| dc.description.abstract | The rapid development of wireless communication systems raises the high demand for high Q factor resonators. To fabricate a high Q factor resonator, a precise membrane is important. It requires fabricating different sizes of membranes of resonators in one device. The motivation of this thesis is to simultaneously deal with 3 problems occurring in the fabrication process: notching issue, aspect ratio dependent etching (ARDE) and off-vertical sidewall.
A membrane was fabricated consisting of SiO2, metal electrodes and AlN in the project. With AlN and bottom metal electrode over an SOI wafer, Cr was deposited on the AlN as the top electrode, afterwards, a photoresist pattern was made by lithography, finally, the membrane was released by DRIE on the Si side. In the process of DRIE, the emphasis is to ensure good verticality, notching-free bottom and less ARDE for different cavities. One group of experiments consisting of 4 trials was conducted to gain enough experience before designing 2 extra optimizing experiments, sequentially the results of the six experiments were analyzed in comparison with the benchmark recipe. The result of the optimizing recipe showed that all three issues were mitigated, more vertical side wall and notching free trench was achieved on the AlN wafer, more small size cavities were open, and the final membrane is characterized by both optics microscope and SEM. The results showed that a multiple-step Bosch process is an effective way to fix these three problems simultaneously, and the mechanism of the DRIE on Si was discussed.
The key to this Bosch process includes 3 aspects: mitigating charge separation at the interface, a controllable passivating layer on the sidewall and a suitable bias power. This is important to achieve a high aspect ratio sidewall with a smooth, vertical sidewall, and nothing-free bottom, as well as the smallest ARDE.
Due to the longtime of tools down and limited lab availability, we only investigated the influence of change in bias power and pulsed bias on the Si trench. The influence of temperature and new gas composition hasn’t been explored.
Finally, it can be concluded that bias power and ICP pulse are two key factors to control these problems, and all these will contribute to a high Q factor resonator. | |