Hello everybody,

As part of my research project, I have been trying to use comb-drives as
resonators for sensor applications. However, I have problems with
measuring the resonant frequency of the comb-drives and hoped that you
could advise me on how to fix the problem.

I used MUMPS for fabrication of the devices. The comb-drives have 9 or
15 moving fingers on each side with 10 or 16 fingers anchored to the
substrate. The thickness of the polysilicon layer is 2um, the finger
width/gap is 3um, the finger length is 40um, and the finger engagement
is 20um. Although the devices seem to operate normally at DC or low
frequencies (<10Hz), I can not make them resonate (or I am not picking
up the signal properly). The transresistance amplifier that I am using
has a gain of 5x10^6. I have tested it with 1nA input signals for the
frequency range of 10Hz to 500kHz and it is stable (although noisy for
such an input signal). Therefore, I do not think that I have problems
with the electronics. The test equipment that I use include
Lock-in-amplifiers, network/spectrum analyzers, function generators, DC
power supplied, and oscilloscopes. I usually sweep the frequency of the
input signal from about 5kHz before to 5kHz after the
calculated/simulated resonant frequency, with a sweeping time of 100sec
or longer.

The methods I have tried are (both in air and in vacuum on packaged or
unpackaged chips):
        1-Direct output signal (current) measurement: What I get using
this
method is a current whose amplitude depends on frequency. If the
substrate is grounded the current decreases with frequency, otherwise it
increases with frequency. My conclusion is this is purely due to the
input signal feedthrough. I do not see any resonant peak or change in
the phase of the output signal for the entire frequency range. The
displacement of the comb-drive, however, is clearly visible at low
frequencies (<15Hz).
        2-Frequency doubling: The same as above; i.e., there is no
resonant peak when the input signal frequency is half the resonant
frequency. Again, the frequency doubling effect can be visually verified
at low frequencies (<10Hz).
        3-Square wave input: I applied a square wave (~100Hz) to the
input of the comb-drive and then monitored the output for a damped sine
wave. However, I could not detect any thing other than a signal which
looked like a damped exponential.
        4-Modulation technique: I applied a high frequency carrier
signal (150kHz to 300kHz) to shuttle (or the input comb) and a sine wave
to the other port. Then, I amplified and demodulated the output signal,
but could not see any resonance peak.

I have tried to follow all the steps in you described in literature. The
signal amplitudes and DC voltage levels are around what people have
reported. I do not think that the devices are damaged, as the operate as
expected in low frequencies.

Could you let me know if I am missing something or doing something
wrong, or if there is another method that I can try (other than using
the strobe light or gated spectrum analysis).

Thank you in advance for your help.

Regards,
Behraad