The demand for high performance, high packing density and low power dissipation integrated circuits (IC) has caused downward scaling of feature sizes in very large-scale integration (VLSI) fabrication. However with the scaling of interconnect feature size; the interconnect delay, which is one of the key factors determining the overall performance of devices is severely penalized [1]. Therefore copper (Cu), which is well known for its lower resistivity and excellent resistance to electromigration (EM) and stress-induced voiding as compared to aluminum (Al) and its alloy, was introduced as the material for nanoscale interconnects [2-9]. However, when the efforts to develop an effective Cu etch process failed, dual damascene which is an elegant technique of inlaying metal for interconnects throughout the back-end-of-line (BEOL) becomes the preferred choice. Even though Cu dual damascene has already been proven to be process-friendly and costs effective, its reliability robustness is strongly process and structural dependent. Starting from 0.13 µm process, yield and reliability have become critical issues for the integration of Cu and low-k dielectrics. Poor yield is observed at wafer sort and even for those apparent good chips; the failure rates are unacceptable during accelerated life testing. Physical analysis has shown the formation of stress-induced voids to be either within or beneath vias [10-35]. As such, understanding on the phenomena of stress-induced voiding in Cu interconnects must be established so as to resolve these issues.