{"id":"2213dbaa-b42d-41af-9802-b67985753064","arxiv_id":"2411.16497","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A 5-band Wannier model and a minimal 3-band model describe the stacking-dependent valence bands of RP1 perovskite monolayers and bilayers for Ca2TiO4, Sr2TiO4, and Ba2TiO4.","lead":"The authors build tight-binding and minimal three-band models of the valence bands of ultrathin perovskite layers and their stacked bilayers, with parameters for three titanates. This gives a foundation for modeling twisted oxide interfaces, extending moiré physics beyond van der Waals materials.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Arbitrary-shift interlayer model is not fully specified: ξ, κ, and fc are never given, so the claimed pathway to twisted-layer tight-binding models is not yet actionable.","rationale":"The reader's formal weakest assumption was the equivalence of AA-type stackings, but the reader's rationale also cited the missing ξ, κ, and fc values as a reason for the conditional verdict. I agree that the direct four-stacking results are credible, and I do not see an internal inconsistency that would overturn them. The more load-bearing issue, in my reading, is that the arbitrary-shift interpolation — the actual bridge from the four high-symmetry calculations to a twisted bilayer — is both under-specified and unvalidated. The paper gives general formulas but no values for the decay/cutoff functions, and Appendix C shows only that the special-stacking parameters can be expressed as particular evaluations of those formulas. This makes the Section V claim that the parametrization 'provides the necessary ingredients for full-scale tight-binding models of twisted layers' premature. I would keep the verdict at CONDITIONAL rather than ACCEPT because this gap is addressable with additional fitting and intermediate-shift DFT data, but it is a genuine missing step, not a cosmetic omission. The AA-equivalence assumption is also untested and would matter for a moiré potential, but it is secondary to the absence of any operational interpolation scheme. My concrete test targets the central gap directly; a second, complementary test would be to compute AAx and AAy stackings for STO and compare their VBM and interlayer splittings with r00, which would settle the AA-equivalence assumption as well.","tokens_in":105,"tokens_out":6360,"duration_ms":130520,"concrete_test":"Compute DFT-derived MLWF interlayer couplings for STO at several intermediate shifts, e.g., r = (0.25,0), (0,0.25), (0.25,0.25), using the same Wannier90 workflow as the paper. Then fit a single set of ξ, κ, t0, and fc from the general expressions Eqs. (26)-(29) and check whether these parameters reproduce all four corner values in Table III (r00, r11, r10, r01) within a defined tolerance such as 10 meV, and whether the resulting band structures at intermediate shifts match DFT near the valence band maximum. If no single parametrization can do this, the configuration-space model is not validated for arbitrary twist.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The four high-symmetry stacking results are credible and directly supported by DFT comparisons. The load-bearing gap is in the extension to arbitrary shifts, which is the stated route to twisted-layer models. In Section IV C, the general interlayer hoppings in Eqs. (26)-(29) depend on decay constants ξ and κ and a cutoff function fc, but no numerical values or functional form for these quantities are provided anywhere in the paper or supplement. Appendix C only fixes combinations: from Eq. (C2), u2 = t0 e^{-κ} and u3 = t0 e^{-κ} e^{-ξ}, so e^{-ξ} = u3/u2 is determined, but t0 and κ remain undetermined because t0 is never specified. Furthermore, no DFT or Wannier-based check at intermediate shifts is reported, so the exponential-plus-cutoff interpolation is an untested postulate. If the deliverable is 'the necessary ingredients for full-scale tight-binding models of twisted layers' as claimed in Section V, this is the critical step: without ξ, κ, fc, and an intermediate-shift validation, the four-stacking Hamiltonians in Table III cannot be assembled into a moiré Hamiltonian. The direct four-stacking results are not affected, but the transferable model is incomplete. The assumed equivalence of the four AA-type stackings in Section IV A is a related but secondary untested assumption, since AAx, AAy, and the unit-cell-center AA stacking are never computed.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript develops tight-binding models for the valence bands of monolayer and bilayer Ruddlesden-Popper (RP1) perovskite oxides, specifically Ca2TiO4, Sr2TiO4, and Ba2TiO4. A five-band maximally localized Wannier function (MLWF) model is constructed for the mirror-odd valence states, and a reduced three-band effective model is fitted to the top of the valence band. For bilayers, the authors compute generalized stacking fault energies and heights, derive interlayer coupling Hamiltonians for four high-symmetry stackings (AA, AB, DWx, DWy) in both the five-band and three-band forms, and compare the resulting band structures with DFT. They then propose a configuration-space parametrization of the interlayer hopping as a function of relative layer shift, intended as the foundation for future twisted-bilayer models. The direct four-stacking results are presented as the main technical achievement, with the configuration-space extension framed as the route to full moiré Hamiltonians.","tokens_in":22539,"tokens_out":2106,"duration_ms":21366,"significance":"If the results hold, this is a valuable step toward extending twistronics beyond van der Waals materials to strongly bonded oxide layers. The paper's concrete strengths are the direct DFT-based construction of MLWF tight-binding models for three materials, the explicit stacking-dependent interlayer Hamiltonians for the four high-symmetry stackings, and the GSFE/GSFH fits. The comparison between the MLWF bands and DFT bands is close (Fig. 3b for STO; Figs. S4-S7 for CTO/BTO), and the paper makes falsifiable predictions for the stacking-dependent valence band structure that can be tested by future experiments or calculations. However, the claimed pathway to twisted-bilayer models rests on an arbitrary-shift interpolation whose parameters are not fully specified, which limits the transferability of the model as presented.","major_comments":[{"comment":"The general interlayer hopping parametrization for arbitrary shifts depends on the decay constants ξ and κ and the cutoff function fc(x), but no numerical values or explicit functional form for these quantities are provided anywhere in the manuscript or supplement. Appendix C only fixes combinations: from Eq. (C2), u2 = t0 e^{-κ} and u3 = t0 e^{-κ} e^{-ξ}, so e^{-ξ} = u3/u2 is determined, but t0 and κ remain undetermined because t0 is never specified. Since the stated deliverable in Section V is \"the necessary ingredients for full-scale tight-binding models of twisted layers,\" these missing constants make the arbitrary-shift model not actionable. The authors should provide explicit values (or a fitting procedure) for ξ, κ, and fc, and should test the exponential-plus-cutoff interpolation against DFT or MLWF calculations at intermediate shifts.","section":"§IV.C, Eqs. (26)-(29) and Appendix C"},{"comment":"The assumption that the four AA-type stackings are essentially equivalent is stated but never tested. AAx, AAy, and the unit-cell-center AA stacking are never computed, even though they are identified as distinct in Fig. 4(b) and the moiré cell reduction relies on their equivalence. A single DFT or MLWF calculation for the AAx or AAy stacking would directly test this load-bearing assumption; without it, the configuration-space model may misrepresent the actual moiré potential.","section":"§IV.A"},{"comment":"The simplified three-band model is numerically fitted to DFT bands (Section III) and its interlayer parameters are described as \"optimized by hand\" (Section IV.C), rather than derived from the MLWF model or from the general interpolation. This means the three-band model has no independent predictive content beyond the fitted stackings; its agreement with DFT in Fig. 7 is a fitting result, not a validation. The manuscript should clarify the status of these parameters and ideally show that they are consistent with the general shift-dependent expressions in Eqs. (26)-(29), or with a Schur-complement reduction of the five-band model.","section":"§III and §IV.C, Table IV"}],"minor_comments":[{"comment":"The caption for Fig. 1 lists panels (d) and (e) as bulk cubic SrTiO3 and bulk RP1 Sr2TiO4 band structures, but the text in Section II does not explicitly reference these panels; adding a pointer would improve readability.","section":"§II, Fig. 1"},{"comment":"The convention {r = 0, Rj = 0} ⇒ xj = 0, yj = 0 is stated, but similar care is needed for the other atoms j that may also produce zero denominators (e.g., when rj = 0 for nonzero r due to symmetry); the current text only mentions the r = 0, Rj = 0 case.","section":"§IV.C, Eq. (24)"},{"comment":"The table lists λD = w7/w6, but w7 and w6 are positive for all materials, while λD is presumably a signed ratio that can affect the band structure; a brief note on the allowed range or physical meaning of λD would be useful.","section":"§IV.C, Table III"},{"comment":"The line-cut labels in Fig. 6 use \"AB / DX\" and \"AA\" in the x-axis ticks, but the text defines DWx and DWy; using consistent notation (DWx/DWy) would avoid confusion.","section":"§IV.B, Fig. 6"}],"recommendation":"major_revision","confidential_remarks":"The reader's concern about the incompleteness of the arbitrary-shift interpolation is well-founded and is the main technical gap. The direct four-stacking results are solid and could support a publication on their own, but the paper's stated goal of providing the foundation for twisted-layer models requires the missing parameters (ξ, κ, fc) and an intermediate-shift validation. The AA-equivalence assumption is also untested and should be demonstrated or qualified. The paper is within scope for the journal and the literature coverage is adequate, but the presentation would benefit from clearer separation between the fitted models and derived models."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick read on arXiv:2411.16497. The four-stacking tight-binding results are the real content, and they are credible. The authors build a symmetry-adapted 5-band MLWF model for the M-odd valence bands of RP1 monolayers of CTO, STO, and BTO, plus a reduced 3-band effective model, and interlayer Hamiltonians for AA, AB, DWx, and DWy stackings. The MLWF bands track DFT nearly indistinguishably for STO, and the parameter tables are a genuine resource. The GSFE/GSFH Fourier fits are also useful. That part is solid.\n\nThe soft spot is the generalization to arbitrary shifts. Equations (26)–(29) define the interlayer hoppings in terms of decay constants ξ, κ and a cutoff fc, but no values or functional forms are given anywhere, and Appendix C only fixes combinations like u2 = t0 e^{−κ} and u3 = t0 e^{−κ} e^{−ξ}, leaving t0, κ, ξ undetermined. No intermediate-shift DFT or Wannier check is reported, so the exponential-plus-cutoff interpolation is an untested postulate. The paper's Section V claims these are the 'necessary ingredients for full-scale tight-binding models of twisted layers.' That claim does not yet hold. Without ξ, κ, fc, and an intermediate-shift validation, the four-stacking tables cannot be assembled into a moiré Hamiltonian.\n\nThe secondary assumptions are more minor: the equivalence of the four AA-type stackings is asserted rather than computed, and the 3-band interlayer parameters are 'optimized by hand' rather than derived. These are plausible and addressable. The circularity concern is not real: fitting to DFT and then comparing is standard model construction, and the paper doesn't overstate it.\n\nBottom line: the four-stacking models are a real contribution; the arbitrary-shift extension is incomplete. The paper deserves a serious referee, but the revision should supply the missing parameters and a check at an intermediate shift. If you work on twisted oxides, the tables are worth citing.","headline":"Solid four-stacking tight-binding models for RP1 perovskites, but the claimed pathway to twisted-layer models is missing the arbitrary-shift parameters.","tokens_in":23012,"tokens_out":2301,"would_cite":true,"duration_ms":20433,"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 stacking-dependent tight-binding model captures the valence bands of perovskite bilayers and supplies the foundation for twisted oxide layers.","keywords":["twistronics","perovskite oxides","Ruddlesden-Popper phases","tight-binding model","maximally localized Wannier functions","moiré patterns","stacking-dependent electronic structure","interlayer coupling"],"falsifier":"Calculate the valence bands of a small-angle twisted Sr2TiO4 bilayer with full ionic relaxation and compare them band-by-band with the model's prediction built from these shift-dependent couplings; if the low-energy bands differ by more than a few tens of meV, or if relaxed AAx and AAy stackings produce materially different interlayer couplings, the configuration-space foundation is falsified.","tokens_in":21893,"feed_emoji":"🔄","tokens_out":7568,"duration_ms":68622,"temperature":0.7,"pith_summary":"Twistronics has mostly been a van der Waals story: weakly bound layers twisted against each other to make moiré patterns. This paper tries to move that program into oxide perovskites, where interlayer bonding is much stronger. It claims that the valence-band physics of a Ruddlesden-Popper RP1 layer, and of bilayer stacks in the four high-symmetry configurations AA, AB, DWx, and DWy, can be captured by a five-band maximally localized Wannier tight-binding model, and at even lower cost by a three-band effective model whose interlayer couplings depend on the relative in-plane shift between the layers. For Sr2TiO4 the model bands are reported to be nearly indistinguishable from the density-functional results, and the same construction is parametrized for Ca2TiO4 and Ba2TiO4. The payoff, if the claim holds, is a ready-made ingredient for modeling twisted perovskite bilayers, where the local stacking at each point of the moiré cell controls the electronic structure.","feed_headline":"Five-band model captures perovskite bilayer stacking physics","feed_subtitle":"A five-band Wannier model plus shift-dependent couplings reproduces DFT bands and sets up oxide moiré physics.","key_machinery":"The load-bearing object is the mirror-odd sector of the oxygen-p/titanium-d valence manifold. Symmetry under reflection through the mid-plane of the BO6 octahedron separates the 17 atomic orbitals into states that couple between layers and states that do not; keeping only the five mirror-odd combinations reduces the valence manifold to a manageable tight-binding basis. The monolayer Hamiltonian is written with functions fk(a)=2cos(k·a) and gk(a)=2i sin(k·a), and the interlayer coupling at each stacking is a matrix whose entries are sums over neighbor atoms weighted by Gaussian decay in in-plane distance and interlayer separation. The minimal model is obtained by a Schur-complement reduction of the lower two bands, leaving a three-band-per-layer Hamiltonian with renormalized parameters. This construction is what makes twistronics tractable: the same shift-dependent coupling matrices can be evaluated at every local stacking in a twisted moiré cell.","core_discovery":"The paper's central discovery is that the entire interlayer physics of ultrathin perovskite bilayers lives in five mirror-odd valence states: the hybrids ψ1,z and φ1,ℓ formed from apical oxygen p orbitals and the titanium dz2 orbital, together with the in-plane oxygen pz combinations χ±1,z. A 15-parameter tight-binding Hamiltonian in this basis reproduces the DFT valence bands of the monolayer essentially exactly for STO and well for CTO and BTO. For bilayers, the paper writes the interlayer coupling as explicit matrices at the four special shifts r00 (AA), r11 (AB), r10 (DWx), and r01 (DWy), with momentum dependence in cosine and sine combinations of the shift, and shows that the resulting bands match the DFT bilayer bands. The full 10-band model can be reduced to three bands per layer by integrating out the lower two mirror-odd states, producing a minimal model that retains the stacking-dependent valence-band maximum away from Γ. The authors state that this shift-dependent parametrization provides the necessary ingredients for full-scale tight-binding and continuum models of twisted perovskite bilayers.","pith_inferences":["If ionic relaxation inside a moiré cell is strong, the paper's assumption that the four AA-type stackings behave equivalently may fail; computing interlayer couplings for relaxed bilayers at AAx and AAy would test this directly, since the paper explicitly leaves relaxation to future work.","The same symmetry decomposition could be applied to the conduction band, which the paper notes is dominated by Ti dxy orbitals and is not modeled here; a paired valence and conduction model would open electron-doped twisted oxides.","The large stacking-energy differences suggest that twisted oxide bilayers could develop reconstructed stacking domains whose electronic properties differ more sharply than in van der Waals moirés; this is an inference, not a result of the paper.","The reported near-independence of the AA stacking's electronic states, combined with large interlayer separation, implies that AA regions in a moiré cell act almost like decoupled monolayers; a continuum model built from these couplings would predict strong spatial modulation of band hybridization across the moiré cell."],"forward_implications":["The tabulated monolayer and interlayer parameters are directly usable to build tight-binding models of twisted RP1 oxide bilayers in arbitrary commensurate moiré cells.","The shift-dependent interlayer couplings provide the moiré potential for continuum models of twisted perovskite oxides, following the same route used for graphene and transition-metal dichalcogenides.","Because the AA stacking is more than 1 eV per unit cell higher in energy than AB and has nearly twice the interlayer separation, local regions of a moiré cell will have sharply different interlayer coupling, so stacking domains should be electronically distinct.","In all four stackings the interlayer interaction moves the valence-band maximum into mirror-odd states away from Γ, so hole-doped twisted perovskite bilayers should show strongly stacking-sensitive transport and optical response."],"supporting_citations":[{"why":"supplies the maximally localized Wannier construction used to define the tight-binding basis from DFT.","marker":"[51]"},{"why":"introduces the configuration-space approach for representing local stacking environments of twisted layers.","marker":"[26]"},{"why":"provides the minimal-model framework for weakly interacting bilayers that the oxide model extends to strongly coupled ionic layers.","marker":"[49]"},{"why":"gives the continuum-model machinery that the shift-dependent parametrization is designed to feed.","marker":"[53]"},{"why":"defines the generalized stacking fault energy used here to map the stacking dependence of energy and interlayer height.","marker":"[57]"},{"why":"demonstrates experimentally that moiré patterns form in twisted stacks of perovskite oxide nanomembranes.","marker":"[41]"},{"why":"reports twisted freestanding BaTiO3 layers with ferroelectric vortex patterns, the experimental system the model targets.","marker":"[42]"},{"why":"provides a prior first-principles treatment of twisted BaTiO3 that the present shift-dependent models complement.","marker":"[45]"}],"fun_headline_variants":["Oxide bilayers get a stacking-dependent band model","Three effective bands capture perovskite bilayer stacking","Stacking control of oxide bands via 15-parameter model","Mirror-odd states set perovskite bilayer interlayer physics","Perovskite twistronics gains a first-principles model"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The derivation depends on the premise that every local environment in a twisted bilayer is faithfully represented by the untilted, unrelaxed stackings computed here, and specifically that the four AA-type stackings have nearly identical electronic effects.","fun_headline_variants_meta":{"raw":{"variants":["Oxide bilayers get a stacking-dependent band model","Three effective bands capture perovskite bilayer stacking","Stacking control of oxide bands via 15-parameter model","Mirror-odd states set perovskite bilayer interlayer physics","Perovskite twistronics gains a first-principles model"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000561,"raw_usage":{"total_tokens":2650,"prompt_tokens":918,"completion_tokens":1732,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":534,"completion_tokens_details":{"reasoning_tokens":1669}},"tokens_in":534,"tokens_out":1732,"duration_ms":11095,"temperature":1.0,"reasoning_tokens":1669,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T13:03:03.348744+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate the valence bands of a small-angle twisted Sr2TiO4 bilayer with full ionic relaxation and compare them band-by-band with the model's prediction built from these shift-dependent couplings; if the low-energy bands differ by more than a few tens of meV, or if relaxed AAx and AAy stackings produce materially different interlayer couplings, the configuration-space foundation is falsified.","supporting_citations":[{"cited_title":"Stacking-engineered ferroelectricity and multiferroic order in van der Waals magnets","cited_arxiv_id":"2405.20069","evidence_quote":"introduces the configuration-space approach for representing local stacking environments of twisted layers."},{"cited_title":"Bennett, D","cited_arxiv_id":null,"evidence_quote":"provides the minimal-model framework for weakly interacting bilayers that the oxide model extends to strongly coupled ionic layers."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the continuum-model machinery that the shift-dependent parametrization is designed to feed."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"defines the generalized stacking fault energy used here to map the stacking dependence of energy and interlayer height."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"demonstrates experimentally that moiré patterns form in twisted stacks of perovskite oxide nanomembranes."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"reports twisted freestanding BaTiO3 layers with ferroelectric vortex patterns, the experimental system the model targets."},{"cited_title":"Square Moir\\'e Superlattices in Twisted Two-Dimensional Halide Perovskites","cited_arxiv_id":"2312.16679","evidence_quote":"provides a prior first-principles treatment of twisted BaTiO3 that the present shift-dependent models complement."}],"review_version":1}