Electric current in conventional semiconductor electronics is controlled by a voltage applied to various parts of any particular electronic component. However, a completely new way of controlling the flow of electric current was proposed some twenty years ago. This new idea is based on the observation that the internal angular momentum of electrons (spin) with its associated magnetic moment is conserved over nanoscale distances. It follows that, when an ultrathin layer structure is prepared, the spin 'remembers' its orientation across the whole thickness of the structure, which means that electrons with different spin orientations do not mix and flow independently as if in two separate wires connected in parallel. If the layer structures contains magnetic layers the two spin channels become inequivalent. Moreoever, it is found that the resistance of electrons with a given spin orientation depends on the magnetic configuration of all the magnetic components in the layer structure. Since the magnetic configuration can be altered by applying a magnetic field, one can control the flow of electrons (electrical current) by applying a magnetic field. With this discovery the era of an entirely new field of condensed matter physics called spintronics had began. Successful applications of the ideas of spintronics depends on our ability to grow ultrathin magnetic layer structures which are near perfect on an atomic scale. Within the last twelve months this has been achieved for layer structures containing ferromagnetic metals (FM) and MgO insulating barrier. However, for multilayers contaning FM layers and semidoncuctor (SC) layers these ideal conditions have yet to be reasloed. Yet the future success of spintronics depends on integration of spintronic components into conventional semiconductor structures. The main goal of this proposal is to combine experimental expertise in the area of classical spintronics and, in particular, expertise in epitaxial growth of layer structures with theoretical insights gained from studying near perfect magnetic junctions with an MgO barrier, We are confident that both experimental and theoretical methods developed for magnetic junctions with an MgO barrier can be transferred to FM/SC systems and thus the outstanding problem of achieving near perfect spin transport across FM/SC interfaces can be solved.