Diffusion bonding criterion based on real surface asperities: Modeling and validation

Abstract A new mathematical model based on real surface morphology is derived to accurately predict the void closure time, bonding ratio and void distribution state in diffusion bonding process, which considers short-wavelength asperities distributing on long-wavelength ones. The wavelength of long-wavelength asperities plays a key role in calculating the contribution of grain boundary diffusion mechanism. New formulas are built to calculate the transferred mass induced by surface source mechanism. The contributions of main mechanisms to void closure can be accurately predicted by using the proposed model. It is verified with a case containing long- and short-wavelength asperities and two cases consisting of the surface asperities with the same height and wavelength. The predicted void size and distribution state are consistent with experimental results in case one. The maximum and standard deviations between prediction and experiment are 3.2 % and 1.0 % for case two, 13.5 % and 7.3 % for case three, respectively. The present model shows a higher prediction accuracy than previous ones.