Establishing the Magnetoelastic Origin of Spin-Wave Routing through Focused Ion Beam Patterning

Spin waves are promising information carriers for analog and wave-based computing, requiring compact and precisely engineered scattering landscapes. Focused ion beam (FIB) irradiation enables such control by locally modifying the spin-wave dispersion in yttrium iron garnet (YIG), yet the underlying crystallographic mechanisms remain unclear. Here, we present an experimentally validated framework that attributes FIB-induced spin-wave steering to magnetoelastic effects arising from irradiation-induced lattice dislocations. Following FIB irradiation and wet-chemical etching, local height profiles were obtained by atomic force microscopy (AFM) and used as fixed geometric constraints in fits of spin-wave dispersion relations measured by time-resolved magneto-optical Kerr effect (trMOKE) microscopy. The dispersion relation was extended by an explicit magnetoelastic field term, treated as a fit parameter. Its evolution reveals three successive deformation regimes, elastic, plastic, and partial amorphization, explaining the observed non-monotonic dependence of the spin-wave wavelength on ion dose. A three-phase deformation scenario based on SRIM simulations reproduces the extracted magnetoelastic field trends, validating the fitting approach. Micromagnetic simulations incorporating strain tensors derived from the experimental magnetoelastic field reproduce the characteristic non-monotonic wavelength behavior. These results establish a physical basis for FIB-engineered graded-index (GRIN) spin-wave landscapes and magnetoelastically programmable magnonic devices.

💡 Research Summary

The paper addresses a critical gap in the understanding of how focused ion beam (FIB) irradiation modifies spin‑wave propagation in yttrium iron garnet (YIG) thin films, a key requirement for building compact, wave‑based computing elements such as graded‑index (GRIN) spin‑wave lenses and routers. The authors combine experimental measurements, analytical modeling, Monte‑Carlo simulations, and micromagnetic calculations to demonstrate that the observed spin‑wave steering originates from magnetoelastic fields generated by irradiation‑induced lattice dislocations.

First, 100 nm YIG films are deposited on gadolinium gallium garnet substrates and patterned with a Ti/Au microstrip line for spin‑wave excitation. Adjacent 50 µm × 50 µm squares are irradiated with Ga⁺ ions at 8 keV, 16 keV, and 30 keV, covering doses from 2 × 10¹² to 60 × 10¹² ions cm⁻² in 2 × 10¹² steps. After irradiation, a brief wet‑chemical etch removes the near‑surface amorphous layer that forms at high doses; atomic force microscopy (AFM) quantifies the resulting local thickness reduction, which is then fixed as a geometric constraint in subsequent analysis.

Time‑resolved magneto‑optical Kerr effect (trMOKE) microscopy records spin‑wave propagation at frequencies between 2.285 GHz and 2.33 GHz under a 250 mT out‑of‑plane field. Fourier analysis of the trMOKE images yields the spin‑wave wavelength λ as a function of ion dose. The λ‑versus‑dose curve exhibits two clear turning points (at ≈12 × 10¹² and 34 × 10¹² ions cm⁻²), dividing the response into three monotonic regimes: decreasing λ (elastic regime), increasing λ (plastic regime), and decreasing λ again (partial amorphization regime).

To explain this non‑monotonic behavior, the authors extend the Kalinikos‑Slavin spin‑wave dispersion relation by adding an explicit magnetoelastic field term H_me = B₁ ε · M, where ε is the strain tensor generated by lattice defects and B₁ is the magnetoelastic coupling constant. By fitting the measured dispersion with H_me as a free parameter, they extract a dose‑dependent magnetoelastic field that mirrors the three‑phase deformation scenario.

The physical origin of the strain is modeled using SRIM Monte‑Carlo simulations of Ga⁺ ion implantation. SRIM provides depth‑resolved damage densities (vacancies, interstitials) for each acceleration voltage. Multiplying these profiles by the applied dose yields a depth‑dependent damage function D_FIB(z). The authors interpret D_FIB(z) in terms of three stages: (I) elastic accumulation of strain due to isolated defect clusters, (II) plastic relaxation as clusters collapse into edge dislocation loops (⊥) that migrate and partially relieve strain, and (III) near‑surface amorphization that removes material (observed as thickness loss) while deeper layers continue to accumulate strain. The SRIM‑derived strain evolution reproduces the experimentally fitted H_me trends, confirming that the magnetoelastic field is a reliable proxy for the underlying defect‑induced strain.

Micromagnetic simulations using MuMax3 incorporate the strain tensors derived from the fitted H_me values (with B₁ = 3.48 × 10⁵ J m⁻³). These simulations reproduce the measured wavelength shifts across all dose regimes, including the re‑entrant decrease in λ at high doses, thereby validating the magnetoelastic origin of the dispersion changes.

Overall, the study provides a self‑consistent experimental‑computational pipeline: (1) AFM‑measured thickness loss fixes geometric parameters, (2) trMOKE yields spin‑wave dispersion, (3) extended Kalinikos‑Slavin fits extract H_me, (4) SRIM predicts defect‑induced strain, and (5) micromagnetics confirm the impact on spin‑wave propagation. This framework establishes that FIB‑engineered magnetoelastic landscapes can be used to design programmable magnonic devices with graded refractive‑index profiles, opening pathways to low‑loss, high‑frequency, sub‑micron spin‑wave circuitry for analog and wave‑based computing.