Dresden 2026 – scientific programme

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

DS: Fachverband Dünne Schichten

DS 20: Poster

DS 20.29: Poster

Thursday, March 12, 2026, 18:30–20:30, P2

Using layered metamaterials in spin - ventil structures as a basis for artificial neural network — •Vladimir Boian¹, Cătălin Cimbir¹, and Vladimir M Fomin² — ¹Technical University of Moldova, Institute of Electronic Engineering and Nanotechnologies, Chisinau, Moldovau — ²IET, Leibniz IFW Dresden, Dresden, Germany

Elaboration of a superconducting artificial neural network (ANN) the most promising solution in the design and development of non-von Neumann architectures. There are two main components of ANN: the nonlinear switch neuron constituting a spin valve, and the linear connection element the synapse. This study presents the results of computer modeling of superconducting spin valves on the base of Josephson Junction with weak link prepared from artificial magnetic metamaterials, and of superconducting synapses, based on hybrid layered structures superconductor/ferromagnet. The proximity effect in a superlattice formed by superconducting Nb nanolayers and ferromagnetic Co layers with different thicknesses and coercive fields is analyzed both theoretically and experimentally. In this sense, they can be applied as tunable kinetic inductors for the design of ANN synapses. Metamaterials based on Nb and Co nanolayers are a very promising class of artificial magnets for superconducting spintronics and quantum computing. The use of artificial neural networks with a radically reduced consumption of electricity are increasingly appreciated worldwide.

Keywords: layered metamaterials; spin - ventil structures; superconducting artificial neural network (ANN)