Linjun Cao

Linjun Cao is a fourth-year Ph.D. student in Materials Science and Engineering Program under the supervision of Prof. Paul Ho. Her research focuses on microstructure analysis of nanoscale Cu interconnects and its effects on electromigration reliability.

Electromigration (EM) is well recognized to be a key reliability concern for Cu damascene interconnects in advanced integrated circuits. As the Cu interconnect lines continue to scale with each technology node, the EM reliability becomes increasingly complex due to changes in Cu microstructure. The emergence of polycrystalline Cu grain structure in fine lines for the 65 nm node technology and beyond markedly reduced the Cu EM reliability due to the addition of grain boundary diffusion to the overall mass transport. This raises important and challenging questions for the future developments of Cu interconnect technology. The primary objective of Linjun’s PhD study is to investigate the effects of scaling and grain structure on EM reliability of Cu interconnects.

To study the microstructure effect on EM, detailed information on grain structures and their contribution to mass transport is required. In Linjun’s research, a recently developed high-resolution electron diffraction technique is used to characterize the grain structure of the 45 nm node Cu interconnects. The results yield detailed information on grain orientation distributions (crystallographic texture), grain size distributions and grain boundary characteristics at a spatial resolution of 1-2 nm. A dominant sidewall growth of {111} grains is observed, reflecting the increasing importance of the interfacial energy in controlling grain growth as Cu interconnects continue to be scaled down in dimensions.

The grain structure statistics are used to identify the local critical flux divergent sites for EM-induced void formation and then to evaluate the statistical behavior of EM lifetime using a microstructure-based model. In the Cu line, void formation occurs at triple junctions bounded by neighboring Cu grains having different sizes and crystallographic orientations. The net mass flow between the adjacent grains determines flux divergence. To evaluate the flux divergence, the orientation-dependent interfacial diffusivity and grain boundary diffusivity are determined from the resistance traces observed in the EM tests. Then, the poly-grain cluster length distributions obtained from high-resolution orientation mapping technique and the extracted interface and grain boundary diffusivities are put into the statistical model to analyze EM lifetime and statistics.

This research establishes a direct correlation between microstructure of Cu nanolines and experimentally observed void formation kinetics as well as EM lifetime distribution. Analysis of the scaling effect on microstructure evolution of nanoscale Cu interconnects will provide the basic information required to develop a better understanding and control of the microstructure for improving EM reliability of future Cu metallizations.

Color-coded orientation map for 70 nm Cu line with its corresponding texture plots along 3 orthogonal directions. Color codes for orientations are represented in the standard stereographic triangle.