Table of contents

Mossbauer spectra of amorphous Fe₇₃Ti₂₇ alloys with various hydrogen contents have been recorded at different temperatures. These spectra are asymmetric, which indicates a correlation between isomer shift and hyperfine field. It appears that both the parameters of this correlation and the mean isomer shift are uniquely determined by the value of the mean hyperfine field. The same result is found concerning the shape of the hyperfine field distributions.

TDPAC (time-differential perturbed angular correlation) spectra of ¹⁸¹Hf were measured for the first time for an amorphous alloy (Cu₅₇Hf₄₃) prepared by melt spinning. The resultant spectra indicate that there are Hf atoms surrounded by a crystalline-like atomic arrangement as well as atoms in a state of disorder. The power spectra of the TDPAC spectra were estimated using the maximum entropy method (MEM). By comparing the power spectra for the Cu₅₇Hf₄₃ amorphous alloy with those for the Cu₃Hf₂ and CuHf₂ intermetallic compounds, the crystalline-like atomic arrangement around the Hf atoms in the amorphous sample was found to be nearly identical to that in the Cu₃Hf₂ intermetallic compound.

Measurements of the electrical resistivity of pure Li and LiMg alloys below 3K indicate that the magnitude of electron-electron scattering in pure Li is determined directly by the band structure of Li rather than indirectly by anisotropic scattering.

The crystallisation of Ni₇₉B₂₁ and Ni₆₂B₃₈ amorphous alloys produced by ion implantation is studied via perturbed angular correlation experiments on radioactive ¹¹¹In probes implanted in the amorphous layer. The hyperfine interaction sensed by ¹¹¹Cd (daughter of ¹¹¹In) provides information about the structure (amorphous or crystalline) of the alloy. The crystallisation temperatures and the nature of the crystallised phases are thus determined; they are known to depend strongly on the alloy composition.

The diffusion of Al in Au:xAl (x=1.6, 3.0, 5.0 at.%) has been investigated by measuring the rotating-frame nuclear spin relaxation time T_{1 rho} of ²⁷Al. It is shown that the diffusion-induced contribution to T_{1 rho} , which we observe in the temperature range between 550 and 1150K, has to be attributed to ²⁷Al diffusion. From the correlation times obtained from a Torrey approximation of the NMR data the temperature dependence of the correlated Al diffusion coefficient in Au:Al has been determined.

A steady creep rate together with a review of the literature on diffusion measurements in beta phases of the alloys Ag-50 at.% Zn, Cu-25 at.% Al, Cu-14.4 at.% Sn and Cu-42.7 to 52 at.% Zn are incorporated in the Dorn equation in order to choose the effective diffusion coefficient D. The thermodynamic approach to the vacancy mechanism allows the introduction of a new 'segregation coefficient', in addition to the stationary, homogeneous and interdiffusion coefficients. This analysis is reinforced by the work of Menon and Copley (1975) on segregation in binary metallic solid solutions due to diffusional creep.

The authors propose a simple local pseudopotential for lithium composed of a Hartree potential due to the nucleus and the core electrons, a core-valence exchange-correlation potential in the local density functional approximation and a repulsive delta -function potential, which represents the orthogonalisation effect due to the s core states. The pseudopotential can be expressed in simple analytic form, specified with only one parameter, both in real and in inverse space. It is shown that a number of the more basic physical quantities, depending upon calculated form factors and effective pair potentials, can be simultaneously described within a rather limited range of the parameter.

Ordering potentials of a liquid Li-Na alloy are obtained from pair interactions calculated by pseudopotential theory. They are compared with an experimentally based potential deduced from the neutron diffraction observations of Ruppersberg and Knoll (1977). The agreement is reasonably good for an appropriate parameter choice. The concentration-concentration structure factor is then calculated in the random-phase approximation, starting from the ab initio pair potentials, and comparison is made with the observed one. It is thus shown that certain characteristic features of the latter can, in some measure, be reproduced theoretically from first principles.

The conduction electron Zeeman splitting has been measured on the X-centred hole pockets in palladium utilising the de Haas-van Alphen technique. Two spin-splitting zero contours have been revealed and a large anisotropy was found for the cyclotron orbit g-factor (g_c). The variation of the g_c-factor on the X pockets is compared with previous results on Fermi surface sheets in palladium and platinum.

The one-electron energy band structures for LaBe₁₃, CeBe₁₃, ThBe₁₃ and UBe₁₃ were calculated by the self-consistent APW method with the local density approximation. It was found that the Be s(p) bands form wide valence bands below a pseudogap at 0.8 eV below the Fermi energy E_F in UBe₁₃; and the Be p bands extend mostly above a pseudogap at 1.9 eV above E_F in UBe₁₃. Between these bands, there are M d-Be p mixed bands. In CeBe₁₃ and UBe₁₃, the f bands are inserted among the d-p mixed bands. Due to the squeezing effect of the p-f mixings from both above and below the f bands, the f band widths in CeBe₁₃ and UBe₁₃ are very small, only 1 eV in UBe₁₃, much narrower than the experimentally observed PES-BIS peak, indicating the importance of the many-body effect. In LaBe₁₃ and ThBe₁₃, the f band widths are much larger, more than twice that in UBe₁₃, due to the reversed p-f mixing effect. The p-f mixing parameters for these materials were also evaluated from the present band calculation.

A method for the calculation of an open shell, strong-scattering metal cohesive energy E_coh in terms of the effective interaction is presented. It involves a so called reference medium defined as an ensemble of weighted atomic configurations which are significant for the problem under consideration. E_coh is written as a sum of n-body interactions up to n=M plus a residual error epsilon _M. The effective forces are determined in a unique way by making epsilon _M statistically small for the reference medium configurations. The M=2 (up to pair-force) description is studied in detail for the bulk, weak defect case. A metal, liquid-like reference medium is then appropriate and the recently developed liquid metal theory is used to derive closed equations for the effective pair interaction within the Kohn-Sham local density approximation. The theory is applied to the study of d-bonding pair interaction in transition metals as a function of interatomic distance R and of d-band charge Z_d, within Andersen's tight-binding atomic sphere approximation. The pair interaction is found to be mainly attractive and to behave as Z_d(10-Z_d)/R¹⁰ when R is near the physical hard-core diameter. The behaviour is more complicated when R is much larger and is explained with the help of an approximate solution of the full equations. Finally the Dick-Overhauser overlap pair interaction, which is valid for closed shell atoms, is extended to the open-shell case.

A series of polycrystalline Co_xSi_100-x (x=0 to 100) samples were studied by positron lifetime and Doppler broadening methods and by X-ray diffractometry. Only one lifetime component could be resolved even in the case of two-phase samples, varying between 220 ps (Si) and 130 ps (Co). The parameter S and the lifetime show similar dependences on the composition. Both tau and S can be interpreted well in terms of the average electron density (including both core and valence electrons). Structural vacancies around CoSi₂ can be excluded. In the range from Co to Co₂Si tau and S saturate with values that can be expected for Co vacancies.

The authors show that the two-peak structure of the valence photo-emission from cerium metal is connected with two local processes of electron emission. One peak appears approximately=2 eV below the Fermi energy epsilon _F and corresponds to the removal of an f electron and simultaneous screening by electrons close to epsilon _F. The other peak, approximately=0.2 eV below epsilon _F, is caused by the removal of a d electron and simultaneous screening by an f electron. The binding energies are obtained from relativistic self-consistent Delta SCF calculations for various configurations of a fictitious lattice atom whose embedding in Ce metal is simulated by using non-integral orbital occupation numbers extracted from a self-consistent band-structure calculation. Although the two peaks are related to the formation of different holes the authors show that they display similar resonance behaviour. In the off-resonance case they calculate their relative spectral intensities within a generalised Kohn-Sham theory. The results agree qualitatively with the observed dependence on the photon energy. The satellite energies of the core spectra can be explained quantitatively within the same effective-lattice-atom approximation.

The authors report resistivity and magnetoresistance measurements (4.2<T<300K; H<or=7.6 T) of extremely pure monocrystalline MoSi₂ (RRR approximately=1800). This metallic compound behaves like a compensated metal (n(h⁺)=n(e^-)) with a very large magnetoresistivity at 4.2K which varies like H² up to 7 T. The anisotropy of the electronic structure accounts well for the anisotropy of the results. Deviations from Matthiessen's rule are also reported.

The Hall coefficient R_H of Hg-Ga liquid alloys has been measured, via the Corbino geometry, using Ga as a reference material. Phase separation was avoided by working in the temperature range 490-500K. Disc specimens 1 mm thick were used to ease the problems of preparation and containment. The results have been compared with R_H(FE)=-(n mod e mod )^-1 which is the value predicted by the free-electron model. It was found that Delta R_H/R_H(FE)=(R_H-R_H(FE))/R_H(FE) is always negative in the Hg-Ga system, a trend that is observed in other mercury alloys. The authors discuss whether the contribution to R_H from skew scattering, caused by spin-orbit interaction, can explain the general behaviour that is found experimentally.

The authors present a theory for the static, paramagnetic, spin susceptibility for metals at finite temperatures. It is based on a 'first principles' mean-field theory of itinerant magnetism and provides an expression which resembles that of a classical Heisenberg model added to an itinerant component. The quantities which occur in this expression all depend on the underlying electronic structure of the disordered local moment (DLM) paramagnetic state and thus demonstrate the subtle combination of the localised and itinerant aspects of this problem. The theory is applied to BCC iron and FCC nickel. A Curie temperature of 1280K and effective Curie constant moment of 1.97 mu _B is obtained for iron in reasonable agreement with experiment. From the calculations of the wavevector dependent susceptibility, which are compared with quasi-elastic neutron scattering data, equal time spatial, magnetic correlation functions are inferred which are consistent with the magnetic structure of the initially imposed (DLM) paramagnetic state. The calculations for nickel provide a very different picture. At temperatures at which the DLM paramagnetic state differs from the conventional 'Stoner' state, it is shown that this state is unstable via the sensitivity of the moments to their orientational environment and that the theory for the paramagnetic state is equivalent to that of the Stoner-Wohlfarth picture. It is concluded that an improved theory for nickel must incorporate a mechanism for 'local' moment formation on several sites.

Applying the theory developed in a previous paper with improved electronic data to Pd and, for the first time, to V the authors calculate the dynamic spin susceptibilities of these transition metals in their full dependence on the wave vectors. In addition a parameter-free theory of the electron-spin fluctuation coupling is given, enabling the evaluation of the spin fluctuation contribution to the electron mass enhancement. Results for (m*/m)_spin are of the right order of magnitude and demonstrate that from the theoretical point of view a large Stoner enhancement is compatible with a relatively low mass enhancement.

The temperature dependence of the magnetic susceptibility of liquid Na and Cs has been re-examined experimentally. The present and earlier measurements indicate that for liquid Li and Cs the magnetic susceptibility decreases with rising temperature, although for liquid Na, K and Rb it increases with temperature. Such experimental results are explained theoretically in terms of the nearly-free-electron model for the magnetic susceptibility of liquid metals. Structural correction effects are very important for accounting for these temperature variations.

The authors have performed some measurements of magnetic susceptibility, magnetisation (H<or=150 kOe), large- and small-angle neutron scattering (SANS) at different temperatures (2.6K<T<300K) on the asperomagnetic amorphous alloy TbCu_3.54. Having analytically described the SANS data, they have carefully distinguished the various scales of observation in real space as a function of the distinct scattering laws in order to reach a full description of the magnetic order over a distance extending from one AAngstrom to a few microns. They observe the existence of large magnetic domains with an average magnetisation of about 1 mu _B/Tb and which have low scattering density boundaries. At intermediate distances they distinguish terbium 'bubbles' (diameter approximately 15 AA) regularly spaced ( approximately 40-50 AA) and having a magnetisation per unit volume ten times stronger than that of the surrounding matrix. They also describe the magnetic order at short distances. They present the variations of these properties with temperature and annealing. A H^-1/2 dependence of the magnetisation at the approach to saturation seems to be related to the presence of the terbium bubbles. The magnetic order exhibiting 'bubbles' with strong intensity of magnetisation is called 'seedy magnetic' order.

Remarkable changes in magnetic properties due to plastic deformation have been found in Co-rich Co₃Ti intermetallic compounds. Measurements of magnetisation as a function of temperature, magnetic field and the degree of plastic deformation have been performed, mainly on a sample containing 22 at.% Ti, using a magnetometer. The spontaneous magnetisation and Curie temperature have been determined from the sigma _g²-H/ sigma _g plot for several temperatures. The temperature variation of the spontaneous magnetisation in the heavily deformed samples shows unusual behaviour. A quantitative correlation of the magnetisation and the Curie temperature with the degree of cold-rolling has been obtained. A qualitative explanation of the changes in the magnetic properties, taking into consideration the local atomic environment and the anti-phase boundary, is given.

The problem of Knight shifts in samarium intermetallic compounds is investigated using a density matrix perturbational approach, which takes into account the strong admixing between the J=⁵/₂ and ⁷/₂ levels of the Sm³⁺ ion. Particular attention is paid to the effects of crystal fields on the White-Van Vleck crossover temperature T_CO, and explicit expressions are given for the shifts in T_CO, induced by hexagonal and cubic crystal fields, correct to (B_n^m/kT)². It is shown that whereas quadratic crystal fields can either raise of lower T_CO, depending on the direction of the applied field B, cubic crystal fields always act so as to lower the White-Van Vleck crossover temperature. In addition, a method is suggested for the determination of both the bare conduction electron Knight shift parameter K₀ and the crystal-field coefficient B₂₀ in hexagonal compounds, which does not involve additional measurements on isostructural compounds containing lanthanum.

The analysis of Mossbauer absorption spectra of ¹⁷⁰Yb in YbBe₁₃ in the temperature range from 0.05 to 75K leads to the following conclusions: (i) Yb is trivalent in the magnetically ordered (T_N=1.28K) as well as in the paramagnetic regions and its ground crystal-field state is the Gamma ₇ doublet; (ii) the fast Yb³⁺ paramagnetic fluctuations are mainly due to the exchange RKKY interaction; (iii) the observed thermal variation of the relaxation rate implies that conduction electron scattering by the Yb ion is also present, and suggests that in YbBe₁₃ the Yb ion has a Kondo behaviour.

High-purity aluminium samples, thermally treated to systematically vary the concentration of monovacancies or prismatic dislocation loops, were used to evaluate the recently developed fast Fourier transform/power spectrum (FFT/PS) deconvolution method. It is shown that the FFT/PS method is capable of automatically correcting for accidental and/or intentional changes in the instrument resolution function. In addition, the results of this study indicate that both the sensitivity and the accuracy of the energy distribution lineshape parameters are improved by deconvolution. Finally, it is demonstrated that if the filtering error introduced by the FFT/PS procedure is controlled, then it is possible to distinguish aluminium samples containing predominantly vacancy-type defects from those with prismatic dislocation loops as the dominant defect type.

Amorphous Ni-Zr powder was prepared by mechanical alloying of crystalline nickel and zirconium powder in an argon flux. X-ray diffraction and magnetic measurements were used to follow the process. The wavenumber of the principal diffraction peak and the crystallisation behaviour of the amorphous powder were compared with results known from studies of melt-spun Ni_xZr_1-x.