We propose to investigate the use of an array of coated surface- acoustic-wave (SAW) microsensors for the identification and quantitation of organic vapors from several different chemical classes. This work is motivated by the need for improved sensor technology in direct-reading industrial-hygiene monitoring equipment. Currently available portable monitoring instruments for organic vapors can neither identify unknown vapors nor discriminate between the components of vapor mixtures. An array of microsensors that provides a unique response pattern for a given vapor can be used to determine the identify and concentration of the vapor alone or in a mixture. The small size and low power requirements of the sensor array will facilitate incorporation into miniaturized instrumentation suitable for real-time personal monitoring and respirator-cartridge breakthrough applications. Eighteen different chemically sensitive materials will be tested as sensor coatings for exposure to 50 organic vapors representing 13 chemical classes. A 158MHz SAW sensor (-0.8cm2 area) fabricated in our laboratory will be coated sequentially with each of the coating materials and exposed to each target vapor over a relevant range of concentrations. The sensor responses will be stored and then analyzed collectively using pattern recognition methods thereby simulating an array of sensors. The ability to identify vapors individually as well as in binary and ternary mixtures will be determined. Experimental verification of the results predicted from the pattern recognition analysis for binary mixtures will be performed on a subset of the test vapors. Most of the coating materials to be used consist of polymers or oligomers selected to provide partial selectivity based on the differential solubility of each vapor. Several room-temperature liquid crystals will also be tested as coating materials. The anisotropic nature of the liquid crystals can provide discrimination based on subtle size and shape features of otherwise similar compounds. In addition to measuring the steady-state sensor response for a given vapor concentration, we will also monitor the response time for each coating/vapor pair to obtain a estimate of the vapor diffusion coefficient. Adding this feature to the pattern recognition analysis should result in further improvements in the selectivity of the array based on differential vapor diffusion rates.
