Mirzanezhad, H. (EE) – Van der Waals materials-based memory devices for neuromorphic computing

The separation of memory and processing in conventional computer architectures leads to significant energy consumption associated with data transfer, known as the von Neumann bottleneck. Two-dimensional (2D) materials provide a promising platform for developing nanoscale memory devices for in-memory and neuromorphic computing because of their atomic-scale thickness, lack of dangling bonds, and ability to form van der Waals structures. In this work, bilayer MoS₂ is investigated as an active material for memory devices based on two different switching mechanisms. The first is a resistive memory based on a vertical junction that exhibits a reproducible hysteresis loop and can operate as a memristor. Possible mechanisms contributing to the observed hysteresis include ion migration, charge trapping, polarization switching, or a combination of these mechanisms. The second is a ferroelectric memory based on the established ferroelectricity of parallel-stacked bilayer MoS₂, where broken inversion symmetry gives rise to switchable out-of-plane polarization. This work therefore follows two parallel directions: a memory device based on the resistive effect and a memory device based on the ferroelectric effect, with the longer-term goal of developing 2D material-based memory devices for neuromorphic computing.

Event Host: Hamid Mirzanezhad, Ph.D. Student, Electrical Engineering 

Advisor: Aiming Yan