{"id":"3588e488-5357-41c2-b4a3-968dd106abce","arxiv_id":"2506.15769","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Quantum dimer models on the Aztec diamond and square-octagon fortress have RK ground states with boundary-shape-dependent phase separation, and a new Kasteleyn method computes vison correlators exactly.","lead":"The authors construct quantum spin models whose ground states, at a fine-tuned 'Rokhsar-Kivelson' point, split into distinct phases that depend on the shape of the sample boundary. This mirrors the classical 'Arctic circle' effect in random dimer tilings, now realized as exact quantum phase separation in spin and dimer models.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The square-octagon fortress example, the only one claimed to be stable away from the RK point, rests on an unverified identification with a weighted Aztec diamond; if that correspondence is inexact, the three-region phase separation and the robust breakdown claim are unsupported.","rationale":"The reader's weakest assumption identifies the same load-bearing point: the fortress result, which is the paper's main candidate for a robust breakdown of the thermodynamic limit, relies on an equivalence with a weighted Aztec diamond that is quoted rather than derived. This is the correct focus because the Aztec diamond example, while rigorous, is confined to the fine-tuned RK point and is explicitly acknowledged as possibly fragile; the square-octagon fortress is what would make the central claim significant beyond a single critical point. The paper does provide real independent support elsewhere: the Kasteleyn-based vison method is a concrete algorithmic contribution, and the rectangular-boundary square-octagon results include an analytic Szego-limit derivation. Those parts do not depend on the questionable equivalence. The conditional verdict is therefore appropriate: the exact Aztec and rectangular-boundary statements stand, but the fortress phase diagram and robustness claims should be tempered or independently verified. A small-size exact-enumeration check would settle the most direct question, namely whether the partition-function correspondence is actually exact for finite fortresses, and would thereby test whether Eq. (45) can legitimately be imported.","tokens_in":39736,"tokens_out":17360,"duration_ms":197758,"concrete_test":"For radii R=1 through 8, exactly enumerate all dimer coverings of the square-octagon fortress and of the two-periodic Aztec diamond with face weights a=b=c=1, d=1/2, and verify the Appendix D partition-function identity (Eqs. D1-D3) including all stated finite-size prefactors. If the identity fails for any R, the imported octic curve Eq. (45) and the three-region phase diagram for the fortress are not established; if it holds for all small R, the equivalence is at least supported at the partition-function level.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is strongest in the square-octagon fortress example, because it is the only example asserted to survive small perturbations away from the RK point. That example depends on the equivalence between the unweighted fortress and a two-periodic weighted Aztec diamond with face weights a=b=c=1 and d=1/2, stated in Appendix D, and on the octic curve Eq. (45) imported from Refs. [7,24]. Appendix D presents the correspondence and the weight assignment, but it does not prove the equivalence or demonstrate that local dimer statistics of the fortress, including the dimer-dimer and vison correlators used in Section V, are governed by the weighted Aztec model. The critical region's power-law behavior is explicitly not resolved by the numerics: Section V C says the numerical plots are 'less conclusive.' If the octic curve or the fortress--Aztec correspondence is inexact, the claimed frozen/critical/gaseous three-region structure, and the 'likely stable to perturbations' conclusion built on it, are unsupported. The Aztec diamond example alone is rigorous, but it is fine-tuned to the RK point and the paper itself concedes it may not survive perturbations.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper constructs local spin-1/2 Hamiltonians on square and square-octagon lattices whose low-energy sectors are Rokhsar-Kivelson quantum dimer models. At the RK point the ground state is the uniform superposition of classical dimer coverings, so classical dimer results transfer. On square-shaped domains the standard critical phase is recovered; on Aztec-diamond domains the Arctic circle produces macroscopic frozen and critical regions, giving a shape-dependent thermodynamic limit. For the square-octagon lattice with rectangular boundaries the authors compute dimer-dimer correlators analytically in the infinite-size limit, obtain exponential decay, and prove via the strong Szegő limit theorem that the vison correlator along an octagon path tends to the nonzero constant sqrt(60)/10, identifying an ordered short-range entangled phase. For the square-octagon fortress they use a correspondence with a weighted Aztec diamond to claim three regions (frozen, critical, gaseous), with the central region behaving like the rectangular case. They also introduce an exact fraction-free LU method for vison correlators.","tokens_in":39894,"tokens_out":7134,"duration_ms":79984,"significance":"The exact square-octagon rectangular results are solid and the vison correlator method is a genuine technical contribution with clear advantages over Monte Carlo. The central conceptual claim—shape-dependent thermodynamic limit at the RK point—is rigorously established for the Aztec-diamond example and is plausible for the fortress conditional on the cited classical equivalence. The paper is careful with Kasteleyn asymptotics, checks the conditions of the strong Szegő theorem, and uses exact integer arithmetic for ill-conditioned fortress matrices, which makes the numerical results reproducible. The main weaknesses are that the fortress three-region structure rests on a not-fully-stated equivalence with a weighted Aztec diamond and that the claimed critical region is explicitly numerically inconclusive; the robustness-away-from-RK claim is also presented as a demonstration despite being explicitly left as future work.","major_comments":[{"comment":"The three-region structure of the fortress ground state rests on the asserted equivalence between the unweighted square-octagon fortress and a two-periodic weighted Aztec diamond with face weights a=b=c=1 and d=1/2. Appendix D states the weight assignment and quotes partition-function identities, but it does not state the precise equivalence theorem or establish that the local dimer statistics used in Section V (dimer probabilities, dimer-dimer correlators, vison correlators) coincide between the two models. Since Eq. (45) and the frozen/critical/gaseous phase assignment are imported through this correspondence, please provide a precise statement of the equivalence with a specific reference or proof, and explain in what sense the Arctic curve and local statistics transfer to the fortress.","section":"Appendix D and Section V B"},{"comment":"The abstract states that the diamond-shaped domain has a region 'exhibiting critical correlations,' but Section V C reports that in the intermediate region 'the numerical plots are less conclusive; however, we expect the dimer-dimer connected correlator to exhibit power-law decay.' No power-law fit, scaling collapse, or analytic argument for the fortress critical region is shown. Because the three-phase structure is a headline result, the claim of a critical phase should either be supported by data or explicitly downgraded to an expectation in both the abstract and the discussion.","section":"Section V C and abstract"},{"comment":"The discussion states that the fortress example 'demonstrates that the breakdown of the thermodynamic limit can persist beyond fine-tuned points,' but no perturbation analysis is performed; the only evidence offered is the expectation that the gapped central region survives, and Future Direction 3 lists an explicit quantum Monte Carlo test of stability as open. Please either present a concrete stability argument (for example perturbation theory or QMC data) or clearly label robust stability as a conjecture rather than a demonstrated result.","section":"Section VI (Discussion)"}],"minor_comments":[{"comment":"The sentence 'These properties can can be derived using Kasteleyn matrix methods' contains a duplicated word; please fix the typo.","section":"Section III A"},{"comment":"The text refers to 'Hsor,0' in the sentence following Eq. (21); this should be 'Hsos,0' for consistency with the notation introduced in Eq. (19).","section":"Section IV A"},{"comment":"The sentence comparing paths 'considered in Section V' should refer to Section IV, since the rectangular-boundary paths are defined there.","section":"Section V C"},{"comment":"The caption contains a garbled phrase, 'square paths a, which show a mixture of constant and exponential decay behavior'; please rephrase and define the labeling of square paths clearly.","section":"Figure 12 caption"},{"comment":"The caption spells 'Kastelen matrix condition number'; the correct spelling is 'Kasteleyn'. The same appendix also refers to 'Kastelen matrix condition number' in the title of the figure.","section":"Figure 22 caption and Appendix C"}],"recommendation":"major_revision","confidential_remarks":"The Aztec and rectangular square-octagon results are strong and likely publishable. The main risk is the fortress section: the equivalence with the weighted Aztec diamond needs a precise citation or proof, the critical-region claim in the abstract overstates the numerical evidence, and the robustness claim is presented too strongly. These are fixable, so I recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this paper contains a real technical contribution and two correct exact statements, but its advertised claim—'breakdown of the thermodynamic limit'—is only rigorously true for effective dimer models in the U→∞ limit at the RK point, not for the underlying spin Hamiltonians at finite coupling. Read it for the Kasteleyn machinery, not for the philosophical conclusion.\n\nWhat's actually new: the authors give an exact, efficient method for computing vison correlators in planar dimer RK wavefunctions by flipping signs in the Kasteleyn matrix and using determinants of small submatrices, with fraction-free LU to control the exponentially ill-conditioned Aztec/fortress geometries. On the square-octagon lattice with rectangular boundaries they compute the inverse Kasteleyn matrix asymptotics, show exponential dimer-dimer decay, and prove via the strong Szegő theorem that the vison correlator along an octagon path tends to sqrt(60)/10. That is a solid, well-checked calculation. The rectangular case is correctly identified as a gapped, short-range entangled ordered phase, not a Z2 spin liquid. The Aztec diamond example is an exact restatement of the classical Arctic circle theorem in quantum language. The fortress example—three regions, central region same as rectangular case—is consistent with the cited mathematics and with their numerics, and the paper honestly says the critical region's power-law is 'less conclusive' in the plots.\n\nSoft spots, in proportion: (1) The abstract and title claim more than is proven. The rigorous statements hold for the dimer model in the U→∞, RK-point limit. Finite-U spin Hamiltonian corrections could, in principle, alter the ground state. (2) The fortress's critical region and octic curve are imported from the weighted-Aztec correspondence (Appendix D), which is cited, not proved. This is acceptable practice, but the paper's central 'robustness' claim rests on a conjecture, explicitly flagged as such. (3) The stability away from RK is pure conjecture (their future direction #3 admits this). These are not fatal—the paper is careful about some of them—but the packaging overpromises.\n\nWho is this for? People working on quantum dimer models, RK wavefunctions, or exact correlator techniques in frustrated magnetism. The vison correlator method alone is worth the read. I'd send it to a competent referee rather than desk reject.\n\nRecommendation: publish after the authors recalibrate the abstract and prominently state the limit in which the thermodynamic-limit breakdown is rigorous.","headline":"A genuinely useful exact method and two clean examples of shape-dependent ground-state phase separation at RK points, wrapped in a title that overstates what is proven.","tokens_in":40479,"tokens_out":3456,"would_cite":true,"duration_ms":35020,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["82B20","82B26","05C70"],"pacs":[],"model":"deepseek-v4-flash","headline":"The thermodynamic limit can fail: the same spin Hamiltonian has different bulk phases on square and diamond domains.","keywords":["thermodynamic limit","quantum dimer models","Rokhsar-Kivelson wavefunction","Kasteleyn matrix","Aztec diamond","square-octagon fortress","vison correlator","phase separation"],"falsifier":"Compare the exact partition function and edge-occupation probabilities of the unweighted square-octagon fortress, computed directly at large radius with integer arithmetic, against the weighted-Aztec-diamond predictions; a discrepancy in the limiting frozen-region probabilities, or a central correlation length differing from $-\\log((3-\\sqrt{5})/2)^{-1} \\approx 1.039$, would falsify the claimed three-region structure.","tokens_in":39470,"feed_emoji":"🧩","tokens_out":7510,"duration_ms":74122,"temperature":0.7,"pith_summary":"At the core of statistical physics sits the assumption that a large system's bulk behavior does not care about the shape of its container. This paper tries to break that assumption with explicit quantum spin Hamiltonians whose exact ground states, at the Rokhsar–Kivelson point, are uniform sums over dimer coverings. On square-shaped domains the ground state is a single uniform phase, while on diamond-shaped domains of the same Hamiltonian macroscopic frozen regions appear at the corners, separated by a critical region (Aztec diamond) or by two additional regions including a gapped central phase (square-octagon fortress). The authors also supply an exact, highly efficient Kasteleyn-matrix method for the nonlocal vison correlator, and use it to show the square-octagon central phase is ordered and short-range entangled, not a Z2 spin liquid. If the claims hold, the thermodynamic limit of these models is not well defined unless the shape of the system and the location of the origin relative to that shape are specified.","feed_headline":"The same quantum model changes phase when the box shape does","feed_subtitle":"Exact ground states of spin-dimer Hamiltonians split into frozen, critical, and gapped regions—or don't—depending only on the boundary.","key_machinery":"The load-bearing object is the Kasteleyn matrix $K$, a signed adjacency matrix whose determinant counts dimer coverings and whose inverse gives all local correlation functions through Kenyon's formula. The paper's most useful identity is the vison correlator $\\langle(-1)^{\\eta_E}\\rangle = \\det(I - 2K'_E K^{-1}_E)$, which turns a nonlocal string correlator into a finite determinant; combined with fraction-free sparse LU decomposition, it yields exact correlators on ill-conditioned graphs. For the square-octagon lattice, the inverse Kasteleyn matrix is expressed by contour integrals of rational functions divided by the characteristic polynomial $P(z,w) = -5 + z + z^{-1} + w + w^{-1}$, whose gap controls exponential decay and whose zero set on the square lattice would give power laws. For the fortress, the three-region structure comes from a quoted equivalence between the unweighted fortress and a two-periodic weighted Aztec diamond with weights $a=b=c=1$ and $d=1/2$, whose Arctic octic curve is imported from the literature. The strong Szegő limit theorem for Toeplitz determinants supplies the infinite-string vison constant.","core_discovery":"The central discovery is that the thermodynamic limit can fail to be shape-independent in exact ground states of local quantum Hamiltonians. For the square lattice, the authors reverse-engineer a spin-1/2 Hamiltonian whose low-energy sector is the Rokhsar–Kivelson dimer model; at the RK point the ground state is the uniform superposition of all perfect matchings. On a rectangular domain this state is critical everywhere, with power-law dimer–dimer correlations. On the Aztec diamond—the same Hamiltonian with different boundary terms—the same state develops frozen staggered regions outside the Arctic circle, occupying about 21 percent of the area, with power-law critical correlations inside. For the square-octagon lattice, the rectangular-boundary model has exponentially decaying dimer correlations and a constant vison correlator along octagon paths, with the infinite-string constant 0.774596669..., indicating a gapped ordered phase. On the square-octagon fortress, the same model separates into three regions: frozen staggered corners, a critical intermediate region, and a central region with the same gapped short-range-entangled phase as the rectangular-boundary case, with phase boundaries given by the octic curve Eq. (45).","pith_inferences":["These constructions suggest a sharper definition of the thermodynamic limit: for shape-dependent phase separation, the limit should be taken along specified shapes, and one could classify Hamiltonians by whether their infinite-volume phase diagram depends on the limiting shape.","The same Kasteleyn determinant method should extend to Wilson-loop observables and to non-bipartite planar graphs, giving an exact probe of confinement in RK-type dimer models beyond the square-octagon lattice.","A Rydberg-atom implementation of the square-octagon dimer model with fortress boundaries could test the predicted constant vison correlator directly, creating coexisting frozen, critical, and gapped regions in a single experiment.","If perturbation away from the RK point leaves the gapped central region intact, shape-induced phase separation may be observable at the accessible sizes of current quantum simulators, since the macroscopic frozen region already appears at moderate radii."],"forward_implications":["If these claims hold, numerical or experimental studies of these Hamiltonians must specify not just lattice and boundary conditions but the global shape of the domain, because the infinite-size limit is not a single object.","The square-octagon fortress shows phase separation can occur between a gapped central region and frozen corners separated by a critical region, so the phenomenon is not confined to a single fine-tuned point; the gapped central region is expected to remain stable under small perturbations.","The exact Kasteleyn-based vison correlator, computed to precision around $10^{-21}$, replaces Monte Carlo estimates at the $10^{-6}$ level for RK wavefunctions on planar lattices, making spin-liquid diagnostics far cheaper.","The central region of the fortress and the square-boundary square-octagon lattice share the same phase, with exponentially decaying dimer correlations and a constant vison correlator; this phase is not a $\\mathbb{Z}_2$ quantum spin liquid but an ordered short-range-entangled state.","The frozen corner regions occupy a finite fraction of the system, about $1-\\pi/4 \\approx 21\\%$ for the Aztec diamond, so the phase separation is macroscopic rather than a boundary-localized effect."],"supporting_citations":[{"why":"Supplies the Rokhsar–Kivelson Hamiltonian whose RK point has a ground state that is the uniform superposition of dimer coverings.","marker":"[1]"},{"why":"Provides the Arctic circle theorem for the Aztec diamond, giving the frozen-versus-critical spatial separation used as the first example.","marker":"[3]"},{"why":"Gives Kenyon's formula connecting the inverse Kasteleyn matrix to dimer probabilities and correlation functions.","marker":"[5]"},{"why":"Introduces generalized domino shuffling and the fortress construction, supplying the square-octagon fortress geometry and its three-region picture.","marker":"[7]"},{"why":"Establishes Kasteleyn's determinant/Pfaffian method for counting dimer coverings, the foundation of all correlator calculations in the paper.","marker":"[9]"},{"why":"Provides the two-periodic Aztec diamond Arctic octic curve, imported as Eq. (45) for the fortress phase boundaries.","marker":"[24]"},{"why":"Gives the characteristic-polynomial criterion that distinguishes exponential from power-law decay of dimer correlations.","marker":"[29]"},{"why":"Supplies the fraction-free sparse LU decomposition used for exact correlator computations on ill-conditioned Kasteleyn matrices.","marker":"[42]"},{"why":"States the strong Szegő limit theorem used to evaluate the infinite-string vison correlator as a nonzero constant.","marker":"[53]"}],"fun_headline_variants":["Boundary shape alone can flip a quantum model's phase","Quantum phases change when the boundary shape does","Thermodynamic limit breaks: box shape dictates quantum phase","Same quantum model, different phase: just change the box","Spin-dimer model's phase is set by the box's shape"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The fortress phase diagram rests on an equivalence, quoted rather than proved here, between the unweighted square-octagon fortress and a specific weighted Aztec diamond; if that equivalence is inexact, the octic-curve region boundaries are unsupported.","fun_headline_variants_meta":{"raw":{"variants":["Boundary shape alone can flip a quantum model's phase","Quantum phases change when the boundary shape does","Thermodynamic limit breaks: box shape dictates quantum phase","Same quantum model, different phase: just change the box","Spin-dimer model's phase is set by the box's shape"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000896,"raw_usage":{"total_tokens":3907,"prompt_tokens":1040,"completion_tokens":2867,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":656,"completion_tokens_details":{"reasoning_tokens":2801}},"tokens_in":656,"tokens_out":2867,"duration_ms":20413,"temperature":1.0,"reasoning_tokens":2801,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T19:31:53.798180+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the exact partition function and edge-occupation probabilities of the unweighted square-octagon fortress, computed directly at large radius with integer arithmetic, against the weighted-Aztec-diamond predictions; a discrepancy in the limiting frozen-region probabilities, or a central correlation length differing from $-\\log((3-\\sqrt{5})/2)^{-1} \\approx 1.039$, would falsify the claimed three-region structure.","supporting_citations":[{"cited_title":"Chhita, K","cited_arxiv_id":null,"evidence_quote":"Gives the characteristic-polynomial criterion that distinguishes exponential from power-law decay of dimer correlations."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the fraction-free sparse LU decomposition used for exact correlator computations on ill-conditioned Kasteleyn matrices."},{"cited_title":"Bhakta and D","cited_arxiv_id":null,"evidence_quote":"States the strong Szegő limit theorem used to evaluate the infinite-string vison correlator as a nonzero constant."}],"review_version":2}