{"id":"a43e31ed-74af-4996-9bf0-6129e430e961","arxiv_id":"2411.13417","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Antiferromagnetic order in the Hubbard model shifts irreducible vertex divergences to stronger coupling and turns their character from charge-dominated to spin-dominated at an exceptional point.","lead":"This paper computes how the mathematical breakdown of perturbation theory, seen as divergences of vertex functions, changes when an antiferromagnet orders spontaneously. It finds that magnetic order pushes the breakdown to stronger interactions but does not remove it, and connects the behavior to the crossover from Slater to Heisenberg magnetism.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The mitigation claim depends on the unverified assumption that the 2D Mermin-Wagner-violating AF DMFT solution is a valid proxy for the 3D AF Hubbard phase; a 3D calculation would settle this.","rationale":"I read the paper as making a well-defined DMFT statement: within the two-particle DMFT solution of the 2D half-filled Hubbard model, allowing for spontaneously broken SU(2) symmetry shifts the vertex divergence lines to higher U relative to the constrained paramagnetic solution. The numerical evidence, including the public dataset and the consistent eigenvalue tracking across the PM-to-AF crossover, supports this as a property of the DMFT equations. The weakest point is the paper's own stated premise that these 2D broken-symmetry results are a faithful proxy for the AF phase of the Hubbard model in 3D, where long-range order is physical. This is not derived; it is asserted in Section II A and used to give physical significance to the central claim. The reader's CONDITIONAL verdict is appropriate because this proxy assumption is testable and unverified. My stress-test agrees with the reader's primary concern but does not elevate it to a rejection: the DMFT calculation is internally consistent, and the paper is transparent about the limitation. A 3D DMFT run would settle the proxy question directly. I do not take the secondary concern about Eq. (43) as load-bearing, because in DMFT the lattice irreducible vertex is local by construction, so the local impurity vertex is the standard diagnostic for vertex divergences. Thus I would keep the verdict CONDITIONAL rather than moving to ACCEPT or REJECT.","tokens_in":31863,"tokens_out":14517,"duration_ms":164722,"concrete_test":"Perform the same two-particle DMFT calculation on a 3D cubic lattice with t = 1/(2*sqrt(6)) (so the noninteracting density-of-states variance matches the 2D case with t = 1/4), compute the first four vertex divergence lines in the AF phase, and compare them to the corresponding PM lines. If the AF lines bend to higher U by a comparable amount, the Mermin-Wagner-violating 2D setup is a valid proxy and the central claim stands. If the shift is absent or reversed, the claimed mitigation is an artifact of the 2D mean-field AF order.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that the onset of AF order shifts the vertex divergence lines to higher U and therefore mitigates, but does not eliminate, the breakdown of self-consistent perturbation theory. This is established by comparing the AF DMFT solution to the PM DMFT solution on a 2D square lattice. However, at any finite T in 2D the Mermin-Wagner theorem forbids long-range AF order; the broken-symmetry solution is an artifact of the local, mean-field-like nature of DMFT. The authors explicitly acknowledge this in Section II A and assume, without derivation, that the 2D results are 'qualitatively similar' to a 3D Hubbard model where AF order is physical. If this proxy fails, the observed 'bending' of the divergence lines is not a property of AF order in the Hubbard model but a spurious effect of imposing broken symmetry in a dimension where it does not occur. The subsequent interpretations in terms of the Slater-Heisenberg crossover and the speculative implications for 2D systems are directly dependent on this assumption. Since the core observable (the shift of the divergence lines) is only physically meaningful if the AF phase it refers to is representative of the actual system, this unverified dimensional-lifting assumption is the most load-bearing weakness.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper extends the study of two-particle irreducible vertex divergences to the antiferromagnetically ordered phase of the half-filled Hubbard model, using dynamical mean-field theory (DMFT) and, for comparison, static mean-field/RPA. The central numerical finding is that, below the Néel temperature, the vertex divergence lines bend toward larger interaction U relative to the paramagnetic phase, indicating that long-range AF order partially mitigates but does not eliminate the breakdown of self-consistent perturbation theory. The authors also identify exceptional points at which divergence lines change from charge-dominated to spin-dominated character, trace real-part zero-crossing (RZ) lines toward the Slater regime, and connect the eigenvalue structure of the local generalized susceptibility to the Slater-Heisenberg crossover. A generalized Bethe-Salpeter analysis is used to discuss possible thermodynamic instabilities in the AF phase, and the final section speculates about implications for two-dimensional systems where Mermin-Wagner prevents true long-range order.","tokens_in":32162,"tokens_out":4941,"duration_ms":61336,"significance":"If the central claim holds, the paper establishes a qualitatively new result: the onset of AF order shifts but does not remove the non-perturbative breakdown of self-consistent perturbation theory. This would extend a decade of paramagnetic-phase studies into symmetry-broken phases and provides a two-particle diagnostic of the Slater-Heisenberg crossover, which is valuable for future diagrammatic extensions of DMFT. The paper has notable strengths: the formal BSE derivation for the Bethe lattice in Appendix E is transparent; the small-magnetization approximation in Section IV is explicitly tested; and the authors provide a public dataset and plotting scripts (Appendix A, Ref. [88]). The central claims are, however, currently supported by a 2D DMFT solution that violates the Mermin-Wagner theorem, by divergence-line locations reported without statistical uncertainty, and by an eigenvector selection criterion whose robustness is not established. These issues affect the strength of the central conclusion and require attention before the paper can be accepted.","major_comments":[{"comment":"The central physical conclusion—that AF order mitigates but does not prevent the breakdown of perturbation theory—is established in a 2D DMFT solution in which long-range AF order at finite temperature is forbidden by the Mermin-Wagner theorem. The authors explicitly acknowledge this and assume in Section II A that their 2D results are 'qualitatively similar' to those of a 3D Hubbard model, without derivation or benchmarking. This assumption is load-bearing: if the broken-symmetry 2D DMFT solution is an artifact of the local mean-field-like approximation, the reported bending of divergence lines below TN may not be a property of AF order in the Hubbard model. I request a concrete test of this proxy, for example a 3D cubic-lattice DMFT calculation at representative (U,T) points, or at least a quantitative justification based on variance-matched density of states and a comparison of the local vertex in the AF phase. The current statement, while honest, is not sufficient to support the paper's main claim about the Hubbard model.","section":"Section II A and Section VI"},{"comment":"The claimed 'clear bending' of the divergence lines below TN and the apparent kink at TN are central results, but the reported lines are not accompanied by error bars on the eigenvalue zero crossings. Since the eigenvalues are obtained from CT-QMC data for the two-particle susceptibility, statistical and systematic noise can shift the zero-crossing locations; without quantified uncertainties, the bending could be within numerical error. In addition, the low-U portion of the TN curve in Fig. 3(b) is estimated from a mean-field result with a manually adjusted rescaling factor ('albeit with a slightly higher factor', Ref. [60]), and the exact factor is not stated. Please specify the rescaling procedure and provide either error bars on the divergence lines or a sensitivity analysis showing that the bending is robust to QMC noise and to the TN normalization.","section":"Figure 3(b) and Section III A"},{"comment":"The interpretation in terms of a Slater-Heisenberg crossover relies on the eigenvalues/eigenvectors 'associated with the lowest Matsubara frequencies', selected by the criterion in Eq. (48) using the lowest five Matsubara frequencies. The paper acknowledges that the quantitative behavior of χc is not reproduced without including more eigenvectors, but it still uses this subset to draw the qualitative conclusion that low-frequency eigenvalues capture the Slater-Heisenberg change. This selection threshold N* is a free parameter, and no convergence test with respect to N* or an alternative objective criterion is reported. Please demonstrate that the qualitative conclusions—negative low-frequency contributions in the Slater AF, less-negative contributions in the Heisenberg AF—are robust to N* and, ideally, corroborate the crossover location with an independent one-particle or thermodynamic diagnostic (e.g., double occupancy or the kinetic-energy gain).","section":"Section III B, Eq. (48), and Fig. 4"},{"comment":"The small-magnetization approximation of Eqs. (57)-(58), which is used to derive the effective renormalization t^2 D_i and the combined stability condition of Eq. (63), is verified in Fig. 8 only for U=3 and for four specific temperatures/magnetizations (m = 0.31, 0.61, 0.78, 0.92). The subsequent phase-separation discussion and the claim that the main effects of SU(2) breaking are general rely on this approximation being valid in the regime where it is applied. Please test the approximation at weak coupling (e.g., U=1, where RZ lines and the Slater regime are located) and either provide an estimate of the neglected off-diagonal couplings or explicitly restrict the conclusions to the parameter range where the approximation has been verified.","section":"Section IV, Eqs. (56)-(63)"}],"minor_comments":[{"comment":"The phrase 'hopping parameters to the next neighbors' should read 'hopping parameters to the nearest neighbors'.","section":"Fig. 1 caption"},{"comment":"The text states Γ = β δΣ[G]/δG, while Eq. (40) contains an explicit factor β/2; please reconcile the notation or add a sentence explaining the factor.","section":"Eq. (40) and surrounding text"},{"comment":"There is a typo in 'particle-holy symmetric' (should be 'particle-hole symmetric').","section":"Section VI"},{"comment":"The term 'exceptional point' is used for a degeneracy of two zero eigenvalues of χ; since 'exceptional point' already has a standard meaning in the context of non-Hermitian matrices, please clarify the relation between the two uses or add a footnote.","section":"Section III A"},{"comment":"Ref. [41] is cited as an arXiv preprint; if it has been published by the time of publication, please update the reference.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The reader's conditional verdict is fair. My main additional concern beyond the Mermin-Wagner proxy is the lack of quantitative uncertainty estimates for the central divergence-line locations and the unspecified low-U TN rescaling in Fig. 3(b). These are fixable within the manuscript's scope: the authors can add error bars, a sensitivity analysis, and a 3D or Bethe-lattice benchmark, or alternatively reframe the central claim as a statement about the DMFT AF solution rather than about the Hubbard model. I do not see grounds for rejection, provided the requested robustness checks are supplied."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new result is the mapping of irreducible vertex divergences inside the AF-ordered DMFT solution: the divergence lines bend to higher U below TN, acquire mixed spin-charge character, pass through exceptional points, and are accompanied by RZ lines that connect to the Slater-Heisenberg crossover. That is a real extension of the established paramagnetic vertex-divergence program, and it is done carefully. The symmetry analysis of the eigenvectors, the Bethe-lattice derivation of the BSE in the AF phase, and the distinction between true divergences and RZ lines are all solid. The data are public, which makes the numerical claims checkable. The paper also earns credit for what it does not overclaim. The Mermin-Wagner limitation is stated plainly in Section II A, and the speculative discussion about the exact 2D system in Section VI is clearly labeled as speculation. Appendix G is a good example of honest negative reasoning: it shows that negative eigenvalues are necessary but not sufficient for the charge-response enhancement. The soft spots are real but not fatal. The zero-crossing locations of the eigenvalues are read off without error bars, so the quantitative bending of the divergence lines is less certain than the figure suggests. The low-frequency eigenvector selection in Eq. (48) is a hand-picked subset, and the authors themselves concede that it is qualitative. The low-U TN rescaling is another free input. On the dimensional-lifting concern: I think the stress-test note overstates the problem. The primary claim is about the broken-symmetry DMFT solution, and within that solution the bending is well-defined; the 3D extrapolation is an assumption, but the authors flag it, and the qualitative similarity to 3D is plausible given how DMFT handles dimensionality. The weaker link is genuinely the unquantified numerical uncertainty, not the MW issue. Who should read this: people working on diagrammatic extensions of DMFT in ordered phases, and anyone tracking the vertex-divergence literature. It deserves a serious referee. I would send it out with a request for error estimates on the divergence-line locations and a tightened statement about what is DMFT-AF physics versus what is expected for the actual 2D model.","headline":"First systematic map of two-particle vertex divergences in the AF-ordered DMFT solution, with a plausible mitigation claim that is weaker where the 2D-to-3D extrapolation is invoked.","tokens_in":724,"tokens_out":679,"would_cite":true,"duration_ms":23783,"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":"Antiferromagnetic order weakens, but does not cure, the breakdown of perturbation theory in the Hubbard model.","keywords":["vertex divergences","irreducible vertex","antiferromagnetic Hubbard model","dynamical mean-field theory","Slater-Heisenberg crossover","generalized susceptibility","Mermin-Wagner theorem","non-perturbative breakdown"],"falsifier":"Run the same two-particle DMFT calculation in the three-dimensional Hubbard model, where Néel order is thermodynamically stable at finite temperature, and check whether the vertex-divergence lines bend toward larger $U$ once inside the ordered phase. If they do not bend, or if a two-dimensional calculation that restores the forbidden long-range order shows divergence lines moving to smaller $U$ instead of accumulating near $T=U=0$, the paper's central claim and its 2D extrapolation would both be contradicted.","tokens_in":31707,"feed_emoji":"🧲","tokens_out":9161,"duration_ms":91054,"temperature":0.7,"pith_summary":"The paper asks a concrete question: does the onset of magnetic order rescue the many-body perturbation expansion in a strongly correlated electron system? The answer, from dynamical mean-field theory applied to the half-filled Hubbard model in its antiferromagnetic phase, is that order helps but does not save. The curves where the irreducible vertex diverges—the formal sign that self-consistent perturbation theory has broken down—shift to larger interaction strengths once Néel order sets in, and they also change character from charge-dominated to spin-dominated at exceptional points. This places a precise boundary on where diagrammatic resummation methods can be trusted in magnetically ordered phases, and it ties the shift directly to the crossover from a weak-coupling Slater antiferromagnet to a strong-coupling Heisenberg antiferromagnet.","feed_headline":"Magnetic order bends, not erases, perturbation-theory failure lines","feed_subtitle":"Antiferromagnetic order shifts the breakdown of perturbative methods to larger interaction strength, but it persists.","key_machinery":"The central object is the local generalized susceptibility matrix $\\chi$, organized in charge and spin blocks over Matsubara frequencies, together with the Bethe-Salpeter rewriting $\\chi_{q\\pm} = [\\chi^{-1} + \\Delta\\chi^{-1}_{0q}]^{-1}$ in which momentum enters only through the inverse-bubble difference. Vertex divergences are detected as zero eigenvalues of $\\chi$, because the irreducible vertex is $\\Gamma = \\chi^{-1} - \\chi_0^{-1}$. In the antiferromagnetic phase the off-diagonal $\\chi_{sc}$ and $\\chi_{cs}$ blocks become nonzero and purely imaginary, coupling spin and charge and making the divergence lines exchange character at exceptional points.","core_discovery":"Below the Néel temperature, the divergence lines of the two-particle irreducible vertex in the Hubbard model bend toward higher $U$ relative to their paramagnetic position, so spontaneous antiferromagnetic order mitigates but does not eliminate the breakdown of the self-consistent perturbation expansion. In the AF phase the charge and spin sectors of the susceptibility couple, and a zero eigenvalue of the generalized susceptibility matrix now produces a divergence in a mixed spin-charge channel; at an exceptional point the corresponding eigenvector switches from predominantly charge to predominantly spin character, and that point coincides with the suppression of the largest spin eigenvalue responsible for the Curie-like magnetic response of the Mott bad metal. In the Slater weak-coupling regime only “real-part zero-crossing” lines, which are complex-conjugate eigenvalue pairs with zero real part, are found just as in RPA, while true vertex divergences appear in the Heisenberg strong-coupling regime. The local charge response is suppressed by AF order in the Slater regime but enhanced in the Heisenberg regime, tracing the same split to a delicate balance between bubble and vertex contributions.","pith_inferences":["The pattern “order mitigates but does not remove the singularity” may be generic: any broken-symmetry state that leaves the underlying interaction strong could show a similar bending of non-perturbative boundaries rather than their disappearance.","If the RZ-to-divergence conversion under restoration of spin symmetry holds, then inexpensive RPA-level calculations in symmetry-broken states could be used to map where non-perturbative physics will appear in more exact treatments.","The eigenvalue-fountain picture suggests a practical diagnostic: the appearance of complex-conjugate eigenvalue pairs in mean-field susceptibility matrices can flag a regime where a fully self-consistent calculation will develop genuine vertex divergences, even though the mean-field vertex itself cannot diverge."],"forward_implications":["Bold diagrammatic resummations and parquet-type schemes aimed at antiferromagnetically ordered phases should use two-particle vertices computed directly in the ordered state, because a paramagnetic starting point would miss the Slater-to-Heisenberg crossover identified here.","The exceptional point where a divergence line changes from charge- to spin-dominated provides a sharp two-particle diagnostic of the crossover between Slater and Heisenberg antiferromagnetism.","In two dimensions, where long-range order is forbidden at finite temperature, the “RZ lines” of the broken-symmetry solution are expected to become actual divergence lines accumulating toward $T=U=0$, implying that the self-consistent perturbation expansion has no convergence radius at zero temperature.","At finite doping the coupled spin-charge instability condition of the AF Bethe-Salpeter equation predicts simultaneous divergences of spin and charge responses, relevant for phase-separation instabilities in doped antiferromagnets."],"supporting_citations":[{"why":"Establishes that the irreducible vertex of the DMFT solution of the Hubbard model diverges at zero eigenvalues of the generalized susceptibility, the phenomenon this paper extends to the AF phase.","marker":"[7]"},{"why":"Provides the paramagnetic $T$-$U$ map of the first vertex divergence lines that the AF calculation is compared against.","marker":"[24]"},{"why":"Shows that vertex divergences reverse the sign of the effective electronic interaction and drive phase-separation instabilities, the physical consequence explored here.","marker":"[18]"},{"why":"Links eigenvalues of the local susceptibility to thermodynamic instabilities of the lattice, the logic generalized to the AF phase.","marker":"[23]"},{"why":"Documents the kinetic-energy gain and optical signature of Heisenberg-type AF order used to interpret the enhanced charge response.","marker":"[36]"},{"why":"Supplies the Bethe-Salpeter formalism for spontaneously SU(2)-broken phases used in the AF calculations.","marker":"[40]"},{"why":"Gives the sign-switch criterion that controls the stability of iterative self-consistent perturbation theory, used to distinguish RZ lines from true breakdown.","marker":"[57]"},{"why":"Supplies the Néel temperature curve used to mark the AF phase boundary in the phase diagram.","marker":"[59]"}],"fun_headline_variants":["Antiferromagnetism bends perturbation breakdown lines, not erases","AF order shifts vertex divergences but can't remove them","Magnetic order tempers, doesn't cure, Hubbard model divergences","From Slater to Heisenberg: AF order reshapes vertex singularities","Magnetic order fails to kill Hubbard model's vertex divergences"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that a two-dimensional dynamical mean-field solution with spontaneously broken spin symmetry, which violates the theorem forbidding long-range magnetic order in two dimensions, is a valid stand-in for the true antiferromagnetic Hubbard model and correctly locates the lattice vertex divergences.","fun_headline_variants_meta":{"raw":{"variants":["Antiferromagnetism bends perturbation breakdown lines, not erases","AF order shifts vertex divergences but can't remove them","Magnetic order tempers, doesn't cure, Hubbard model divergences","From Slater to Heisenberg: AF order reshapes vertex singularities","Magnetic order fails to kill Hubbard model's vertex divergences"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001092,"raw_usage":{"total_tokens":4590,"prompt_tokens":1003,"completion_tokens":3587,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":619,"completion_tokens_details":{"reasoning_tokens":3498}},"tokens_in":619,"tokens_out":3587,"duration_ms":27305,"temperature":1.0,"reasoning_tokens":3498,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:25:13.014645+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same two-particle DMFT calculation in the three-dimensional Hubbard model, where Néel order is thermodynamically stable at finite temperature, and check whether the vertex-divergence lines bend toward larger $U$ once inside the ordered phase. If they do not bend, or if a two-dimensional calculation that restores the forbidden long-range order shows divergence lines moving to smaller $U$ instead of accumulating near $T=U=0$, the paper's central claim and its 2D extrapolation would both be contradicted.","supporting_citations":[{"cited_title":"Chalupa, T","cited_arxiv_id":null,"evidence_quote":"Provides the paramagnetic $T$-$U$ map of the first vertex divergence lines that the AF calculation is compared against."},{"cited_title":"Reitner, P","cited_arxiv_id":null,"evidence_quote":"Shows that vertex divergences reverse the sign of the effective electronic interaction and drive phase-separation instabilities, the physical consequence explored here."},{"cited_title":"Dagotto, J","cited_arxiv_id":null,"evidence_quote":"Documents the kinetic-energy gain and optical signature of Heisenberg-type AF order used to interpret the enhanced charge response."},{"cited_title":"Sangiovanni, A","cited_arxiv_id":null,"evidence_quote":"Supplies the Bethe-Salpeter formalism for spontaneously SU(2)-broken phases used in the AF calculations."}],"review_version":1}