Ferroelectric/Antiferroelectric HfZrO<sub><i>x</i></sub> Artificial Synapses/Neurons for Convolutional Neural Network-Spiking Neural Network Neuromorphic Computing.
Authors
Affiliations (6)
Affiliations (6)
- School of Integrated Circuits, Shandong University, Jinan 250100, China.
- Suzhou Research Institute of Shandong University, Suzhou 215123, China.
- College of Integrated Circuits and Micro-Nano Electronics, School of Microelectronics, State Key Laboratory of Integrated Chip and System, Fudan University, Shanghai 200433, China.
- National International Innovation Center, Shanghai 201203, China.
- Key Laboratory of Computational Neuroscience and Brain-Inspired Intelligence (Fudan University), Ministry of Education, Shanghai 200433, China.
- State Key Laboratory of Crystal Materials, Shandong University, Jinan 250100, China.
Abstract
Brain-inspired neuromorphic computing offers significant potential for efficient and adaptive computational platforms. Emerging ferroelectric and antiferroelectric HfZrO<sub><i>x</i></sub> devices provide key roles in convolutional neural network (CNN) and spiking neural network (SNN) computing with unique polarization switching characteristics. Here, we present ferroelectric/antiferroelectric HfZrO<sub><i>x</i></sub> devices to realize functions of artificial synapse/neurons by element doping engineering. The HfZrO<sub><i>x</i></sub>-based ferroelectric and antiferroelectric devices exhibit excellent endurance characteristics of 1 × 10<sup>9</sup> cycles. Based on the non-volatile polarization switching and spontaneous depolarization nature of ferroelectric and antiferroelectric devices, integrate-and-fire behaviors were constructed for neuromorphic computing. For the first time, a complementary ferroelectric/antiferroelectric HfZrO<sub><i>x</i></sub> artificial synapse/neuron-based hybrid CNN-SNN framework was constructed for energy-efficient cardiac magnetic resonance imaging (MRI) classification. The hybrid neural network breaks the limitation of pure SNN in 3D image recognition and improves the accuracy from 82.3 to 92.7% compared to pure CNN, highlighting the potential of composition-engineered ferroelectric materials to implement high-efficiency neuromorphic computing.