An investigation of polymer-coated surface acoustic wave microsensor arrays for the determination of organic solvent vapors.

摘要

Several aspects of surface acoustic-wave (SAW) microsensor technology are explored. Arrays of polymer-coated SAW microsensors can be used to recognize and quantify individual vapors and the components of simple vapor mixtures found in workplace environments. Incorporation of SAW sensor arrays into portable analytical systems for such applications presents a number of challenges.; Integration of multiple sensors on the same substrate requires a process for depositing different absorbent polymer layers on designated area of each sensor. To address this need, in-situ UV-photopolymerization of vapor-phase monomers was explored. Thin films of two silylmethacrylate polymers were successfully grown without a free-radical initiator on working SAW resonators and subsequently exposed to a series of organic vapors. Reproducible response profiles and response patterns were obtained.; A commercial hand-held 'electronic nose' employing an array of four polymer-coated SAW sensors was then extensively modified to allow for the quantitative determination of various solvent vapors individually and in selected binary and ternary mixtures. Results based on experimental data, Monte Carlo simulations, and pattern recognition analyses indicate that with proper design revisions this instrument could serve as a personal multi-vapor exposure monitor in previously characterized occupational environments.; The 'limit of recognition' (LOR) has been used as a criterion for evaluating the performance of multi-sensor arrays for determining individual vapors. This study explored how the LOR could be applied to binary and ternary vapor mixtures. A general approach to defining and evaluating LORs in terms of the absolute and relative concentrations of the mixture components was developed. Correlations with the Euclidean distance(s) allowed reasonably accurate predictions of the mixture LORs. Results are considered in the context of using microsensor arrays as vapor recognition components in microanalytical systems.; Finally, artificial neural networks (ANN) were coupled with linear solvation energy relationship (LSER) models to develop a means of classifying unknown vapors from sensor array response patterns. Results show that the ANN-LSER model can classify previously uncalibrated vapors, but only if the ANN is trained with vapors from similar functional groups. Successful classification of more than 100 unknown vapors from eight functional-group categories was demonstrated on the basis of a 33-vapor calibration response library.