Reliability modeling for ULSI interconnects

Electromigration (EM) and stress-induced voiding (SIV) are the two major reliability concerns for metal interconnects in integrated circuits. In particular, with the dimensions of interconnect scaled into nano regime and the inclusion of low-k materials as dielectrics, the reliability aspects for on chip interconnects are becoming more challenging. In this work, both EM and SIV failure physics are investigated using finite element based modeling and verified through experiments. The most fundamental aspect of EM is diffusion, where electron wind force (EWF) is considered as the only source of driving force for EM mass transport. As interconnect line width becomes narrow, the effects of the surrounding materials on EM can no longer be accounted for by the modification of the atomic diffusivity as in the traditional diffusion path approach. In this work, a modified EM modeling methodology is proposed to improve the EM modeling accuracy. This methodology is based on the driving force approach and its mathematical formulations are derived based on Green’s theorem. Three important driving forces are considered in this work: electron wind force (EWF), temperature gradient induced driving force (TGIDF), and thermo-mechanical stress gradient induced driving force (SGIDF). The formulations are implemented through finite element analysis (FEA), and the EM void nucleation and its growth can be simulated through the developed static and dynamic simulation FEA codes. It is found that the thermo-mechanical stress gradient induced driving force is the dominant driving force for EM failure in M2 test structures. The modeling results are consistent with our experimental results on reservoir effect structures.DOCTOR OF PHILOSOPHY (EEE