{"id":"b7172303-e35e-462d-a365-e976ef6d28ae","arxiv_id":"2608.08104","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Real-time D-GW simulations predict momentum-selective electronic heating, a quench-like antiferromagnetic response with frozen correlation length, and a local-moment melting signature in the photoexcited Hubbard model.","lead":"Using a real-time simulation of the photoexcited 2D Hubbard model, this paper predicts that nodal electrons heat up faster than antinodal electrons, while antiferromagnetic spin fluctuations heat up strongly but retain their correlation length. The results give a unified microscopic framework for momentum-dependent ultrafast signals seen in photoemission, Raman, and RIXS experiments on cuprates.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central effective-temperature comparisons (N vs AN, AFM vs electronic) rely on linear fits whose frequency windows are not specified; for the strongly nonthermal M-point magnon distribution, the reported overheating may be a fitting artifact.","rationale":"I considered the reader's identified weakness (NCA accuracy for two-particle observables) but found it less decisive for the current paper: the authors explicitly acknowledge the need for three-point vertex corrections for quantitative agreement (Sec. III D), so the qualitative conclusions are presented as approximate. A more targeted, immediately checkable weakness is the extraction of effective temperatures. The central claims are phrased in terms of T_eff inequalities (Fig. 3g-i), yet the fit windows are not specified for the bosonic modes, and the electron window (around the quasiparticle peak) is problematic where a pseudogap suppresses that peak. Because the M-point magnon distribution is demonstrably nonthermal (Fig. 5c: frozen peak position while the width tracks the electron T), a linear fit to F^m_M(ω) need not represent any meaningful temperature; the reported excess T_eff may simply reflect the fitting convention. This is not a question of method accuracy but of interpretation of the computed spectra, and it can be tested without new simulations by re-analyzing the existing distribution functions. If the ordering is robust to the window, the claim survives; if not, it should be reframed as a qualitative nonthermal redistribution rather than a temperature excess. The caption error in Fig. 4 and the missing code/data are secondary. Overall, the reader's CONDITIONAL verdict remains appropriate, possibly with an added requirement to report the fit windows.","tokens_in":23113,"tokens_out":7399,"duration_ms":75946,"concrete_test":"Using the stored Wigner-transformed distribution functions F^m_M(ω), F^m_Γ(ω), F^e_N(ω), and F^e_AN(ω) at the probe times shown in Fig. 3, re-extract T_eff with two protocols: (1) a narrow window |ω|≤0.02D around zero frequency, and (2) a wide window covering the full spectral peak (e.g., |ω|≤0.1D). If the ordering T_eff^N > T_eff^AN and T_eff^M > T_eff^e, and the magnitude of the AFM excess, are not preserved between protocols, then the reported momentum-selective heating is an artifact of the unspecified fit window. This test uses only data already computed and would settle whether the headline claim is convention-dependent.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline results in Fig. 3(g-i) compare effective temperatures extracted from distribution functions F(t,ω)=ln(A^</A^>) via Eq. (15). For electrons the fit window is described as 'the frequency window surrounding the quasiparticle peak at ω=0' (Sec. II B), but for the bosonic modes at Γ and M no window is specified, and the data are 'omitted at early and late times' because the spin susceptibility decays slowly. This is load-bearing: the AFM mode is shown to be strongly nonthermal (peak width follows the electron temperature while the peak position is frozen, Fig. 5), so F^m_M(ω) need not be linear; a single slope fitted over an unspecified window can produce an arbitrarily high T_eff. The same applies to the N-AN electron difference: fitting a pseudogapped AN spectrum near ω=0 may return a low T_eff simply because the quasiparticle weight is depleted, not because the mode is colder. Without specifying the fit windows and demonstrating robustness to the window choice, the central quantitative claims ('N hotter than AN', 'AFM T_eff >> electron T_eff') are not uniquely defined. The paper's own demonstration of nonthermal dynamics (Fig. 5c) makes this ambiguity concrete rather than hypothetical.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a real-time D-GW study of the photoexcited half-filled single-band Hubbard model on the square lattice, with the impurity problem solved by NCA. It claims three main results: (i) a transient nodal–antinodal anisotropy in effective electronic temperature, with nodal quasiparticles heating faster than antinodal ones; (ii) a strongly nonthermal magnetic response in which the M=(π,π) antiferromagnetic mode heats far above the electronic temperature while its peak position (correlation length) stays nearly frozen, followed by a momentum-space magnon cascade; and (iii) a dynamical criterion for local-moment formation based on a slow, interaction-dependent peak in the real-time local spin susceptibility, whose photoinduced melting is tracked. The paper compares these predictions qualitatively with tr-ARPES, ultrafast Raman, and time-resolved RIXS experiments on cuprates.","tokens_in":23212,"tokens_out":4011,"duration_ms":41498,"significance":"If substantiated, the paper would provide a unified microscopic framework for momentum-selective ultrafast dynamics in correlated electron systems, a topic of active experimental interest. The strength of the work is that it goes beyond DMFT by including nonlocal spin fluctuations in real time, and it produces several specific, falsifiable predictions: the N-AN temperature anisotropy, the frozen AFM peak position with hot magnon distribution, and the slow-χ_loc signature of local moments. The paper ships no code, but the method and equilibrium phase diagram are taken from the authors' prior published work, and the numerical setup is described in sufficient detail to be reproduced. The main significance rests on the reliability of the NCA-based two-particle spectra and on the effective-temperature extraction, both of which need scrutiny before the quantitative claims can be accepted.","major_comments":[{"comment":"The central quantitative claims that N quasiparticles are hotter than AN excitations and that the AFM mode reaches T_eff far above the electronic temperature depend on linear fits to the distribution functions F(t,ω) via Eq. (15), but the fit windows are not specified. For electrons the text states a window 'surrounding the quasiparticle peak at ω=0,' yet for the bosonic modes at Γ and M no window is given, and the data are explicitly omitted at early and late times because the spin susceptibility decays slowly. Since the M-point distribution is shown to be strongly nonthermal in Fig. 5(c) (peak position frozen while the width follows the electronic temperature), F^m_M(ω) need not be linear, and a single slope extracted over an unspecified window can yield an arbitrary T_eff. The same ambiguity applies to the AN spectrum, where the pseudogap depletes spectral weight near ω=0. The authors should specify the exact frequency intervals used for every T_eff extraction, show representative fits and their goodness, and demonstrate that the N-vs-AN and AFM-vs-electron differences are robust to the window choice. Without this, the headline numbers in Fig. 3(g-i) are not uniquely defined.","section":"Sec. II.B, Eq. (15), Fig. 3(g-i)"},{"comment":"The load-bearing two-particle observables—the momentum-resolved spin susceptibility, the magnon spectra A^m_q(t,ω), and the local spin susceptibility χ^m_loc(t,τ)—are computed with the non-crossing approximation (NCA) as the impurity solver within D-GW. NCA is an approximate real-time solver whose accuracy for two-particle response functions is not established, and the authors themselves state in Sec. III.D that quantitative agreement with equilibrium D-TRILEX would require three-point vertex corrections and a more accurate impurity solution. Because the frozen AFM correlation length, the excessive AFM effective temperature, and the slow component of χ^m_loc are all inferred from NCA-based two-particle spectra, the possibility that these features are numerical artifacts needs to be addressed. I request at least one equilibrium benchmark of χ^m(q,ω) or χ^m_loc(τ) against a more reliable solver (e.g., OCA/IPT, or exact diagonalization on a small cluster) at the relevant U and T, or, alternatively, a clear downgrading of the quantitative T_eff comparisons to qualitative statements in the abstract and conclusions.","section":"Sec. III.D, Sec. II.C"},{"comment":"The proposed dynamical criterion for local-moment formation is defined only qualitatively. The green LMM crossover line in Fig. 2(a) is said to be determined from the criterion introduced in Sec. III.D, but that section describes the slow peak only as 'well-separated' and does not specify a quantitative threshold or algorithm (e.g., a minimum peak position in τ, a minimum weight ratio between the slow and fast components, or a curvature condition). The extraction of the LMM melting point in the nonequilibrium trajectory (green square in Fig. 2a) is therefore not reproducible. A precise operational definition of the criterion is needed, including how the peak position and separation are measured on the logarithmic τ axis and how the crossover is assigned when the slow peak shifts and merges with the fast peak during the pump.","section":"Sec. III.D, Fig. 2(a)"},{"comment":"The paper presents the results as a 'microscopic explanation' for momentum-dependent phenomena observed in cuprates, but the calculations are for the half-filled single-band Hubbard model, while the cited tr-ARPES and Raman experiments (Refs. [21,22,44]) are on optimally doped cuprates with a hole-doped Fermi surface. The half-filled model possesses particle-hole symmetry and its pseudogap is driven purely by commensurate AFM fluctuations; the doped cuprate pseudogap involves a different Fermi-surface topology and additional mechanisms. The paper should state this mapping limitation explicitly and separate the model-independent mechanism (AFM fluctuation-driven momentum-selective heating) from the direct material comparison. As written, the abstract and conclusion claim more experimental relevance than the model can support.","section":"Sec. I, Sec. III.B"}],"minor_comments":[{"comment":"The caption labels the left column as 'U=1.15' and the right column as 'U=0.85' in one place and 'weakly correlated metal (U=1.15)' in the same caption, which is contradictory; the weakly correlated case is U=0.85 and the correlated metallic regime is U=1.15. Please correct the caption and ensure panel labels match the text.","section":"Fig. 4 caption"},{"comment":"Equation (15) contains a typographical error: an extra closing parenthesis appears after 'µ_eff(¯t)'.","section":"Eq. (15)"},{"comment":"The notation T^loc_eff(¯t) appears in the Fig. 4 caption and in Sec. III.B but is not defined in Sec. II.B, where only T_eff(¯t) is introduced. Please clarify whether the reference spectrum is evaluated at the local (momentum-averaged) effective temperature and define this quantity explicitly.","section":"Sec. II.B, Fig. 4 caption"},{"comment":"The legend labels for the magnon effective temperatures, 'A(m)_Γ' and 'A(m)_M', are confusing because A is used both for spectral functions and for these curves; using T^Γ_eff(¯t) and T^M_eff(¯t) would be clearer.","section":"Fig. 3(g-i)"},{"comment":"The text says 'we extract effective temperature of spin excitations'; this should be 'effective temperatures of spin excitations'.","section":"Sec. III.C"}],"recommendation":"major_revision","confidential_remarks":"The paper builds on the authors' own previous D-GW work (Ref. [40]) and reuses the equilibrium phase diagram from that paper; this is appropriate but means the novelty is concentrated in the nonequilibrium predictions. The reported comparison with experiment relies on two unpublished or private references (Ref. [13] with a placeholder-style DOI, and Ref. [45] as private communication), which should be verified before publication. The central quantitative claims are vulnerable to the effective-temperature fit ambiguity and to the known limitations of NCA for two-particle spectra; if the authors can provide the requested robustness checks, the paper would be a solid contribution, but in its current form the quantitative statements outrun the evidence."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know before reading further. The paper is not another equilibrium-phase-diagram study; it is a real-time D-GW simulation of a photoexcited half-filled 2D Hubbard model that makes concrete predictions: nodal quasiparticles heat faster than antinodal ones, the M-point antiferromagnetic mode reaches an effective temperature far above the electronic one while its peak position stays frozen, magnetic spectral weight cascades to long wavelengths, and the slow tail of the local spin susceptibility tracks local-moment melting. The predictions are new for this method and, importantly, nothing is fitted to the experimental observations they are meant to explain, so the circularity burden is genuinely low. The dimer analysis in the appendix is a nice analytical anchor for the susceptibility-peak interpretation.\n\nWhere the paper is soft: the effective-temperature machinery is under-specified. Eq. (15) is fit to a window described only as 'surrounding the quasiparticle peak at ω=0'; for the bosonic modes the window is not given at all, and the data are omitted at early and late times. Given that the paper itself shows the AFM distribution is strongly nonthermal (frozen peak, moving width), a single slope over an unspecified range can return an almost arbitrary T_eff. The N-AN difference is less sensitive, but a pseudogapped AN spectrum could bias the same extraction. This does not sink the qualitative story, which is supported by the spectral-weight redistribution and the peak-position/width comparison, but the headline numbers should either be replaced by a nonthermal metric or backed by explicit frequency intervals and a window-robustness check.\n\nSecond, the two-particle spin response is computed with the NCA impurity solver, and the authors explicitly state that quantitative accuracy would require three-point vertex corrections. That is an honest caveat, but it applies precisely to the quantities the central claim depends on. A cross-check of the M-point results at one or two parameters with a better solver, or at least a demonstration that NCA artifacts do not drive the frozen-peak behavior, would raise confidence.\n\nMinor: Fig. 4 swaps the U labels between caption and text, and Fig. 3 uses A-m notation for temperatures. No code or data is released, which is frustrating for a computational paper.\n\nWho this is for: theorists in nonequilibrium DMFT and experimentalists doing tr-ARPES/Raman/RIXS on cuprates. The predictions are specific enough to test. The paper deserves serious peer review — it is not a desk reject — but the referees should ask for the fit-window specification, robustness checks, and a careful rewrite of the cuprate mapping.","headline":"The paper makes credible, testable predictions for momentum-selective ultrafast dynamics, but the quantitative effective-temperature claims are under-specified and need to be pinned down before publication.","tokens_in":23918,"tokens_out":5825,"would_cite":false,"duration_ms":62164,"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":"The paper predicts that in a photoexcited correlated metal, nodal quasiparticles heat to higher effective temperatures than antinodal ones, while antiferromagnetic fluctuations heat far above the electrons with a nearly frozen correlation…","keywords":["ultrafast dynamics","Hubbard model","momentum-selective response","nodal-antinodal anisotropy","spin fluctuations","local magnetic moments","time-resolved spectroscopy","real-time D-GW"],"falsifier":"Recalculate the same pump-pulse protocol with an impurity solver that is systematically improvable, or measure time-resolved RIXS on a square-lattice antiferromagnet, and check whether the zone-corner magnon peak position shifts with the electronic temperature during the pulse; if it shifts immediately rather than remaining frozen, the claimed quench-like magnetic response fails.","tokens_in":22766,"feed_emoji":"⚡","tokens_out":15626,"duration_ms":142452,"temperature":0.7,"pith_summary":"The paper argues that the momentum-dependent electronic structure of a correlated metal controls how it responds to an ultrafast pump pulse. In the half-filled Hubbard model on a square lattice, it predicts that nodal quasiparticles — the coherent parts of the Fermi surface — absorb energy more efficiently than antinodal, pseudogapped parts, so a transient temperature gap opens across the Fermi surface. It further claims that antiferromagnetic fluctuations at the zone corner are driven far above the electronic temperature while their correlation length stays frozen, making the magnetic response look like a quench rather than thermal heating. The same real-time calculation identifies the nonthermal transfer of spectral weight into the antinodal pseudogap and a slow component of the local spin susceptibility that can be read as a dynamical signature of local-moment formation and its photoinduced melting. If correct, these results place the nodal–antinodal anisotropy seen in time-resolved photoemission and Raman, and the nonthermal magnons seen in time-resolved RIXS, on a common microscopic footing.","feed_headline":"Pump heats nodal quasiparticles faster than antinodal ones","feed_subtitle":"The coherent node absorbs the pump more readily than the pseudogapped antinode, producing a measurable temperature gap.","key_machinery":"The argument is carried by a real-time extension of the dual-GW (D-GW) approach, a nonequilibrium many-body method that solves a local impurity problem nonperturbatively on the L-shaped Kadanoff–Baym contour and then dresses the propagators with a GW-like nonlocal self-energy built from renormalized charge and spin fluctuations. This gives momentum- and frequency-resolved single-particle spectra and two-particle susceptibilities directly in real time. Wigner-transformed spectral functions define effective temperatures and distribution functions for electrons and for magnetic excitations, and the two-component structure of the local spin susceptibility — a fast hopping-dominated piece and a slow superexchange-dominated piece, confirmed by an exact Hubbard-dimer analysis — carries the proposed dynamical criterion for local-moment formation.","core_discovery":"Resonant photoexcitation of the half-filled Hubbard model on a square lattice does not merely heat the system as a whole. In the metallic and pseudogap regimes, the nodal (coherent) parts of the Fermi surface heat faster than the antinodal (pseudogapped) parts, so the quasiparticle effective temperature at the node exceeds that at the antinode until the antinodal pseudogap collapses. At the same time, the antiferromagnetic mode at $M=(\\pi,\\pi)$ reaches an effective temperature well above the electronic one while its spectral peak, and therefore the magnetic correlation length, stays almost unchanged during the pulse; long-wavelength modes at the zone center remain tied to the electrons. After the pulse, magnetic spectral weight cascades from high momenta down to long-wavelength modes. The paper also claims that a slow, interaction-dependent peak in the real-time local spin susceptibility, whose timescale is set by the superexchange energy, provides an experimentally accessible dynamical signature of local-moment formation and its melting under photoexcitation.","pith_inferences":["If the zone-corner magnon stays frozen while heating, then the relaxation bottleneck in gapless antiferromagnets is the electron–magnon coupling at the zone corner, so pump shaping could transiently populate or deplete long-wavelength magnetic modes; the paper does not pursue this control scenario.","The local-moment criterion should be testable in multi-orbital Hund metals, where the same slow susceptibility component should separate itinerant and local-moment spin dynamics; the authors do not make this extension.","Because the method's two-particle sector rests on the non-crossing approximation, a small-cluster exact-diagonalization check of the frozen magnon peak position would be a natural next step; the paper itself notes that quantitative agreement would require three-point vertex corrections.","The mapping of a nonthermal trajectory onto the equilibrium $U$–$T$ phase diagram implies that simple effective-temperature analyses of pump–probe data can overestimate heating, since the transient crossover boundaries sit above the equilibrium ones."],"forward_implications":["Time-resolved photoemission and Raman measurements should see a genuine nodal–antinodal quasiparticle temperature gap whenever the antinodal pseudogap is present, not an experimental artifact.","Transient antinodal in-gap states seen in cuprate pump–probe photoemission can be interpreted as correlation-driven spectral-weight transfer from the Hubbard bands, not simple gap filling.","Two-temperature and three-temperature models of ultrafast magnetism would need a momentum-selective magnetic channel, because the zone-corner antiferromagnetic mode can sit far above the electronic temperature while its correlation length is frozen.","The slow component of the real-time local spin susceptibility gives a pump–probe-accessible signature of local-moment formation and melting, extending previous equilibrium or imaginary-time criteria to nonequilibrium experiments."],"supporting_citations":[{"why":"Introduces and applies the real-time D-GW method that all calculations in this paper use.","marker":"[40]"},{"why":"Establishes the D-GW description of nonlocal correlation effects and the need for vertex corrections beyond DMFT.","marker":"[43]"},{"why":"Provides the experimental observation of transient antinodal in-gap states that the nonthermal spectral-weight transfer is said to explain.","marker":"[21]"},{"why":"Reports time-resolved RIXS paramagnon dynamics with a frozen peak position that the paper's frozen AFM correlation length is compared with.","marker":"[13]"},{"why":"Supplies the ultrafast Raman thermometry evidence that magnetic effective temperatures rise above the electronic one.","marker":"[45]"},{"why":"Provides the tr-ARPES nodal heating data that frame the nodal–antinodal anisotropy question.","marker":"[44]"},{"why":"Gives an ultrafast Raman probe of cuprate transitions used as a comparison for momentum-selective response.","marker":"[15]"},{"why":"Reports a symmetry-resolved Raman measurement finding higher quasiparticle temperature in the nodal channel.","marker":"[16]"},{"why":"Identifies the two-component local spin susceptibility in a semiclassical equilibrium setting that motivates the new dynamical LMM criterion.","marker":"[46]"},{"why":"Provides the real-time non-crossing approximation used to solve the impurity problem in the simulations.","marker":"[52]"}],"fun_headline_variants":["Nodes heat faster than antinodes in photoexcited Hubbard model","Ultrafast pump creates transient nodal-antinodal temperature gap","Antiferromagnetic mode heats above electrons, keeps correlation length","Local-moment melting leaves dynamical signature in spin susceptibility"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire momentum-resolved picture rests on the non-crossing approximation used inside the real-time impurity solver; if that approximation distorts two-particle spin fluctuations in the pseudogap regime, the reported temperature gap, frozen correlation length, and moment-melting signature could be numerical artifacts rather than physics.","fun_headline_variants_meta":{"raw":{"variants":["Nodes heat faster than antinodes in photoexcited Hubbard model","Ultrafast pump creates transient nodal-antinodal temperature gap","Antiferromagnetic mode heats above electrons, keeps correlation length","Local-moment melting leaves dynamical signature in spin susceptibility"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000397,"raw_usage":{"total_tokens":2113,"prompt_tokens":1015,"completion_tokens":1098,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":631,"completion_tokens_details":{"reasoning_tokens":1029}},"tokens_in":631,"tokens_out":1098,"duration_ms":12112,"temperature":1.0,"reasoning_tokens":1029,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T00:24:56.263099+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recalculate the same pump-pulse protocol with an impurity solver that is systematically improvable, or measure time-resolved RIXS on a square-lattice antiferromagnet, and check whether the zone-corner magnon peak position shifts with the electronic temperature during the pulse; if it shifts immediately rather than remaining frozen, the claimed quench-like magnetic response fails.","supporting_citations":[{"cited_title":"Dasari, H","cited_arxiv_id":null,"evidence_quote":"Introduces and applies the real-time D-GW method that all calculations in this paper use."},{"cited_title":"Dasari, H","cited_arxiv_id":null,"evidence_quote":"Establishes the D-GW description of nonlocal correlation effects and the need for vertex corrections beyond DMFT."},{"cited_title":"Cilento, G","cited_arxiv_id":null,"evidence_quote":"Provides the experimental observation of transient antinodal in-gap states that the nonthermal spectral-weight transfer is said to explain."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Reports time-resolved RIXS paramagnon dynamics with a frozen peak position that the paper's frozen AFM correlation length is compared with."},{"cited_title":"Gatuingt and Y","cited_arxiv_id":null,"evidence_quote":"Supplies the ultrafast Raman thermometry evidence that magnetic effective temperatures rise above the electronic one."},{"cited_title":"Parham, H","cited_arxiv_id":null,"evidence_quote":"Provides the tr-ARPES nodal heating data that frame the nodal–antinodal anisotropy question."},{"cited_title":"Gatuingt, A","cited_arxiv_id":null,"evidence_quote":"Gives an ultrafast Raman probe of cuprate transitions used as a comparison for momentum-selective response."},{"cited_title":"Gallais, Tracking photo-induced superconducting to normal state transition in the cuprate Bi2Sr2CaCu2O8, inAdvances in Ultrafast Condensed Phase Physics V, Vol","cited_arxiv_id":null,"evidence_quote":"Reports a symmetry-resolved Raman measurement finding higher quasiparticle temperature in the nodal channel."},{"cited_title":"Sayad, R","cited_arxiv_id":null,"evidence_quote":"Identifies the two-component local spin susceptibility in a semiclassical equilibrium setting that motivates the new dynamical LMM criterion."}],"review_version":1}