Worst-case test vectors generation using genetic algorithms for the detection of total-dose induced leakage current failures

In this paper, we develop a methodology for generating the worst-case test vectors (WCTV) necessary to detect leakage current failures in a standard-cell based ASIC device exposed to a total ionizing dose. The methodology is based on using the genetic algorithm technique and as such it produces a near worst-case vector. The methodology is validated experimentally by applying the generated vectors on a test chip after exposure to total dose. In terms of total-dose induced leakage current failures, experiments shown that the near worstcase vector results are very close to those of the worst-case vector.