Development of a MEMS-based Dielectric Mirror A Report Submitted for the Henry Samueli School of Engineering Research Scholarship Program By ThanhTruc Nguyen June 2001 Faculty Supervisor Richard Nelson Electrical and Computer Engineering
Development of a MEMS-based Dielectric Mirror ThanhTruc Nguyen June 2001 The objective of this project was to develop processes for and then fabricate a dielectric mirror. In fact, this objective was exceeded and the project expanded to fabricate a Fabry-Perot etalon, which incorporates two dielectric mirrors. This type of device is useful for extracting the energy from narrow spectral regions of infrared radiation; ergo, this is a spectrometer. The device can be applied to many different applications. Generally the extracted energy contains desired information. In the case of fiber-optic communications, simultaneous transmitting a range of separately encoded wavelengths increases the information bandwidth, a technique known as dense wavelength division multiplexing (DWDM). The Fabry-Perot etalon is a candidate to accomplish the demultiplexing. Another application is gas sensing. It was decided to focus on building an etalon that is applicable to sensing of hydrocarbons in the middle infrared (approximately 3.3 micrometers wavelength). The etalon was fabricated as a solid dielectric (silicon dioxide) with a mirror on each surface. Each mirror is a stack of alternating silicon dioxide and silicon layers; each layer is 1/4 wavelength thick. The upper mirror consists of three ¼ wavelength layers of Si/SiO 2 /Si and the lower mirror consists of four ¼ wavelength layers of Si/SiO 2 /Si/ SiO 2. Figure 1 shows the reflection spectrum of a dielectric mirror that was fabricated on this project. For hydrocarbon gas sensing, the reflection must be high over the range of 3 to 4 micrometers for various gasses. Note that this mirror meets this requirement nicely. As shown in figure 1, a maximum reflectance of 97% reflectance occurs around 3.3 µm,
which is perfect for the desired detection of hydrocarbon gasses. That is an indication of a good mirror. Figure 1. Reflection spectrum of dielectric mirror. Figure 2 shows the transmission spectrum for the first etalon that was fabricated. As shown in the figure, there are two undesired sidebands, one is at short wavelength and one is at longer wavelength. The appearance of the two sidebands is due to errors in the thickness of the layers used in the two mirrors. The main peak appears at shorter wavelength, 2.7 µm, than the desired wavelength of approximately 3.3 µm., which is
optimal for the detection of methane or propane. The peak is shifted to a shorter wavelength because the cavity is thinner than the required thickness. Figure 2. IR transmission spectrum of an early etalon. Finally, the etalon was successfully fabricated. Figure 3 shows 85% transmittance at 3.373 µm, the desired wavelength. The position of the transmittance peak overlays the propane absorption peak, which indicates that the fabricated etalon is ideal to use for propane detection.
Figure 3. IR Transmission for Final etalon is ideal. Several methods were attempted for depositing the thin films. The best was RF sputtering. This method allowed all of the films to be sequentially deposited in one machine in one vacuum pump down and gave very smooth surfaces. The disadvantages were slow deposition (each etalon took 12 hours to produce) and highly stressed films. I believe that the stress can be reduced using thermal annealing. All of the processing was performed in the Integrated Nanosystems Research Facility (INRF) in the School of Engineering. I plan to continue this work during the summer quarter to explore new fabrication methods for the device and to development suspension springs for the upper mirror of the etalon.