{"id":"5d1beb69-a7e0-45fc-b0ca-9c7794fc3505","arxiv_id":"2505.21287","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The paper proves existence and, in a smaller parameter range, uniqueness of solutions to a Matsubara frequency discretization of the IPT-DMFT equations, analyzes the Hubbard dimer, and reports numerical evidence that discretized solutions can fail the Pick interpolation criterion.","lead":"This mathematics paper studies a standard numerical discretization of the IPT-DMFT equations from condensed matter physics, proving when the discretized equations have solutions and when those solutions are unique. It also shows, on the Hubbard dimer, that converged discretized solutions can fail to correspond to any analytic Green's function, a caution for practical DMFT calculations.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The Section 4.2 algebraic reduction to the polynomial system (46)/(57) is asserted without derivation; the Nω=1 solution-count claims and sparsity results depend on it exactly.","rationale":"The reader's verdict is CONDITIONAL, with the algebraic reduction listed as an issue but the discrete-continuous convergence chosen as the weakest assumption. I agree with the conditional verdict but put the weight on the algebraic reduction: it is a gap in a proof of a claimed result, not a conjectural limitation. The existence/uniqueness theorems appear rigorous; the minor boundary issue (strict vs non-strict inequality in Theorem 3.1) and the 1/β vs β factor in Lemma 3.4 are typographical and do not change the conclusions. The TRIQS mismatch is honestly disclosed and does not affect the theorems. Therefore the paper should be accepted subject to providing the full derivation of (46)/(57) and the verification of the Nω=1 counts, which is exactly the CONDITIONAL status.","tokens_in":25985,"tokens_out":29837,"duration_ms":272488,"concrete_test":"Use a computer algebra system to re-derive (57) from (19)-(20) for the Hubbard dimer, Nω=1, in the variables x_0,x_1. Verify symbolically that the resulting system is exactly (58)-(59). Then, for a grid of (a,b) in [0,10]×[0,25], solve both the derived system and the original discretized equations (19)-(20) numerically (e.g., with HomotopyContinuation.jl or a multi-start Newton method), and compare the numbers of complex and admissible (real in (0,1]^2) solutions. Any mismatch invalidates the §4.2 reduction.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4.2 asserts that for bipartite systems, the purely imaginary admissible solutions of (19)-(20) are exactly the real solutions of the algebraic system (46), with dimer specialization (57). The derivation is compressed into 'it follows that' and no proof or appendix is given. This is load-bearing for the claimed 'some results for Nω=1': the 16-complex-solution count and 1-5 admissible solutions reported in §4.3.2 come from applying homotopy continuation to (58)-(59), not to the original equations. If the variable change x_n=1/(1+y_n) or the clearing of denominators introduces spurious solutions or drops boundary solutions (e.g., y_n=0), those numbers would not describe the IPT-DMFT solutions. The theorem statements 3.1/3.5 do not depend on this reduction, so this is not a challenge to those proofs, but it is an unverified cornerstone of the algebraic-characterization claims.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper studies the Matsubara-frequency (MaF) discretization of the IPT-DMFT equations for translation-invariant Hubbard models. In the dimensionless formulation (19)–(20), the authors prove via Brouwer's fixed-point theorem that for each cutoff N_ω there exists t_{N_ω} > 0 such that for 0 ≤ t ≤ t_{N_ω}/sqrt(deg(G_H)) the discretized equations admit a solution (Δ,Σ) with nonpositive imaginary parts (Theorem 3.1). Under the additional smallness condition t^2 u^2 deg(G_H) < η_{N_ω}, they prove uniqueness and linear convergence of the fixed-point iteration (Theorem 3.5). For bipartite graphs, particle-hole symmetry reduces the problem to a polynomial system in the variables x_n = 1/(1+y_n); the dimer case N_ω = 0 is fully characterized as a function of α = U/T (Theorem 4.2), and for N_ω = 1 numerical algebraic geometry suggests 16 complex solutions and 1–5 admissible solutions. The final section reports TRIQS and Julia simulations showing a conductor-to-insulator transition and cases where the converged values violate the Pick criterion.","tokens_in":26127,"tokens_out":12412,"duration_ms":130220,"significance":"These appear to be the first existence and uniqueness results for the discretized IPT-DMFT equations, and they are obtained from explicit fixed-point estimates rather than by fitting parameters. The N_ω = 0 dimer theorem is a complete, rigorous characterization of the admissible solutions in the low-temperature limit. The algebraic-geometry viewpoint in Section 4 is promising and, if fully justified, would give a systematic way to count and classify solutions for larger N_ω. The numerical experiments are clearly labelled as illustrations and include a careful high-precision check of the Pick criterion, which is a useful caution for practitioners of the MaF discretization. The main limitations are that the algebraic reduction in Section 4.2 is not proved in the manuscript, and the convergence of discretized solutions to the continuous IPT-DMFT solutions is only conjectured (Remark 5.1).","major_comments":[{"comment":"The passage from the rational equations (19)–(20) to the polynomial system is asserted with 'it follows that' and no proof is given. This equivalence is load-bearing for the N_ω = 1 results in Section 4.3.2, because all solution counts are computed on the polynomial system (58)–(59), not on the original equations. A rigorous derivation must justify the change of variables x_n = 1/(1+y_n), the spectral decomposition of h_perp^0, and the clearing of denominators, and it must rule out spurious roots introduced by that clearing and the possible loss of boundary solutions (e.g., x_n = 1, i.e., y_n = 0). Please provide the full derivation in the text or in an appendix.","section":"Section 4.2, Eqs. (46) and (57)"},{"comment":"The Lipschitz constant L_{N_ω} is defined as a supremum over D_t^{N_ω}, which depends on the parameter t. The subsequent identification η_{N_ω} = π^2/L_{N_ω} therefore makes η_{N_ω} depend on t, contrary to the theorem's statement that η_{N_ω} depends only on N_ω. The argument can be repaired by taking the supremum over the largest admissible set, D_{t_{N_ω}/sqrt(deg(G_H))}^{N_ω}, which depends only on N_ω, but as written the proof is incomplete.","section":"Proof of Theorem 3.5"},{"comment":"The assertions that (58)–(59) have 16 complex solutions and that the number of admissible solutions ranges from 1 to 5 are obtained from HomotopyContinuation.jl, not from a proved theorem. Similarly, the separation line b = 3(1 + a/10)^3 is stated without computation. If these are numerical observations, the text should label them as such; if they are intended as rigorous 'some results for N_ω = 1', the manuscript must supply proofs or certificates. This matters because the abstract promises results for N_ω = 1.","section":"Section 4.3.2, Eqs. (58)–(59)"}],"minor_comments":[{"comment":"The Matsubara frequency in the definition of F_DMFT_Nω is written 'i2(n+1)π' rather than 'i(2n+1)π'; this is inconsistent with all other occurrences and appears to be a typo.","section":"Eq. (34)"},{"comment":"The statement says F_Nω(D_t^{Nω}) ⊂ −C_+ for all 0 ≤ t ≤ t_Nω, but because D_t^{Nω} = t^2 deg(G_H) C_Nω, the proof actually requires t ≤ t_Nω / sqrt(deg(G_H)); the statement and proof should be aligned.","section":"Lemma 3.4"},{"comment":"The caption writes 'ω_n = 2(n+1)π/β'; the correct Matsubara frequency is ω_n = (2n+1)π/β.","section":"Figure 2 caption"},{"comment":"The bullet states that there is a unique solution 'in C+' with iΔ∞_{α,t} → +∞; for the admissible set −C_+ this should read 'in −C_+' so that the stated limit is consistent.","section":"Theorem 4.2, first bullet"},{"comment":"The quantity φ(n,Nω) = π^3 Im(F_Nω(0)_n) is described as a 'positive rational number', but it is negative (for example φ(0,0) = −3); the word 'positive' should be removed or the sign convention explained.","section":"Proof of Lemma 3.4"},{"comment":"The phrase 'it is highlited' should be 'it is highlighted'.","section":"Section 1, last paragraph"}],"recommendation":"major_revision","confidential_remarks":"The main obstacle is the unproved algebraic reduction in Section 4.2; without it, the advertised N_ω = 1 results are not supported. Theorems 3.1 and 3.5 are essentially sound, with the η_{N_ω} dependence in Theorem 3.5 being a fixable gap. The numerical N_ω = 1 counts should be explicitly framed as computational observations rather than theorems. The paper otherwise fits the journal's scope and would be a useful contribution once the derivation is supplied."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper does what it says: for a Matsubara-frequency discretization of the IPT-DMFT equations, it proves existence of admissible solutions for t deg(G)^(1/2) below a cutoff (Theorem 3.1) and uniqueness plus linear convergence of the fixed-point iteration under a stronger smallness condition (Theorem 3.5). These are real theorems with explicit bounds and clean proofs via Brouwer and Picard. The dimer analysis for Nω=0 is strong: Theorem 4.2 completely characterizes the number of solutions in the low-temperature limit as a function of α=U/T, with a threshold at 3π/2. The numerical observation that the discretized solution can violate the Pick criterion is also useful and honestly reported.\n\nThe soft spots are real but mostly not load-bearing for the main theorems. Section 4.2's reduction of the bipartite equations to the polynomial system (46) is asserted with 'it follows that' and no derivation. The Nω=1 solution counts in Section 4.3.2 come from homotopy continuation applied to (58)-(59), not to the original equations, so if the variable change or clearing denominators introduced spurious solutions, those counts would not describe IPT-DMFT solutions. The theorems in Section 3 do not depend on this reduction, so the paper's central existence and uniqueness claims stand. A second issue: the numerical phase transition in Section 5.1 uses TRIQS with tail-fitting, which the authors themselves note differs from the equations analyzed. The Julia simulations in Section 5.2 are faithful to the discretization, and that is where the Pick-violation result is obtained. Finally, the link between the discretized and continuous equations is only conjectured (Remark 5.1); the paper does not prove convergence as Nω→∞. That limits the physical interpretation of the results, but the mathematical claims are about the discretized system, and they hold.\n\nOverall: this is a solid, careful paper. The main theorems are correct and the dimer characterization is new and nontrivial. It belongs in peer review; the referees should ask for a derivation of (46) (an appendix would do) and clearer separation of the TRIQS-based numerics from the analyzed equations.","headline":"Rigorous existence and uniqueness for the MaF-discretized IPT-DMFT equations, with a real but non-fatal gap: the algebraic reduction behind the Nω=1 solution counts is asserted, not proved.","tokens_in":26663,"tokens_out":2427,"would_cite":true,"duration_ms":27763,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":["65H10","65H20","82B80"],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper proves that the discretized IPT-DMFT equations admit physically admissible solutions for small hopping at any finite Matsubara cutoff, unique under a stronger smallness condition, with multiple admissible solutions possible for…","keywords":["dynamical mean-field theory","IPT-DMFT equations","Matsubara frequency discretization","existence and uniqueness","Hubbard dimer","Pick criterion","particle-hole symmetry","algebraic solution counting"],"falsifier":"At a single parameter pair with $\\alpha=U/T$ fixed just above $3\\pi/2$ and $t$ large, solve the $N_\\omega=0$ dimer equation (48) with arbitrary precision; Theorem 4.2 predicts exactly three admissible solutions, so any count other than three would falsify the characterization.","tokens_in":25767,"feed_emoji":"⚛️","tokens_out":12854,"duration_ms":116890,"temperature":0.7,"pith_summary":"This paper proves that the Matsubara-frequency discretization of the IPT-DMFT equations—the version of dynamical mean-field theory that solves the impurity model by second-order perturbation theory—is mathematically well posed. For each cutoff $N_\\omega$, existence of a physically admissible solution (components in $-C_+$, i.e. with nonpositive imaginary part) is shown under the condition that the dimensionless hopping $t=\\beta T$ is not too large, with the allowed range shrinking with the graph degree. Uniqueness, and linear convergence of the natural fixed-point iteration, holds when $t^2u^2\\deg(G_H)$ is below a cutoff-dependent constant. For bipartite systems the equations reduce to a sparse real algebraic system; for the Hubbard dimer at $N_\\omega=0$ the number of admissible solutions is fully characterized in the low-temperature limit, with a transition at $\\alpha=U/T=3\\pi/2$. Numerical simulations exhibit a conductor–insulator transition and show that, for large enough $U$ at small $N_\\omega$, the converged discrete Green's function cannot be interpolated by a Pick function.","feed_headline":"DMFT on a finite frequency grid is proven solvable","feed_subtitle":"Existence and uniqueness of IPT-DMFT solutions are established; the Hubbard dimer's solution count jumps at U/T = 3π/2.","key_machinery":"The Matsubara-frequency (MaF) discretization: the unknown functions $\\Delta$ and $\\Sigma$, originally analytic on the upper half-plane, are represented by their values at the lowest $N_\\omega+1$ imaginary Matsubara frequencies $i\\omega_n$, $\\omega_n=(2n+1)\\pi/\\beta$. The nonlocal IPT equation is replaced by the rational map $F_{N_\\omega}$ defined by a convolution over triples of frequencies (equation (18)), and the full fixed-point map is $F^{\\mathrm{DMFT}}_{N_\\omega}(\\Delta)_n = t^2 w^T (i(2n+1)\\pi - t h^0_\\perp - u^2[F_{N_\\omega}(\\Delta)]_n)^{-1} w$. Existence is proved by Brouwer's fixed-point theorem on the compact set $D_t^{N_\\omega}=\\{z: |z_n|\\le t^2\\deg(G_H)/((2n+1)\\pi)\\}$, while uniqueness and linear convergence follow from the contraction estimate of Lemma 3.6. For bipartite systems, the map preserves the imaginary axis, and in the variables $x_n=1/(1+y_n)$ the equations reduce to the sparse polynomial system (57), whose solution count is studied with symbolic and numerical algebraic geometry.","core_discovery":"The central claim is Theorem 3.1: for every Matsubara cutoff $N_\\omega$ there is a positive constant $t_{N_\\omega}$ such that, whenever $0\\le t \\le t_{N_\\omega}/\\sqrt{\\deg(G_H)}$, equations (19)–(20) have a solution $(\\Delta,\\Sigma)$ with all components in the physically admissible cone $-C_+$. The obstruction this overcomes is that the discretized IPT map $F_{N_\\omega}$ does not preserve $-C_+$ for $N_\\omega>0$ (the counterexample is equation (23)); the proof uses Brouwer's fixed-point theorem on the compact set $D_t^{N_\\omega}$. Theorem 3.5 shows uniqueness of that solution—and linear convergence of the simple fixed-point iteration (33)–(34)—under the additional smallness condition $t^2u^2\\deg(G_H)<\\eta_{N_\\omega}$. For bipartite graphs, particle-hole symmetry forces solutions onto the imaginary axis and the discretized equations become a sparse polynomial system in variables $x_n\\in(0,1]$; the paper completely solves the $N_\\omega=0$ Hubbard dimer in the low-temperature limit, locating a transition in the count of admissible solutions at $\\alpha=3\\pi/2$, and reports that for $N_\\omega=2$ and $5$ the converged numerical solution can have a negative Pick-matrix eigenvalue, meaning no Pick function interpolates the discrete Green's function. The paper conjectures (Remark 5.1) that as $N_\\omega\\to\\infty$ the Pick-criterion violations disappear and the discrete solutions converge to the continuous ones.","pith_inferences":["Editorial inference: the sharp threshold $\\alpha=3\\pi/2$ at $N_\\omega=0$, where the count of admissible solutions jumps from 1 to 3, is a plausible finite-cutoff precursor of the Mott transition; tracking how this threshold moves with $N_\\omega$ would give a quantitative test of the paper's convergence conjecture.","Editorial inference: the lowest eigenvalue of the Pick matrix computed from a converged discrete Green's function could serve as a practical a posteriori error indicator, telling a user how many Matsubara frequencies are needed before the discrete data become analytically continuable.","Editorial inference: because the existence proof only covers small $t$, the strong-coupling regime where the numerics find insulator-like solutions may lie outside the range of guaranteed admissibility—an open question whether those solutions are physical or artifacts of the finite cutoff."],"forward_implications":["For any finite Matsubara cutoff, the IPT-DMFT fixed-point loop has a guaranteed starting point in $D_t^{N_\\omega}$ from which a physically admissible solution exists, as long as $t$ is below the graph-degree-adjusted bound.","In the uniqueness regime $t^2u^2\\deg(G_H)<\\eta_{N_\\omega}$, the simple fixed-point iteration converges linearly, with a rate controlled by the same constants used in the existence proof.","For the Hubbard dimer at $N_\\omega=0$ and low temperature, the number of admissible solutions changes at $\\alpha=U/T=3\\pi/2$: below this ratio only one solution exists (escaping to infinity with $t$), while above it two additional bounded solutions appear.","Numerical simulations on the Hubbard dimer at $\\beta=1$ show the spectral density at zero frequency $\\rho(0)$ vanishing between $U\\approx 6$ and $8$, identifying a conductor–insulator transition in the IPT-DMFT approximation.","For $N_\\omega=2$ and $5$, the converged discrete Green's function can have a negative Pick-matrix eigenvalue, proving that no Pick function interpolates the discrete data; the paper conjectures this artifact disappears as $N_\\omega\\to\\infty$.","The physical reading of the theorems rests on the unproved assumption that solutions of the discretized equations converge to solutions of the continuous IPT-DMFT equations as the Matsubara cutoff grows; the paper only conjectures this in Remark 5.1."],"supporting_citations":[{"why":"Supplies the continuous IPT-DMFT equations, the definition of the IPT solver $F^{\\mathrm{IPT}}_\\beta$, and the Pick-function framework that the discretization extends.","marker":"[5]"},{"why":"Provides the Brouwer fixed-point theorem used to prove existence in Theorem 3.1.","marker":"[35]"},{"why":"States the Nevanlinna–Pick interpolation theorem and the Pick-matrix criterion tested in Section 5.2.","marker":"[28]"},{"why":"The TRIQS library used for the numerical simulations of the conductor–insulator transition.","marker":"[29]"},{"why":"Provides the homotopy-continuation methods used to count complex solutions of the $N_\\omega=1$ dimer system.","marker":"[2]"},{"why":"Used to compute the dimension of the algebraic variety of solutions for the $N_\\omega=1$ dimer system.","marker":"[12]"}],"fun_headline_variants":["Discretized IPT-DMFT equations proven to have solutions","Existence and uniqueness for discretized IPT-DMFT equations","Particle-hole symmetry yields purely imaginary IPT-DMFT solutions","Solution count for Hubbard dimer jumps at U/T = 3π/2"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The physical reading of the theorems rests on the unproved assumption that solutions of the discretized equations converge to solutions of the continuous IPT-DMFT equations as the Matsubara cutoff $N_\\omega$ grows; the paper only conjectures this in Remark 5.1.","fun_headline_variants_meta":{"raw":{"variants":["Discretized IPT-DMFT equations proven to have solutions","Existence and uniqueness for discretized IPT-DMFT equations","Particle-hole symmetry yields purely imaginary IPT-DMFT solutions","Solution count for Hubbard dimer jumps at U/T = 3π/2"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00051,"raw_usage":{"total_tokens":2616,"prompt_tokens":1211,"completion_tokens":1405,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":827,"completion_tokens_details":{"reasoning_tokens":1328}},"tokens_in":827,"tokens_out":1405,"duration_ms":12938,"temperature":1.0,"reasoning_tokens":1328,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:30:28.926414+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"At a single parameter pair with $\\alpha=U/T$ fixed just above $3\\pi/2$ and $t$ large, solve the $N_\\omega=0$ dimer equation (48) with arbitrary precision; Theorem 4.2 predicts exactly three admissible solutions, so any count other than three would falsify the characterization.","supporting_citations":[{"cited_title":"A mathematical analysis of IPT-DMFT, June","cited_arxiv_id":null,"evidence_quote":"Supplies the continuous IPT-DMFT equations, the definition of the IPT solver $F^{\\mathrm{IPT}}_\\beta$, and the Pick-function framework that the discretization extends."},{"cited_title":"Shapiro.A Fixed-Point Farrago","cited_arxiv_id":null,"evidence_quote":"Provides the Brouwer fixed-point theorem used to prove existence in Theorem 3.1."},{"cited_title":"The Nevanlinna-Pick Interpolation Problem","cited_arxiv_id":null,"evidence_quote":"States the Nevanlinna–Pick interpolation theorem and the Pick-matrix criterion tested in Section 5.2."},{"cited_title":"TRIQS: A toolbox for research on interacting quantum systems.Computer Physics Communications, 196:398–415, 2015","cited_arxiv_id":null,"evidence_quote":"The TRIQS library used for the numerical simulations of the conductor–insulator transition."},{"cited_title":"HomotopyContinuation.jl: A Package for Homotopy Continuation in Julia","cited_arxiv_id":null,"evidence_quote":"Provides the homotopy-continuation methods used to count complex solutions of the $N_\\omega=1$ dimer system."}],"review_version":1}