Search

An array of 20 compositionally different carbon black−polymer composite chemiresistor vapor detectors was challenged under laboratory conditions to discriminate between a pair of extremely similar pure analytes (H_(2)O and D_(2)O), compositionally similar mixtures of pairs of compounds, and low concentrations of vapors of similar chemicals. Several discriminant algorithms were utilized, including k nearest neighbors (kNN, with k = 1), linear discriminant analysis (LDA, or Fisher's linear discriminant), quadratic discriminant analysis (QDA), regularized discriminant analysis (RDA, a hybrid of LDA and QDA), partial least squares, and soft independent modeling of class analogy (SIMCA). H_(2)O and D_(2)O were perfectly classified by most of the discriminants when a separate training and test set was used. As expected, discrimination performance decreased as the analyte concentration decreased, and performance decreased as the composition of the analyte mixtures became more similar. RDA was the overall best-performing discriminant, and LDA was the best-performing discriminant that did not require several cross-validations for optimization.

A statistical metric, based on the magnitude and standard deviations along linear projections of clustered array response data, was utilized to facilitate an evaluation of the performance of detector arrays in various vapor classification tasks. This approach allowed quantification of the ability of a 14-element array of carbon black−insulating polymer composite chemiresistors to distinguish between members of a set of 19 solvent vapors, some of which vary widely in chemical properties (e.g., methanol and benzene) and others of which are very similar (e.g., n-pentane and n-heptane). The data also facilitated evaluation of questions such as the optimal number of detectors required for a specific task, whether improved performance is obtained by increasing the number of detectors in a detector array, and how to assess statistically the diversity of a collection of detectors in order to understand more fully which properties are underrepresented in a particular set of array elements. In addition, the resolvi...