(231d) Selectivity Enhancement of Nanowire Gas Sensors Using Impedance Spectroscopy and Artificial Neural Network

There is a significant market growth opportunity for sensors with better performances in the gas detection segment driven largely by more stringent safety regulations in monitoring and controlling industrial and residential environment. Because nanowires have ultra-high surface to volume ratio, organic nanowire sensors show high sensitivity, and fast response, coupled with low power consumption and low cost. However, their poor selectivity is a major challenge that prevents organic nanowire sensors from being efficiently used. In this talk, a new nanowire sensor/sensor array fabrication method, and enhancement of selectivity by using artificial neural network (ANN) models will be presented. Our previous and ongoing research show that copper tetracyanoquinodimethane (CuTCNQ) and tetrathiafulvalene charge transfer salts (TTF) can grow into nanowires from solution on photolithographic gold pattern in ambient temperature and pressure by simply applying overpotential. After drying in air, the CuTCNQ sensor can detect sub-ppm level ammonia gas by measuring the impedance spectroscopy. Equivalent electric circuit models will be used to analyze the conduction and detection mechanisms of the nanowires as well as the effect of their geometry and contact junctions. By employing the same fabrication method, different sensing materials can be deposited onto one devices working as a sensing array. The impedance data of different material in the sensing array will be coupled with ANN models to enhance the selectivity of the sensors. The ANN-enhanced sensors can selectively detect different gases, such as ammonia and ethanol in the ppm level.