{"id":"ecba77a9-4c87-4150-bcdb-cbff5e28bda5","arxiv_id":"2505.19852","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A dynamical exchange-correlation field, defined as the Coulomb potential of a 'dynamical xc hole,' replaces the self-energy in the Green function equation of motion, and cluster-extrapolated approximations reproduce benchmark spectra of the Hubbard, Heisenberg, and Anderson models.","lead":"A condensed-matter theorist proposes replacing the self-energy, the standard but heavy tool for computing electron spectra, with a local 'exchange-correlation field' built from a dynamical hole in the electron density. The paper argues this field obeys helpful sum rules and demonstrates the recipe on model chains, an impurity model, and the electron gas.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The practical claim rests on cluster-extrapolated Vxc, which is uncontrolled: the paper's own Sec. XVIII A admits nearest-neighbour truncation degrades small-k Hubbard spectra, yet no convergence test in cluster size is provided.","rationale":"The reader's weakest_assumption identifies the cluster-extrapolation premise as the central vulnerability, and my reading agrees: the formal part is a tautological but sound rewriting, whereas every demonstrated application relies on transferring Vxc from a small cluster to the infinite system with fit parameters derived from known benchmark boundaries. The paper is honest about the nearest-neighbour limitation in the Hubbard case, but it does not quantify convergence, and the Heisenberg and Anderson sections fit several parameters (ωsp, B, the SIAM ansatz coefficients) to target data. This does not invalidate the exact formalism, but it makes the practical claim 'simple but accurate approximations' depend on an untested assumption of short-rangedness. The most decisive check would be a systematic cluster-size convergence study, which is currently absent. Since the reader already assigned CONDITIONAL on this basis, no change in verdict is needed; my stress-test confirms that the conditional verdict is the right one. The independent reproductions against Bethe ansatz, DMRG, and NRG are real evidence that the approximations are not vacuous, but they cannot substitute for a direct test of the transferability assumption because the fitted parameters already encode the target spectral features.","tokens_in":53770,"tokens_out":2500,"duration_ms":32275,"concrete_test":"Compute the exact Vxc from exact diagonalization of 1D Hubbard clusters with N=2, 4, 6, 8, 10 at a fixed U/Δ (e.g., U/Δ=4), and extract the site-off-diagonal components V_{i,i+d}(t) as a function of distance d. Check whether max_t |V_{i,i+d}(t)| / max_t |V_{i,i+1}(t)| decays with d and stabilizes as N grows; if the d=2 component is non-negligible or grows with N, the dimer/nearest-neighbour extrapolation is uncontrolled. Separately, generate the spectral function from the N=6 and N=8 cluster extrapolations and compare both to the DMRG benchmark of Benthien and Jeckelmann; agreement that improves monotonically with N would support the extrapolation, while non-monotonic or stationary disagreement would falsify it.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The exact rewriting of the equation of motion (Eqs. 30, 40, 43, 44, 46) appears internally consistent: Vxc is defined as the Coulomb potential of the dynamical xc hole, and the sum rule and on-top constraint follow from operator identities. The load-bearing step is the transfer of Vxc from a small cluster to the infinite system. In Sec. XVIII A, the dimer Vxc is used for the 1D Hubbard chain and the paper explicitly states it 'clearly neglects components of Vxc beyond nearest neighbours', attributing the small-k discrepancy with DMRG to that neglect. A six-site cluster improves agreement, but no evidence is given that the Vxc matrix elements decay with distance or that the N=2,4,6 results are converging. In Sec. XVIII B, the Heisenberg ansatz VD(k,t)=A(k)e^{-iωsp(k)t}+B(k) (Eq. 424) uses ωsp fitted to the known two-spinon boundary (Eq. 428) and B adjusted by hand for finite-size effects, so the agreement with DMRG and KCuF3 is partly a fit to the target spectrum. In Sec. XVIII C, the SIAM Vxc ansatz (Eq. 464) is fitted to a 50-site cluster and then benchmarked against NRG; this tests the analytic form, not the transferability of an uncontrolled approximation. None of these checks isolates the central assumption: that a short-range, cluster-derived Vxc converges to the exact Vxc of the infinite system. If Vxc has long-range components, the practical program loses its foundation even though the formal framework is exact.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript develops \"density-functional Green function theory\", in which the one-particle Green function obeys the local-in-time equation of motion [i∂t − h(r) − Vxc(r,r′;t)]G(r,r′;t)=δ(r−r′)δ(t) (Eq. 44), with Vxc defined as the instantaneous Coulomb potential of a dynamical exchange-correlation hole ρxc (Eq. 46). The hole is shown to satisfy the sum rule (Eq. 40) and the on-top constraint (Eq. 43), both derived from exact operator identities. The paper presents analytic and numerical Vxc for the hydrogen atom, the Holstein model, the Hubbard dimer, and the homogeneous electron gas within RPA/plasmon-pole approximations; derives quasiparticle equations, a total-energy formula, and thermal/nonequilibrium extensions; and applies cluster-extrapolated Vxc to the 1D Hubbard chain, the 1D antiferromagnetic Heisenberg chain, and the single-impurity Anderson model, comparing against Bethe ansatz, DMRG, NRG, and KCuF3/Na data. The formal part is internally consistent, but the load-bearing practical step is the assumption that a small-cluster Vxc can be extrapolated to infinite systems; this assumption is stated in Sec. XVIII but never validated by a systematic convergence study.","tokens_in":54118,"tokens_out":6513,"duration_ms":78790,"significance":"If established, the formalism would be conceptually attractive and computationally useful: local-in-time propagation avoids the convolution in the Dyson equation, and the xc-hole picture provides a natural link between Green function theory and DFT. The formal identities—Eqs. (40), (43), (44), and (46)—are correct and are derived cleanly. The Hubbard gap αU is a genuine parameter-free prediction and agrees remarkably well with the Bethe ansatz over a wide range of U (Fig. 23); the DMRG, NRG, KCuF3, and Na comparisons are real and nontrivial. However, the applications rest on uncontrolled cluster-to-lattice extrapolations and on several fitted parameters (e.g., γ, Z, η1 in Sec. XVII E; A, B, ωsp in Sec. XVIII B; λ, ω1, C in Sec. XVIII C). The central practical claim—that a short-range, cluster-derived Vxc can replace the self-energy without significant loss—is therefore not yet established. The exactness of the rewriting does not by itself validate any approximation scheme.","major_comments":[{"comment":"The central practical claim that a cluster-computed Vxc transfers to the infinite lattice is not supported by any convergence test. The dimer Vxc is truncated at nearest-neighbour level; the paper explicitly states it \"clearly neglects components of Vxc beyond nearest neighbours\" and attributes the small-k disagreement with DMRG to this neglect. The six-site calculation (Fig. 24) improves agreement, but no N=2,4,6 convergence sequence, no data on the distance decay of Vxc matrix elements, and no error estimate are provided. Since all applications in this section inherit this transferability assumption, the benchmark agreement, including the parameter-free gap αU, cannot be taken as evidence that the infinite-system Vxc is sufficiently short-ranged.","section":"XVIII A, Eqs. (361)-(370)"},{"comment":"The Heisenberg ansatz VD(k,t)=A(k)e^{-iωsp(k)t}+B(k) is not derived from the cluster calculation: ωsp(k) is fitted to the exact two-spinon boundary (Eq. 428) and B(k) is adjusted by hand to remove finite-size effects. This puts part of the target spectrum into the input, so the agreement with DMRG and KCuF3 is partly a consistency check of the ansatz rather than a predictive test of Vxc transferability. A protocol that determines A, B, and ωsp from cluster data alone, together with a sensitivity analysis, is needed before the comparison can be interpreted as support for the extrapolation premise.","section":"XVIII B, Eqs. (424)-(428)"},{"comment":"The SIAM ansatz Vxc(t)=[λ(ω1+C)+(1−λ)Ce^{iω1t}]/[λ+(1−λ)e^{iω1t}] has its parameters determined from the same 50-site cluster whose physics it is then used to represent; benchmarking against NRG validates the functional form in the wide-band limit but does not test whether a Vxc computed on a small cluster converges to the infinite-system Vxc. No cluster-size study is reported for this model, and the wide-band-limit parameters are effectively fitted, so Fig. 31 does not isolate the transferability assumption.","section":"XVIII C, Eq. (464)"},{"comment":"The HEG-based LDA is not parameter-free: the unoccupied-momentum cutoff is set to about 1.5kF to reproduce the static correlation hole (Sec. XVII C), and γ(rs), Z(rs), and η1(rs) are extracted from GW calculations or treated as fitting parameters (Eqs. 339, 346, 347). The Na photoemission comparison therefore demonstrates that a GW-calibrated model can reproduce known spectra, but it does not establish a density-functional approximation for inhomogeneous systems; no inhomogeneous test of the LDA prescription in Eq. (100) is reported.","section":"XVII C-E, Eqs. (302)-(348)"}],"minor_comments":[{"comment":"The bosonic spectral representation of the dynamic part of Vxc is introduced as a conjecture; since it may be used in later constructions, it should be labelled as an assumption and its range of validity stated.","section":"V B, Eq. (56)"},{"comment":"The renormalized weights Ae0(q) and Ah0(q) are allowed to deviate from unity, but the paper does not state whether the resulting spectral functions still satisfy the relevant sum rules; a brief check would be helpful.","section":"XVIII A, Eqs. (371)-(377)"},{"comment":"The notation g(r,r′,r′′;t) is introduced as a correlator, but its physical meaning and the regime in which it can be complex are not discussed until much later; a short comment near Eq. (30) would improve readability.","section":"IV, Eqs. (30)-(31)"},{"comment":"The validity condition for the q-independent plasmon-pole approximation (q≤qc with qc values given for rs=3,4,5) is stated without derivation or reference; a citation or one-line justification is needed.","section":"XVII C, Eqs. (302)-(303)"},{"comment":"The definition of Vxc(k,t) as a two-index object is stated without explaining how the four-index object of Eq. (49) reduces to it in the translationally invariant chain; an explicit reduction would prevent confusion.","section":"XVIII B, Eq. (416)"},{"comment":"The manuscript is a synthesis of the author's previous publications (Refs. [35,44,45,58,66,84,86]); a short paragraph at the start of Sec. XIX summarizing what is new relative to those works would help readers place the contributions.","section":"Various"}],"recommendation":"major_revision","confidential_remarks":"The formal identities and parameter-free Hubbard gap are real and worth publishing. My main concern is the gap between the exact-by-construction rewriting and the practical extrapolations: the paper's own acknowledgements of missing long-range components and hand-adjusted parameters show that the benchmark agreements are not yet evidence for the central transferability claim. I would encourage the editor to require either a systematic cluster-size convergence study or a clear downgrading of the practical claims to \"illustrative ansatze\". The manuscript also appears to be largely assembled from prior papers; a novelty assessment may be relevant for scope, though I did not base my recommendation on it."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a review-style consolidation of the author's earlier Vxc papers, not a new result. The formal core — define a correlator from the two-particle Green function, subtract the density to get a dynamical xc hole, and let its Coulomb field Vxc reproduce the exact Green function — is an exact rewriting, not new physics. The sum rule and on-top constraint follow cleanly from operator identities, and I do not find a problem with the algebra where it matters. If the paper were only that, it would be a useful but modest formal note.\n\nWhat earns credit is the unified presentation and the benchmarking. The Hubbard dimer Vxc is analytic and parameter-free for the gap, the αU gap tracks the Bethe ansatz, and switching from the dimer to a six-site cluster visibly improves the k-resolved spectra against DMRG. The Heisenberg, SIAM, HEG and Na sections are honestly labeled as parametrized or cluster-extrapolated, and the comparisons to DMRG, NRG, neutron scattering and photoemission are real.\n\nThe soft spot is exactly what the stress test flags. The practical program stands or falls on whether a small-cluster Vxc transfers to the infinite system, and the paper never gives a convergence test in cluster size. Section XVIII A admits the dimer Vxc \"clearly neglects components beyond nearest neighbours,\" and the small-k spectra degrade as a result; the six-site cluster does better, but there is no evidence that the Vxc matrix elements decay with distance or that N = 2, 4, 6 is converging. The Heisenberg ansatz fits ω_sp to the known two-spinon boundary and adjusts B by hand; the SIAM ansatz is fitted to a 50-site cluster; the HEG parameters come from GW. So the benchmarks demonstrate internal consistency of the ansatze, not the central transferability claim. None of this undermines the exact formal rewriting, but it does mean the paper does not yet make the case that Vxc is a practical replacement for self-energy. No code or data is shipped, which matters here because the cluster extrapolations are hard to reproduce from the text alone.\n\nWho is this for? People who want a one-stop overview of the Vxc programme and experts who can test the cluster-convergence question. It deserves a serious referee: the formal core is coherent, the benchmarking is honest, and the underlying question — whether a local-in-space-time xc field can approximate the self-energy — is worth referee time. I would not cite it in my own work as a new result; it is a consolidation with thin new material.","headline":"A well-written consolidation of the author's own Vxc program; the formal rewriting is sound and the benchmarks are honestly presented, but the practical transferability claim rests on uncontrolled cluster extrapolations.","tokens_in":54835,"tokens_out":2564,"would_cite":false,"duration_ms":32084,"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":"This paper argues that the one-particle Green function can be obtained from a local dynamical exchange-correlation field instead of the nonlocal self-energy, with simple approximations already matching benchmark spectra for strongly…","keywords":["Green function","dynamical exchange-correlation field","dynamical xc hole","self-energy","strongly correlated electrons","Hubbard chain","Anderson impurity model","homogeneous electron gas"],"falsifier":"Calculate the exact $V_{\\rm xc}$ for the one-dimensional Hubbard chain at $U=4$ using a large reference calculation and compare with the dimer- and six-site-extrapolated forms: if the difference at small $k$ and finite $t$ does not shrink as the cluster grows, the extrapolation premise fails.","tokens_in":53312,"feed_emoji":"⚛️","tokens_out":8696,"duration_ms":78037,"temperature":0.7,"pith_summary":"The paper argues that the one-particle Green function of an interacting electron system can be found from a local-in-time equation of motion that contains a dynamical exchange-correlation field, $V_{\\rm xc}(r,r';t)$, instead of the traditional nonlocal self-energy. $V_{\\rm xc}$ is the Coulomb potential of a dynamical exchange-correlation hole $\\rho_{\\rm xc}$, which obeys a sum rule and an exact on-top constraint inherited from density-functional theory. The paper presents this replacement as exact and shows that simple approximations for $\\rho_{\\rm xc}$ -- extrapolated from small clusters or parametrized from the homogeneous electron gas -- reproduce benchmark spectra for the one-dimensional Hubbard and Heisenberg chains, the single-impurity Anderson model, and sodium photoemission. If the claim holds, spectral calculations that currently require expensive self-energy construction could be replaced by a local-in-time propagation problem.","feed_headline":"Green-function theory swaps self-energy for a local field","feed_subtitle":"A dynamical xc hole reproduces Hubbard, Heisenberg, Anderson, and Na spectra.","key_machinery":"The central object is the dynamical exchange-correlation hole $\\rho_{\\rm xc}(r,r',r'';t)$, defined by writing the two-particle Green function as $G^{(2)}(r,r',r'';t)=[\\rho(r'')+\\rho_{\\rm xc}(r,r',r'';t)]G(r,r';t)$. Its Coulomb potential is $V_{\\rm xc}$; the sum rule and on-top constraint make $V_{\\rm xc}$ act like a local density-dependent field, and only the spherical average of $\\rho_{\\rm xc}$ and its first radial moment enter $V_{\\rm xc}$. The paper also uses a quasiparticle-effective-field decomposition $\\Xi_q(r,t)\\approx\\Xi^S_q(r)+\\Xi^D_q(r)e^{i\\Omega t}$ to turn the equation of motion into a quasiparticle picture of a static energy shift plus a dynamic satellite-generating term.","core_discovery":"The central claim is that there exists an exact local-in-time equation of motion for the time-ordered Green function, $[i\\partial_t - h(r) - V_{\\rm xc}(r,r';t)]G(r,r';t)=\\delta(r-r')\\delta(t)$, where $V_{\\rm xc}$ is generated by the dynamical exchange-correlation hole through $V_{\\rm xc}(r,r';t)=\\int dr''\\, v(r-r'')\\rho_{\\rm xc}(r,r',r'';t)$. The hole satisfies $\\int d^3r''\\, \\rho_{\\rm xc}(r,r',r'';t)=-\\delta_{\\sigma\\sigma''}\\theta(-t)$ and $\\rho_{\\rm xc}(r,r',r;t)=-\\rho(r)$, the same normalization and on-top conditions as the static exchange-correlation hole used in ground-state density-functional theory. The paper reports that approximate $V_{\\rm xc}$ built from the Hubbard dimer and a six-site cluster produces the spinon-holon structure and the $\\alpha U$ gap of the half-filled Hubbard chain; a four-site spin cluster extrapolated with a two-spinon energy reproduces the Heisenberg dynamic structure factor; an ansatz matched to a 50-site Anderson cluster reproduces the Hubbard side bands and Kondo resonance; and an electron-gas parametrization reproduces sodium photoemission with plasmon satellites.","pith_inferences":["An extension the paper leaves implicit is that diagrammatic intuition could be redirected from self-energy expansions to constructing $\\rho_{\\rm xc}$ directly, for example from density-response data.","The cluster-transfer success suggests a nearsightedness test: computing $V_{\\rm xc}$ for systematically larger clusters and checking convergence would turn the practical claim into a quantitative statement about correlation length.","The electron-gas local-density model could be probed on a second simple metal such as aluminum; the paper shows the sodium spectrum but does not test whether the same parametrization transfers.","Since only the spherical average and first radial moment of the xc hole determine $V_{\\rm xc}$, approximate holes with incorrect higher moments may still yield accurate spectra, which could be exploited in data-driven constructions of $V_{\\rm xc}$."],"forward_implications":["Spectra for photoemission and inverse photoemission could be computed by propagating the Green function pointwise in time, because $V_{\\rm xc}$ is local in time and no self-energy convolution is needed.","Density-functional-style approximations such as a local-density approximation built from the homogeneous electron gas become available for one-particle excitation spectra, not just ground-state energies.","For strongly correlated models, a $V_{\\rm xc}$ parametrized from small clusters can capture correlation gaps, spinon continua, and Kondo features at much lower cost than exact diagonalization or impurity solvers.","The static-plus-dynamic form of the effective quasiparticle field gives a systematic language for describing satellites: the dynamic term couples electrons to the main collective mode and transfers spectral weight to a replica peak.","Because hole and electron see different $V_{\\rm xc}$, the gap underestimation of ground-state density-functional theory is explained without invoking a derivative discontinuity."],"supporting_citations":[{"why":"Foundational paper proposing the time-dependent exchange-correlation potential that this review develops.","marker":"[35]"},{"why":"Derives the Hubbard-dimer Vxc and applies it to the 1D Hubbard chain, providing the gap and spectral comparisons.","marker":"[44]"},{"why":"Supplies the RPA calculation of the electron-gas dynamical xc hole and potential used for the LDA-style parametrization.","marker":"[45]"},{"why":"Extends the formalism to spin systems and provides the four-site Heisenberg Vxc and DMRG comparisons.","marker":"[84]"},{"why":"Gives the Anderson impurity Vxc ansatz and the NRG benchmark spectra reproduced here.","marker":"[86]"},{"why":"Provides the DMRG benchmark spectra for the 1D Hubbard chain used to assess the cluster-extrapolated Vxc.","marker":"[79]"},{"why":"Gives the exact gap of the 1D Hubbard model against which the calculated alpha-U gap is compared.","marker":"[81]"},{"why":"Supplies the inelastic neutron scattering data for KCuF3 used to compare the Heisenberg dynamic structure factor.","marker":"[85]"},{"why":"Provides the measured sodium photoemission spectrum reproduced by the electron-gas model.","marker":"[76]"},{"why":"Gives the cumulant-expansion spectral functions for Na and Al used as a reference for plasmon satellites.","marker":"[73]"}],"fun_headline_variants":["Local xc field replaces self-energy in Green functions","Dynamical xc hole makes Green function theory local","Self-energy swapped for a local field in many-electron theory","Exact local Green function equation via dynamical xc hole","Green function theory goes local with a dynamical xc field"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The practical results stand on the assumption that a small cluster's exchange-correlation field, with a few fitted parameters, accurately represents the infinite system's field.","fun_headline_variants_meta":{"raw":{"variants":["Local xc field replaces self-energy in Green functions","Dynamical xc hole makes Green function theory local","Self-energy swapped for a local field in many-electron theory","Exact local Green function equation via dynamical xc hole","Green function theory goes local with a dynamical xc field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000311,"raw_usage":{"total_tokens":1835,"prompt_tokens":1075,"completion_tokens":760,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":691,"completion_tokens_details":{"reasoning_tokens":680}},"tokens_in":691,"tokens_out":760,"duration_ms":6727,"temperature":1.0,"reasoning_tokens":680,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:08:00.333207+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Calculate the exact $V_{\\rm xc}$ for the one-dimensional Hubbard chain at $U=4$ using a large reference calculation and compare with the dimer- and six-site-extrapolated forms: if the difference at small $k$ and finite $t$ does not shrink as the cluster grows, the extrapolation premise fails.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Foundational paper proposing the time-dependent exchange-correlation potential that this review develops."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Derives the Hubbard-dimer Vxc and applies it to the 1D Hubbard chain, providing the gap and spectral comparisons."},{"cited_title":"Jacob and S","cited_arxiv_id":null,"evidence_quote":"Supplies the RPA calculation of the electron-gas dynamical xc hole and potential used for the LDA-style parametrization."},{"cited_title":"Wang and J","cited_arxiv_id":null,"evidence_quote":"Extends the formalism to spin systems and provides the four-site Heisenberg Vxc and DMRG comparisons."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Gives the Anderson impurity Vxc ansatz and the NRG benchmark spectra reproduced here."},{"cited_title":"Particle conservation in the single-particle Green's function","cited_arxiv_id":"2101.00704","evidence_quote":"Provides the DMRG benchmark spectra for the 1D Hubbard chain used to assess the cluster-extrapolated Vxc."},{"cited_title":"Bergersen, F","cited_arxiv_id":null,"evidence_quote":"Gives the exact gap of the 1D Hubbard model against which the calculated alpha-U gap is compared."},{"cited_title":"Aryasetiawan, L","cited_arxiv_id":null,"evidence_quote":"Supplies the inelastic neutron scattering data for KCuF3 used to compare the Heisenberg dynamic structure factor."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the measured sodium photoemission spectrum reproduced by the electron-gas model."}],"review_version":1}