New approaches to injecting spin polarized currents into semiconductors
Final Report Abstract
Our goal in this project was to experimentally realize and test the Spin-FET proposed by Gurzhi et al. To accomplish this we split the task up in 4 steps: Simulation of the device with realistic parameters Development of robust gating technology Fabrication of a preliminary device with MOS gating for proof of principle Fabrication of the full Spin-FET The simulation of the device was necessary to find the right dimensions for the sample and the right conductivities for the individual layers to get a highly spin polarizes current. The FES simulation was carried out with an in-house program. The results of the simulation are very promising because they show us, that the necessary conductivities (and therefor dopings) are well in an established range. Also the dimensions of the sample are accessible by standard EBL. One the downside we learned that the difference in actual electron numbers for the different spin polarizations will be too small to be detected by MOKE technique. To fabricate robust gates with high k dielectrics we introduced the ALD technology to our cleanroom. Advantages of this technique are the high precision of the deposited layer thickness its uniformity. It was additionally interesting for our group, because it allows to deposit at rather low temperatures. This is important for the intensely studied material HgTe witch is sensitive to thermal load. Many spin injection experiments make use of thin tunnel barriers. Thin, high resistive and loophole free layers can be fabricated especially well controlled by ALD technology. We performed tests on our in-house built ALD system to find robust growth parameters and typical growth rates. The deposited layers were characterized by XRR- and AFM-measurements, as well as electrically probed in terms of leakage and gate efficiency. The layers show high quality and the values we find a in good approximation with the once reported in the community. Since we found the MOKE technique to be unfeasible for our experiment, we tried a different approach. By using ferromagnetic contacts we would get all electric devices. So far none of our experiments with the ferromagnetic metal Cobalt exhibited convincing spin detection, but exploration of these ideas is ongoing.
Publications
- “Removal of GaAs growth substrates from II–VI semiconductor heterostructures”, Semicond. Sci. Technol. 29, 045016 (2014)
S Bieker, P R Hartmann, T Kießling, M Rüth, C Schumacher, C Gould, W Ossau and L W Molenkamp
(See online at https://doi.org/10.1088/0268-1242/29/4/045016)
