{"id":"faef398a-cd80-4652-8936-883e607bade6","arxiv_id":"2411.14428","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Even a 10 mT magnetic field reduces non-local magnon signals in YIG by roughly 20%, and the influence of a weak in-plane uniaxial anisotropy, tuned by device orientation, governs the zero-field limit.","lead":"This paper measures how a small external magnetic field suppresses the flow of magnons, the magnetic wave quanta used in data signaling, in thin films of the insulating magnetic material YIG, and it shows the signal is only maximal at exactly zero field. The authors then use a simple magnetic model plus the orientation of the YIG crystal to explain a subtle in-plane magnetic anisotropy that becomes important for future field-free magnonic devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central field-sweep claim (V1f rises >20% from 10 mT to 150 µT) is not cleanly separated from the field-driven rotation of M relative to the fixed Pt spin polarization, so the attribution to magnon diffusion length is not yet established.","rationale":"The paper's headline result is that the known high-field suppression of magnon diffusion length continues down to zero field, evidenced by a >20% increase in normalized V1f between 10 mT and 150 µT. But V1f is not a bare transport observable: Equation (2) makes injection/detection efficiency a function of σ·M, and at zero field M is not aligned with B but with the easy axis. In the low-field window the magnetization rotates, so the efficiency term has a field dependence of its own. The V2f field-independence check is not a sufficient control for this window: V2f is a detector-local signal, and its constancy is asserted for fields up to roughly 50 mT, whereas the relevant deviations appear below about 1 mT (Fig. 2). Without an explicit accounting for η(B) in Fig. 1(c), the central claim is underdetermined. The monodomain concern raised by the reader is related but different: even a perfect monodomain with one well-defined M still has this confound, because the single macrospin rotates with field. The SW fit itself is impressive independent support for the anisotropy parameters and for the V2f-as-local-magnetometer idea, but it does not by itself validate the transport attribution. A decisive check is a field sweep along the easy axis, where η is constant by symmetry, or, failing that, a published decomposition of V1f(B) into the SW-predicted η(B) and the residual transport factor. Pending that, CONDITIONAL is appropriate.","tokens_in":12094,"tokens_out":5231,"duration_ms":52784,"concrete_test":"Perform a field sweep of V1f from 10 mT to 150 µT with Bext applied exactly along the fitted easy axis (φB = φEA = 30°). In the SW model this keeps M parallel to the easy axis at all fields, so η = g(σ·M) is constant and any residual V1f increase directly measures the transport/diffusion-length contribution. If V1f still rises by more than 20%, the central claim is supported; if it is flat or much smaller, the observed low-field rise is dominated by magnetization reorientation rather than by magnon diffusion length.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"Equation (2) makes the injection/detection efficiency η = g(σ·M_YIG) depend on the angle between the fixed Pt spin polarization and the local magnetization. For any field direction not parallel to the easy axis (the device has φEA ≈ 30°, while the sweeps in Fig. 2 use φB = 0° and 90°), reducing Bext from 10 mT to 150 µT rotates M away from B and toward the easy axis. This changes η_inj and η_det over exactly the field window in which the headline 20% V1f increase is reported (Fig. 1c). The paper's only control for efficiency is the statement that V2f is field-independent up to Bext ≈ 50 mT. That statement applies to a range where M is already aligned with B; it does not cover the 150 µT–10 mT range, where Fig. 2 shows V2f itself deviating from cos(φB) and exhibiting jumps and hysteresis. The SW model fit to angular data is then used to account for the anisotropy, but no explicit decomposition of the measured V1f(B) into η(B) and a transport factor is shown. If the 20% rise is instead mostly the efficiency term, the central claim that magnon diffusion length suppression persists to zero field is not established; what remains is a real but re-interpreted anisotropy effect.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports non-local magnon transport measurements in Pt/YIG devices at in-plane fields from 0.5 T down to about 150 µT. The central claim is that the field-induced suppression of the non-local first-harmonic voltage V1f persists down to the zero-field limit, so that the highest signal amplitude is achieved only at zero field; a Stoner-Wohlfarth macrospin model with fitted in-plane uniaxial anisotropy (K = 16 ± 2 J/m³, φ_EA = 30° ± 5°) is used to explain the low-field angular deviations, hysteresis, and coercivity differences, and to predict device geometries for field-free operation. The paper also reports a strong increase of the anisotropy at low temperatures, with K and the effective anisotropy field growing by factors of about 15 and 10, respectively, below 100 K.","tokens_in":12416,"tokens_out":7828,"duration_ms":76451,"significance":"If the central claim is correct, the paper establishes a practical constraint for diffusive magnonic devices: even a modest 10 mT field attenuates the non-local signal by about 20%, and the in-plane uniaxial anisotropy controls the zero-field response. The measurements are careful, and the macrospin model reproduces the striking angular jumps and hysteretic features of both harmonic signals. The proposed use of the second-harmonic signal as an embedded local magnetometer is creative and potentially useful. However, the paper does not quantitatively separate the field dependence of the injection/detection efficiency from that of the magnon diffusion length, and the uniqueness of the fitted anisotropy parameters is not demonstrated. The significance of the headline claim depends on this separation, so the paper needs substantial revision before its conclusions can be accepted.","major_comments":[{"comment":"The central claim that the >20% rise in V1f between 10 mT and 150 µT reflects the field dependence of the magnon diffusion length is not separated from the field-dependent injection/detection efficiency. Equation (2) makes η_inj,det = g(σ·M_YIG), and because M_YIG is not aligned with the field in this window (as the paper itself shows through the angular jumps and hysteresis in Fig. 2), both η_inj and η_det change as B is reduced. The stated control—V2f field independence up to about 50 mT—does not cover the 150 µT–10 mT range, where V2f itself is strongly field dependent (Fig. 2(c,d)). To support the headline claim, the authors should provide a decomposition of V1f(B) into an efficiency factor (for example, computed from the simultaneously measured V2f or from φ_YIG(B) obtained with the SW model) and a transport factor, or measure V1f(B) along the easy axis where Δφ_EA = 0.","section":"Eq. (2) and Fig. 1(c)"},{"comment":"The fitted values K = 16 ± 2 J/m³ and φ_EA = 30° ± 5° are not shown to be uniquely determined. The fitting procedure uses K, φ_EA, the prefactor g in Eq. (2), and the sample-mounting angle correction described in the Methods as adjustable inputs; no residuals, parameter correlations, or confidence regions are shown. Because these same parameters are then used to generate the device configurations in Fig. 4, the agreement for the measured device is an interpolation rather than an independent validation of the model. A parameter sensitivity analysis and, ideally, measurements on devices with different twist angles are needed before the Fig. 4 design rules can be considered predictive.","section":"Fig. 3 and Methods (SW simulation)"},{"comment":"The claim that V2f measured at the detector represents the magnetization across the entire 1 µm transport channel rests on the monodomain assumption supported only by a citation to Ref. 27. The paper provides no domain imaging or local magnetization measurement for this particular film, and the Pt wires themselves could modify the local magnetic state through strain or Joule heating. If the magnetization under the injector and the detector differ, the product η_inj η_det cannot be factorized from V2f, and the interpretation of low-field V1f as transport-dominated is not justified. The authors should either provide direct evidence for monodomain behavior over the device footprint or discuss how their conclusions would change if the magnetization varies along the channel.","section":"Paragraph after Eq. (2)"}],"minor_comments":[{"comment":"The caption contains the sentence \"Adjust main text for the experimental data point on Panel (e)\", which appears to be a leftover editing note; it should be removed and the experimental data point should be described in the main text.","section":"Fig. 4(e) caption"},{"comment":"The crystallographic directions \"[1126]\" and \"[1160]\" are written without overbars; the standard notation for negative indices (e.g., [112̄] and [11̄0] or an equivalent explicit convention) should be used to avoid ambiguity.","section":"Eq. (1) and crystallographic notation"},{"comment":"The statement that \"the equilibrium position is identified when both derivatives are positive\" should be corrected to \"the first derivative vanishes and the second derivative is positive\" for an energy minimum.","section":"Methods, SW simulation"},{"comment":"The abstract says a 10 mT field attenuates the non-local spin voltage by about 20%, while the main text says the V1f signal increases \"by more than 20%\" when the field is reduced from 10 mT to 150 µT; these numbers should be reconciled and the normalization of the data in Fig. 1(c) stated explicitly.","section":"Abstract vs. main text"},{"comment":"Figure 5 uses a device with an injector–detector spacing of about 0.5 µm, whereas the main text and Figs. 1–4 use a device with about 1 µm spacing; the possible dependence of the extracted K and B_K on device geometry should be discussed.","section":"Fig. 5"},{"comment":"Ref. 27 is cited for the statement that magnetic domains in YIG films extend over hundreds of micrometers; since that reference concerns ultrathin YIG/Pt bilayers, the authors should verify that the cited domain size applies to their 80 nm PLD-grown film and to the specific device geometry used here.","section":"Ref. 27"}],"recommendation":"major_revision","confidential_remarks":"The paper fits the scope of the journal and the experimental work appears to be of good quality. The main risk is that the central claim reinterprets an anisotropy-driven efficiency change as a transport effect; this is a fixable analysis issue rather than an unfixable flaw. I would encourage the editor to seek a referee with expertise in non-local magnon transport to assess whether the proposed decomposition of V1f and V2f is sufficient after revision."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper shows that the non-local magnon signal in YIG keeps rising as the field drops, with a claimed >20% increase from 10 mT to 150 µT, and it pins the low-field angular and hysteretic behavior on an in-plane uniaxial anisotropy, quantified with a Stoner-Wohlfarth macrospin (K ≈ 16 J/m³, easy axis ≈ 30°). The core observation is plausible and the fits to the angular and hysteresis data are genuinely good. That is worth something.\n\nWhat is new: careful low-field measurements down to 150 µT, extraction of the uniaxial anisotropy from transport rather than static magnetization, and a temperature-dependent study showing a tenfold enhancement of the anisotropy field at low temperatures. The twist-angle device concept in Figure 4 is a nice design idea, though it is an extrapolation of the fitted model.\n\nThe soft spots are real but fixable. The main one is the stress-test concern: Equation (2) makes the injection/detection efficiency η depend on the angle between the fixed Pt spin polarization and the local magnetization. In the field window where the 20% rise is reported (10 mT→150 µT), M is rotating toward the easy axis, so η itself changes. The paper's control—that V2f is field-independent up to 50 mT—is not decisive for this range, because Figure 2 shows V2f deviating from cos(φB) and exhibiting jumps at 150 µT. The authors say they account for the anisotropy in the SW fit, but they never show an explicit decomposition of V1f(B) into η(B) and a transport factor. Without that, the central claim that the magnon diffusion length suppression persists to zero field is not fully established. Also, K and φEA are fit outputs, not independently calibrated; the monodomain assumption is reasonable but cited from large-domain films and not verified for this exact device footprint. The “>20%” reads from normalized data without error bars. Minor: the Figure 4 caption contains an editorial artifact, so the manuscript is not in final form.\n\nNone of this kills the paper. The data look self-consistent and the anisotropy interpretation is likely correct. But the key quantitative claim needs a cleaner separation of efficiency and transport, plus error bars and preferably an independent magnetization check. This is a solid candidate for peer review after revisions.\n\nWho is this for: anyone working on magnon transport, YIG devices, or field-free spintronics. I would bring it to the reading group and cite it if the decomposition gets sorted out.","headline":"Useful low-field magnon transport data with a credible anisotropy model, but the headline 20% effect is not cleanly separated from field-dependent injection/detection efficiency.","tokens_in":12986,"tokens_out":3008,"would_cite":true,"duration_ms":30163,"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":"The non-local magnon signal in yttrium iron garnet is maximal only at zero field, and a weak in-plane uniaxial anisotropy controls the low-field response.","keywords":["magnon transport","yttrium iron garnet","non-local spin voltage","magnon diffusion length","in-plane uniaxial anisotropy","Stoner-Wohlfarth model","spin Hall effect","field-free magnonics"],"falsifier":"Measure the first-harmonic voltage of the same device for fields below 150 µT, down to a few microtesla: if $V_{1f}$ saturates at a finite value before the field reaches zero, or if it turns over and decreases again as $B \\to 0$, the claim that the signal is maximal only at zero field is falsified. A complementary check is to image the domain structure of the 80 nm film at 150 µT: if multiple domains with different in-plane orientations coexist within the ~1 µm channel, the monodomain macrospin interpretation, and with it the fitted $K$ and $\\varphi_{EA}$, would not transfer across the device.","tokens_in":11786,"feed_emoji":"🧲","tokens_out":11364,"duration_ms":103127,"temperature":0.7,"pith_summary":"Diffusive magnon transport in yttrium iron garnet is normally measured under an external magnetic field, but this paper shows that the field itself is the enemy of signal: the non-local spin voltage rises monotonically as the field is lowered and is maximal only at zero field. Lowering the field from 10 mT to 150 µT increases the first-harmonic voltage by more than 20% in a 1 µm transport channel, so any nonzero field, however small, costs signal. A Stoner-Wohlfarth macrospin model — a single magnetization arrow whose equilibrium balances the Zeeman energy against a weak in-plane uniaxial anisotropy — reproduces the abrupt, hysteretic jumps observed in the angular sweeps and quantifies the anisotropy as $K = 16 \\pm 2$ J/m$^3$ with easy axis at $30^\\circ \\pm 5^\\circ$ from the $[11\\bar{2}]$ direction. The same model predicts how much zero-field signal survives for a given wire orientation, and the paper uses the second harmonic as an embedded local magnetometer to show that this anisotropy strengthens more than tenfold at cryogenic temperatures.","feed_headline":"Magnon signals peak only at zero magnetic field","feed_subtitle":"Even 10 mT cuts the spin signal by 20%; weak in-plane anisotropy governs the zero-field response.","key_machinery":"The load-bearing object is the Stoner-Wohlfarth macrospin model, which represents the whole YIG channel by a single magnetization direction $\\varphi_{YIG}$ and finds its equilibrium by minimizing the energy per unit volume $E/V_{\\mathrm{eff}} = -M_S B_{\\mathrm{ext}} \\cos(\\varphi_{YIG}-\\varphi_B) + K \\sin^2(\\varphi_{YIG}-\\varphi_{EA})$, a Zeeman term competing with a uniaxial in-plane anisotropy. Fitting this model to the measured angular scans yields the two parameters that carry the paper's argument: the anisotropy density $K$ and the easy-axis angle $\\varphi_{EA}$. The second essential mechanism is harmonic lock-in detection, which separates the electronically injected first-harmonic signal (transport through the whole channel) from the thermally generated second-harmonic signal (localized at the detector); since $V_{2f}$ is proportional to the projection of the local magnetization on the fixed platinum spin polarization, it functions as a built-in magnetometer. These pieces convert the abrupt low-field jumps and hysteresis from unexplained artifacts into a quantitative measure of YIG's in-plane anisotropy and a geometric rule for zero-field signal retention.","core_discovery":"The central discovery is that the magnetic-field penalty on diffusive magnon transport in YIG does not vanish at small fields: the non-local first-harmonic voltage grows asymptotically as the external field is reduced, reaching its maximum only in the zero-field limit, and even 10 mT costs about 20% of the signal in a 1 µm channel. The paper shows that a weak in-plane uniaxial anisotropy, normally neglected, controls the low-field response: a Stoner-Wohlfarth macrospin fit gives $K = 16 \\pm 2$ J/m$^3$ and an easy-axis direction $30^\\circ \\pm 5^\\circ$ from the $[11\\bar{2}]$ crystal axis, reproducing the abrupt, hysteretic jumps that appear in the angular dependence of $V_{1f}$ and $V_{2f}$ below about 1 mT. The fitted model also predicts that the zero-field remnant of the first-harmonic signal is set by the wire-to-easy-axis twist angle, scaling as $\\cos^2(\\Delta\\varphi_{EA})$, with parallel alignment preserving the full signal and perpendicular alignment quenching it. Finally, using the second harmonic as a local magnetization probe, the paper reports that the anisotropy energy and effective anisotropy field rise from about 2.7 J/m$^3$ and 20 µT at room temperature to about 42 J/m$^3$ and 220 µT below 100 K, a more than tenfold enhancement it ties to strain-related growth effects.","pith_inferences":["If the same mechanism is at work in other low-damping magnetic insulators, optimal signal at zero field is a generic design rule, and the roughly 20% loss per 10 mT becomes a benchmark for comparing materials.","A direct test of the monodomain assumption is sub-micron magnetic imaging of the film at 150 µT; a multi-domain channel would broaden or split the sharp $V_{2f}$ jumps that the single-macrospin fit currently explains.","Because $V_{1f}$ samples the whole channel while $V_{2f}$ is localized at the detector, a joint field- and angle-resolved analysis of the two harmonics could separate interface coupling from bulk magnon-diffusion contributions to the field dependence—a decomposition the paper leaves implicit.","The strain hypothesis predicts that growth parameters such as oxygen pressure or substrate mismatch shift $K$ and the easy-axis angle $\\varphi_{EA}$, so a growth-series study would turn the fitted anisotropy from a fitting parameter into a controlled design input."],"forward_implications":["Zero field becomes the preferred operating point for magnonic devices: it maximizes the non-local spin voltage and removes the need for an external magnet.","Geometric layout is a control knob: alignment between the platinum wire's spin polarization and the film's easy axis sets the zero-field remnant signal from full retention to zero, following a $\\cos^2(\\Delta\\varphi_{EA})$ law.","The second harmonic signal can be read as a local, device-embedded magnetometer, giving access to switching fields and anisotropy without sacrificing the transport measurement.","Cryogenic magnonics must account for a much stronger in-plane anisotropy: below 100 K the effective anisotropy field exceeds 200 µT, so anisotropy effects dominate the low-field response rather than vanishing."],"supporting_citations":[{"why":"It establishes the Pt/YIG non-local magnon transport platform and the diffusive transport regime that this paper operates in.","marker":"[19]"},{"why":"It documents the high-field reduction of magnon diffusion length that this paper extends down to the zero-field limit.","marker":"[20]"},{"why":"It provides the harmonic lock-in procedure that separates the first-harmonic magnon signal from the second-harmonic spin-Seebeck signal.","marker":"[25]"},{"why":"It supports the monodomain assumption over the device footprint and reports the same easy-axis orientation from spin magnetoresistance measurements.","marker":"[27]"},{"why":"It reports the same easy-axis orientation from magneto-optical Kerr measurements and links it to growth-induced strain in pulsed-laser-deposited films.","marker":"[28]"},{"why":"It defines the spin-mixing conductance at the Pt/YIG interface that enters the coupling-efficiency factor $g$.","marker":"[26]"},{"why":"It quantifies the magnon spin conductivity of ultrathin YIG, supporting the interpretation of the measured voltage as diffusive magnon transport.","marker":"[13]"},{"why":"It shows that at high fields the GGG substrate can contribute paramagnons, one of the reasons high-field operation is undesirable.","marker":"[21]"}],"fun_headline_variants":["Zero field is the only magnon signal peak","Even 10 mT cuts magnon signal by 20%","Weak in-plane anisotropy governs zero-field magnonics","Tenfold anisotropy boost at cryo enables zero-field magnonics"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central argument assumes the film's magnetization is uniform—a single magnetic domain—across the whole device, so the local direction at the detector wire represents the entire transport channel.","fun_headline_variants_meta":{"raw":{"variants":["Zero field is the only magnon signal peak","Even 10 mT cuts magnon signal by 20%","Weak in-plane anisotropy governs zero-field magnonics","Tenfold anisotropy boost at cryo enables zero-field magnonics"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000604,"raw_usage":{"total_tokens":2881,"prompt_tokens":1072,"completion_tokens":1809,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":688,"completion_tokens_details":{"reasoning_tokens":1743}},"tokens_in":688,"tokens_out":1809,"duration_ms":14771,"temperature":1.0,"reasoning_tokens":1743,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T15:10:24.672994+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the first-harmonic voltage of the same device for fields below 150 µT, down to a few microtesla: if $V_{1f}$ saturates at a finite value before the field reaches zero, or if it turns over and decreases again as $B \\to 0$, the claim that the signal is maximal only at zero field is falsified. A complementary check is to image the domain structure of the 80 nm film at 150 µT: if multiple domains with different in-plane orientations coexist within the ~1 µm channel, the monodomain macrospin interpretation, and with it the fitted $K$ and $\\varphi_{EA}$, would not transfer across the device.","supporting_citations":[],"review_version":1}