Recently, the research team led by Professor Huo Nengjie from the School of Electronic Science and Engineering (School of Microelectronics) at the Faculty of Engineering, South China Normal University, has made a breakthrough in edge-computing devices for 3D integrated electronics and low-power neuromorphic vision. The related work has been published in ACS Nano (IF: 16.0), a distinguished journal in CAS Zone 1, under the title "Ferroelectric Gate-All-Around Transistors for 3D-Integrated Electronics and Neuromorphic Vision." Ms. Chen Wenjie, a master's student (Class of 2023) from the School, serves as the first author, with Professor Huo as the corresponding author, and South China Normal University as the primary affiliation. This study, through an innovative device architecture based on two-dimensional ferroelectric gate-all-around transistors, achieves the synergistic integration of ultra-low-power logic switching, 3D integrated electronics, and artificial neuronal functionalities. It establishes a novel paradigm for multifunctional devices tailored to edge intelligent computing, thereby charting a new path for the development of next-generation high-energy-efficiency edge intelligence hardware.

Architectural Innovation: Ferroelectric Gate-All-Around Transistors and 3D Logic Integration

    The rapid advancement of artificial intelligence and the Internet of Things imposes ever-increasing demands for high performance, low power consumption, and high integration density in electronic devices. However, conventional silicon-based transistors are confronting fundamental physical scaling limits and power dissipation bottlenecks. To address these challenges, the research team has developed a ferroelectric gate-all-around field-effect transistor (Fe-GAA-FET), which employs two-dimensional ferroelectric CuInP₂S₆ (CIPS) as the gate-all-around dielectric to achieve robust electrostatic control over the MoS₂ channel. Leveraging the ferroelectric negative-capacitance effect, the device successfully breaks the room-temperature Boltzmann limit, demonstrating an ultra-steep subthreshold swing of 25.3 mV dec⁻¹, an on/off current ratio of 10⁸, and a field-effect mobility of 310 cm² V⁻¹ s⁻¹. Furthermore, by exploiting the natural van der Waals stacking properties of two-dimensional materials, the team constructed vertically integrated complementary CMOS inverters and NOR logic gates. Compared with conventional planar architectures, this 3D integration scheme reduces the device footprint by approximately 50%, exhibiting excellent logic switching performance and substantial potential for high-density integration.

Application Prospects: A New Paradigm for Neuromorphic Computing and Edge Intelligence

    Building upon the underlying device, the research team further explored its application in neuromorphic computing. Taking advantage of the gradual ferroelectric polarization switching in CIPS and the favorable capacitive matching of the device, the Fe-GAA-FET can emulate leaky-integrate-and-fire (LIF) neuronal dynamics without requiring additional capacitor or reset circuitry, achieving an ultra-low energy consumption of only 1.09 pJ per spike. In practical validation, a spiking neural network (SNN) model based on these devices was constructed, attaining a recognition accuracy of 92.71% on the DVS128 Gesture dataset.

    This work provides a novel pathway to overcome the limitations of the conventional von Neumann architecture. The resulting technologies are expected to play a significant role in frontier applications such as smart terminals, wearable devices, robotic vision, and industrial surveillance systems.

Acknowledgments and Intellectual Property

    This research has been continuously supported by the National Natural Science Foundation of China, the Guangdong Provincial Natural Science Foundation, and other funding sources. One invention patent related to this technology has been filed.

Paper Link: https://url.scnu.edu.cn/record/view/index.html?key=ca42c96b0c8467e67d2886294caa5953