Search

With the aggressive scaling of complementary metal-oxide semiconductor (CMOS) nearing an inevitable physical limit and its well-known power crisis, the quest for an alternative/augmenting technology that surpasses the current semiconductor electronics is needed for further technological progress. Spintronic devices emerge as prime candidates for Beyond CMOS era by utilizing the electron spin as an extra degree of freedom to decrease the power consumption and overcome the velocity limit connected with the charge. By using the nonvolatility nature of magnetization along with its direction to represent a bit of information and then manipulating it by spin-polarized currents, routes are opened for combined memory and logic. This would not have been possible without the recent discoveries in the physics of nanomagnetism such as spin-transfer torque (STT) whereby a spin-polarized current can excite magnetization dynamics through the transfer of spin angular momentum. STT have expanded the available means of switching the magnetization of magnetic layers beyond old classical techniques, promising to fulfill the need for a new generation of dense, fast, and nonvolatile logic and storage devices. All-spin logic (ASL) is among the most promising spintronic logic switches due to its low power consumption, logic-in-memory structure, and operation on pure spin currents. The device is based on a lateral nonlocal spin valve and STT switching. It utilizes two nanomagnets (whereby information is stored) that communicate with pure spin currents through a spin-coherent nonmagnetic channel. By using the well-known spin physics and the recently proposed four-component spin circuit formalism, ASL can be thoroughly studied and simulated. Previous attempts to model ASL in the linear and diffusive regime either neglect the dynamic characteristics of transport or do not provide a scalable and robust platform for full micromagnetic simulations and inclusion of other effects like spin Hall effect and spin-orbit torque. In this thesis, we propose an improved stochastic magnetization dynamics/time-dependent spin transport model based on a finite-difference scheme of both the temporal and spatial derivatives to capture the key features of ASL. The approach yields new finite-difference conductance matrices, which, in addition to recovering the steady-state results, captures the dynamic behavior. The new conductance matrices are general in that the discretization framework can be readily applied and extended to other spintronic devices. Also, we provide a stable algorithm that can be used to simulate a generic ASL switch using the developed model.