{"id":"ce09e3fe-6437-4b5f-9175-bff062f87938","arxiv_id":"2607.19885","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Ferroelectricity in CaBi2B2O9 (B=Ta,Nb) is shown to arise from cooperative trilinear coupling of polar, rotation, and tilting modes plus interlayer sliding, explaining high Curie temperatures and resistivity.","lead":"This paper uses density-functional theory and group-theoretical mode analysis to explain why the Aurivillius ferroelectrics CaBi2Ta2O9 and CaBi2Nb2O9 have very high Curie temperatures, large polarization, and high electrical resistivity. It argues that a cooperative coupling of three lattice distortion modes—and an interlayer sliding between Bi2O2 layers and perovskite blocks—underlies these properties, offering design guidance for high-temperature piezoelectrics.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Layer-number-independent interlayer-sliding mechanism is asserted from only two n=2 compounds; no n=1 or n=3 test is provided.","rationale":"After reading the paper, I agree with the reader's conditional verdict. The central quantitative claim for the two compounds—cooperative condensation of Γ5−, X2+, X3− with dominant trilinear stabilization—is well supported: the Landau expansion has R2>0.99, coefficients are tabulated, and the decomposition is consistent with the total DFT energy. The mode amplitudes, frozen-phonon curves, and Berry-phase polarizations are internally consistent. However, the abstract and the 'Polarization properties' section generalize the interlayer-sliding mechanism to all Aurivillius oxides based on topology alone. This is the weakest link because the paper provides no calculation for n=1 or n=3, and known results for Bi4Ti3O12 (ref. [20]) indicate that A-site Bi in the perovskite block is a major polar driver, which would break the naive layer-number independence. An additional secondary concern is the reuse of piezoelectric constants from prior work (Table VI) and a ~20% discrepancy between the Berry-phase polarization (0.59 C/m²) and the Born-charge decomposition (0.71 C/m²) for CBNO, which is not discussed; but these do not alter the qualitative mechanism. Therefore, the verdict remains CONDITIONAL: the n=2 mechanism is credible, but the universal claim needs verification on other layer numbers.","tokens_in":22149,"tokens_out":11404,"duration_ms":120861,"concrete_test":"Perform the same PBE symmetry-mode decomposition and frozen-phonon analysis for n=1 (Bi2WO6) and n=3 (Bi4Ti3O12) using identical VASP settings. Specifically, (1) relax the polar ground state from the parent phase and decompose the distortion into mode amplitudes; (2) compute the frozen-phonon energy curve for rigid relative sliding of the Bi2O2 layer against the perovskite block in the parent phase; (3) compute the overlap of the softest polar Γ5−-type eigenvector with the rigid-sliding displacement pattern. If in Bi4Ti3O12 the polar eigenvector has <50% overlap with rigid sliding, or the sliding well depth is much shallower than for n=2, the layer-number-independent claim fails. An alternative quantitative test: compute the layer-shift (P_shift) vs intralayer (P_intra) decomposition for Bi4Ti3O12 using the same Born-charge method; if P_intra/P_shift is markedly different from the n=2 va","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's most distinctive claim—'interlayer sliding as a general, layer-number-independent structural mechanism for ferroelectricity in Aurivillius oxides'—is not supported by the presented evidence. Only two n=2 compounds (CBTO, CBNO) are computed. The topological assertion that alternating Bi2O2/perovskite stacking implies a universal mechanism ignores known changes in polar-mode character with layer number: in n=3 Bi4Ti3O12, the perovskite-block A-site is Bi3+, and ref. [20] shows this Bi participates strongly in the ferroelectric instability; in n=1, there is no A-site cation in the block. The paper's own Table V shows that intralayer polarization (P_intra = 0.317/0.542 C/m²) exceeds the layer-shift contribution (P_shift = 0.146/0.170 C/m²) in both compounds, so even for n=2, rigid sliding is not the dominant polarization source. The frozen-phonon evidence (Fig. 4c,f) demonstrates that rigid Bi2O2 sliding is a soft channel in the parent phase, but the actual Γ5− mode couples this sliding to substantial internal distortions (δu_a of 0.015–0.052 frac in Table IV). No n=1 or n=3 calculation is provided to test the claimed layer-number independence, so the generalization is a model hypothesis, not a demonstrated result.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a first-principles study of the two-layer Aurivillius ferroelectrics CaBi2B2O9 (B = Ta, Nb). Using the I4/mmm phase as a reference, the authors decompose the A21am ferroelectric distortion into Γ1+, Γ5-, X2+, and X3- symmetry modes, fit a multimode Landau expansion to DFT total energies, and conclude that the Γ5-X2+X3- trilinear term dominates the energy stabilization (61.4% in CBTO, 80.7% in CBNO). Berry-phase calculations give spontaneous polarizations of 0.44 and 0.59 C/m2; a Born-effective-charge layer decomposition attributes a large share of the polarization to Bi displacements and to relative sliding between Bi2O2 layers and perovskite blocks. Piezoelectric constants are reported for the two compounds, and band-structure/effective-mass analyses are used to explain their high intrinsic resistivity. On this basis the paper advances a broad claim that interlayer sliding is a general, layer-number-independent ferroelectric mechanism in all Aurivillius oxides.","tokens_in":22423,"tokens_out":7064,"duration_ms":77717,"significance":"If the central results hold, the paper gives a valuable microscopic picture of ferroelectricity in two representative high-TC Aurivillius compounds, going beyond the single-soft-mode paradigm of conventional perovskites. The DFT calculations are carefully specified, the mode decomposition is standard and reproducible, and the Landau fit achieves R2>0.99. The layer-resolved Born-charge decomposition is a useful contribution, as is the identification of Bi2O2 sliding as an energetically soft channel. The main limitation is that the paper's most distinctive generalization—layer-number independence—is not tested by any calculation beyond the two n=2 compounds, and the own data in Table V show that intralayer distortions, not rigid sliding, dominate the polarization magnitude. The piezoelectric tensor table also appears to quote prior work rather than report new calculations.","major_comments":[{"comment":"The claim that interlayer sliding is 'a general, layer-number-independent structural mechanism for ferroelectricity in Aurivillius oxides' is not supported by the evidence in the manuscript. Only two n=2 compounds (CBTO, CBNO) are computed; no n=1, n=3, or n=4 member is investigated or cited with quantitative calculations. Moreover, Table V shows that the intralayer polarization (P_intra = 0.317 C/m2 in CBTO and 0.542 C/m2 in CBNO) exceeds the layer-shift contribution (P_shift = 0.146 and 0.170 C/m2), so even for these n=2 compounds rigid interlayer sliding is not the dominant source of polarization. The known n=3 compound Bi4Ti3O12 (ref. [20]) has a polar mode in which the perovskite-block Bi participates strongly, indicating that the polar character can change with layer number. Please either add explicit calculations for at least one other layer number, or restrict the conclusion to t","section":"Abstract; Conclusion; 'Polarization properties', paragraph beginning 'Based on the above analysis...'"},{"comment":"Table VI reports d31, d32, d33, d24, and d15 for PBE and PBEsol, with the footnote 'a,b Obtained by Tan et al. from DFT calculations[16,44].' The Methods section states 'The piezoelectric tensor is evaluated by a finite-difference approach,' and the text says 'The piezoelectric constants are calculated using the direct finite-stress method... as summarized in Table VI.' This is internally inconsistent. If the tensor components are taken from Refs. [16] and [44], the text and Methods should clearly say so, and the original contribution of this paper is then limited to the decomposition in Table VII/Eq. (4), not the d_ij values themselves. If the values were recomputed here, the footnote is incorrect. Either way, the manuscript must be corrected because the current wording misattributes the piezoelectric tensor data.","section":"Table VI and 'Piezoelectric properties'"},{"comment":"The energy-decomposition percentages in Table III need a precise definition and consistent sign convention. For CBTO, the Γ1+ single-mode entry is listed as '+233.9 (−30.9%)' while the Γ5− entry is '−159.8 (21.1%)'; for the trilinear term the entry is '−464.4 (61.4%).' It is not stated whether the percentage is the term value divided by the total stabilization energy (756.2 meV/f.u. for CBTO) or by something else, and the signs of the percentages do not match the signs of the energy contributions. Since the paper's headline conclusion that the trilinear term contributes 61.4% and 80.7% of the stabilization rests on this table, please define the normalization explicitly and ensure that positive-energy contributions are labeled consistently (e.g., as destabilizing) relative to the total energy lowering.","section":"Table III and Eq. (1)"}],"minor_comments":[{"comment":"The caption says 'at fixed cubic lattice parameters,' but the reference structure throughout is the tetragonal I4/mmm phase. This should read 'tetragonal I4/mmm lattice parameters.'","section":"Table II caption"},{"comment":"The effective-mass formula is written as ∂2E(k)/∂2k; it should be ∂²E(k)/∂k². Also define which Cartesian direction is used for each effective mass more explicitly.","section":"Eq. (6)"},{"comment":"In the paragraph discussing the perovskite block, the text says 'The intralayer polarization contribution P shif t from the perovskite block is calculated to be 0.211 and 0.412 C/m2...' This should be P_intra, not P_shift, based on the definitions in the same section.","section":"Text near Table V"},{"comment":"The superscripts 'a,b' in Table VI are not explained in the caption. If the values are from Refs. [16] and [44], use a standard citation format and state this in the text as well as the footnote.","section":"Table VI footnote"},{"comment":"Several sentences contain grammatical slips, e.g., 'whose the Γ5−X2+X3− trilinear coupling,' 'the all the biquadratic coupling terms,' and 'Since the Bi2O2 layer and the perovskite block constitute... this mechanism is expected to be generally applicable.' These should be corrected in a careful language pass.","section":"Grammar and wording"},{"comment":"The general Aurivillius formula is written as Bi2mAn−mBnO3(m+n); the conventional notation is usually Bi2O2(A_{n−1}B_nO_{3n+1}). Please check the formula for typographical accuracy.","section":"Introduction"}],"recommendation":"major_revision","confidential_remarks":"The underlying DFT calculations for the two n=2 compounds appear sound and the Landau energy decomposition is a legitimate, useful analysis. The main obstacle to acceptance is the unsupported universal claim in the abstract/conclusion, which is easy to fix either by adding a test at another layer number or by clearly labeling the generalization as a hypothesis. The ambiguity about whether the piezoelectric tensor values in Table VI are original or quoted from prior work also needs resolution before publication; this is a matter of attribution and internal consistency rather than a technical flaw in the calculations. With those changes, the paper could be suitable for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a serious DFT paper on two Aurivillius ferroelectrics, and the quantitative Landau decomposition is worth engaging with, but the headline generalization about layer-number-independent sliding is not supported by the evidence presented. The core result is that the A21am phase is stabilized by trilinear coupling among Gamma5-, X2+, X3-, with that term contributing around 61% and 81% of the energy lowering in CBTO and CBNO. The fit is good (R^2>0.99), the frozen-phonon energy profiles are informative, and the layer-center decomposition of polarization is a genuinely useful addition; it lets them separate rigid sliding from intralayer distortion and show that Nb off-centering drives the polarization difference between the two compounds. That is real new content.\n\nSoft spots, in proportion. First, the claim that interlayer sliding is a layer-number-independent mechanism for all Aurivillius oxides is asserted from two n=2 compounds. No n=1 or n=3 calculation or citation tests it. The topological argument is plausible but it is a hypothesis, and the paper presents it as a conclusion. This is the main issue. Second, the paper contains an internal tension: early on it calls sliding 'the dominant polarization mechanism', but Table V shows P_intra is more than twice P_shift in both compounds. The quantitative decomposition gives a more nuanced picture—sliding sets the polar channel, intralayer distortion carries the polarization—and the text should say that consistently. Third, the piezoelectric tensor in Table VI is taken from the authors' previous work, not computed here; that is fine if presented as reuse, but the paper does not make that clear in the table. Fourth, the P values are not checked against experiment or functional dependence; a short PBEsol or SCAN check on polarization would help. None of these are fatal. The central mechanism is well supported. The layer-number claim needs actual tests on other members, and the sliding-vs-intralayer language needs tightening.\n\nWho this is for: people working on layered oxide ferroelectrics, and anyone who wants a worked example of a multimode Landau fit to DFT. I would send it to peer review, because the quantitative decomposition is useful and the overgeneralization is fixable. I would not cite the layer-number claim in its current form, but I would cite the CBTO/CBNO energy decomposition.","headline":"Quantitative trilinear-coupling analysis of CBTO/CBNO is solid and useful, but the layer-number-independent sliding claim outruns the evidence.","tokens_in":22989,"tokens_out":1750,"would_cite":true,"duration_ms":17354,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The ferroelectric phase of the Aurivillius oxides CaBi2Ta2O9 and CaBi2Nb2O9 is stabilized not by a single polar soft mode but by the cooperative condensation of a polar displacement, an in-phase octahedral rotation, and an anti-phase octahe","keywords":["Aurivillius oxides","CaBi2Ta2O9","CaBi2Nb2O9","ferroelectricity","trilinear coupling","octahedral rotation/tilting","interlayer sliding","charge-transfer insulator"],"falsifier":"Calculate the layer-resolved sliding amplitude and the fitted trilinear coupling coefficient for a one-layer Aurivillius compound such as Bi2WO6 or a three-layer compound such as Bi4Ti3O12. If the rigid interlayer-sliding contribution is absent, or if the Gamma5-X2+X3- trilinear term no longer dominates the stabilization energy, the universal layer-count-independent mechanism is falsified.","tokens_in":1713,"feed_emoji":"⚡","tokens_out":6103,"duration_ms":124628,"temperature":0.7,"pith_summary":"The paper sets out to explain three experimentally puzzling facts about the Aurivillius ferroelectrics CaBi2Ta2O9 and CaBi2Nb2O9: their ferroelectric order survives to very high temperature, the polarization is large despite the small Ca cation, and the resistivity is exceptionally high. The proposed answer is that ferroelectricity here does not work like a conventional perovskite. A polar displacement mode, an in-phase octahedral rotation, and an anti-phase octahedral tilt condense together, and the trilinear term coupling all three dominates the energy balance—61% in the Ta compound and 81% in the Nb compound—creating a deep double well that explains the high Curie temperatures. The same layer-resolved analysis traces the polar displacement to a rigid in-plane sliding of the Bi2O2 layer relative to the perovskite block, which the authors argue is intrinsic to the alternating-layer topology and therefore general to Aurivillius oxides. The high resistivity, in turn, follows from the wide O 2p–metal d charge-transfer gap and weakly dispersive band edges, giving a single microscopic picture connecting structure, polarization, piezoelectricity, and insulation.","feed_headline":"Sliding layers, not a soft mode, drive Aurivillius oxides","feed_subtitle":"In CaBi2Ta2O9 and CaBi2Nb2O9, a sliding Bi2O2 layer plus octahedral rotation and tilt sets the polarization and explains 1200 K ordering.","key_machinery":"The central object is the trilinear coupling term C·Q(Gamma5-)·Q(X2+)·Q(X3-) in a Landau free-energy expansion for the I4/mmm to A21am transition. Q(Gamma5-), Q(X2+), and Q(X3-) are the amplitudes of the polar, in-phase rotation, and anti-phase tilt modes; the fitted coefficient C is negative and large, so the three otherwise competing distortions cooperate, deepening the ferroelectric double well and binding the three modes into a single condensation event. The supporting analytical tool is a layer-center decomposition of atomic displacements, which separates each atom's motion into rigid layer displacement (interlayer sliding) and internal distortion, making it possible to assign the polar","core_discovery":"The A21am ferroelectric ground state of CaBi2B2O9 (B=Ta,Nb) is not produced by a lone polar soft mode. The I4/mmm parent is unstable in three channels at once—polar Gamma5-, in-phase octahedral rotation X2+, anti-phase octahedral tilt X3-—and freezing all three together lowers the energy far more than any single mode. The fitted Landau expansion attributes 61% (CBTO) and 81% (CBNO) of the total stabilization to the trilinear term C·Q(Gamma5-)·Q(X2+)·Q(X3-). Decomposing the polar displacement shows a rigid interlayer sliding between the charged Bi2O2 layer and the perovskite block plus internal distortions inside each block; because the sliding is intrinsic to the alternating-layer stacking t","pith_inferences":["A direct calculation of layer-resolved rigid displacements in a one-layer Aurivillius member such as Bi2WO6 or a three-layer member such as Bi4Ti3O12 would test the 'layer-number-independent' claim; if the sliding amplitude or the trilinear coefficient changes character, the universality statement would need qualification.","The same rigid-block sliding channel might operate in other layered oxides with alternating charged blocks, such as Dion-Jacobson or Ruddlesden-Popper families, where an analogous interlayer displacement could be a generic route to polarity.","If octahedral tilts are the primary well-deepening degrees of freedom, epitaxial strain that modifies the tilt system should affect the Curie temperature more strongly than conventional polar-mode strain engineering; this is a testable prediction.","The orbital picture suggests a design rule: B-site cations with higher d states widen the charge-transfer gap and improve insulation, while cations with stronger off-centering improve polarization; layered architectures could in principle separate these two functions."],"forward_implications":["Because octahedral tilt and rotation, not the polar mode alone, control the depth of the ferroelectric well, the Curie temperature of Aurivillius oxides should be tunable through chemical pressure: replacing Ca with larger Sr or Ba weakens the tilts and lowers Tc, consistent with the reported 1196 K, 573 K, and 333 K series.","The spontaneous polarization has two distinct sources: interlayer sliding, which is nearly identical in CBTO and CBNO, and intralayer B-site off-centering, which is much stronger for Nb than Ta; this explains CBNO's larger polarization and d33 and isolates a common sliding baseline.","Bi ions and the Bi2O2 layer contribute around 45% (CBTO) and 35% (CBNO) of the spontaneous polarization, so treating the Bi2O2 layer as a rigid, inert block will misestimate both polarization and piezoelectric response.","The intrinsic insulating character is controlled by the O 2p–metal d charge-transfer gap and by flat, heavy band edges; the Ta compound's wider gap accounts for its experimentally higher resistivity, suggesting that raising the d-state energy is a design route to better insulation.","For polycrystalline ceramics, the intrinsic d33 and d24 coefficients are the dominant response channels, with the large d24 arising from easy polarization rotation in a flat in-plane energy landscape; texturing and poling should target these channels."],"fun_headline_variants":["Interlayer sliding, not soft modes, sparks ferroelectricity in Aurivillius","Three coupled distortions, not one, explain Aurivillius polarization","Bi2O2 sliding drives high-T ferroelectricity in layered oxides","Ferroelectricity in CaBi2B2O9 traces to Bi2O2 layer sliding","Layered oxide polarity from interlayer sliding, not a single mode"],"cache_read_input_tokens":24192,"weakest_assumption_plain":"The claim that interlayer sliding is a general, layer-number-independent mechanism for all Aurivillius oxides rests on calculations of only two n=2 compounds; no n=1, n=3, or n=4 member is computed to verify that the same sliding mode survives as the perovskite block thickness changes.","fun_headline_variants_meta":{"raw":{"variants":["Interlayer sliding, not soft modes, sparks ferroelectricity in Aurivillius","Three coupled distortions, not one, explain Aurivillius polarization","Bi2O2 sliding drives high-T ferroelectricity in layered oxides","Ferroelectricity in CaBi2B2O9 traces to Bi2O2 layer sliding","Layered oxide polarity from interlayer sliding, not a single mode"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000407,"raw_usage":{"total_tokens":2057,"prompt_tokens":957,"completion_tokens":1100,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":701,"completion_tokens_details":{"reasoning_tokens":1005}},"tokens_in":701,"tokens_out":1100,"duration_ms":9215,"temperature":1.0,"reasoning_tokens":1005,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T11:24:20.290263+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate the layer-resolved sliding amplitude and the fitted trilinear coupling coefficient for a one-layer Aurivillius compound such as Bi2WO6 or a three-layer compound such as Bi4Ti3O12. If the rigid interlayer-sliding contribution is absent, or if the Gamma5-X2+X3- trilinear term no longer dominates the stabilization energy, the universal layer-count-independent mechanism is falsified.","supporting_citations":[],"review_version":1}