Evaluating Soft Error Reliability of Combinational Circuits Using a Monte Carlo Based Method

The aggressive scaling of nanometric VLSI technologies increases the integrated circuits susceptibility to radiation, especially to single event effects. The reduction of supply voltage and transistors dimensions increases the probability of such events occurring if an ionizing particle hits the device near a reverse-biased PN junction. Depending on the hit location and particle energy, this may generate transients of distinct magnitudes. This work proposes a Monte Carlo-based procedure to analyze circuit reliability against radiation-induced transient faults. The procedure evaluates the robustness of circuits to transient faults through a fault injection campaign identifying the radiation-sensitive nodes and the error rate of the circuit under evaluation. Additionally, it explores the physical design characteristics of logic gate layouts. This is important to estimate better the error rate, relating the total sensitive area with the expected particle flux. With enhanced accuracy, the proposed method can be applied to estimate the reliability of combinational circuits against transient faults in different technologies. The effectiveness of the proposed methodology is proven by applying it to several benchmark circuits with distinct sizes, which are mapped to different libraries in terms of gate complexities.