{"id":"c386eaa3-ae55-4f48-9156-d6ce4104f430","arxiv_id":"1908.10748","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Dual fermion calculations show that K-point antiferromagnetic spin fluctuations of the half-filled triangular Hubbard model persist from the insulator into the metal, and the computed spectra match qualitative features of neutron and NMR experiments.","lead":"Computing spin and charge response functions of the triangular-lattice Hubbard model with the dual fermion method reveals strong low-energy spin fluctuations at the K point in the Mott insulator that also persist, at higher energy, in the metallic phase. The calculated spectra are compared to neutron scattering and NMR data on frustrated triangular-lattice materials, providing a direct theoretical handle on these experiments.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The metallic-phase K-point spin-fluctuation claim rests on an unbenchmarked ladder dual-fermion susceptibility; a concrete cross-check is needed before accepting the crossover.","rationale":"After reading the paper and the reader's verdict, the most defensible central claim is the continuous evolution of K-point spin fluctuations from the insulator into the metal. The evidence is a single method (ladder DF) plus MaxEnt continuation. I looked for internal inconsistencies or places where the inference overreaches. The paper is transparent about the approximation and even lists known artifacts (the nonzero Gamma-point peaks). The comparison to experiments is explicitly qualitative, with a fitted hopping scale; that weakens but does not break the claim. The load-bearing point is that the metallic-phase onset near U=7t is a small spectral shift in an unbenchmarked method. The known square-lattice benchmark shows momentum-dependent errors, and there is no triangular-lattice benchmark; the K point is precisely where a 120-degree AFM vertex would be strongest, so overestimation is plausible. This is not a disagreement with consensus; it is a correctness-risk issue. A concrete independent cross-check (DCA/DMRG S(K, omega) or systematic DF ladder-order variation) would settle it. Since the reader already made this conditional-acceptance point, my stress-test does not change the verdict.","tokens_in":10334,"tokens_out":4617,"duration_ms":57544,"concrete_test":"Compute the K-point spin response S(K, omega) or Im chi_m(K, omega) for U=6t, 7t, and 8.2t at T=t/6 with an independent method, e.g. DCA/CDMFT with cluster sizes Nc=4, 8, 12 or DMRG on a sufficiently wide cylinder. If the low-energy K-point enhancement disappears or shifts by more than about 0.1t between U=6t and 7t relative to the DF result, the metallic-phase claim would not survive outside the ladder DF approximation; if it reproduces the same trend, the concern is settled.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is that K-point (120-degree AFM) spin fluctuations persist into the metallic phase and 'start to condense' around U=7t. The only evidence is the ladder dual fermion (DF) susceptibility, analytically continued by MaxEnt with no error bars. The authors themselves state in the Method section that DF is 'uncontrolled in practice', and the square-lattice benchmark (Ref. 49) found momentum-dependent errors and an overall energy-scale factor; no analogous benchmark exists for the triangular lattice. This matters specifically because the DF ladder violates spin conservation, producing the acknowledged nonzero Gamma-point spin and charge excitations in Figs. 1 and 5. The K point is not protected from this Ward-identity violation, and the claimed sharp drop in omega_m(K) between U=6t and 7t is a small energy shift in a quantity with no estimated error. If the DF vertex overestimates magnetic correlations near the metal-insulator crossover, the metallic K-point peak and its 'condensation' around U=7t could be an artifact, while the insulating-phase result would remain qualitatively correct. This does not make the paper unsound, but it leaves the headline crossover result quantitatively unverified.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies the half-filled triangular-lattice Hubbard model at T=t/6 using the ladder dual fermion approximation with maximum-entropy analytic continuation. It presents momentum- and energy-resolved spin and charge susceptibilities for U=6t (metal), U=8.2t (crossover), and U=12t (insulator). The central claim is that strong K-point spin fluctuations corresponding to 120-degree antiferromagnetic fluctuations exist not only in the insulator but also in the metallic phase, moving to higher energy as U decreases, and that they begin to condense around U=7t. The paper also presents simulated neutron spectra, spin-lattice relaxation rates, comparisons to Ba8CoNb6O24 and κ-(ET)2Cu2(CN)3, and charge susceptibilities relevant to momentum-resolved electron-loss spectroscopy.","tokens_in":10565,"tokens_out":6409,"duration_ms":70300,"significance":"The metallic-phase persistence of K-point spin fluctuations and their evolution across the metal-insulator crossover would be a useful result beyond the usual strong-coupling spin-model description, and the momentum-resolved spectra are valuable for interpreting neutron and NMR experiments on triangular compounds. The paper is commendably transparent: it explicitly states that the ladder DF approximation is uncontrolled in practice, acknowledges the spurious nonzero Gamma-point spin and charge excitations, identifies the energy-scale uncertainty in the Ba8CoNb6O24 comparison, and includes a supplementary exact-diagonalization check of the Hubbard versus Heisenberg spin response. If the central crossover claim is supported by an additional cross-check, the paper would make a solid contribution; currently the evidence for that claim is not quantitatively established.","major_comments":[{"comment":"The metallic-phase part of the central claim, namely that strong K-point spin fluctuations persist into the metal and condense around U=7t, rests solely on the ladder dual fermion susceptibility, which the Method section describes as 'uncontrolled in practice'. The only cited benchmark, Ref. 49, is for the square lattice and demonstrates momentum- and interaction-dependent scaling effects; no triangular-lattice benchmark is provided. Because the same calculation produces the acknowledged spurious nonzero spin and charge excitations at Gamma (Fig. 1(d)-(f); Fig. 5(d)-(f)), the K-point result is not protected by a conservation law. I therefore cannot determine whether the drop in omega_m(K) between U=6t and U=7t is physical or an artifact of the vertex approximation. A concrete cross-check at one metallic U, such as cluster DMFT, a finite-size QMC susceptibility, or a quantitative transfer of the Ref. 49 error estimates to the triangular lattice, is needed to support the central claim.","section":"Method; Figs. 1 and 2"},{"comment":"The claim that spin fluctuations 'start to condense' around U=7t is based on omega_m(K), which is the peak position of Im chi_m(K,omega) obtained from maximum-entropy analytic continuation and is shown without error bars. The sharp drop between U=6t and U=7t is a relatively small energy shift on the scale of the spectra, and MaxEnt peak positions carry systematic uncertainty from the choice of default model and from Monte Carlo noise. Please provide uncertainty estimates, for example by varying the MaxEnt default model or bootstrap resampling the raw QMC data, or alternatively phrase the condensation claim more cautiously as a qualitative observation.","section":"Fig. 2 and surrounding text"}],"minor_comments":[{"comment":"The caption contains a typo: 'pannel' should be 'panel'; in the supplement text 'To valid this simplification' should read 'To validate this simplification'.","section":"Supplement, Fig. S3 caption"},{"comment":"The supplement refers to 'Fig. 1(b)' for spin excitations and 'Fig. 4(b)' for charge excitations in the main text; the correct references appear to be Fig. 1(d)-(f) and Fig. 5(d)-(f), respectively.","section":"Supplement, bare susceptibility discussion"},{"comment":"The phrase 'is invisible in the insulator (U=12t)' is ambiguous; the charge susceptibility is small but not necessarily zero, so 'suppressed below the scale of the plot' or 'negligible' would be more precise.","section":"Main text, Fig. 5 discussion"},{"comment":"The density of states is plotted together with omega_m(K) but the DOS axis is not labeled in the panel; the caption should state explicitly which curve and scale refer to the DOS, as the right axis appears unlabeled.","section":"Fig. 2(a)"},{"comment":"The affiliation line on the supplementary material contains a formatting error: '1,2Center for Computational Quantum Physics' duplicates the numbering and should be cleaned up.","section":"Supplement, author affiliation line"}],"recommendation":"major_revision","confidential_remarks":"The paper is within scope of the journal and the authors are unusually candid about the limitations of their method. The main blocking issue is not the use of an approximate method per se, but the absence of any independent or error-controlled support for the specific metallic-phase crossover claim that forms the paper's headline. One clean cross-check at a single metallic interaction strength would be sufficient to make the claim credible; I would not require a full benchmark of the method. The experimental comparison section is more limited than the abstract suggests, but it is presented honestly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a solid, honest computational paper, and the main new claim—that K-point 120-degree spin fluctuations survive into the metallic phase—is worth taking seriously even though it rests on an approximation the authors themselves label 'uncontrolled in practice.' The paper should be peer-reviewed; the referee should ask for one concrete cross-check or a sharper caveat.\n\nWhat is new: previous cluster DMFT studies looked at static susceptibilities or phase boundaries; this is the first systematic dual-fermion (DF) calculation of momentum- and frequency-resolved spin and charge response across the metallic, crossover, and Mott regimes of the triangular Hubbard model. The K-point fluctuation result is new, as is the comparison to Ba8CoNb6O24 and κ-(ET)2Cu2(CN)3. They use the open DF code, ALPS MaxEnt, and the supplement adds an ED comparison of Hubbard vs Heisenberg spectra showing the 4t^2/U mapping fails at U=12t—that is honest and useful. The Gamma-point artifacts are flagged as conservation-law violations of DF, not hidden.\n\nWhere I'd push: the load-bearing result—the drop in ω_m(K) between U=6t and 7t and the persistence of K fluctuations in the metal—comes from a ladder DF susceptibility with no triangular-lattice benchmark and no error bars on the MaxEnt peak positions. The square-lattice benchmark (Ref. 49) found momentum-dependent scaling errors; that does not automatically transfer. The K point is not protected from the Ward-identity violation, so a spurious enhancement there is possible. The energy-scale uncertainty is acknowledged, and the experimental fit uses t=3 meV, which is transparent. These are real limitations but not fatal: the paper's conclusions are phrased as suggestions, and the insulating-phase result is consistent with known 120-degree order.\n\nBottom line: the paper is a useful contribution for anyone working on frustrated triangular systems or on two-particle response functions from DF. It deserves a serious referee. I'd recommend acceptance after the authors either add a benchmark (e.g., compare to a small-cluster ED susceptibility or a DCA result at the same parameters) or explicitly bracket the metallic-phase result as provisional. The core is sound; the cross-check is a matter of measurement.","headline":"A transparent DF study of triangular-lattice spin and charge response whose headline metallic-phase spin-fluctuation result is plausible but quantitatively unverified; worth refereeing.","tokens_in":11059,"tokens_out":2839,"would_cite":true,"duration_ms":29488,"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":"Strong spin fluctuations at the K point exist in both the Mott insulator and the metallic phase of the half-filled triangular Hubbard model, moving to higher energy as the interaction weakens.","keywords":["triangular lattice Hubbard model","dual fermion approximation","spin susceptibility","120-degree antiferromagnetic order","Mott metal-insulator crossover","spin-lattice relaxation","charge susceptibility","neutron spectroscopy"],"falsifier":"Compute the dynamic magnetic susceptibility at $U=6t$ and $7t$, $T=t/6$, with an unbiased method that does not rely on the dual fermion truncation, for example exact diagonalization on the largest accessible cluster or another controlled cluster method; if the K-point peak of $\\mathrm{Im}\\,\\chi_m(\\mathbf{K},\\omega)$ disappears or becomes no stronger than the bare-bubble response while the density of states still shows a metal, the paper's central claim is refuted. Experimentally, momentum-resolved inelastic neutron scattering on a metallic half-filled triangular compound with $U$ near the crossover should show a resolvable K-point spin excitation above the single-particle continuum; its absence would falsify the claim.","tokens_in":10110,"feed_emoji":"🧲","tokens_out":8012,"duration_ms":76536,"temperature":0.7,"pith_summary":"The paper aims to establish that the half-filled triangular-lattice Hubbard model develops strong momentum- and energy-resolved spin fluctuations at the K point, the 120-degree antiferromagnetic wavevector, not only in the Mott insulating regime but also on the metallic side of the metal-insulator crossover. Using the ladder dual fermion approximation, the authors compute static and dynamic spin and charge susceptibilities at $T=t/6$ for $U=6t$, $8.2t$, and $12t$, and find that K-point spin fluctuations persist at $U=6t$--$7t$ while their characteristic energy rises as $U$ decreases. If the claim holds, the magnetic response of the model evolves continuously through the crossover, and spin-model descriptions that apply only at large $U$ miss an entire metallic regime of strong magnetic correlations. The paper also connects these spectra to inelastic neutron scattering on Ba$_8$CoNb$_6$O$_{24}$ and to NMR relaxation measurements on an organic triangular compound, and presents charge susceptibilities as predictions for momentum-resolved electron-loss spectroscopy.","feed_headline":"Spin fluctuations persist from Mott insulator into the metal","feed_subtitle":"Half-filled triangular Hubbard model keeps 120-degree magnetic fluctuations alive in the metallic phase.","key_machinery":"The load-bearing object is the two-particle magnetic susceptibility $\\chi_m(\\mathbf{q},\\omega)$ resolved on a $24\\times24$ momentum grid of the triangular Brillouin zone, computed in the ladder dual fermion approximation, a diagrammatic extension of dynamical mean-field theory that keeps local correlations nonperturbative and adds nonlocal spin and charge correlations at the ladder level. The K point, the corner of the hexagonal Brillouin zone, is the wavevector conjugate to the 120-degree antiferromagnetic spin pattern, so the location and energy of the peak in $\\chi_m(\\mathbf{K},\\omega)$ diagnose the ordering tendency. The same two-particle machinery produces the spin-lattice relaxation rate and the bare-bubble comparison that exposes the role of vertex corrections. The authors stress that total spin and charge conservation are violated at the $\\Gamma$ point in this approximation, producing spurious nonzero $\\Gamma$-point excitations.","core_discovery":"At $T=t/6$ the static magnetic susceptibility $\\chi_m(\\mathbf{q},0)$ shows a dominant peak at the K point for $U=12t$, signalling incipient 120-degree antiferromagnetic order. The dynamic susceptibility shows that this K-point spin excitation persists at $U=8.2t$ and $U=6t$, but moves to higher energy; the maximum intensity at the K point grows slowly between $U=6t$ and $7t$ and rapidly beyond, while the density of states at the Fermi level shows the system is still metallic at $U=7t$. The authors conclude that strong spin fluctuations begin to condense at the K point around $U\\sim 7t$ inside the metallic phase. The spin-lattice relaxation rate $(T_1T)^{-1}$ rises with $U$, approaching magnetic-order behaviour near $T\\approx0.125t$ at $U=12t$, and the computed neutron spectra at $U=12t$ reproduce the K-point intensity pattern seen in Ba$_8$CoNb$_6$O$_{24}$ when $t=3$ meV. Charge susceptibilities are suppressed by $U$, show uniform charge response in the metal, and require vertex corrections beyond the bare bubble.","pith_inferences":["If K-point spin fluctuations persist into the metal, then transport quantities such as the resistivity or the optical scattering rate should show signatures of a finite magnetic correlation length that grows as $U$ approaches the crossover; the paper does not compute these, but they are direct consequences of the susceptibility it reports.","The spurious $\\Gamma$-point intensity in the dual fermion results suggests the quantitative energy scale of the metallic K-point fluctuations may shift under a more accurate method; a benchmark against exact diagonalization on small clusters at the same temperature would separate the physical peak from the approximation artifact.","The collapse of the K-point excitation energy between $U=6t$ and $7t$ resembles the behavior one would expect if a broad crossover separates a weakly correlated metal from a spin-fluctuation-dominated metal; identifying whether the associated length scale diverges would require larger clusters and lower temperatures.","Because the Heisenberg mapping fails below $U\\sim20t$, effective exchange couplings extracted by fitting spin models to neutron data on correlated triangular materials may be systematically biased; re-fitting with the Hubbard model at finite $U$ could change the inferred $J$ values."],"forward_implications":["The metal at $U\\lesssim8t$ already carries incipient 120-degree antiferromagnetic correlations, so the magnetic response does not switch on abruptly at the Mott transition; it strengthens continuously as $U$ grows.","The spin-excitation energy at the K point drops sharply between $U=6t$ and $7t$ while the system remains metallic, locating the onset of strong K-point fluctuations inside the metallic phase rather than at the insulator boundary.","At $U=12t$, the computed neutron scattering pattern compares well with Ba$_8$CoNb$_6$O$_{24}$, and the Hubbard model is argued to be a more appropriate starting point than a Heisenberg model, which only becomes equivalent for $U>20t$.","The temperature dependence of $(T_1T)^{-1}$ for $U=6t$, $8.2t$, and $12t$ tracks the pressure dependence seen in $\\kappa$-(ET)$_2$Cu$_2$(CN)$_3$, suggesting interaction strength plays the role of pressure.","The momentum-resolved charge susceptibility in the metal is concentrated at the Brillouin-zone boundary at high energy and at the zone center statically, giving a concrete prediction for electron-energy-loss measurements."],"supporting_citations":[{"why":"Supplies the inelastic neutron scattering data on Ba8CoNb6O24 that the U=12t spin susceptibility is compared against.","marker":"[9]"},{"why":"Provides DMRG evidence for 120-degree antiferromagnetic order in the insulating triangular Hubbard model, justifying the K-point focus.","marker":"[8]"},{"why":"The square-lattice dual fermion error assessment that the authors rely on for the claim that the approximation's momentum dependence is accurate.","marker":"[49]"},{"why":"Introduces the dual fermion formalism on which the ladder approximation used throughout is built.","marker":"[35]"},{"why":"Documents the violation of total spin and charge conservation in dual fermion that explains the spurious Gamma-point spectral weight.","marker":"[57]"},{"why":"Previous dual fermion study placing the metal-insulator crossover of the triangular Hubbard model, used to select U=6t, 8.2t, and 12t as representative.","marker":"[53]"},{"why":"Provides the 1H NMR spin-lattice relaxation data on kappa-(ET)2Cu2(CN)3 under pressure that the calculated (T1T)^{-1} curves are compared to.","marker":"[2]"},{"why":"The supplementary material containing the bare susceptibility, the quasiparticle weight, and the Hubbard-versus-Heisenberg comparison that supports the vertex-correction argument.","marker":"[55]"}],"fun_headline_variants":["K-point spin fluctuations survive metallic crossover in triangular Hubbard model","Triangular Hubbard model: spin fluctuations outlive Mott transition","Spin fluctuations persist from insulator to metal in triangular lattice","Spin fluctuations move to higher energy as system metallizes"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire result rests on the ladder dual fermion approximation being accurate enough in momentum-resolved magnetic response at $T=t/6$; the paper itself calls this approximation 'uncontrolled in practice', and known momentum-dependent errors such as the spurious $\\Gamma$-point response mean the metallic K-point fluctuations could in principle be an artifact.","fun_headline_variants_meta":{"raw":{"variants":["K-point spin fluctuations survive metallic crossover in triangular Hubbard model","Triangular Hubbard model: spin fluctuations outlive Mott transition","Spin fluctuations persist from insulator to metal in triangular lattice","Spin fluctuations move to higher energy as system metallizes"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000627,"raw_usage":{"total_tokens":2908,"prompt_tokens":964,"completion_tokens":1944,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":580,"completion_tokens_details":{"reasoning_tokens":1879}},"tokens_in":580,"tokens_out":1944,"duration_ms":14871,"temperature":1.0,"reasoning_tokens":1879,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:34:35.520246+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the dynamic magnetic susceptibility at $U=6t$ and $7t$, $T=t/6$, with an unbiased method that does not rely on the dual fermion truncation, for example exact diagonalization on the largest accessible cluster or another controlled cluster method; if the K-point peak of $\\mathrm{Im}\\,\\chi_m(\\mathbf{K},\\omega)$ disappears or becomes no stronger than the bare-bubble response while the density of states still shows a metal, the paper's central claim is refuted. Experimentally, momentum-resolved inelastic neutron scattering on a metallic half-filled triangular compound with $U$ near the crossover should show a resolvable K-point spin excitation above the single-particle continuum; its absence would falsify the claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the inelastic neutron scattering data on Ba8CoNb6O24 that the U=12t spin susceptibility is compared against."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"The square-lattice dual fermion error assessment that the authors rely on for the claim that the approximation's momentum dependence is accurate."},{"cited_title":"Yoshioka, A","cited_arxiv_id":null,"evidence_quote":"Documents the violation of total spin and charge conservation in dual fermion that explains the spurious Gamma-point spectral weight."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Previous dual fermion study placing the metal-insulator crossover of the triangular Hubbard model, used to select U=6t, 8.2t, and 12t as representative."},{"cited_title":"Kurosaki, Y","cited_arxiv_id":null,"evidence_quote":"Provides the 1H NMR spin-lattice relaxation data on kappa-(ET)2Cu2(CN)3 under pressure that the calculated (T1T)^{-1} curves are compared to."},{"cited_title":"The white dashed line rep- resents the boundary of the Brillouin zone","cited_arxiv_id":null,"evidence_quote":"The supplementary material containing the bare susceptibility, the quasiparticle weight, and the Hubbard-versus-Heisenberg comparison that supports the vertex-correction argument."}],"review_version":1}