Ammonia detection based on Pd/Rh-GaN and recognition of disease markers of nitrogen compounds assistant by deep learning

Bimetallic has received a lot of attention due to its excellent catalytic performance. Herein, bimetallic nanofilms were successfully modified on GaN epitaxial wafer by a co-reduction method in polyol solution, and different concentrations of Pd/Rh-GaN were obtained. Meanwhile, the morphology and structure of Pd/Rh-GaN were subjected to a series of measurements, and the gas sensors modified with bimetallic modification were also studied for gas sensitivity. The results of the gas-sensitive properties show that the Pd/Rh-GaN-0.25 sensor has outstanding gas-sensing performance with the highest sensitivity, the lowest limit of detection, and the fastest response/recovery time. In addition, we extended other bimetallic nanoparticles to Pd/Ru and Pt/Ru, and deep learning was combined with the sensor array to recognize three markers of nitrogen compounds (NH3, NO2, TMA) in Exhaled breath, achieving high precision recognition with up to 96 % accuracy. This work offers a new approach for the design and preparation of gas sensors for early warning of chronic kidney disease.