Microcircuit Defect and Counterfeit Detection via Silicon Geometry

Microcircuit testing, counterfeit and malicious implant detection face increasingly difficult challenges. The level of sophistication incurred by threat actors has steadily increased and we should expect this trend to continue. Existing techniques either fall short given the growth in complexity and density of modern ICs, or are costly or destructive. We propose a novel method, named SilGeo, for non-destructive assessment of authenticity, integrity, and reliability of ICs of arbitrary complexity and structure. Our technique uses high-frequency, analog-domain signaling to assess an unpowered target device. Our experimental evaluation confirms the effectiveness of the approach, with 100% accuracy in datasets of 200+ samples.