Ion Beam Engineering of YIG for Spin Wave based Computing Devices
Focused ion beam (FIB) irradiation is a versatile and powerful tool for micro- and nanoscale patterning. In magnonics, FIB can be used to locally modify the crystal lattice of yttrium iron garnet (YIG), introducing strain in a highly controlled manner.
Through magnetoelastic coupling, this strain alters the local magnetic properties and, consequently, the dispersion of propagating spin waves. By tailoring the spin-wave dispersion spatially, FIB-patterned regions can act as an effective refractive-index landscape for magnons. This enables the realization of optics-inspired magnonic devices for spin-wave routing and novel approaches to analog and neuromorphic computing.
Beyond conventional optical elements such as lenses and mirrors, graded-index (GRIN) structures are particularly attractive for magnonic information processing. Since their effective refractive-index profile can be precisely tailored by FIB patterning, increasingly complex and non-trivial analog computing tasks can be realized using inverse-designed scattering landscapes.
In this approach, a backpropagation-based machine-learning algorithm optimizes the spatial scattering landscape while accounting for the spin-wave dynamics described by the Landau–Lifshitz–Gilbert equation. By tailoring how spin waves propagate, interfere, and scatter within the patterned region, such structures can be designed to perform a wide range of wave-based signal-processing and analog-computing tasks.
The two-port perceptron shown here serves as an illustrative example of this concept: depending on the amplitudes applied at the two input ports, the inverse-designed scattering landscape routes the spin-wave signal toward one of two outputs, thereby realizing a physical classification task. More generally, this approach provides a flexible platform for implementing complex, application-specific spin-wave transformations directly in the magnetic medium.
Funding and Duration
Deutsche Forschungsgemeinschaft (DFG); 2027
Further Information
Principal Investigator: Markus Becherer
Doctoral Candidate: Felix Naunheimer
Collaboration Partners
Prof. Gyorgy Csaba, Pazmany Peter Catholic University
Dr. Adam Papp, Pazmany Peter Catholic University