HIGHLY SELECTIVE MEMS AMMONIA GAS SENSOR FOR RENAL FAILURE SCREENING AND MONITORING HEMODIALYSIS ENDPOINT

Breath gas analysis provides a way to diagnose diseases by detection of special
odors in the breath. It has been proven as a non-invasive tool for diagnosis. Some
gases from breath have been determined to be highly characteristic of specific
diseases. For instance, patients with renal failure exhaled ammonia gas with a higher
concentration than the normal level, because their kidneys failed to purify the blood
from waste materials. This dissertation reports the design, analytical modeling,
simulation, fabrication, and characterization of a MEMS ammonia gas sensor for
screening renal failure and determining hemodialysis endpoint. The sensor is designed
and fabricated following the standard PolyMUMPs process technology. The device is
operated in dynamic mode using electro-thermal actuation method and capacitive
sensing technique is used to measure the output voltage via MS3110 universal readout
circuit. A sensitive and selective material (PANI/MWCNTs/TiO2 nano-composite) is
coated on top of the moving plate. The mass of the moving plate changes according
to the amount of ammonia gas molecules adsorbed on the sensitive coating material.
The theoretical parameters of the device are derived based on analytical modeled
equations, verified with simulation and validated experimentally. Experimental results
show that the sensor is capable of detecting ammonia gas in air with concentration
ranging from a lower detection limit of 1 ppm to 10 ppm. This is within the range of
2 ppm and above for the abnormal level of ammonia in exhaled breath of patients
with renal failure. The sensitivity of the sensor is found to be 17.85 mV/ppm after
amplification and 0.088 mV/ppm prior to amplification. Finally, the response of the
sensor is measured towards various gases found in the exhaled breath including
89.80 %RH relative humidity, 10 ppm of acetone, methanol, and ethanol
concentration. The sensor is found to be highly selective towards ammonia gas
compared to the other gasses. The response of the sensor is found to be 0.78 mV/s
while the recovery is found to be 0.67 mV/s.