Soft X-ray Emission Spectra from Aluminium-niobium and Aluminium-palladium Alloys

A1 Lz.3 emission spectra from AI-Nb and Al-Pd alloys have been measured and compared with the A1 K emission spectra from the same alloys. Information regarding the density of states in the vicinity of aluminium ions has been deduced. 1.0 Introduction. Soft X-ray spectrometry is a well established technique for investigating the density of electron states throughout the valence band of metals or alloys. Nevertheless, the results are difficult to interpret because the effects of transition-probabilities can not be estimated without an apriori knowledge of both the initial and final state wave functions of the electrons involved in the emission process and because the broadening of spectra due to many body effects can not easily be estimated. Aluminium is one of few metals for which the soft X-ray emission spectrum is fairly well understood [I] and as such, it provides a good basis for the investigation of binary alloys. However, although the shape of the pure aluminium emission spectrum may be explained, the explanation of the aluminium spectrum from alloys again becomes difficult since it is hard to predict to what extent changes in the spectrum shape are due to changes in the density of states and to changes in transition probabilities. The latter has received little theoretical attention although it has been pointed out by Stott [2] that, in dilute alloys, the transition probability for a solvent atom depends on its position relative to the solute atom. A further aspect which must be borne in mind when attempting to draw conclusions from the emission spectra of alloys is that the emission from, in our case, the aluminium atom is a function of the density of states and the symmetry of the valence band wave functions in the vicinity of the aluminium ion core. Recently Soven [3] has predicted that the density of states in the vicinity of the solvent and solute atoms will differ. The results presented here support this prediction. In effect, then, the soft X-ray emission spectra give information about the density of states in a localised region about the emitting atom ; and this must be a function of the configuration of the different types of atom surrounding the emitting atom. In the case of random substitutional solid solutions, the emission spectrum will reflect an average of the above effects and changes in the emission spectrum would be expected to vary smoothly with changes in the concentration of the alloy. This was found to be the case in the aluminium-silver system [4]. For ordered intermetallic compounds, however, each emitting atom for a particular compound will be in more or less the same environment and one might expect that for different intermetallic compounds of the same alloy system, the changes in the emission spectrum from each component might be quite considerable due to changes in the environment of the emitting atoms. We shall see in the results presented here, that this is the case in at least one alloy system. 2.0 Experimental. -The work presented here is complementary to the work reported in [5 ] and 161. The alloys (prepared in Kiev ; Ukr. Ac. Sci., Institute of Metal Physics) were made in an argon atmosphere in an electric arc furnace having a tungsten electrode. They were then annealed at a suitable temperature to homogenise them. The apparatus has been described in detail elsewhere [7]. Briefly, the spectrometer uses a ruled concave diffraction grating at grazing incidence with a scanning photomultiplier detector. The grating has a radius of curvature of 99.88 cm, a blaze angle of 20 4, 600 lineslmm. and is coated with platinum to enhance reflection. The detector is a Bendix M 306 Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1971459 C4-326 L. M. WATSON, Q. S. KAPOOR AND V. V. NEMOSHKALENKO photomultiplier with a copper-beryllium photocathode. The spectrometer chamber is maintained at a pressure of 2 x torr, during operation, by a mercury diffusion pump and two liquid nitrogen vapour traps. The X-ray tube consists of a rotatable, water-cooled, stainless steel anode on to which four specimens are attached using a conducting cement. X-rays are produced by electron bombardment from an electron gun consisting of a barium oxide coated nickel filament and a slit which focuses the beam on to an area of the target of approximately 0.05 cm by 1 cm. A mechanical scraper is incorporated into the tube so that, by rotating the target to face the scraper, the surface of the target may be scraped to expose clean metal. The X-ray tube is maintained at a pressure of 2 x torr during operation by a separate mercury diffusion pump and liquid nitrogen vapour trap system. A potential of 2 kV is applied to the target and a bombarding current of 5 to 9 mA is normally used. The power supplies are stabilised to better than 0.1 %. The results were obtained by repeated scanning of the same spectrum until enough counts had been accumulated to reduce the statistical counting error to an acceptable level. Before each scan, the target was scraped to prevent the build up of contaminants on the surface. This was not completely successful as we discuss later. The data was collected on punched tape for subsequent processing by computer. 2.1 PROCESSING. -The results are presented as processed and unprocessed. The processed results have undergone the following treatment. a) The background intensity, which is assumed to vary linearly between suitable points beyond the high and low energy limits of the band, has been subtracted. b) Each point in the spectrum has been divided by the fifth power of the frequency (y5), y Z because the spectrometer measures I(l) d l , where 1(A) is the intensity at wavelength l and d l is the wavelength interval embraced by the detector slit width, and converting this to I(E) dE, where E is the energy of the radiation, introduces the factor of y 2 . Theremaining y3 term is to account for the frequency dependent factor of the transition probabilities. c) Intensity loss, due to the build up of contaminants on the surface of the target during each scan, is measured by measuring the intensity at some fixed point in the spectrum before and after each scan. Using these data, each point in the spectrum is corrected for intensity fall off by assuming the fall off to be linear in time. (This has been measured and was found to be a good approximation.) Because the above corrections, in particular (a), may be open to question, the results are also presented in the unprocessed form. In this case only the conversion from I(A) dA to I(E) dE has been made. In both cases the error bars, representing the 67 % confidence levels, have been drawn. 3 .0 Resuits and discussion. Figures l a and 4a show the processed Al L,,, emission spectra from the alloys drawn one vertically above the other. The straight vertical lines are at 1 eV separation. Figures Ib and 4b show the unprocessed spectra and, since the back-