Gas sensors have been used as the detection unit of gas analyzers. A low-cost and easy to fabricate gas analyzer can be made by integration a general purpose gas sensor with a microfluidic channel on a polymer substrate. However, ambient fluctuations influence the gas sensor characteristics and operation. In essence, these devices are vulnerable to drift since the interaction of the sensing pallet of the gas sensor with the surrounding air is a temperature- and humidity-dependent process which introduces drift terms to the output signal. These drift terms can also be considered as a controllable input while the sensor itself is considered as a multiple input and output system. In the present study, a methodology based on statistical techniques is introduced to perform the sensitivity analysis and to determine the contribution of each of the input parameters in the output response of the sensor used in a gas analyzer. A regression model is also applied to predict the response of the sensor to the temperature, humidity and the gas concentration values that have not been experimentally tested. Also, the effect of these environmental conditions on the three mentioned factors is studied using design of experiments (DOX) methods.
A low-cost nitric oxide gas sensor based on bonded gold wires
