Highly conductive GaN-Cu3(HITP)2/PANI gas sensor based on GaN double heterojunction coupled with deep learning for accurate trimethylamine mixed gas detection

Metal-Organic Frameworks (MOFs) have attracted extensive attention in the field of gas sensing. However, single MOFs materials suffer from shortcomings such as poor electrical conductivity and easy agglomeration, leading to limited performance. Herein, the highly conductive GaN-M₃(HITP)₂/PANI nanocomposites were synthesized through in-situ method on GaN-HP film. Due to the higher degree of protonation caused by hierarchically porous architecture, the GaN-Cu₃(HITP)₂/PANI sensor demonstrates better TMA sensing performance compared with GaN-Ni₃(HITP)₂/PANI sensor. The GaN-Cu₃(HITP)₂/PANI sensor achieves a lower detection limit of 500 ppb, higher response value for 100 ppm TMA (91.4 %), faster response/recovery rate (21 s/56 s) at room temperature. Moreover, the GaN-Cu₃(HITP)₂/PANI sensor exhibits good long-term stability and humidity resistance. Systematic analysis revealed that the enhanced sensing performance in the GaN-Cu₃(HITP)₂/PANI composite stems from the formation of coupled heterointerface system and its hierarchically porous architecture. The Douglas-Peucker (DP) algorithm was used for the first time to extract the response features of GaN-Cu₃(HITP)₂/PANI sensor, and combined with a deep learning algorithm, successfully achieving exact recognition of the mixed gas of TMA and MH₃, with an identification rate as high as 94.1 %.Thus, the GaN-Cu₃(HITP)₂/PANI sensor provides a feasible solution for developing portable TMA detection systems embedded into smart wearable devices.