculation but the agreement is not as good for the Pd-Si pairs. This may be due to the fact that the chosen model only takes into account the size effect between the particles and does not include their chemical interaction.

The nature of the inhomogeneities existing in a computer-built model amorphous metal has been investigated by analysing the spatial correlation of various atomic level parameters. It was found that (i) clustering of atomic internal stresses is due mostly to the statistical fluctuation of the random distribution of the atomic level parameter; (ii) spatial fluctuations in the atomic shear modulus, the atomic volume and the geometrical coordination number extend over a longer range (several atomic distances) than was expected from the random distribution; and (iii) the spatial fluctuations of the shear modulus and the atomic volume are quite variable while the stress fluctuation is practically unaltered for the structural change.

Measurements of the low-temperature specific heat, in different magnetic fields, are presented for the metallic glass Cu_0.46Zr_0.54 in the as-quenched and fully relaxed states. The data are analysed in terms of the theory of type II superconductivity and the variations of the relevant parameters are discussed. Structural relaxation changes the electron density of states at the Fermi level, the Debye temperature and the average of the square of the electron-phonon matrix elements in the opposite direction to that of crystallisation. The most strongly affected quantity is the electron-phonon interaction. In the two different glassy states, fluctuation contributions to the specific heat are observed in the critical region which exhibit contrasting behaviour in an applied magnetic field just below the transition points. An explanation is suggested in terms of a partial suppression by the structural relaxation of structural defects of the type and size of those which have been identified recently in glassy metals by a computer simulation.

Doppler-broadening measurements of positron annihilation in quenched single-crystal Al-1.5*10^-2at%Si show a reversible temperature dependence of the annihilation characteristics which varies according to the defect state of the crystal. This effect seems to be largely able to account for an apparent discrepancy between data for quenched samples measured at 78K and at higher temperatures.

The martensitic phase transition onset is studied for noble metal binary beta brasses. The transformation mechanism, which is electronic, is determined by electron-phonon coupling in such metals. The composition dependence of structure stability is investigated, together with the connected role of ternary additions to the alloys. The donor-like and acceptor-like electronic behaviour is tested for a number of noble metal beta brasses. The results are in very good agreement with the available experimental data.

The single-crystal elastic moduli of HoCo₂ have been determined using ultrasonic velocity techniques in the temperature range 4-200K, and in magnetic fields of up to 6 T. The observed properties can be satisfactorily interpreted in terms of crystal-field effects acting on the Ho³⁺ ion.

The Fermi surface of NiTi, which has the B2 structure, was calculated. The results were used to find the nesting conditions epsilon _i(k)=- epsilon _i(k+Q). A theory of phase transitions in NiTi based on the temperature-dependent Green function method is suggested. The theory agrees satisfactorily with experimental results on the observed martensitic transformations.

Using a simple nearest-neighbour model, published measurements of the thermodynamic activity and the formation enthalpy of CoG are fitted. In combination with vacancy measurements each fit yields a set of four interaction energies, which agree very well. The interaction energy between two gallium nearest neighbours is estimated by use of the Miedema model (1979, 1980) of cohesion in alloys. Knowing all the interaction energies in CoGa the vacancy concentration on the Ga sublattice and the Ga atom concentration of the Co sublattice can be calculated. This leads to very reasonable values. By comparing triple-defect disorder with substitutional disorder it is shown that the interaction-energy parameter set is able to account for the special defect structure of CoGa.

A comprehensive description of the thermodynamic properties of liquid Na-Cs is presented, all derived from the EMF, E(c,T,p), of a concentration cell in which sodium beta alumina was the separator. The alloy heat capacity was found to vary linearly between the pure component values, and the small departures of the expansivity and compressibility from ideal behaviour are attributable to the volume contraction on mixing. There is no evidence for the existence of Na₂Cs clusters in the liquid in either the heat capacity data or the variation of the correlation function S_cc(0) with concentration.

The principal methods of formulating the single site approximations (both average t matrix and coherent potential approximations) for the equilibrium properties of disordered alloys are discussed. It is shown that the recent method based on the use of a projection operator can be used to derive the other formulations simply and more generally.

The recently introduced projection technique in the theory of alloys is used to develop in a simple way the cluster expansions of the self-energy for an arbitrary reference medium. Three choices of the reference medium, self-consistent and non-self-consistent, are considered and pair site approximations (PSA) are developed. It is shown that the 'optical theorem' of the self-energy can be used to prove that for these PSA the sign of the density of states is not guaranteed to be non-negative.

X-ray measurements of the structure factors of paired reflections (330)-(411) and (442)-(600) have been performed on three Fe-Si alloys, having approximate compositions of Fe₃Si, Fe₈₆Si₁₄ and Fe₉₃Si₇. In addition an absolute measurement of the (110) atomic scattering factor of Fe₈₆Si₁₄ has been performed. The (110) scattering factor of Fe atoms in the 14 at% Si alloy is, within experimental uncertainty, the same as in pure Fe. The ratios of the structure factors of the paired reflections show that the charge density of all alloys is very nearly spherical, in contrast to that of Fe metal, whose charge density of all alloys is very nearly spherical, in contrast to that of Fe metal, whose charge density is preferentially contained in the t_2(g) sub-band. A comparison with existing band calculations for Fe₃Si has shown notable discrepancies between experiment and calculations. Combining the results of charge and spin density measurements, possible mechanisms responsible for the observed effects have been indicated.

Using the concept of the complex band structure the scattering of conduction electrons by an intrinsic stacking fault in Cu was investigated. In order to match the solutions of the two half-crystals at the fault a bivariational procedure was applied which made the reflection coefficient stationary. It was found that apart from neck states the Bloch waves could penetrate the stacking fault with a rather small probability of being scattered.

The potentialities of combined electrical resistivity measurements and quench experiments to follow the evolution of local order and infer atom jump frequencies in short-range ordering alloys are recalled. Application to two amorphous alloys, Cu₅₀Ti₅₀ and Ni₃₅Ti₆₅, is presented. In either case, a reversible variation of the resistivity with quench temperature could be discriminated from an irreversible evolution linked to the metastability of these materials with respect to the 'ideal' glass. The reversible term is thought to be reflection of the thermodynamic equilibrium state of chemical short-range order. This effect is discussed in the light of recent neutron diffraction and EXAFS studies, which have produced clear evidence that the metal atoms have a preference for unlike neighbours in both alloy glasses considered. From the observed pattern of the resistivity ordering curves, the existence of ordering domains each characterised by a specific relaxation rate and evolution degree of short-range order, was inferred. Atom jump rates were also derived.

Pulsed field de Haas-van Alphen measurements of the Gamma centred part of the Fermi surface in ZrZn₂ are presented for fields up to 35T. The spin-split Fermi surface was found to be field dependent. It is argued that the observed de Haas-van Alphen frequency is different from the frequency defined by the Onsager relation. An expression is derived based on the Stoner-Wohlfarth model which describes the difference in spin-up and spin-down frequencies as a function of field strength. The observed difference in the effective mass for the two spin directions can also be explained by this model and it compares well with band structure calculations. An anomalous behaviour of the second harmonic of the majority surface de Haas-van Alphen frequency was observed. A search for other frequencies was carried out but in spite of the high sensitivity, in terms of m*T_D, no other orbits were seen.

Valence-band photoelectron spectra have been measured for alpha manganese over the range of excitation energy from 30 to 130 eV. The observed spectra show the two-peak structure in agreement with the reported partial density-of-states curve obtained by the tight-binding approximation, although no quantitative agreement is found regarding the peak positions and the width of the band. The 3p-3d resonance is observed at excitation energies above the 3p threshold. A satellite appears at 5.8 eV below the Fermi edge.

For pt.II see ibid., vol.11, p.1133 (1981). The semiclassical model of the relaxation of conduction electrons after the absorption of an X-ray photon in a metal is used to calculate the intensity of plasmon satellites accompanying core level lines in XPS. Within the Born approximation and with purely classical arguments the authors describe the wavefunctions of the hole and the photoelectron resulting from their interaction with the plasmons of the Fermi gas, over the whole energy range, including the destructive interference effect at small velocity. They critically question whether an XPS experiment may provide information on the intrinsic processes of screening. They then discuss the validity of describing the photoelectron as a point particle or as a wave packet, and show that these concepts may only be used when the velocity of the photoelectron is large in comparison to the Fermi energy.

The structural transformation has been investigated through resistivity measurements on the ternary systems V₅₀Ta_xRu_50-x and V_50-xTa_xRu₅₀, together with the binary system V_50+xRu_50-x. The transition temperature T_L drops sharply with an increase of concentration x which gives rise to a decrease of the electron-atom ratio e/a. On the other hand, T_L does not vary with x in the system V_50-xTa_xRu₅₀, where e/a is independent of x. These results may suggest that the structural transformation is electronically driven. A phenomenological expression to describe the electronic transition can well explain the resistivity anomalies which are caused by the structural transformation.

The electrical resistance of metallic glasses was measured as a function of temperature between 1.5 and 300K under pressures of up to 12 GPa (120 kbar). Depending on the system, the resistivity is either found to increase (Mg₇₀Zn₃₀) or to decrease (Cu₅₇Zr₄₃, Ti₅₀Be₄₀Zr₁₀). This behaviour is in contrast to the negative temperature coefficient of resistivity which all the amorphous alloys considered here exhibit. The results are compared with theoretical predictions.

The electrical resistivity rho and the thermoelectric power S of liquid Zn_1-cSb_c alloys have been measured as a function of temperature and composition up to 700 degrees C. The resistivity isotherms present a neat maximum in the middle of the phase diagram, whereas the thermoelectric power varies in a sigmoid fashion, with a change of sign around c=0.3. Comparison is made between the experimental results and the theoretical predictions of the nearly free electron model. The theory is unable to reproduce the essential features of the concentration dependence of rho and S, even when use of structural geometrical models is made, to take account of ordering effects. It is concluded that the behaviour of the structure factors alone is probably unable to explain the experimental results and that at least some electronic rearrangement, versus concentration, should be included in the theory.

An analysis is presented of the low-temperature electrical resistivity of the noble metals. The deviations from Matthiessen's rule (DMR) are calculated by means of a variational solution to the Boltzmann equation, employing an electron distribution function that contains both an angular dependence and an energy dependence. Quantitative agreement with experiment is obtained for all the noble metals. The temperature dependence of the electrical resistivity is also calculated. It is shown that including the electron-electron scattering term provides the explanation for the unexpected T⁴ behaviour observed recently for the resistivity of all the noble metals in the temperature range 2<or approximately=T<or approximately=6K.

For pt.I see ibid., vol.12, no.72, p.2985 (1982). An analysis is presented of the effect of electron-dislocation scattering on the temperature-dependent part of the electrical resistivity of the noble metals at low temperatures. The basic idea is that for any given sample, the degree to which the relaxation time varies over the Fermi surface is determined by the 'competition' between small-angle electron-dislocation scattering and large-angle electron-impurity scattering. Detailed calculations of the low-temperature resistivity are carried out, based on the variational method, that yield quantitative agreement with the recent resistivity data for Ag and Cu. The analysis also leads to the following results. (i) For the ultra-pure samples, the residual dislocations are shown to eliminate the T⁴ dependence that characterises the resistivity of the less pure samples. (ii) Electron-dislocation scattering is shown to produce a variety of possible deviations from Matthiessen's rule (DMR), including large negative DMR, small positive DMR or even a complete absence of DMR, in agreement with the resistivity data for strained samples of Ag and Al. (iii) Dislocations are shown to increase the magnitude of the electron-electron scattering term for the noble metals, in agreement with the recent Kaveh-Wiser theory (1980-1982).

A general procedure is given for calculating the low-temperature magnetoresistivity of impurities in metals taking into account both the Bloch wave character of the electronic states and multiple scattering effects. The linearised Boltzmann equation is solved for anisotropic Fermi surfaces by expanding the distribution function in sets of polynomials of the Cartesian components of the electronic velocity orthonormalised on the Fermi surface (Fermi surface harmonics). This is valid for arbitrary shapes of the Fermi surface and for all magnetic field strength with the exception of the quantum limit.

The ultrasonic attenuation due to dislocations was studied in the range 3.6-30 MHz in a single crystal of tin doped with 0.0016 at% indium at 1.3K in the normal and superconducting states. The attenuation at various pulse amplitudes shows peaks in both states. The attenuation peaks due to the breakaway of the dislocations from minor pinning points were used to calculate the dislocation resonance frequency. The breakaway stress is found to decrease with increasing frequency in both states. The attenuation shows resonance-type behaviour as a function of frequency. In the superconducting state unpinning starts at a lower pulse amplitude and there is a systematic decrease of the resonance frequency with increasing amplitude. In the normal state there is evidence of unpinning at higher values of pulse amplitude. The nature of the variation of attenuation with amplitude and frequency in both the states has been explained on the basis of Rogers theory (1962).

Within the alloy analogy approximation it is shown that the disordered local moments (DLM) can be formed inside the degenerate d band. The Coulomb correlation, if sufficiently strong, produces two sub-bands, in which the DLM phase can appear. The temperature-dependent susceptibility in the Hubbard model is obtained and the temperature stability of different phases discussed. The results of the Stoner model for susceptibility versus temperature dependence are critically evaluated. It is shown that the Curie-Weiss law holds only in very small regions around the band edges. From the zeros of the susceptibility denominator the transition temperatures between either ferromagnetic, antiferromagnetic or DLM phases and the paramagnetic phase are obtained. It is shown that the DLM phase is stable high above the Curie temperature in ferromagnetic iron but is unlikely to be stable above the Neel temperature in antiferromagnetic chromium.

For a sufficiently large correlation (u>or approximately=1), the authors show that the static susceptibility, specific heat, resistivity and thermoelectric power, calculated for the asymmetric Anderson model by means of the determinantal perturbation expansion of Yosida and Yamada (1975), all show universal behaviour in the low-temperature, low-energy region, characterised by a single scaling constant theta _r= Delta / pi chi up arrow up arrow sin ( delta _sigma ). As the asymmetry increases, the onset of the universality shifts towards larger u values. The thermoelectric power (TEP) takes the form S=S_N((T/ theta _r)-K_s(T/ theta _r)³) where K_s is a parameter-independent constant and S_N=(2 pi k_B/3 mod e mod )cos( pi Z_d/2(2l+1)) is the normalisation constant determined by the number Z_d of electrons in the virtual bound state (VBS). Thus, the TEP changes sign as one moves across the 3d or 4f series. Comparison with the experimentally obtained universal TEP curve shows our results to be valid up to T approximately= 0.2 theta _r. The Z_d values calculated from S_N as given by the TEP data are consistent with the values obtained by different experimental methods.

The magnetic properties of the intermetallic compound Ce(Fe_0.8Al_0.2)₂ are studied by means of magnetisation measurements as functions of temperature and field. The low-temperature magnetisation versus temperature curve shows a peak at 12.5K with remanent effects characteristic of spin-glass behaviour. At higher temperatures a second peak was observed at 165K. The nature of this second peak is qualitatively discussed.

The concentration dependence, temperature dependence and external field dependence of the ⁵⁷Fe Mossbauer spectra are reported in the mixed magnetic interaction system Fe_1-xMn_xBe₂. ⁵⁵Mn nuclear resonance data have been obtained for concentrations x<or=0.50. These results, together with dynamical susceptibility at selected concentrations, are combined with bulk magnetic measurements to elucidate the magnetic phases and transitions occurring in this system. The low-temperature (4.2K) hyperfine field distribution reveals a zone of applicability of the binomial distribution law in the Fe rich region (x<or=0.24) and a progressive increase of a low or zero field component with increasing Mn concentration. In combination with the effect of an external field and with susceptibility behaviour, this trend is attributed to increasing frustration effects. The temperature dependence of the hyperfine field distribution reveals a heterogeneous character of the magnetic ordering which increases with Mn concentration. Coexistence of significant localised and itinerant contributions to magnetism is demonstrated for these alloys. The resulting distribution of exchange interactions together with the instability of the Mn coupling produce severe frustration effects. The magnetic inhomogeneity is assigned to ferromagnetic and mictomagnetic clusters whose proportions vary sensitively with temperature and concentration.

Reports the results of magnetic measurements performed on U(Fe_xCo_1-x)₂ compounds in the temperature range 4.2-500K. For the Fe content x<0.62, the compounds show an exchange-enhanced paramagnetism. The magnetic susceptibility chi ₀, is not temperature dependent. The chi ₀ values increase with increasing Fe content. The compounds with x>0.62 are ferromagnetic. The saturation magnetisation decreases with increasing Co content. Above the Curie point, the data can be described by considering a Pauli paramagnetic term in addition to a Curie-Weiss behaviour. The composition dependence of the saturation magnetisation is analysed using the local environment model.

Magnetisation in the Pd₂MnSn Heusler alloy was observed to be drastically reduced by plastic deformation. The reduction is able to be explained in terms of the dislocation theory with the consideration of ferromagnetic and antiferromagnetic Mn-Mn interactions. Plastic deformation makes Mn atoms on the regular site substitute to Sn and Pd sites. The number of these Mn atoms is proportional to the dislocation density, rho , or its square root, and is estimated to be 10²⁰ cm^-3, when rho =10¹² cm^-2, which is in good agreement with the experimental results.

An explanation is given of the negative g shift in concentrated, bottlenecked systems under the assumption of positive exchange coupling between the localised spin sub-system and the band electrons. The effective spin Hamiltonian for d band electrons is derived, on the basis of which it is shown that the g factor for d band electrons differs from the free electron value. The effective g factor in Gd rich compounds is studied. The theoretical calculations are in good agreement with the experimental data.