Impact of nuclear reaction models on neutron-induced soft error rate analysis

Abe, Shinichiro   

Terrestrial neutron-induced soft error rate (SER) analyses are performed by means of multi-scale Monte Carlo simulation with different nuclear reaction models used in PHYSERD. It is clarified that a choice of nuclear reaction models has a great effect on the SER prediction. Through validation of the nuclear reaction models, the most suitable nuclear reaction models for the soft error simulation are presented. Next, neutron-induced soft error is analyzed based on the multiple sensitive volume (MSV) model using the PHITS code. The result is compared with those obtained by PHYSERD based on event-by-event TCAD simulation and by simple sensitive volume (SSV) model using the PHITS code. From the comparisons of the collected charges between TCAD simulation and SV model calculations for each single event, it is found that the accuracy of the SER calculation based on the MSV model is improved by considering the spatial dependence of the charge collection efficiency.