Temperature and humidity compensation improves PANI/SnS 2 ammonia concentration prediction
Zhemin Jia, Chen Wang, Yuru Wang, David Hui, Zhenyu WangAbstract
Quantitative ammonia (NH 3 ) sensing with chemiresistive devices is strongly affected by temperature and relative humidity (RH), compromising calibration reliability under variable conditions. Here, we establish an environmental compensation framework using a PANI/SnS 2 device for NH 3 detection, a PEDOT:PSS/SnO 2 device for temperature sensing, and a MoS 2 device for RH sensing. The temperature-sensing device exhibits a temperature coefficient of resistance of −1.38 % °C −1 over 20–50 °C, while the RH-sensing device shows a sensitivity of 1.53 %/% RH over 25–90 % RH. Temperature, RH, and the PANI/SnS 2 response are subsequently incorporated into five regression models for NH 3 concentration prediction. Gradient boosting regression performs best on an independent test set, achieving R 2 = 0.95, a mean absolute error of 2.18 ppm, and a root-mean-square error of 3.57 ppm. Compared with the uncompensated model, environmental compensation reduces the root-mean-square error by 47.3 %. These results demonstrate that incorporating environmental information substantially improves quantitative NH 3 estimation under variable environmental conditions and provides a basis for future development of integrated NH 3 monitoring systems.