{"id":"fbbca27c-eeb4-40a4-984b-c76439b3f5be","arxiv_id":"2504.17980","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"A robust single-pulse poling recipe is demonstrated for thin-film lithium tantalate, and a pole-after-etch ridge waveguide converts 1550 nm to 775 nm with 208 %W^-1 cm^-2 normalized efficiency.","lead":"This paper reports a robust fabrication recipe for periodically poled thin-film lithium tantalate and uses it to build a waveguide that converts 1550 nm light to 775 nm with 208 %W^-1 cm^-2 normalized efficiency. The result strengthens lithium tantalate as a platform for quantum photonics, sensing, and spectroscopy.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (3) omits how the measured 178.7-nm poling depth enters, so the 208-vs-244 agreement relies on an unstated overlap model and a single destructive cross-section; the headline efficiency claim needs independent support.","rationale":"The reader's weakest_assumption identifies the same load-bearing point: the efficiency claim depends on a single measured poling depth and the unshown mode-overlap model behind Eq. (3). I agree with that assessment. The fabrication-robustness portion of the paper is comparatively well supported: Table I, SHM images, and SEM cross-sections give a systematic parameter sweep, and the rectangular-domain recipe is plausible, even if a quantitative rectangularity metric would be useful. The efficiency claim is the load-bearing quantitative result because the wording 'in line with the theoretical value' elevates the device from a demonstration to a validation of the process. With no error budget on 208 %, no statement of how the 178.7 nm depth enters Eq. (3), and only one destructive cross-section, the agreement with 244 % is a single uncheckable point. This does not warrant rejection: the measured efficiency can stand as a reported value, and the fabrication study is valuable. It does warrant conditional acceptance with the specific requirement that the authors provide the overlap-model parameters/code and multi-position or multi-device depth and efficiency data. Since the reader already assigned CONDITIONAL, my stress-test does not change the verdict.","tokens_in":9316,"tokens_out":11532,"duration_ms":120854,"concrete_test":"Recompute the theoretical Fig. 3(c) curve from the full 2D SHG overlap integral: for the reported geometry (500 nm film, ~300 nm etch, 61° sidewall, 3.908 µm period, 55% duty cycle), set d(x,y)=d33 only in the inverted-domain region and evaluate η at 178.7 nm depth. Also take two additional differential-etch cross-sections along the 1 cm device. If the recomputed theoretical efficiency differs from 244 %W^-1cm^-2 by more than ~15%, or the measured depth varies by more than ±10% along the device, the claimed 'in line with theory' agreement is not independently established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative headline—208 %W^-1cm^-2 'in line with' 244 %W^-1cm^-2—rests on Eq. (3), but Eq. (3) is not sufficient as printed. It gives ηtheory ∝ d_eff^2 A_sh/A_fh^2 and identifies d_eff with the nonlinear tensor; it does not state how the measured 178.7 nm poling depth enters. In the pole-after-etch device the nonlinear polarization exists only in a portion of the cross-section, so the proper theory is a 2D overlap integral over the poled region, not a global mode-area ratio with bulk d_eff. The depth dependence shown in Fig. 3(c) therefore depends on an unstated model. Moreover, the experimental value is a single device, the depth is measured on one post-measurement cross-section, and no propagation loss, duty-cycle variation along the 1 cm interaction length, or repeated-device statistics are reported. The 208/244 agreement could be coincidental or the product of a self-consistent but untested model; nothing currently tests whether that one cross-section and that simplified formula represent the whole interaction region.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports a systematic study of periodic poling in thin-film lithium tantalate (TFLT), identifying a single-pulse recipe (peak-voltage hold time of 10 ms or less, 90 s ramp-down, field strength around 50 kV/mm, and a 33% poling-finger fill factor) that produces rectangular domains across acoustic-grade and optical-grade films, various electrode metals, and oxide interlayers, as validated by second-harmonic microscopy and differential-etch SEM. The authors then apply this recipe in a pole-after-etch process to fabricate a ridge waveguide and report a normalized second-harmonic generation (SHG) efficiency of 208 %W^-1 cm^-2 at 1552 nm, which they state is in line with a theoretical value of 244 %W^-1 cm^-2 computed from Eq. (3).","tokens_in":9595,"tokens_out":5970,"duration_ms":62310,"significance":"If the central claims hold, the paper provides a transferable, robust poling process for thin-film lithium tantalate and demonstrates a working telecom-band frequency converter on that platform, which is significant for integrated nonlinear photonics and quantum photonics. The robustness claim is supported by Table I, which lists 11 chip variations with SHM and SEM confirmation of domain quality. The theoretical efficiency is not fitted to the measured value; it uses independently measured waveguide geometry, literature d33, and a measured poling depth, which is a strength. However, the quantitative efficiency comparison is underdocumented: Eq. (3) as printed does not connect the measured poling depth to the efficiency, and the experimental value lacks uncertainty and repetition, so the headline agreement needs revision before the claim is fully established.","major_comments":[{"comment":"The printed theoretical formula does not specify how the measured 178.7 nm poling depth enters the calculation. Equation (3) is a bulk, fully-poled expression: it contains only mode areas, refractive indices, and the material d_eff, with no overlap integral over the poled portion of the cross-section and no depth variable. The text states that the 244 %W^-1 cm^-2 value is calculated for the measured depth, and Fig. 3(c) plots efficiency versus depth, but the underlying model is not given. Please provide the explicit overlap integral or a complete simulation description, including how the partial-depth domain, 61-degree sidewall, duty cycle, and material anisotropy are handled, so that the 208-versus-244 agreement can be independently reproduced.","section":"Section III, Eq. (3) and Fig. 3(c)"},{"comment":"The experimental efficiency is reported as a single maximum value from one device with no uncertainty, and the paper does not report propagation losses, duty-cycle variation along the 1 cm interaction length, or repeated-device statistics. The poling depth is inferred from one differential-etch cross-section after the measurement, so the comparison with theory implicitly assumes that this cross-section and the simplified model represent the entire interaction region. Please provide error bars for the efficiency, a loss correction or an upper bound on propagation loss, and at least one additional device or cross-section to support the claim that the measured value is 'in line with' the theoretical value.","section":"Section III, Eq. (2), Fig. 3(a)"},{"comment":"The robustness claim is not fully quantified. Table I lists poling field strengths ranging from 41.5 to 70 kV/mm, and the text notes that higher field strengths are required for thicker oxide interlayers, which is at odds with the statement that 'the optimal poling recipe is insensitive to these variations.' Please specify the acceptance criteria used for each condition, such as duty-cycle tolerance, domain shape, and poling depth, and clarify whether the recipe is robust at fixed field strength or whether the field must be re-optimized for each stack.","section":"Section II, Table I"}],"minor_comments":[{"comment":"The phrase 'poled regionds' should read 'poled regions.'","section":"Table I caption"},{"comment":"The parameters A and B are defined in Fig. 1(b) but are used without restating their definitions in the text and in the captions of Fig. 1(e) and (f); please define them at first use in the results text.","section":"Section II, Fig. 1(b)-(f)"},{"comment":"The text refers to 'the fundamental TE mode at the second harmonic' when defining n_sh and A_sh; the second harmonic at 775 nm may not be the fundamental mode for that wavelength, so the wording should be corrected to 'the TE mode at the second-harmonic wavelength' or the mode order should be explicitly verified.","section":"Section III, Eq. (3)"},{"comment":"The experimental spectrum is plotted as a line without markers or error bars; please indicate the wavelength step, measurement uncertainty, and whether the line is a raw sweep or a fit, to allow the reader to assess the spectral shape and the uncertainty of the 208 %W^-1 cm^-2 peak.","section":"Section III, Fig. 3(a)"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: the real contribution is the systematic poling study, and it is solid enough on its own. Table 1 spans acoustic- and optical-grade film, several electrode metals, oxide interlayers, and both lithography routes, all giving near-50% duty cycle; SHM and SEM cross-sections back the domain-quality claims. The pulse recipe (short peak hold, 90 s ramp-down, ~50 kV/mm) is clearly described, and the empirical relation between ramp-down time and poling depth is exactly the kind of practical information the TFLT fabrication community needs.\n\nThe SHG part is a nice demonstration, but the quantitative headline is softer than it looks. Equation (3) is not sufficient as printed: the text says the theoretical efficiency uses the measured 178.7 nm poling depth, yet the depth never appears in the formula. So the 208 vs 244 %W^-1cm^-2 agreement rests on an unstated overlap model for a partially poled cross-section. Figure 3(c) plots efficiency versus depth, but the model behind that curve is not shown. That is more than cosmetic: if the calculation uses a global mode-area ratio rather than a 2D overlap integral over the poled region, the agreement could be fortuitous. On the experimental side there is one device, one measured cross-section, no propagation loss, no duty-cycle variation along the 1 cm interaction, no repeated-device statistics, and no error bars. Those are ordinary for a first demonstration, but they mean the efficiency claim should be treated as promising, not confirmed.\n\nMinor soft spots: the robustness to oxide interlayers is qualified by the need for higher field with thicker or ambient-annealed oxide, and the data are only available on request. Neither is disqualifying.\n\nBottom line: take the robustness recipe at face value. Treat the numerical efficiency as a target for follow-up, and ask the authors to show the overlap integral, define the mode areas in the partially poled geometry, and give measurement uncertainty. This paper deserves serious peer review; it is a useful experimental contribution for the TFLT community. Send it out.","headline":"A credible robustness recipe for PP-TFLT with a useful fabrication study, paired with an SHG efficiency claim that is plausible but quantitatively under-supported until the overlap model is made explicit.","tokens_in":10168,"tokens_out":2004,"would_cite":true,"duration_ms":23387,"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 single high-voltage pulse with a 10 ms peak and 90 s ramp-down creates stable rectangular poling domains in thin-film lithium tantalate, and a pole-after-etch waveguide converts telecom light with 208 %W⁻¹cm⁻² efficiency.","keywords":["thin-film lithium tantalate","periodic poling","second-harmonic generation","quasi-phase matching","telecom-band frequency conversion","ferroelectric domain engineering","ridge waveguide","pole-after-etch"],"falsifier":"Measure SHG from identically fabricated waveguides with poled lengths of, say, 0.5, 1.0, and 1.5 cm, and etch multiple cross-sections along each device; if the efficiency does not scale as $\\eta \\propto L^2$ or the domain depth varies along the ridge, the claimed $208\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ is not a stable property of the whole device.","tokens_in":9129,"feed_emoji":"⚡","tokens_out":11974,"duration_ms":104939,"temperature":0.7,"pith_summary":"The paper establishes a single-pulse poling recipe for thin-film lithium tantalate that is transferable across wafer grade, electrode metal, and oxide interlayer: a fast ramp to about 500 V, a peak-voltage time of 10 ms or less, and a 90 s ramp-down produce rectangular inverted domains with near-50% duty cycle. The pulse shape is the core of the process, since the long ramp-down prevents domain back-switching and the short flat top avoids triangular domains. Using this recipe on a ridge waveguide etched before poling, the authors achieve second-harmonic generation from 1550 nm to 775 nm with normalized conversion efficiency $208\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$, matching the $244\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ predicted from the measured 178.7 nm poling depth. This matters because periodic poling of TFLT, previously a material-sensitive step, becomes a one-shot process for a platform with lower photorefraction and birefringence than lithium niobate.","feed_headline":"One poling pulse makes lithium tantalate double telecom light","feed_subtitle":"The same pulse works across film grades and metals, hitting 208 % per W cm² conversion from 1550 to 775 nm","key_machinery":"The load-bearing mechanism is the poling pulse waveform: a linear ramp-up to roughly 500 V, a flat top of 10 ms or less, and a linear 90 s ramp-down starting from half the peak voltage. The long ramp-down stabilizes inverted domains against back-switching, while the short flat top prevents the asymmetric triangular domains seen at longer hold times. The secondary mechanism is the pole-after-etch flow, in which the waveguide is etched first and the poling period is set from measured geometry by $\\Lambda = 2\\pi/(k_{\\mathrm{sh}} - 2k_{\\mathrm{fh}})$, absorbing thickness and etch-depth variations before poling. The theoretical benchmark is $\\eta_{\\mathrm{theory}} = 2\\omega^2 d_{\\mathrm{eff}}^2/(n_{\\mathrm{fh}}^2 n_{\\mathrm{sh}} \\varepsilon_0 c^3) \\cdot A_{\\mathrm{sh}}/A_{\\mathrm{fh}}^2$, evaluated at the measured poling depth; this supplies the target the experiment must match.","core_discovery":"The central discovery is that domain shape and depth in thin-film lithium tantalate are controlled primarily by the poling pulse envelope rather than by material specifics. With poling fields near 50 kV/mm and fingers occupying about one-third of the poling period, the same single pulse gives rectangular domains of roughly 50% duty cycle in acoustic-grade and optical-grade films, with Cr, Ni, Ti, or NiCr electrodes, and with or without an oxide interlayer. In the pole-after-etch waveguide, the pulse yields 178.7 nm deep domains across a 1 cm region, and the measured SHG efficiency of $208\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ is close to the theoretical $244\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ computed from that depth. The authors note that full-depth poling in the same geometry would theoretically raise the peak efficiency to $2314\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$.","pith_inferences":["The pulse recipe should transfer to other quasi-phase-matched processes in TFLT, such as spontaneous parametric down-conversion or sum-frequency generation, since the poling itself does not care about the target wavelength.","A practical production question left open by the paper is whether the 90 s ramp-down can be shortened without losing domain depth; if not, it sets the throughput limit for wafer-scale poling.","If the shallow 178.7 nm domain depth is the main efficiency bottleneck, then applying this same pulse recipe with sidewall poling or poling before the ridge etch should approach the full-depth theoretical efficiency of $2314\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$."],"forward_implications":["The same poling pulse works for acoustic- and optical-grade films, Cr/Ni/Ti/NiCr electrodes, and with or without an oxide interlayer, so processes do not need to be retuned for each material variant.","In the pole-after-etch flow, the poling period is computed from measured post-etch geometry, so thickness and etch-depth variations are absorbed into the period and quasi-phase matching is accurate on the first try.","A 1 cm PP-TFLT ridge waveguide converts 1550 nm light to 775 nm with normalized efficiency $208\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$, matching the theoretical $244\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ for the measured domain depth.","Duty cycle is controlled by the poling peak voltage, giving a direct tuning knob for nonlinear device performance.","For the same waveguide geometry, full-film-depth poling would theoretically raise the peak conversion efficiency to $2314\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$."],"supporting_citations":[{"why":"Previous demonstration of periodic poling in thin-film lithium tantalate that this paper systematically varies and extends.","marker":"[9]"},{"why":"Earlier continuous-wave second-harmonic generation in PP-TFLT, the baseline this device's conversion efficiency advances.","marker":"[10]"},{"why":"Introduces the pole-after-etch fabrication flow used here to set the poling period from measured post-etch geometry.","marker":"[21]"},{"why":"Bulk lithium tantalate poling study whose pulse-shaping insight motivates the long ramp-down time for domain stabilization.","marker":"[25]"},{"why":"PP-TFLN poling study that supports the roughly one-third finger-fill factor used to reach near-50% duty cycle.","marker":"[24]"},{"why":"Standard nonlinear-optics reference supplying the quasi-phase-matching and conversion-efficiency framework behind the theoretical formula.","marker":"[22]"}],"fun_headline_variants":["Single pulse poling gives robust telecom doubling in thin-film LiTaO3","One pulse works across grades: robust PP-TFLT SHG at 1550 nm","208% W−1cm−2 telecom doubling from one robust poling pulse","Single pulse, any electrode: robust thin-film LiTaO3 frequency doubler"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The efficiency match assumes the 178.7 nm poling depth observed on one etched cross-section, together with the simplified mode-overlap ratio in the theoretical formula, represents the entire 1 cm waveguide.","fun_headline_variants_meta":{"raw":{"variants":["Single pulse poling gives robust telecom doubling in thin-film LiTaO3","One pulse works across grades: robust PP-TFLT SHG at 1550 nm","208% W−1cm−2 telecom doubling from one robust poling pulse","Single pulse, any electrode: robust thin-film LiTaO3 frequency doubler"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000871,"raw_usage":{"total_tokens":3775,"prompt_tokens":954,"completion_tokens":2821,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":570,"completion_tokens_details":{"reasoning_tokens":2733}},"tokens_in":570,"tokens_out":2821,"duration_ms":20541,"temperature":1.0,"reasoning_tokens":2733,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:27:47.309746+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure SHG from identically fabricated waveguides with poled lengths of, say, 0.5, 1.0, and 1.5 cm, and etch multiple cross-sections along each device; if the efficiency does not scale as $\\eta \\propto L^2$ or the domain depth varies along the ridge, the claimed $208\\,\\%\\mathrm{W}^{-1}\\mathrm{cm}^{-2}$ is not a stable property of the whole device.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Introduces the pole-after-etch fabrication flow used here to set the poling period from measured post-etch geometry."},{"cited_title":"Xue , author X","cited_arxiv_id":null,"evidence_quote":"Bulk lithium tantalate poling study whose pulse-shaping insight motivates the long ramp-down time for domain stabilization."},{"cited_title":"Suntsov , author C","cited_arxiv_id":null,"evidence_quote":"PP-TFLN poling study that supports the roughly one-third finger-fill factor used to reach near-50% duty cycle."}],"review_version":1}