Switching Behavior in NbO2-Based Memristive Devices

To meet the growing needs of computing, we depend on the fabrication of increasingly miniaturized transistors and the denser integration onto semiconductor chips. This trend would not continue as we approach fundamental physical limits due to the extremely small feature size. Studies suggest computational energy needs will exceed global energy production by 2040.



Modern computing devices are based on the Von Neumann architecture, and a significant amount of energy is spent moving data back and forth between memory and the processor. Memristor-based brain-inspired neuromorphic architecture can help with this fundamental challenge. The new architecture can do in-memory computations eliminating the energy spent on data transfer. Memristors can dynamically change their resistance in response to electrical stimuli making them perfect for mimicking the switching behavior of neurons.



We fabricated niobium dioxide (NbO2) based memristive devices and examined their characteristics. These devices showed excellent switching behavior making them highly favorable candidates for future research. We are currently exploring techniques to optimize and further improve the switching behavior of the memristors. Our work contributes to advancing the understanding of memristive devices for applications such as artificial intelligence, non-volatile memory, and neuromorphic computing.