Simulating Electromigration in 3-D Microelectronics

Faculty Mentor Information

Dr. Mahmood Mamivand, Boise State University; and Dr. Ashkan Farazin, Boise State University

Presentation Date

7-16-2026

Abstract

In this study, we present a multiphysics finite element simulation that replicates a recent electrical-thermal-stress coupling framework based on atomic flux divergence for electromigration (EM). The physical field distributions across the 3-D interconnect geometry show strong structural and qualitative agreement with the reference model, successfully locating extreme current crowding, Joule heating, and thermomechanical stress mismatches at the critical TSV–RDL junction. Quantitatively, the simulated maximum temperature is in agreement with minimal error, while the peak current density aligns closely within the same order of magnitude. Although the peak structural stress showed higher variance compared to the reference data, a difference attributed to microstructural constraints and boundary condition sensitivities, the tracked maximum atomic depletion successfully characterizes localized, early-stage EM void incubation rather than the final macroscopic area failure threshold.

This document is currently not available here.

Share

COinS
 

Simulating Electromigration in 3-D Microelectronics

In this study, we present a multiphysics finite element simulation that replicates a recent electrical-thermal-stress coupling framework based on atomic flux divergence for electromigration (EM). The physical field distributions across the 3-D interconnect geometry show strong structural and qualitative agreement with the reference model, successfully locating extreme current crowding, Joule heating, and thermomechanical stress mismatches at the critical TSV–RDL junction. Quantitatively, the simulated maximum temperature is in agreement with minimal error, while the peak current density aligns closely within the same order of magnitude. Although the peak structural stress showed higher variance compared to the reference data, a difference attributed to microstructural constraints and boundary condition sensitivities, the tracked maximum atomic depletion successfully characterizes localized, early-stage EM void incubation rather than the final macroscopic area failure threshold.