Variability Cancellation to Improve Diagnostic Performance of Testing through Silicon Vias in Power Distribution Network of 3D-IC

Abstract

A novel method is proposed to improve diagnostic performance and to decrease probability of false diagnoses in testing open defects of TSVs (Through Silicon Vias) in PDN (Power Distribution Network) of 3D-ICs (Three Dimensional Integrated Circuits). The conventional method measures resistances between microbumps placed exactly under TSVs and detects defects of the TSVs by changes of the resistances. But it suffers from manufacturing variabilities which also cause changes of the resistances even without any defect, and its diagnostic performance is not enough for practical applications. In this paper, a method is proposed which measures resistance of another microbump pair to cancel the variabilities in addition to each resistance for detecting an open defect, and improves its diagnostic performance by cancelling the variabilities in the detecting resistance. Experimental simulations are conducted using a 3D-IC example with two dies. The results show that the diagnostic performances by the conventional method are improved by the proposed variability cancellation and reach the practical level.