Memory and Computing

Resistive Memory

Lateral and vertical two-terminal devices with various channel and electrode materials exhibit programmable memory behaviors. We study resistive memories based on ionic, filamentary and atomically switchable conductors. We focus on the mechanism, dynamic range, stability of their read, write and erase operation.

Transistor Memory

Three-terminal memory can eliminate the need for a selector in a resistive memory circuit. We incorporate low-dimensional floating gates or charge-trap dielectrics into the thin-film transistor architecture and optimize their memory characteristics by controlling the processing, properties and programming of these elements.

Neuromorphic Computing

Neuromorphic hardware mimics the functioning of our brain in enabling highly parallelized and low power computation. It is attracting increasing interest for application in artificial intelligence (AI). Our group studies electronic neurons, synapses and their physical connection to build a configurable neural network.