{"id":"4c164542-c69a-4e56-a1a9-318bcfe29617","arxiv_id":"1908.04776","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"The dynamical charge susceptibility of the 2D Hubbard model in the pseudogap regime shows weak momentum and temperature dependence and no clear pseudogap signature, based on eight-site dynamical cluster approximation calculations.","lead":"This paper computes how charge fluctuations behave in the Hubbard model, a standard model for high-temperature superconductors, using a numerical approximation called the dynamical cluster approximation. It finds that these fluctuations show little momentum or temperature dependence and no clear connection to the pseudogap, and predicts the result is directly measurable in electron energy-loss spectroscopy.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eight-site DCA cannot resolve longer-period or incommensurate charge fluctuations, so the negative charge-fluctuation conclusion is not yet established at pseudogap temperatures.","rationale":"The reader's weakest-assumption identification is exactly the load-bearing concern: the eight-site DCA cluster has too few independent momenta to rule out charge fluctuations that couple to the pseudogap, and the cluster-size checks reported in the paper are not performed in the relevant low-temperature regime. I agree with the reader that this warrants a CONDITIONAL verdict rather than ACCEPT or REJECT. The paper is otherwise careful: the impurity solver is numerically exact, the charge-conservation Ward identity is satisfied, the analytical-continuation caveat is explicitly stated and shown not to alter the dominant peak, and the momentum dependence is contrasted with the magnetic case. None of these facts removes the need for a larger-cluster check at β=15. I do not find a stronger objection: the doping evolution and the fluctuation diagnostics are internally consistent, and the paper's conclusion is phrased as a statement about the studied parameter range. However, that range is defined by an eight-site cluster, and the negative result is precisely a statement about what is absent at the sampled momenta. The concern is therefore not about consensus or interpretation but about whether the observable studied can resolve the physics the conclusion addresses. A 16-site or larger DCA run at pseudogap temperatures would settle the issue; if the new momenta also show only broad, high-energy, weakly temperature-dependent peaks, the conclusion would be substantially strengthened. Until then, the conditional verdict is appropriate.","tokens_in":11155,"tokens_out":2583,"duration_ms":29758,"concrete_test":"Perform 16-site (or larger) DCA simulations at β=15, δ=-0.05 and -0.11, using a cluster whose reciprocal lattice includes Q=(π/2,0), Q=(π/4,π/4), and other incommensurate vectors absent from the eight-site set. Compute χch(Q,iΩ_n), analytically continue Imχch(Q,Ω), and compare with the eight-site results at the same doping and temperature. If a low-energy peak or a strongly temperature-dependent feature develops at the newly sampled Q while the antinodal spectral function remains pseudogapped, the central negative claim fails. Also report bootstrap error bars on both the eight- and sixteen-site data so the comparison is quantitative.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central claim is a negative one: in the eight-site DCA solution of the 2D Hubbard model, the dynamical charge susceptibility shows no clear pseudogap signature, and charge fluctuations are not a good way to describe pseudogap physics. That conclusion requires that the calculation actually samples the charge fluctuations that could be pseudogap-relevant. Eight-site DCA provides only four independent cluster momenta: Q=(0,0), (π/2,π/2), (π,0), and (π,π). This resolution cannot capture incommensurate or longer-period stripe-like charge fluctuations, whose wave vectors are precisely the ones proposed to be coupled to the pseudogap in the literature cited in the manuscript. The authors explicitly acknowledge this in the conclusion: 'due to the limited momentum resolution, DCA is insensitive to stripes with periods larger than our cluster size.' The supporting 4-site and 16-site checks are described only as performed at 'select points and high temperature,' not in the low-temperature pseudogap region where the negative finding is claimed. The fluctuation-diagnostics argument (Fig. 6) is shown at β=10, whereas the pseudogap regime is entered on cooling toward β=15. The paper's own limitation statement therefore flags a missing cluster-size convergence check at the parameter regime that matters most. This is not an internal inconsistency, and the CT-AUX solver is numerically exact, but the central conclusion is underdetermined because the relevant charge-ordering momenta are not sampled at the relevant temperatures. Without a larger-cluster check, the reported absence of a pseudogap signature in χch could be an artifact of the eight-site cluster's momentum resolution.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents dynamical cluster approximation (DCA) calculations of the charge susceptibility χch(Q,Ω) for the two-dimensional Hubbard model at U/t=7, t'/t=-0.15, using a numerically exact continuous-time auxiliary-field impurity solver on an eight-site cluster. The authors study the momentum, doping, and temperature dependence of the susceptibility and report that it can be described by a single peak at a characteristic frequency, with little momentum and temperature dependence, while doping shifts the peak to lower frequencies. They observe no clear signature of the pseudogap in the charge channel, in contrast to the magnetic channel, and use fluctuation diagnostics to argue that charge fluctuations are not a good way to describe pseudogap physics. The paper also compares with the bare (vertex-free) susceptibility, showing that vertex corrections suppress and flatten the momentum dependence, and suggests the results are relevant for momentum-resolved electron energy-loss spectroscopy.","tokens_in":11511,"tokens_out":6011,"duration_ms":56543,"significance":"If the central claims hold, this paper provides a useful reference calculation of the dynamical charge susceptibility in a regime relevant to cuprate high-temperature superconductors, filling a gap noted by the authors: previous theoretical results were available mostly at high temperature, while the pseudogap regime was unexplored in this observable. The use of an exact impurity solver and the explicit discussion of vertex corrections are strengths, as is the proposal of a direct comparison with M-EELS experiments. The fluctuation-diagnostics analysis also connects two-particle observables to the single-particle self-energy, which is a valuable methodological contribution. The main importance lies in the negative result: if charge fluctuations are not a good descriptor of the pseudogap at accessible momenta, this constrains theoretical scenarios linking charge order and the pseudogap.","major_comments":[{"comment":"The central negative claim—that charge fluctuations show no pseudogap signature—is supported only at the four independent cluster momenta Q=(0,0), (π/2,π/2), (π,0), and (π,π). The authors themselves note in the Conclusion that DCA is insensitive to stripes with periods larger than the cluster size, and the 4-site and 16-site checks mentioned in Section 2 are described only at 'select points and high temperature', not at the low temperatures where the pseudogap develops. Since the pseudogap regime is the regime in which the negative claim is made (e.g., β=15 at δ=-0.05), the cluster-size dependence of the central result is not established. Please extend the 16-site (or at least a systematic 4- vs 8-site) comparison to β=15 for the underdoped and optimally doped cases, or explicitly restrict the main conclusion to the momentum resolution of the eight-site cluster throughout the abstract and conclusion.","section":"Section 2 (paragraph beginning 'Eight-site DCA yields...') and Conclusion"},{"comment":"No statistical error bars are reported for χch(Q,iΩ_n) or for the analytically continued real-frequency spectra. The statements 'little temperature dependence' (Fig. 4) and 'no clear signature of the pseudogap' are based on differences of order 10–20% between β=5 and β=15 (e.g., the low-frequency Matsubara values in Fig. 4). Without error bars or tabulated values with uncertainties, the reader cannot assess whether the observed differences are statistically significant or whether the apparent absence of a pseudogap signature is an artifact of noise. Please provide error bars at least for the static zero-frequency susceptibilities and for the peak positions of the analytically continued spectra, and state the Monte Carlo statistics used for the two-particle Green's functions.","section":"Figs. 2–5 and related text"},{"comment":"The fluctuation diagnostics that motivate the conclusion 'charge fluctuations are not a good way to describe pseudogap physics' are presented only at β=10. According to the text around Fig. 4 and the phase diagram in Fig. 1, at δ=-0.05 the system 'gradually enter[s] the pseudogap regime' on cooling from β=5 to β=15, so β=10 may be on the boundary of, rather than fully inside, the pseudogap regime. To support the conclusion in the pseudogap region, please show the same diagnostics at β=15 for an underdoped doping (or explicitly state that β=10 is inside the pseudogap phase and justify the representativeness). In addition, the pie charts in Fig. 6 sum only the first 10 Matsubara frequencies; please justify that higher-frequency contributions do not change the qualitative conclusion.","section":"Fig. 6 and Eqs. (6)–(7)"}],"minor_comments":[{"comment":"Typographical inconsistencies: 'Figure. 2', 'Figure. 4', and similar should be 'Fig. 2', 'Fig. 4', etc.; 'Pi chart' in the Fig. 6 caption should be 'Pie chart'.","section":"Throughout"},{"comment":"The sentence 'A Ward identity requires the frequency dependence to be identically zero in systems that conserve total charge' is unclear. What vanishes for Ω≠0 at Q=(0,0) is χch(Q=0,iΩ_n), not its frequency dependence in general; please rephrase to state the Q=(0,0) result explicitly.","section":"Page 3, left column (Ward identity sentence)"},{"comment":"The authors state the results 'should be directly measurable' in M-EELS experiments, but M-EELS typically measures the loss function Im[-1/ε], which involves the dielectric function rather than the bare charge susceptibility. A brief statement relating χch to the loss function or to the dielectric function would make the proposed comparison more concrete.","section":"Abstract and closing paragraph"},{"comment":"The sentence 'approximately δ → −δ for t' → −t''' is not explained; consider adding a one-sentence clarification of the approximate particle-hole relation used here.","section":"Page 4, bottom (t' dependence)"},{"comment":"It is not clear whether the magnetic contributions shown in Fig. 6 (lower panels) are new results from this paper or taken from reference 29; please clarify the source of the spin-channel data.","section":"Fig. 6 caption and text"}],"recommendation":"major_revision","confidential_remarks":"The authors are transparent about the momentum-resolution limitation of DCA, which is commendable, but the abstract and the final sentence of the conclusion are stronger than the evidence presented: the cluster-size check at pseudogap temperatures is missing, and the absence of error bars makes the 'little temperature dependence' claim difficult to evaluate. These issues are fixable within the manuscript's scope, so I recommend major revision rather than rejection. The paper is within the scope of the journal and is likely to be of interest to the strongly correlated electron community."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take.\n\nThe genuinely new thing is that this paper computes the dynamical charge susceptibility in the pseudogap and near-superconducting regime of the 2D Hubbard model, a regime explicitly not covered by the earlier high-temperature DCA work of Kung et al. It gives the doping, temperature, and momentum dependence of χch(Q, Ω), including vertex corrections, and it presents fluctuation diagnostics that separate charge and spin contributions to the self-energy. The data should be directly useful for M-EELS and RIXS people who need a theory curve to compare against.\n\nWhat the paper does well, concretely. The CT-AUX impurity solver is numerically exact for the cluster problem; the DCA self-consistency is the only approximation. The calculation respects the Ward identity at Γ, which many diagrammatic methods get wrong. The vertex corrections are shown to be huge—they kill the (π,π) enhancement and flatten the momentum dependence. The fluctuation diagnostics (Fig. 6) are a nice addition: they show the pseudogap is dominated by short-range magnetic fluctuations, and that charge contributions come in with comparable weight from all momenta. That is a meaningful statement, not a null result in the usual sense.\n\nWhere I would push back. The central negative claim—'charge fluctuations are not a good way to describe pseudogap physics in the entire parameter range studied here'—is stronger than what the momentum resolution supports. Eight-site DCA has four independent bosonic momenta: Γ, (π/2,π/2), (π,0), and (π,π). It cannot see incommensurate or long-period stripe fluctuations, and the paper says so. The 4-site and 16-site checks are only described as 'select points and high temperature,' not at β = 10–15 where the pseudogap actually opens. So the no-signature conclusion is established for the resolved momenta, but not for the stripe-like fluctuations that the literature often couples to the pseudogap. That is a real soft spot, though the authors flag it themselves.\n\nTwo smaller things. There are no statistical error bars on any of the plotted susceptibilities. For a numerical paper whose main message is about the absence of a feature, error bars matter. And the analytical continuation is acknowledged as uncontrolled; I am fine with their caution, but the real-frequency peak positions should be read as qualitative. Minor, not fatal.\n\nIs this a serious paper? Yes. It is careful, honest, and fills an actual gap. I would send it to a referee, with the expectation that the referee asks for error bars and a low-temperature cluster-size check if it is doable, and asks the authors to soften the 'entire parameter range' phrasing. The math and citation pattern look solid; plenty of self-citations, but they are the relevant method and phase-diagram papers, and the central result is not fitted to anything.\n\nIn short: a good, citable calculation with a caveat in proportion to its main claim.","headline":"First DCA calculation of the dynamical charge susceptibility in the Hubbard model's pseudogap regime; clean negative result for an eight-site cluster, but cluster-size convergence at low temperature is the key caveat.","tokens_in":11976,"tokens_out":2866,"would_cite":true,"duration_ms":28727,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["71.10.Fd","71.27.+a","74.72.-h"],"model":"deepseek-v4-flash","headline":"Charge fluctuations show no signature of the pseudogap in the eight-site Hubbard model, the authors argue, with the dynamical charge susceptibility reducing to a single featureless peak.","keywords":["Hubbard model","dynamical charge susceptibility","pseudogap","dynamical cluster approximation","cuprates","charge fluctuations","fluctuation diagnostics","M-EELS"],"falsifier":"Compute the dynamical charge susceptibility $\\chi_{ch}(Q,\\Omega)$ on a 16-site or larger DCA cluster at the same $U/t=7$, $t'/t=-0.15$ parameters and at low temperatures ($\\beta \\gtrsim 10$) inside the pseudogap regime; if a significant momentum-dependent low-frequency peak or a $Q=(\\pi/2,\\pi/2)$ or commensurate charge response emerges that is absent on the eight-site cluster, the negative finding is a cluster-size artifact. A corresponding experiment would be to measure M-EELS on an underdoped cuprate near $\\delta \\approx 0.1$ and look for a sharp low-energy charge mode at a nonzero momentum that the single-peak prediction of this paper would miss.","tokens_in":10921,"feed_emoji":"🧲","tokens_out":1864,"duration_ms":18764,"temperature":0.7,"pith_summary":"This paper asks whether charge fluctuations can explain the pseudogap in the two-dimensional Hubbard model, the standard theoretical stand-in for the cuprate high-temperature superconductors. Using an eight-site dynamical cluster approximation, the authors compute the dynamical charge susceptibility across momenta, dopings, and temperatures, including the pseudogap regime. They find that the susceptibility is featureless: a single peak at a characteristic frequency with almost no momentum or temperature dependence and only a mild doping dependence. They conclude that charge fluctuations are not a good way to describe pseudogap physics in the parameter range studied, and that short-ranged antiferromagnetic spin fluctuations, not charge modes, carry the self-energy changes associated with the pseudogap.","feed_headline":"Charge fluctuations show no pseudogap signature in Hubbard model","feed_subtitle":"In eight-site DCA, the charge susceptibility is one featureless peak; spin fluctuations carry the pseudogap instead.","key_machinery":"The central object is the dynamical charge susceptibility $\\chi_{ch}(Q,\\Omega)$, defined from the two-particle Green's function in the density channel and obtained from the eight-site dynamical cluster approximation (DCA) with a numerically exact continuous-time auxiliary-field impurity solver. The DCA coarse-grains momentum space into cluster momenta, giving four independent momentum transfers: $Q=(0,0)$, $(\\pi/2,\\pi/2)$, $(\\pi,0)$, and $(\\pi,\\pi)$. The paper also uses fluctuation diagnostics, which expresses the single-particle self-energy in terms of two-particle quantities through the exact equation of motion, separating the self-energy into charge and magnetic fluctuation contributions; this is the machinery that lets the authors attribute the pseudogap to spin rather than charge fluctuations.","core_discovery":"The central claim is that in the eight-site dynamical cluster approximation of the two-dimensional Hubbard model at U/t = 7 and t'/t = -0.15, the dynamical charge susceptibility chi_ch(Q, $\\Omega$) shows no clear signature of the pseudogap. As a function of frequency it is well represented by a single peak at a characteristic frequency; the peak position and weight show little momentum or temperature dependence, while doping shifts the peak to lower frequencies and sharpens it. The static uniform susceptibility rises with doping and is not suppressed upon entering the pseudogap, in contrast to the magnetic susceptibility. Vertex corrections are essential: without them the bare susceptibility has a dominant (pi,pi) contribution, whereas the full vertex-suppressed result shows comparable weight at all momenta. Fluctuation diagnostics decompose the self-energy into charge and spin channels; the pseudogap is well described by short-ranged Q = (pi,pi) magnetic fluctuations, while charge modes contribute from all momenta and many frequencies with comparable strength. The authors therefore state that charge fluctuations are not a good way to describe pseudogap physics in the entire parameter range studied here.","pith_inferences":["The paper's negative finding is a constraint on cluster size: a larger cluster that resolves stripe-like charge order with periods longer than eight sites could reveal charge fluctuations tied to the pseudogap that this calculation cannot see, so the claim is only as strong as the cluster momentum resolution.","A testable extension would be to compute the same quantities on larger clusters (e.g., 16- or 32-site DCA) at the low temperatures where the pseudogap opens, checking whether the single-peak structure and the absence of momentum dependence survive when longer-wavelength charge modulations are resolvable.","The similar magnitude of $\\chi_{ch}$ at $(\\pi,0)$ and $(\\pi/2,\\pi/2)$ suggests an approximate isotropy of charge fluctuations in momentum space that could be a generic feature of the doped Mott insulator, but confirming this would require systematic study at other $U$ and $t'$ values.","The authors' conclusion that the doping evolution of $\\chi_{ch}$ is not caused by the pseudogap implies that the doping-dependent peak shift could instead track the single-particle bandwidth renormalization or the Mott scale, an attribution they do not make explicitly."],"forward_implications":["If the central claim is correct, the pseudogap in the cuprate-relevant Hubbard model should not be interpreted as a charge-fluctuation phenomenon, and theories that build the pseudogap from charge order or charge modes would need to explain why the eight-site susceptibility shows no accompanying charge signature.","The computed $\\chi_{ch}(Q,\\Omega)$ provides a direct target for momentum-resolved electron energy-loss spectroscopy (M-EELS) and resonant inelastic X-ray scattering (RIXS) on cuprates, and a mismatch between those experiments and the single-peak prediction would indicate physics beyond the one-band Hubbard model at these parameters.","Because vertex corrections are essential in suppressing the $(\\pi,\\pi)$ charge response, low-order diagrammatic approximations that neglect vertex corrections will overestimate charge fluctuations and their role in the self-energy; conserving approximations are required for quantitative statements.","The fluctuation diagnostics result implies that any description of the pseudogap in terms of bosonic modes should be dominated by short-ranged antiferromagnetic fluctuations, and that charge-mode descriptions would require summing contributions from all momenta and a broad frequency range, making them impractical.","The near temperature independence of $\\chi_{ch}$ across the pseudogap crossover means that charge susceptibility measurements alone are unlikely to serve as a pseudogap thermometer, unlike the magnetic susceptibility."],"supporting_citations":[{"why":"Defines the dynamical cluster approximation as the momentum-space self-energy approximation with full frequency dependence, including the coarse-graining and the lattice susceptibility inversion used in Eq. (5).","marker":"48"},{"why":"Supplies the continuous-time auxiliary-field quantum impurity solver used as the numerically exact impurity solver in the DCA self-consistency.","marker":"63"},{"why":"Establishes the eight-site DCA phase diagram with the pseudogap metal regime at these parameters, the region the paper probes.","marker":"53"},{"why":"Provides the Ward identity guaranteeing that the DCA charge susceptibility conserves total charge, so the zero at $Q=(0,0)$ and the frequency structure are exact constraints on the data.","marker":"69"},{"why":"Introduces fluctuation diagnostics, the equation-of-motion decomposition of the self-energy into two-particle fluctuations that the paper uses to attribute the pseudogap to spin rather than charge modes.","marker":"74"},{"why":"Supplies the magnetic susceptibility results used for comparison and the magnetic-channel fluctuation diagnostics shown in Figure 6.","marker":"29"},{"why":"Provides the high-temperature doping evolution of the dynamical charge susceptibility that the paper extends into the pseudogap regime.","marker":"46"},{"why":"Momentum-resolved electron energy-loss spectroscopy reference establishing the experimental technique that the paper says could directly measure the computed susceptibilities.","marker":"7"}],"fun_headline_variants":["Charge susceptibility: one peak, no pseudogap","No pseudogap signature in Hubbard charge fluctuations","Charge channel featureless while spin carries pseudogap","Charge susceptibility flat without pseudogap in DCA","Hubbard model: charge susceptibility ignores pseudogap"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion rests on the assumption that the eight-site DCA cluster gives enough momentum resolution to see the charge fluctuations that matter for the pseudogap, an assumption the authors themselves qualify by noting that the method is insensitive to stripes with periods larger than the cluster.","fun_headline_variants_meta":{"raw":{"variants":["Charge susceptibility: one peak, no pseudogap","No pseudogap signature in Hubbard charge fluctuations","Charge channel featureless while spin carries pseudogap","Charge susceptibility flat without pseudogap in DCA","Hubbard model: charge susceptibility ignores pseudogap"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00024,"raw_usage":{"total_tokens":1483,"prompt_tokens":873,"completion_tokens":610,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":489,"completion_tokens_details":{"reasoning_tokens":534}},"tokens_in":489,"tokens_out":610,"duration_ms":6037,"temperature":1.0,"reasoning_tokens":534,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:32:28.462605+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute the dynamical charge susceptibility $\\chi_{ch}(Q,\\Omega)$ on a 16-site or larger DCA cluster at the same $U/t=7$, $t'/t=-0.15$ parameters and at low temperatures ($\\beta \\gtrsim 10$) inside the pseudogap regime; if a significant momentum-dependent low-frequency peak or a $Q=(\\pi/2,\\pi/2)$ or commensurate charge response emerges that is absent on the eight-site cluster, the negative finding is a cluster-size artifact. A corresponding experiment would be to measure M-EELS on an underdoped cuprate near $\\delta \\approx 0.1$ and look for a sharp low-energy charge mode at a nonzero momentum that the single-peak prediction of this paper would miss.","supporting_citations":[{"cited_title":"Gull , author P","cited_arxiv_id":null,"evidence_quote":"Supplies the continuous-time auxiliary-field quantum impurity solver used as the numerically exact impurity solver in the DCA self-consistency."},{"cited_title":"Hafermann , author E","cited_arxiv_id":null,"evidence_quote":"Provides the Ward identity guaranteeing that the DCA charge susceptibility conserves total charge, so the zero at $Q=(0,0)$ and the frequency structure are exact constraints on the data."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the magnetic susceptibility results used for comparison and the magnetic-channel fluctuation diagnostics shown in Figure 6."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the high-temperature doping evolution of the dynamical charge susceptibility that the paper extends into the pseudogap regime."}],"review_version":1}