材料科学
三元运算
晶体管
碳纳米管
逻辑门
逆变器
纳米技术
瓶颈
光电子学
互连
场效应晶体管
电路复杂度
纳米电子学
制作
碳纳米管场效应晶体管
神经形态工程学
电子线路
电子工程
功率(物理)
和大门
功率消耗
导电体
异质结
调制(音乐)
作者
Yu Teng,Jian Yao,Qinan Wang,Yanyan Zhao,Lin Geng,Pin Zhao,Kunjie Wang,Zongjie Shen,Chun Zhao,Lixing Kang,Mario Lanza
标识
DOI:10.1002/adma.202523297
摘要
As circuit integration continues to advance, power consumption has become a critical bottleneck limiting further development. Multi-valued logic (MVL) has garnered extensive attention due to its potential to reduce interconnect complexity and switching losses. Single-walled carbon nanotubes (SWCNTs), with their superior electrical properties, ultra-small dimensions, and controllable aligned array growth, offer unique advantages for the large-scale fabrication of high-density MVL circuits. However, progress in this field using SWCNTs remains relatively lagging compared to two-dimensional materials, primarily due to device stability issues arising from challenges in precise doping control. Here, we demonstrate a system consisting of acetylacetonate metal molecules encapsulated within SWCNTs (M(acac)x@s-SWCNT), in which carrier concentration can be dynamically modulated under an applied electric field. Transistors based on this platform validate that this electric-field-controlled modulation yields three well-defined logic states: 0, 1, and 2. These transistors demonstrate good uniformity and stable operation, showing a static power consumption of 8.2 pW and dynamic power consumption of 0.31 nJ (state 0 to 1) and 0.35 µJ (state 1 to 2). The ternary inverter based on this heterostructure exhibits rail-to-rail output capability, enabling the accurate execution of MVL operations. Ternary weight networks (TWNs) built with these transistors reduce computational complexity and storage, enabling efficient neuromorphic computing.
科研通智能强力驱动
Strongly Powered by AbleSci AI