k-Selective Electrical-to-Magnon Transduction with Realistic Field-distributed Nanoantennas

TL;DR

The paper addresses the challenge of predicting how realistic on-chip nanoantennas couple electrical drive to propagating spin waves in YIG. It introduces an end-to-end framework that couples frequency-domain FE electromagnetic simulations with FD micromagnetics, importing the complex near-field as a drive to compute the spin response m(k, ω) and the antenna k-weighting W(k, ω). The study demonstrates quantitative agreement with AEPSWS measurements on a 48 nm YIG film, revealing how tapering, return paths, and geometry shape the accessible k-space and dispersion, and it derives practical design rules for CPW vs. stripline transducers. This approach provides actionable guidance for on-chip magnonic transducer design with implications for low-power operation and future quantum magnonics implementations.

Abstract

The excitation and detection of propagating spin waves with lithographed nanoantennas underpin both classical magnonic circuits and emerging quantum technologies. Here, we establish a framework for all-electrical propagating spin-wave spectroscopy (AEPSWS) that links realistic electromagnetic drive fields to micromagnetic dynamics. Using finite-element (FE) simulations, we compute the full vector near-field of electrical impedance-matched, tapered coplanar and stripline antennas and import this distribution into finite-difference (FD) micromagnetic solvers. This approach captures the antenna-limited wave-vector spectrum and the component-selective driving fields (perpendicular to the static magnetisation) that simplified uniform-field models cannot. From this coupling, we derive how realistic current return paths and tapering shapes, k-weighting functions, for Damon-Eshbach surface spin waves in yttrium-iron-garnet (YIG) films are, for millimetre-scale matched CPWs and linear tapers down to nanometre-scale antennas. Validation against experimental AEPSWS on a $48\,nm$ YIG film shows quantitative agreement in dispersion ridges, group velocities, and spectral peak positions, establishing that the antenna acts as a tunable k-space filter. These results provide actionable design rules for on-chip magnonic transducers, with immediate relevance for low-power operation regimes and prospective applications in quantum magnonics.

Figures (5)

• Figure 1: The investigated system as an image section of the FE electromagnetic simulation box. Device stack and RF layout used for AEPSWS. The 48 nm YIG film (grey) on GGG is contacted by CPW transmission lines with linear tapers to a nanoantenna (stripline or CPW; Sec. \ref{['sec:results']})), as shown. Yellow boxes indicate the 50-$\Omega$ lumped ports used to excite/detect the RF signal, to model a quasi-TEM (Transverse Electromagnetic) transmission line. Coordinate axes are overlaid with $\hat{\mathbf y}\parallel \mathbf H$ (Damon–Eshbach), $\hat{\mathbf x}\parallel \mathbf k$, and $\hat{\mathbf z}$ out of plane. (Colour-coded rendering.)
• Figure 2: Sample fabrication and results. (a) Atop a 48 yttrium-iron-garnet (YIG) film grown on a 500 thick gadolinium-gallium-garnet (GGG) substrate, the antenna structures are deposited with a nanofabrication process. Using electron-beam lithography and thin-film metal evaporation, the antennas are fabricated. The antennas consit of Ti(5)/Au(85). For more information, please refer to the main text. SEM image of (b) the stripline antenna and the CPW tapers, and (c) the CPW antenna with the CPW tapers.
• Figure 3: FE and FD simulation results. Magnetic field norm on the surface of the simulated (a) stripline antenna and (b) coplanar waveguide antenna in A/m represented by the false colour bar. The components of the magnetic field from the excitation antenna are exported in the region of the micromagnetic finite-difference (FD) box. Averaged over the width of the FD box, the magnetic field norm is plotted for the (c) stripline antenna and (d) CPW antenna . In the FD box, micromagnetic simulations are conducted using the excitation field gained from the electromagnetic simulations. The normalised spin-wave excitation intensities are extracted from the micromagnetic simulations for (e) the stripline antenna and (f) the CPW antenna.
• Figure 4: All-electrical propagating spin-wave spectroscopy results compared with the simulation. The linear magnitude and imaginary part of the $\Delta S_{21}$ parameter compared to the simulated spin-wave excitation in Demon-Eshbach mode at an external magnetic field of 150 for (a) the CPW antenna and (d) the stripline antenna are shown. The dispersion relation was obtained by the numerical simulation using the excitation fields for (b) the CPW antenna and (e) the stripline antenna, illustrated by the colour map. Experimentally obtained group velocities for (c) the CPW antenna and (f) the stripline antenna, compared with the theoretically obtained group velocities by the derivative of the dispersion relation obtained by the Kalinikos-Slavin model Kalinikos_1986 (solid black line).
• Figure 5: Comparison between numerical simulation, TetraX calculations and calculations obtained by the Kalinikos-Slavin model. The Hue of the colourmap represents the dispersion relation obtained by the numerical simulations. Dispersion relations obtained by TetraX calculations and Kalinikos-Slavin model are represented by the dashed green and orange lines, respectively.