This work demonstrates the potential of micromechanical resonators as compact human breath detectors. Such sensors can be mounted on robotic platforms used in search and rescue missions for detection of survivors in disaster areas. Human breath carries higher concentrations of both moisture and carbon dioxide (CO 2 ) compared to surrounding air. To implement the sensors, single crystalline silicon resonators with frequencies in the MHz range were coated by thin films of polyethyleneimine (PEI), which is a strong absorbent of both moisture and CO 2 . The detection mechanism is based on the frequency shift caused by the added mass from the absorbed molecules. The resonators were tested in a humidity controlled chamber showing sensitivity towards humidity as high as 79 ppm/%RH from 30 to 58 %RH and 1.2 286 ppm/%RH from 58 to 85 %RH at room temperature. The response of the same coated resonators to CO 2 was also characterized showing an overall frequency shift of 3010 ppm. Finally, the resonators were exposed to human breath samples and showed frequency shifts as high as 1020ppm with response time constants around 10 seconds.

This work demonstrates the proof of concept for a selfcontrolled electronic frequency trimming technique for micromechanical single crystalline silicon resonators. Trimming is carried out via localized self-induced thermal oxidation occurring during extreme thermal actuation of the resonant structures. The cooling effect at resonance, discussed for the first time in this work, allows the localized oxidation to stop automatically as soon as the resonator frequency reaches the targeted actuation frequency applied to the structure. The viability of this concept is demonstrated by application of both off-resonance and atresonance actuation signals to fabricated resonators showing that as opposed to the off-resonance signals, the at-resonance signals with the same intensity do not lead to frequency trimming.