{"id":"28d810a5-1496-4e52-b8e9-8dd99db96d7a","arxiv_id":"2608.05927","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Enlarging the variational cluster in mean-field theory reveals additional incompressible phases of hard-core bosons on the square-kagome lattice, including a robust 3/4-density plateau corresponding to half-magnetization.","lead":"Using a cluster-based mean-field method, this paper maps out incompressible density phases of hard-core bosons on the square-kagome lattice and finds extra magnetization plateaus beyond the standard six-site unit cell picture. The work links these plateaus to two real copper-based quantum magnets and predicts a new half-magnetization plateau that future high-field experiments could test.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed intrinsic phases at 13/24, 17/24, and 19/24 rest only on 24-site cluster mean-field; the two 24-site clusters disagree, and densities match 24-site quantization, so finite-cluster artifacts are the more parsimonious explanation.","rationale":"The reader's weakest assumption identified the same load-bearing weakness: the cluster mean-field ansatz is not converged, and the new phases are only seen in 24-site clusters with setup-dependent differences. My concern sharpens this by noting the exact quantization of the new plateau densities (denominator 24 equals cluster size), which makes finite-size artifacts a direct possibility, and by contrasting the lack of independent confirmation for 13/24, 17/24, and 19/24 with the independent ED support for 3/4. The proposed test—larger clusters or DMRG/ED on the full model—would settle whether the phases are intrinsic. This does not change the reader's CONDITIONAL verdict; the paper remains publishable as a conditional result pending such checks.","tokens_in":10238,"tokens_out":6778,"duration_ms":63267,"concrete_test":"Run HMFT with a 36-site or 48-site cluster (or a non-commensurate cluster) and check whether plateaus at 13/24, 17/24, and 19/24 persist or shift to n/36 or n/48 values. Alternatively, perform density-matrix renormalization group on cylinders of width 4–6 and length up to 48 for the same Hamiltonian, computing density versus chemical potential; if no incompressible interval appears at these densities in the thermodynamic limit, the claim that these phases are intrinsic is falsified. A positive result—plateaus at the same physical densities with converged cluster size—would support the claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the square-kagome lattice intrinsically hosts incompressible phases at ρ=13/24, 17/24, and 19/24 is not established by the presented numerics. Evidence consists solely of cluster mean-field calculations with a 24-site variational cluster, and the two 24-site clusters give different plateau sets: Setup 2 adds a plateau at 17/24 that is absent in Setup 1 and reduces the width of the 19/24 plateau (Sec. V, Fig. 4a). In HMFT, an incompressible phase on a 24-site cluster corresponds to an integer number of particles per cluster, so densities of 13/24, 17/24, and 19/24 are precisely the values one would obtain from a 24-site discretization even if the thermodynamic limit has no such gap. No larger cluster (36 or 48 sites) and no exact method for the full model is applied to these densities. The paper's Sec. III B concedes that only structures compatible with the chosen cluster can be represented. The 3/4 phase is independently supported by earlier exact-diagonalization signatures and appears in all clusters, but the new phases have no such support. The most parsimonious reading is that these plateaus are artifacts of the finite variational cluster, not intrinsic features of the lattice.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies hard-core bosons on the square-kagome lattice with nearest-neighbor hopping and repulsion, using a hierarchical cluster mean-field method (HMFT) on 12-site and two distinct 24-site clusters. It reports that the conventional six-site unit-cell description captures only the known 2/3 and 5/6 incompressible phases, while enlarged clusters reveal additional plateaus at 3/4, 13/24, 17/24, and 19/24, which map via the Matsubara-Matsuda transformation to magnetization plateaus of a spin-1/2 XXZ model. The paper also applies the method to two square-kagome compounds with first-principles exchange parameters, reproducing observed 1/3 and 2/3 plateaus in KCu6AlBiO4(SO4)5Cl and predicting a 1/4 plateau near 75 T, while Na6Cu7BiO4(PO4)4Cl3 is argued to show plateaus at 1/3 and 5/6.","tokens_in":10418,"tokens_out":3964,"duration_ms":42550,"significance":"If the central claim is correct, the paper would show that the square-kagome lattice supports a substantially richer hierarchy of incompressible crystals than previously recognized, and it would provide concrete, falsifiable magnetization-plateau predictions for two candidate materials. The strongest point in favor of the paper is the rho=3/4 phase: it appears in both 12-site and 24-site calculations and is independently supported by earlier exact-diagonalization signatures on 24-, 36-, and 48-site clusters, even though those signatures were not interpreted as a robust plateau. The material predictions are also genuinely falsifiable, and the phase diagrams involve no free parameters fitted to target plateaus. These strengths are, however, concentrated on the 3/4 phase; the claimed intrinsic character of the 13/24, 17/24, and 19/24 phases rests on much weaker evidence, as detailed in the major comments.","major_comments":[{"comment":"The central claim that the 13/24, 17/24, and 19/24 phases are intrinsic to the infinite lattice is not established because the two 24-site clusters give different plateau sets: Setup 2 adds a 17/24 plateau that is absent in Setup 1 and reduces the width of the 19/24 plateau. Since these densities correspond to integer particle numbers on a 24-site cluster, they are exactly the values one would expect from finite-cluster quantization even if the thermodynamic limit has no such gap. The paper does not provide a 36- or 48-site calculation, no exact-diagonalization data at these densities, and no estimate of the charge gap as a function of cluster size. I request either larger-cluster HMFT results, an independent method (e.g., exact diagonalization or tensor-network simulation) for at least the 13/24 and 19/24 densities, or an explicit quantitative argument that the 24-site discretization error cannot account for the plateaus.","section":"Sec. V, Figs. 3(b) and 4(a)"},{"comment":"Section III B states that only spatial structures compatible with the chosen cluster geometry can be represented. The clusters were deliberately chosen to accommodate the candidate localized states, including the approximate 20-site state with finite leakage shown in Fig. 1(c). Finding that state in the enlarged variational space is therefore partly a consequence of the variational ansatz, not an unbiased search. To support the intrinsic character of the new phases, the paper should demonstrate that these states remain lowest in energy when represented in a cluster-independent basis, for example by comparing against a larger cluster that does not have the 20-site construction built in, or by quantifying the leakage amplitude and showing that it decays with cluster size.","section":"Sec. III B"},{"comment":"The rho=3/4 phase is the best-supported new phase, but its status as a robust incompressible phase would be strengthened by a direct measurement of the charge gap and its scaling with cluster size. The paper cites exact-diagonalization signatures on 24-, 36-, and 48-site clusters, but those earlier works attributed the feature to finite-size effects, and the present manuscript does not provide a convergence analysis of the gap or the local density pattern across cluster sizes. Without such quantitative support, the claim that rho=3/4 is a true thermodynamic plateau remains plausible but not fully demonstrated.","section":"Sec. V, rho=3/4 paragraph"},{"comment":"The predicted 1/4 magnetization plateau in KCu6AlBiO4(SO4)5Cl at roughly 75 T is computed with Setup 2 only. Given that Setup 1 and Setup 2 already disagree on the plateau structure of the ideal model at nearby densities (e.g., 17/24 and 19/24), the material prediction should be checked with at least one alternative cluster geometry, and preferably with a small perturbation of the exchange parameters, before it is presented as a concrete experimental target.","section":"Sec. VI A, Fig. 5(a)"}],"minor_comments":[{"comment":"There is a typo in the introduction: 'Heisenebrg' should be 'Heisenberg'.","section":"Sec. I"},{"comment":"The chemical formula of Na6Cu7BiO4(PO4)4Cl3 is written inconsistently: the text uses Na6Cu7BiO4(PO4)4Cl3, while the figure caption and one occurrence in the text use 'Na6Cu7BiO4(PO)4Cl3'. Please unify the notation.","section":"Sec. VI B and Fig. 5"},{"comment":"The caption states that the 20-site localized state has 'only weak hopping processes extending beyond the cluster boundary', but no numerical measure of the leakage is given. Defining a leakage parameter and reporting its value would make the approximate-localization claim quantitative and testable.","section":"Fig. 4(b)"},{"comment":"The lower-density region of the Setup 2 phase diagram is not shown and is described only as 'qualitatively identical' to Setup 1. Since the discrepancy between Setups 1 and 2 is a central issue, the full phase diagram or a quantitative comparison of plateau boundaries should be provided.","section":"Sec. V, Fig. 4(a)"},{"comment":"The convergence criterion is stated as a change of all mean fields below 10^-6, but there is no mention of the number of self-consistent iterations, the dependence on initial conditions, or the possible presence of multiple local minima. A brief discussion of these numerical aspects would improve reproducibility.","section":"Sec. III A"}],"recommendation":"major_revision","confidential_remarks":"The manuscript addresses a timely topic and contains a useful, self-contained derivation of localized-magnon states and a falsifiable material prediction. My main concern is that the paper's most novel quantitative claim—that the 13/24, 17/24, and 19/24 phases are intrinsic to the lattice—rests on a single cluster size with two mutually inconsistent cluster geometries. The authors should be encouraged to add larger-cluster or exact checks before publication, and the material prediction should be flagged as provisional if those checks are not feasible. The manuscript otherwise fits the scope of cond-mat.str-el."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth reading for one result: the 3/4 plateau on the square-kagome lattice looks real. It appears in both the 12-site and 24-site clusters, and it matches earlier exact-diagonalization signatures in Refs. [10] and [23]. The reinterpretation of that finite-size feature as a genuine incompressible phase is plausible and is the paper's main contribution.\n\nWhat's genuinely new is the HMFT treatment with clusters beyond the six-site unit cell, which turns up plateaus at 13/24, 17/24, and 19/24. These have not been reported before. The paper also makes a concrete, falsifiable prediction: a 1/4 magnetization plateau in KCu6AlBiO4(SO4)5Cl at around 75 T. That is a nice touch and gives experimentalists something to shoot for.\n\nThe soft spot is exactly where the stress-test note points. These new densities rest entirely on the 24-site cluster mean-field. There is no 36-site or 48-site check, no exact diagonalization or tensor-network cross-check for those fillings. And the two 24-site setups do not fully agree: Setup 2 adds 17/24 that Setup 1 lacks, and shrinks the 19/24 plateau. In HMFT, an incompressible phase on a 24-site cluster corresponds to an integer number of particles per cluster, so densities like 13/24 and 19/24 are exactly what you would expect from the discretization even if no true gap survives in the thermodynamic limit. The paper is honest about the cluster-compatibility limitation (Sec. III B), but that honesty does not solve the convergence problem. So the new phases are intriguing, not established.\n\nThe 3/4 phase, in contrast, appears in clusters of two different sizes and geometries, and it has prior ED support. That part of the paper holds up. The material comparison is qualitative – the high-field experimental data are sparse – but the reproduction of the 1/3 and 2/3 plateaus is encouraging.\n\nWho is this for? People working on square-kagome magnets, flat-band localized magnons, and cluster mean-field methodology. It is a clear, well-organized paper with a solid core and a speculative edge. A serious referee should see it, but should insist on either a larger cluster or an exact method applied to the 13/24, 17/24, and 19/24 fillings before those are sold as intrinsic phases. With that revision, the paper becomes a stronger contribution.\n\nRecommendation: send it to peer review. It is not a desk reject by any means.\n\nBest,\n[Your name]","headline":"A useful HMFT study that makes a solid case for the 3/4 plateau but leaves the new 24-site plateaus unproven; deserves review with a demand for converged numerics.","tokens_in":11037,"tokens_out":3216,"would_cite":true,"duration_ms":29285,"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":"The paper argues that the square-kagome lattice hosts a hierarchy of incompressible phases beyond the conventional six-site-unit-cell description, with a stable $\\rho=3/4$ phase and additional phases at $\\rho=13/24$, $17/24$, and $19/24$…","keywords":["square-kagome lattice","hard-core bosons","compact localized states","magnetization plateaus","hierarchical cluster mean-field theory","XXZ model","frustrated magnetism","incompressible phases"],"falsifier":"A concrete test is to compute the same hard-core boson or XXZ model on the square-kagome lattice with an unbiased method on larger systems, such as a tensor-network or exact-diagonalization calculation on 36- or 48-site clusters: if the $13/24$, $17/24$, $19/24$, or $3/4$ plateaus disappear, the central claim fails. For the material prediction, a magnetization measurement of KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl up to roughly 200 T that shows no plateau at $m/m_s=1/4$ near 75 T would falsify the predicted additional plateau.","tokens_in":9906,"feed_emoji":"🧲","tokens_out":8613,"duration_ms":72957,"temperature":0.7,"pith_summary":"This paper argues that the square-kagome lattice supports more incompressible phases than the conventional six-site-unit-cell description reveals. Using a hierarchical cluster mean-field approximation, the author reproduces the known localized-magnon crystal phases at densities $2/3$ and $5/6$ and finds additional incompressible states at $13/24$, $3/4$, $17/24$, and $19/24$. The $3/4$ phase persists in both 12-site and 24-site clusters and is identified as a robust half-magnetization plateau in the spin-$1/2$ XXZ model. The enlarged-cluster results are then applied to two square-kagome compounds, reproducing measured plateaus and predicting new ones, notably a $1/4$ plateau in KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl.","feed_headline":"Hidden plateaus appear once square-kagome clusters grow","feed_subtitle":"Larger cluster calculations uncover a stable 3/4 plateau and predict a 1/4 plateau in a real compound.","key_machinery":"The argument rests on hierarchical cluster mean-field theory (HMFT), in which the lattice is partitioned into identical clusters, intra-cluster correlations are treated exactly by numerical diagonalization, and inter-cluster couplings are decoupled through self-consistent mean fields. Three cluster geometries are used: a 12-site cluster and two 24-site clusters (Setup 1 and Setup 2) chosen to accommodate the compact localized states and the approximate 20-site localized configuration. The Matsubara-Matsuda mapping converts hard-core boson densities into spin magnetizations, so each incompressible density plateau becomes a magnetization plateau. The key object is the approximate 20-site localized state, a nearly destructive-interference eigenstate with finite leakage that is not an exact compact localized state but nonetheless drives the $17/24$, $3/4$, and $19/24$ phases.","core_discovery":"The central claim is that the conventional six-site unit-cell description of the square-kagome lattice is incomplete: it captures only the compact-localized-state phases at $\\rho=2/3$ and $5/6$, whereas a hierarchical cluster mean-field treatment with larger clusters stabilizes a richer set of incompressible crystals. The $\\rho=3/4$ phase, corresponding to the $m/m_s=1/2$ magnetization plateau, appears already at the 12-site level and survives both 24-site cluster geometries, making it a robust property of the model rather than a finite-size artifact. The phases at $\\rho=13/24$, $17/24$, and $19/24$ appear only when the variational cluster is enlarged, and they are associated with an approximately localized 20-site configuration rather than an exact compact localized state. Through the Matsubara-Matsuda mapping these phases correspond to magnetization plateaus at $m/m_s=1/12$, $5/12$, and $7/12$ in the spin-$1/2$ XXZ model.","pith_inferences":["One can test the $3/4$ plateau's stability directly by running the same HMFT hierarchy on 36- and 48-site clusters; if it persists while the neighboring $13/24$ and $19/24$ phases shift or vanish, the hierarchy suggests a critical cluster-size scale for each phase.","The same beyond-unit-cell logic could be applied to other flat-band lattices with multiple compact-localized-state sizes, where enlarged clusters may reveal similar hierarchies of approximate localized crystals.","If the 20-site approximate localized state is the true microscopic origin, the plateau widths should scale with $V/t$ in a predictable way, so varying the XXZ anisotropy along the $t'$ axis provides a tunable test of the mechanism."],"forward_implications":["The six-site unit-cell picture is incomplete; enlarged-cluster treatments are needed to enumerate the incompressible phases of the square-kagome lattice.","The $\\rho=3/4$ phase corresponds to a stable $m/m_s=1/2$ magnetization plateau that should appear in the spin-$1/2$ square-kagome Heisenberg and XXZ models.","The ideal square-kagome model should also show plateaus at $m/m_s=1/12$, $5/12$, and $7/12$, whose stability depends on the localized configurations accessible at larger clusters.","KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl should show a magnetization plateau at $m/m_s=1/4$ around 75 T, between the existing low-field data and the observed $1/3$ plateau.","Strong diagonal exchange in Na$_6$Cu$_7$BiO$_4$(PO$_4$)$_4$Cl$_3$ suppresses the low-field plateau and shifts the highest plateau to $m/m_s=5/6$, qualitatively different from the ideal lattice."],"supporting_citations":[{"why":"It supplies the previously established magnetization-jump plateaus on the square-kagome lattice that the 12-site calculation reproduces.","marker":"[10]"},{"why":"It is the earlier square-kagome study that identified the localized-magnon plateau phases used as the reference for the present phase diagram.","marker":"[24]"},{"why":"It established the $2/3$ and $5/6$ compact-localized-state phases of the square-kagome antiferromagnet that the enlarged clusters build upon.","marker":"[27]"},{"why":"It introduced the enlarged-cluster approximate-localized-state construction that motivates the new phases beyond exact compact localized states.","marker":"[29]"},{"why":"It provides the Matsubara-Matsuda mapping that connects hard-core boson densities to spin magnetizations and hence to plateaus.","marker":"[30]"},{"why":"It is the foundational hierarchical mean-field theory method that the paper's cluster calculations extend.","marker":"[34]"},{"why":"It applied hierarchical cluster mean-field theory to hard-core bosons on the kagome lattice, giving the staircase-of-crystal-phases framework used here.","marker":"[39]"},{"why":"It supplies the experimental magnetization data and the exchange parameters for KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl.","marker":"[31]"},{"why":"It supplies the exchange parameters for Na$_6$Cu$_7$BiO$_4$(PO$_4$)$_4$Cl$_3$ used to compute its magnetization process.","marker":"[40]"}],"fun_headline_variants":["Square-kagome clusters grow, hidden plateaus emerge","Beyond unit cells: new incompressible phases in square-kagome","Larger clusters expose missing square-kagome plateaus","Half-magnetization plateau robust in square-kagome model","Square-kagome: bigger clusters unlock hidden crystals"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the product-state cluster mean-field ansatz on the 12- and 24-site clusters represents the true infinite-lattice ground state of these localized crystal phases, so the new plateaus are physical rather than artifacts of the chosen cluster size and geometry.","fun_headline_variants_meta":{"raw":{"variants":["Square-kagome clusters grow, hidden plateaus emerge","Beyond unit cells: new incompressible phases in square-kagome","Larger clusters expose missing square-kagome plateaus","Half-magnetization plateau robust in square-kagome model","Square-kagome: bigger clusters unlock hidden crystals"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000218,"raw_usage":{"total_tokens":1442,"prompt_tokens":948,"completion_tokens":494,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":410}},"tokens_in":564,"tokens_out":494,"duration_ms":158513,"temperature":1.0,"reasoning_tokens":410,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T20:55:27.173868+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A concrete test is to compute the same hard-core boson or XXZ model on the square-kagome lattice with an unbiased method on larger systems, such as a tensor-network or exact-diagonalization calculation on 36- or 48-site clusters: if the $13/24$, $17/24$, $19/24$, or $3/4$ plateaus disappear, the central claim fails. For the material prediction, a magnetization measurement of KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl up to roughly 200 T that shows no plateau at $m/m_s=1/4$ near 75 T would falsify the predicted additional plateau.","supporting_citations":[{"cited_title":"Niggemann, N","cited_arxiv_id":null,"evidence_quote":"It supplies the exchange parameters for Na$_6$Cu$_7$BiO$_4$(PO$_4$)$_4$Cl$_3$ used to compute its magnetization process."},{"cited_title":"Nakano, Y","cited_arxiv_id":null,"evidence_quote":"It supplies the previously established magnetization-jump plateaus on the square-kagome lattice that the 12-site calculation reproduces."},{"cited_title":"Richter and J","cited_arxiv_id":null,"evidence_quote":"It is the earlier square-kagome study that identified the localized-magnon plateau phases used as the reference for the present phase diagram."},{"cited_title":"Matter Phys.12, 507 (2009)","cited_arxiv_id":null,"evidence_quote":"It established the $2/3$ and $5/6$ compact-localized-state phases of the square-kagome antiferromagnet that the enlarged clusters build upon."},{"cited_title":"Ortiz and C","cited_arxiv_id":null,"evidence_quote":"It is the foundational hierarchical mean-field theory method that the paper's cluster calculations extend."},{"cited_title":"Huerga, S","cited_arxiv_id":null,"evidence_quote":"It applied hierarchical cluster mean-field theory to hard-core bosons on the kagome lattice, giving the staircase-of-crystal-phases framework used here."},{"cited_title":"Fujihala, K","cited_arxiv_id":null,"evidence_quote":"It supplies the experimental magnetization data and the exchange parameters for KCu$_6$AlBiO$_4$(SO$_4$)$_5$Cl."}],"review_version":1}