Hostname: page-component-8448b6f56d-m8qmq Total loading time: 0 Render date: 2024-04-23T20:28:59.213Z Has data issue: false hasContentIssue false

Kinetics of Void Drift in Copper Interconnects

Published online by Cambridge University Press:  01 February 2011

Zung-Sun Choi
Affiliation:
zung@mit.edu, Massachusetts Institute of Tecnology, Materials Science and Engineering, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States, 617-253-8563, 617-258-6749
Reiner Mönig
Affiliation:
moenig@mit.edu, Massachusetts Institute of Tecnology, Materials Science and Engineering, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States
Carl V. Thompson
Affiliation:
cthomp@mit.edu, Massachusetts Institute of Tecnology, Materials Science and Engineering, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States
Michael Burns
Affiliation:
burns@rowland.harvard.edu, Rowland Institute at Harvard, 100 Edwin H. Land Boulevard, Cambridge, MA, 02142, United States
Get access

Abstract

We have observed the real-time behavior of electomigration-induced voids in both passivated and unpassivated copper interconnects in a Scanning Electron Microscope (SEM), and correlated void nucleation, growth, drift and stagnation with post-electromigration crystallographic microanalyses carried out using Electron Back-Scattered Diffraction (EBSD) analysis. Voids that nucleate at various locations along the interconnects often drift toward the cathode, where they grow, coalesce, and eventually cause electrical failure. In-situ SEM observations allowed for the tracking of void shapes and drift rates over long (multi-grain) distances. Changes in the size and the velocity of the voids were observed when the voids passed through different grains. These changes are attributed to the difference in diffusivity for different grain orientations. In passivated lines, voids were often trapped at individual grain boundaries, where they grew to cause failure, or de-trapped to continue to drift toward the cathode. In unpassivated lines, voids did not drift, but instead always nucleated and grew and grain boundaries. Locations at which voids grew in unpassivated lines, or at which voids were trapped and grew in passivated lines, were correlated with the crystallographic orientations of “upwind” and “downwind” grains. From these analyses, we find that the average electromigration interface diffusivities (z*D) as a function of grain orientation are ordered according to {100} > {111} > {110}. Quantitative analysis of void dynamics, correlated with crystallographic microanalyses, provides important data for modeling of electromigration-induced failure, and for process-optimization for improved reliability.

Type
Research Article
Copyright
Copyright © Materials Research Society 2006

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

[1] Int'l. Technology Roadmap for Semicondutors (http://public.itrs.net/).Google Scholar
[2] Hu, C.K. et al, Proceedings of IEEE International Interconnect Technology Conference (IEEE, Piscataway, NJ, 1999), p267.Google Scholar
[3] Lloyd, J.R., Clemens, J., Snede, R., Microelectronics Reliability, 39, 1595 (1999).Google Scholar
[4] Hau-Riege, S.P., J. Appl. Phys. 91, 2014 (2002).Google Scholar
[5] Hau-Riege, C.S., Marathe, A.P., and Pham, V., Proceedings of the 41st Annual IEEE Int'l. Rel. Phys. Symp., Dallas, 2003 (IEEE, New York, 2003), p.173.Google Scholar
[6] Hau-Riege, C.S. and Thompson, C.V., Appl. Phys. Lett. 78, 3451 (2001).Google Scholar
[7] Zschech, E., Meyer, M.A., and Langer, E., MRS Symp. Proc. 812, F7.5.1 (2004).Google Scholar
[8] Borgensen, P., Korhonen, M.A., and Li, C.Y., AIP Symp. Proc. 263, 219 (1992).Google Scholar