A MEMS-based approach that uses temperature-dependent sensing responses to recognize chemical targets in untrained backgrounds

A major practical challenge for solid state microsensors is the detection of trace chemical species over time and in complex gas-phase backgrounds. We describe a MEMS-based, chemiresistive technology that has succeeded in such problems by combining oxide nanomaterials on microscale platforms, acquisition of dense temperature-dependent response data streams, and novel signal processing methods. Unlike the operation and analysis employed with many electronic noses, our higher dimensional approach captures surface electronic implications of changing adsorptive/reactive phenomena caused by rapid thermal cycling. Here, we demonstrate new capabilities for recognizing toxic targets over extended time periods, even in untrained backgrounds that contain aggressive contaminants at higher concentrations. The approach involves a moving-window, correlation-based methodology to identify chemical events and decouple the foreground conditions from the background.