Physics-based Simulation Algorithm Research For Electromigration Reliability Of Integrated Circuits

As integrated circuits(IC)enter the nanoscale process,the problem of electromigration on metal interconnects has become the most important factor limiting the development of integrated circuits.Existing IC electromigration reliability analysis algorithms are generally based on the Black equation or the Korhonen physical model.However,due to the complexity of IC structure and the diversity of factors,it is difficult for existing algorithms to consider all of them.These results in the calculation results of the existing algorithms being too conservative compared with the actual electromigration failure time.In addition,the scale off the IC is huge,which makes the time cost of the algorithm difficult to control.This paper analyzes and studies the power supply network with the most significant electromigration phenomenon in integrated circuits.The basic structure of the power supply network,the multi-segment interconnected metal wire,is used as the basic research object to analyze the cause of the phenomenon of electromigration.Based on the Korhonen physical model,a transient stress analysis method based on integral transformation is proposed.The experimental results were compared with COMSOL to verify the accuracy of the integral transformation method.Then,the integral transformation algorithm is extended to the whole chip power supply network.By modeling the power supply network as a time-varying resistor network,whether the maximum voltage drop of the power supply network caused by electromigration exceeds the threshold is used as the criterion for determining the failure of the power supply network,and the average failure time of the circuit is calculated.The experimental sample selects the actual design example provided by IBM as the existing method.By comparing the experimental results with the existing methods,the correctness and efficiency of the simulation algorithm are verified.