Our research in neuromorphic sensing, memory and computing is driven by the pursuit of advanced optoelectronic intelligent systems capable of emulating the remarkable efficiency observed in the human brain. To achieve this ambitious objective, we explore novel two-dimensional (2D) quantum materials with versatile and tunable properties, which can be controlled through both optical and electric fields. These materials serve as the foundation for developing cutting-edge systems that have the potential to revolutionize various applications, from artificial intelligence to sensory technologies. Through our research, we aim to not only expand our fundamental understanding of these materials but also harness their unique characteristics to significantly enhance the diversity and performance of next-generation neuromorphic intelligence.

Towards 2D Materials and Functional Devices (二维材料与功能器件)

Innovative semiconductor materials, with a particular focus on emerging 2D materials, have garnered significant attention. These materials, due to their reduced dimensionality (atomic layer thickness), exhibit remarkable tunable field-effect electrical properties in comparison to traditional silicon-based devices. In our research group, we study fundamental properties of 2D semiconductors as well as limits to their performance in nano-electronic and nano-optoelectronic devices. After the demonstration of diverse heterostructure-based devices, we have realized artificial synapses, logic operations, and neural networks.

Reference:

X. Pan, et al. “Parallel perception of visual motion using light-tunable memory matrix” Science Advances 9, eadi408 (2023)

X. Pan,  et al. “Surface Charge Transfer Doping Enabled Large Hysteresis in van der Waals Heterostructures for Artificial Synapse” ACS Materials Letters 3, 235-242 (2021)

Towards Neuromorphic Intelligent Market (类脑感存算智能应用)

The neuromorphic computing market is witnessing rapid expansion, with increasing opportunities across data centers, automotive systems, and various espects of modern life. This momentum shall be further accelerated by the recognition of advancements in artificial intelligence through the 2024 Nobel prize in physics and chemistry. In our research group, we focus on pioneering in-sensor computing by mapping machine learning algorithms and computational models directly onto device architectures. Our work aims to optimize sensory data processing by embedding intelligence within the sensors themselves, enhancing efficiency and performance. 

Reference:

X. Pan, et al. “Parallel perception of visual motion using light-tunable memory matrix” Science Advances 9, eadi408 (2023)

X. Pan, et al. “Nonvolatile van der Waals Heterostructure Phototransistor for Encrypted Optoelectronic Logic Circuit” ACS Nano 16, 4528-4535, (2022)