Hybrid structure with a ferromagnetic film and an array of magnetic molecules for deep-nanoscale reprogrammable magnonics
Oleksandr Pastukh, Piotr Graczyk, Mateusz Zelent, Lukasz Laskowski, Maciej Krawczyk
TL;DR
This work introduces a deep-nanoscale, reprogrammable magnonic platform formed by decorating a YIG film with a regularly spaced Mn$_{12}$ single-molecule magnets array. Through micromagnetic and FEM analyses, the authors demonstrate resonant dipolar coupling between propagating spin waves and molecular moments, producing an anti-crossing gap that suppresses transmission and is tunable via external field, molecular density, arrangement, and AFM clustering. The results show the system can operate in the GHz regime with controllable gap position and width, enabling dense, reprogrammable magnonic networks with potential neuromorphic and quantum-magnonic interfaces. The work outlines practical routes to fabrication and orientation control and highlights the key role of inter-molecular spacing in determining coupling strength, offering a viable path toward hardware neural networks at the nanoscale.
Abstract
Miniaturization is an essential element in the development of information processing technologies and is also one of the main determinants of the usability of the tested artificial neural networks. It is also a key element and one of the main challenges in the development of magnonic neuromorphic systems. In this work, we propose a new platform for the development of these new spin-wave-based technologies. Using micromagnetic simulations, we demonstrate that magnetic molecules regularly arranged on the surface of a thin ferromagnetic layer enable resonant coupling of propagating spin waves with the dynamics of the molecules' magnetic moments, opening a gap in the transmission spectrum up to 150 MHz. The gap, its width, and frequency can be controlled by an external magnetic field or the arrangement of molecules on the ferromagnetic surface. Furthermore, the antiferromagnetic arrangement of the magnetic moments of molecules or clusters of molecules allows for control of the gap's position and width. Thus, the proposed hybrid structure offers reprogrammability and miniaturization down to the deep nanoscale, operating frequencies in the range of several GHz, key properties for the implementation of artificial neural networks.
