In this article we extend the currently established diffusion theory of spin-dependent electrical conduction by including spin-dependent thermoelectricity and thermal transport. Using this theory, we propose new experiments aimed at demonstrating nov
el effects such as the spin-Peltier effect, the reciprocal of the recently demonstrated thermally driven spin injection, as well as the magnetic heat valve. We use finite-element methods to model specific devices in literature to demonstrate our theory. Spin-orbit effects such as anomalous-Hall, -Nernst, anisotropic magnetoresistance and spin-Hall are also included in this model.
We measured the anomalous-Nernst effect and anisotropic magnetoresistive heating in a lateral multiterminal Permalloy/Copper spin valve using all-electrical lock-in measurements. To interpret the results, a three-dimensional thermoelectric finite-ele
ment-model is developed. Using this model, we extract the heat profile which we use to determine the anomalous Nernst coefficient of Permalloy Rn=0.13 and also determine the maximum angle of theta=8 degrees of the magnetization prior to the switching process when an opposing non-collinear 10$^{circ}$ magnetic field is applied.
We present time-resolved Kerr rotation measurements of electron spin dynamics in a GaAs/AlGaAs heterojunction system that contains a high-mobility two-dimensional electron gas (2DEG). Due to the complex layer structure of this material the Kerr rotat
ion signals contain information from electron spins in three different layers: the 2DEG layer, a GaAs epilayer in the heterostructure, and the underlying GaAs substrate. The 2DEG electrons can be observed at low pump intensities, using that they have a less negative g-factor than electrons in bulk GaAs regions. At high pump intensities, the Kerr signals from the GaAs epilayer and the substrate can be distinguished when using a barrier between the two layers that blocks intermixing of the two electron populations. This allows for stronger pumping of the epilayer, which results in a shift of the effective g-factor. Thus, three populations can be distinguished using differences in g-factor. We support this interpretation by studying how the spin dynamics of each population has its unique dependence on temperature, and how they correlate with time-resolved reflectance signals.
