Table of contents

Recent work by other investigators showing solid xenon and argon precipitates in aluminium has been extended in the present Letter to an examination of the structure of krypton bubbles in copper, nickel and gold. As in previous work, the inert-gas bubbles are found in the solid face-centred cubic phase epitaxial with the host crystal. Electron microscopy diffraction patterns have been used to show the solid to gas transition during heating, and the subsequent recrystallisation on cooling. The results on the solid inert-gas bubbles are a dramatic indication of the high pressure involved; these pressures, which vary with the shear modulus of the metal, are a direct consequence of the bubble growth processes.

Using elementary principles, the six components of the symmetric valence electric field gradient tensor q_ij^v are calculated for a defect in a cubic metal. This leads to results differing from the earlier ones along the non-symmetric directions, although the results agree along the symmetric directions. The theory is applied to several impurities in aluminium.

The spin-resolved density of states of Fe-Ni ferromagnetic metallic glasses (Fe₈₂B₁₂Si₆, Fe₆₀Ni₂₀B₂₀ and Fe₄₄Ni₃₇B₁₉) has been determined by spin-resolved UV photoemission spectroscopy (h nu =21.2 eV). The samples show 'common band' behaviour. It is suggested that they are weak ferromagnets. There is a trend towards strong ferromagnetism with increasing Ni content, in contrast to recent conjectures of strong ferromagnetism in Fe₈₃B₁₇.

Measurements of the magnetocaloric effect in two different spin glasses, CuMn and LaGdAl₂ are presented and compared with the predictions of equilibrium thermodynamics. It is shown that LaGdAl₂ is an extremely reversible system that exhibits a relatively large amount of adiabatic cooling below T_G and very little irreversibility. CuMn, on the other hand, is much more irreversible and a definite observation of adiabatic cooling has not been made. It is proposed on the basis of these results that the spin-glass phenomenon is fundamentally an equilibrium one that is often obscured by secondary irreversible properties.

The formation of He-stabilised voids in aluminium single crystals injected with helium is monitored by positron lifetime measurements. The bubbles are observed to grow during annealing from 300 to 930K. The bubbles are extremely stable and survive annealing up to the melting point. Positron lifetime data are used to discuss the He density inside the bubbles and their growth mechanism.

Positron annihilation Doppler-broadening spectra were measured and lineshape parameters were estimated during isochronal aging of dilute alloys of Al-1.0 at.%Si and Al-0.1 at.%Si quenched from 773 and 823K to 183K. Two peaks in the lineshape parameters were observed at about 213 and 243K in both alloys. It is suggested that the increase and subsequent decrease of the parameters around 213K are due to the growth of vacancy clusters and the collapse of their open volumes, respectively. Positron annihilation experiments showed that the collapse of vacancy clusters in quenched pure Al seems to occur in a temperature region below about 213K. The size distribution of Si precipitates was determined by electron microscopy in various aging treatments of Al-1.0 at.%Si alloys quenched from 823K. Taking into account the results from electron microscopy, it is deduced that the change in lineshape parameter around 243K seems to be related to the nucleation of Si precipitates.

The temperature dependence of the elastic constants C'=¹/₂(C₁₁-C₁₂),C₄₄ and B of BCC iron is calculated within the tight-binding approximation using the single-site spin-fluctuation theory of band magnetism. The theory reproduces the observed anomalous 'softening' of the tetragonal shear constant C', relating it to the contribution of the e_g orbitals. The negative slope with temperature of the trigonal shear constant C₄₄ and bulk modulus B is shown to arise from the lattice expansion. The influence of magnetic moments on the elastic constants of ferromagnetic and antiferromagnetic, BCC transition metals is discussed in general.

The author has studied the effect of di-interstitials on phonon dispersion curves in FCC metals. The calculated frequency shift of the transverse (100) phonon in Al shows resonance behaviour at wavevector zeta =0.1 ( nu =0.86 THz), a result corroborated from the local frequency spectrum of the di-interstitial. The calculation highlights the importance of di-interstitials and other small clusters in understanding the perturbed phonons in neutron-irradiated metals.

Self-diffusion in solid Li was studied via the spin-lattice relaxation of polarised radioactive ⁸Li nuclei using their asymmetric beta -decay radiation. The diffusion-induced spin-lattice relaxation rate 1/T_1diff was measured as a function of temperature from the melting point T_m=454K down to about ¹/₂T_m and as a function of the magnetic field from 8 to 700 mT. 1/T_1diff can be explained assuming the nuclear dipole-dipole interaction alone. The data are analysed in the frame of the encounter model of correlated self-diffusion. The macroscopic self-diffusion coefficient D^SD is determined over almost seven decades. A deviation from Arrhenius behaviour at elevated temperatures is interpreted in terms of a combined monovacancy/divacancy (1V/2V) mechanism. The correlation factor obtained from D^SD and tracer self-diffusion coefficients D^T, measured by mass spectroscopy, has a temperature dependence which is consistent with the 1V/2V model. The question of the possible existence of a non-classical isotope effect cannot be settled by a comparison of the present D^SD values with those known from conventional nuclear magnetic resonance on ⁷Li.

A simple two-phase model for non-stoichiometric palladium hydride (PdH_x, 0<or=x<or=1) is described. This model is applied within the cluster-Bethe formalism. Local densities of states, charge densities, total energy and heat of formation are calculated, and the results are compared with the authors' previous one-phase theory and with experiment. The two-phase model gives results which are closer to experimental findings than those of the one-phase theory, especially for lower values of x(x<0.5).

Electron scattering by interstitial hydrogen or deuterium in palladium is investigated by means of measurements of de Haas-van Alphen Dingle temperatures, T_D, and electrical resistivity, rho . For Fermi-surface orbits with a strong d character dT_D/dx (x is the number of interstitial atoms per Pd atom) is approximately twice as large in PdD_x as in PdH_x. For electrons on the large Gamma -centred Fermi-surface sheet, however, dT_D/dx=750K for both H and D. This value is much smaller than values predicted by average-t-matrix calculations. The experimental value for dT_D/dx is, however, confirmed by electrical resistivity measurements on homogeneous alpha -PdH_x and alpha -PdD_x samples prepared by rapidly cooling down the samples from the hydrogen (deuterium) loading temperature (typically 600K) to helium temperature.

The authors have measured polycrystalline Fe, polycrystalline B and a series of Fe-B metallic glasses in the amorphous and crystalline states by means of high-resolution UV photoemission. Auger electron and energy-loss spectroscopy. The density of states of the amorphous metallic glasses is characterised by a maximum immediately at the Fermi level which may be assigned to the Fe 3d electrons. No definite changes in the electronic structure were observed in the case of single-phase crystallisation (Fe₇₅B₂₅). For the other samples they have found effects due to the precipitation of alpha -Fe. A comparison with photoemission results in the literature showed discrepancies which they explain mainly by chemical differences in the sample surfaces.

For pt.I see ibid., vol.12, p.281 (1982). The Compton spherical profile of the conduction electrons in copper is calculated employing the one-particle Green function to determine the spectral weight function of Cu. Many-body effects are introduced through the energy-dependent self-energy. The momentum and energy distributions are calculated and dispersion relations are obtained by plotting the energy of the peaks of the momentum and energy density against p. Two calculations are performed, one with and one without the energy-dependent part of the self-energy. The Compton profile J₀(q) with no energy-dependent self-energy is calculated and the difference between the two profiles is reported. The Compton profile calculated with energy-dependent self-energy is compared with the available experimental data and the effect of the energy-dependent self-energy is investigated.

The RKKY indirect exchange interaction in metals of arbitrary continuous dimension d>or=1 is evaluated exactly in the Kaneyoshi model of the electronic conduction states. The interaction reduces to known forms in the limiting case corresponding to free electrons. Otherwise there is exponential damping of the oscillating interaction strength corresponding to ranges of the order of the electronic mean free path as well as a phase shift of the oscillation which also depends on the mean free path.

Measurements have been made of the residual resistivity and thermoelectric power in dilute alloys of CuNi, CuPd, CuPt and CuRh. A measurement has also been made of the UPS spectrum in CuRh. These have been used to obtain the parameters associated with the resonant d states in these alloys using the phase-shift model of Gupta and Benedek (1979). de Haas-van Alphen measurements made by other workers have been re-analysed, using the new residual resistivities, within the phase-shift model of Coleridge, Holzwarth and Lee (1974). Comparison is made between the parameters obtained from these two approaches. Various theoretical calculations of the de Haas-van Alphen Dingle temperature and frequency change on alloying are discussed in relation to the measured values.

The resistivity of a series of (PdNi)Mn alloys containing between 0.5 and 3.0 at.%Mn has been measured from 1.5 to 25K. Various estimates for the magnetic ordering temperature are extracted from these data and are compared with those found previously from AC susceptibility measurements on this system. The incremental resistivity data enable estimates for spin-dependent and spin-independent coupling constants between local Mn moments and conduction electrons to be made, together with values for the acoustic spin-wave stiffness D.

The electrical resistivity and magnetoresistance (MR) of single-crystal antiferromagnetic GdRh_1.07Sn_4.21 have been studied from 4.2 to 25K and in a magnetic field up to 30 kOe. Below T_N the resistivity can be attributed to electron-spin-wave scattering. A positive MR at temperatures below T_N is believed to arise from the field-induced enhancement of the spin fluctuations. The field dependence of the MR is quadratic at lower fields and it has an inverse temperature dependence below T_N.

A collation of data on the resistivity of aluminium foils is presented. The T² term caused by the surface, which is now well established, is examined and found to exist over a large range in sample thickness and observational temperature. Contrary to previous assertions, the magnitude of this T² term does not appear to vary simply as the inverse of the sample thickness. It is suggested that there are two ways in which a surface-caused T²-like term may arise. In metals with simple Fermi surfaces conventional surface scattering theories yield such a term from a combination of a bulk T⁵ term and surface scattering. By contrast, for a metal such as aluminium with a disconnected Fermi surface an extra T² term arising directly from surface scattering processes may appear.

Electronic structures of d electrons in the intermetallic compounds ScCo₂, TiCo₂, ZrCo₂, LuCo₂, HfCo₂ and ScNi₂ with cubic Laves phase structure are calculated in the tight-binding approximation. By using the calculated density-of-states curves and taking the effect of spin fluctuations into account, the temperature dependences of the paramagnetic spin susceptibility in these compounds are calculated. The magnetic field dependence of the induced magnetic moment in LuCo₂ at 0K is also calculated. Good agreement between the calculated and observed results is obtained. As observed in ZrCo₂ and HfCo₂, moreover, an excess Co atom in the Co-rich non-stoichiometric compounds is shown to have a local magnetic moment.

A neutron diffraction study ( lambda =1.140 AA) of the cubic intermetallic compound DyBe₁₃ has shown that below T_N=10+or-0.5K DyBe₁₃ exhibits a commensurate magnetic structure with a propagation vector tau =¹/₃c^* and magnetic moments of the Dy³⁺ ions (M₁, M₂ and M₃ with M₂=M₃) perpendicular to the crystallographic c axis. As powder neutron diffraction data can only yield two relations between M₁, M₂=M₃ and phi =(M₁,M₂), identical Dy³⁺ magnetic moments were assumed (M=M₁=M₂=M₃). A Mossbauer effect investigation of the same compound is reported in the temperature range 1.5-15K. Below the magnetic ordering temperature T_N=10K the results are interpreted within a model including either two magnetic sites (M₁ not=M₂=M₃, T=1.5K) or one site (M₁=M₂=M₃) and relaxation. A self-consistent analysis of the magnetic structure, with a Hamiltonian including exchange and crystalline electric field components, shows that M₁ and M₂ are neither identical nor significantly different ( Delta M/M<or=1%).

Ab initio calculations for the electronic structure of pairs of magnetic impurities like Cr, Mn and Fe in Cu and Ag are presented. The calculations are based on density functional theory and apply the KKR Green function method. The authors calculate the local densities of states and the local moments for both the ferromagnetic and antiferromagnetic pair configurations. In particular, they estimate the energy difference between these configurations for different interatomic impurity distances. They show that this RKKY-type energy can be obtained in first order from the difference between the single-particle energies of these configurations calculated with the same 'frozen' impurity potential. In both hosts, two Fe impurities couple strongly ferromagnetically on nearest-neighbour sites, whereas Cr impurities show a strong antiferromagnetic interaction. Mn impurities are an intermediate case, interacting weakly antiferromagnetically.

The magnetic susceptibility of a range of CuMn alloys containing 10-90 at.% Mn was measured over the temperature range 4-400K. A spin-glass-type freezing was observed for all compositions, and alloys containing more than 70 at.% Mn also showed an anomaly at their Neel temperature. A magnetic phase diagram is proposed. When alloys containing less than 60 at.% Mn were aged at 100 degrees C the magnetic susceptibility was found to increase, with the maximum effect being observed at 20 at.% Mn. The susceptibilities of alloys containing 60-70 at.% Mn decreased with aging. Both these observations are consistent with the short-range order known to exist in these alloys.

The magnetic transition at the Neel temperature, T_N, of the heavy rare-earth metal holmium has been examined by specific heat measurements. The transition at T_N=(132.245+or-0.080)K is found to be continuous and the data around T_N have been analysed in terms of the equation C^+or-=(A^+or-/ alpha ^+or-) mod t mod ^{- alpha +or-}(1+E^+or- mod t mod ^x+or-)+B^+or-+D^+or-t where t=(T-T_N)/T_N. The value of the critical exponent alpha is 0.27+or-0.02 for the data in the critical regions -3.8<log₁₀ mod t mod <-1.2 below T_N and -3.9<log₁₀ mod t mod <-2.9 above T_N. The authors find that the scaling law alpha ⁺= alpha ^- is valid and the condition T_N⁺=T_N^- is obeyed in the critical regions. The value of alpha (=0.27) that they obtain is not close to the predictions of renormalisation group theory for magnetic systems with short-range exchange interactions. A cross-over from the critical region to a region with different behaviour is observed at log₁₀ mod t mod =-2.9 above but not below T_N. An unidentified anomaly is observed in the specific heat of holmium at 98K.

EXAFS measurements on the Zn K edge are reported for crystalline and amorphous Mg₇₀Zn₃₀ alloys. The spectra have been found to be identical in the two phases and have been well reconstructed using a multishell modelling. The local structure around the Zn atom is therefore similar in the glassy and crystalline states and shows preferential chemical ordering.

The surface core-level shift has been measured for Cu atoms at the (100) surface of a single crystal of a Ni-5% Cu alloy. Surface-sensitive high-resolution measurements of AlK_alpha excited photoemission from the Cu 2p_3/2 core level show surface-shifted features at a reduced binding energy 0.42 eV below the bulk binding energy. The magnitude of this shift is interpreted in terms of a quasichemical approach. The relative magnitudes of the surface and bulk components of the 2p_3/2 peak measured by low-angle electron emission exhibit a dependence upon the emission azimuth related to the surface crystal directions of the (100) plane. A simple atomic shadowing or electron blocking model is proposed to account for this. Peak-shape and peak-broadening factors are discussed for both the surface and bulk components.

A fairly thin crystalline layer was proved to be formed on the free surface of amorphous Fe-rich Fe-Zr alloys during the melt-spinning process. This crystalline surface layer can be easily removed by polishing lightly with emery paper. X-ray diffraction results for melt-spun amorphous Fe_100-xZr_x alloys showed that the crystalline surface layer consists of BCC Fe for 8<or=x<or=10 and Fe₃Zr for x=11. BCC Fe on the free surface of the amorphous Fe₉₁Zr₉ ribbon has a strong texture with its (100) plane parallel to the ribbon surface. The bulk crystallisation of amorphous Fe₉₁Zr₉ alloy is significantly affected by the preexisting surface crystallites; the crystallised ribbon shows a strong (100) texture, the same as that for the surface crystallites. The results of X-ray diffraction and differential thermal analysis for the amorphous Fe₉₁Zr₉ alloy showed that the surface crystallites decompose into alpha -Fe and Fe₃Zr before the onset of bulk crystallisation.

Recently Trebin and Bross (1984) obtained valence-electron densities of titanium carbide using the modified augmented plane-wave (MAPW) method. This density differs in symmetry around Ti from previous results obtained by Blaha and Schwarz (1983) who have employed the linearised APW scheme. This discrepancy is important in connection with the chemical bonding in this compound. It is shown that the electron densities obtained by LAPW and APW agree very well with each other and thus should be reliable. New LAPW calculations suggest that the MAPW results may be in error because in the latter work the k-space integration was limited to only high-symmetry directions.

For original paper see ibid., vol.14, p.2097 (1984). The author criticises the efforts of Oota et al. to analyse observed changes in the superconducting critical temperature Tc and residual resistivity rho ₀ in terms of the electron-to-atom ratio e/a for a range of pseudobinary A15 compounds based on V₃Ge, on the grounds that in such compounds, Tc is strongly dependent on both the exact composition and the degree of crystallographic long-range order. The addition of third elements, he maintains, may influence either of these factors. He also questions the claim of Oota et al. that their specimens were of a single A15 phase, and suggests metallographic examination would have shown they contained approximately 10% V₅Ge₃.