Responses of nanostructured sensing array to mixed VOCs: feature extraction and pattern recognition analysis

This paper investigates the response characteristics of a sensing array constructed from nanostructured thin films on interdigitated microelectrodes (IMEs). It also demonstrates the array's capability to identify and quantify gas mixtures. The films were prepared from dithiol (PRDT, PDT, ODT, NDT, and DDT) and mercaptoundecanoic acid (MUA) linked assemblies of 2-nm sized gold nanoparticles. These films exhibit sensitive responses to volatile organic compounds (VOCs). The characteristics of the stationary response signals are studied. The static responses are extracted as the features for identifying VOCs by pattern recognition. The Principal Component Analysis (PCA) shows the patterns of binary mixture presented in the static features are linear combinations of the individual analytes' patterns. Pattern classification based on the selected features is performed with Artificial Neural Network (ANN), which shows satisfied detection and quantification performance.