Correction of Single Event Latchup Rate Prediction Using Pulsed Laser Mapping Test

Most classical approaches of single event effect rate prediction are based on the rectangular parallelepiped model of sensitive volume. However it is not clear about the number of sensitive volume in the device when predicting the in-flight single event latchup rate. As for memory device, there are two empirical practices to deal with the latchup sensitive volume number: one assumes that there is only one sensitive volume in the whole device; another assumes that there are as much sensitive volumes as the number of memory cells. The latchup sensitive volume number of a 4M-bits SRAM is determined as 63360 using pulsed laser mapping test in this work first. Based on the two assumed and measured sensitive volume number, the single event latchup rates of the device are calculated and compared. The results show that pulsed laser could be a powerful tool to obtain the real sensitive volume number in the device, which is significant for single event latchup rate prediction. The latchup rate will be either overestimated or underestimated with the assumption of the sensitive volume number as one or as much as the number of memory cells.