Comparison of Predictive Methods for Analysis of Electromigration

Abstract

Two techniques, lifetime test and temperature‐ramp resistance analysis to characterize electromigration (TRACE), were employed in electromigration studies of unpassivated Al‐1% Si metallizations at the wafer level. The objectives of the study were to: (i) apply the TRACE technique at the wafer level and compare the results with those obtained from life testing, and (ii) determine the impact of failure criterion on estimated kinetic parameters. In addition to the open‐circuit condition applied during life testing, a finite increase in resistance has been considered as a failure criterion to analyze its effect on estimated kinetic parameters. It was found that the open‐circuit criterion predicts a failure time of 779 min in good agreement with the experimentally determined value of 729 min. As the percentile increase in resistance (considered as failure) is decreased the estimated failure time decreases. TRACE experiments have been conducted at the wafer level for the first time. Failure times calculated from the TRACE data are compared with those obtained from the life‐test data at various failure criteria. It was found that calculated failure times overestimate measured failure times at a common failure criterion.