Space-Confined Growth of Ultrathin 2D β-Ga2O3 Nanoflakes for Artificial Neuromorphic Application

Abstract

In recent years, wide-bandgap semiconductor β-Ga₂O₃ material has been widely studied because of its excellent properties. Simultaneously, 2D metal oxides (2DMOs) have also become a focus of research owing to their superior stability and unique physical properties arising from quantum confinement effects. Therefore, the exploration of 2D β-Ga₂O₃ is expected to reveal its novel electrical properties in electronic applications. However, the synthesis of high-quality 2D β-Ga₂O₃ remains a formidable challenge. Herein, a confined space is constructed to synthesize high-quality 2D β-Ga₂O₃ nanoflakes by enhancing the control of the kinetics of chemical vapor deposition process. In the device results, it is shown that the grown nanoflakes have excellent switching properties and potential artificial synaptic response characteristics. Based on this premise, an artificial recognition system for handwritten numerals is developed, achieving a peak recognition accuracy of approximately 96%. This system holds significant potential for application within an emerging neuromorphic recognition framework tailored for advanced driver-assistance systems. In this work, a new feasible pathway is provided for the synthesis of 2D non-layered oxides and the potential of 2D oxides in the field of neuroanalog electronics and recognition is shown, thereby advancing the fields of 2D β-Ga₂O₃ electronics and 2DMOs electronics.