Pith. sign in

REVIEW 3 major objections 6 minor 31 references

Giant electro-optic coefficient in single crystal barium titanate on oxide insulator based Mach-Zehnder interferometer

T0 review · 3 major / 6 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A wet-etched barium titanate thin-film waveguide reaches a Pockels coefficient of 1268 pm/V, the largest reported in thin-film BTO and close to the bulk value.

desk verdict The paper claims a record thin-film BTO Pockels coefficient, but the extraction model has an unvalidated scalar-confinement approximation that could move the number; the claim should be conditional on a full overlap re-analysis. read the letter →

arxiv 2505.21927 v2 pith:DRZE3HGN submitted 2025-05-28 physics.optics physics.app-ph

Hong-Lin Lin , Pragati Aashna , Yu Cao , Aaron Danner This is my paper
classification physics.opticsphysics.app-ph
keywords bariumtitanatePockelseffectelectro-opticcoefficientMach-Zehnderinterferometerthin-filmphotonicswetetchingmodeconfinementintegrated
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper reports the largest electro-optic (Pockels) coefficient measured in a thin-film barium titanate (BTO) platform: r42 = 1268 pm/V, close to the bulk single-crystal value of 1300 pm/V. The value is extracted from the voltage-induced resonance shift of an unbalanced Mach-Zehnder interferometer made from a wet-etched single-crystal BTO ridge on an oxide (DSO) insulator, in which the fundamental optical mode carries 74.3% of its power inside the BTO. The authors argue that full optical confinement in the active electro-optic layer is what allows the thin-film material to exhibit near-bulk Pockels strength, in contrast to hybrid silicon or silicon nitride waveguides that interact with BTO only through evanescent overlap. If correct, this would make BTO-on-insulator a practical platform for compact, low-voltage, high-speed modulators and other nonlinear photonic devices.

What carries the argument

The extraction rests on the Pockels-induced perturbation of BTO's permittivity tensor, given in Supplementary Eq. (2): an applied DC field E adds off-diagonal terms -Γ_BTO ($n_o^{2}$ $n_e^{2}$ r42 E) to the otherwise diagonal relative permittivity, where Γ_BTO is the simulated optical power confinement in the BTO layer and n_o, n_e are the ordinary and extraordinary indices. Diagonalizing the reduced 2x2 matrix via a rotation matrix yields voltage-dependent principal indices, which feed a theoretical MZI transmission spectrum (Eq. 6) whose resonance shift is matched to the measured 41.3 pm/V shift to determine r42. The load-bearing step is treating the mode-overlap factor as a scalar multiplier on the off-diagonal Pockels term.

What would settle it

Fabricate the same MZI geometry with a different etch depth (for example 150 nm instead of 250 nm) and extract r42; if the scalar-confinement model is adequate, the extracted value should stay near 1268 pm/V, while a model error would show up as a systematic shift with geometry. Alternatively, measure the half-wave voltage Vπ of a straight BTO phase shifter with independently known field and overlap and compare the implied r42 to 1268 pm/V.

Watch

Extended reading notes

Core claim

On the paper's own terms: single-crystal BTO grown by pulsed laser deposition on DSO can be wet-etched into single-mode ridge waveguides without degrading the electro-optic response, and an unbalanced MZI built from such a waveguide tunes at 41.3 pm/V under applied voltage. From that tuning efficiency, combined with COMSOL cross-sectional mode and field simulations and an interference-spectrum fit, the paper extracts r42 = 1268 pm/V, which is very close to the bulk value of 1300 pm/V and, to the authors' knowledge, the largest reported for thin-film BTO. The result is attributed to the 74.3% optical mode confinement in the BTO ridge, which lets the strong material Pockels effect be used directly rather than diluted by evanescent coupling in hybrid platforms.

Load-bearing premise

The measured giant coefficient depends on the assumption that the optical confinement factor can be applied as a simple multiplier to the electro-optic effect in the waveguide; if the real light and electric-field distributions do not behave that way, the extracted number could be systematically off.

Editorial extensions

If this is right

  • If correct, BTO ridge waveguides can combine near-bulk Pockels strength with strong optical confinement, implying a much lower voltage-length product than lithium niobate modulators.
  • The wet-etch process (0.03% HF, 250 nm etch in a 500 nm film) is simple and avoids dry-etch damage, so it could be adopted for other BTO-on-insulator photonic circuits.
  • Device designs based on partial evanescent overlap with BTO become unnecessary; full-confinement geometries can exploit the material's full r42.
  • The 41.3 pm/V tuning efficiency in this specific MZI geometry could be scaled up by reducing the electrode gap or increasing the electrode length.
  • BTO-on-insulator becomes a credible route to ultra-low-power, compact Mach-Zehnder modulators for datacom and coherent systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • If r42 is genuinely bulk-like in this thin-film form, then at a fixed voltage a fully confined BTO phase shifter should beat lithium niobate (r33 ≈ 30 pm/V) by roughly an order of magnitude in phase efficiency, barring microwave-optical velocity mismatch and loss.
  • The scalar-Γ_BTO assumption could be tested by measuring r42 at several etch depths or waveguide widths and checking that the extracted value does not shift systematically with geometry.
  • The same platform may also enable electro-optic polarization rotation (using r51) or tuning of microring resonators, where high confinement in the active layer directly improves tuning efficiency.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This manuscript reports a Mach-Zehnder interferometer fabricated in single-crystal barium titanate-on-insulator (BTO/DSO) using an optimized wet-etching process, and extracts a Pockels coefficient of r42 = 1268 pm/V from a measured resonance tuning efficiency of 41.3 pm/V. The device consists of a 500 nm BTO film with a 250 nm wet-etched ridge, an electrode length of 3500 μm, and an arm length difference of 430 μm. The fundamental TE mode has a reported 74.3% power overlap with the BTO layer and an effective index of 2.066. The Pockels coefficient is obtained by matching the measured voltage-dependent transmission spectrum to a model described in Supplementary Note S1, where a trial r42 is used to modify the BTO permittivity matrix and the resulting MZI spectrum is simulated. The paper compares the achieved coefficient and mode overlap with prior BTO devices and claims the largest reported thin-film BTO Pockels coefficient, close to the bulk value of 1300 pm/V.

Significance. If the extraction is robust, the result is significant: it would demonstrate that single-crystal thin-film BTO can retain essentially the bulk Pockels coefficient while simultaneously providing strong optical confinement in the active layer, addressing a known limitation of evanescent BTO-on-Si or BTO-on-SiN devices. The wet-etched single-crystal platform, the 2.52 nm RMS sidewall/surface roughness, the linear tuning data, and the comparative Table 1 are useful contributions. However, the central record claim is the output of a fitting model that relies on two simplifications in Supplementary Eq. (2) and Supplementary Eq. (6), and the manuscript reports no error bars, no repeated devices, and no control measurements. Because the quoted value is only about 2.5% below the bulk value, systematic model bias of even a few percent is decisive for the record claim. The paper would be strengthened by a full vectorial overlap re-extraction and an explicit comparison of the modeled and measured free spectral range.

major comments (3)
  1. [Supplementary Note S1, Eq. (2)] The extraction multiplies the off-diagonal Pockels term of the BTO permittivity matrix by the scalar power confinement factor Γ_BTO = 0.743. For a ridge waveguide mode, the first-order change in effective index is Δn_eff = (c ε0 / 2P) ∫ E_opt*(r) · Δε(r) · E_opt(r) dA, and the weighting function for an off-diagonal perturbation is proportional to the cross-term between orthogonal field components (e.g., E_y*E_z), not to the Poynting-vector power fraction defined in Eq. (1). These two weights can differ substantially for a strongly guided ridge mode, and the difference affects the extracted r42 directly. Since r42 is the free parameter used to match the measured 41.3 pm/V tuning, the quoted value is model-dependent. The authors should re-extract r42 using the full vectorial overlap integral with the simulated DC field distribution and report how the result changes.
  2. [Supplementary Note S1, Eq. (6)] The MZI output field expression uses bare material indices n_a, n_b, and n_o in all phase terms, but the phase accumulated by a waveguide arm should be computed with the modal effective index, which the authors report as n_eff = 2.066 at 0 V in Fig. 2b. In particular, the reference-arm term exp(−i2π n_o (L+D)/λ) ignores waveguide dispersion and mode-index reduction, which changes both the modeled FSR and the mapping from the 0.288 nm shift at 5 V to r42. The measured FSR of 2.563 nm should be used as a consistency check: if the model with the reported geometry cannot reproduce the measured FSR using modal effective indices, then the fitted r42 of 1268 pm/V is not reliable. This issue is load-bearing because the claimed value is only about 2.5% below the bulk value, so even a few percent bias in the model changes the record claim.
  3. [Main text, Figs. 2c–2d and Methods] The tuning efficiency and the extracted r42 are based on a single device measured at DC voltages from 0 to 20 V in 5 V steps. No error bars, repeated devices, sign-reversal measurement, or thermal control are reported. For a claim of the largest reported Pockels coefficient in thin-film BTO, repeated measurements on several devices and a quantitative estimate of systematic uncertainty arising from the model simplifications in Eq. (2) and Eq. (6) are necessary. A sign-reversal check would also help confirm that the observed shift is the linear electro-optic effect rather than a slow pyroelectric, bolometric, or charge-related artifact.
minor comments (6)
  1. [Fig. 1 caption] The caption states '45° angle view (b); top view (c)', but panel b shows the AFM image, while panels c and d show SEM images; the correspondence between panels and views should be corrected.
  2. [Table 1, row 7] Row 7 is labeled 'SiN ridge' but cites reference 10 (Tang et al., BaTiO3 thin-film waveguide modulator), which does not appear to describe a SiN ridge waveguide; the entry should be checked against the cited reference.
  3. [References] References 27 and 31 are the same paper (Gill et al., Applied Physics Letters 71, 1783) listed twice under different numbers; one duplicate should be removed.
  4. [Main text, substrate size] The text says '100 mm2 DSO substrate', which is ambiguous; this should be written as '100 mm²', '10 mm × 10 mm', or the intended wafer size in standard units.
  5. [Throughout] There are several typographical errors, including 'availabe' in the introduction, 'wet-eching' on page 2, and 'respextively' in the Fig. 3 caption; these should be corrected during revision.
  6. [Fig. 2b / measured FSR] The measured FSR of 2.563 nm and the simulated effective index of 2.066 are not explicitly reconciled in the text; a short discussion comparing the measured FSR with the model would support the extraction procedure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: r42 = 1268 pm/V is obtained by standard parameter extraction from measured MZI resonance shifts; self-citations to prior growth/etching work are not load-bearing.

full rationale

The central claim is an extracted Pockels coefficient, and the extraction chain is not circular. The measured transmission spectrum supplies the tuning efficiency (41.3 pm/V), and Supplementary Note S1 varies a trial r42 in Eqs. (2) and (6) until the modeled resonance shift matches the measured shift; r42 is therefore an output of a fit, not an input or a renamed prediction. The scalar Gamma_BTO weighting in Eq. (2) and the use of material indices in Eq. (6) are modeling simplifications that could bias the fitted value, but that is a numerical/experimental correctness and uncertainty concern, not a circularity: no equation is defined in terms of the quantity it purports to determine, and no fitted parameter is relabeled as an independent prediction. The self-citations (Refs. 13 and 15 for PLD growth and wet etching, Ref. 29 for a prior BTO ridge device) are process, design, and comparison references; they do not supply the r42 result. The comparison with the external bulk value of 1300 pm/V (Ref. 9) is used only as context, not as an input to the extraction. No circular step can be exhibited from the paper's own equations or citation chain.

Assumptions & free parameters 1 free parameters · 5 assumptions · 0 invented entities

The central extraction depends on one fitted parameter (r42), four domain assumptions about the electro-optic model, crystal orientation, field simulation, and absence of parasitic effects, and a standard interference formula.

free parameters (1)
  • r42 (Pockels coefficient) = 1268 pm/V
    Trial r42 values are used in the simulation and the value that matches the measured 5 V resonance shift (0.288 nm, 41.3 pm/V tuning) is selected. This is the central measured output, not an unconstrained fit.
assumptions (5)
  • domain assumption The electro-optic permittivity change is described by SI Eq. (2), with gamma_BTO multiplying the off-diagonal Pockels term.
    This perturbation-style treatment assumes the mode confinement factor can be applied directly to the material index change; it is used to extract r42.
  • domain assumption The BTO film is single-crystal with c-axis out of plane, enabling r42 through coplanar electrodes.
    Verified in the paper, but the crystal quality and orientation determine the validity of the r42 extraction.
  • domain assumption The measured resonance shift is entirely due to the Pockels effect.
    No control measurements are shown to exclude thermo-optic or piezoelectric contributions to the observed shift.
  • domain assumption The DC electric field distribution is accurately given by the COMSOL simulation (3.49 kV/cm effective).
    Any error in the simulated field directly scales the extracted r42.
  • standard math The MZI interference formula (SI Eq. 6) correctly accounts for the polarization rotation and phase accumulation.
    Standard Jones matrix treatment, but assumes no depolarization or loss differences between arms.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Giant electro-optic coefficient in single crystal barium titanate on oxide insulator based Mach-Zehnder interferometer." pith.science (2026). https://pith.science/paper/DRZE3HGN

@misc{pith2026250521927,
  author       = {Pith},
  title        = {Pith review of: Giant electro-optic coefficient in single crystal barium titanate on oxide insulator based Mach-Zehnder interferometer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DRZE3HGN}},
  note         = {Machine review of arXiv:2505.21927}
}
read the original abstract

Electro-optic modulators are indispensable components of modern day photonic integrated circuits (PICs). Recently lithium niobate has emerged as a key material to realize large-bandwidth high-speed modulation, but next-generation modulators require high-density integration, low cost, low power and high performance simultaneously, which are difficult to achieve with established integrated lithium niobate photonics platforms due to its limited electro-optic coefficient. Leveraging its exceptional Pockels coefficient, barium titanate (BTO) in the thin film form has emerged as a promising alternative but the electro-optic coefficients reported in thin-film BTO often fall short of bulk values due to challenges in film growth and waveguide fabrication. Here, we report, to the best of our knowledge, the largest Pockels coefficient (r42) of 1268 pm/V in thin film BTO platform, which is very close to the bulk value. We measure it by using an unbalanced Mach-Zehnder interferometer, fabricated by an optimized wet-etching method for realising single-mode waveguides in single-crystal barium titanate-on-insulator grown by pulsed laser deposition. This giant r42 is extracted from a device in which the optical mode is fully confined within a single-crystal BTO waveguide. This approach contrasts with previous designs where the core material - typically silicon or silicon nitride - supports only partial confinement, resulting in an evanescent overlap with a multi-crystalline BTO layer. This highly confined BTO on insulator electro-optic modulation technology may significantly advance the field of ultra-low-power integrated photonic devices and allows for the realization of next-generation efficient and compact photonic circuits.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

31 extracted references · 29 canonical work pages

  1. [1]

    Bogaerts, W. et al. Programmable photonic circuits. Nature 586, 207–216 (2020)

    2020

  2. [2]

    Shekhar, S. et al. Roadmapping the next generation of silicon photonics. Nat Commun 15, 751 (2024)

    work page 2024

  3. [3]

    T., Mashanovich, G., Gardes, F

    Reed, G. T., Mashanovich, G., Gardes, F. Y. & Thomson, D. J. Silicon optical modulators. Nature Photon 4, 518– 526 (2010)

    work page 2010

  4. [4]

    Liang, G. et al. Robust, efficient, micrometre-scale phase modulators at visible wavelengths. Nat. Photon. 15, 908– 913 (2021)

    work page 2021

  5. [5]

    Berciano, M. et al. Strained silicon photonics for Pockels effect based modulation. in 2017 IEEE 14th International Conference on Group IV Photonics (GFP) 57–58 (IEEE, Berlin, Germany, 2017). doi:10.1109/GROUP4.2017.8082194

    work page arXiv 2017

  6. [6]

    Wang, C. et al. Integrated lithium niobate electro-optic modulators operating at CMOS-compatible voltages. Nature 562, 101–104 (2018)

    work page 2018

  7. [7]

    Boes, A. et al. Lithium niobate photonics: Unlocking the electromagnetic spectrum. Science 379, eabj4396 (2023)

    work page 2023

  8. [8]

    Feng, H. et al. Integrated lithium niobate microwave photonic processing engine. Nature 627, 80–87 (2024)

    2024

Show all 31 references
  1. [9]

    Zgonik, M. et al. Dielectric, elastic, piezoelectric, electro-optic, and elasto-optic tensors of BaTiO3 crystals. Phys. Rev. B 50, 5941–5949 (1994)

    1994

  2. [10]

    L., Towner, D

    Tang, P., Meier, A. L., Towner, D. J. & Wessels, B. W. BaTiO3 thin-film waveguide modulator with a low voltage– length product at near-infrared wavelengths of 0.98 and 1.55 μm

  3. [11]

    J., Hamano, T., Meier, A

    Tang, P., Towner, D. J., Hamano, T., Meier, A. L. & Wessels, B. W. Electrooptic modulation up to 40 GHz in a barium titanate thin film waveguide modulator. Opt. Express 12, 5962 (2004)

    2004

  4. [12]

    Abel, S. et al. Large Pockels effect in micro- and nanostructured barium titanate integrated on silicon. Nature Mater 18, 42–47 (2019)

    2019

  5. [13]

    Cao, Y. et al. A Barium Titanate-on-Oxide Insulator Optoelectronics Platform. Advanced Materials 33, 2101128 (2021)

    2021

  6. [14]

    Xiong, C. et al. Active Silicon Integrated Nanophotonics: Ferroelectric BaTiO3 Devices. Nano Lett. 14, 1419–1425 (2014)

    2014

  7. [15]

    & Danner, A

    Lin, H.-L., Aashna, P., Cao, Y. & Danner, A. Single -mode waveguides in barium titanate -on-insulator platform fabricated by wet etching. in CLEO 2024 SW4R.6 (Optica Publishing Group, Charlotte, North Carolina, 2024). doi:10.1364/CLEO_SI.2024.SW4R.6

    2024 doi

  8. [16]

    Eltes, F. et al. A BaTiO3 -Based Electro-Optic Pockels Modulator Monolithically Integrated on an Advanced Silicon Photonics Platform. J. Lightwave Technol. 37, 1456–1462 (2019)

    2019

  9. [17]

    Abel, S. et al. A strong electro-optically active lead-free ferroelectric integrated on silicon. Nat Commun 4, 1671 (2013)

    2013

  10. [18]

    Castera, P. et al. Electro-Optical Modulation Based on Pockels Effect in BaTiO 3 With a Multi-Domain Structure. IEEE Photon. Technol. Lett. 28, 990–993 (2016)

    2016

  11. [19]

    & Buchal, Ch

    Petraru, A., Schubert, J., Schmid, M., Trithaveesak, O. & Buchal, Ch. Integrated optical Mach Zehnder modulator based on polycrystalline BaTiO3. Opt. Lett. 28, 2527 (2003). 9

    2003

  12. [20]

    Eltes, F. et al. Low-Loss BaTiO3–Si Waveguides for Nonlinear Integrated Photonics. ACS Photonics 3, 1698–1703 (2016)

    2016

  13. [21]

    J., Meier, A

    Tang, P., Towner, D. J., Meier, A. L. & Wessels, B. W. Low -Loss Electrooptic BaTiO 3 Thin Film Waveguide Modulator. IEEE Photon. Technol. Lett. 16, 1837–1839 (2004)

    2004

  14. [22]

    J., Meier, A

    Tang, P., Towner, D. J., Meier, A. L. & Wessels, B. W. Low -voltage, polarization -insensitive, electro -optic modulator based on a polydomain barium titanate thin film. Applied Physics Letters 85, 4615–4617 (2004)

    2004

  15. [23]

    Kormondy, K. J. et al. Microstructure and ferroelectricity of BaTiO 3 thin films on Si for integrated photonics. Nanotechnology 28, 075706 (2017)

    2017

  16. [24]

    Dong, Z. et al. Monolithic Barium Titanate Modulators on Silicon -on-Insulator Substrates. ACS Photonics 10, 4367–4376 (2023)

    2023

  17. [25]

    Winiger, J. et al. PLD Epitaxial Thin-Film BaTiO3 on MgO-Dielectric and Electro-Optic Properties. Adv Materials Inter 11, 2300665 (2024)

    2024

  18. [26]

    & Tuller, H

    Kim, I.-D., Avrahami, Y., Socci, L., Lopez-Royo, F. & Tuller, H. L. Ridge waveguide using highly oriented BaTiO3 thin films for electro-optic application. Journal of Asian Ceramic Societies 2, 231–234 (2014)

    2014

  19. [27]

    M., Conrad, C

    Gill, D. M., Conrad, C. W., Ford, G., Wessels, B. W. & Ho, S. T. Thin -film channel waveguide electro -optic modulator in epitaxial BaTiO3. Applied Physics Letters 71, 1783–1785 (1997)

    1997

  20. [28]

    Ortmann, J. E. et al. Ultra-Low-Power Tuning in Hybrid Barium Titanate-Silicon Nitride Electro-optic Devices on Silicon. ACS Photonics 6, 2677–2684 (2019)

    2019

  21. [29]

    D., Lin, H.-L., Tan, S

    Cao, Y., Ng, J. D., Lin, H.-L., Tan, S. L. & Danner, A. An active straight waveguide polarization rotator in single crystal thin film barium titanate-on-insulator. Applied Physics Letters 122, 031106 (2023)

    2023

  22. [30]

    Leroy, F. et al. Guided-wave electro-optic characterization of BaTiO3 thin films using the prism coupling technique. Opt. Lett. 38, 1037 (2013)

    2013

  23. [31]

    M., Conrad, C

    Gill, D. M., Conrad, C. W., Ford, G., Wessels, B. W. & Ho, S. T. Thin -film channel waveguide electro -optic modulator in epitaxial BaTiO3. Applied Physics Letters 71, 1783–1785 (1997). 10 Methods BTO thin film growth BTO films were grown by pulsed laser deposition with a KrF ...

    1997

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.