60 nm Pixel-size pressure piezo-memory system as ultrahigh-resolution neuromorphic tactile sensor for in-chip computing. (September 2021)

• Record Type: Journal Article
• Authors: Jiang, Chengming; Tan, Dongchen; Sun, Nan; Huang, Jijie; Ji, Ruonan; Li, Qikun; Bi, Sheng; Guo, Qinglei; Wang, Xiaohu; Song, Jinhui
• Is Part Of: Nano energy. Volume 87(2021)
• Publication Date: 2021-09
• Subjects: Piezo-electronics -- Nanowires -- Memristors -- Tactile Sensor -- In-chip computing; Nanoscience -- Periodicals; Nanotechnology -- Periodicals; Nanostructured materials -- Periodicals; Power resources -- Technological innovations -- Periodicals; Nanoscience; Nanostructured materials; Nanotechnology; Power resources -- Technological innovations; Periodicals; 621.042
• DOI: 10.1016/j.nanoen.2021.106190
• Languages: English
• ISSNs: 2211-2855

Abstract

Abstract: Neuromorphic tactile sensors provide considerable feasibility to simulate essential capabilities of the human tactile system even beyond the bio-perceptible range. Nevertheless, the complicated circuitry of artificial tactile sensors based on electrical-skin, memory, and processing components show serial barriers in terms of system resolution, integration, and power consumption. Here, a high-resolution pressure piezo-memory system (HPPMS) is reported with 60 nm pixel-size as a competent neuromorphic tactile sensor, that presents non-volatile force resistive conversion and force-tunable synaptic functions. The piezo-memory pixel arrays realize nanoscale force-image sensing and memory actions as same as neuromorphic tactile preprocessing by enhanced processing efficiencies and recognition rates. The proof-of-concept system affords the possibility to achieve ultra-small force distributions, simplify tactile sensor circuits, and develop advanced applications in in-sensor computing and artificial intelligence. Graphical Abstract: ga1 Highlights: The HPPMS works with the piezo-NWs and resistive-memristors as a neuromorphic tactile sensor. The device exhibits a nanoscale pixel-size of 60 nm with resolution-size of 130 nm. The pressure-stability of the HPPMS can allow a long storage time of >2 months. The HPPMS works for in-chip computing with the improved processing efficiency (36.4%) and image recognition rate (0.99).