Concentration recognition of gas sensor with organic field-effect transistor assisted by artificial intelligence

Abstract

Many gas analytes have been detected with organic field-effect transistor (OFET) sensors. However, most of the analytes are detected with known concentrations and few OFET sensors have been used to quantify analytes. Here, we integrated the multiple independent parameters of OFETs with artificial neural network (ANN) to quantify toxic H2S. The precise concentration recognition was confirmed as the difference between real and predicted values was less than 5%. The H2S sensors also showed high responsivity over 6500% at 150 ppm, and quick response in ten seconds. Moreover, the multiple parameters provided the possibility to differentiate eight different analytes. Thus, the artificial intelligence combined with OFET sensors realized precise quantification and fingerprint recognition of gas analytes for real applications.