Table of contents

The authors have made measurements between 2K and 16K of the thermopower S of pure gold samples, with iron impurity removed by oxidation. When the main features of the observed temperature dependence of S in pure Au, Ag and Cu are interpreted as a change with temperature of the dominant electron scattering mechanism, the outstanding result is that S for the ideally pure metal is strongly negative below 10K.

A new single-site effective medium method for liquid metals is proposed. The effective medium is determined by a generalised Wigner-Seitz condition. Numerical results for the density of states of liquid iron show the broadening of the atomic d resonance in good agreement with the bandstructure calculation.

It is shown that the presence of nonmagnetic inclusions in gadolinium crystals has a significant effect on the first anisotropy constant K₁, leading to a variation of easy direction with inclusion content. The observed concentration of hexagonal oxide platelets is sufficient to explain differences in the values of anisotropy constants and easy direction reported by various workers. The authors conclude that the most reliable measurements are those of Corner and Tanner (1976) and Smith et al. (1977) on characterised, high-purity gadolinium crystals produced by solid state electrotransport.

The authors have derived the free-energy functional of an extended Hubbard model Hamiltonian with interatomic interactions. They have shown that as far as the ferromagnetic transition is concerned universality holds and that introducing a short-range interaction between neighbours yields the same free-energy functional as for the simple Hubbard model if irrelevant terms are discarded.

Spin echo NMR data on stoichiometric and substituted samples of Ni₂MnSn and Pd₂MnSn (i) show that Mn on Sn sites is polarised antiferromagnetically, and (ii) give information on the conduction electron polarisation produced by the Mn moments. The results are in good agreement with a d resonance model applied in the preasymptotic range.

The theoretical image contrast from a small finite edge dislocation loop in an elastically isotropic medium has been studied in the two-beam theory of electron diffraction. Computer image simulation has been used to investigate the influence which image plane normal, diffracting vector, foil thickness and dislocation loop Burgers vector, orientation, position and size have on so-called black/white intensity distributions. The choice and range of the parameters investigated have been dictated by experimental work on radiation damage in BCC molybdenum published previously by the authors. The present results have indicated that a variety of images are possible for this simple elastic defect model but that useful image classifications can be made in terms of modulus g.b where g is the reciprocal lattice vector of the operating reflection and b the dislocation Burgers vector.

For pt.I see ibid., no.7, p.1359 (1977). Previously the authors presented theoretical images of small, edge dislocation loops in an isotropic BCC crystal. The results demonstrated the wide range of image contrast that is obtained as a function of loop geometry and location in the crystal and diffracting conditions. In the case of loops located within the layer structure, it was shown that image symmetry is a sensitive function of modulus g.b and loop orientation. Image contrast was classified into four basic types corresponding to (1) mod g.b mod =0 (loops inclined), (2) mod g.b mod =0 (loops approximately edge on), (3) mod g.b mod <or=1 and (4) mod g.b mod >3. The authors now explore the practical usefulness of this approach to the determination of Burgers vector directions of small loops in irradiated molybdenum and copper crystals. More specifically, a range of examples of experimental images in these two metals have been selected for comparison with the theoretical predictions.

An atomistic monovacancy model, specially designed for variational calculations, is presented for an FCC metal crystal. The total electron distribution is given as a superposition of several, differently centred, spherically symmetrical densities. The formation energy E_VF, expressed as a sum of electrostatic integrals and statistical approximations for exchange, correlation and kinetic energy is minimized by variation of the electron density function in the vacancy region. A simple method, based on series expansion of the densities involved, is suggested for the variational calculation. Restrictions on and span of the suggested method are discussed. Resulting formation energies, 1.21 eV for Cu and 1.09 eV for Ag, are in close agreement with experimental values. The resulting conduction electron density in the vacant cell is considerably lower than those values predicted by other electron theoretical vacancy models.

Peak counting rates in the angular correlation were measured as a function of temperature above room temperature in the solid and liquid phases of Cu and in the solid phase of Ni. The peak counting rates in Cu increased about 17% upon melting. Temperature dependence of the peak counting rates in the liquid phase shows a hump at 1150 degrees C. Two possible explanations were proposed for this anomaly. The data in the solid phases were numerically analysed by the trapping model in which it was assumed that only single vacancies and divacancies are effective. The data at lower temperatures, where the effect of divacancies is sufficiently small, were analysed by the conventional trapping model of single vacancies. The results of both analyses are consistent. The characteristic temperature T_c in the trapping model was defined. It was demonstrated that T_c is proportional to E_1V^F. The experimental values of T_c were plotted against E_1V^F for various FCC metals and linear relationship between E_1V^F and T_c was also experimentally found.

The contribution of phonons to the excess entropy of concentrated alloys A_cB_1-c in a semiempirical model is studied. Numerical results are presented for a number of alloys. It is shown that the variations of the local entropies with the concentration mainly determine the excess entropy and that the local environment effects are not negligible; discussion of these results in relation to the surrounded atom model is presented.

The authors examine, in a Hartree-Fock approximation, the stabilisation of the gamma and alpha phase of iron by transition impurities on the right and left (respectively) of the periodic table. They show that the electronic free energy resulting from the difference of the screening potentials due to the different behaviour of the alpha and gamma density of states is considerable and roughly proportional to Z, the impurity matrix excess charge.

For pt.III see ibid., vol.5, p.1849 (1975). Formulae, which include the full nonlocality of the pseudopotential, are derived for the calculation of the electrical resistivity and thermoelectric power of simple liquid metals. Each of these formulae contains a significant correction to that obtained from the nearly-free-electron theory of Ziman. This correction tends to increase the magnitudes of the electrical resistivity and thermoelectric power and becomes more important in going from a light metal to a heavier metal, due to the increase of the nonlocal strength of the pseudopotential. Also, the theoretical results for the electrical resistivity and thermoelectric power, obtained from the presently derived formulae using the energy-dependent model pseudopotential developed recently and liquid structure factors available in the literature, agree reasonably well with the experimental results for the metals of interest.

For pt.IV see ibid., vol.7, p.1439 (1977). The authors have applied the previously published results of an energy independent model pseudopotential calculation to evaluate the density of states effective masses, m*, at the Fermi level for the alkali metals in both solid and liquid phases. The results thus obtained are used to reduce the range of the uncertainties (i) in the existing theoretical values for m* and the electron-phonon coupling parameters and (ii) in the existing experimental values for the electronic specific-heat coefficients and the paramagnetic spin susceptibilities for the heavy solid alkali metals. Also, these results show that the change in the Fermi level density of states effective mass on melting is larger for the heavy alkali metals than for the lighter alkali metals.

The model potential theory of Heine and Abarenkov (1964) and Shaw (1968) has been applied to the calculation of the energy bandstructure and density-of-states for aluminium. The use of Shaw's optimised model potential parameters is found to produce energies which differ from those of the ab initio calculations by approximately 0.03 Ryd. The model potential procedure due to Bunyan (1977), however, leads to improved agreement with the ab initio calculations to within 0.01 Ryd, provided the potential is modelled to agree with that used in the ab initio calculation. The results indicate that basic bandstructure data for the simple metals can be obtained from the model potential theory to an accuracy comparable to that achieved in more complex first-principles calculations.

Grain boundary electrical resistivities are calculated on the assumption that the scattering is due primarily to the dislocations which constitute these boundaries. Dislocation densities in grain boundaries have been estimated using Frank's formula as modified to take account of lattice symmetry. Dislocation resistivities are calculated ab initio assuming resonance scattering at the Fermi surface, and measured dislocation resistivities are also used where these are available. The calculated grain boundary resistivities are in satisfactory agreement with experimental data for thirteen of the fourteen metals from seven subgroups of the periodic table, for which data is available.

The authors have measured the transverse magnetothermopower of a single crystal of Mo at liquid ⁴He temperatures with the heat current along a (110) direction and the magnetic field in a (110) plane. The results appear to contradict the predictions of the high-field semiclassical theory but a careful analysis shows this is not so. It is emphasised that the separation of diffusion and phonon drag components can be misleading since the temperature dependence of many nonthermoelectric terms will effect a variation in the thermopower which can be confused with phonon drag.

The distance dependence of the strength of the magnetic interaction between Mn atoms dissolved in Pd is proposed on the basis of the experimental results for the spontaneous magnetisation at T=0, the 'step' in the electrical resistivity and for magnetic entropy as deduced from specific heat. Using this interaction a Monte-Carlo calculation of the specific heat and other properties has been carried out and compared with the experimental data.

Detailed studies of low-field susceptibility chi (T), thermoremanence and isothermal remanence in gold-iron alloys containing 0.25, 2, 4, 7at% Fe are presented. The susceptibilities show sharp but not discontinuous reversals of slope at T_f. Below the maximum in chi (T), the zero-field cooled susceptibility shows a marked shoulder at 0.6T_f which coincides with a maximum in the incremental isothermal remanence as a function of temperature. Below T_f the total magnetisation decreases slowly, flattening out to a constant value of approximately 0.9M(T_f). Thermoremanence increases almost linearly with decreasing temperature from zero at T_f. Both total magnetisation M(T)/M(T_f) and TRM/M(T_f) follow functions of temperature which are almost independent of composition. The results suggest a finite susceptibility at T=0, approximately 0.35 chi (T_f) indicating the presence of a reversible component to the magnetisation at absolute zero. Analogies with the behaviour of rock magnetism are indicated and the variation of TRM is analysed in terms of a spectrum of blocking temperatures deduced from isothermal remanence experiments.

Starting from electronic APW wavefunctions the author derives the expression of the direct Coulombic contribution of the conduction electrons to the crystalline electric field in rare earth intermetallics. The results on the compounds DyRh, DyCu, LaAg, DyZn and YMg explain partially the variation and the amplitude of the fourth-order CEF parameters in the corresponding series.

The substitution of V and Cr for Ga in the equiatomic alloy CoGa results in the appearance of ferromagnetism beyond a critical concentration which appears to be related to an average electron-atom ratio rather than to any obvious feature of the crystallographic ordering. Neutron diffraction results are consistent with a model in which at low concentrations the alloys order with the B2 structure but as the transition metal concentration increases double ordering of the L2₁ type is observed. Despite the similarity in bulk magnetic properties the moment distribution in the V system appears evenly distributed on the Co sublattice whereas in the Cr system the Cr atoms also possess a moment. These may be involved in the formation of superparamagnetic assemblies which diffuse scattering measurements suggest exist in the Co(GaCr) alloys.

The s-d Hamiltonian has been factorised and by using an identity of the functional-integral technique, the Hamiltonian is linearised. The author obtains the partition function of the system in the static approximation by solving the equations of motion of the single-particle Green functions. From this partition function specific heat of the system is calculated. The specific heat against temperature curve shows the well known peak.

Detailed experiments on temperature variations of magnetic susceptibility and NMR Knight shift are reported. A linear relationship is found between the concentration-independent susceptibility and the concentration-independent Knight shift. The T² temperature variation at T<<T_K is clearly observed. All the past data including those of Van Dam et al. are shown to be contaminated by the interaction effect. Compared with the Yosida-Yamada-Yoshimori theory, the correlation energy U and the Kondo temperature T_K are estimated.

The magnetisation of the Kondo systems CuFe, AgFe, AuFe and MoFe has been obtained from Mossbauer experiments for 0.03K<T<200K and external fields up to 60 kG. The results for the different fields are compared with each other and to recent theoretical ideas of universal scaling, developed by Wilson on the basis of renormalisation group theory.

The effects of introducing dilute quantities of Pd(<3at%) on the characteristic antiferromagnetic properties of chromium have been investigated using neutron diffraction. The results support earlier magnetic susceptibility measurements in suggesting that Pd does not conform to the general pattern exhibited by the nonmagnetic solutes which are to the right of chromium in the periodic table. In fact a decrease in the parameters mu , T_N, Q and T_SF is observed. The results can be explained if it is assumed that the Pd d electrons are not contributed to the band and that the Pd configuration approaches a complete d shell.

The electron spin resonance of gadolinium was measured on polycrystalline samples of the cubic metallic monopnictides of type MX (M=Sc, Y, La, Lu; X=P, As, Sb, Bi). For most of these systems the spectra show a characteristic line structure which clearly reflects the influence of the crystal field. The cubic crystal field parameter b₄ of Gd is known from previous measurements on single crystals for YSb, YBi, LaSb and LaBi. The spectra of Gd are compared for LaSb powder samples and single crystals. This yields a method for deriving the crystal field parameter b₄ from spectra of polycrystalline samples. In this way the values of b₄ of Gd are determined for ScAs, YAs, LaP, LaAs, LuP and LuAs; for ScP and YP an upper limit can be given. The variation of b₄ is discussed in terms of lattice parameter and covalency throughout this series of compounds.

Low-energy (h(cross) omega =21.2 eV) photoemission measurements of liquid gallium made over the temperature range 320K to 820K are reported. The optical density of states derived from the spectra contrasts with a free-electron curve and reflects the main features observed in high-energy X-ray-induced spectra for solid gallium films. Little variation is found in the spectra with change of temperature.

The fine structure (K1, K2) in the main X-ray K absorption edge of rubidium in the metallic state and in some rubidium compounds is reported for the first time. The chemical shifts of the K edge of rubidium in the compounds Rb₂O RbOH, RbCN, RbF, RbCl, RbBr, RbI, RbNO₃ and Rb₂SO₄ have been measured. The extended fine structures have been observed and analysed to give information on the average bond lengths.

The authors performed optical measurements consisting of near normal incidence reflectance (2-11 eV) and multiangle reflectance measurements (11-15 eV) on gallium films condensed on cooled glass substrates in a vacuum of 2*10^-10 Torr. Optical constants were determined with the help of a Kramers-Kronig analysis of the reflectance data. The optical conductivity sigma , real part of the dielectric constant epsilon ₁ and energy loss function -Im(1/ epsilon ) were deduced. epsilon ₁ against 1/(h(cross) omega )² shows a linear dependence; -Im(1/ epsilon ) peaks at (14.05+or-0.05) eV.

The results of polarized neutron diffraction measurements on ferromagnetic BCC Fe-V alloys are analysed within the multiband CPA theory. A Hubbard Hamiltonian with intra- and interatomic exchange interaction is used. Both Hubbard and exchange terms are treated in the Hartree-Fock approximation. A good agreement for the magnetic moment and asphericity of its distribution is found. It is shown that the increase of the asphericity parameter with vanadium content is caused by the transfer of e_g electrons from vanadium to iron sites. This transfer is responsible also for the occurrence of a negative moment on vanadium confirmed in diffuse scattering measurements.