{"id":"6d740583-a471-4d9e-94e2-a02e192bc520","arxiv_id":"2502.02978","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Microstructuring YIG films so they sit only in the homogeneous region of the GGG substrate's stray field eliminates the asymmetric FMR linewidth broadening at cryogenic temperatures.","lead":"This paper shows that cutting a YIG magnetic film into a small square or stripe removes a cryogenic linewidth-broadening effect caused by the GGG substrate's stray field, making the ferromagnetic resonance peak symmetric again. The measurements also suggest that below 10 K the magnetic damping no longer follows the usual Gilbert linear-in-frequency behavior, for reasons not yet identified.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Non-Gilbert linewidth decrease above 18 GHz rests on unvalidated linewidth extraction; independent fitting and conversion checks are needed.","rationale":"The reader's weakest assumption identifies the split-Lorentzian fit as the fragile link, and I agree with that in part. The fit is indeed unvalidated and unaccompanied by uncertainties. However, the more precise load-bearing issue is the whole extraction pipeline: the lineshape model plus the frequency-to-field conversion, which depends on assumed YIG magnetization and GGG stray-field corrections at low temperature. A false non-monotonic linewidth could enter through either stage, and the two are entangled in the analysis. This makes the non-Gilbert claim less secure than the microstructuring result, which has independent support from the geometry dependence and FEMME simulations. I would keep the conditional verdict: the microstructuring demonstration is credible, but the non-Gilbert interpretation should be treated as preliminary until an independent extraction and conversion-robustness check is performed. This is not a rejection: the paper is honest about the unresolved mechanism, and the raw data appear capable of supporting the proposed re-analysis. If the check confirms the trend, the claim would be strongly supported; if not, the conclusion should be weakened to a report of an apparent non-Gilbert behavior pending further study or explained as an artifact.","tokens_in":7798,"tokens_out":6156,"duration_ms":113278,"concrete_test":"Re-analyze the raw subtracted S21 spectra for configurations (c) and (d) at T<10 K: extract ΔB with a standard symmetric Lorentzian fit and a model-independent estimator (e.g., numerical FWHM or integrated-area linewidth), without using the split-Lorentzian model; then repeat the Δf→ΔB conversion while varying M_s(T) and the GGG stray-field magnitude within ±10%. If the maximum near 18 GHz and the subsequent decrease survive all combinations, the non-Gilbert claim is supported; if the trend shifts or disappears, it is an extraction artifact.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The manuscript's most novel and broad claim is the non-Gilbert-like behavior below 10 K: a nonlinear increase of ΔB to a maximum near 18 GHz, followed by a decrease with frequency, asserted to be independent of the inhomogeneous GGG stray field. This claim rests entirely on ΔB values extracted with the split-Lorentzian fitting model and the subsequent Δf→ΔB conversion (Methodology). The fitting model is not validated against an independent lineshape, no uncertainty estimates are given, and the conversion depends on temperature-dependent YIG saturation magnetization and on the GGG stray-field correction, both of which are model inputs with no stated uncertainties. A small systematic error in these corrections at low temperature and high field could generate an apparent non-monotonic ΔB(f) even if the intrinsic damping is Gilbert-like. The paper itself states that the physical mechanisms are 'not yet clear' and 'requires further theoretical considerations', and the observed oscillations in ΔB versus frequency suggest scatter that is not quantified. If the decrease above 18 GHz is an extraction artifact, the central claim beyond the microstructuring demonstration collapses; if it survives independent extraction, the claim is substantially strengthened. The asymmetric-broadening elimination itself is better supported by the geometry-dependent spectra and FEMME simulations and is not the load-bearing weakness.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports temperature-dependent ferromagnetic resonance measurements on 150 nm-thick YIG films grown on GGG, comparing four geometries: a 4x4 mm plane film, a 4x0.5 mm stripe parallel to the bias field, the same stripe orthogonal to the bias field, and a 0.5x0.5 mm square. The central claim is that microstructuring the YIG film so that the probed region lies in the homogeneous part of the GGG substrate stray field eliminates the asymmetric FMR linewidth broadening observed at cryogenic temperatures. A second, broader claim is that below 10 K the extracted linewidth Delta-B shows non-Gilbert behavior: it increases nonlinearly to a maximum near 18 GHz and then decreases with increasing frequency, independent of the inhomogeneous GGG stray field. The authors support the first claim with spectra at 293 K and 2 K together with FEMME simulations of the GGG stray field, and support the second claim with Delta-B versus frequency curves extracted using a split-Lorentzian fitting model introduced in Section II.","tokens_in":8074,"tokens_out":4580,"duration_ms":43241,"significance":"If the geometry result holds, the paper provides a practical and simple route to suppress substrate-induced FMR broadening in YIG-GGG, which is directly relevant to cryogenic magnonic and hybrid quantum devices. The same-wafer control, the four-configuration comparison, and the FEMME simulation support are genuine strengths, and the reported elimination of asymmetric broadening is visually clear and internally consistent. The broader non-Gilbert claim below 10 K, however, rests entirely on linewidth values extracted with an unvalidated fitting model and on model-dependent conversion corrections, with no uncertainty estimates. Because this claim is presented in the abstract and conclusion as a main result, the paper's overall significance is contingent on whether the linewidth extraction and conversion survive scrutiny. The paper is honest in stating that the physical mechanisms are not yet clear, but that admission does not replace the need for quantitative error analysis.","major_comments":[{"comment":"The non-Gilbert claim below 10 K — a nonlinear increase in ΔB to a maximum near 18 GHz and then a decrease with frequency — rests entirely on ΔB values obtained from the split-Lorentzian fit. No confidence intervals, fit residuals, or repeatability estimates are reported for any extracted linewidth. Because the 2–10 K regime is also where the GGG background is strongest, the reader cannot distinguish a physical linewidth drop from an extraction artifact. Please provide uncertainty estimates for every ΔB point and validate the split-Lorentzian extraction against at least one independent analysis, for example a direct fit to the raw absorption with a different lineshape or a comparison with a standard Lorentzian fit in the symmetric geometries.","section":"Section III, Fig. 2"},{"comment":"The conversion from the measured frequency linewidth Δf to the magnetic linewidth ΔB depends on the temperature-dependent saturation magnetization of YIG and on the antiparallel GGG stray-field correction, both supplied as model inputs without stated uncertainties. These corrections are largest at low temperature and high field, precisely the regime in which the claimed decrease of ΔB above 18 GHz appears. Please provide a sensitivity analysis showing how uncertainties in M_s(T) and in the GGG stray-field magnitude propagate into ΔB(f); if the non-monotonic behavior survives all plausible variations, the claim would be substantially strengthened.","section":"Section II, Methodology"},{"comment":"The linewidth data contain unquantified oscillations that the paper attributes to thickness inhomogeneities and strain effects, but no supporting evidence is provided for that attribution. The 'steady decrease' above 18 GHz is claimed against this oscillating background, so an apparent drop could partly be scatter. Please quantify the scatter, for example through standard deviations of repeated measurements or a smoothness metric, and show that the decrease after the maximum exceeds this scatter at each temperature.","section":"Section III, Figs. 2(a)–(d)"},{"comment":"The statement that the non-Gilbert behavior is 'independent of the inhomogeneous GGG stray field' is strictly supported by the geometry comparison, but the stronger conclusion in Section IV that 'the physical reason behind the non-linear nature is not related to the substrate-induced magnetic field' goes beyond the data. The microstructured samples remain on the full GGG substrate and still experience the homogeneous part of the stray field and other substrate-induced damping contributions. Please rephrase the conclusion to the supported statement or provide an additional control, such as YIG transferred to a nonmagnetic substrate, to separate homogeneous GGG effects from intrinsic YIG behavior.","section":"Abstract and Section IV"}],"minor_comments":[{"comment":"The text contains typographical spacing issues, for example 'B0 up to 1 .3 T' and 'V oronov' in the author list; please correct these throughout.","section":"Section II, first paragraph"},{"comment":"The colored boxes in panels (e)–(f) and the curve colors in panels (g)–(j) are described in the text but not identified in the caption; please add a legend or explicit color key.","section":"Figure 1 caption"},{"comment":"The frequency intervals 5–10, 10–15, 15–20, and 30–35 GHz are mentioned in the text but not indicated in the figure; consider adding vertical guides or a table with band edges to improve readability.","section":"Figure 3"},{"comment":"Reference [19] lists 'P.C.S.' as an author, which appears to be an incomplete or incorrect set of initials; please check and correct the citation.","section":"Reference [19]"},{"comment":"The sentence beginning 'Continuous advances in nanofabrication technology... focused attention' has a subject–verb agreement problem; please revise for clarity.","section":"Introduction, first paragraph"},{"comment":"The list of observed FMR frequencies is useful but would be more informative with measurement uncertainties, especially because the central claim involves differences in linewidth rather than in resonance position.","section":"Figure 1(g)–(j) caption"}],"recommendation":"major_revision","confidential_remarks":"To the editor: this manuscript fits the journal's scope, and the core geometry-based demonstration of removing asymmetric stray-field broadening is likely correct and practically useful. The main risk is the unvalidated and unquantified linewidth extraction that underlies the non-Gilbert claim. The requested revisions — uncertainty propagation, an independent lineshape check, and a sensitivity analysis of the Δf-to-ΔB conversion — are local to the analysis rather than requiring new physics, so a careful revision could make the paper acceptable. I do not recommend rejection at this stage."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: the microstructuring result is real, and the non-Gilbert claim is not yet proven. The four-geometry comparison is a clean, controlled experiment, and the spectra make the case that asymmetric broadening comes from the GGG stray-field inhomogeneity. Shrinking the YIG to the homogeneous region removes that broadening. That is a simple and useful trick for cryogenic YIG-GGG work, and the FEMME simulations are consistent with the story.\n\nWhat is new: the demonstration that geometry alone can suppress the substrate-induced broadened linewidth, at least for the asymmetric part. The observation of a non-Gilbert linewidth decrease above about 18 GHz below 10 K is also new, but it is presented as 'evidence' while the paper itself says the mechanism is unclear.\n\nThe soft spots are concentrated in the non-Gilbert claim. Every linewidth comes from the split-Lorentzian fit, and there are no error bars, no residuals, and no comparison with an independent extraction method. The conversion from Δf to ΔB depends on temperature-dependent saturation magnetization and the GGG stray-field correction, and both enter without any stated uncertainty. A small systematic error in those corrections could turn an intrinsically Gilbert-like damping into an apparent non-monotonic ΔB(f). The oscillations in the linewidth are attributed to thickness inhomogeneities and strain, which is plausible but not tested. The paper is honest about the lack of mechanism, but honesty does not substitute for validation.\n\nThat said, the central geometry comparison is not at risk. The asymmetric broadening disappears exactly when the sample is in the homogeneous stray field, and that conclusion does not depend on the details of the fitting model. The non-Gilbert behavior is an add-on, and it needs heavier lifting.\n\nWho this is for: anyone doing cryogenic FMR or quantum magnonics on YIG-GGG. The microstructuring trick is practically useful, and the open question about linewidth behavior at low temperatures is worth having on the table.\n\nRecommendation: send it to peer review. The microstructuring result deserves publication, and the non-Gilbert claim needs a referee to push for an uncertainty budget or an independent check. I would not let the paper through without that, but it is a legitimate candidate for review.","headline":"A clean demonstration that microstructuring YIG into the homogeneous GGG stray-field region removes asymmetric FMR broadening, plus an intriguing but unproven non-Gilbert linewidth claim that needs an independent extraction check.","tokens_in":8557,"tokens_out":2663,"would_cite":true,"duration_ms":22537,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"By patterning the YIG film into a small square or stripe, the paper removes the GGG substrate's inhomogeneous stray field from FMR linewidth measurements, and reports a non-Gilbert linewidth peak near 18 GHz below 10 K.","keywords":["YIG-GGG","ferromagnetic resonance","FMR linewidth","GGG stray field","microstructuring","Gilbert damping","cryogenic magnonics"],"falsifier":"Remove the GGG substrate from a microstructured YIG sample, or transfer the YIG onto a nonmagnetic substrate, and repeat the 2 to 10 K FMR sweep up to 40 GHz; if the linewidth no longer peaks near 18 GHz and then decreases with frequency, the claimed non-Gilbert behavior depends on the substrate after all or is an artifact of the split-Lorentzian fit rather than an intrinsic YIG property.","tokens_in":1825,"feed_emoji":"🧲","tokens_out":2415,"duration_ms":67046,"temperature":0.7,"pith_summary":"This paper reports temperature-dependent ferromagnetic resonance measurements from 293 K down to 2 K, up to 40 GHz, on 150 nm yttrium iron garnet (YIG) films on gadolinium gallium garnet (GGG) substrates, comparing a plain film with microstructured stripes and a square. The authors show that the asymmetric FMR linewidth broadening seen at cryogenic temperatures originates from the inhomogeneous stray field of the partially magnetized GGG substrate at the film edges, and that patterning the YIG so the probed region sits only in the homogeneous part of the stray field removes this broadening. Below 10 K, they find the linewidth no longer follows the usual Gilbert linear-in-frequency behavior: it rises nonlinearly to a maximum near 18 GHz and then decreases with frequency, even when the inhomogeneous stray field is absent. If correct, the result gives a simple design rule for low-loss YIG-GGG devices at millikelvin temperatures and points to an intrinsic, substrate-independent damping mechanism at cryogenic temperatures.","feed_headline":"Microstructuring eliminates substrate-induced FMR linewidth broadening","feed_subtitle":"A small patterned YIG film avoids the GGG substrate's uneven stray field, exposing a damping peak near 18 GHz.","key_machinery":"The load-bearing experimental device is the set of four sample geometries: the plain 4 x 4 mm film, a 0.5 x 4 mm stripe parallel to $B_0$, the same stripe orthogonal to $B_0$, and a 0.5 x 0.5 mm square. Because the GGG stray field is strongest and most nonuniform near the edges of the substrate, the first two geometries expose the YIG to the inhomogeneous field while the last two place the probed YIG only in the homogeneous central region; comparing the four isolates the stray-field contribution. The associated analysis machinery is the split-Lorentzian lineshape, which fits the left and right sides of the asymmetric FMR peak separately, together with a cubic spline background subtraction and an ellipticity correction converting the frequency linewidth into $\\Delta B$. This combination is what lets the paper extract a linewidth from asymmetric spectra and then compare geometries.","core_discovery":"The paper's central claim is that the cryogenic FMR linewidth broadening in YIG-GGG is dominated by the spatially inhomogeneous magnetic stray field of the GGG substrate, and that this parasitic contribution can be eliminated by microstructuring the YIG film into a 0.5 mm square or a 0.5 x 4 mm stripe oriented so the measured YIG sits only in the homogeneous central region of the substrate field. In the plane film and in the stripe parallel to the bias field, the FMR peak broadens asymmetrically toward lower frequencies at 2 K; in the orthogonal stripe and the square, the peak is symmetric. Fitting the spectra with a split-Lorentzian model, the authors extract the magnetic linewidth $\\Delta B$ versus the FMR frequency $f_{\\mathrm{FMR}}$ for all geometries and find that below 10 K the linewidth increases nonlinearly with frequency up to about 18 GHz and then decreases, a non-Gilbert trend that persists when the inhomogeneous stray field is removed and therefore is not caused by the stray field inhomogeneity.","pith_inferences":["A natural extension is to the propagating-spin-wave regime: the same edge-dominated stray field that broadens FMR should suppress spin-wave transmission at millikelvin temperatures, and microstructuring may offer a path to low-loss magnon conduits.","The occurrence of the linewidth maximum near 18 GHz suggests a resonant or crossover mechanism tied to the GGG magnetization rather than a simple relaxation; a test would vary the GGG composition, thickness, or field orientation and track whether the peak frequency shifts.","Because the split-Lorentzian model is the sole extractor of the reported linewidths, an independent cross-check, such as a different fitting form or a direct time-domain damping measurement, would be needed to confirm that the decrease above 18 GHz is physical and not a fitting artifact.","If the non-Gilbert behavior survives such a check, cryogenic YIG damping would not be capturable by a single Gilbert alpha, which would matter for the design of magnon-based quantum memories."],"forward_implications":["YIG-GGG devices intended for millikelvin operation should pattern the YIG into islands smaller than the homogeneous region of the GGG stray field, avoiding the edge-dominated broadening without changing the magnetic material.","Reported cryogenic linewidths from unpatterned YIG-GGG samples include a substrate-geometry-dependent contribution, so future measurements should specify the lateral position and orientation of the film relative to the substrate edges.","The persistent non-monotonic linewidth below 10 K, with a maximum near 18 GHz, implies that the Gilbert model is incomplete for YIG-GGG at low temperatures and high fields; a quantitative damping parameter must account for this frequency dependence.","The reference-subtraction and split-Lorentzian extraction procedure can be applied to other thin-film magnetic systems on paramagnetic substrates to separate substrate artifacts from intrinsic damping."],"supporting_citations":[{"why":"Supplies the characterization of GGG magnetization and the numerical simulation of the substrate stray field that motivates and is compared with the microstructuring experiment.","marker":"[28]"},{"why":"Provides the split-Lorentzian fitting model used to extract linewidths, along with the prior report of stray-field-induced broadening that this work extends.","marker":"[30]"},{"why":"Supplies the cubic spline background subtraction used before fitting the FMR peaks.","marker":"[33]"},{"why":"Supplies the conversion from frequency linewidth to magnetic linewidth accounting for the ellipticity of the spin precession.","marker":"[34]"},{"why":"Supplies the low-loss liquid phase epitaxy YIG film growth used for all samples.","marker":"[27]"},{"why":"Documents the cryogenic damping increase in YIG films, framing the problem that the microstructuring approach addresses.","marker":"[29]"}],"fun_headline_variants":["Micro-patterned YIG eliminates GGG stray field broadening","Tiny YIG squares remove substrate-induced FMR linewidth blur","Structured YIG films reveal non-Gilbert damping at low T","Shaping YIG into stripes sharpens cryogenic FMR spectra"],"cache_read_input_tokens":10752,"weakest_assumption_plain":"The split-Lorentzian fitting model faithfully extracts the physical FMR linewidth from the asymmetric absorption spectra, so the non-monotonic frequency dependence below 10 K is real damping behavior rather than an artifact of the fit; the paper does not validate the model against an independent measurement or provide uncertainty estimates.","fun_headline_variants_meta":{"raw":{"variants":["Micro-patterned YIG eliminates GGG stray field broadening","Tiny YIG squares remove substrate-induced FMR linewidth blur","Structured YIG films reveal non-Gilbert damping at low T","Shaping YIG into stripes sharpens cryogenic FMR spectra"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000313,"raw_usage":{"total_tokens":1781,"prompt_tokens":953,"completion_tokens":828,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":569,"completion_tokens_details":{"reasoning_tokens":753}},"tokens_in":569,"tokens_out":828,"duration_ms":8032,"temperature":1.0,"reasoning_tokens":753,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-09T10:19:43.269441+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Remove the GGG substrate from a microstructured YIG sample, or transfer the YIG onto a nonmagnetic substrate, and repeat the 2 to 10 K FMR sweep up to 40 GHz; if the linewidth no longer peaks near 18 GHz and then decreases with frequency, the claimed non-Gilbert behavior depends on the substrate after all or is an artifact of the split-Lorentzian fit rather than an intrinsic YIG property.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the characterization of GGG magnetization and the numerical simulation of the substrate stray field that motivates and is compared with the microstructuring experiment."},{"cited_title":"Damping Enhancement in YIG at Millikelvin Temperatures due to GGG Substrate","cited_arxiv_id":"2412.02827","evidence_quote":"Provides the split-Lorentzian fitting model used to extract linewidths, along with the prior report of stray-field-induced broadening that this work extends."},{"cited_title":"Herrera-Gomez, D","cited_arxiv_id":null,"evidence_quote":"Supplies the cubic spline background subtraction used before fitting the FMR peaks."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the conversion from frequency linewidth to magnetic linewidth accounting for the ellipticity of the spin precession."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the low-loss liquid phase epitaxy YIG film growth used for all samples."},{"cited_title":"Kosen, A","cited_arxiv_id":null,"evidence_quote":"Documents the cryogenic damping increase in YIG films, framing the problem that the microstructuring approach addresses."}],"review_version":1}