{"id":"7cff434e-0b8f-4c9e-ae00-8c9cb11159de","arxiv_id":"2607.03690","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"low","formal_verification":"none","parameter_count":5,"one_line_summary":"Monolithic BTOI photonic crystals and Fabry–Pérot cavities reach ~0.5 dB/cm loss, Q up to 230k, reff ≈ 154 pm/V, and ~11 GHz material-limited EO bandwidth at the PhC band edge.","lead":"Researchers fabricated low-loss monolithic photonic crystals and high-Q cavities in barium titanate thin films, then measured electro-optic modulation up to ~11 GHz. This matters because BTO’s large Pockels effect plus subwavelength nanostructures could shrink high-speed optical modulators for computing and communications.","discovery_kind":"new_method","skeptic_critique":{"model":"grok-4.5","headline":"No significant objection identified beyond the reader's already-flagged RF packaging caveat.","rationale":"The paper's primary contribution is the demonstration of a low-loss anisotropic etch that enables high-contrast monolithic BTOI photonic crystals and high-Q FP cavities, together with a measured reff consistent with prior hybrid devices. These results rest on SEM, transmission spectra, SHG poling verification, and DC tuning, all of which are internally consistent and do not rely on the contested RF packaging. The bandwidth claim is secondary and already caveated by the authors' own S11 data and smoothing; the reader correctly flags it without over-weighting it into a rejection. No additional load-bearing flaw (e.g., in the reff extraction formula, Q extraction, or bandgap contrast) rises to the same level. Therefore the CONDITIONAL verdict stands unchanged.","tokens_in":15037,"tokens_out":450,"duration_ms":4198,"concrete_test":"Re-measure S21 on the same PhC band-edge device after replacing the 6 µm-gap, 200 nm Au contacts with a coplanar waveguide geometry whose |S11| remains <–10 dB to 25 GHz; if the 3 dB point still sits near 11 GHz and tracks the published r42(f) curve, the material-limited claim is confirmed; if it moves substantially higher, packaging was dominant.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The reader's weakest assumption correctly isolates the softest point: that the PhC-band-edge S21 roll-off (Fig. 4b) is material-limited by BTO r42 rather than electrode parasitics. The paper itself reports large S11, contact ringing, and LOESS smoothing, so the material interpretation is only partially isolated. That caveat is already reflected in the CONDITIONAL verdict and does not undermine the fabrication/optical core (anisotropic etch, >40 dB PhC contrast, Qi~1M, Q_L=230k FP, reff~154 pm/V). No stronger internal inconsistency or hidden assumption that would collapse the central claim was found.","agreement_with_reader":"agree"},"referee_report":{"model":"grok-4.5","summary":"The manuscript reports an optimized anisotropic dry-etch process for monolithic barium titanate-on-insulator (BTOI) that enables low-loss 1-D photonic crystals and photonic-crystal Fabry–Pérot cavities. Racetrack resonators reach intrinsic Q near 1 million and propagation loss of ~0.45 dB/cm; PhCs show >40 dB bandgap contrast; FP cavities reach loaded Q up to 230k. After SHG-verified poling, DC tuning yields reff ≈ 154 pm/V (racetracks) and ~145 pm/V (FPs). Microwave modulation at a PhC band edge gives a 3 dB EO bandwidth of 11 GHz (6 dB at 21 GHz), interpreted as material-limited by the frequency dependence of r42, together with demonstrations of sideband-resolved/unresolved resonator modulation, quasi-single-sideband generation, and high-power asymmetric comb-like spectra.","tokens_in":15218,"tokens_out":919,"duration_ms":21345,"significance":"Monolithic, high-contrast, low-loss nanostructures have been a missing capability for BTOI relative to LNOI. Demonstrating Qi ~1M, >40 dB PhC contrast, and Q_L = 230k FP cavities with a usable reff ~150 pm/V is a concrete materials-and-process advance that strengthens BTOI as a CMOS-compatible EO platform. The optical metrics are supported by spectra, length-extrapolated loss with error bars, SEM sidewall angle, and SHG poling maps; reff is extracted via a standard formula with stated parameters. These results open a path to compact, low-energy modulators, dispersion-engineered devices, and visible-wavelength EO components, even if the absolute bandwidth claim remains partially RF-packaging limited.","major_comments":[{"comment":"The central claim that the PhC-band-edge S21 roll-off (Fig. 4b) is set by BTO’s intrinsic r42 frequency dependence rather than cavity lifetime is only partially isolated from electrode parasitics. The paper reports large S11 reflections, contact ringing, and LOESS smoothing of S21 (Fig. 4b and Methods). A de-embedded or better-matched electrode measurement (or a clear upper bound on contact-limited bandwidth independent of the optical device) is needed before the 11 GHz / 21 GHz numbers can be presented as material-limited without qualification.","section":null},{"comment":"Eq. (1) for reff depends on a simulated electro-optic overlap Γ_eo = 0.51 (Supplementary Fig. 3) and on n_eff = 1.9. The manuscript should state the uncertainty on Γ_eo (mesh, index, electrode geometry) and show that the extracted 154 pm/V remains consistent under reasonable variation of these inputs; otherwise the comparison to literature values is less robust than claimed.","section":null}],"minor_comments":[{"comment":"Fig. 4b caption and Methods should explicitly state the LOESS parameters and whether any contact de-embedding was applied, so the smoothed curve can be reproduced.","section":null},{"comment":"Inconsistent notation for lattice constant (Λ vs a) appears in the Fig. 2 caption and body text; standardize.","section":null},{"comment":"Fig. 1d y-axis includes negative loss values; clarify that these are fit residuals or adjust the plot range.","section":null},{"comment":"The abstract and introduction cite bulk r42 > 1200 pm/V; a brief sentence contrasting bulk vs thin-film effective coefficients would help non-specialist readers.","section":null},{"comment":"Methods: state the number of devices/resonances averaged for the Qi ~1M and loss figures, and whether the 230k FP Q is a single peak or representative.","section":null},{"comment":"Typographical inconsistencies (e.g., “Fabry–Pérot” vs “Fabry-Perot”, missing spaces around units) should be cleaned in a revision.","section":null}],"recommendation":"minor_revision","confidential_remarks":"The fabrication and optical-quality core is solid and publishable; the RF-bandwidth interpretation is the softest point but is already partially flagged by the authors. Minor revision requiring clearer qualification of the material-limited claim (and Γ_eo uncertainty) is appropriate. Fit for a strong optics/photonics journal is good if the packaging caveat is handled transparently."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"This is a clean experimental device paper that delivers what the title promises. The new piece is a working anisotropic dry etch for monolithic BTOI that supports ~100 nm features, >40 dB PhC bandgaps, and PhC Fabry–Pérot cavities with loaded Q up to 230k, on top of racetracks that reach Qi ~1M and ~0.5 dB/cm. That combination was missing; prior BTO work was mostly hybrid Si/SiN stacks or simpler monolithic rings/MZIs. They also show the expected reff ~154 pm/V after SHG-verified poling and a set of modulation demos (sideband-resolved, band-edge quasi-SSB, comb-like spectra under strong drive).\n\nWhat they do well is straightforward: SEM sidewall angle (~75°), racetrack length extrapolation with error bars, Lorentzian Q fits, SHG poling maps, and a standard reff extraction with stated Γ_eo. The methods section is usable. The optical and DC-EO core looks solid and reproducible enough that other groups can build on the process.\n\nThe soft spot is exactly the one the reader flagged and the stress-test confirmed: the claim that the PhC-band-edge S21 (3 dB ~11 GHz) is material-limited by r42 roll-off. They report large S11, contact ringing, and LOESS smoothing, so electrode parasitics are not cleanly de-embedded. That weakens the strongest bandwidth interpretation but does not touch the fabrication or optical results. Energy-consumption language is aspirational; they do not report VπL or fJ/bit. Those are minor relative to the platform advance.\n\nThis is for people building EO modulators, slow-light devices, or visible/quantum components who need a higher-Pockels CMOS-compatible film with subwavelength patterning. Citation pattern is appropriate. Math and data are standard experimental practice, not circular. I would send it to referees; it deserves a serious look and will likely improve with tighter RF packaging discussion. Worth reading if you work in the area; I would cite the etch/PhC results.","headline":"Solid experimental platform paper: first real low-loss monolithic BTOI 1-D PhCs and 230k FP cavities, with reff~154 pm/V; RF bandwidth claim is only partially isolated from contacts.","tokens_in":15884,"tokens_out":540,"would_cite":true,"duration_ms":4941,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"An anisotropic etch yields low-loss monolithic barium titanate photonic crystals and high-Q cavities with material-limited electro-optic bandwidth near 11 GHz.","keywords":["barium titanate-on-insulator","monolithic nanophotonics","photonic crystals","Fabry–Pérot cavities","Pockels effect","electro-optic modulation","high-Q resonators","ferroelectric domain poling"],"falsifier":"A re-measurement of the same band-edge modulator with improved 50-ohm electrode geometry that drives S11 well below –10 dB across 1–25 GHz; if the 3 dB electro-optic bandwidth then moves substantially higher than 11 GHz, the material-limited interpretation is false.","tokens_in":15901,"feed_emoji":"⚡","tokens_out":1082,"duration_ms":9290,"temperature":0.7,"pith_summary":"Barium titanate on insulator has a very large Pockels coefficient, so it can make compact, low-voltage electro-optic modulators, but only if the optical mode sits inside the barium titanate itself and the film can be etched into sub-wavelength features without destroying optical quality. This paper shows that an optimized dry etch produces 75-degree sidewalls and low roughness, enough to support single-mode waveguides, one-dimensional photonic crystals, and Fabry–Pérot cavities entirely in the barium titanate film. Racetrack resonators reach intrinsic quality factors near one million and propagation losses of about 0.5 dB/cm; photonic crystals open bandgaps with more than 40 dB contrast; and photonic-crystal-mirror cavities reach loaded quality factors up to 230k. After poling, the effective electro-optic coefficient is about 154 pm/V. By modulating at a photonic-crystal band edge rather than a high-Q resonance, the authors measure a 3 dB electro-optic bandwidth of 11 GHz that tracks the known frequency roll-off of the material coefficient rather than cavity lifetime. The same platform also produces asymmetric sidebands, frequency-comb-like spectra under strong drive, and the usual sideband-resolved and unresolved resonant modulation regimes.","feed_headline":"Low-loss BTO photonic crystals hit 11 GHz electro-optic speed","feed_subtitle":"Anisotropic etch yields Q~230k cavities and material-limited modulation in monolithic barium titanate","key_machinery":"The anisotropic (≈75°) dry etch of thin-film barium titanate that simultaneously preserves high index contrast for photonic-crystal unit cells and low sidewall scattering, enabling both deep bandgaps and high-Q cavities in a single monolithic platform.","core_discovery":"A carefully optimized anisotropic dry etch of commercial barium titanate-on-insulator films produces low-loss monolithic nanostructures—one-dimensional photonic crystals with >40 dB bandgap contrast and photonic-crystal Fabry–Pérot cavities with loaded Q up to 230k—while racetracks reach intrinsic Q near 1 million and ~0.5 dB/cm loss. After domain poling, the effective Pockels coefficient is ~154 pm/V, and microwave modulation at the photonic-crystal band edge yields a material-limited 3 dB bandwidth of 11 GHz (6 dB at 21 GHz).","pith_inferences":["Because domain writing energy scales with the volume of material that is poled, the same nanostructures that raise light–matter interaction should also lower the energy cost of non-volatile ferroelectric phase shifters or reconfigurable photonic gates.","If contact parasitics can be eliminated, the residual material roll-off near 10 GHz becomes the next target for materials engineering (stoichiometry or strain) rather than device geometry.","The demonstrated feature sizes and Q values are already in the range needed for Purcell-enhanced coupling to visible solid-state emitters once the platform is extended below 800 nm."],"forward_implications":["Monolithic BTOI photonic crystals and Fabry–Pérot cavities can shrink modulator capacitance and switching energy relative to hybrid silicon or silicon-nitride overcladding designs.","Band-edge and cavity Q can be designed independently, giving a route to slow-light or bandwidth-engineered modulators whose speed is not forced by photon lifetime.","The same etch enables visible-wavelength electro-optic devices and dispersion-engineered cavities for nonlinear and quantum photonics on a CMOS-compatible film.","Asymmetric sideband generation and multi-line comb spectra at a sharp band edge become available as native functions of the platform without extra filtering optics."],"fun_headline_variants":["BTO photonic crystals hit material-limited 11 GHz EO bandwidth","Low-loss BTO PhCs with 230k-Q cavities and 11 GHz modulation","Monolithic BTO nanophotonics yield 11 GHz electro-optic speed","Anisotropic BTO etch enables 230k-Q PhC cavities and 11 GHz EO","BTO PhC band-edge modulation reaches 11 GHz 3-dB bandwidth"],"cache_read_input_tokens":128,"weakest_assumption_plain":"The claim that the measured 11 GHz roll-off is set by the intrinsic frequency dependence of barium titanate’s electro-optic coefficient, rather than by electrode reflections, contact parasitics, or optical filtering at the band edge.","fun_headline_variants_meta":{"raw":{"variants":["BTO photonic crystals hit material-limited 11 GHz EO bandwidth","Low-loss BTO PhCs with 230k-Q cavities and 11 GHz modulation","Monolithic BTO nanophotonics yield 11 GHz electro-optic speed","Anisotropic BTO etch enables 230k-Q PhC cavities and 11 GHz EO","BTO PhC band-edge modulation reaches 11 GHz 3-dB bandwidth"]},"model":"grok-4.5","effort":"low","cost_usd":0.00478,"raw_usage":{"total_tokens":1504,"prompt_tokens":965,"num_sources_used":0,"completion_tokens":99,"cost_in_usd_ticks":47800000,"prompt_tokens_details":{"text_tokens":965,"audio_tokens":0,"image_tokens":0,"cached_tokens":256},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":440,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":965,"tokens_out":99,"duration_ms":3844,"temperature":1.0,"reasoning_tokens":440,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-12T00:38:51.955769+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"A re-measurement of the same band-edge modulator with improved 50-ohm electrode geometry that drives S11 well below –10 dB across 1–25 GHz; if the 3 dB electro-optic bandwidth then moves substantially higher than 11 GHz, the material-limited interpretation is false.","supporting_citations":[],"review_version":1}