Fully Self-Powered Gas Sensor through Fe-Ion Doping Engineering in Ni₂P for Ultrasensitive and Visualized NO₂ Sensing

Self-powered gas sensors based on zinc-air batteries (ZABs) integrate battery and gas sensing functions. Various gaseous reduction reactions at the battery cathode will induce a measurable and reproducible change in the battery's output voltage, eliminating the need for any external power supply. However, they still face significant challenges in achieving ultrasensitive, selective, and drift-resistant sensing under complex conditions. Herein, we report a NO₂ gas sensor based on a ZAB, utilizing Fe-doped nickel phosphide (FNP) as the gas-sensitive material and incorporating deep learning algorithms to enhance the sensing performance. In the context of FNP gas-sensitive layers, the charge carrier mobility is significantly enhanced, owing to the electron delocalization and redistribution induced by the Fe dopants. Furthermore, the shift in the d-band center toward the Fermi level induced by Fe dopants facilitates stronger O 2p/Fe d orbital hybridization. As a result, the adsorbate-substrate interaction is enhanced and the Gibbs free energy of the NO₂ reduction reaction is reduced. The charge density difference also indicates the facilitated electron transfer from FNP to NO₂ and the electron accumulation at the O sites. Consequently, the sensors exhibit a high response (0.22 V @ 20 ppm), a low limit of detection (LOD: 61.8 ppb), and fast sensing speed (14 s). After that, the introduction of the InceptionTime model and wavelet transformation algorithm enables the sensor to achieve remarkable gas recognition and concentration quantification, along with a significantly reduced LOD of 36.9 ppb. Finally, a smart sensing device is constructed with the sensors, microcontrollers, and electrochromic devices for remote and visualized gas detections.