{"id":"8552fbb6-f104-43bc-a1cf-9a3bedbaa688","arxiv_id":"2412.02827","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"The 13-fold rise in effective magnetic damping in a YIG film on GGG at millikelvin temperatures is mostly caused by the inhomogeneous stray field of the partially magnetized GGG substrate, which can broaden the resonance linewidth up to 6.7 times.","lead":"The paper measures magnetic damping in a 97 nm yttrium iron garnet film on a GGG substrate down to 30 mK and attributes the large linewidth increase to the substrate's own magnetic field. The result points to a fixable materials problem for quantum magnonics rather than an intrinsic limit of YIG.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Dominance claim hinges on dismissing dynamic YIG-GGG coupling via an unreviewed companion preprint's k=0 argument; the strongly inhomogeneous GGG stray field makes the FMR mode nonuniform, so that argument may not apply.","rationale":"The central claim is that the non-uniform quasi-static GGG stray field dominates the observed FMR broadening at millikelvin temperatures. The evidence for this is a set of micromagnetic simulations and a semianalytical independent-resonator model that reproduce the linewidth enhancement factor at 2 K and 250 mT. A critical step is the dismissal of dynamic YIG-GGG dipolar coupling, which is based entirely on the companion preprint Ref. [61]. The preprint's k=0 argument is not automatically applicable here because the strongly inhomogeneous GGG field makes the coupled mode nonuniform in space, so it has a finite wavevector content even when driven by a uniform microwave field. The reader's weakest assumption flags dynamic coupling as a concern; this stress-test sharpens that concern by identifying the exact place where the paper's justification is insufficient. The concern is concrete and testable: adding a dynamic GGG susceptibility to the simulation would reveal whether the static-field-only model overestimates the inhomogeneous broadening while a dynamic loss mechanism fills the gap. The paper has independent strengths: the room-temperature simulation matches experiment, the GGG magnetization is measured, and the analytical and numerical approaches agree with each other. The concern does not disprove the mechanism, but it is a sufficient reason to require the additional check before accepting the dominance attribution. Therefore the existing CONDITIONAL verdict is appropriate; no change is needed.","tokens_in":14328,"tokens_out":9112,"duration_ms":101720,"concrete_test":"Compute the FMR response for the same 2 K geometry and 250 mT bias using the static GGG stray field plus a dynamic GGG contribution, modeled as a frequency-dependent complex susceptibility or by coupling the YIG LLG to a paramagnetic spin bath with the measured GGG EPR linewidth (about 400 mT). Compare the resulting FWHM to the static-field-only simulation (2.42 mT) and the experiment. If the dynamic contribution changes the linewidth by more than about 0.5 mT (about 20%) at the matching point, the static-field mechanism is not dominant and the central claim weakens.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central attribution — that the quasi-static inhomogeneous GGG stray field is the dominant source of the low-temperature FMR broadening — depends on excluding dynamic dipolar coupling between YIG and the paramagnetic GGG spins. This exclusion is made in Sec. III by citing the companion preprint Ref. [61], which reportedly shows that the coupling is pronounced only for propagating magnons with nonzero wavevector and vanishes for the FMR (k=0) mode. That step is load-bearing because if dynamic GGG losses contribute at the uniform mode, the simulated 6.7x enhancement is an upper bound and the mechanism is misattributed. The k=0 argument does not obviously transfer to this experiment: at 2 K the GGG stray field varies from about 8 mT at the center to about 33 mT at the edges (Fig. 1b), so the mode excited by a uniform microwave field in this inhomogeneous bias is spatially nonuniform and contains finite-wavevector components. The paper's own micromagnetic simulations produce a spatially resolved magnetization response rather than a plane-wave k=0 mode. Moreover, Ref. [61] is an unreviewed companion preprint, and no independent check is provided that dynamic GGG-mediated damping of the quasi-uniform mode is negligible. Without such a check, the single-point factor-6.7 match (Fig. 3c) could be the result of overestimating the static-field inhomogeneity while an unmodeled dynamic contribution fills the experimental deficit.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports FMR spectroscopy on a 97-nm YIG film on a 500-µm GGG substrate from 300 K down to 30 mK. It attributes the strong low-temperature broadening of the FMR linewidth to the inhomogeneous static stray field produced by the partially magnetized paramagnetic GGG substrate. The authors compute this stray field with a finite-element solver using measured GGG magnetization curves, feed it as a static bias into micromagnetic simulations and into a semianalytical independent-resonator model, and compare the simulated linewidth with experiment at 300, 20, 8, and 2 K. The simulation reproduces the 300-K Gilbert behavior and produces a linewidth enhancement factor of up to 6.7 at 2 K and 250 mT, which they compare with the measured factor of about 6.6. They also report a deviation from Gilbert damping above 18 GHz and discuss implications for quantum magnonics, including substrate engineering. The central claim is that the GGG stray-field inhomogeneity is the dominant mechanism behind the tenfold linewidth increase at millikelvin temperatures, with additional unresolved processes accounting for part of the 2-K data.","tokens_in":14690,"tokens_out":8402,"duration_ms":87776,"significance":"If the central claim is established, the result is significant for quantum magnonics because it separates an extrinsic, geometry-dependent substrate contribution from intrinsic YIG damping, and it provides a falsifiable prediction (the 6.7-fold enhancement and its dependence on MGGG) that can guide substrate engineering. The paper's methodology has clear strengths: the low-temperature linewidth is not used to fit model constants; the GGG magnetization is measured and the stray field is computed independently; and the 300-K micromagnetic simulation matches experiment, validating the pipeline. The semianalytical and numerical results agree with each other. However, the dominance claim depends on two assumptions that need quantitative support: the exclusion of dynamic YIG-GGG coupling through a companion preprint, and the independent-resonator treatment of the collective FMR response. The recognized 2-K discrepancy and the unexplained >18-GHz behavior also mean the claim is currently supported by a small subset of the data.","major_comments":[{"comment":"The exclusion of the dynamic YIG-GGG dipolar/EPR channel is load-bearing and is currently carried by the companion preprint Ref. [61]. The text asserts that this mechanism \"is pronounced for propagating magnons with nonzero wavenumbers k ≠ 0 and should vanish for FMR,\" but the FMR mode studied here is not an ideal k = 0 plane wave: at 2 K the computed GGG stray field changes from 8 mT in the center to 33 mT at the edges (Fig. 1b), so the driven response is spatially inhomogeneous and contains finite-wavevector components. Because Ref. [61] is an unreviewed preprint and no quantitative estimate of EPR-mediated damping of the quasi-uniform mode is supplied, the single-point factor-6.7 agreement in Fig. 3(c) could be accidental. Please add an independent check (for example, a finite-k dynamic susceptibility estimate or a simulation that includes GGG dynamics) or reduce the claim to \"major contribution\" rather than \"dominant mechanism.\"","section":"Sec. III (mechanism discussion around Fig. 2)"},{"comment":"The semianalytical model treats the film as independent local resonators and justifies this by the short spin-wave mean free path. This argument addresses propagation of long-wavelength magnons but not the collective nature of the FMR response in an inhomogeneous static field: neighboring regions can lock through exchange and dipolar fields, which would partially average the local-field distribution. The agreement of Eq. (7) with the micromagnetic simulations in Fig. 3(a) is reassuring, but both approaches use the same static-field input and the comparison covers only three low-temperature datasets. A direct test would be to vary the exchange stiffness in the micromagnetic simulations (including the zero-exchange limit) and show that the FWHM is unchanged, and to state how the artificially increased damping near the boundaries affects the excitation region.","section":"Sec. II-C, Eq. (7)"},{"comment":"The data in Fig. 3(b) show that at 2 K the simulated linewidth saturates at 2.42 mT while the measured linewidth reaches 4 mT, and above 18 GHz the measured linewidth is below the simulated values with an oscillatory trend. The text acknowledges these deviations and labels them as requiring further investigation, but the abstract and conclusion nevertheless assert that the GGG stray field is the dominant mechanism. The dominance claim should be backed by a quantitative decomposition of the 4 mT linewidth into the computed 2.42 mT stray-field part and the residual part, and by a discussion of whether the high-frequency deviations invalidate the Gilbert fits from which the experimental enhancement factor of about 6.6 is derived.","section":"Sec. III, Fig. 3(b)"},{"comment":"The model input for the intrinsic YIG damping is the room-temperature linewidth ΔB ≈ 0.06 + 0.006 f (mT), which is used at all temperatures. If the intrinsic damping increases at low temperatures—as might be expected from rare-earth impurity relaxation (Refs. [39], [56]–[60])—then the simulated 6.7-fold enhancement is an upper bound on the GGG contribution, and part of the observed low-temperature increase could be misattributed. Please state explicitly whether any temperature dependence of the intrinsic YIG damping was considered, and if not, justify the assumption or test its sensitivity.","section":"Sec. II-B and II-C"}],"minor_comments":[{"comment":"The reference-field subtraction is described verbally, but the magnitude of the offset (15–40 mT) and its possible effect on the extracted linewidth are not discussed; a short sensitivity statement would help.","section":"Sec. II-A"},{"comment":"The connection between P(f) ∼ f·Im[∫ dt (m·b)] in Eq. (1) and P(f) ∼ f²·Im[∑ mz exp(2π i f tj)Δt] in Eq. (2) is not obvious; the sign of the exponential and the origin of the extra factor of f should be stated explicitly.","section":"Eqs. (1) and (2)"},{"comment":"The split-Lorentzian fitting model is mentioned but its functional form is not given; a reference is cited, but the exact model used for the FWHM extraction should be specified.","section":"Sec. II-A"},{"comment":"There are a few typographical and notation artifacts, including \"V oronov\" in the author list and \"ByGGG\" in the text, which should be written as B_y^GGG or similar.","section":"Fig. 1 and Sec. III"}],"recommendation":"major_revision","confidential_remarks":"The load-bearing exclusion of dynamic YIG-GGG coupling rests on Ref. [61], a companion preprint that appears to be from the same group and is not peer-reviewed. The authors should be asked to supply an independent estimate or to weaken the dominance claim. I would also encourage the authors to make the measured linewidth data and simulation output available, rather than only \"upon reasonable request,\" given that the paper's central quantitative claim depends on those data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe thing to know about this paper is that it does a real, quantitative job on a known nuisance: the paramagnetic GGG substrate's stray field broadens YIG FMR linewidth at millikelvin temperatures. The prior papers from this group (Refs. 42 and 43) established the field's existence, magnitude, and inhomogeneity; this one adds the linewidth consequence and a concrete 6.7x enhancement factor. That's a legitimate extension, not a new mechanism.\n\nThe 300 K validation is the strongest part: micromagnetic simulations reproduce measured linewidths at room temperature, and the semianalytical model uses the room-temperature damping as input, so the low-T enhancement isn't fit to the data. The agreement at 20 K and 8 K within scatter, and the honest reporting of the 2 K underprediction (simulated 2.42 mT vs measured up to 4 mT) and the above-18 GHz non-Gilbert behavior, all speak well of the authors' care.\n\nThe soft spot is exactly where the stress test points. The claim that the stray field is the dominant mechanism rests on dismissing dynamic dipolar coupling between YIG and GGG as negligible for the FMR mode, citing companion preprint [61] whose k=0 argument may not transfer here. The stray field varies by 8–33 mT across the film, so the mode excited by a uniform microwave field is not a plane-wave k=0 mode; it contains finite-wavevector components. That makes the companion preprint's conclusion, even if correct for a genuinely uniform mode, not obviously applicable. Since [61] is unreviewed and no independent estimate is provided, the 'dominant' language is stronger than the evidence supports. The 2 K shortfall reinforces this: at least one additional broadening channel is active, so the simulated 6.7x is likely an upper bound on the static-field contribution.\n\nAlso worth flagging: no data or code release, and the simulated linewidths have no error bars. That matters for a paper whose central number is a simulation–experiment comparison.\n\nNone of this is fatal. The static stray-field broadening is demonstrably real and important. The practical conclusion—that microstructuring the YIG or resizing the substrate reduces low-temperature broadening—survives even if the dynamic channel is non-negligible, because the gradient is the controllable part.\n\nThis deserves peer review. A good referee should push for a direct estimate of the dynamic coupling for the nonuniform mode, error bars on the simulated linewidths, and a softened 'dominant' claim given the unexplained 2 K and high-frequency residuals. I'd take it with those revisions.","headline":"A quantitative step forward on GGG stray-field broadening, but the 'dominant mechanism' claim overreaches because the dynamic coupling channel is dismissed via an unreviewed companion preprint.","tokens_in":15255,"tokens_out":2403,"would_cite":true,"duration_ms":23190,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["76.50.+g","75.70.-i"],"model":"deepseek-v4-flash","headline":"At millikelvin temperatures, the tenfold broadening of a YIG film's magnetic resonance is traced to stray fields from its GGG substrate.","keywords":["ferromagnetic resonance","YIG film","GGG substrate","millikelvin temperatures","magnetic damping","stray field","quantum magnonics","Gilbert damping"],"falsifier":"Take the same 97 nm YIG film, transfer it to a diamagnetic substrate (or remove the GGG entirely) and measure FMR at 2 K and 250 mT: if the linewidth stays near its room-temperature value, the stray-field mechanism is confirmed; if it still broadens by several times, an intrinsic low-temperature loss mechanism is at play.","tokens_in":14175,"feed_emoji":"🧲","tokens_out":8206,"duration_ms":75821,"temperature":0.7,"pith_summary":"At millikelvin temperatures, the magnetic resonance of a thin yttrium iron garnet (YIG) film—the material with the longest magnon lifetimes—broadens by up to a factor of ten, which directly limits its use in quantum magnonics. The authors argue that most of this broadening is not intrinsic to the YIG film but comes from the gadolinium gallium garnet (GGG) substrate underneath it: at low temperatures the paramagnetic GGG becomes partially magnetized and produces a non-uniform stray field that weakens the internal bias field of the film differently at different positions. Because the measured resonance is the sum of these locally shifted resonances, the line appears broadened and asymmetric. Measurements down to 30 mK, together with simulations and a semi-analytical model, yield a predicted stray-field-driven broadening of up to 6.7 times, close to the measured 6.6-fold increase at 2 K and 250 mT. If this is right, the extra damping can be engineered away by shaping the substrate or replacing GGG with a diamagnetic alternative.","feed_headline":"Stray field from GGG substrate inflates YIG linewidth tenfold","feed_subtitle":"The paramagnetic substrate's nonuniform field, not YIG itself, dominates the low-temperature broadening.","key_machinery":"The load-bearing mechanism is the quasi-static, coordinate-dependent stray field of the partially magnetized GGG substrate, expressed through the mutual demagnetization tensor $N_{yy}(y)$ of the YIG/GGG bilayer. The local resonance field at position $y$ is $B_{\\mathrm{loc}}(y) = B_0 - \\mu_0 M_{\\mathrm{GGG}} N_{yy}(y)$, and the local FMR frequency follows the standard in-plane field-versus-frequency relation for a thin film. The measured linewidth is then obtained by adding up independent local Lorentzian resonances over the sample width (Eq. 7), so the collective mode is treated as a sum of uncoupled resonators sampling different fields. The model's validity depends on the spin-wave mean free path being much shorter than the 5 mm sample, so that cooperative effects average nothing out.","core_discovery":"The paper's central claim is that the dominant physical mechanism for the observed tenfold increase in ferromagnetic resonance (FMR) linewidth in YIG/GGG at millikelvin temperatures is the non-uniform bias magnetic field generated by the partially magnetized paramagnetic GGG substrate. At 2 K and 250 mT the GGG-induced stray field varies from about 8 mT at the film center to 33 mT at the edges, opposing the external field. Simulating the FMR absorption in this spatially varying static field—and fitting the synthetic spectra with the same split-Lorentzian procedure used for the experiment—produces a linewidth increase of up to 6.7 times relative to 300 K, in line with the experimental factor of about 6.6. The paper also reports that the effective Gilbert damping and inhomogeneous broadening extracted from fits up to 18 GHz both grow with decreasing temperature (up to 13-fold and 5-fold, respectively) and saturate below about 500 mK, tracking the saturation of GGG magnetization. Above 18 GHz the measured linewidth departs from the viscous Gilbert model, which the authors attribute to the field-dependent magnetization of GGG and leave as an open problem.","pith_inferences":["The same stray-field mechanism should broaden the magnetic resonance of any ferromagnetic film grown on a paramagnetic substrate at low temperature, so the result generalizes beyond YIG/GGG to other garnet and oxide heterostructures.","If the independent-resonator model is right, the measured effective linewidth should shrink as the lateral size of the YIG film is reduced toward the stray-field gradient scale; this could be tested directly by measuring microstructured films of different widths.","The non-Gilbert behavior above 18 GHz may reflect relaxation channels the quasi-static model does not include, such as field-dependent coupling to GGG spin dynamics; identifying those channels is a natural next step."],"forward_implications":["At low fields and low frequencies, the dominant 'damping' seen in YIG/GGG at millikelvin temperatures is an artifact of a non-uniform static field, not a loss in YIG itself; true magnon lifetimes can be recovered by removing that field variation.","Microstructuring the YIG film or shaping the GGG substrate to flatten the stray-field gradient should substantially narrow the measured FMR linewidth at cryogenic temperatures.","The only complete fix is a diamagnetic substrate with a lattice match close to YIG, such as YSAG spacers or other garnets; YAG alone suffers from lattice mismatch.","Gilbert damping parameters extracted below about 18 GHz at low temperatures are strongly contaminated by stray-field broadening, so reported $\\alpha_{\\mathrm{eff}}$ values in this regime are upper bounds on the intrinsic damping.","The saturation of the linewidth increase below about 500 mK is a signature of GGG magnetization saturation, so further cooling alone will not reduce the broadening."],"supporting_citations":[{"why":"Supplies the temperature-dependent GGG magnetization curves, the stray-field profile, and the YIG anisotropy fields used as inputs for both numerical and analytical linewidth calculations.","marker":"[43]"},{"why":"Establishes the earlier finding that magnetized paramagnetic GGG creates a stray field that shifts the YIG FMR field, the effect this paper develops into a linewidth-broadening mechanism.","marker":"[42]"},{"why":"Shows that dipolar coupling to the GGG substrate's EPR is pronounced only for propagating magnons with nonzero wavenumber and vanishes for the uniform FMR mode, which justifies treating the stray field as static.","marker":"[61]"},{"why":"Provides the finite-difference micromagnetic solver used to compute the FMR absorption spectra of the YIG film in the static GGG stray field.","marker":"[48]"},{"why":"Supplies the finite-element/boundary-element magnetostatic solver used to compute the GGG substrate's stray field at the YIG position.","marker":"[49]"},{"why":"Gives the demagnetizing-field expression for a nonuniform rectangular prism, on which the analytical mutual demagnetization tensor $N_{yy}(y)$ in Eqs. (5)-(6) is based.","marker":"[52]"},{"why":"Gives the Lorentzian absorption formula $P \\sim \\mathrm{Im}[(\\omega_{\\mathrm{loc}} + i\\Gamma - \\omega_e)^{-1}]$ used to sum independent local resonators.","marker":"[54]"}],"fun_headline_variants":["GGG stray field, not YIG, widens FMR line tenfold at mK","Substrate's stray field drives tenfold YIG damping boost","Millikelvin damping blowup traced to GGG's nonuniform field","Paramagnetic GGG, not YIG, causes tenfold linewidth leap","At mK, GGG's field, not YIG, inflates linewidth 10x"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes the YIG film's many local regions precess independently enough that the measured resonance is merely the sum of locally shifted resonances, and that the GGG substrate's dynamic magnetic response does not drain energy from the uniform precession mode.","fun_headline_variants_meta":{"raw":{"variants":["GGG stray field, not YIG, widens FMR line tenfold at mK","Substrate's stray field drives tenfold YIG damping boost","Millikelvin damping blowup traced to GGG's nonuniform field","Paramagnetic GGG, not YIG, causes tenfold linewidth leap","At mK, GGG's field, not YIG, inflates linewidth 10x"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000239,"raw_usage":{"total_tokens":1560,"prompt_tokens":1040,"completion_tokens":520,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":656,"completion_tokens_details":{"reasoning_tokens":428}},"tokens_in":656,"tokens_out":520,"duration_ms":4533,"temperature":1.0,"reasoning_tokens":428,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T23:04:13.044135+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the same 97 nm YIG film, transfer it to a diamagnetic substrate (or remove the GGG entirely) and measure FMR at 2 K and 250 mT: if the linewidth stays near its room-temperature value, the stray-field mechanism is confirmed; if it still broadens by several times, an intrinsic low-temperature loss mechanism is at play.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the temperature-dependent GGG magnetization curves, the stray-field profile, and the YIG anisotropy fields used as inputs for both numerical and analytical linewidth calculations."},{"cited_title":"& Suess, D","cited_arxiv_id":null,"evidence_quote":"Provides the finite-difference micromagnetic solver used to compute the FMR absorption spectra of the YIG film in the static GGG stray field."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the finite-element/boundary-element magnetostatic solver used to compute the GGG substrate's stray field at the YIG position."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the demagnetizing-field expression for a nonuniform rectangular prism, on which the analytical mutual demagnetization tensor $N_{yy}(y)$ in Eqs. (5)-(6) is based."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Lorentzian absorption formula $P \\sim \\mathrm{Im}[(\\omega_{\\mathrm{loc}} + i\\Gamma - \\omega_e)^{-1}]$ used to sum independent local resonators."}],"review_version":1}