{"id":"c2aeb557-089a-4c99-88b3-30e1e7bee4ae","arxiv_id":"2411.19646","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A YIG-GGG resonator with thin YIG couples magnetic and acoustic modes strongly enough to build a simplified, frequency-tunable oscillator that achieves up to 30 dB lower phase noise on acoustic modes.","lead":"Researchers built a small oscillator from a crystal made of yttrium iron garnet (YIG) on a gadolinium gallium garnet (GGG) substrate, combining magnetic and acoustic resonances. The oscillator tunes its frequency with a weak magnetic field and shows up to 30 dB lower phase noise when locked to acoustic modes, which could simplify low-noise microwave sources for communications and radar.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Appendix B's quoted HBAR edge frequencies imply Q≈7015, not the stated 9018, so the 'agreement' between the measured 30 dB phase-noise improvement and the Q-factors is unsupported.","rationale":"The central claim of the paper reports a 30 dB phase-noise improvement that is explicitly said to be 'in agreement' with Q_HBAR≈9018 and Q_FMR≈446. This agreement is the quantitative backbone connecting the improved oscillator design to the stated performance. However, the edge frequencies quoted in Appendix B for the HBAR mode imply Q≈7015, which changes the expected improvement from about 26 dB to about 24 dB. This is a concrete, internal inconsistency: it does not rely on contested external theory, only on arithmetic with the paper's own numbers. If the Q value is 9018, the frequency gap must be ~140 kHz; if the gap is 180 kHz, the Q is 7015. Either way the text needs correction, and the Leeson comparison needs to be re-evaluated. The issue is load-bearing because the 'agreement' statement is used to validate the phase-noise measurements and the design choices. It does not necessarily invalidate the demonstrated oscillator behavior (the 30 dB improvement could still be real), but it removes the stated supporting evidence and requires the authors to explain the extra 4-6 dB. This does not change the CONDITIONAL verdict; it sharpens the condition: the authors must provide raw S21 data, corrected Q values, and a Leeson model that accounts for any power or noise-figure differences. No other concern seems as directly tied to the central quantitative claim.","tokens_in":13211,"tokens_out":8549,"duration_ms":67038,"concrete_test":"Recompute Q_HBAR from the raw S21 measurement around the 385th HBAR mode using the Kajfez construction in Appendix B, explicitly listing the two blue-point frequencies. Then check whether the frequency gap is 140 kHz (Q≈9018) or 180 kHz (Q≈7015). Independently, re-fit the Leeson lines in Fig. 5(b) using the corrected Q and a single noise figure for both regimes: if the 30 dB separation at 1 kHz offset cannot be reproduced, the claimed agreement with the Q-factors fails and the authors must state the additional mechanism (e.g., difference in carrier power or noise figure).","verdict_should_be":"UNCHANGED","load_bearing_attack":"In Appendix B, the loaded Q of the 385th HBAR is computed as Q = f0/(fA2 − fA1) (Eq. B3), with fA1 = 1.26264 GHz and fA2 = 1.26282 GHz. Taking f0 ≈ 1.26273 GHz, the quoted frequency separation of 180 kHz yields Q ≈ 7015, not 9018. The FMR line uses fFMR1 = 1.25518 GHz and fFMR2 = 1.25800 GHz, giving Q ≈ 446, consistent with the statement; the discrepancy is specific to the acoustic mode. If Q_HBAR is actually 7015, the maximum phase-noise improvement from the Q ratio alone, 20 log10(7015/446), is about 23.9 dB, making the reported 30 dB improvement even harder to attribute to the measured quality factors. The text's 'in agreement with the measured FMR and HBAR Q-factors' is thus not supported by the numbers given. The Leeson fits in Fig. 5(b) would need an additional 4–6 dB from a difference in noise figure or carrier power between the two regimes; this is not stated or justified. The discrepancy could be a typographical error, but it is exactly the quantitative link the central claim depends on, so it must be resolved.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports a two-port magneto-acoustic oscillator (MAO) based on a YIG-GGG composite resonator with reduced YIG thickness (1.08 µm) that achieves ~1 MHz magneto-elastic coupling in the 1-2 GHz range. The simplified design eliminates the circulator and preselector required in the authors' previous work. Two operational regimes are demonstrated: a low-phase-noise regime locking only to high-Q HBAR modes with discrete frequency tunability in ~3.281 MHz steps, and a complex regime locking to both HBAR and FMR modes with continuous tunability and variable phase noise. The authors claim a 30 dB phase-noise improvement when locking to HBAR modes relative to FMR locking, stated to be in agreement with measured Q factors (Q_HBAR ≈ 9018, Q_FMR ≈ 446).","tokens_in":13405,"tokens_out":6158,"duration_ms":51083,"significance":"If the claims hold, the work is a useful step toward practical tunable low-phase-noise oscillators, replacing bulky external components with a single composite resonator. The qualitative evidence—avoided crossings in S-parameters, spectrograms, and phase-noise maps—is direct and compelling. The thickness optimization is grounded in a parameter-free coupling model from prior literature, which strengthens the design rationale. However, the quantitative link between the 30 dB phase-noise improvement and the measured Q factors is not supported by the numbers presented, and the missing Leeson-model details prevent verification of the central quantitative claim.","major_comments":[{"comment":"The HBAR Q-factor is inconsistent with the quoted edge frequencies. The text gives fA1 = 1.26264 GHz and fA2 = 1.26282 GHz, i.e., a linewidth of 180 kHz. With f0 ≈ 1.26273 GHz, Eq. (B3) yields Q ≈ 7015, not 9018. Even if the stated 9018 were correct, the expected phase-noise improvement from the Q ratio alone is 20 log10(9018/446) ≈ 26 dB, and with Q ≈ 7015 it is ≈ 24 dB. The abstract and Section V claim a 30 dB improvement 'in agreement with the measured FMR and HBAR Q-factors'; this agreement is not justified. The authors must correct the Q value or the edge frequencies and must account for the remaining 4-6 dB, for instance by specifying different noise figures or signal-power levels in the two regimes.","section":"Appendix B, Eq. (B3)"},{"comment":"The Leeson-model comparison is not reproducible. The text refers to 'Leeson's equation (??)' without displaying the expression, and the dashed lines are said to be in 'Fig. 6(b)' although the phase-noise plots appear in Fig. 5(b). The caption states that the output power (-26 dBm) and the power at the resonator output (-51 dBm) were used, but the actual equation and the parameter values for both regimes are not given. Without these, the claimed correspondence between the measured 30 dB improvement and the Leeson estimates cannot be checked. Please provide the Leeson equation used and the numerical values of the noise figure and power for the FMR-locked and HBAR-locked cases.","section":"Section V"},{"comment":"The Darko Kajfez Smith-chart extraction is applied to the strongly asymmetric HBAR resonance shown in Fig. 6(b,d). This method assumes a circular resonance loop, and the strongly asymmetric line shape raises questions about the validity of the extracted Q-factor. The choice of the 'blue points' at the intersection of the auxiliary diameter line with the S21 curve is subjective for a non-circular loop. The authors should provide an uncertainty estimate for the extracted Q and demonstrate that the result is robust against reasonable variations in the auxiliary-line placement, because the Q ratio is the central basis for the claimed phase-noise improvement.","section":"Appendix B / Section III"}],"minor_comments":[{"comment":"The text states that the 6 dB attenuator setting corresponds to 'a green horizontal line in Fig. 4(b,c)', but no green line appears in those panels; the relevant reference is likely Fig. 3(b)-(e), and the line color should be specified in the figure or caption.","section":"Section V"},{"comment":"The caption says 'applied magnetic field between 10.29 and 10.52 MHz'; the units should be mT, not MHz.","section":"Fig. 3(a) caption"},{"comment":"The theoretical slope from Eq. (3) is 62.84 MHz/mT, while the measured value is 55.65 MHz/mT. The text calls these 'close', but an 11% deviation may merit a brief comment, for example on demagnetizing effects or the approximate nature of the Taylor expansion.","section":"Section II"},{"comment":"The DOI listed for reference [48] appears to belong to a Physical Review B article rather than the Journal of Physics D article cited; please verify and correct the reference.","section":"Reference [48]"}],"recommendation":"major_revision","confidential_remarks":"The experimental platform is interesting and likely of value to the applied-physics and microwave-oscillator communities, but the central quantitative claim (30 dB improvement 'in agreement' with Q factors) is currently unsupported by the numbers in Appendix B, and the missing Leeson equation prevents verification. These issues are fixable and do not appear to undermine the qualitative findings. I would encourage the editor to request a revision that corrects the Q calculation, supplies the Leeson model details, and clarifies the comparison between regimes."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a real incremental step: a two-port YIG-GGG magneto-acoustic oscillator that removes the circulator and preselector of the earlier design, achieves ~1 MHz magneto-elastic coupling by thinning YIG to 1.08 µm, and demonstrates two operating regimes — low-phase-noise discrete HBAR locking and a complex regime with continuous tuning. The measured phase noise at 1 kHz offset improves from about -35 to -65 dBrad²/Hz when the oscillator switches from FMR to HBAR locking. That 30 dB improvement is directly measured and is the strongest result in the paper. The coupling model (Eq. 1) is parameter-free and taken from previous work; it isn't fitted to the output, so the circularity burden is low.\n\nThe soft spots are quantitative, and they matter. The abstract says the 30 dB improvement is \"in agreement with the measured FMR and HBAR Q-factors,\" but the numbers don't support that. The two Q values quoted (446 and 9018) give 20 log10(9018/446) ≈ 26 dB, not 30 dB. Worse, Appendix B's own edge frequencies for the HBAR — f_A1 = 1.26264 GHz and f_A2 = 1.26282 GHz — imply Q ≈ 7015, not 9018. That makes the predicted improvement from the Q ratio about 24 dB, leaving 6 dB unexplained. The Leeson fits shown in Fig. 5(b) use an effective noise figure that is never stated, and the Leeson equation itself is missing from the text (it appears as \"(??)\"). So the quantitative link between the measured Q-factors and the phase-noise improvement is not established. The improvement may well be real, but the explanation needs more work.\n\nThere's also a clear copy-paste error in Appendix A: it describes the composite two-port MAR with YIG thickness 9.75 µm and GGG thickness 364 µm, contradicting Section III's 1.08 µm and 543 µm. That has to be fixed. No uncertainties are given for any of the Q or phase-noise values, which is a bit thin for a device paper.\n\nNone of this is fatal to the qualitative story. The two-regime demonstration and the simplified architecture are solid engineering contributions, and the measurements are directly reported. The problems are in the interpretation and the arithmetic, and they're fixable.\n\nI'd send this to review — it deserves referee time — but my own verdict would be conditional until the Q inconsistency, the missing Leeson details, and the Appendix A errors are resolved. It's a \"maybe\" for the reading group, mostly because the fixable issues make it a useful object lesson in checking claimed agreements. I'd cite it if I were working on magneto-acoustic oscillators, once it's cleaned up.","headline":"A useful two-port YIG-GGG oscillator with a directly measured 30 dB phase-noise improvement, but its own Q-factor arithmetic doesn't explain that improvement.","tokens_in":14052,"tokens_out":2736,"would_cite":true,"duration_ms":21212,"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":"Thinning YIG to 1.08 µm strengthens magneto-elastic coupling enough to simplify and quiet a microwave oscillator.","keywords":["magneto-acoustic oscillator","YIG-GGG resonator","magneto-elastic coupling","HBAR modes","phase noise","frequency tunability","ferromagnetic resonance","two-port oscillator"],"falsifier":"Measure the phase-noise difference on the same oscillator between an FMR-locked state and an adjacent HBAR-locked state at equal output power and compare it with $20\\log_{10}(9018/446) \\approx 26$ dB predicted from the quoted Q-factors; a difference far from 30 dB, or Q-factors re-measured by an independent method that disagree, would overturn the claimed mechanism.","tokens_in":12937,"feed_emoji":"📡","tokens_out":7741,"duration_ms":60630,"temperature":0.7,"pith_summary":"The paper claims that cutting the YIG layer in a YIG-GGG composite to 1.08 µm raises the magneto-elastic coupling between the ferromagnetic resonance and the acoustic thickness modes to about 1 MHz in the 1-2 GHz band. That strong coupling makes the acoustic resonances large and asymmetric enough that a two-port oscillator loop can lock onto them without a circulator or an extra YIG preselector. Depending on the loop gain, the oscillator locks either to high-Q acoustic modes, where phase noise drops by about 30 dB, or to both acoustic and ferromagnetic modes, where tuning becomes continuous. The practical payoff would be a compact, field-tunable oscillator whose frequency step and noise floor are set by acoustic rather than magnetic modes.","feed_headline":"Thinner YIG film cuts oscillator phase noise by 30 dB","feed_subtitle":"Slimming YIG to 1.08 µm lets a two-port oscillator lock to high-Q acoustic modes without a circulator.","key_machinery":"The load-bearing object is the composite magneto-acoustic resonator (MAR), a 1.08 µm YIG film on a 543 µm GGG substrate, whose coupling strength is governed by $\\kappa^2 = \\frac{\\gamma b^2}{8\\pi^3 M_s \\lambda d \\sqrt{c_{44}\\rho}}\\left(1-\\cos\\frac{\\pi\\lambda d}{L}\\right)^2$. The mechanism is the thickness-dependent mode overlap: the uniform FMR mode couples to the $\\lambda$-th acoustic mode through the factor $\\left(1-\\cos\\frac{\\pi\\lambda d}{L}\\right)^2$, so thinning YIG pushes strong coupling into the 1-2 GHz range where GGG acoustic losses are still acceptable. This coupling reshapes the S21 response into asymmetric, phase-steep resonances, and it is the phase steepness that lets a two-port loop satisfy the Barkhausen criterion at selected HBAR frequencies.","core_discovery":"The central discovery is that reducing the YIG film thickness from 9.75 µm to 1.08 µm increases the overlap between the uniform FMR profile and the standing acoustic modes, raising the coupling constant $\\kappa$ from 0.47 MHz to 1.0 MHz. In the two-port resonator this appears as pronounced avoided crossings and strongly asymmetric S21 resonances. The authors show that this permits a simplified ring oscillator in which the Barkhausen amplitude and phase conditions can be satisfied around HBAR modes alone. In the low-phase-noise regime the oscillator emits only at discrete fields, producing steps of 3.281 MHz between adjacent acoustic modes; in the higher-gain regime it oscillates at any field, locking to the FMR or to HBAR modes depending on field. Phase noise at 1 kHz offset is about -35 dBrad²/Hz when locked to FMR and about -65 dBrad²/Hz when locked to an HBAR, a 30 dB improvement the authors attribute to the measured Q-factors, 446 for FMR and 9018 for the 385th HBAR mode.","pith_inferences":["Beyond the paper: because the 3.281 MHz step is set by the GGG substrate thickness, choosing a different substrate thickness should allow the channel spacing of the discrete tuning regime to be engineered.","Beyond the paper: the same asymmetric-resonance condition could be exploited as a magnetically reconfigurable filter or switch, since the Barkhausen phase condition selects one acoustic mode at a time.","Beyond the paper: an independent test is to sweep the field at fixed attenuation in the low-noise regime and verify that the oscillator's discrete locking windows coincide exactly with the avoided-crossing centers in the S21 spectrogram; any offset would indicate the locking is not purely mode-determined."],"forward_implications":["A two-port magneto-acoustic oscillator can be built without a circulator or preselector, shrinking the footprint relative to the earlier one-port design.","Low-phase-noise operation comes with discrete frequency tuning in 3.281 MHz steps matching the HBAR mode spacing, which suits channelized frequency synthesis.","Switching the loop gain into the complex regime gives continuous field tuning while the phase noise varies by up to 30 dB depending on which mode is locked.","The in-plane magnetization geometry lowers the required field from roughly 220 mT to about 10 mT and doubles the tuning slope to 55.65 MHz/mT."],"supporting_citations":[{"why":"Provides the earlier one-port MAO design and its 20 dB improvement, the baseline this work simplifies and improves.","marker":"[15]"},{"why":"Gives the coupled magnon-phonon theory used for the coupling-constant expression in Eq. (1).","marker":"[29]"},{"why":"Supplies the magnetoelastic interaction formalism underlying the coupling calculation.","marker":"[30]"},{"why":"Reviews spin-insulatronics magnetoelastic coupling used to justify the coupling scaling with mode overlap.","marker":"[31]"},{"why":"Demonstrates magnon-phonon avoided crossings in similar garnet systems, the signature used here to identify strong coupling.","marker":"[32]"},{"why":"Shows bright and dark states of strongly coupled magnon-phonon modes, backing the avoided-crossing interpretation of the S21 data.","marker":"[33]"},{"why":"Supplies the Hilbert-transform phase-noise measurement approach used for the reported phase-noise spectra.","marker":"[50]"},{"why":"Supplies the Smith-chart resonance-loop method used to extract loaded Q-factors for FMR and HBAR resonances.","marker":"[53]"},{"why":"Details the geometric Q-factor extraction procedure used to obtain Q values of 446 and 9018.","marker":"[54]"}],"fun_headline_variants":["1.08 µm YIG cuts oscillator phase noise 30 dB","Magneto-acoustic oscillator needs no circulator","Thin YIG boosts magneto-elastic coupling","30 dB quieter oscillator with thinner YIG","YIG-GGG oscillator tunes in 3.28 MHz steps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes that the 30 dB phase-noise difference is actually caused by the resonator Q-factors extracted from Smith-chart resonance loops (446 and 9018), through ordinary Leeson scaling, and that those Q values remain valid for strongly asymmetric magneto-acoustic resonances.","fun_headline_variants_meta":{"raw":{"variants":["1.08 µm YIG cuts oscillator phase noise 30 dB","Magneto-acoustic oscillator needs no circulator","Thin YIG boosts magneto-elastic coupling","30 dB quieter oscillator with thinner YIG","YIG-GGG oscillator tunes in 3.28 MHz steps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001035,"raw_usage":{"total_tokens":4442,"prompt_tokens":1114,"completion_tokens":3328,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":730,"completion_tokens_details":{"reasoning_tokens":3249}},"tokens_in":730,"tokens_out":3328,"duration_ms":21673,"temperature":1.0,"reasoning_tokens":3249,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T05:58:55.872869+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the phase-noise difference on the same oscillator between an FMR-locked state and an adjacent HBAR-locked state at equal output power and compare it with $20\\log_{10}(9018/446) \\approx 26$ dB predicted from the quoted Q-factors; a difference far from 30 dB, or Q-factors re-measured by an independent method that disagree, would overturn the claimed mechanism.","supporting_citations":[{"cited_title":"Litvinenko, R","cited_arxiv_id":null,"evidence_quote":"Provides the earlier one-port MAO design and its 20 dB improvement, the baseline this work simplifies and improves."},{"cited_title":"Verba, I","cited_arxiv_id":null,"evidence_quote":"Gives the coupled magnon-phonon theory used for the coupling-constant expression in Eq. (1)."},{"cited_title":"Lisenkov, A","cited_arxiv_id":null,"evidence_quote":"Supplies the magnetoelastic interaction formalism underlying the coupling calculation."},{"cited_title":"Brataas, B","cited_arxiv_id":null,"evidence_quote":"Reviews spin-insulatronics magnetoelastic coupling used to justify the coupling scaling with mode overlap."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates magnon-phonon avoided crossings in similar garnet systems, the signature used here to identify strong coupling."},{"cited_title":"Litvinenko, P","cited_arxiv_id":null,"evidence_quote":"Supplies the Hilbert-transform phase-noise measurement approach used for the reported phase-noise spectra."},{"cited_title":"Kajfez and P","cited_arxiv_id":null,"evidence_quote":"Supplies the Smith-chart resonance-loop method used to extract loaded Q-factors for FMR and HBAR resonances."},{"cited_title":"Kajfez and E","cited_arxiv_id":null,"evidence_quote":"Details the geometric Q-factor extraction procedure used to obtain Q values of 446 and 9018."}],"review_version":1}