A quality quantitative method of silicon direct bonding based on wavelet image analysis

The rapid development of MEMS (micro-electro-mechanical systems) has received significant attention from researchers in various fields and subjects. In particular, the MEMS fabrication process is elaborate and, as such, has been the focus of extensive research inquiries. However, in MEMS fabrication, component bonding is difficult to achieve and requires a complex approach. Thus, improvements in bonding quality are relatively important objectives. A higher quality bond can only be achieved with improved measurement and testing capabilities. In particular, the traditional testing methods mainly include infrared testing, tensile testing, and strength testing, despite the fact that using these methods to measure bond quality often results in low efficiency or destructive analysis. Therefore, this paper focuses on the development of a precise, nondestructive visual testing method based on wavelet image analysis that is shown to be highly effective in practice. The process of wavelet image analysis includes wavelet image denoising, wavelet image enhancement, and contrast enhancement, and as an end result, can display an image with low background noise. In addition, because the wavelet analysis software was developed with MATLAB, it can reveal the bonding boundaries and bonding rates to precisely indicate the bond quality at all locations on the wafer. This work also presents a set of orthogonal experiments that consist of three prebonding factors, the prebonding temperature, the positive pressure value and the prebonding time, which are used to analyze the prebonding quality. This method was used to quantify the quality of silicon-to-silicon wafer bonding, yielding standard treatment quantities that could be practical for large-scale use. (paper)