Correlation of SAW (Surface Acoustic Wave) Coating Responses with Solubility Properties and Chemical Structure Using Pattern Recognition

摘要

Twelve surface acoustic wave (SAW) device coatings were exposed to eleven chemical vapors and responses were correlated with solubility properties and coating structure to determine possible vapor/coating interaction mechanisms. Hydrogen bonding ability is implicated as a significant vapor/coating interaction mechanism. Pattern recognition schemes applied to the preliminary data aided in solubility property/response correlations. Principal component analysis demonstrated good separation of different classes of chemical vapors tested. Hierarchical clustering provided additional evidence of the correlations between solubility properties and the observed clustering. In addition, pattern recognition methods were used to determine potential selectivity of an array detector using these coatings. Learning techniques show that one fourth of the sensors can adequately separate compounds of interest from chemically similar interferences.