SPICE Workshop on Nanomagnetism in 3D, April 30th - May 2nd 2024
Andrii Chumak
The field of magnonics, in which magnons, the quanta of spin waves, carry and process information, has shown promise for energy-efficient computing [1,2]. Recently, many proof-of-concept magnonic devices have been developed, paving the way for the development of complex magnonic circuits [3]. The extension of magnonic circuits to the third dimension opens up new opportunities in terms of access to novel physical phenomena and novel concepts for data transport and processing [4]. In the first part of my talk I will give a short overview of the main achievements in the field of 3D magnonics in the last years.
The second part of my talk will be devoted to the complementary direction of inverse-design magnonics. In inverse design, a functionality is first defined and a feedback loop algorithm provides a configuration in a complex reconfigurable magnonic medium to achieve this functionality [4, 5]. The drawback of the standard inverse-design approach is that the inverse problem has to be solved numerically first, which can be time and energy consuming depending on the complexity of the functionality, and only afterwards the obtained design can be reproduced experimentally [6]. In addition, a perfect calibration between numerics and experiment is required. In my talk, we report on a conceptually different approach where the inverse problem is solved experimentally.
The magnonic inverse design processor is based on an 18 µm thick rectangular YIG film with a 7x7 omega-shaped DC loop array placed on the YIG film. The loops are connected to 49 independent current source channels with currents limited to the range of -300 to +300 mA. Three input and three output antennas are used to access a wide range of functionalities. A genetic optimization algorithm is used to implement the notch filter functionality, and the attenuation within the bandwidth of about 33 dB is clearly observed. A 2-port frequency demultiplexer has been successfully realized and tested using the same processor.
[2] A. V. Chumak, P. Kabos, M. Wu, et al., “Advances in Magnetics Roadmap on Spin-Wave Computing,” IEEE Trans. Magn. 58, 0800172 (2022).
[3] Q. Wang, G. Csaba, R. Verba, A. V. Chumak, P. Pirro, “Perspective on Nanoscaled Magnonic Networks”, arXiv:2311.06129 (2023).
[4] “Three-Dimensional Magnonics”, edited by Gianluca Gubbiotti, Jenny Stanford Publishing, 2019
[5] Q. Wang, A. V. Chumak, and P. Pirro, “Inverse-design magnonic devices,” Nat. Commun. 12, 2636 (2021).
[6] A. Papp, W. Porod, and G. Csaba, “Nanoscale neural network using non-linear spin-wave interference,” Nat. Commun. 12, 6422 (2021).
[7] M. Kiechle, L. Maucha, V. Ahrens, et al., “Experimental Demonstration of a Spin-Wave Lens Designed with Machine Learning,” IEEE Magn. Lett., 13, 6105305 (2022).