Naval Research Laboratory, Washington, DC
The electronís spin angular momentum is one of several alternative state variables under consideration on the International Technology Roadmap for Semiconductors for processing information in the fundamentally new ways which will be required beyond end-of-roadmap CMOS technology. Electrical injection / transport of spin-polarized carriers is prerequisite for developing such an approach. While significant progress has been realized in GaAs , little has been made in Si, despite its overwhelming dominance of the semiconductor industry. We report successful injection of spin-polarized electrons from an Fe film into Si(001) n-i-p doped heterostructures, and spin transport across the Si/AlGaAs interface (1). The circular polarization of the electroluminescence (EL) due to radiative recombination in the Si tracks the Fe magnetization, confirming that the electrons originate from the Fe. The polarization reflects Fe majority spin, consistent with the common delta_1-symmetry of the Fe majority and Si(001) conduction band (CB). The spin polarization in the Si is ~30% at 5K, with significant polarization extending to at least 125K. In Si/AlGaAs/GaAs quantum well (QW) structures, the spin polarized electrons drift under applied field from the Si and recombine in the GaAs QW, where the polarized EL can be quantitatively analyzed, yielding an electron spin polarization of 10%. Spin transport occurs despite the poor crystalline quality of Si epilayers on GaAs, the 0.3 eV CB offset (Si band lower), and the fact that the sample surface was exposed to air before growth of the Si on the AlGaAs/GaAs. This approach injects spin-polarized electrons near the Si CB edge with near unity conversion efficiency and modest bias voltages. The realization of efficient electrical injection and significant spin polarization using a simple tunnel barrier compatible with "back-end" Si processing should greatly facilitate development of Si-based spintronics. Prospects for electrical detection in silicon in a lateral transport geometry are also discussed. This work was supported by ONR and core NRL programs.
1. B.T. Jonker et al, Nature Physics 3, 542 (August 2007)