Regensburg 2022 – scientific programme

Parts | Days | Selection | Search | Updates | Downloads | Help

DS: Fachverband Dünne Schichten

DS 16: Thin Film Applications 2

DS 16.4: Talk

Wednesday, September 7, 2022, 11:45–12:00, H14

Redox-based Memristive Devices for Neuromorphic Systems — •Benjamin Spetzler, Seongae Park, Anna Linkenheil, Tzvetan Ivanov, and Martin Ziegler — Technical University Ilmenau, Ilmenau, Germany

Redox-based memristive devices have demonstrated promising properties for their application as synaptic elements in neuromorphic computing systems. The device characteristics are the product of a variety of complex mechanisms, and electronic and ionic processes need to be precisely tuned, which requires a deep understanding of the underlying physical mechanisms and control of the fabrication parameters. We present redox-based memristive elements and show how their properties can be tailored by systematic design variations for applications in neuromorphic computing architectures. In this context, the influence of different oxide layer systems and electrode materials on the device characteristics is analyzed to assess their properties for neuromorphic computing. The experimental findings are supported by a numerical device model, which connects the physical processes with technology parameters, and permits a deeper understanding of the origin of the current-voltage hysteresis. Furthermore, we discuss the system integration of memristive devices and present memristive device arrays.

This work was partially funded by the Carl-Zeiss Foundation via the Project MemWerk and the German Research Foundation (DFG) through the Collaborative Research Centre CRC 1461 "Neurotronics- Bio-Inspired Information Pathway".