Center for Functional Nanomaterials Colloquium

Advancing electronics beyond Moore's Law requires a revolutionary approach that encompasses new materials, novel devices, disruptive technologies, and alternative computing architectures. A memristor (short for memory resistor) is an emerging device that encodes digital or analog information through distinct resistance states. When built into circuits, memristive crossbar arrays enable massively parallel multiply–accumulate operations by harnessing fundamental physical laws, boosting both power efficiency and computational throughput. Leveraging ionic motion, memristive devices also serve as bio-realistic emulators for artificial neurons and synapses, making them ideal for neuromorphic applications. In this talk, I will start by introducing a computing memristor designed for analog in-memory computing. I will then showcase an AI accelerator that brings together memristive crossbar arrays and CMOS technology on a single chip, highlighting our recent lab-to-fab transfer initiative supported by the CHIPS Act. Lastly, I will delve into diffusive memristors and their applications in neuromorphic engineering.