S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

APPLIED SCIENCES AND ENGINEERING Copyright © 2024 The

Authors, some rights
Parametric magnon transduction to spin qubits reserved; exclusive
licensee American
Association for the
Mauricio Bejarano1,2, Francisco J. T. Goncalves1, Toni Hache1,3, Michael Hollenbach1,4, Advancement of
Christopher Heins1, Tobias Hula1,5, Lukas Körber1,4, Jakob Heinze1, Yonder Berencén1, Science. No claim to
Manfred Helm1,4, Jürgen Fassbender1,4, Georgy V. Astakhov1, Helmut Schultheiss1* original U.S.
Government Works.
The integration of heterogeneous modular units for building large-­scale quantum networks requires engineering Distributed under a
mechanisms that allow suitable transduction of quantum information. Magnon-­based transducers are especially Creative Commons
attractive due to their wide range of interactions and rich nonlinear dynamics, but most of the work to date has Attribution
focused on linear magnon transduction in the traditional system composed of yttrium iron garnet and diamond, License 4.0 (CC BY).
two materials with difficult integrability into wafer-­scale quantum circuits. In this work, we present a different ap-
proach by using wafer-­compatible materials to engineer a hybrid transducer that exploits magnon nonlinearities
in a magnetic microdisc to address quantum spin defects in silicon carbide. The resulting interaction scheme
points to the unique transduction behavior that can be obtained when complementing quantum systems with
nonlinear magnonics.

INTRODUCTION architecture, which could lead to enhanced transduction behavior

The seminal proposal of quantum computers by R. Feynman in the and tunability, has remained largely unexplored.

Downloaded from https://www.science.org on March 24, 2024

1980s (1) spurred a string of research and engineering developments In this work, we fill this gap by presenting a quantum transducer
in quantum computing hardware that spans from the first experi- that exploits magnon nonlinearities in a ferromagnetic microdisc to
mental realization of a quantum bit, or qubit, with trapped ions (2) to mediate the microwave interaction with spin qubits in silicon car-
the latest 1121-­superconducting qubit processor by IBM (3). Despite bide (SiC). Our hybrid system harnesses parametric magnon effects
the remarkable progress in implementing qubits in a variety of hard- to downconvert the microwave driving frequency and address off-­
ware platforms such as superconducting circuits (4), trapped ions resonant ensembles of spin qubits. This indirect scheme minimizes
(5), defect centers (6, 7), and semiconductor quantum dots (8), cur- the microwave footprint by using highly-­confined magnon stray
rent error rates hamper a monolithic approach to building scalable fields to drive the spin qubits at room temperature. We envision our
quantum systems capable of processing, storage, and transmission of results as a proof of principle of a wafer-­compatible hybrid quan-
quantum information (9, 10). This has motivated research into hybrid tum system that could help expand the quantum engineer’s toolbox
architectures for quantum information networks and leveraging the by enhancing quantum systems with the rich nonlinear physics of
complementary advantages of modular building blocks based on dis- magnons.
tinct physical systems (11, 12).
Under this heterogeneous approach, engineering interfaces that
allow tunable transduction between the distinct physical components RESULTS
become increasingly important. So far, research efforts have explored Hybrid nonlinear magnon-­quantum system
realizing these interfaces via magnetic fields, phonons (the quanta of Our proposed hybrid system, shown in Fig. 1, is composed of two
lattice vibrations), and, to a lesser extent, magnons (the quanta of col- main elements: a nonlinear magnonic system as the transducing
lective spin excitations in magnetic materials) (9, 13). Magnon-­based component and an ensemble of silicon-­vacancy (VSi) defects in SiC
transducers, despite being comparatively unexplored in the field, are as the quantum component. The magnonic system is a 50-­nm-­thick
especially promising for quantum computing due to the unique func- permalloy (Ni81Fe19) disc with a 5.1-­μm diameter in a flux-­closure
tionalities provided by their wide range of magnon interactions and vortex state (Fig. 1B) (26). The magnon spectrum of this system is
intrinsic nonlinear phenomena (14–16). Nevertheless, most work in quantized into degenerate doublet states described by their indices
quantum magnonics has focused instead on linear magnon dynamics (n, m) with n = 0,1,2, … the number of nodes in the radial direction
to mediate the interaction between microwave photons and super- and m = 0, ±1, ±2, … the number of periods in the azimuthal direc-
conducting or spin defect qubits (14–24). Furthermore, the vast ma- tion (fig. S3) (27–30). Examples of these vortex eigenmodes are
jority of the reported hybrid systems are composed of yttrium iron shown in Fig. 1A. We use an on-­chip omega-­shaped antenna that
garnet (YIG) and nitrogen-­vacancy (NV) defects in diamond, both of surrounds the disc to inductively couple to these vortex modes (see
which present fabrication challenges for wafer-­scale integration (14, Fig. 1B). Because of the rotational symmetry of this excitation field,
15, 25). The introduction of magnon nonlinearities into the quantum only modes with m = 0 are efficiently excited by the antenna.
The quantum component (Fig. 1, C and E) is a 200-­nm-­deep VSi-­
rich layer in the 4H SiC polytype substrate (fig. S1). The electronic
1 spin state associated with this defect center has been extensively ex-
Helmholtz-­Zentrum Dresden-­Rossendorf, Institute for Ion Beam Physics and Ma-
terials Research, 01328 Dresden, Germany. 2Faculty of Electrical and Computer En- plored as an atomic-­scale spin qubit with long coherence times even
gineering, Technical University of Dresden, 01062 Dresden, Germany. 3Max Planck at room temperature (31, 32). Figure 1E shows the energy levels and
Institute for Solid State Research, 70569 Stuttgart, Germany. 4Faculty of Physics, optical transitions of the VSi occupying the quasicubic lattice site,
Technical University of Dresden, 01062 Dresden, Germany. 5Institute of Physics,
Technical University of Chemnitz, 09107 Chemnitz, Germany. referred to as VSi(V2) in the literature (33), which is the one we fo-
*Corresponding author. Email: h.​schultheiss@​hzdr.​de cus on in this work. The energy levels of this defect are quadruplets

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 1 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Magnonic system Quantum system

B
C
A Mdyn 2

Min Max
B
Si

c Axis (1000)
f0 VSi(V2) C
(0, 0)
(n, m)
D
FFT magn. (arb. units)
f0 Min Max E
f−
2
(0, 12) ES f
|+3/2
1
(n, m) f+ 2
ISC
|+1/2
ZFS

|−1/2
f+ f− 1

(n, m) GS |−3/2

Downloaded from https://www.science.org on March 24, 2024

(1, ±12) B =0
0 B

Fig. 1. Parametric quantum transducer based on vortex magnons. (A) Parametric generation of magnons via three-­magnon splitting. The initial pump mode f0 splits
into secondary modes f+ and f− following specific selection rules. (n, m) are the radial and azimuthal indices of the quantized vortex magnons. Mdyn denotes the dynamic
component of the magnetization vector. (B) A permalloy (Py) disc with 5.1-­μm diameter and 50-­nm thickness on top of a SiC substrate hosting an ensemble of VSi(V2)
defects. An on-­chip antenna surrounding the disc is used to excite the vortex magnons. (C) Crystallographic structure of the 4H SiC polytype showing the VSi(V2) defect
residing at the quasicubic lattice site. (D) General coupling principle. The ν1,2 resonances of the VSi(V2) cross the secondary magnon modes at two distinct field ranges
(dashed circles) allowing pure-­magnon addressability of the spin qubits. The observed branching of the secondary f+ and f− modes with increasing out-­of-­plane field
∣B∣ corresponds to the exchange-­induced splitting of the degenerate doublets as the vortex transitions into a vortex-­cone state (29, 64). (E) Ground state energy level
structure and optical transitions of VSi(V2). Intersystem crossing (ISC) allows spin initialization and readout. At ∣B∣ ≠ 0, energy levels split due to Zeeman interaction. ZFS,
zero-­field splitting. The spatial profiles of the modes in (A) and their FFT (fast Fourier transform) spectra in (D) were obtained using micromagnetic simulations.

with S = 3/2, split at both ground and excited states into ms = ±1/2, ±3/2 into the doublets (0, ∓12) at f = 2.55 GHz and (1, ±12) at f = 3.55 GHz
(31, 34–36). A spin-­dependent intersystem crossing enables spin (see Fig. 1A) (30). Notably, as the vortex transitions into a vortex-­
initialization and readout via photoluminescence (PL). At zero mag- cone state with increasing out-­of-­plane fields ∣B∣, the secondary
netic field, the energy levels are spin-­split by the zero-­field splitting modes show minimal change, facilitating frequency crossing points
of 2D = 70 MHz, where D is the crystal field constant. At ∣B∣ ≫ 2.5 mT, with the electron spin resonance of VSi(V2). At these crossings, the
the VSi(V2) spin resonance frequencies ν1,2 shift linearly with ∣B∣ VSi(V2) are degenerate only to the secondary modes, leaving the pump
due to Zeeman splitting following mode f0 and the antenna microwaves off-­resonant. Figure 1D sum-
marizes this indirect coupling scheme between the secondary
ν1,2 = ν0 ± D(3cos2 θ − 1) (1) parametric magnon modes and the VSi(V2).
with ν0 = γ∣B∣, ∣γ∣ = 28 MHz/mT being the electron gyromagnetic A remarkable consequence of our coupling mechanism is that
ratio, and θ being the angle between the magnetic field and the c axis the interaction between the magnons and the VSi(V2) is unidirec-
(31, 34, 37). tional. While 3MS is itself a reversible process [three-­magnon con-
At the core of this hybrid system is a “three-­wave mixing” process fluence (41, 42)], the generation of the initial pump mode f0 by the
involving three vortex modes: an initial pump mode f0 with m = 0 VSi(V2) spins is not possible as it would require simultaneous excita-
and two secondary modes f+ and f− with m ≠ 0. Above a certain tion of the secondary modes which, in turn, lie at different frequen-
microwave power threshold the pump mode f0 becomes unstable cies f− and f+.
and spontaneously splits into the two secondary modes (38, 39). The
selection rules of this three-­magnon splitting (3MS) process require Magnon-­driven VSi spins
the secondary magnon modes to have opposite azimuthal index m, We start by experimentally demonstrating the driving of VSi(V2) spins
but different radial indices n which, in return, leads to distinct fre- by the parametric magnons in the vortex disc. Given the magnon-­
quencies f− = f0/2 − Δf and f+ = f0/2 + Δf (40, 41). For the particular driven regime occurs at the two crossing points between the VSi(V2)
case of the 5.1-­μm-­diameter vortex disc, it was shown that high-­ resonances and the f− and f+ modes (dashed circles in Fig. 1D), we
power microwave excitation at fexc = 6.1 GHz results in the sponta- use the magnetic field as an independent parameter to tune the
neous splitting of the directly-­excited pump mode with indices (0, 0) VSi(V2) into and out of resonance with these modes while tracking

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 2 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

the PL from the defects through optically detected magnetic reso- We attribute this downconversion to the parametric generation of
nance (ODMR) at room temperature. As a reference experiment, we secondary magnons in the disc via the 3MS process detailed in the
first exclude the influence of magnons on the ODMR spectra by col- previous section. In good agreement with the simulation data shown
lecting the PL of VSi(V2) located at the center of an empty antenna, in Fig. 1D, the parametric magnon modes, with frequencies f− and
i.e., without the vortex disc. The resulting ODMR spectra, shown in f+, are resonant to the spin defects at ∣B∣ = 90 mT and ∣B∣ = 115 mT,
Fig. 2A, display the expected linear field dependence of the VSi(V2) respectively, and couple to the VSi(V2) with a spin-­magnon coupling
spin resonances as previously described by Eq. 1 for θ = 0, i.e., with strength g. The absence of a level anti-­crossing around fexc ∼ 6 GHz
out-­of-­plane magnetic fields. reveals that the intrinsic decoherence mechanisms from the ma-
The ODMR spectra of VSi(V2) defects below the vortex disc are gnons, κm, and from the VSi ensembles, κq, overcome the coupling
shown in Fig. 2B. Here, we distinguish two distinct resonances of strength, κm, κq > g, putting the system in the weak coupling regime.
different origins: the resonances increasing linearly with the field 4g 2
This regime is defined by cooperativities C < 1, with C ≡ κ κ .
and the off-­diagonal resonances around fexc ∼ 6 GHz. The linear m q
The large ODMR linewidth observed at fexc ∼ 6 GHz in Fig. 2C
resonances show similarities with the antenna-­driven resonances in
can have multiple causes. On one hand, the 9 dBm of microwave
Fig. 2A, however, are a result of the combined action of the micro-
power used to drive the 3MS is sufficiently high to broaden the
wave antenna fields and the dynamic stray fields from the linear vor-
ODMR linewidth and distort its lineshape (fig. S6). On the other
tex dynamics. Given that the microwave power delivered by the
hand, the counter-­propagating secondary modes can induce Doppler
antenna is kept constant, the increase of the ODMR intensity along
broadening, as has been observed with whispering gallery modes
the diagonal is an indication of the concomitant increase of the dy-
coupled to atoms (43). In addition, inhomogeneous broadening
namic stray fields, which together result in overall larger microwave
caused by the interaction with the thermal-­magnon bath could also
driving fields. This increase in dynamic stray fields occurs because
play an important role.

Downloaded from https://www.science.org on March 24, 2024

the vortex magnetization is more resonantly excited when the exci-
tation frequency fexc approaches the resonant frequency of the pump
mode (fig. S4). Threshold activation of magnon-­driven VSi
In contrast, the off-­diagonal resonances around ∣B∣ = 90 mT and Now, we demonstrate the threshold activation of the VSi(V2) spins
∣B∣ = 115 mT in Fig. 2B are a result of a purely-­magnonic driving of driven by the parametric magnons in the vortex disc. We complement
the VSi(V2) spin transitions. To illustrate this, in Fig. 2C, we show the ODMR measurements on VSi(V2) with micro-­focused Brillouin
ODMR spectrum at ∣B∣ = 90 mT as extracted along the dashed line light scattering microscopy (μBLS) on the disc to elucidate the con-
in Fig. 2B. At this field, the VSi(V2) resonances are centered around current magnon dynamics driving the changes in the PL of VSi(V2)
ν0 ∼ 2.5 GHz (see second x axis on top of Fig. 2B). The apparent con- centers. Figure 3A shows a power-­dependent ODMR spectrum at
tradicting presence of ODMR intensities at fexc ∼ 6 GHz suggests a ∣B∣ = 99 mT, where the f− magnon mode drives the VSi(V2) spin
microwave downconversion to the VSi(V2) resonance frequencies. transitions (as seen in Fig. 2B). While typical ODMR spectra are

A B C

E (GHz)
fexc
f f+ f0
5.9

3MS
g
2.5
f VSi
Purely magnon
driven

E (GHz)

ve n fexc
- dri ri ven
f exc + f 0-d g
f exc 2.5

f0 VSi

B (mT) B (mT)

Fig. 2. Optical detection of magnon-­driven VSi(V2) spin transitions. (A) Reference ODMR spectra as a function of the external magnetic field ∣B∣ measured at the cen-
ter of a microwave antenna without the vortex disc (see inset). The diagonal ΔPL/PL intensity corresponds to the field-­dependent VSi(V2) resonances. The discrete-­like
resonances are due to coarse increments in the applied magnetic field (see the Supplementary Materials). (B) ODMR spectra as a function of external magnetic field
∣B∣ for VSi(V2) ensembles below the vortex disc (see inset). The off-­diagonal signal around fexc ∼ 6 GHz results from VSi(V2) spin transitions purely driven by the parametric
magnon modes f− and f+. (C) ODMR spectrum at ∣B∣ = 90 mT, as extracted along the dashed line in (B). The energy diagram illustrates the origin of the two distinct features
in the spectrum. g denotes the spin-­magnon coupling strength. The spectra in (A) and (B) were both obtained by applying 9 dBm of microwave power. 3MS, three-­
magnon splitting.

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 3 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

continuous with increasing microwave driving power (fig. S6), the

A
measured ODMR spectra depict a threshold-­like power depen-
dence. This nonlinear onset of ODMR intensity at the off-­resonant
fexc ∼ 6.1 GHz suggests that the appearance of the resonant magnon
f− occurs after the microwave power exceeds a critical threshold am-
plitude. This critical amplitude is associated with the instability
threshold of the pump mode f0 as discussed before.
To further clarify the evolution of this nonlinear process, we ex-
tract the average ODMR contrast over a 20-­MHz-­frequency win-
PL/PL (%)

0.07 dow centered at fexc = 6.24 GHz and identify three separate power
ranges: (I) below threshold, (II) above threshold, and (III) saturation
(Fig. 3B). For each range, Fig. 3 (C and D) shows the average ΔPL/
PL of the VSi(V2) spins and the average μBLS spectrum from the
0 B = 99 mT disc, respectively. Below threshold (I), there is no change in the PL
from the VSi(V2) spins as there is no magnon mode at the VSi(V2)
B resonance frequencies. In this regime, only the pump mode f0, which
is nonresonant to the VSi(V2) spins, is present. Above threshold (II),
fexc = 6.24 GHz there is a nonzero μBLS signal intensity at the resonance frequency
of the secondary magnon f− (Fig. 3D). As the magnon mode f−
matches the VSi(V2) resonance frequencies, it drives VSi(V2) spin
I II III

Downloaded from https://www.science.org on March 24, 2024

transitions which induces a change in PL. At saturation (III), the
intensity of the secondary mode f− is the largest, which gives rise to
the high change in PL observed in Fig. 3C. By further increasing the
excitation power, higher-­order interaction processes between the
secondary magnons, such as four-­magnon scattering, become rele-
C vant. Consequently, this leads to decoherence, limiting the energy
flux between the pump and secondary magnons (30, 38) and result-
I II III ing in the observed saturation in ΔPL/PL.

Room temperature spin-­magnon coupling

Having experimentally demonstrated the working principle of our
hybrid transducer, we now characterize the room temperature spin-­
magnon coupling between the secondary vortex modes and the
VSi(V2) spins. We build upon the theoretical work of Candido et al.
(44) and define the spin-­magnon coupling strength g as follows
√
D 2
g(rVSi ) = γμ0 hrf⊥ (rVSi ) (2)
VSi I VSi II VSi III 4
f0 f0 with μ0 being the vacuum permeability, hrf⊥ being the magnon stray
fields perpendicular to the VSi(V2) spin axis, and rVSi = (x, y, z) being
f0 the coordinates of the VSi(V2) defect. Using a combination of micro-
f f+ f f+ magnetic simulations and analytical calculations (see the Supplemen-
tary Materials), we obtain the magnon stray fields hrf⊥ and use Eq. 2 to
derive the spatial profile of the spin-­magnon coupling strength g be-
low the vortex disc. Figure 4 (B and C) shows the calculated coupling
strength for the secondary magnon modes f− and f+ at a plane 175 nm
below the disc (depicted in Fig. 4A), which lies within the VSi(V2)-­
Fig. 3. Threshold process of the parametric driving scheme. (A) ODMR spectra
rich layer in our sample (fig. S1). Figure 4D shows the variation of the
for increasing microwave excitation powers at ∣B∣ = 99 mT for VSi(V2) below the disc
(see inset). (B) ODMR contrast averaged over a 20-­MHz-­frequency window cen-
coupling strength in Fig. 4 (B and C) along a line at x = 0 μm. Because
tered at fexc= 6.24­GHz showing the characteristic threshold behavior of the three-­ of the high spin density of permalloy, which results in a high satura-
magnon splitting. Three distinct microwave power ranges can be identified as tion magnetization μ0MS,Py = 1 T (45, 46), we obtain coupling
follows: (I) below threshold, (II) above threshold, and (III) saturation. For each range, strengths as large as 10-­MHz. For comparison purposes, the coupling
the average ODMR spectrum is shown in (C) and the corresponding average BLS strengths of other reported NV/YIG hybrid systems is around 1-­MHz
spectrum is shown in (D). VSi in (D) refers to ν0. Solid line in (B) is a sigmoidal fit to for spin-­magnon distances in the order of tens of nanometers (47, 48).
the data. These lower coupling strengths are a result of the much lower
spin density of YIG, which yields a smaller saturation magnetization
μ0MS,YIG = 0.170 T (49).
We now show the coupling strength for the f+ and f− modes
across the thickness of the SiC substrate in Fig. 4 (F and G) (see

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 4 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

g/2 (MHz)
A Py B C Hz D
Hz M
3
M 3 f+
B
Hz Hz
M M
3 3

z z = −17
5 nm
f
y x f f+

g/2 (MHz)
E Py F G H
q = 15 M
B Hz
q = 15 M Hz

f−

Downloaded from https://www.science.org on March 24, 2024

Hz
3M
Hz

z f
3M

q
y=0
y x f f+

Fig. 4. Room-­temperature coupling between the vortex magnons and the VSi(V2) spins. (A) Schematic depicting plane located 175-­nm below the bottom surface of
the disc, where the spin-­magnon coupling strength is calculated. (B and C) Coupling strength g for the parametric modes f− and f+ at ∣B∣ = 90­mT and ∣B∣ = 130­mT, re-
spectively, at the plane shown in (A). (D) Coupling strength extracted along x = 0­ μm from the intensity maps shown in (B) and (C). (E) Schematic depicting cross-­sectional
plane located at y = 0 ­μm starting 25­nm below the disc. (F and G) Coupling strength g for the parametric magnons f− and f+ at ∣B∣ = 90­mT and ∣B∣ = 130­mT, respec-
tively, at the plane shown in (E). (H) Coupling strength for the f− and f+ modes along x = −1.05 μ­ m and x = 0.6­ μm as extracted from (F) and (G), respectively.

cross-­sectional plane in Fig. 4E) and compare it with the intrinsic DISCUSSION
decoherence rates of the system. We estimate κq = 2π × 15 MHz for In summary, the hybrid quantum transducer presented in this work
the spin ensemble, and κm,− = 2π × 89 MHz and κm,+ = 2π × 73 MHz exploits magnon nonlinearities to enhance microwave transduction
for the damping rates of the f− and f+ modes, correspondingly (see the to spin qubits. The resulting spin-­magnon coupling scheme points
Supplementary Materials). Here, it is evident that the spin-­magnon to a transduction behavior different from the one observed in hy-
coupling strengths at the depth of the VSi-­rich layer are smaller than brid systems with linearly-­excited magnons. Whereas the dipolar
the decoherence rates from the magnons and from the spin ensemble, coupling between the spin centers and the linear magnons is always
confirming that the system is in the weak coupling regime. None- “on,” the parametric processes at the core of our hybrid system rep-
theless, we can also observe an important consequence of the pure resent a way to selectively tune the spin-­magnon coupling “on” and
magnon-­spin coupling mechanism. As the antenna microwaves are “off ”. With such control of the coupling, it would be possible to pro-
off-­resonant to the spin defects, the microwave driving fields are of tect the spin centers against any resonant magnon noise-­induced
purely magnonic origin. These magnon stray fields decay quickly in decoherence. Moreover, given the nature of the resulting coupling is
space with a 1/r3 dependence, confining the microwave control to a purely of magnonic origin, the microwave driving fields are con-
submicrometer volume below the disc. This is in stark contrast to fined to submicrometer volumes, in stark contrast with the micro-
the ∼1-­mm3-­sized volumes typically achieved with transmission wave fields from antennas, which typically extend over hundreds of
line microwave antennas (9). As shown in Fig. 4 (F and G), the ma- micrometers on the sample plane (51). Looking forward, this could
gnon stray fields result in g/2π > 15 MHz confined to the top ∼100 nm be an interesting path toward reducing the spillover cross-­talk (52)
close to the disc. To expand on this microwave confinement, in as the number of qubits scales in future quantum chips.
Fig. 4H, we plot the spatial decay of the coupling strength at x = Our proposed system represents a departure from the traditional
–1.05 μm (for f−; Fig. 4F) and at x = 0.6 μm (for f+; Fig. 4G). Al- NV/YIG composition by using permalloy and silicon vacancies in sili-
though spin defects at z = –175 nm (like the defects in our sample) con carbide, two materials highly compatible with industry-­standard
experience coupling strengths in the range of 5 to 7 MHz, there is a fabrication protocols and easier integrability into heterogeneous quan-
broad depth-­range to engineer defects with stronger coupling. Precise tum architectures (53–55). Despite the large spin-­magnon coupling
control of the implantation energies allows the creation of defects closer strengths calculated, our system is not in the strong coupling regime,
to the surface to achieve large spin-­m agnon coupling strengths due to the large damping rates from the vortex magnons and the VSi
but still deep enough to avoid detrimental surface effects (50). spin ensemble. We expect that coupling to single spin defects instead of

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 5 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

spin ensembles will result in C > 1 (fig. S7), therefore bringing the sys- For the ODMR measurements, the sample was placed on a sam-
tem into the strong coupling regime. Moreover, by going to cryogenic ple holder that allowed for laser illumination via the backside of the
temperatures, the damping processes intrinsic to metallic ferromag- SiC substrate, to enable probing of the VSi defects directly below the
netic thin films should be reduced (56). Another interesting perspec- permalloy disc. The ΔPL/PL signal was obtained using a lock-­in de-
tive is exploiting the nonlinear 3MS process to generate squeezed tection scheme, with the photovoltage from the photodiode sampled
magnon states (57–59). Such nonclassical states have the potential to by a lock-­in amplifier (Signal Recovery 7265) locked to a reference
exponentially enhance the coupling strengths and cooperativities in sinusoidal signal which modulates the amplitude of the microwaves
hybrid quantum systems (60, 61). being fed to the sample.
Quantum magnonics focused thus far on coupling long-­range For both confocal and μBLS microscopes, a permanent magnet
magnon propagations to mediate microwave interactions with spin on a motorized stage is used to deliver out-­of-­plane magnetic fields
qubits, naturally directing research efforts into YIG-­based hybrid along the crystal axis of the SiC substrate. All the experiments in this
systems due to its low damping coefficient. This, however, left a large work were performed at room temperature.
body of magnon phenomena unexplored. By introducing a nonlin-
ear magnonic system, we provide alternative perspectives for engi- Micromagnetic simulations
neering quantum interfaces to spin qubits and motivate further The micromagnetic simulations for Fig. 1 (A and D) were obtained
research into uncovering the interesting phenomena lying at the with MuMax3 (63). The disc was discretized into 512 × 512 × 1 rect-
intersection of nonlinear magnonics and quantum systems. angular cells, with lengths Δx = 10 nm, Δy = 10 nm, and Δz = 50 nm.
The standard material parameters for permalloy were used: satura-
tion magnetization MS = 810 kA/m, exchange stiffness Aex = 13 pJ/mm,
MATERIALS AND METHODS Gilbert damping parameter αG = 0.007, and gyromagnetic ratio γ =
Sample fabrication

Downloaded from https://www.science.org on March 24, 2024

29.6 GHz/T. The equilibrium configuration was set to be a vortex
The 10-­mm × 10-­mm sample was diced from a high-­purity semi-­ state with chirality c = −1 and polarity p = 1.
insulating 4H-­SiC wafer with a crystal axis along the surface nor- The magnetization dynamics were excited by a spatially homoge-
mal. To create the VSi defects, the entire sample was irradiated with neous time-­varying sinusoidal magnetic field of 5-­mT amplitude and
protons at a fluence of 1 × 1015 cm−2 with an energy of 30 keV. This 6.1 GHz in the out-­of-­plane direction. The time evolution of the mag-
procedure yields a theoretical penetration depth of the protons netization was recorded every 10 ps for 100 ns, for a total of 10,000
around 200 nm, which was obtained with Monte Carlo simulations samples. To obtain the power spectrum of the modes in the disc, a fast
in the software SRIM (fig. S1). After irradiation, electron beam Fourier transform (FFT) in time for each cell was calculated and the
lithography, sputter deposition, and subsequent liftoff were per- resulting spectrum was spatially averaged over the whole disc. The
formed to fabricate the Ni81Fe19 disc (5.1-­μm diameter and 50-­nm spatial mode profiles in Fig. 1A were obtained by performing a back-
thick) and the on-­chip antenna (6 nm Cr/150 nm Au). ward windowed FFT around the frequency of interest.

Experimental setups Calculation of the magnon stray fields

To measure ODMR, a custom-­built confocal scanning microscope To calculate the magnon magnetic fields, we consider the N cells of
was used (fig. S2). The VSi defects were optically excited with a 785-­nm the discretized disc as N point dipoles located at the center of each
continuous-­wave laser diode (Thorlabs LP785-­SF100) which is fo- cell and calculate the total dipolar magnetic field at an arbitrary
cused onto the sample by a microscope objective (Zeiss, LD Plan point P in space, as the resulting collective contribution from
Neofluar 63×/NA 0.75). This objective has an adjustable coverglass each ith cell
correction to compensate for the aberrations introduced by illumi-
nating the VSi defects through the SiC substrate. The emitted PL, N
∑ μ0 3(mi ⋅ ̂ri )̂ri − mi
with a spectrum centered at 900-­nm, is collected through the same μ0 H(m, r) = (3)
i
4π ri3
objective and then filtered by appropriate dichroic mirrors and long-
pass filters before finally being measured by a silicon avalanche pho- with mi being the vector dipole moment of the ith cell and ̂ri being
todetector (Thorlabs, APD440A). the normalized distance vector from the ith cell to point P. By re-
A dedicated μBLS microscope was used to detect the magnon peating this algorithm for all the time samples of the micromagnetic
dynamics within the permalloy disc. μBLS uses the inelastic scatter- simulation, we obtain the time-­dependent magnetic field μ0H(t, m,
ing of photons with magnons to perform a space-­and frequency-­ r) at point P. We use this algorithm for calculating the dipolar mag-
resolved detection of the magnon intensity (62). For this purpose, a netic fields at the evaluation planes shown in Fig. 4. The resulting
continuous-­wave 532-­nm solid-­state laser (Novanta Photonics, To- fields will contain the contributions from the overall magnetization
rus 532) is focused onto the sample using a microscope objective dynamics in the disc. To isolate the contribution from a specific vor-
(Zeiss, LD EC Epiplan Neofluar 100×/NA 0.75) yielding a spot size tex magnon, f+ or f−, we perform an FFT in time for each evaluation
of approximately 320 nm. The scattered light is collected by the same point and then a backward windowed FFT around the frequency f+
objective and analyzed with a high-­resolution tandem Fabry-­Perot or f−. The magnetic fields used in Eq. 2 are the transversal compo-
interferometer (JRS Scientific Instruments, TFP-­2). nents of these reconstructed magnon magnetic fields.
To excite the magnetization dynamics in the permalloy disc and
drive spin transitions in the VSi defects, microwaves from a signal
generator (Keysight, 5183B) were fed to the on-­chip antenna through Supplementary Materials
nonmagnetic RF probes (GGB Industries Inc., Model 40A), which This PDF file includes:
are positioned onto the sample using xyz micro-­positioning stages. Sections S1 to S5

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 6 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

Figs. S1 to S7 28. M. Buess, T. P. J. Knowles, R. Höllinger, T. Haug, U. Krey, D. Weiss, D. Pescia, M. R. Scheinfein,
References C. H. Back, Excitations with negative dispersion in a spin vortex. Phys. Rev. B 71, 104415
(2005).
29. B. Ivanov, C. Zaspel, High frequency modes in vortex-­state nanomagnets. Phys. Rev. Lett.
REFERENCES AND NOTES 94, 027205 (2005).
1. R. P. Feynman, Simulating physics with computers. Int. J. Theor. Phys. 21, 467–488 (1982). 30. K. Schultheiss, R. Verba, F. Wehrmann, K. Wagner, L. Körber, T. Hula, T. Hache, A. Kákay,
2. J. I. Cirac, P. Zoller, Quantum computations with cold trapped ions. Phys. Rev. Lett. 74, A. A. Awad, V. Tiberkevich, A. N. Slavin, J. Fassbender, H. Schultheiss, Excitation of
4091–4094 (1995). whispering gallery magnons in a magnetic vortex. Phys. Rev. Lett. 122, 097202 (2019).
3. D. Castelvecchi, IBM releases first-­ever 1,000-­qubit quantum chip. Nature 624, 238–238 31. H. Kraus, V. A. Soltamov, D. Riedel, S. Väth, F. Fuchs, A. Sperlich, P. G. Baranov, V. Dyakonov,
(2023). G. V. Astakhov, Room-­temperature quantum microwave emitters based on spin defects
4. Y. Nakamura, Y. A. Pashkin, J. S. Tsai, Coherent control of macroscopic quantum states in a in silicon carbide. Nat. Phys. 10, 157–162 (2014).
single-­Cooper-­pair box. Nature 398, 786–788 (1999). 32. M. Widmann, S.-­Y. Lee, T. Rendler, N. T. Son, H. Fedder, S. Paik, L.-­P. Yang, N. Zhao, S. Yang,
5. C. D. Bruzewicz, J. Chiaverini, R. McConnell, J. M. Sage, Trapped-­ion quantum computing: I. Booker, A. Denisenko, M. Jamali, S. A. Momenzadeh, I. Gerhardt, T. Ohshima, A. Gali,
Progress and challenges. Appl. Phys. Rev. 6, 021314 (2019). E. Janzén, J. Wrachtrup, Coherent control of single spins in silicon carbide at room
6. F. Jelezko, J. Wrachtrup, Single defect centres in diamond: A review. Phys. Status Solidi A temperature. Nat. Mater. 14, 164–168 (2015).
203, 3207–3225 (2006). 33. E. Sörman, N. T. Son, W. M. Chen, O. Kordina, C. Hallin, E. Janzén, Silicon vacancy related
7. P. G. Baranov, A. P. Bundakova, A. A. Soltamova, S. B. Orlinskii, I. V. Borovykh, R. Zondervan, defect in 4H and 6H SiC. Phys. Rev. B 61, 2613–2620 (2000).
R. Verberk, J. Schmidt, Silicon vacancy in SiC as a promising quantum system for 34. E. Janzén, A. Gali, P. Carlsson, A. Gällström, B. Magnusson, N. T. Son, The silicon vacancy in
single-­defect and single-­photon spectroscopy. Phys. Rev. B 83, 125203 (2011). SiC. Phys. B Condens. Matter 404, 4354–4358 (2009).
8. D. Loss, D. P. DiVincenzo, Quantum computation with quantum dots. Phys. Rev. A 57, 35. N. Mizuochi, S. Yamasaki, H. Takizawa, N. Morishita, T. Ohshima, H. Itoh, J. Isoya,
120–126 (1998). Continuous-­wave and pulsed EPR study of the negatively charged silicon vacancy with S
9. D. S. Wang, M. Haas, P. Narang, Quantum interfaces to the nanoscale. ACS Nano 15, = 3/2 and C3v symmetry in n -­type 4H − SiC. Phys. Rev. B 66, 235202 (2002).
7879–7888 (2021). 36. S. A. Tarasenko, A. V. Poshakinskiy, D. Simin, V. A. Soltamov, E. N. Mokhov, P. G. Baranov,
10. S. Pirandola, S. L. Braunstein, Physics: Unite to build a quantum Internet. Nature 532, V. Dyakonov, G. V. Astakhov, Spin and optical properties of silicon vacancies in silicon
169–171 (2016). carbide—A review. Phys. Status Solidi B 255, 1700258 (2018).

Downloaded from https://www.science.org on March 24, 2024

11. C. Monroe, R. Raussendorf, A. Ruthven, K. R. Brown, P. Maunz, L.-­M. Duan, J. Kim, 37. D. Simin, F. Fuchs, H. Kraus, A. Sperlich, P. G. Baranov, G. V. Astakhov, V. Dyakonov,
Large-­scale modular quantum-­computer architecture with atomic memory and photonic High-­precision angle-­resolved magnetometry with uniaxial quantum centers in silicon
interconnects. Phys. Rev. A 89, 022317 (2014). carbide. Phys. Rev. Applied 4, 014009 (2015).
12. Z.-­L. Xiang, S. Ashhab, J. Q. You, F. Nori, Hybrid quantum circuits: Superconducting circuits 38. V. S. L’vov, Nonlinear dynamics and kinetics of magnons, in Nonlinear Waves 3, A. V.
interacting with other quantum systems. Rev. Mod. Phys. 85, 623–653 (2013). Gaponov-­Grekhov, M. I. Rabinovich, J. Engelbrecht, Eds. (Springer Berlin Heidelberg,
13. N. Lauk, N. Sinclair, S. Barzanjeh, J. P. Covey, M. Saffman, M. Spiropulu, C. Simon, 1990), pp. 224–239.
Perspectives on quantum transduction. Quantum Sci. Technol. 5, 020501 (2020). 39. H. Suhl, The theory of ferromagnetic resonance at high signal powers. J. Phys. Chem. Solid
14. D. D. Awschalom, C. R. Du, R. He, F. J. Heremans, A. Hoffmann, J. Hou, H. Kurebayashi, Y. Li, 1, 209–227 (1957).
L. Liu, V. Novosad, J. Sklenar, S. E. Sullivan, D. Sun, H. Tang, V. Tyberkevych, C. Trevillian, 40. R. E. Camley, Three-­magnon processes in magnetic nanoelements: Quantization and
A. W. Tsen, L. R. Weiss, W. Zhang, X. Zhang, L. Zhao, C. W. Zollitsch, Quantum engineering localized mode effects. Phys. Rev. B 89, 214402 (2014).
with hybrid magnonic systems and materials (Invited Paper). IEEE Trans. Quantum Eng. 2, 41. R. Verba, L. Körber, K. Schultheiss, H. Schultheiss, V. Tiberkevich, A. Slavin, Theory of
1–36 (2021). three-­magnon interaction in a vortex-­state magnetic nanodot. Phys. Rev. B 103, 014413
15. Y. Li, W. Zhang, V. Tyberkevych, W.-­K. Kwok, A. Hoffmann, V. Novosad, Hybrid magnonics: (2021).
Physics, circuits, and applications for coherent information processing. J. Appl. Phys. 128, 42. V. L’vov, Wave Turbulence under Parametric Excitation: Applications to Magnets, Springer
130902 (2020). Series in Nonlinear Dynamics (Springer Berlin Heidelberg, 2012).
16. D. Lachance-­Quirion, Y. Tabuchi, A. Gloppe, K. Usami, Y. Nakamura, Hybrid quantum 43. D. W. Vernooy, A. Furusawa, N. P. Georgiades, V. S. Ilchenko, H. J. Kimble, Cavity QED with
systems based on magnonics. Appl. Phys. Express 12, 070101 (2019). high-­​Q whispering gallery modes. Phys. Rev. A 57, R2293–R2296 (1998).
17. D. Lachance-­Quirion, Y. Tabuchi, S. Ishino, A. Noguchi, T. Ishikawa, R. Yamazaki, 44. D. R. Candido, G. D. Fuchs, E. Johnston-­Halperin, M. E. Flatté, Predicted strong coupling of
Y. Nakamura, Resolving quanta of collective spin excitations in a millimeter-­sized solid-­state spins via a single magnon mode. Mater. Quantum Technol. 1, 011001 (2021).
ferromagnet. Sci. Adv. 3, e1603150 (2017). 45. K. Ounadjela, H. Lefakis, V. S. Speriosu, C. Hwang, P. S. Alexopoulos, Thickness
18. D. Lachance-­Quirion, S. P. Wolski, Y. Tabuchi, S. Kono, K. Usami, Y. Nakamura, dependence of magnetization and magnetostriction of NiFe and NiFeRh films. J. Phys.
Entanglement-­based single-­shot detection of a single magnon with a superconducting Colloques 49, C8-­1709–C8-­1710 (1988).
qubit. Science 367, 425–428 (2020). 46. P. E. Mijnarends, S. Sahrakorpi, M. Lindroos, A. Bansil, Angle-­resolved photoemission
19. Y. Tabuchi, S. Ishino, A. Noguchi, T. Ishikawa, R. Yamazaki, K. Usami, Y. Nakamura, Coherent spectra, electronic structure, and spin-­dependent scattering in Ni1–xFex Permalloys. Phys.
coupling between a ferromagnetic magnon and a superconducting qubit. Science 349, Rev. B 65, 075106 (2002).
405–408 (2015). 47. T. Neuman, D. S. Wang, P. Narang, Nanomagnonic cavities for strong spin-­magnon
20. P. Andrich, C. F. de las Casas, X. Liu, H. L. Bretscher, J. R. Berman, F. J. Heremans, P. F. Nealey, coupling and magnon-­mediated spin-­spin interactions. Phys. Rev. Lett. 125, 247702
D. D. Awschalom, Long-­range spin wave mediated control of defect qubits in (2020).
nanodiamonds. Npj Quantum Inf. 3, 28 (2017). 48. D. S. Wang, T. Neuman, P. Narang, Spin emitters beyond the point dipole approximation
21. T. van der Sar, F. Casola, R. Walsworth, A. Yacoby, Nanometre-­scale probing of spin waves in nanomagnonic cavities. J. Phys. Chem. C 125, 6222–6228 (2021).
using single electron spins. Nat. Commun. 6, 7886 (2015). 49. M. A. Gilleo, S. Geller, Magnetic and crystallographic properties of substituted
22. F. Casola, T. van der Sar, A. Yacoby, Probing condensed matter physics with magnetometry yttrium-­iron garnet, 3Y2O3 xM2O3 (5–x)Fe2O3. Phys. Rev. 110, 73–78 (1958).
based on nitrogen-­vacancy centres in diamond. Nat. Rev. Mater. 3, 17088 (2018). 50. L. Rondin, G. Dantelle, A. Slablab, F. Grosshans, F. Treussart, P. Bergonzo, S. Perruchas,
23. M. Fukami, D. R. Candido, D. D. Awschalom, M. E. Flatté, Opportunities for long-­range T. Gacoin, M. Chaigneau, H.-­C. Chang, V. Jacques, J.-­F. Roch, Surface-­induced charge
magnon-­mediated entanglement of spin qubits via on-­and off-­resonant coupling. PRX state conversion of nitrogen-­vacancy defects in nanodiamonds. Phys. Rev. B 82, 115449
Quantum 2, 040314 (2021). (2010).
24. Y. Li, V. G. Yefremenko, M. Lisovenko, C. Trevillian, T. Polakovic, T. W. Cecil, P. S. Barry, 51. N. P. de Leon, K. M. Itoh, D. Kim, K. K. Mehta, T. E. Northup, H. Paik, B. S. Palmer, N. Samarth,
J. Pearson, R. Divan, V. Tyberkevych, C. L. Chang, U. Welp, W.-­K. Kwok, V. Novosad, S. Sangtawesin, D. W. Steuerman, Materials challenges and opportunities for quantum
Coherent coupling of two remote magnonic resonators mediated by superconducting computing hardware. Science 372, eabb2823 (2021).
circuits. Phys. Rev. Lett. 128, 047701 (2022). 52. M. Sarovar, T. Proctor, K. Rudinger, K. Young, E. Nielsen, R. Blume-­Kohout, Detecting
25. S. Kosen, A. F. van Loo, D. A. Bozhko, L. Mihalceanu, A. D. Karenowska, Microwave magnon crosstalk errors in quantum information processors. Quantum 4, 321 (2020).
damping in YIG films at millikelvin temperatures. APL Mater. 7, 101120 (2019). 53. T. Kimoto, J. A. Cooper, Fundamentals of Silicon Carbide Technology: Growth,
26. T. Shinjo, T. Okuno, R. Hassdorf, K. Shigeto, T. Ono, Magnetic vortex core observation in Characterization, Devices and Applications (John Wiley & Sons, 2014).
circular dots of permalloy. Science 289, 930–932 (2000). 54. S. Castelletto, A. Boretti, Silicon carbide color centers for quantum applications. J. Phys.
27. M. Buess, R. Höllinger, T. Haug, K. Perzlmaier, U. Krey, D. Pescia, M. R. Scheinfein, D. Weiss, Photonics 2, 022001 (2020).
C. H. Back, Fourier transform imaging of spin vortex eigenmodes. Phys. Rev. Lett. 93, 55. S. Mathuna, T. O’Donnell, N. Wang, K. Rinne, Magnetics on silicon: An enabling
077207 (2004). technology for power supply on chip. IEEE Trans. Power Electron. 20, 585–592 (2005).

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 7 of 8

S c i e n c e A d v a n c e s | R e s e ar c h A r t i c l e

56. Y. Zhao, Q. Song, S.-­H. Yang, T. Su, W. Yuan, S. S. P. Parkin, J. Shi, W. Han, Experimental S.-­Y. Lee, J. Wrachtrup, Quantum properties of dichroic silicon vacancies in silicon carbide.
investigation of temperature-­dependent gilbert damping in permalloy thin films. Sci. Phys. Rev. Applied 9, 034022 (2018).
Rep. 6, 22890 (2016). 67. R. Verba, V. Tiberkevich, A. Slavin, Damping of linear spin-­wave modes in magnetic
57. H. Y. Yuan, Y. Cao, A. Kamra, R. A. Duine, P. Yan, Quantum magnonics: When magnon nanostructures: Local, nonlocal, and coordinate-­dependent damping. Phys. Rev. B 98,
spintronics meets quantum information science. Phys. Rep. 965, 1–74 (2022). 104408 (2018).
58. M. Elyasi, Y. M. Blanter, G. E. W. Bauer, Resources of nonlinear cavity magnonics for
quantum information. Phys. Rev. B 101, 054402 (2020). Acknowledgments: We thank K. Schultheiss for fruitful discussions and for carefully reading
59. L.-­A. Wu, H. J. Kimble, J. L. Hall, H. Wu, Generation of squeezed states by parametric down the manuscript, as well as A. Henschke for support with the experimental setup. We thank
conversion. Phys. Rev. Lett. 57, 2520–2523 (1986). U. Kentsch and the Ion Beam Center (IBC) at Helmholtz-­Zentrum Dresden-­Rossendorf (HZDR) for
60. C. Leroux, L. C. G. Govia, A. A. Clerk, Enhancing cavity quantum electrodynamics via the proton implantation. Support by the Nanofabrication Facilities Rossendorf (NanoFaRo) at
antisqueezing: Synthetic ultrastrong coupling. Phys. Rev. Lett. 120, 093602 (2018). IBC is gratefully acknowledged. Funding: This work was supported by the German Research
61. W. Qin, A. Miranowicz, P.-­B. Li, X.-­Y. Lü, J. Q. You, F. Nori, Exponentially enhanced Foundation (DFG) under the grant SCHU 29224-­1 awarded to H.S. Author contributions:
light-­matter interaction, cooperativities, and steady-­state entanglement using Conceptualization: M.B., F.J.T.G., T.Ha., T.Hu., L.K., Y.B., M.He., J.F., G.A., H.S. Formal analysis: M.B.
parametric amplification. Phys. Rev. Lett. 120, 093601 (2018). Funding acquisition: H.S. Investigation: M.B., F.J.T.G., M.Ho., C.H., Y.B., H.S. Methodology: M.B.,
62. T. Sebastian, K. Schultheiss, B. Obry, B. Hillebrands, H. Schultheiss, Micro-­focused brillouin F.J.T.G., T.Ha., M.Ho., T.Hu., J.H., Y.B., G.A., H.S. Project administration: M.He., J.F., G.A., H.S.
light scattering: Imaging spin waves at the nanoscale. Front. Phys. 3, 10.3389/ Resources: T.Ha., M.Ho., Y.B. Software: M.B., H.S. Supervision: F.J.T.G., M.He., H.S. Validation: M.B.,
fphy.2015.00035 (2015). F.J.T.G., C.H., L.K., Y.B., H.S. Visualization: M.B., M.Ho., H.S. Writing–original draft: M.B.
63. A. Vansteenkiste, J. Leliaert, M. Dvornik, M. Helsen, F. Garcia-­Sanchez, Writing–review & editing: M.B., F.J.T.G., T.Ha., C.H., L.K., Y.B., J.F., H.S. Competing interests: The
B. Van Waeyenberge, The design and verification of mumax3. AIP Adv. 4, 107133 authors declare that they have no competing interests. Data and materials availability: All
(2014). data needed to evaluate the conclusions in the paper are present in the paper and/or the
64. B. A. Ivanov, G. M. Wysin, Magnon modes for a circular two-­dimensional easy-­plane Supplementary Materials.
ferromagnet in the cone state. Phys. Rev. B 65, 134434 (2002).
65. L. Körber, “Theory and simulation on nonlinear spin-­wave dynamics in magnetic vortices,” Submitted 10 April 2023
thesis, Technische Universität Dresden, Dresden (2019). Accepted 15 February 2024
66. R. Nagy, M. Widmann, M. Niethammer, D. B. R. Dasari, I. Gerhardt, O. O. Soykal, Published 20 March 2024

Downloaded from https://www.science.org on March 24, 2024

M. Radulaski, T. Ohshima, J. Vučković, N. T. Son, I. G. Ivanov, S. E. Economou, C. Bonato, 10.1126/sciadv.adi2042

Bejarano et al., Sci. Adv. 10, eadi2042 (2024) 20 March 2024 8 of 8