{"id":"9ed13764-3cb6-4763-9858-eed1a653a41f","arxiv_id":"2506.13209","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"high","formal_verification":"none","parameter_count":3,"one_line_summary":"A new functional-independent DFT workflow computes the clamped Pockels tensor of tetragonal BaTiO3 and finds that the largest coefficient r51 rises steeply as the local titanium off-centering decreases.","lead":"This paper computes the Pockels tensor of tetragonal barium titanate, the coefficient that controls its use in fast optical modulators, using a flexible first-principles workflow. It finds that the largest Pockels coefficient depends strongly on the microscopic off-centering of titanium atoms and can be tuned into agreement with experiments by adjusting that off-centering.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed recovery of r51 depends on tuning Ti off-centering to 0.45%, while the P4bm ground state at 0.466% gives 391 pm/V; without independent evidence that 0.45% is physical, the central claim overreaches.","rationale":"I read the paper as proposing an automated finite-difference route to clamped Pockels tensors with arbitrary functionals, and using it to recover the large r51 of tetragonal BTO. The method is credible: it builds on established electric-enthalpy finite-field techniques, is implemented in open-source Vibroscopy, and the data are archived, which are real supporting artifacts. The main scientific claim, however, is the recovery of the experimental r51 range, and that claim rests on Section III C, where the Ti off-centering is scanned. The ground-state structure gives r51 = 391 pm/V; the experimental range is reached only after moving Ti to 0.45% displacement, and the physical justification for that magnitude is not supplied. The paper itself states that refs 55 and 56 show local off-centering exists but not its value. This makes the abstract's 'as a result ... recovered' overstate what is demonstrated. The reader's weakest-assumption analysis identified exactly this: the magnitude of Ti off-centering is unconstrained experimentally, and the result is extremely sensitive to it. I agree. I do not see an internal algebraic error in Eq. (4) or the finite-difference implementation that would constitute a separate fatal issue. The r13 sign problem is a further symptom that the calculation is incomplete, but it is not the central claim; the same missing-physics concern reinforces the conditional verdict rather than changing it. The requested revision, either independent evidence for 0.45% or reframing as a sensitivity study, is feasible, so CONDITIONAL remains appropriate.","tokens_in":16456,"tokens_out":5201,"duration_ms":58936,"concrete_test":"Determine the actual Ti off-centering magnitude in the measured clamped films, e.g., by Ti K-edge EXAFS or diffuse-scattering refinement on the films of ref. 14, and recompute r51 at that value using the same Vibroscopy/PBEsol workflow. If the refined displacement is 0.466% within uncertainty, Table IV implies r51 ≈ 391 pm/V, outside the experimental range, and the abstract's recovery claim fails; if it is 0.45% or lower, the claim is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is Section III C and Table IV. The abstract states that 'the correct value range of r51 ... is recovered' as a result of the method, but the ground-state P4bm structure has Ti off-centering 0.466% and yields r51 = 391.2 pm/V, 46% below the experimental 730 ± 150 pm/V. The recovery is obtained only after manually scanning the Ti displacement: at 0.45% r51 = 667 pm/V, and at 0.425% r51 = 1614.7 pm/V. Thus a 3.4% relative reduction in the off-centering parameter (0.466% to 0.45%) changes r51 by about 70%, and a 8.8% reduction changes it by about a factor of four. No independent evidence is given that the real clamped BTO films have 0.45% rather than 0.466% off-centering; refs 55 and 56 establish local off-centering and its direction, not its magnitude. Without that, 'recovery' is a post-hoc parameter selection, not a prediction. The same section concedes that r13 is negative at 0.45% and that a physical contribution is missing, further weakening the claim that the tensor is quantitatively captured. The monotonic trend r51 versus off-centering is a useful sensitivity result, but it does not by itself validate the headline.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents an automated finite-difference DFT workflow for computing clamped Pockels tensors, based on the modern theory of polarization, the electric-enthalpy functional, and finite-difference derivatives of forces and polarizations, implemented in the open-source Vibroscopy/AiiDA ecosystem. It applies this workflow to tetragonal BaTiO3 using PBEsol, RRKJ ultrasoft pseudopotentials, and a P4bm 2x2x1 supercell with local titanium off-centering, which stabilizes the soft phonon modes that LDA-based DFPT cannot describe. The authors compute r13, r33, and r51, test an extended-Hubbard correction (PBEsol+U+V), and study the dependence of the Pockels tensor on the magnitude of titanium off-centering. They report that r51 increases as titanium off-centering decreases and claim that at 0.45% displacement r51 enters the experimental range, recovering the measured value.","tokens_in":16781,"tokens_out":7320,"duration_ms":86912,"significance":"If the central claim were actually predictive, the paper would be significant: a functional-independent, reproducible workflow that avoids imaginary soft modes in a technologically relevant ferroelectric and reproduces the largest Pockels coefficient of BaTiO3 would be a useful contribution to electro-optic materials design. The strengths of the paper are its open-source implementation, the availability of data on the Materials Cloud Archive, the self-consistent first-principles determination of Hubbard parameters, and the clean sensitivity analysis showing that r51 increases with decreasing titanium off-centering. However, the quantitative recovery of r51 is currently a post-hoc parameter selection rather than a prediction, and the full clamped tensor is not reproduced at the selected geometry because r13 has the wrong sign. These issues affect the paper's headline claim and require reframing or additional evidence before publication.","major_comments":[{"comment":"The central claim that r51 is recovered rests on a post-hoc choice of titanium off-centering. The P4bm ground state at 0.466% displacement gives r51 = 391.2 pm/V, 46% below experiment; reducing the displacement to 0.45% gives 667 pm/V and to 0.425% gives 1614.7 pm/V. The paper selects the intermediate value because it falls inside the experimental window of 730 ± 150 pm/V. References [55] and [56] establish the existence and direction of local titanium off-centering but do not pin down its magnitude, so the paper provides no independent evidence that real clamped films have 0.45% rather than 0.466% displacement. A 3.4% relative change in this parameter changes r51 by about 70%, and an 8.8% change multiplies it by about four, making the agreement at 0.45% highly sensitive to an unconstrained input. The abstract and conclusion should either be reworded to present this as a sensitivity study, or independent evidence for the 0.45% magnitude must be supplied.","section":"Section III C and Table IV"},{"comment":"Even at the tuned geometry, the full clamped Pockels tensor is not captured: at 0.45% displacement the computed r13 is -1.5 pm/V, whereas the experimental value is 9 ± 2 pm/V. The manuscript itself concedes, in the final paragraph of Section III C, that the negative r13 suggests a missing physical contribution. Thus the statement that the correct value range of r51 is recovered cannot be extended to a claim that the Pockels tensor is quantitatively reproduced. The authors should identify and correct the missing contribution, or explicitly limit the paper's quantitative claim to r51.","section":"Section III C, final paragraph and Table IV"},{"comment":"The claim that symmetry requires r13, r33, and r51 to have the same sign is incorrect. For point group 4mm, the independent electro-optic coefficients are not related to one another by symmetry, and the group-theoretical tensor tables cited as reference [57] do not impose equal signs on these coefficients. Consequently, the inference that a negative and small r13 must indicate a missing physical contribution does not follow from symmetry alone. The discrepancy with the experimental r13 remains, but the stated symmetry-based reasoning should be removed or corrected.","section":"Section III B"}],"minor_comments":[{"comment":"Reference [40] is the 1996 PBE paper, not the PBEsol functional; the PBEsol functional should be cited to Perdew et al., Phys. Rev. Lett. 100, 136406 (2008).","section":"Reference [40]"},{"comment":"The overestimation of r33 is stated as 26% in Section III B and as 35% in the Conclusion, while Table II gives 58.4 pm/V against an experimental value of 43 pm/V, i.e., about 36%; these numbers should be made consistent.","section":"Section III B and Conclusion"},{"comment":"The statement 'r51 within 4% of the experimental value at 0.45% displacement' is difficult to reconcile with Table IV, where 667 pm/V is about 8.6% lower than the quoted experimental center value of 730 pm/V; the comparison basis (center value, near boundary, or corrected reference) should be specified.","section":"Conclusion"},{"comment":"The left panel in Fig. 8 is labeled r31 in the caption, but the discussion concerns r13; this label should be corrected.","section":"Fig. 8 caption"},{"comment":"The sentence comparing Kim et al.'s r51 = 1844 pm/V with the present 1614.7 pm/V is confusing, because the phrase 'is even smaller than our result' reads as if the coefficient were smaller when numerically it is larger; the intended comparison is with the titanium displacement, and the sentence should be rephrased.","section":"Section III C, comparison with Kim et al."}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Read the Pockels paper. The workflow is genuinely new and the paper is honest about its own limitations, but the abstract oversells the r51 result. The match with experiment comes from tuning the Ti off-centering to 0.45%, not from the ground-state structure (0.466% gives 391 pm/V, 46% below experiment; 0.425% gives 1614 pm/V). Without independent evidence that real clamped films have 0.45% off-centering, this is post-hoc selection, not prediction. The paper itself frames it as a sensitivity study in Section III C, but the abstract and conclusion say 'recovered', which is too strong.\n\nWhat is actually good: the functional-independent finite-difference pipeline in Vibroscopy is a real contribution—it lets you compute clamped Pockels tensors beyond LDA, and they demonstrate it with PBEsol and a P4bm supercell that stabilizes the soft modes. The analysis of why PBEsol+U+V fails (Hubbard correction changes the soft-mode eigenvectors away from Slater type, so the large Raman/polarity terms vanish) is careful and useful. The r51-vs-off-centering trend is new and physically plausible, consistent with earlier strain-dielectric work. Data and code are open, which is a plus.\n\nSoft spots, in order: (1) The main quantitative claim is a fit. The sensitivity to the off-centering is extreme—a 3.4% change in displacement changes r51 by 70%. Refs 55/56 establish local off-centering exists and its direction, but not its magnitude. So the 'recovery' should be framed as a demonstration of sensitivity, not a zero-parameter prediction. (2) r13 is negative and tiny, against both symmetry expectations and experiment; the paper concedes a missing physical contribution. That's okay as a limitation, but it means the tensor is not quantitatively captured. (3) Minor: the comparison with Kim et al. involves a factor-of-two on displacement percentages that could confuse readers.\n\nMy take: this deserves a serious referee. The method is solid and the sensitivity study is worth publishing. The revisions are within reach: reframe the r51 claim, or give independent evidence for 0.45% off-centering, and address r13. I'd send it out.","headline":"Useful functional-independent Pockels workflow with an honest sensitivity study, but the headline r51 'recovery' is a tuned parameter, not a prediction.","tokens_in":17336,"tokens_out":2563,"would_cite":true,"duration_ms":27460,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["78.20.Jq","71.15.Mb","77.80.-e"],"model":"deepseek-v4-flash","headline":"An ab initio workflow that works with any exchange-correlation functional recovers the large Pockels coefficient r51 of tetragonal BaTiO3 and traces it to titanium off-centering.","keywords":["Pockels effect","barium titanate","density-functional theory","electro-optic modulators","soft phonon modes","titanium off-centering","clamped Pockels tensor","ferroelectrics"],"falsifier":"Measure the local titanium displacement magnitude in clamped tetragonal BTO films (e.g., by extended X-ray absorption fine structure or diffuse scattering that resolves the $\\langle110\\rangle$ component); if it comes out at 0.466% or larger rather than 0.45%, the reported recovery of $r_{51}$ fails, since the paper's own ground-state calculation gives $r_{51}\\approx391$ pm/V, well below the experimental $730\\pm150$ pm/V.","tokens_in":16274,"feed_emoji":"⚡","tokens_out":11081,"duration_ms":104751,"temperature":0.7,"pith_summary":"The paper presents a way to compute the clamped Pockels tensor of ferroelectric crystals from density-functional theory without locking in a particular exchange-correlation functional, using finite-difference derivatives of polarizations and forces. Applied to tetragonal barium titanate, the workflow stabilizes the soft phonon modes that control the largest Pockels coefficient, $r_{51}$, which earlier LDA-based calculations could not reach. The authors show that $r_{51}$ depends strongly on the local off-centering of titanium atoms along $\\langle110\\rangle$: reducing the off-centering from the calculated ground-state value of 0.466% to 0.45% raises $r_{51}$ from 391 pm/V to 667 pm/V, inside the experimental range of 730 ± 150 pm/V. If correct, this gives a predictive, functional-flexible route for designing electro-optical modulators around barium titanate films.","feed_headline":"BaTiO3's r51 matches experiment at 0.45% titanium off-centering","feed_subtitle":"The workflow stabilizes the soft phonons behind the largest electro-optical response in tetragonal BaTiO3.","key_machinery":"The load-bearing object is a $2\\times2\\times1$ P4bm supercell of tetragonal BTO with local titanium off-centering along the $\\langle110\\rangle$ directions: four titanium atoms displaced in a vortex-like pattern that keeps the macroscopic polarization along [001] while replacing one mirror plane by a glide plane. Combined with PBEsol and RRKJ ultrasoft pseudopotentials, this structural prototype turns the imaginary soft phonon modes of the P4mm cell into positive, Slater-type modes, making it possible to evaluate the phonon contribution to the Pockels tensor. The calculation itself is carried by automated first- and second-order finite-difference derivatives of the polarization and of the Hellmann-Feynman forces, together with the modern theory of polarization through the electric-enthalpy functional; the titanium displacement percentage is the tunable parameter that controls $r_{51}$.","core_discovery":"The central claim is that the clamped Pockels tensor of tetragonal BTO can be obtained at the ab initio level for any exchange-correlation functional, and that the experimentally large coefficient $r_{51}$ is recovered once the structure is described with the P4bm supercell whose titanium atoms are displaced along $\\langle110\\rangle$, using PBEsol and ultrasoft pseudopotentials. At the ground-state 0.466% titanium displacement the calculation gives $r_{51}=391.2$ pm/V, outside the experimental $730\\pm150$ pm/V; moving the titanium positions toward the high-symmetry P4mm geometry so that the displacement is 0.45% yields $r_{51}=667$ pm/V, while 0.425% yields 1614.7 pm/V. The same trend connects the soft phonon frequency (2.0, 1.5, and 1.0 THz at 0.466%, 0.45%, and 0.425%) and the in-plane dielectric constant, matching the physical picture that decreasing titanium off-centering softens the low-energy modes and enhances the electro-optic response. The paper also argues that the PBEsol+U+V correction, although it improves the band gap and dielectric tensor, destroys the Slater-type character of the soft mode eigenvectors and therefore is not suitable for Pockels predictions.","pith_inferences":["Beyond the paper: if the monotonic relation between titanium off-centering and $r_{51}$ holds in films, strain or defect engineering that reduces the off-centering would become a concrete design lever for electro-optic modulators.","Beyond the paper: the X-ray experiments cited establish that local titanium off-centering exists but not its magnitude, so a direct measurement of the displaced amount would decide whether 0.45% is a prediction or an adjustable fit.","Beyond the paper: applying the same finite-difference workflow to other ferroelectric perovskites would test whether softer Slater-type modes generically produce larger Pockels coefficients, as suggested by the BTO results.","Beyond the paper: the persistent sign mismatch for $r_{13}$ hints that physics omitted here, such as anharmonicity, finite temperature, or strain coupling, is needed before the full tensor can be trusted."],"forward_implications":["With a titanium off-centering of 0.45% along $\\langle110\\rangle$, the computed clamped $r_{51}$ of 667 pm/V falls inside the experimental range of 730 ± 150 pm/V, while the soft phonon frequency drops from 2.0 to 1.5 THz.","Starting from the P4bm ground state, moving titanium toward the P4mm geometry monotonically increases $r_{51}$; at 0.425% displacement it reaches 1614.7 pm/V, so small structural changes produce large electro-optic changes.","The smallest coefficient $r_{13}$ remains vanishingly small and negative in the tuned structures, even though tetragonal symmetry requires $r_{13}$, $r_{33}$, and $r_{51}$ to share a sign, indicating a missing contribution in the calculation.","PBEsol+U+V should not be used for Pockels predictions in BTO: it raises the soft mode frequency to 3.16 THz and changes the eigenvectors away from the Slater type, driving $r_{51}$ down to 65.5 pm/V despite improving the band gap and dielectric constants.","Because the workflow is exchange-correlation functional independent, it can be applied to strained or defective ferroelectric films, the configurations relevant for integrated electro-optical modulators."],"supporting_citations":[{"why":"Derives the clamped Pockels tensor formula and its DFPT implementation; the LDA-based calculation could not report r51 because of unstable soft phonon modes.","marker":"[21]"},{"why":"Prior DFPT calculation with averaged titanium-displaced cells that reported r51=820.8 pm/V, the main comparison for the titanium-displacement dependence.","marker":"[24]"},{"why":"Introduces the structural prototype idea that motivates the P4bm 2x2x1 supercell with local titanium off-centering.","marker":"[43]"},{"why":"The exchange-correlation functional used for all DFT calculations; the text labels it PBEsol.","marker":"[40]"},{"why":"Supplies the automated finite-difference machinery that turns force and polarization derivatives into phonon frequencies, Raman tensors, mode polarities, and dielectric susceptibilities.","marker":"[30]"},{"why":"Gives the modern theory of polarization used to obtain polarizations and their derivatives.","marker":"[27]"},{"why":"Introduces the finite-electric-field electric-enthalpy approach used to compute polarization derivatives with respect to the field.","marker":"[38]"},{"why":"X-ray evidence that local titanium displacements are real and tilted away from [111], grounding the off-centering supercell in experiment.","marker":"[55]"},{"why":"Provides the experimental clamped Pockels coefficients, including the r51=730±150 pm/V target.","marker":"[15]"},{"why":"Provides the experimental soft phonon frequency near 1.14 THz used to judge the stabilized modes.","marker":"[51]"}],"fun_headline_variants":["BaTiO3's r51 matches experiment at 0.45% off-centering","Titanium off-centering tunes BaTiO3's Pockels r51 to experiment","Reduced Ti shift boosts BaTiO3's electro-optic coefficient","Ab initio Pockels tensor recovers BaTiO3 r51 beyond LDA","Soft phonons and Ti displacement control BaTiO3's r51"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that real tetragonal BTO films have a titanium off-centering near the tuned 0.45% displacement along $\\langle110\\rangle$, rather than the 0.466% of the computed P4bm ground state, because the cited X-ray experiments show off-centering exists but do not fix its magnitude.","fun_headline_variants_meta":{"raw":{"variants":["BaTiO3's r51 matches experiment at 0.45% off-centering","Titanium off-centering tunes BaTiO3's Pockels r51 to experiment","Reduced Ti shift boosts BaTiO3's electro-optic coefficient","Ab initio Pockels tensor recovers BaTiO3 r51 beyond LDA","Soft phonons and Ti displacement control BaTiO3's r51"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000958,"raw_usage":{"total_tokens":4145,"prompt_tokens":1072,"completion_tokens":3073,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":688,"completion_tokens_details":{"reasoning_tokens":2969}},"tokens_in":688,"tokens_out":3073,"duration_ms":25514,"temperature":1.0,"reasoning_tokens":2969,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T00:35:58.298601+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the local titanium displacement magnitude in clamped tetragonal BTO films (e.g., by extended X-ray absorption fine structure or diffuse scattering that resolves the $\\langle110\\rangle$ component); if it comes out at 0.466% or larger rather than 0.45%, the reported recovery of $r_{51}$ fails, since the paper's own ground-state calculation gives $r_{51}\\approx391$ pm/V, well below the experimental $730\\pm150$ pm/V.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the clamped Pockels tensor formula and its DFPT implementation; the LDA-based calculation could not report r51 because of unstable soft phonon modes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Prior DFPT calculation with averaged titanium-displaced cells that reported r51=820.8 pm/V, the main comparison for the titanium-displacement dependence."},{"cited_title":"Umari and A","cited_arxiv_id":null,"evidence_quote":"Introduces the structural prototype idea that motivates the P4bm 2x2x1 supercell with local titanium off-centering."},{"cited_title":"Togo, First-principles phonon calculations with Phonopy and Phono3py, J","cited_arxiv_id":null,"evidence_quote":"The exchange-correlation functional used for all DFT calculations; the text labels it PBEsol."},{"cited_title":"Bastonero and N","cited_arxiv_id":null,"evidence_quote":"Supplies the automated finite-difference machinery that turns force and polarization derivatives into phonon frequencies, Raman tensors, mode polarities, and dielectric susceptibilities."},{"cited_title":"Veithen, X","cited_arxiv_id":null,"evidence_quote":"Gives the modern theory of polarization used to obtain polarizations and their derivatives."},{"cited_title":"Souza, J","cited_arxiv_id":null,"evidence_quote":"Introduces the finite-electric-field electric-enthalpy approach used to compute polarization derivatives with respect to the field."},{"cited_title":"Scalabrin, A","cited_arxiv_id":null,"evidence_quote":"X-ray evidence that local titanium displacements are real and tilted away from [111], grounding the off-centering supercell in experiment."},{"cited_title":"U Bremen Excellence Chair Program","cited_arxiv_id":null,"evidence_quote":"Provides the experimental clamped Pockels coefficients, including the r51=730±150 pm/V target."},{"cited_title":"Timrov, N","cited_arxiv_id":null,"evidence_quote":"Provides the experimental soft phonon frequency near 1.14 THz used to judge the stabilized modes."}],"review_version":1}