One of the fundamental goals of current spintronics-based research is the efficient and controllable injection of electrons with a particular spin alignment into Si. Although recent success has generated interest in the possibilities of advanced silicon devices<SUP>1</SUP>, much of the physics involved with the spin injection process is not yet fully understood. To this end, gallium phosphide is well suited as it is the closest lattice matched III-V material to silicon (0.37% misfit) and has a large band gap relative to silicon.
The GaP/Si system has been the subject of interest for decades due to its potential as a gateway to III-V/Si integration. Recent work on MBE grown GaP/Si has resulted in high quality crystalline thin films<SUP>2</SUP>. By use of this process, which includes Si homoepitaxy and migration enhanced epitaxy III-V nucleation, GaP films free of defects related to the heterovalent interface can be grown to the precise thicknesses required for spin injection (~1-10nm). Much work remains to be done, however, on the characterization of these thin heteroepitaxial GaP/Si films and the GaP/Si interface. In addition to the fairly uncharacterized GaP/Si system, there is also little known about the nature of metal/GaP Schottky contacts, which are necessary for metal/GaP/Si tunnel-barrier spin-injection. To this end, Au/GaP and Fe/GaP Schottky diodes were fabricated to facilitate the characterization of the resulting metal-semiconductor contacts. Current-voltage data shows similar, strong rectifying behavior for both metals, and internal photoemission (IPE) experiments reveal Schottky barriers for Au and Fe to be 1.15 and 1.14eV, respectively. The negligible difference in resultant barrier heights compared to the 0.6eV difference in work functions between the two metals indicates a Fermi level pinning mechanism for Schottky barrier formation on GaP. Therefore, because Fe, a ferromagnetic metal, makes a good Schottky barrier with GaP, and the fact that the epitaxial GaP can be grown to desired doping levels and thicknesses, the effective tunneling barrier can be tailored to promote efficient spin injection into silicon. In order to fully understand the efficacy of the Fe/GaP/Si system as a spin injection tunnel barrier, we must first characterize all of the materials and interfaces involved. To this end, work will continue on the characterization of metal/GaP Schottky barriers, which in turn enables the materials level characterization of the GaP epitaxial layers through the use of such as advanced techniques as deep level transient and optical spectroscopies. Additionally, we intend to probe and will report the effects of GaP/Si interfacial quality on the barrier heights and other properties of resultant metal/GaP/Si devices. <SUP>1</SUP>Dash et al., Nature, 462, 491-494 (2009); <SUP>2</SUP> Grassman et al., Appl. Phys. Lett. 94, 232106 (2009).