Low-temperature, wafer-level, gold thermocompression bonding: Modeling of flatness deviations and associated process optimization for high yield and tough bonds

Konstantinos Stamoulis, Christine H. Tsau, S. Mark Spearing

Research output: Contribution to journalArticleAcademicpeer-review

5 Citations (Scopus)

Abstract

Wafer-level, thermocompression bonding is a promising technique for MEMS packaging. The quality of the bond is critically dependent on the interaction between flatness deviations, the gold film properties and the process parameters and tooling used to achieve the bonds. The effect of flatness deviations on the resulting bond is investigated in the current work. The strain energy release rate associated with the elastic deformation required to overcome wafer bow is calculated. A contact yield criterion is used to examine the pressure and temperature conditions required to flatten surface roughness asperities in order to achieve bonding over the full apparent area. The results are compared to experimental data of bond yield and toughness obtained from four-point bend delamination testing and microscopic observations of the fractured surfaces. Conclusions from the modeling and experiments indicate that wafer bow has negligible effect on determining the variability of bond quality and that the well-bonded area is increased with increasing bonding pressure. The enhanced understanding of the underlying deformation mechanisms allows for a better controlled trade-off between the bonding pressure and temperature.

Original languageEnglish
Article number06
Pages (from-to)42-52
Number of pages11
JournalProgress in Biomedical Optics and Imaging - Proceedings of SPIE
Volume5716
DOIs
Publication statusPublished - 21 Jul 2005
Externally publishedYes
EventReliability, Packaging, Testing and Characterization of MEMS/MOEMS IV - San Jose, CA, United States
Duration: 24 Jan 200525 Jan 2005

Fingerprint Dive into the research topics of 'Low-temperature, wafer-level, gold thermocompression bonding: Modeling of flatness deviations and associated process optimization for high yield and tough bonds'. Together they form a unique fingerprint.

Cite this