UC Berkeley

Thermal infrared detectors based on MEMS bimorph beams have the potential to exceed the performance of current uncooled thermal infrared cameras both in terms of sensitivity and cost. These cameras are part of a rapidly growing industry are used for a vast array of applications such as military and civilian night vision, industrial monitoring, and medical imaging. Many researchers have explored the use of metal-ceramic MEMS bimorphs for this application even though it has long been acknowledged that polymer-ceramic bimorphs would be superior. However, because of the difficulties of designing and fabricating MEMS systems based on polymer-ceramic bimorphs, little progress has been made towards their development.

This dissertation describes the initial design, fabrication, and testing of thermo-mechanical infrared sensors based on MEMS polymer-ceramic bimorph beams. Sensors based on bimorphs composed of both the biopolymer chitin on poly-silicon and OCG-825 photoresist on poly-silicon were fabricated and tested. Chitin bimorphs were fabricated using a novel photolithographic chitosan process previously developed for this research.

A sensor design based on a residual stress and ambient temperature compensating geometry and which includes novel features such as vertically aligned thermal isolation regions and selective shielding is presented. Simplified sensors were tested using an optical readout method where the deformation of the sensors was observed as variations in the intensity of visible light reflected to a digital camera. In order to obtain quantitative measurements, image analysis was performed. While the feasibility of simply observing the average brightness of the light reflected from a sensor was demonstrated, several image processing algorithms were tested and shown to increase the signal to noise ratio. An IR source approximating a blackbody was combined with a series of filters and lenses to limit transmission of light to the sensor to wavelengths from approximately 1.0 to 3.6 microns.

A periodic signal was produced by coupling a mechanical chopper wheel with the IR source. The sensor was able to detect these signals at frequencies of at least 5 Hz. By comparing the sensor signal to a known rate of warming of the IR source and the measured noise level at equilibrium, a noise equivalent temperature difference of as low as 360 mK was measured. In light of this encouraging and clear proof of concept, suggestions for achieving performance gains and developing novel imaging systems based on polymer-ceramic bimorphs through future research efforts are offered.