科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ Materials horizons2026-08-26

Synaptic memristors fabricated based on 2D layered CuInP2S6 through chemical intercalation for neuromorphic computing.

Xiao Chen, Xiushan Tu, Yan Xie, Mengdie Li, Hao Kan, Didi Zhang, Ludan Xu, Yujie Liu, Wenjing Jie

原始摘要(英文原文)· Original abstract
Chemical intercalation provides a versatile atomic-scale route for the rational design of molecular hybrid architectures within the van der Waals (vdW) layered hosts, enabling precise modulation of their structure and electrical, optical, and chemical functionalities. Therefore, it provides us a feasible way to achieve modified two-dimensional (2D) vdW materials for stable analog resistive switching (RS) behaviors which are highly desirable for synaptic devices and neuromorphic computing. In this work, we synthesized benzyltriethylammonium-intercalated CuInP2S6 (TEBA-CIPS) hybrid crystals by using benzyltriethylammonium chloride (TEBA-Cl) as the intercalation agent to provide anions of TEBA+ into the vdW gaps of CIPS, where the charge balance can be maintained by partial substitution of Cu+. The intercalated TEBA-CIPS samples can be exfoliated for fabrication of Au/TEBA-CIPS/Au vertical memristors and the fabricated devices exhibit stable analog RS behavior and successfully emulate diverse synaptic plasticity. In neural network simulations, it achieved 90.3% recognition accuracy for handwritten digits, and the recognition accuracy can be enhanced to 93.8% by using step pulses to stimulate the synaptic memristors based on TEBA-CIPS. Furthermore, the device effectively simulates the classical honeybees' proboscis extension reflex (PER). This work demonstrates a feasible strategy to modulate the crystal structure of the CIPS through a chemical intercalation method to achieve stable and reliable RS behaviors which are suitable for artificial synapses and neuromorphic computing.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Synaptic memristors fabricated based on 2D layered CuInP2S6 through chemical intercalation for neuromorphic computing. — 科研速览 Science Skim