Determining the current polarization in Al/Co nanostructured point contacts

We present a study of the Andreev reflections in superconductor/ferromagnet nanostructured point contacts. The experimental data are analyzed in the frame of a model with two spin-dependent transmission coefficients for the majority and minority charge carriers in the ferromagnet. This model consistently describes the whole set of conductance measurements as a function of voltage, temperature, and magnetic field. The ensemble of our results shows that the degree of spin polarization of the current can be unambiguously determined using Andreev physics. PACS numbers: 74.45.+c, 72.25.-b, 74.78.Na The field of spintronics is largely based on the ability of ferromagnetic materials to conduct spin-polarized currents [1]. Thus, the experimental determination of the degree of current polarization has become a key issue. Recently the analysis of Andreev reflections in superconductor/ferromagnet (S/F) point contacts has been used to extract this spin polarization in a great variety of materials [2, 3, 4, 5, 6, 7]. The underlying idea is the sensitivity of the Andreev process to the spin of the carriers, which in a spin-polarized situation is manifested in a reduction of its probability [8]. The theoretical analysis of these S/F point-contact experiments has been mainly carried out following the ideas of the Blonder-Tinkham-Klapwijk (BTK) theory [9]. Different generalizations of this model to spin-polarized systems have been proposed, in which with an additional phenomenological parameter P, the spin polarization of the ferromagnet, excellent fits to the experimental data have been obtained [2, 3, 4, 5, 6, 7]. However, a microscopic justification of these models is lacking [10, 11, 12]. Recently, Xia et al. [13] have combined ab initio methods with the scattering formalism to analyze the Andreev reflection in spin-polarized systems. Their main conclusion is that, in spite of the success in fitting the experiments, these modified BTK models do not correctly describe the transport through S/F interfaces. Therefore, at this stage several basic questions arise: what is the minimal model that describes on a microscopic footing the Andreev reflection in spin-polarized systems? And, more importantly, can the current polarization be experimentally determined using Andreev physics? In this paper we address these questions both experimentally and theoretically. We present measurements of the differential resistance of nanostructured Al/Co point contacts as a function of voltage, temperature, and magnetic field. To analyze the experimental data we have developed a model based on quasiclassical Green functions, the main ingredients of which are two transmission coefficients accounting for the majority and minority spin bands in the ferromagnet. We show that this