Rapid Detection of Analytes with Improved Selectivity Using Coated Microcantilever Chemical Sensors and Estimation Theory

Abstract

Rapid detection of analytes with improved selectivity is achieved though the use of estimation theory to analyze the response of polymer-coated microcantilever chemical sensors. In general, chemical sensors exhibit partial selectivity and can have relatively long response times. Using estimation theory, it is possible to make short-term response predictions from past data. This makes it possible to use the transient information (response time), often unique to an analyte/coating pair, to achieve an improvement in analyte species recognition while simultaneously allowing for a reduction in the time required for identification and quantification. An extended Kalman filter is used as a recursive online approach to refine the estimate of the sensor's future response. Both identification and quantification are thus possible as soon as the filter estimate achieves a high confidence level. Also, with improved selectivity, identification is possible using fewer sensors in an array.