Variation-Aware Physics-Based Electromigration Modeling and Experimental Calibration for VLSI Interconnects

Abstract

Electromigration (EM) has emerged as a major reliability concern for interconnects in advanced technology nodes. The existing EM models are mostly either based on empirical formulation or calibrated based on simulations using finite element analysis. Moreover, the existing models do not consider variations in material parameters causing variations in the EM induced failure times, as typically observed in measurements. In this work, we first calibrate a physics-based EM model using measured failure times of damascene copper at accelerated test conditions. We then extend the model to obtain the distribution of material parameters leading to observed variations in failure times. This model, being physics-based, can be used for material and dimension exploration while being able to model and predict the variations in the failure times at various operating conditions.