Spin accumulation at in-situ grown Fe/GaAs(100) Schottky barriers measured using the three- and four-terminal methods

We examined the spin accumulation in Fe/n-GaAs Schottky barriers to evaluate the accuracy of the three-terminal (3T) and four-terminal (4T) measurement geometries. A fully epitaxial Fe/n-GaAs junction was grown in situ using cluster molecular beam epitaxy without breaking the vacuum to exclude the formation of an oxide layer or surface roughness at the interface during intermixing. The spin resistance of the 4T nonlocal spin valve (ΔR_NLSV = 0.71 Ω) was twice the value obtained using the 4T Hanle effect method (ΔR_4TH = 0.35 Ω) at 10 K, as predicted theoretically, and this value remained constant over the temperature range examined, from 10 K to 77 K. The temperature-dependent spin lifetimes measured using the 3T and 4T Hanle effects exhibited similar behaviors. Although the spin resistance obtained using the 3T Hanle effect was enhanced compared with that obtained using the 4T effect, it was reasonable to conclude that the spin signals obtained from the 3T and 4T measurements originated from spin accumulation in n-GaAs due to the absence of an oxide tunnel barrier or a well-defined interface in our samples. These results completely ruled out any other sources of artifacts.