Spin Manipulation in Semiconductors and Spintronic Ultrafast Nanodevices

Keywords:
spin injection, spin extraction, spin transistor, semiconductors, ferromagnets, Schottky barrier, delta doping

Abstract:
Current ITRS-2003 semiconductor roadmap suggests that by the year 2012-2015 current CMOS technology will run out of steam, when the feature size will be about 20-16 nm and the Moore’s law would end. Industry, therefore, starts paying increasing attention towards new technologies enabling it to stay on the roadmap beyond next 10 years. One of very promising candidate technologies is the spin injection ultrafast low power devices, some of them described below. A consistent microscopic theory of injection of spin polarized electrons from a ferromagnet (FM) into a semiconductor (S) is suggested. It describes tunneling and emission of the electrons through modified FM-S Schottky barrier [1, 2]. The efficient spin injection can be achieved in a system with very thin heavily doped semiconductor layer (delta-doped layer) formed at the FM-S interface. The barrier is made practically transparent to electrons at energies above a certain threshold, close to the peak in the (minority) density of states in the ferromagnet slightly above the Fermi level. We calculate the nonlinear I-V characteristic of the junction and a spin accumulation in S. We show that in a nonlinear regime of current saturation the maximal spin polarization of current and electron density (spin accumulation) can approach 100% at room temperature and relatively low current density in a nondegenerate high-resistance semiconductor [1, 2]. The condition for an efficient spin injection through the modified FM-S Schottky is opposite to the Rashba condition and no problem with the so-called conductivity mismatch actually arises. The correct condition on the tunnel resistance of the mosified Schottky barrier reads, where is the spin-flip length and is the conductivity of a semiconductor, which is easy to meet in lightly doped semiconductors, not heavily doped implied, opposite to what is suggested by the incorrect “conductivity mismatch” theory. Several spin injection nanodevices can be built on this effect, including an ultra-high frequency spin valve, an amplifier, a frequency multiplier, and a square-law detector. The spin precession of the electrons in the semiconductor layer is produced by a magnetic field produced by a base current in an adjacent nanowire. The emitter current can be controlled with frequencies up to several 100 GHz [1, 3, 4]. It is interesting to explore the spin transport devices based on low atomic weight and organic materials, where the spin-relaxation time may be extraordinarily long (~ms). Related issues pertaining to transport in organic/molecular systems will also be explored [5]. [1] A.M. Bratkovsky and V.V.Osipov, Phys. Rev. Lett. 92, 098302 (2004). [2] V.V.Osipov and A.M.Bratkovsky, cond-mat/0307030. [3] V.V.Osipov and A.M.Bratkovsky, Appl. Phys. Lett. 84, 2118 (2004). [4] A.M.Bratkovsky and V.V.Osipov, cond-mat/0307656. [5] B. Larade and A.M. Bratkovsky, Phys. Rev. B 68, 235305 (2003).

Back to Program

Nanotech 2005 Conference Program Abstract