Two-Dimensional Markov Chain Analysis of Radiation-Induced Soft Errors in Subthreshold Nanoscale CMOS Devices

Radiation-induced soft errors have been a reliability concern for logic integrated circuits since their emergence. Feature-size and supply-voltage reduction require the analysis of soft-error sensitivity as a function of technology scaling. In this paper, an analytical framework based on Markov chains and queue theory is presented for computation of alpha-particle-induced soft-error rates of a flip-flop operated in the subthreshold regime. The proposed framework is capable of reflecting the technology parameters such as supply voltage Vdd, channel length, process-induced threshold variation, and operating temperature. As an example, the framework is used to investigate the mean time to error of flip-flops built in a 32 nm fully-depleted silicon-on-insulator technology operating in the subthreshold regime subject to two limiting fluxes of alpha particle radiation: high at 100 (α/h.cm²) and ultra-low alpha (ULA) emission 0.002 (α/h.cm²).