{"id":"64d358a2-d877-4dad-9760-8fb5e0962d30","arxiv_id":"2411.18202","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A CoFeB/Ru/CoFeB synthetic antiferromagnet grown on LiNbO3 shows FMR damping, linewidth, dispersion, and group velocity comparable to SAFs on standard substrates, supporting a new acousto-magnonic platform.","lead":"This paper measures spin waves in a synthetic antiferromagnet grown on a piezoelectric lithium niobate substrate and finds its magnetic quality matches standard substrates. It suggests this material platform could couple magnons to surface acoustic waves for tunable, non-reciprocal devices.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. 5's phase-to-wavevector conversion assumes the V-shaped, reciprocal, k-independent-attenuation dispersion that Fig. 4 reports, so the extracted group velocities and the 'line with theory' comparison are not model-free; the central quantitative claim needs an independent dispersion check.","rationale":"The paper is best read as a materials-characterization study whose central claim is comparative and quantitative. The FMR linewidth, the time-of-flight group delay, and the phase-based dispersion are internally coherent, and the use of multiple antenna distances is a genuine consistency check. The most load-bearing weakness is that the phase-to-wavevector inversion in Eq. 5 assumes exactly the V-shaped, reciprocal, wavevector-independent-attenuation dispersion that the paper then reports as a measured result. The time-of-flight value vg ~ 4.2 km/s and the independently measured FMR linewidth provide cross-checks, so the work is not circular in an obvious or disqualifying way; the issue is that the quantitative dispersion and its field dependence, which are used to claim agreement with theory, are not model-free. This is addressable by a generalized fit allowing curvature and k-dependent damping. I agree with the reader that this is the weakest assumption, and I also agree that the absence of a same-run non-piezoelectric control makes the 'as good as standard substrates' comparison less secure. Neither issue, however, justifies rejecting the paper: the evidence strongly suggests the platform works, and the requested checks are feasible. Thus the conditional verdict stands unchanged.","tokens_in":7125,"tokens_out":9952,"duration_ms":102899,"concrete_test":"Fit the raw, time-gated field-derivative phase data with a generalized dispersion omega(k) = omega0 + vg k + beta k^2 + gamma k^3 and allow Latt(k) = Latt0 (1 + eta k^2); test whether beta, gamma, and eta are statistically distinguishable from zero. If they are not, Eq. 5's assumptions are validated for the measured bandwidth. If they are, recompute the dispersion and group velocities and compare with the quoted vg = 4.422 +/- 0.062 km/s; a shift larger than the stated error would mean the headline values are model-dependent. A co-deposited control SAF on Si/SiO2 with identical antennas would separately settle the 'as good as standard substrates' comparison.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim, that the spin-wave properties are in line with theory and that the material is as good as on standard substrates, rests heavily on the dispersion extracted through Eq. 5. That equation is derived in Sec. IV.B for reciprocal waves with a strictly V-shaped (linear) dispersion, and footnote 30 explicitly assumes that the attenuation length Latt is independent of wavevector. The measured phase is converted into spin-wave wavevector using this assumed form, and the group velocities in Fig. 4(b) are then obtained from linear fits to those converted points. The 'quasi-linear V-shape' of Fig. 4(a) is therefore partly an output of the model rather than an independent measurement. If the true dispersion contains curvature, for example from exchange or dipolar corrections of order k^2 or k^3, or if Latt varies with k, the reported vg, Latt, and their field dependence would be biased. The time-of-flight estimate vg ~ 4.2 km/s and the FMR linewidth are useful independent anchors, so this is a concern about quantitative accuracy rather than data integrity, but the assumption is not separately checked. In addition, the 'as good as standard non-piezoelectric substrates' part of the claim relies on literature values rather than a same-run control sample, which weakens the comparative statement even if the extraction were model-free.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The authors report VNA-FMR and propagating spin-wave spectroscopy measurements on a CoFeB/Ru/CoFeB synthetic antiferromagnet (SAF) grown on a Y-cut LiNbO3 substrate. They extract the uniform acoustic-mode resonance frequency and linewidth, the acoustic spin-wave dispersion and group velocity, and an attenuation length, and they conclude that the magnetic quality is comparable to SAFs on standard non-piezoelectric substrates and consistent with macrospin theory. The paper proposes the LiNbO3/SAF combination as a platform for microwave acousto-magnonics.","tokens_in":7348,"tokens_out":5758,"duration_ms":55192,"significance":"If the quantitative conclusions hold, this is a useful materials advance: it demonstrates low Gilbert damping (alpha = 0.006 +/- 0.001), narrow FMR linewidths, and spin-wave group velocities near 4.4 km/s in a piezoelectric-substrate platform, with good device-to-device consistency across three antenna separations (standard deviation ~200 m/s). The paper benefits from an independent time-of-flight group-velocity estimate (~4.2 km/s) and from careful field-dependent FMR linewidth analysis. However, the central quantitative claims are supported by a model-dependent phase-to-wavevector conversion and by literature-based rather than same-run control comparisons, so the headline statements need either independent verification or appropriate qualification.","major_comments":[{"comment":"Equation (5) is derived under the explicit assumption of a reciprocal, strictly V-shaped (linear) dispersion and a wavevector-independent attenuation length (footnote 30). The phase-to-wavevector conversion and the subsequent linear fits in Fig. 4(b) therefore cannot by themselves establish that the dispersion is linear or that vg = 4.422 +/- 0.062 km/s is a model-free measurement; the quoted uncertainty reflects only the fit precision within the assumed model. Because the abstract and Section V use the dispersion and group velocity as primary evidence that the SAF on LiNbO3 is \"as good as\" standard substrates and \"in line with theory,\" this assumption is load-bearing. I ask the authors to provide an independent check of the linear-dispersion hypothesis, for example a frequency-resolved time-of-flight measurement over a wider bandwidth or a comparison with a model-free inversion of the phase using a more general dispersion ansatz, and to quantify how exchange or dipolar k^2/k^3 corrections and a k-dependent Latt would bias vg and Latt.","section":"IV.B, Eq. (5) and Fig. 4"},{"comment":"The comparative claim that the magnetic properties are \"as good as when grown on standard non-piezoelectric substrates\" is based on literature values for SAFs on silicon (refs. 17 and 25) rather than on a control sample measured in the same run. While the reported FMR linewidth and damping are indicative of good film quality, a same-run control wafer would materially strengthen the comparison. Absent that control, the conclusion should be qualified to \"consistent with literature values for standard substrates\" rather than presented as a direct equivalence.","section":"V and Abstract"},{"comment":"The field dependence of the group velocity is one of the main comparisons with theory, but the data points are obtained from the same model-dependent conversion discussed above. The \"semi-quantitative agreement\" with the 2-macrospin model would be more convincing if the theoretical curve and the measured vg(Hx) were plotted together with an error band that includes the model-parameter uncertainties (Hj, Ms, and the demagnetizing factors), rather than only the fit uncertainties quoted in the text.","section":"IV.C, Fig. 4(b)"}],"minor_comments":[{"comment":"The caption contains a duplicated \"of\": \"Scanning electron microscopy image of of a PSWS device.\"","section":"Fig. 1 caption"},{"comment":"The notation alternates between \"LiNbO 3\" and \"LiNbO3\"; please use a consistent formatting.","section":"Throughout"},{"comment":"The figure-caption style is inconsistent (\"Fig 3.(a)\" versus \"Fig. 3(a)\" and \"Fig 4.(a)\"); please standardize.","section":"Figure captions"},{"comment":"The Data Availability statement says the data are \"available within the article from the corresponding author upon reasonable request,\" which is ambiguous; please clarify whether the data are contained in the article or available on request, and consider providing a repository link.","section":"Data Availability"},{"comment":"The time-of-flight estimate vg = 4.2 km/s is used as an independent cross-check, but no uncertainty is given for this estimate; please state the estimated uncertainty or at least the width of the observed wavepacket arrival.","section":"III and IV.A"}],"recommendation":"major_revision","confidential_remarks":"This is a well-executed materials characterization study, and I do not see a fundamental error in the measurements or the internal consistency checks. My main concern is that the headline quantitative claims are extracted through a model that assumes the very dispersion the paper reports, and the \"standard substrate\" comparison is not a same-run control. Both issues are fixable by additional analysis or by appropriately qualifying the claims. I would also encourage the authors to make the raw phase and transmission data accessible, because the circularity concern can be independently checked only if the data are available."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nShort version: this is a genuine materials-platform paper, not a repackaged methods exercise. It reports the first VNA-FMR and propagating-spin-wave spectroscopy of a CoFeB/Ru/CoFeB synthetic antiferromagnet grown on LiNbO3, and the numbers are internally consistent. The acoustic-mode damping of 0.006, a time-of-flight group velocity around 4.2 km/s, the phase-fit value of 4.4 km/s, and the agreement among three antenna distances within ~200 m/s all point to a clean, high-quality SAF film on a piezoelectric substrate. The field dependence of the group velocity also tracks the expected SAF behavior: rising through the scissor state and saturating near Hj. That is real evidence, not curve fitting.\n\nThe paper's own framing — that the magnetic properties are 'as good as on standard substrates' — is the softest part. There is no same-run control sample on silicon or sapphire; the comparison leans on literature values from the same group (refs. 17, 25, 26) and one external preprint. That is a solid benchmark in practice, but it is not a controlled measurement, and the claim is stated more strongly than the evidence supports.\n\nThe more technical soft spot is the dispersion extraction. Equation 5 assumes a strictly V-shaped, reciprocal dispersion and a k-independent attenuation length (footnote 30). The phase-to-wavevector conversion and the linear fits in Fig. 4 are therefore partly outputs of that assumed model. If there is curvature or k-dependent damping, the reported vg and Latt would be biased. The paper does give one independent anchor — the time-of-flight group delay — and it matches within ~5%, so this is a quantitative-accuracy concern, not a sign of fabricated data. Still, the model-dependence should be disclosed more explicitly and, ideally, checked with a model-free inversion of the phase data or a dispersion measurement over a wider k-range.\n\nWho is this for? Magnonics and acousto-magnonics people choosing a substrate/platform will get concrete numbers. It deserves a serious referee and is citable once the extraction assumptions are clarified or a control sample is added. My sense is a conditional accept with a request to soften the comparative claim and add an explicit statement of what Eq. 5 assumes.\n\nRecommendation: send it to review.","headline":"Solid first characterization of SAF spin waves on LiNbO3, with addressable model-dependence in the dispersion extraction and a control-sample gap.","tokens_in":7966,"tokens_out":2255,"would_cite":true,"duration_ms":19241,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A synthetic antiferromagnet grown on a piezoelectric lithium niobate substrate shows spin-wave properties as good as those on standard substrates, opening a route to acousto-magnonic devices.","keywords":["synthetic antiferromagnet","lithium niobate","spin waves","acousto-magnonics","ferromagnetic resonance","propagating spin wave spectroscopy","magnetic damping","dispersion relation"],"falsifier":"Growing a CoFeB/Ru/CoFeB stack on a silicon substrate in the same deposition run and measuring the same two-antenna transmission would test the equivalence claim directly: if the linewidth or attenuation length differs by more than the reported uncertainties, the central claim fails. Likewise, a wavevector-resolved measurement of the attenuation length, for example by varying the antenna distance over a wider range, would show whether the constant-Latt assumption in Eq. 5 hides a real k dependence.","tokens_in":6858,"feed_emoji":"🧲","tokens_out":5298,"duration_ms":44781,"temperature":0.7,"pith_summary":"This paper shows that a CoFeB/Ru/CoFeB synthetic antiferromagnet, where two magnetic layers are coupled so their magnetizations point opposite, retains its high-quality magnetic dynamics when grown on a piezoelectric lithium niobate substrate. Using ferromagnetic resonance and propagating spin-wave spectroscopy, the authors measure the acoustic spin-wave branch and extract a Gilbert damping near 0.006, linewidths around 370 MHz at 50 mT, group velocities near 4.4 km/s, and an attenuation length of about 3.6 micrometers. These numbers match values reported for the same stack on standard non-piezoelectric substrates and agree with a two-macrospin model, so the paper argues that this platform is ready for microwave acousto-magnonics beyond single-layer magnets.","feed_headline":"SAF on lithium niobate spins as well as on silicon","feed_subtitle":"0.006 damping and 3.6-micron attenuation match standard substrates, enabling acousto-magnonic devices.","key_machinery":"The carrying objects are the acoustic spin-wave branch of the CoFeB/Ru/CoFeB synthetic antiferromagnet and the phase-to-wavevector conversion formula, Eq. 5, that turns two-antenna transmission spectra into a dispersion relation. The SAF provides two eigenmodes, optical (out-of-phase) and acoustic (in-phase), and the acoustic branch is the one that propagates with a V-shaped, frequency-reciprocal dispersion. Equation 5, derived from the antenna response model in the thin-antenna limit, links the phase of the field derivative of the reciprocal transmission to the wavevector k through kr minus a distance-independent correction, plus integer multiples of 2π. That identity is what lets the authors convert measured phase into k and thus extract group velocities and attenuation lengths. The analysis also relies on the demagnetizing factors for dots versus stripes to translate the uniform resonance frequency measured on dots into the stripe value.","core_discovery":"On the paper's own terms, the discovery is that the dynamical spin-wave properties of a synthetic antiferromagnet are essentially unchanged when the stack is deposited on Y-cut LiNbO3, a piezoelectric with a large electromechanical coupling coefficient. The uniform acoustic resonance in 4-micron dots sits at 6.49 GHz at 50 mT with a field-independent linewidth whose damping is about 0.006. The dispersion relation measured on stripes from two-antenna transmission is V-shaped, with a group velocity that rises from 4.4 km/s at 50 mT to about 5.2 km/s near 85 mT and then saturates, consistent with the expected scissors-state behavior of a two-macrospin SAF. From the same data the attenuation length is deduced to be 3.6 micrometers. The paper takes these measurements as evidence that high-quality SAFs can be grown on LiNbO3 and used for acousto-magnonic devices, such as nonreciprocal surface acoustic wave delay lines.","pith_inferences":["If the equivalence to standard substrates holds, the immediate next test is direct SAW-SW coupling: a strong magnetoelastic interaction would show nonreciprocal microwave transmission at the frequencies measured here, something the paper does not attempt.","The claimed parity with standard substrates rests on literature values for SAFs on silicon; a same-run control sample on a non-piezoelectric substrate would either confirm or weaken the claim more directly.","The analysis assumes a strictly linear, reciprocal V-shaped dispersion and a wavevector-independent attenuation length; at wavevectors beyond the measured range, any curvature or k-dependent Latt would alter the extracted group velocities and device design.","The 3.6-micrometer attenuation length is comparable to the antenna separations used, so tuning antenna spacing could trade off transmitted signal strength against signal-to-noise, a practical consideration for building functional magnonic devices on this platform."],"forward_implications":["Piezoelectric LiNbO3 can serve as a substrate for SAF magnonic devices without degrading spin-wave quality, removing a major materials obstacle for acousto-magnonics.","The measured damping of about 0.006 and attenuation length of 3.6 micrometers imply that acoustic spin waves can carry information across micrometer-scale distances, enough for on-chip magnonic circuits.","Group velocities of 4 to 5 km/s match the V-shaped dispersion predicted for SAFs, so the two-macrospin model can be used to design future devices on this platform.","This platform enables coupling of spin waves to surface acoustic waves, opening a route to nonreciprocal SAW delay lines and other acousto-magnonic functions that single-layer ferromagnets cannot easily provide.","The phase-to-wavevector extraction method works for reciprocal spin waves on SAFs, making it a reusable tool for characterizing other multilayer magnonic waveguides."],"supporting_citations":[{"why":"Provides the two-macrospin SAF model and the experimental benchmark of SAFs on silicon substrates that the paper compares against.","marker":"[17]"},{"why":"Reports resonance frequencies and linewidths of SAFs on standard substrates, serving as the reference for the 'as good as standard substrates' claim.","marker":"[25]"},{"why":"Supplies the antenna response model and the expression linking S21 phase to wavevector, from which Eq. 5 is derived.","marker":"[29]"},{"why":"Gives the linewidth-to-Gilbert-damping conversion used to obtain alpha approximately 0.006.","marker":"[22]"},{"why":"Provides the time-of-flight spectroscopy method used for time-gating the transmission data.","marker":"[27]"},{"why":"Establishes the field-differentiation data processing and the reciprocal-wave condition for the same stack geometry.","marker":"[18]"},{"why":"Theoretically proposes SAW nonreciprocity via magnetoelastic coupling with a SAF, the application this platform is claimed to enable.","marker":"[31]"}],"fun_headline_variants":["Spin waves in SAF on LiNbO3 match standard substrates","SAF on lithium niobate: spin waves as good as on silicon","Piezoelectric LiNbO3 hosts high-quality SAF spin waves","Acousto-magnonic platform: SAF on LiNbO3 meets standards"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The reported dispersion and velocities assume the spin-wave branch has an exactly linear, reciprocal V-shape with a wavevector-independent attenuation length, and the comparison with standard substrates rests on published values rather than a control sample measured in the same run.","fun_headline_variants_meta":{"raw":{"variants":["Spin waves in SAF on LiNbO3 match standard substrates","SAF on lithium niobate: spin waves as good as on silicon","Piezoelectric LiNbO3 hosts high-quality SAF spin waves","Acousto-magnonic platform: SAF on LiNbO3 meets standards"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000575,"raw_usage":{"total_tokens":2693,"prompt_tokens":899,"completion_tokens":1794,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":515,"completion_tokens_details":{"reasoning_tokens":1715}},"tokens_in":515,"tokens_out":1794,"duration_ms":12235,"temperature":1.0,"reasoning_tokens":1715,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T11:25:05.327603+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Growing a CoFeB/Ru/CoFeB stack on a silicon substrate in the same deposition run and measuring the same two-antenna transmission would test the equivalence claim directly: if the linewidth or attenuation length differs by more than the reported uncertainties, the central claim fails. Likewise, a wavevector-resolved measurement of the attenuation length, for example by varying the antenna distance over a wider range, would show whether the constant-Latt assumption in Eq. 5 hides a real k dependence.","supporting_citations":[{"cited_title":"Millo , author J.-P","cited_arxiv_id":null,"evidence_quote":"Provides the two-macrospin SAF model and the experimental benchmark of SAFs on silicon substrates that the paper compares against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports resonance frequencies and linewidths of SAFs on standard substrates, serving as the reference for the 'as good as standard substrates' claim."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the antenna response model and the expression linking S21 phase to wavevector, from which Eq. 5 is derived."},{"cited_title":"Devolder , author S.-M","cited_arxiv_id":null,"evidence_quote":"Gives the linewidth-to-Gilbert-damping conversion used to obtain alpha approximately 0.006."},{"cited_title":"Devolder , author G","cited_arxiv_id":null,"evidence_quote":"Provides the time-of-flight spectroscopy method used for time-gating the transmission data."},{"cited_title":"Thiancourt , author S","cited_arxiv_id":null,"evidence_quote":"Establishes the field-differentiation data processing and the reciprocal-wave condition for the same stack geometry."},{"cited_title":"Verba , author V","cited_arxiv_id":null,"evidence_quote":"Theoretically proposes SAW nonreciprocity via magnetoelastic coupling with a SAF, the application this platform is claimed to enable."}],"review_version":1}