{"id":"1f183de7-8504-4489-bd73-97387e863612","arxiv_id":"2504.20140","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Allowing a small fraction of electrons to occupy higher bands converts the predicted one-third fractional Chern insulator in rhombohedral pentalayer graphene into a K charge density wave, and basis optimization does not restore the FCI.","lead":"This paper shows in a model of five-layer graphene that the fractional Chern insulator at one-third filling is destroyed by tiny multi-band mixing and replaced by a charge density wave. It also introduces an iterative method to reduce the truncation error in multi-band exact diagonalization, which still fails to restore the fractional phase.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The FCI-to-CDW transition rests on small occupation caps, while the exact three-band calculation shows neither FCI nor CDW, so the central claim is not yet established.","rationale":"The reader's weakest assumption identifies exactly the truncation dependence of the CDW conclusion: the small caps {1,0} and {2,0} are used to claim an FCI-to-CDW transition, while the full {6,6} calculation shows no gap at either FCI or CDW counting. My read agrees with that assessment. The most load-bearing single concern is not the existence of a CDW at small caps—which the PES and structure factor support internally—but the inference from those capped results to the physical ground state. Since truncation breaks basis invariance, the CDW could be an artifact of the HF basis and the occupation cap. The ED iteration method, which the authors honestly test, does not restore agreement with the exact ground states; it even slightly lowers the overlap. Thus the strongest version of the central claim, that the true 1/3 CN ground state is a K-CDW and that current models fail, is not established. The paper remains a careful and honest study, and the CONDITIONAL verdict from the reader remains appropriate; my stress-test does not move that verdict.","tokens_in":73423,"tokens_out":4173,"duration_ms":44858,"concrete_test":"For the 9×2, ν=1/3 CN system, compute at caps {1,0}, {2,0}, {3,2}, and {6,6}: the K-peak height S(q=K), the PES gap below 45 CDW states, and the overlap between each truncated ground-state manifold and the exact {6,6} manifold. If S(K) and the CDW gap extrapolate to zero as the exact limit is approached, or if the {1,0}/{2,0} manifolds have low overlap with {6,6}, then the CDW transition is a truncation artifact rather than a property of the model. Repeating the same check on 15×1 and 21×1 untruncated systems would test system-size dependence.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that in the CN scheme at ν=1/3 the single-HF-band FCI is replaced by a K-CDW once a small fraction of electrons occupies higher bands. The evidence is the PES at truncations {1,0} and {2,0} (Figs. 3b–4c), where a gap opens at CDW counting N_CDW(6,2)=45 and S(q) peaks at K. The load-bearing assumption is that these small caps represent the physical weak-band-mixing regime. The paper itself provides the strongest counter-evidence: for the 9×2 system the full three-band calculation at {6,6} is feasible, and there the PES with N_A=2 is not gapped at either FCI or CDW counting, concomitant with the absence of any gap in the energy spectrum (Sec. III A). No finite-truncation extrapolation is given; the CDW gap for {1,0}→{2,0} is not shown to converge to the exact limit, and in fact the sequence to {6,6} loses the gap. Moreover, because truncation breaks basis invariance (Sec. II C), the CDW could be an artifact of the HF basis used, not of the Hamiltonian. The ED iteration method was introduced to address this, but it only modestly raises n_0,tot (~6%) and slightly decreases the overlap with the exact ground states (Fig. 9 insets, App. E2), so it does not rescue the small-cap CDW as the exact ground state. The honest conclusion is that within this continuum model the 1/3 CN ground state is neither FCI nor CDW at the accessible exact limit; the FCI-to-CDW transition is a statement about a truncated Hilbert space unless truncation dependence is established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies multi-band exact diagonalization of rhombohedral pentalayer graphene on hBN at fractional fillings 1/3 and 2/3, in both the charge-neutrality (CN) and average (AVE) interaction schemes. Working in a Hartree-Fock basis and truncating the many-body Hilbert space by occupation caps on the two higher HF bands, the authors find that at 1/3 filling in the CN scheme the particle entanglement spectrum (PES) of the three lowest states at the FCI momenta develops a gap at CDW counting already for the smallest caps, {1,0} and {2,0}, and that the structure factor peaks at K. They interpret this as a transition from the single-band FCI to a K-CDW under weak band mixing. They also propose an iterative basis-optimization method (ED iteration) intended to maximize the lowest-band occupation of the ground-state manifold, and they report that even after this optimization the FCI gap remains absent once higher bands are partially occupied. The broader conclusion is that current continuum models fail to explain the experimentally observed FCIs in this system.","tokens_in":73758,"tokens_out":4763,"duration_ms":50381,"significance":"If substantiated, the central claim would be important: it would show that single-HF-band ED predictions for this material are not reliable and that weak band mixing can qualitatively change the ground state from an FCI to a CDW. The paper has several strengths: it uses literature PES counting rules rather than fitting parameters to identify phases; it reports the fully untruncated three-band {6,6} result for the 9x2 system as a benchmark; and it is unusually candid about limitations, including non-convergence of the ED iteration and cases where the lowest-band occupation decreases after iteration. The detailed appendix with an analytical toy model for iteration failure is also a useful contribution. However, the central FCI-to-CDW claim currently rests on heavily truncated Hilbert spaces at small system sizes, and the paper's own exact limit does not reproduce the CDW. The significance of the paper is therefore conditional on demonstrating that the small-occupation-cap regime represents the physical weak-band-mixing limit.","major_comments":[{"comment":"The central claim that the 1/3 CN ground state becomes a K-CDW under small band mixing is not supported by the accessible exact limit. For the 9x2 system, the paper states that the full three-band calculation at {6,6} gives a PES that is not gapped at either FCI or CDW counting and, concomitantly, no gap in the energy spectrum. The CDW identification is made at {1,0} and {2,0}, with no finite-truncation extrapolation connecting these small caps to the {6,6} limit. Since the sequence {1,0} -> {2,0} -> {6,6} loses the CDW gap rather than converging toward it, the conclusion that a small fraction of higher-band occupation drives an FCI-to-CDW transition is not established for the untruncated problem.","section":"Sec. III A, Fig. 3"},{"comment":"Because the truncation of band occupations breaks one-body basis invariance, the CDW found at small caps could be an artifact of the particular HF basis used rather than a property of the Hamiltonian. The ED iteration method was introduced to address this issue, but the paper's own results show that it raises n0,tot only modestly (about 6%) and that the overlap between truncated and exact ground states slightly decreases during the iteration (Fig. 9 insets and App. E2). This does not rescue the small-cap CDW as the exact ground state. The authors should either provide evidence that the CDW is stable under basis changes at fixed truncation, or explicitly reframe the claim as a statement about the truncated HF-basis model.","section":"Sec. II C and Sec. V, Fig. 9 and App. E2"},{"comment":"The PES is always computed from the three lowest states at the FCI momenta, even in cases where these are not the absolute ground states of the system. For the AVE scheme at 1/3 filling, the text acknowledges that the states used for the PES are excited states, and then uses the absence of an FCI gap in this excited-state PES to conclude that no FCI appears with band mixing. This is a weaker diagnostic than a ground-state phase assignment. The authors should separate ground-state statements from statements about states at fixed FCI momenta, particularly in the AVE 1/3 discussion, or justify why the excited-state PES is informative for the phase diagram.","section":"Sec. III A, Sec. III C, and Eq. (20)"}],"minor_comments":[{"comment":"The phrase 'a small fraction of electrons is allowed to occupy the higher bands' refers to an occupation cap, not to an actual physical fraction. The actual occupation in the ground state at {1,0} is n1,tot = 0.24 out of 6 particles (~4%). Clarifying the distinction between the cap and the realized occupation would prevent a misreading.","section":"Abstract and Sec. III A"},{"comment":"The convergence criterion R < 1e-3 is presented as a practical stopping rule, but the paper does not discuss how sensitive the conclusions are to this threshold. A sentence on the dependence of the final basis and observables on the chosen tolerance would be useful.","section":"Sec. IV, Eq. (37)"},{"comment":"The structure factor plots in Fig. 4 would benefit from explicitly marking the K and K' points in the hexagon boundary, since the text refers to peaks at K,K' and the reader must identify these locations by eye.","section":"Sec. III A, Fig. 4 caption"},{"comment":"Several appendix figures use 'before' and 'after' symbols without a legend in the figure panels; the captions define them, but adding a small legend or consistent marker definitions in the figures would improve readability.","section":"App. E 3, Figs. 22-34"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of the journal and contains a useful methodological contribution in the ED iteration method, together with an honest and valuable exact {6,6} benchmark. The main technical weakness is truncation dependence of the central CDW claim; this is fixable by either providing a systematic finite-truncation extrapolation or substantially reframing the claim as a statement about the truncated model. I do not see a novelty or citation-pattern problem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Things you should know before reading: this is a serious multi-band ED study, but the paper's headline claim—that the 1/3 FCI in the CN scheme becomes a K-CDW under minimal band mixing—is undercut by the paper's own untruncated calculation. At {6,6} (no occupation cap) the PES is gapped at neither FCI nor CDW counting and the energy spectrum has no gap. So the CDW is observed only in truncated Hilbert spaces at small caps; the transition is not established for the full model.\n\nWhat is genuinely new and worth credit: the PES as a diagnostic in this setting, showing that the threefold degeneracy at FCI momenta is not sufficient and that the small-cap ground states have CDW counting; and the ED iteration method, a reasonable attempt to mitigate the basis dependence of truncated ED. They test it carefully, including non-converging cases and cases where it decreases the lowest-band occupation. That is honest, reproducible methodology, even if it doesn't rescue the CDW.\n\nThe soft spots are in proportion. The central claim rests on truncation caps {1,0} and {2,0}; no finite-truncation extrapolation is given, and the sequence to {6,6} loses the gap. Because truncation breaks basis invariance, the CDW could be an artifact of the HF basis rather than of the Hamiltonian. The paper's conclusion that 'all current models fail to explain the experimental emergence' is too strong given it tests one twist angle, one stacking, and two interaction schemes; the choice V=22 meV in the AVE scheme is not justified in the text, unlike V=28 meV which is tied to the experimental displacement field. None of these are fatal to the work, but they mean the abstract overstates what is shown: the robust statement is that the single-HF-band FCI is destabilized by band mixing, and the nature of the phase in the exact three-band limit is not identified.\n\nWho this is for: people working on FCI stability in rhombohedral graphene and on truncation schemes for multi-band ED. It deserves a serious referee and a major revision; the authors should narrow the abstract and either soften the CDW claim or provide a truncation-extrapolation argument. I would engage with it as a referee.","headline":"Careful multi-band ED study with an honest new method, but the headline FCI-to-CDW transition is undercut by the paper's own untruncated calculation showing neither phase.","tokens_in":74298,"tokens_out":3284,"would_cite":true,"duration_ms":32234,"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 1/3 fractional Chern insulator predicted in pentalayer graphene becomes a charge density wave once higher bands are partially occupied.","keywords":["fractional Chern insulator","rhombohedral pentalayer graphene","multi-band exact diagonalization","particle entanglement spectrum","charge density wave","band mixing","Hartree-Fock basis","ED iteration method"],"falsifier":"Extend the PES analysis to the untruncated three-band Hilbert space at larger system sizes (15×1 and 21×1 at 1/3 filling in the CN scheme, where the FCI counting is 75 and 168 states respectively) and examine whether an FCI-counting gap reappears as caps are raised; a reappearing FCI gap, or a disappearing CDW-counting gap, would show the small-cap transition is a truncation artifact.","tokens_in":73166,"feed_emoji":"🧲","tokens_out":6860,"duration_ms":66670,"temperature":0.7,"pith_summary":"Fractional Chern insulators in moiré pentalayer graphene are usually predicted by projecting interactions onto a single Hartree-Fock band. This paper argues that such predictions break down as soon as electrons are allowed into higher bands: at 1/3 filling in the charge-neutrality scheme, even minimal band mixing converts the apparent FCI into a K-point charge density wave, diagnosed by the particle entanglement spectrum. At 2/3 filling in the average scheme the FCI is more resistant but is still destroyed by stronger mixing. The paper also introduces an iterative basis-optimization method that concentrates the ground state in the lowest band, yet after convergence the FCI gap remains absent. The broader conclusion is that current three-band continuum models cannot account for the experimentally observed FCIs in this material.","feed_headline":"Band mixing flips a predicted fractional Chern insulator into a CDW","feed_subtitle":"Multi-band exact diagonalization shows the 1/3 FCI in the CN scheme dies at minimal band mixing; the 2/3 AVE FCI is sturdier.","key_machinery":"The load-bearing machinery is a three-band exact diagonalization in the Hartree-Fock basis, with the Hilbert space truncated by caps $\\{N_{\\mathrm{band1}}, N_{\\mathrm{band2}}\\}$ on how many particles may sit in the two higher bands. The diagnostic is the particle entanglement spectrum (PES): for a Laughlin-like FCI at 1/3 on 18 momentum points with $N_A=2$, there are 117 entanglement levels below the gap, while a K-CDW gives 45; reading which gap is present tells the phase apart even when energy spectra look similar. The paper's second tool is an ED iteration method that diagonalizes the momentum-resolved one-body density matrix of the ground-state manifold at each step and uses its eigenvectors as a new single-particle basis, thereby maximizing the weight of the ground states on band 0; convergence is declared when a relative change $R$ falls below $10^{-3}$. This method is exact for product states and is intended to reduce truncation error, though convergence is not guaranteed in general.","core_discovery":"The central discovery is that the FCI at $\\nu=1/3$ in the CN scheme, which looks clean in a single HF band with an entanglement gap at the FCI counting (117 levels below the gap on a $9\\times2$ system with $N_A=2$), is not robust. At truncation $\\{1,0\\}$—allowing just one electron in the first remote HF band—the PES gap at FCI counting closes and a new gap opens with 45 levels, matching a K-CDW; the structure factor develops peaks at $K$ and $K'$. The 2/3 FCI in the AVE scheme behaves differently: its PES, computed after projection onto the lowest HF band and particle-hole transformation, keeps the FCI counting gap under moderate band mixing, but the gap collapses with stronger mixing. In the fully untruncated three-band limit $\\{6,6\\}$, the 1/3 PES is gapped at neither FCI nor CDW counting. Thus the paper concludes that single-band ED is biased toward FCIs and that all current continuum models fail to explain the experimental emergence of these phases.","pith_inferences":["Editorial inference: the same small-cap PES test could be applied to other nearly gapless moiré systems, such as twisted MoTe2, where band mixing is also suspected to destabilize FCIs.","Editorial inference: the ED iteration method amounts to a variational basis optimizer; a natural extension is to use it as a convergence check for any truncated multi-band ED calculation, not only FCI searches.","Editorial inference: because the paper finds neither FCI nor CDW in the fully untruncated 1/3 CN limit, the true ground state of the three-band model may be a different state, and identifying it would clarify whether the experimental FCI needs bands beyond the lowest three."],"forward_implications":["Single-band HF-projected ED spectra for rhombohedral pentalayer graphene cannot be taken as evidence of FCIs; multi-band checks with PES are required.","If the CN-scheme 1/3 state is really a CDW, experimental observation of an FCI at that filling implies the CN scheme, the displacement field, or the screening parameters miss crucial physics.","The AVE scheme at 2/3 is the most favorable case among those studied, so future model refinements can be tested against that phase first.","Even after optimizing the single-particle basis, no FCI gap reappears, suggesting the failure is intrinsic to the model rather than a fixable basis artifact."],"supporting_citations":[{"why":"Supplies the previous multi-band ED setup showing FCIs collapse under band mixing; this paper extends it with PES and the iteration method.","marker":"[40]"},{"why":"Provides the particle entanglement spectrum counting formula used to identify FCI phases.","marker":"[3]"},{"why":"Provides the thin-torus CDW counting used to recognize the 45-level entanglement gap.","marker":"[69]"},{"why":"Introduces the entanglement spectrum method generalized by the PES diagnostic.","marker":"[59]"},{"why":"Supplies the Hartree-Fock phase diagram and HF basis that the multi-band ED is built on.","marker":"[39]"},{"why":"One of the single-band ED studies whose FCI prediction is shown to be unstable under band mixing.","marker":"[36]"},{"why":"Another single-band-based approach predicting anomalous Hall crystals and FCIs that motivates the re-examination.","marker":"[41]"},{"why":"Provides the continuum model and parameters for rhombohedral multilayer graphene used in the simulations.","marker":"[37]"}],"fun_headline_variants":["Band mixing flips 1/3 FCI into charge density wave","Fractional Chern insulator not robust to band mixing","Single-band ED bias revealed: FCI becomes CDW","Minimal band mixing destabilizes predicted 1/3 FCI"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The FCI-to-CDW conclusion depends on the assumption that small occupation caps such as $\\{1,0\\}$ and $\\{2,0\\}$ capture the physical weak-band-mixing regime; at the full $\\{6,6\\}$ limit the PES shows no gap at either counting, so a systematic finite-truncation extrapolation is missing.","fun_headline_variants_meta":{"raw":{"variants":["Band mixing flips 1/3 FCI into charge density wave","Fractional Chern insulator not robust to band mixing","Single-band ED bias revealed: FCI becomes CDW","Minimal band mixing destabilizes predicted 1/3 FCI"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00027,"raw_usage":{"total_tokens":1685,"prompt_tokens":1062,"completion_tokens":623,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":678,"completion_tokens_details":{"reasoning_tokens":553}},"tokens_in":678,"tokens_out":623,"duration_ms":5907,"temperature":1.0,"reasoning_tokens":553,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T05:36:28.387444+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Extend the PES analysis to the untruncated three-band Hilbert space at larger system sizes (15×1 and 21×1 at 1/3 filling in the CN scheme, where the FCI counting is 75 and 168 states respectively) and examine whether an FCI-counting gap reappears as caps are raised; a reappearing FCI gap, or a disappearing CDW-counting gap, would show the small-cap transition is a truncation artifact.","supporting_citations":[],"review_version":1}