{"id":"69c70ebe-b4a1-42fd-a83d-2ed32ffaa605","arxiv_id":"2411.18870","paper_version":1,"verdict":"REJECT","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"high","formal_verification":"none","parameter_count":2,"one_line_summary":"A TFLN microdisk using polygon modes achieves 48.08% SHG conversion efficiency at 4.599 mW on-chip pump, surpassing prior domain-inversion-free schemes.","lead":"This paper reports second harmonic generation with a claimed 48% conversion efficiency in a thin-film lithium niobate microdisk, using polygon-shaped cavity modes and no domain inversion. The result, if correct, would be a new record for domain-inversion-free on-chip frequency conversion, approaching poled PPLN resonators.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 48.08% SHG efficiency is internally inconsistent with the paper's own 7.5%/mW normalized-efficiency fit at 4.599 mW, so the headline conversion figure is not self-consistent.","rationale":"The reader's weakest-assumption is that on-chip pump calibration is unspecified, which is a real problem. My stress-test finds a sharper, more load-bearing version: the paper's own two numbers—7.5%/mW normalized efficiency and 48.08% absolute efficiency at 4.599 mW—are mathematically inconsistent under the stated linear dependence. This inconsistency does not depend on external calibration assumptions; it is internal to Section 2.3. If the linear fit is correct, the point efficiency should be about 34.5%, not 48.08%; if the point efficiency is correct, the fitted slope should be about 10.45%/mW, not 7.5%/mW. The paper's comment that the SHG output is underestimated due to low collection efficiency further confuses the definitions rather than resolving the discrepancy. I therefore agree with the reader's negative verdict, but I locate the decisive problem in the self-inconsistency of the efficiency numbers, with the missing on-chip calibration as a compounding issue.","tokens_in":6946,"tokens_out":3280,"duration_ms":32544,"concrete_test":"Reconstruct Figure 2(c,d) from the raw data: plot measured SHG power versus stated on-chip pump power, overlay the 4.599 mW, 2.211 mW point and the fitted 7.5%/mW line, and require a single linear slope (with error bars) to describe all points. Additionally, supply the full calibration chain from the power meter before the taper to the on-chip power, including taper transmission, mode-coupling efficiency, and the SHG collection efficiency from the microdisk to the detector. If the high-power point deviates from the low-power fitted line by more than 10%, or if the calibration chain is missing, the 48.08% claim is not established.","verdict_should_be":"REJECT","load_bearing_attack":"The central claim—absolute SHG conversion efficiency of 48.08% at 4.599 mW on-chip pump—fails an internal consistency check in Section 2.3 and Figure 2(c,d). The paper states that from the linear fit of Figure 2(d) the normalized conversion efficiency is 7.5%/mW, and that at an on-chip pump power of 4.599 mW the SHG power was 2.211 mW. If efficiency is defined as P_SHG/P_pump, then 2.211 mW / 4.599 mW = 48.08%, which corresponds to a slope of 10.45%/mW, about 39% larger than the reported 7.5%/mW fit. Conversely, 7.5%/mW predicts 34.5% at 4.599 mW. The text also says the 7.5%/mW value is underestimated because only the SHG output from the tapered fiber was collected with low coupling efficiency; that statement makes the inconsistency more severe, because the same collected SHG power is used for the 48.08% absolute efficiency without any stated collection-efficiency correction. Moreover, the on-chip pump power is never derived: the paper reports a loaded Q of 2.69e6 and 74.1% transmission but does not specify how input power was calibrated to the microdisk, nor how taper transmission and mode-coupling efficiency were measured. Thus the headline efficiency is not robustly supported even before assessing the phase-matching mechanism.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports ultra-efficient second harmonic generation (SHG) in an X-cut thin-film lithium niobate microdisk, claiming 48.08% absolute conversion efficiency at an on-chip pump power of 4.599 mW. The mechanism is attributed to cavity polygon modes with a square-shaped intensity pattern, which the authors argue provides natural quasi-phase-matching through sign reversals of the effective nonlinear coefficient d_eff as light circulates. The paper also reports high loaded and intrinsic Q factors, a high modal overlap factor, and compares the result favorably with prior domain-inversion-free and PPLN microresonator SHG demonstrations.","tokens_in":7272,"tokens_out":5824,"duration_ms":52298,"significance":"If the central claims were fully supported, this would be a notable advance in on-chip frequency conversion without electric-field poling, leveraging d_33 and high-Q polygon modes. The paper includes direct spectral measurements, CCD images of the modal patterns, and Q-factor characterizations, which are useful experimental evidence. However, the absolute efficiency number is internally inconsistent with the reported normalized efficiency, the on-chip pump calibration is not described, and the proposed quasi-phase-matching picture is quantitatively inconsistent with the stated phase mismatch. The significance of the work is therefore contingent on resolving these load-bearing issues.","major_comments":[{"comment":"The reported absolute efficiency is internally inconsistent with the reported normalized efficiency. The text states that the linear fit in Fig. 2(d) gives 7.5%/mW and that at an on-chip pump power of 4.599 mW the SHG power was 2.211 mW, yielding 48.08%. If conversion efficiency is defined as P_SHG/P_pump, then the same data point implies a slope of 10.45%/mW relative to the zero-pump origin, about 39% larger than the stated fit; alternatively, 7.5%/mW predicts 34.5% at 4.599 mW. The authors must reconcile this discrepancy, for example by reporting the fit range, stating whether the 4.599 mW point was included in the fit, or providing the full dataset with residuals.","section":"§2.3, Fig. 2(c,d)"},{"comment":"The text says that the 7.5%/mW normalized efficiency is underestimated because only the SHG power output from the tapered fiber was collected with low coupling efficiency, yet the same collected power is used without correction to compute the 'absolute conversion efficiency' of 48.08%. This is self-contradictory: if the measured SHG power is only a fraction of the generated SHG, then the measured ratio 2.211 mW/4.599 mW is a lower bound, not an absolute efficiency, unless an unstated collection-efficiency correction was applied. The paper needs to specify what quantity is being reported and how it was calibrated.","section":"§2.3"},{"comment":"The on-chip pump power of 4.599 mW is never derived. Section 2.4 reports a loaded Q of 2.69×10^6 and a transmission of 74.1%, but this is insufficient: the taper–microdisk coupling efficiency, fiber insertion loss, and any power calibration against a reference detector are not described. Without this calibration, the absolute efficiency and the pump-power dependence in Fig. 2(c,d) cannot be assessed.","section":"§2.3, §2.4, Fig. 2"},{"comment":"The proposed natural quasi-phase-matching mechanism is quantitatively inconsistent with the paper's own phase-mismatch value. With Δk = 1.4159×10^6 m^-1, the coherence length is L_c = π/Δk ≈ 2.22 μm. The sign of d_eff changes only twice per round trip, so the sign-reversal period is the full cavity perimeter (π × 57.28 μm ≈ 180 μm), or at best half that if one counts both sign changes of the two-level amplitude. This period is tens of coherence lengths, so the momentum provided by the sign flips (2π/180 μm ≈ 3.5×10^4 m^-1) is more than an order of magnitude too small to compensate Δk. The authors' own statement that |E_SHG| oscillates with ~18.5 periods within each half-cycle confirms that no net quasi-phase-matched growth occurs on the scale of the sign-flip period; the mechanism as described cannot explain the claimed efficiency.","section":"§2.5, Eq. (2), Fig. 4"}],"minor_comments":[{"comment":"The modal overlap factor is given as ~80% in the abstract but 86% in Section 2.3; the discrepancy should be reconciled.","section":"Abstract and §2.3"},{"comment":"The phrase 'play important poles' should read 'play important roles'.","section":"§2.4"},{"comment":"There is a typo: 'carried put' should be 'carried out'.","section":"§2.1"},{"comment":"In Eq. (2), the notation Δk(z) is used while z also appears as the integration variable; please define the convention clearly, e.g., whether Δk is treated as piecewise constant over each segment.","section":"§2.5, Eq. (2)"},{"comment":"The figures would benefit from axis labels, error bars, and a statement of which points were included in the linear fit shown in Fig. 2(d).","section":"Fig. 2(c,d)"}],"recommendation":"reject","confidential_remarks":"The manuscript draws heavily on the authors' prior work (Refs. 15, 19–22) for the polygon-mode and natural-QPM concepts, and the present QPM argument fails quantitatively against the paper's own phase-mismatch value. The headline efficiency also has an internal inconsistency and an unverified on-chip power calibration. I recommend rejection, though a substantially revised manuscript with corrected numbers, a full calibration description, and a valid phase-matching model could merit reconsideration."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know about this one. The headline 48.08% SHG efficiency, if real, would be a genuine advance for domain-inversion-free TFLN microresonators, roughly doubling the best prior numbers (23% and 30%). The polygon-mode approach is clever, and the reported intrinsic Q of ~3.9e7 for the pump mode is very good. The device work appears careful, and the high modal overlap among the square modes is a plausible reason for the strong nonlinear response.\n\nThe problem is that the central number does not survive contact with the paper's own data. The text states that a linear fit to the conversion efficiency gives a normalized value of 7.5%/mW. At the quoted on-chip pump power of 4.599 mW, that predicts 34.5% efficiency, not 48.08%. The 48.08% figure is obtained from 2.211 mW of SHG power divided by 4.599 mW, which corresponds to a slope of 10.45%/mW—about 40% higher. The note that the 7.5% value is underestimated because only the taper-collected SHG was measured makes it worse, because the same collected power is used to compute the 48.08%. Either the fit is wrong, the point is wrong, or the efficiency is being defined two different ways. As written, the headline is not self-consistent.\n\nTwo other soft spots. First, the on-chip pump power is never derived. The paper reports a 74.1% transmission dip and a loaded Q, but not how input power translates to power circulating in the disk; a calibration error here directly inflates the efficiency. Second, the natural quasi-phase-matching explanation is quantitatively shaky. The stated phase mismatch Δk = 1.42e6 m^-1 implies a coherence length of about 2.2 μm, while each half-side of the square mode is on the order of tens of microns. The sign of d_eff flips every half-cycle, i.e. every ~20 coherence lengths, not at the QPM period. The paper shows oscillations within each half-cycle but still claims 'remarkable gain' without computing the integrated effect. That needs a proper model.\n\nSo, a promising device and a potentially important result, but the paper is not acceptable in its current form. The efficiency inconsistency and missing calibration are fixable with a clear definition and a careful accounting of coupled power; the QPM argument needs a quantitative comparison. I would send this to peer review rather than desk reject, because the experimental claim is significant and the flaws are checkable. If the authors can reconcile the numbers, this could become a strong letter.\n\nFor now, I would not cite the 48% number.","headline":"Potentially significant efficiency result, but the 48% claim contradicts the paper's own fit; needs major revision.","tokens_in":7819,"tokens_out":4729,"would_cite":false,"duration_ms":43022,"reading_group":"yes","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Cavity polygon modes in a lithium niobate microdisk achieve 48.08% second-harmonic conversion efficiency without any domain engineering.","keywords":["second harmonic generation","thin-film lithium niobate","microdisk resonator","polygon modes","natural quasi-phase matching","d33 nonlinear coefficient","high-Q microcavity","modal overlap"],"falsifier":"Measure the fiber-to-microdisk coupling efficiency independently (for example, by fitting the transmission dip or adding a drop port) and collect the second harmonic emitted in all directions with an integrating sphere; if the true on-chip pump power is below 4.599 mW or the total generated SHG differs from the collected 2.211 mW, the reported 48.08% absolute conversion efficiency would not hold.","tokens_in":6755,"feed_emoji":"💡","tokens_out":11737,"duration_ms":87060,"temperature":0.7,"pith_summary":"The paper claims that second harmonic generation in a monocrystalline X-cut thin-film lithium niobate microdisk can reach an absolute conversion efficiency of 48.08% without any ferroelectric domain engineering. The key move is to form 'polygon modes'—cavity modes with a square spatial pattern whose two parallel sides sit perpendicular to the crystal's optical axis—so the largest nonlinear coefficient $d_{33}$ is used, and the sign of the effective nonlinearity reverses every half circuit around the cavity, mimicking periodic poling. This natural quasi-phase matching, together with measured intrinsic Q factors above $3\\times10^7$ and a modal overlap around 80%, lets a 4.599 mW on-chip pump at 1560.6 nm produce 2.211 mW of second harmonic at 780.3 nm. The authors claim this is the highest efficiency reported for domain-inversion-free SHG in thin-film lithium niobate microresonators, approaching the 52% record of periodically poled devices.","feed_headline":"Microdisk doubles light's frequency with 48% efficiency","feed_subtitle":"Domain-inversion-free lithium niobate resonator reaches record-level efficiency on just 4.6 mW of pump power.","key_machinery":"The central object is the cavity polygon mode, a square-shaped superposition of whispering-gallery modes created by the weak perturbation of the tapered fiber. The carrying mechanism is the quarter-cycle sign flip of the effective nonlinear coefficient $d_{\\text{eff}}(\\theta) = -d_{22}\\cos^3\\theta + 3d_{31}\\cos^2\\theta\\sin\\theta + d_{33}\\sin^3\\theta$, which takes the values $-2.46$, $-41.7$, $2.46$, and $41.7$ pm/V in successive quarter-circumferences of the square pattern. This periodic sign inversion acts like the domain inversion of a periodically poled crystal, providing an additional momentum that compensates the natural phase mismatch $\\Delta k = 2k_F - k_{\\text{SHG}} \\approx 1.416\\times10^6$ m$^{-1}$. The high modal overlap (about 80–86%) between the pump and second-harmonic polygon modes and the ultrahigh Q factors (loaded pump Q of $2.69\\times10^6$ and intrinsic Q of $3.86\\times10^7$) allow this quasi-phase-matched gain to accumulate to the reported efficiency.","core_discovery":"In an X-cut thin-film lithium niobate microdisk, a weak perturbation from a coupled tapered fiber recombines near-degenerate whispering-gallery modes into polygon modes with a square intensity profile. Because two sides of the square are aligned perpendicular to the crystal's optical axis, TE-polarized light experiences the large $d_{33}$ coefficient along those segments, and the effective nonlinear coefficient $d_{\\text{eff}}$ changes sign every half cycle of the round trip. The resulting natural quasi-phase matching supplies the momentum that compensates the phase mismatch between the 1560 nm pump and the 780 nm second harmonic, while the polygon modes sit away from the rough sidewall and retain intrinsic Q factors of about $3.86\\times10^7$ at the pump and $3.43\\times10^7$ at the harmonic. The authors report a measured SHG power of 2.211 mW from an on-chip pump power of 4.599 mW, giving an absolute conversion efficiency of 48.08%, with a linearly fitted normalized efficiency of 7.5% per milliwatt that they state is underestimated because it counts only the light collected by the tapered fiber.","pith_inferences":["If the on-chip power calibration is confirmed, the same polygon-mode scheme could be applied to other $\\chi^{(2)}$ processes such as spontaneous parametric down-conversion or optical parametric oscillation, where high Q, high overlap, and $d_{33}$ access could yield low-threshold sources.","The formation of polygon modes depends on the tapered-fiber perturbation, which suggests the method might transfer to other crystal cuts or nonlinear materials if the square pattern can be oriented to the dominant tensor component.","Since the reported efficiency was limited by collecting only the fiber-coupled SHG, a device with an optimized out-coupler or drop port could realistically approach or exceed the 52% PPLN record without domain inversion.","The linear growth of conversion efficiency with pump power reported here is typical of unsaturated SHG; testing at higher pump powers should reveal saturation or roll-off and would map the device's power-handling range."],"forward_implications":["Domain-inversion-free SHG in thin-film lithium niobate reaches efficiencies comparable to quasi-phase-matched PPLN microresonators (48.08% versus 52%) at sub-5 mW on-chip pump powers.","Polygon modes provide a route to access $d_{33}$ in X-cut lithium niobate without poling, avoiding domain-wall fabrication complexity and associated scattering loss.","The measured intrinsic Q factors, about $3.86\\times10^7$ at 1560 nm and $3.43\\times10^7$ at 780 nm, make the polygon-mode microdisk a strong platform for other cavity nonlinear optics at low power.","Because only the SHG collected by the tapered fiber was counted, the authors' fitted 7.5%/mW normalized efficiency is a lower bound, implying the internal conversion efficiency is higher than the measured 48.08%."],"supporting_citations":[{"why":"This reference supplies the natural quasi-phase-matching concept in X-cut monocrystalline microdisks and the expression for the effective nonlinear coefficient that the polygon-mode scheme exploits.","marker":"[15]"},{"why":"This reference reports the earlier natural quasi-phase-matching result whose low modal overlap limited efficiency, and serves as the baseline this work improves on.","marker":"[19]"},{"why":"This reference demonstrates the mode-phase-matching scheme with reverse-polarized double-layer TFLN microresonators, cited for the 23% domain-inversion-free efficiency this work surpasses.","marker":"[2]"},{"why":"This reference is cited alongside [2] and [19] as the domain-inversion-free SHG result at 30% efficiency that this work surpasses.","marker":"[13]"},{"why":"This reference reports the 52% SHG efficiency record in quasi-phase-matched PPLN microresonators that this work approaches.","marker":"[23]"},{"why":"This reference provides the fabrication technique used to create the monocrystalline TFLN microdisks.","marker":"[24]"},{"why":"This reference shows that weak perturbation recombines quasi-degenerate whispering-gallery modes into polygon modes, the formation mechanism underlying the experiment.","marker":"[25]"},{"why":"This reference gives the theoretical treatment of mode recombination under weak perturbation used to explain polygon-mode formation.","marker":"[26]"}],"fun_headline_variants":["Microdisk doubles light's frequency with 48% efficiency","48% SHG from polygon modes in a lithium niobate microdisk","Polygon modes enable 48% frequency doubling in microdisk","Monocrystalline microdisk hits 48% SHG efficiency","48% conversion efficiency from microdisk polygon modes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The 48.08% absolute efficiency depends on the unshown calibration that the on-chip pump power was 4.599 mW and that the collected 2.211 mW of second harmonic represents the generated signal; if coupling or collection was misestimated, the efficiency would change.","fun_headline_variants_meta":{"raw":{"variants":["Microdisk doubles light's frequency with 48% efficiency","48% SHG from polygon modes in a lithium niobate microdisk","Polygon modes enable 48% frequency doubling in microdisk","Monocrystalline microdisk hits 48% SHG efficiency","48% conversion efficiency from microdisk polygon modes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000757,"raw_usage":{"total_tokens":3411,"prompt_tokens":1038,"completion_tokens":2373,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":654,"completion_tokens_details":{"reasoning_tokens":2288}},"tokens_in":654,"tokens_out":2373,"duration_ms":15064,"temperature":1.0,"reasoning_tokens":2288,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T10:48:17.361438+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the fiber-to-microdisk coupling efficiency independently (for example, by fitting the transmission dip or adding a drop port) and collect the second harmonic emitted in all directions with an integrating sphere; if the true on-chip pump power is below 4.599 mW or the total generated SHG differs from the collected 2.211 mW, the reported 48.08% absolute conversion efficiency would not hold.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference supplies the natural quasi-phase-matching concept in X-cut monocrystalline microdisks and the expression for the effective nonlinear coefficient that the polygon-mode scheme exploits."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference reports the earlier natural quasi-phase-matching result whose low modal overlap limited efficiency, and serves as the baseline this work improves on."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference demonstrates the mode-phase-matching scheme with reverse-polarized double-layer TFLN microresonators, cited for the 23% domain-inversion-free efficiency this work surpasses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference is cited alongside [2] and [19] as the domain-inversion-free SHG result at 30% efficiency that this work surpasses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference reports the 52% SHG efficiency record in quasi-phase-matched PPLN microresonators that this work approaches."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference provides the fabrication technique used to create the monocrystalline TFLN microdisks."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"This reference shows that weak perturbation recombines quasi-degenerate whispering-gallery modes into polygon modes, the formation mechanism underlying the experiment."},{"cited_title":"Farajollahi, Z","cited_arxiv_id":null,"evidence_quote":"This reference gives the theoretical treatment of mode recombination under weak perturbation used to explain polygon-mode formation."}],"review_version":1}