A Monte Carlo model was proposed to evaluate the reliability for single-event transient (SET) on combinational circuits. The masking effects, including logical, electrical, and timing masking effects, and the randomicity of SET are considered in the proposed model. Based on the reported model, an electrical masking model is also presented to apply for the reliability model. For ISCAS'85 benchmark circuits, the effects of SET duration, the masking effects, and the multi-SETs on the reliability are analyzed by the proposed model. The results are reported showing that the proposed model can be effectively used to analyze the reliability of a complex circuit.

In this paper, we propose a performance optimization scheme for all-spin logic device (ASLD); this scheme makes use of asymmetric tunneling layer to overcome the conductivity mismatch problem. Besides, for the material selection of device, this scheme makes advantage of silver as an energy efficient interconnect material to reduce the switching delay. The simulation results indicate that the high spin injection efficiency can be obtained by increasing the thickness of the transmitter-side tunneling layer or decreasing the thickness of the receive-side tunneling layer. Therefore, the structure of asymmetric tunneling layer is expected to be more effective for spin injection. Compared with the traditional interconnect that copper or aluminum is used as a channel material, the time for magnetic moment reversal and the spin current flowing into the output magnet are predicated to be optimized when silver is chosen as the channel material, which can attribute to the fact that the conductivity of silver material is larger, the spin diffusion length is longer, and the spin-orbit interaction is weaker. Meanwhile, when the critical switching current requires for the magnetization reversal is applied, the reliable working length of silver channel is significantly longer than that of copper or aluminum channel. These abovementioned results provide a new method for the performance optimization of ASLD.