Transport properties arising from local spin fluctuations

Abstract

Various authors have calculated the self-energy of a localised electron in the one-band Wolff model as a first step in finding the transport properties of dilute transition metal-transition metal alloys. However they have each stated that the other's equation is incorrect. To resolve this difference we make our own detailed calculation of the self-energy and find that all the authors use effectively the same equation but in different approximations. We go on to calculate the transport properties - extending previous results by including the effect of the host in such a way as to avoid the approximations of having the host's or the impurity's contribution dominant. When there is no impurity potential scattering the fits to the electrical resistivity data (from RhFe and IrFe alloys) are quite good. With potential scattering added the resistivity curves develop minima at low temperatures which are not seen in the experimental data and thus make the fits much worse than before. Looking at the theoretical diffusion thermopower curves for the same two alloys we see a qualitative similarity to the experimental data but the predicted peaks are an order of magnitude greater than the experimental peaks. Comparing our results to the results from the two-band Wolff model (which does not include the effects of potential scattering) reveals a qualitative similarity with the major difference being that our prediction of the spin fluctuation characteristic temperature is 4 to 5 times greater than the two-band prediction for the same quantity.