Search

All-spin logic (ASL) is a novel approach for digital logic applications wherein spin is used as the state variable instead of charge. One of the challenges in realizing a practical ASL system is the need to ensure non-reciprocity, meaning the information flows from input to output, not vice versa. One approach described previously, is to introduce an asymmetric ground contact, and while this approach was shown to be effective, it remains unclear as to the optimal approach for achieving non-reciprocity in ASL. In this study, we quantitatively analyze techniques to achieve non-reciprocity in ASL devices, and we specifically compare the effect of using asymmetric ground position and dipole-coupled output/input isolation. For this analysis, we simulate the switching dynamics of multiple-stage logic devices with FePt and FePd perpendicular magnetic anisotropy materials using a combination of a matrix-based spin circuit model coupled to the Landau–Lifshitz–Gilbert equation. The dipole field is included in this model and can act as both a desirable means of coupling magnets and a source of noise. The dynamic energy consumption has been calculated for these schemes, as a function of input/output magnet separation, and the results show that using a scheme that electrically isolates logic stages produces superior non-reciprocity, thus allowing both improved scaling and reduced energy consumption.

We investigate theoretically the behavior of electron spin states near graphene edges at low temperature in the presence of an external electric field. The graphene Hamiltonian is solved directly in the low energy regime including the Rashba spin-orbit interaction. Spin propagation in graphene and spin reflection at the edge are explored by calculating the probability of spin flips between the final and initial states. The spin scattering at the edge is affected by the initial location and the propagating direction of the electron with respect to the edge. Results for an example device structure including the effects of the non-uniform electrostatic field and potential near the graphene edge are presented.