2D Neuromorphics

Background

The growing demand for data processing motivates energy-efficient computing approaches. Inspired by the human brain, neuromorphic computing brings memory and computation closer together through programmable devices such as memristors and memtransistors, potentially reducing data-transfer energy and latency. 

Ferroelectric materials enable non-volatile memory through electrically switchable polarization that persists after the applied field is removed and can modify semiconductor responses. Combining this functionality with the electronic and excitonic properties of 2D semiconductors provides a platform for exploring energy-efficient electronic and optoelectronic neuromorphic devices.

Our Research

Our research investigates ferroelectric materials and van der Waals interfaces for non-volatile electronic and optoelectronic devices. 

We study how polarization and interfacial interactions influence charge transport and light emission in 2D semiconductors, and use these relationships to develop programmable device functions. By combining materials synthesis, device fabrication and electrical and optical characterization, we connect material properties with memory behavior and explore device concepts for energy-efficient neuromorphic applications.

Outlook

The development of AI hardware is bringing memory technologies, device engineering, and computing architectures closer together. Progress toward practical systems will draw on expertise in ferroelectric materials, interface control, device reliability, and circuit-level evaluation, connecting emerging-device research with memory and AI semiconductor development.

Publications