Quantitative detection of multi-component chemical gas via MXene-based sensor array driven by triboelectric nanogenerators with CNN-GRU model

Precise identification of multi-component chemical gas poses a significant challenge. In this work, a MXene-based gas sensor array driven by triboelectric nanogenerators (TENGs) was constructed and combined with the neural network model to achieve accurate detection of multi-component chemical gas mixture. The wind-driven TENG array was prepared by Ti3C2Tx MXene and acetate fiber, which was used to powered the gas sensor array. The peak-to-peak value of open-circuit voltage and output power of a single TENG can reach 269 V and 1.2 mW. The gas sensor array was prepared by microelectronic printing and nano-sensing technology. The prepared sensor array exhibited better gas-sensing properties due to the synergistic effect between MXene and metal oxides. The gas-sensing response of MXene/metal oxide nanocomposites is 6.1-9.3 times better than that of the pure MXene. The MXene-based gas sensor array was constructed by integrating the TENG array with the gas sensor array. By combining the signal processing technology of MXene-based gas sensor array and convolutional neural network-gated recurrent unit (CNN-GRU) neural network mode, the composition identification and concentration prediction in NH3-SO2-NO2 three-component chemical gas mixtures were successfully realized with a mean relative error of less than 0.7%, which provides a universal solution for more complex multi-component chemical gas mixture detection.