Herein, we designed a memoristor-type gas sensor (gasistor) with Ag/ordered TiO2 nanorods/FTO sandwich structure for recognizing ethanol concentration in ethanol-methanol mixtures. The gas adsorption characteristics of different TiO2 substrates were first investigated by first-principle calculations. The results showed that the TiO2 surface in the low resistance state exhibited better immunity to other interfering gas molecules. However, similar adsorption energies and charge transfers calculated for adsorption of methanol and ethanol mean that the interference of methanol to the response of ethanol cannot be completely ignored. In next experiments, the Ag/TiO2/FTO gasistor with well-aligned TiO2 nanorods as the resistive layer was constructed successfully and demonstrated the stable resistive switching function and excellent retention characteristic. Further device tests confirm the ability of the gasistor in the low resistance state to discriminate between ethanol and methanol, as well as better anti-interference for other organic vapors. Also, an ultra-short recovery time (∼1.8 s) was observed by applying a direct current scanning voltage to the gasistor. Finally, to improve the recognition rate of ethanol in ethanol-methanol mixed atmosphere, a neural network algorithm based on back-propagation was incorporated and good prediction accuracy was obtained at ethanol concentrations ≥ 5 ppm.
Read full abstract