{"id":"26f7c58c-f8c3-4f87-abe2-af1ca3bf5a31","arxiv_id":"2607.18386","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Every symmetry-indicated fragile phase in any 2D wallpaper group is adiabatically connected to an atomic insulator in some finite symmetry-compatible supercell, with the minimal supercell index tabulated for every symmetry class.","lead":"Fragile topological phases in two-dimensional crystals become ordinary atomic insulators after a sufficiently large, symmetry-preserving enlargement of the unit cell — for every wallpaper group, with or without spin-orbit coupling or time-reversal symmetry. This finite 'stability depth' means physical responses that survive unit-cell refinement cannot uniquely identify fragile topology, and commensurate moiré or charge-density-wave potentials should trivialize it.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The dynamic conclusion (adiabatic connectivity / small-perturbation trivialization) does not follow from the symmetry-data membership test of Eq. (30); Sec. IV and Ref. [67] document exactly the missing completeness assumption.","rationale":"The reader's weakest_assumption identifies exactly this inference from symmetry-data atomicity to adiabatic connectivity, supported by the paper's own Sec. IV disclaimer and Ref. [67]. I agree this is the most load-bearing concern because it is the bridge from the proven Diophantine criterion to the advertised physical conclusion. The alternative mechanical concern (unverified vanishing of BR·M_orb·m_ker in Eq. (27)) is important but secondary: it affects the correctness of the N* tables, while the completeness concern affects whether those tables have the physical meaning claimed in the abstract. The paper's worked examples (p2 and p4mm) are internally consistent and give real confidence in the symmetry-data computation, so I do not call for rejection. However, because the abstract and title state a result stronger than the proof, the conditional verdict should stand until the completeness step is either proven or the central claim is restated as a symmetry-data trivialization theorem.","tokens_in":36938,"tokens_out":7828,"duration_ms":68617,"concrete_test":"Construct a tight-binding model in spinless p2 with the fragile symmetry-data vector of Eq. (6) and a nonzero Euler-class obstruction protected by C2zT (e.g., a real two-band Hamiltonian with that symmetry), and apply the 2×1 supercell enlargement and the symmetry-preserving perturbation of Sec. S3. Numerically compute the Wilson-loop spectrum and a real-space Euler marker in the supercell after the perturbation that makes the symmetry data atomic. If a nonzero Euler invariant remains, the phase is not adiabatically connected to an atomic insulator, refuting the inference from Eq. (30) to the abstract's claim. If all such models trivialize, the concern is resolved and the tables can be trusted as dynamical statements.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in the abstract — 'every symmetry-indicated fragile phase is adiabatically connected to an atomic insulator... trivialized by an arbitrarily small symmetry-preserving perturbation' — is a dynamical statement about Hamiltonians. The proof, however, only establishes Eq. (30): after a suitable supercell enlargement, the symmetry-data vector M_v·v_FR lies in the atomic cone. That is a Diophantine condition on irrep multiplicities at high-symmetry momenta. It certifies that the symmetry data is realizable by some atomic-limit model, but it does not certify that every Hamiltonian with that data is continuously deformable to an atomic model without closing the gap or breaking symmetry. The paper itself flags this in Sec. IV: 'Since we focus on symmetry-indicated phases, we do not consider the possibility of additional topology not captured by symmetry data'; and Sec. VI leaves non-symmetry-indicated topology (e.g., Euler class) as an open problem. The cited Ref. [67] ('Topology invisible to eigenvalues') shows that eigenvalue data can under-determine the phase. Thus the finite-stability conclusion applies to the symmetry-data vector, not necessarily to the physical phase: a fragile phase carrying additional non-indicated invariants could remain non-adiabatic to an atomic insulator even after the indicated data becomes atomic. Without a completeness theorem for the wallpaper groups considered, the title's sweeping 'fragile topology is unstable' is an overstatement.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper develops a Hilbert-basis framework for analyzing the fate of symmetry-indicated fragile topology under symmetry-preserving unit-cell enlargement in all 2D wallpaper groups. It represents unit-cell enlargement as an integer linear map on symmetry-data vectors (Eq. (28)), then checks, for each fragile root, whether its image lies in the atomic cone (Eq. (30)). The main quantitative output is the minimal enlargement index N* for each symmetry class (Tables I–II), with the p2 case worked out in detail: the fragile root of Eq. (6) becomes atomic in the 2x1 supercell via the decomposition of Eq. (34). The authors conclude that every symmetry-indicated fragile phase has finite stability under translation refinement and can be trivialized by an arbitrarily small symmetry-preserving perturbation.","tokens_in":37123,"tokens_out":4879,"duration_ms":44105,"significance":"If the symmetry-data statement is taken as the result, the paper provides a systematic and useful classification: for every 2D wallpaper group, all symmetry-indicated fragile roots become atomic at some finite, explicitly bounded supercell index. The p2 example is fully checkable and internally consistent, the Hilbert-basis method is rigorous and builds on the external classification of Ref. [60], and the N* tables are concrete and falsifiable. The physical electron–positron picture in Sec. II is pedagogically illuminating. However, the paper's central advertised conclusion is broader than the proof: the proof establishes a condition on irrep multiplicities, not adiabatic connectivity of Hamiltonians. The paper itself acknowledges this limitation in Sec. IV and Sec. VI, which makes the abstract's dynamical phrasing an overstatement.","major_comments":[{"comment":"The proof establishes only that M_v · v_FR lies in the atomic cone. This is a Diophantine condition on symmetry-data vectors, not a statement about Hamiltonians. The abstract's conclusion that every symmetry-indicated fragile phase 'is adiabatically connected to an atomic insulator ... and can therefore be trivialized by an arbitrarily small symmetry-preserving perturbation' requires completeness of the symmetry data for the phase. The paper itself disclaims this in Sec. IV ('we do not consider the possibility of additional topology not captured by symmetry data') and Sec. VI leaves Euler-class fragile phases open; Ref. [67] shows eigenvalue data can under-determine topology. A phase carrying an additional non-indicated invariant could remain non-adiabatic to an atomic insulator even after its indicated data becomes atomic. Please either restrict the claims to symmetry-data trivializatio","section":"Abstract and Sec. IV, Eq. (30)"},{"comment":"The vanishing of BR·M_orb·m_ker is asserted as 'explicitly verified' for all wallpaper groups and enlargements, but no proof or code is provided. This property is load-bearing because Eq. (28) defines M_v only when this term vanishes; the exhaustive N* tables in Tables I–II depend on it for every group. The same applies to the consistency condition BR·M_orb·q_adia = 0 in Eq. (29), which is stated without proof. Please provide a general argument (e.g., via the Smith normal form of BR) or ship the verification code/data so that the exhaustive computational claim is independently checkable.","section":"Sec. IV, Eq. (27)"}],"minor_comments":[{"comment":"The text says 'Consistency requires that BR·M_orb·q_adia = 0' but does not explain why this condition follows from the physical requirement that v' be unchanged under adiabatic deformations. A short derivation would improve clarity.","section":"Sec. IV"},{"comment":"The example solution m = (-1,0,1,0,1,0,1,0,0) and the kernel vector are correct, but the statement 'every representative obtained by adding a kernel vector still contains at least one negative multiplicity' is not proved for all kernel vectors. A one-line argument or a reference to the general Hilbert-basis criterion would help.","section":"Sec. II, Eqs. (7)-(8)"},{"comment":"The notation N_FR appears as NFR in the tables; please unify. Also, the caption 'Allowed N_sc' lists the smallest three nontrivial indices, but the text should state this explicitly in the main body as well.","section":"Tables I–II"},{"comment":"There are several grammatical slips, e.g., 'fragile topology has is eventually unstable' in Sec. VI and 'Hilbert-basis' inconsistent hyphenation. These do not affect the physics.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The core symmetry-data computation appears sound and is a useful contribution, but the advertised adiabatic-connectivity claim is not proven. The lack of a verifiable computational artifact for Eq. (27) is a reproducibility concern for the exhaustive tables. I recommend major revision: either substantially soften the dynamical claims or provide the missing completeness proof/code."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague — read this one if you care about fragile topology. The core theorem is real: for all 2D wallpaper groups, every symmetry-indicated fragile phase becomes atomic under some finite symmetry-preserving unit-cell enlargement, with a tabulated minimal index N*. That is new, as far as I can tell; Ref. [60] classified fragile phases but did not treat enlargements systematically. The framework — Hilbert bases for atomic/fragile symmetry data, unit-cell enlargement as an integer linear map — is an honest extension of existing tools. The p2 example is genuinely checkable: I worked through BR·m, the 2×1 supercell map, and the decomposition of v′ into atomic generators, and it all checks out. The p4mm example in the SM is also detailed and gives a concrete picture of why the N=2 enlargement fails (C4 forces all four 4d orbitals toward one maximum). The N* tables are the kind of thing people will want to cite.\n\nSoft spots. First, the title says \"Fragile Topology is Unstable\", but the theorem is about symmetry-indicated fragile topology. The paper is honest about this in Sec. IV — it does not consider additional topology not captured by symmetry data — and Sec. VI leaves non-indicated topology (Euler class, etc.) open. Ref. [67] shows eigenvalue data can under-determine the phase. So the abstract's phrase \"adiabatically connected to an atomic insulator\" is a jump from the proven statement, which is that the symmetry-data vector becomes atomic. For a phase carrying a non-indicated invariant, the enlargement may trivialize the indicated data while the phase remains non-atomic. That is a real gap, and the title should be qualified. Second, the kernel-vanishing condition Eq. (27) is asserted as \"explicitly verified\" for all groups, with no proof or code, and the full enumeration behind Tables I–II is not shipped. I cannot verify the p4gm N*=9 or p6 N*=7 entries without the per-group data. This makes me conditional rather than accepting outright.\n\nNone of this is a demonstrated error. The p2 and p4mm examples give real confidence the machinery works. I would send this to a serious referee — the result is significant enough to deserve the time, and the missing pieces (code/data, a tightened title, a clearer statement of the completeness assumption) are fixable in revision. If I were working on fragile topology, I would cite it with the \"symmetry-indicated\" qualifier.","headline":"Solid, checkable result on symmetry-indicated fragile phases in 2D, but the title overstates what the math actually proves.","tokens_in":37795,"tokens_out":3268,"would_cite":true,"duration_ms":28525,"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":"Every symmetry-indicated fragile phase in a two-dimensional crystal becomes an atomic insulator after a finite unit-cell enlargement.","keywords":["fragile topology","wallpaper groups","symmetry indicators","band representations","unit-cell enlargement","Hilbert basis","atomic insulators","translation symmetry refinement"],"falsifier":"Construct an explicit tight-binding model for one of the listed wallpaper groups at the predicted minimal supercell index whose symmetry-data vector becomes atomic, but whose Wilson-loop winding or corner charge remains nontrivial under every small symmetry-preserving perturbation; that would refute the claimed adiabatic trivialization.","tokens_in":36608,"feed_emoji":"🔁","tokens_out":5213,"duration_ms":45258,"temperature":0.7,"pith_summary":"This paper asks whether fragile topological phases survive when the lattice's unit cell is enlarged while all crystalline symmetries are preserved. Treating enlargement as band folding, it shows that every symmetry-indicated fragile phase in every two-dimensional wallpaper group—with or without spin-orbit coupling or time-reversal symmetry—becomes atomic after a finite supercell enlargement. An arbitrarily small symmetry-preserving perturbation can then connect it to an ordinary atomic insulator. The proof reduces this to integer linear algebra: each enlargement induces a linear map on symmetry-data vectors, and trivialization means the image lands in the cone of atomic data. The paper tabulates the minimal supercell index at which all fragile roots vanish for each symmetry class, showing that fragile topology has only finite stability under translation refinement.","feed_headline":"Unit-cell enlargement trivializes 2D fragile phases","feed_subtitle":"For all wallpaper groups, symmetry-indicated fragile bands become atomic at a finite supercell size.","key_machinery":"The machinery is the Hilbert-basis decomposition of symmetry-data vectors. Atomic phases form a cone generated by a finite Hilbert basis; atomic-plus-fragile phases form a larger cone whose extra generators are called fragile roots. Unit-cell enlargement is represented by an integer matrix that reorganizes Wannier orbitals into the supercell, inducing a linear map on irrep-multiplicity vectors. Trivialization is decided by a Diophantine criterion: the image of each fragile root must decompose as a nonnegative integer combination of atomic generators. A complementary electron–positron picture supplies the physical intuition for why annihilation is possible after enlargement.","core_discovery":"The central claim is that symmetry-indicated fragile topology in 2D is always finitely unstable under translation-symmetry refinement. A fragile phase is one whose symmetry data is a formal difference of band representations with a mandatory negative term—an 'electron–positron' pair in real space. Enlarging the unit cell folds bands and re-labels Wyckoff positions, which the paper encodes as an integer linear map on orbital multiplicities and hence on symmetry-data vectors. The criterion for trivialization is whether applying this map to every fragile root produces a nonnegative integer combination of atomic Hilbert-basis generators. Computing this for all wallpaper groups, spinless and spin","pith_inferences":["If this is right, 'fragile' is not an intrinsic property of a band structure but a property relative to a chosen translation group; any complete classification should state the unit-cell scale alongside the phase label.","A testable extension: driving a real material through a symmetry-preserving superlattice potential at the predicted index should collapse fragile-response signatures such as Wilson-loop winding while keeping the gap open.","The same Hilbert-basis and cone method could be applied to symmetry-changing enlargements that connect different wallpaper groups, where the fragile-to-atomic mapping may differ from the symmetry-preserving case.","For non-symmetry-indicated fragile phases the paper offers only case-by-case adiabatic analysis, so a general theorem covering all fragile phases remains an open, plausible conjecture."],"forward_implications":["A symmetry-indicated fragile phase can always be destroyed by a commensurate superlattice or charge-density-wave potential whose period realizes the predicted supercell index, while preserving the wallpaper symmetry.","Any physical response that is invariant under symmetry-compatible unit-cell enlargement cannot uniquely characterize fragile topology, since the phase becomes indistinguishable from an atomic insulator at that scale.","In higher-symmetry groups such as spinless p6, fragile phases survive small enlargements but vanish at a larger finite index, so the trivialization scale is explicit and can be targeted in experiment.","Euler-class fragile phases protected by C2zT symmetry are expected to trivialize under sufficiently large enlargement, consistent with the general framework.","Whether finite enlargement stability persists in three-dimensional space groups remains an open question."],"fun_headline_variants":["2D fragile phases always trivialize in larger supercells","Translation refinement kills every fragile phase in 2D","Fragile topology dies at finite supercell size for all wallpaper groups","Fragile bands: always atomic in a suitable supercell"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that having the same symmetry-indicated irrep multiplicities as an atomic insulator in the supercell guarantees an adiabatic, gap-preserving path to that atomic insulator—an assumption the paper explicitly leaves unproven and which is known to fail for topology invisible to eigenvalues.","fun_headline_variants_meta":{"raw":{"variants":["2D fragile phases always trivialize in larger supercells","Translation refinement kills every fragile phase in 2D","Fragile topology dies at finite supercell size for all wallpaper groups","Fragile bands: always atomic in a suitable supercell"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000235,"raw_usage":{"total_tokens":1350,"prompt_tokens":767,"completion_tokens":583,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":511,"completion_tokens_details":{"reasoning_tokens":513}},"tokens_in":511,"tokens_out":583,"duration_ms":5311,"temperature":1.0,"reasoning_tokens":513,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-01T15:34:05.234331+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Construct an explicit tight-binding model for one of the listed wallpaper groups at the predicted minimal supercell index whose symmetry-data vector becomes atomic, but whose Wilson-loop winding or corner charge remains nontrivial under every small symmetry-preserving perturbation; that would refute the claimed adiabatic trivialization.","supporting_citations":[],"review_version":1}