Decoupling thermal effects in GaN photodetectors for accurate measurement of ultraviolet intensity using deep neural network

Abstract

Precise and reliable monitoring systems under various harsh conditions are important to realize an overall future automation system for diverse industries. In this study, a facile fabrication method for a gallium nitride (GaN)-based ultraviolet (UV) photodetector was developed as a system-on-chip decoupling system, and a novel decoupling system was proposed to separate the thermal effects from the GaN-based sensor using a deep neural network (DNN)-based regression model. To predict the true values of UV intensity, the currents of both UV- and thermal-responsive GaN film and thermal-responsive copper film were used as input features without additional measurement devices. In addition, the performance of the DNN regression model was evaluated using the mean absolute percentage error. Consequently, the proposed model showed a prediction accuracy of ∼97.86% with two input features. Furthermore, the thermal decoupling method using the proposed sensor was successfully demonstrated, even when the thermal conditions changed significantly. These results support a novel decoupling method for reliable real-time monitoring systems for various harsh environmental industries, such as aerospace, ocean, subterranean, power plants, and combustion engines.