Reconfigurable Carbon Nanotube FET Integrated with Phase-Change Memory for Logic-in-Memory & Neuromorphic Computing Architectures

The von Neumann bottleneck and the energy-profile issue of conventional CMOS architecture, work is being done on the exploration of other computing paradigms. This paper proposes a hybrid nanoelectronic platform that integrates reconfigurable Carbon Nanotube Field-Effect Transistors (CNT-FETs) with Phase-Change Memory (PCM) elements to perform logic-in-memory (LiM) and neuromorphic computing. Based on the excellent electrical properties of single-walled carbon nanotubes (SWCNTs) and the nonvolatile switching property of PCM, the proposed architecture is capable of low-energy, scalable, and CMOS-compatible computation. Simulation results also validate that the platform supports reconfigurable logic operations and memory storage in the same physical location, thereby significantly reducing data movement and increasing computational throughput. Thus, while providing a great substrate for advanced LiM and neuromorphic architectures, this integration also paves the way for low-power, high-density nanoelectronic systems that could fundamentally shape future's embedded and edge computing technologies.