Investigation and detection of crack formations during Micro-electromechanical Systems (MEMS) resonator manufacturing and use

Abstract

Time reference devices have been widely used in the industry for the past decades due to their important role in different devices applications such as portable and wearable device. With the rapid technology improvements in the semiconductor industry, Micro-electromechanical systems (MEMS) resonators were presented as a time reference devices which offer a reduced cost and small size that facilitate the integration with the other electronics. Because of the microfabrication processes and operational modes for MEMS-resonators, defects are expected to be formed in different stages of the manufacturing and use. Those defects would be costly if have not been detected in early stages, hence, industry dedicates time and resources for better detection of defects. With the mass production of MEMS-resonator, manual inspection of defects is becoming an impractical solution, that led to develop automation systems for defects’ detection. This work studies the defects that may occur in MEMS-resonator in different stages and a studies defects detection methods. The study aims to investigate the microfabrication induced defects, the defects formation during the device operation and to develop an automatic defects inspection tool. To fulfil the study objectives, different MEMS-resonator devices were designed in order to study the effect of different parameters such as the geometry and dimensions on the robustness of the device. Standard microfabrication processes were implemented to observe the risk of defects formation during manufacturing. For the operation stage, FEM simulations were executed, and different mode shapes were selected to study the likelihood of mode shapes to produce defects for the device fragile parts. The last stage for defects inspection included implementation of automatic optical inspection (AOI) tool with a hardware setup for image collection and a software part for image processing and defection/robustness decision making, the result of the AOI was validated by the percentage of the detected defects and the occurrence of false-positives. The MEMS-resonators designs considered 2 frequencies of 24MHz and 32MHz, 3 different anchors and 7 different beam widths with average resulting quality factor(Q) of 11.68x10e3. Common defects of AlN particles residuals were observed with SEM as a microfabrication caused defects. 8.4% of the devices were defected electrically by resonating in specific spurious modes, 19.89% of the resonator were defected after performing the tape-peeling test. AOI tool was successful to reach 85.5% detection rate with 7.1% false-positive occurrences.