Space-Confined-Grown 10 nm 2D β-Ga2O3 Nanoflakes for Electrode-Scalable Ultraviolet Photodetector and Neuromorphic Computing

In the post-Moore era, next-generation optoelectronics and neuromorphic computing systems require ultrawide bandgap semiconductors with tunable properties. β-Ga2O3 stands as a promising candidate, possessing a bandgap of 4.5–4.9 eV and a high critical electric field of ∼ 8 MV/cm. However, its practical application is hindered by difficulties in synthesizing high-quality 2D forms and optimizing metal–semiconductor contacts. In this study, highly crystalline ultrathin 2D β-Ga2O3 nanoflakes with improved interface quality and enhanced carrier transport dynamics were synthesized using a space-confined CVD approach. The nanoflakes exhibit excellent structural and crystalline quality, as evidenced by the characteristic peaks in Raman spectroscopy and XRD, along with the clear lattice structure observed via TEM. Metal–semiconductor–metal devices with Ti–Ti, Pd–Pd, and Ti–Pd electrode configurations were fabricated and evaluated through current–voltage and time-resolved current–time measurements. The results reveal that the Schottky barrier height at the metal–Ga2O3 interface plays a decisive role in device performance. Notably, the Ti–Ti device exhibited a distinctive two-stage photocurrent decay after light illumination ceased, attributed to defect-mediated trapping and release of photogenerated carriers. By incorporating this unique photoresponse behavior into a deep neural network, promising image recognition accuracy was achieved on the Modified National Institute of Standards and Technology data set. This work not only offers a reliable pathway for synthesizing high-quality 2D β-Ga2O3 but also demonstrates its potential for application in high-performance optoelectronics and neuromorphic computing.