{"id":"c83ad4ea-77dd-4d36-b980-b127969f0967","arxiv_id":"2607.23414","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Multiconfigurational mixed quantum-classical dynamics captures multi-exciton spatial correlations that mean-field Ehrenfest misses, while a mean-field path approximation recovers early-time correlations efficiently.","lead":"A new simulation method treats many interacting excitons with a multiconfigurational quantum wavefunction while phonons move classically. It shows mean-field diffusion is roughly right but spatial correlations need the multiconfigurational treatment, and a cheaper path approximation recovers early correlations.","discovery_kind":"new_method","skeptic_critique":{"model":"moonshotai/kimi-k3","headline":"The pair-density decay attributed to \"phonon-induced decoherence\" is produced only by Ehrenfest trajectory averaging, whose dephasing rates are unvalidated; every phonon-coupled result (Figs. 3–6) is benchmarked only against itself, never against an exact quantum-phonon reference.","rationale":"My concern is the same one the reader flagged as the weakest assumption — absence of any exact or higher-level quantum-phonon benchmark for the 400-site phonon-coupled systems that carry the abstract-level claims. I sharpen it in one respect the reader left implicit: the specific quantity most at risk is not the MSD (Ehrenfest MSDs for high-T, weak-coupling Holstein models are usually semi-quantitatively fine, and the paper's MSD differences between M values are modest anyway) but the *decay dynamics of the pair-density correlations*, which the paper explicitly attributes to phonon-induced decoherence. Since Ehrenfest's only dephasing mechanism is ensemble averaging over classical trajectories, the transient lifetime of the cross-peaks — arguably the paper's most novel physical observation — is the quantity least anchored to anything exact. The M=1-misses-correlations half of the claim is close to structural and safe. This does not warrant a verdict change: the reader's CONDITIONAL already hinges on exactly this gap (\"conditional on code/ensemble details and clearer limits of Ehrenfest for the correlation claims\"), the regime chosen is the most forgiving one for the method, and the derivation itself (Eqs. 5–20, gauge fixing, regularized inverse) reads as standard, internally consistent MCTDHB-plus-Ehrenfest with no broken step identified. The proposed test is cheap relative to what the authors already compute — it reuses their own 21-site benchmark geometry and asks only for a truncated-Fock or tensor-network reference at matched parameters — and would either convert the CONDITIONAL toward acceptance or precisely delimit where the correlation claims are reliable.","tokens_in":17101,"tokens_out":3037,"duration_ms":95602,"concrete_test":"Run the Fig. 3–4 setup scaled down to the Fig. 1 geometry (N=21 sites, N_ex=3, same τ, U=0.01 a.u., ω₀=5 meV, γ, T=150 K) with a numerically exact phonon treatment: propagate the full exciton-phonon wavefunction with each site oscillator truncated to n_max phonons (converge in n_max), or use ML-MCTDH/t-DMRG on an equivalent chain. Compare MSD(t) and ρ⁽²⁾(i,j;t) at t=0.05–0.48 ps against M=1,3,6 and MFPA. If the exact cross-peak decay time and the M=1-vs-M>1 contrast match the Ehrenfest results within ~20%, the central claim stands; if exact dynamics shows a different decay time, different cross-peak structure, or a different M-ranking, the \"phonon-induced decoherence\" narrative and the MFPA accuracy window are artifacts of trajectory averaging.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim has two halves: (i) M>1 captures spatial pair correlations that M=1 misses, and (ii) these correlations and their ~sub-ps decay under phonons are physical, with MFPA reproducing them. Half (i) is internally secure — it is nearly structural (for M=1, ρ⁽²⁾ factorizes so g⁽²⁾ is flat, as the authors admit \"as expected\"), and the phonon-free/static-disorder exact benchmarks (Figs. 1–2) show genuine M-convergence of the excitonic propagator. Half (ii) is where the load sits, and it is least secure. The phonon-coupled figures compare M=1,2,3,6 and MFPA only against each other within the same multiconfigurational-Ehrenfest framework. But the physically meaningful content of Figs. 4–6 is the *decay* of pair-density cross-peaks between 50 fs and 480 fs, explicitly attributed to \"phonon-induced dynamical disorder and decoherence.\" Mean-field Ehrenfest has no true decoherence mechanism: the apparent decay of ρ⁽²⁾ features arises solely from averaging over an ensemble of classical nuclear trajectories (Wigner-sampled), i.e., classical disorder dressed as dephasing. Ehrenfest is documented to give quantitatively wrong coherence-decay times in Holstein/spin-boson-type models, and here it also couples to a force (Eq. 20) built only from ρ⁽¹⁾, so nuclear back-reaction is insensitive to the very two-body correlations being measured. The chosen regime (T=150 K ≫ ℏω₀=5 meV, weak γ=1.24e-5 a.u., harmonic bath) is the friendliest possible for Ehrenfest, which mitigates but does not settle this — the *quantitative* decay time of the cross-peaks, the M=6 \"convergence\" claim for g⁽²⁾, and the MFPA accuracy window all inherit this unvalidated dephasing. The one exact benchmark is phonon-free, 21 sites, 3 excitons, 0.02 ps; the transport/correlation claims are 400 sites, up to 20 excitons, 4.8 ps — no anchor connects the two regimes. Additionally, trajectory counts for the Wigner-sampled ensembles are not reported, so statistical convergence of ρ⁽²⁾ maps is unverifiable.","agreement_with_reader":"agree"},"referee_report":{"model":"moonshotai/kimi-k3","summary":"The manuscript introduces a multiconfigurational mixed quantum-classical method for finite-temperature multi-exciton dynamics coupled to phonons. The excitonic subsystem is propagated as a linear combination of bosonic permanents over M time-dependent single-particle functions (MCTDHB-like), derived via the Dirac–Frenkel variational principle with gauge fixing, while the phonons evolve quasi-classically under a mean force built from the multiconfigurational one-body density (Eq. 20). The method reduces to multi-trajectory Ehrenfest at Nex=1 and to MCTDHB without phonons. Benchmarks against exact Schrödinger propagation for 3 excitons on 21 sites, with and without static disorder (Figs. 1–2), show systematic convergence in M. On 400-site phonon-coupled systems, the authors report that mean-field (M=1) gives semi-quantitative MSD/diffusion but misses spatial pair correlations, that M>1 captures transient ρ⁽²⁾/g⁽²⁾ cross-peaks that decay on a sub-ps timescale under phonon-induced disorder (Figs. 3–6), and that a mean-field path approximation (MFPA) reproduces the early-time pair correlations at reduced cost even though it fails the MSD.","tokens_in":17643,"tokens_out":3252,"duration_ms":100969,"significance":"If the phonon-coupled results hold, the work is a useful and timely contribution: it provides a systematic, convergent-in-M route to many-body correlations in exciton transport beyond Gross–Pitaevskii/mean-field treatments, with direct relevance to multi-exciton phenomena in organic semiconductors and, prospectively, polaritonic systems. Specific strengths: the EOMs are derived cleanly from the Dirac–Frenkel principle with explicit gauge fixing and a stated regularization of the inverse one-body density matrix; the method has the correct reduction limits (MTE at Nex=1; MCTDHB without phonons); and the excitonic propagator is benchmarked against exact Schrödinger dynamics with demonstrated M-convergence, with and without static disorder (Figs. 1–2). The finding that MFPA reproduces early-time pair correlations from precomputed mean-field trajectories, if quantitatively established, would be a practically valuable cost-saving result. However, the phonon-coupled half of the claims currently rests entirely on internal cross-M comparisons within an unvalidated Ehrenfest nuclear treatment, which caps the present evidentiary weight of Figs. 3–6.","major_comments":[{"comment":"The physical content of Figs. 4–6 is the ~sub-ps decay of the pair-density cross-peaks, explicitly attributed to 'phonon-induced dynamical disorder and decoherence.' Within multi-trajectory Ehrenfest, however, there is no true decoherence mechanism: ensemble averaging over Wigner-sampled classical nuclear trajectories produces dephasing that is classical disorder dressed as decoherence, and mean-field Ehrenfest is documented to give quantitatively inaccurate coherence-decay times in Holstein/spin-boson-type models. Moreover, the nuclear force in Eq. (20) depends only on the one-body density ρ⁽¹⁾ᵢᵢ, so the nuclear back-reaction is blind to the two-body correlations whose decay is the claimed observable. Every phonon-coupled result (Figs. 3–6) is benchmarked only against other runs of the same framework (M=1 vs 2 vs 3 vs 6, MFPA vs M=3). This is load-bearing for half of the central claim.","section":"§III, Figs. 3–6; Eq. (20)"},{"comment":"The abstract and §IV state that MFPA 'can reproduce the spatial correlations to a good accuracy,' and this is the paper's main practical deliverable. The supporting evidence (Fig. 5(d) window-averaged ρ⁽²⁾, Fig. 5(e)–(l) maps, Fig. 6) is qualitative. Two gaps: (a) no quantitative error metric is given for 'good accuracy' (e.g., relative L2 error of g⁽²⁾ or of the windowed ρ⁽²⁾ trace vs. the M=3/M=6 reference as a function of time), and Fig. 5(d) itself shows visible deviations that are not characterized; (b) no mechanistic explanation is offered for why MFPA — which demonstrably fails the MSD (Fig. 5(a)) — nonetheless succeeds for ρ⁽²⁾, i.e., why early-time pair correlations are insensitive to the nuclear back-reaction error that corrupts transport. Since MFPA is an ad-hoc approximation introduced in this paper, its domain of validity should be established quantitatively (at minimum, the","section":"§III, Fig. 5 and Fig. 6; Eq. (29)"},{"comment":"Fig. 3 (MSD and diffusion coefficients on 400 sites) and Figs. 4–5 use M=3 as the de facto reference, and Fig. 6 uses M=6, but convergence in M for the phonon-coupled observables is asserted only by cross-M comparison within Ehrenfest: no M-dependence beyond M=3 is shown for the MSD/diffusion coefficients, and M=6 is shown only for g⁽²⁾ at two times. For Nex=15–20 with U=0.01 a.u., the statement (Fig. 4 caption text) that 'M=2 produces reasonably converged results' is not demonstrated for the transport observables. Please show M=4 (or M=6) MSD and diffusion-coefficient curves for at least the largest Nex, or state clearly which observables are converged at which M. Relatedly, the number of nuclear trajectories used in the phonon-coupled simulations is not stated anywhere I can find, and no statistical error bars appear on the MSD, D(Nex) curves, or Fig. 5(d); these are essential for judg","section":"§III, Figs. 3–6"},{"comment":"The normalization of the initial Gaussian SPF, ϕ₀ᵢ(0) = (1/√N)exp[−(i−i_mid)²/2σ²], writes the total excitation population implicitly; it should be stated explicitly how Nex excitations are distributed (i.e., that the permanent coefficient carries all Nex in mode 0) and how ρ⁽¹⁾ᵢᵢ(0) relates to Nex|ϕ₀ᵢ|². Also, in Eq. (9) the symbol Ĥ is redefined as Ĥ − Ĥ_p ('refers to the matrix element of the many-body Hamiltonian Ĥ = Ĥ − Ĥ_p'), overloading the Hamiltonian symbol defined in Eq. (1); a distinct symbol would avoid confusion, particularly because the coupling term Ĥ_e−p must still contribute through the classical Ri(t) in h_ij.","section":"§II, Eqs. (9) and (25)"}],"minor_comments":[{"comment":"Typo: 'forin silicoinvestigation' — missing spaces around 'in silico.'","section":"§I"},{"comment":"Hopping parameter τ is quoted in a.u. in Figs. 1–2 but in cm⁻¹ (300 cm⁻¹) in Fig. 3; please use consistent units or give the conversion. Similarly 'σ = 50 units' should read '50 lattice sites' with the lattice constant a defined (used later in MSD units of a²).","section":"Fig. 3 caption"},{"comment":"The averaging window in Fig. 5(d) (half-width 5, centered at (0,−10)) appears arbitrary; please motivate the choice (e.g., centered on the dominant cross-peak) and show that conclusions are insensitive to window placement.","section":"Fig. 5(d)"},{"comment":"Fig. 4 uses Nex=10 while Fig. 5 uses Nex=15 and Fig. 3 varies Nex; a short table summarizing parameters (N, Nex, M, U, number of trajectories/realizations) per figure would greatly aid readability and reproducibility.","section":"§III"},{"comment":"The clarification that the method is unrelated to Shalashilin's 'Multiconfigurational Ehrenfest' [52] is helpful; consider also citing ML-MCTDH-based exciton-phonon dynamics work as alternative fully quantum benchmarks to motivate the requested validation.","section":"§I"},{"comment":"Eq. (12): the projector P is defined on |ϕk⟩ while the sum runs over eigenstates |ϕ̃k⟩; please make the notation consistent.","section":"§II, Eq. (12)"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a competent method extension squarely within the journal's scope. The principal scientific risk is the one identified in Major Comment 1: all phonon-coupled physics claims rest on multi-trajectory Ehrenfest, validated only against itself. A small-system benchmark with quantum phonons would substantially raise confidence and is feasible for this group. Self-citation to the group's prior mean-field MQC work is background and appropriate, not circular."},"author_rebuttal":null,"desk_editor":{"model":"grok-4.5","letter":"The one thing worth knowing: they built a usable MCTDHB-style permanent expansion for interacting excitons on top of multi-trajectory Ehrenfest phonons, and they show cleanly that mean-field (M=1) can give decent MSD while completely missing spatial pair correlations—then that precomputed M=1 paths (MFPA) recover early-time g(2) cheaply. That split is the paper’s real payload.\n\nWhat is new is the combination, not either piece alone. Dirac–Frenkel EOMs, gauge fixing, ρ(1)/ρ(2) forces, and the stated limits (MTE at Nex=1; MCTDHB without phonons) are standard and internally consistent. Figs. 1–2 against exact Schrödinger on 21 sites with and without static disorder are honest convergence checks. The transport trends with U and Nex (Fig. 3) are readable, and the pair-density maps make the mean-field failure structural and obvious rather than hand-waved. Self-citation to their prior mean-field polariton MQC is background, not a loop.\n\nSoft spot, in proportion: every phonon-coupled claim that matters (Figs. 3–6, the ~sub-ps decay of cross-peaks, MFPA accuracy, M=6 “convergence” of g(2)) is only compared inside the same Ehrenfest ensemble. There is no exact or higher-level quantum-phonon anchor on the 400-site systems. Apparent decoherence is classical trajectory averaging of Wigner-sampled harmonic baths; Ehrenfest is known to mis-time coherences in Holstein-type models, and the force (Eq. 20) only sees ρ(1), so nuclei never feel the two-body correlations being plotted. Trajectory counts are unreported. That does not break the method paper, but it does mean the physical attribution of the decay and the quantitative MFPA window are provisional. The pure-exciton half of the story is on firmer ground than the dissipative half the abstract leads with.\n\nThis is for people who already run MQC exciton/polariton dynamics and care about correlation observables, not for a general quantum-dynamics audience. Math and citation pattern look solid for the subfield. I would send it to peer review; a serious referee should demand ensemble sizes, clearer Ehrenfest limits, and tempered language on “phonon-induced decoherence.” Engage if you work on multi-exciton or polariton transport numerics; skim the MFPA result if you only need the practical takeaway.","headline":"Solid beyond-mean-field MQC method for multi-exciton pair correlations; the phonon-coupled decay claims rest on unbenchmarked Ehrenfest averaging, but the core methodological step is real and worth refereeing.","tokens_in":18515,"tokens_out":635,"would_cite":true,"duration_ms":23065,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"grok-4.5","headline":"A multiconfigurational mixed quantum-classical method captures multi-exciton spatial correlations that mean-field misses, while a mean-field path shortcut recovers early-time correlations cheaply.","keywords":["mixed quantum-classical dynamics","multiconfigurational wavefunction","multi-exciton transport","phonon-induced disorder","spatial pair correlations","mean-field path approximation","Bose-Hubbard-Holstein"],"falsifier":"On a phonon-coupled multi-exciton chain small enough for a fully quantum or higher-level nuclear treatment, compare multi-mode multiconfigurational pair densities and MSD against that benchmark; a large mismatch would falsify the claim that the quasi-classical multiconfigurational forces and MFPA are reliable.","tokens_in":18115,"feed_emoji":"⚛️","tokens_out":1000,"duration_ms":22991,"temperature":0.7,"pith_summary":"Simulating many interacting excitons coupled to phonons is hard: full quantum dynamics is intractable, and mean-field treatments lose the spatial correlations that matter for quantum materials. This paper builds a mixed approach in which the excitons are a multiconfigurational wavefunction—a linear combination of bosonic permanents over several time-dependent modes—while the phonons move quasi-classically under forces taken from that correlated density. On a dissipative lattice it shows that phonon disorder and exciton–exciton repulsion together set how transport and pair correlations depend on excitation number. Mean-field (one mode) still gives semi-quantitatively right diffusion, but only multi-mode runs produce the pair-density signatures; precomputing nuclear paths at mean-field and then evolving the multiconfigurational excitons (MFPA) recovers those early-time correlations at much lower cost. The practical claim is a usable route to correlated multi-exciton physics at finite temperature without a full quantum phonon treatment.","feed_headline":"Mean-field misses exciton pair maps; a cheap path fix recovers them","feed_subtitle":"Multi-mode mixed quantum-classical dynamics captures correlations; precomputed mean-field trajectories often suffice early on.","key_machinery":"Multiconfigurational many-body Ehrenfest: the excitonic state is expanded in time-dependent permanents built from M single-particle modes (recovering MCTDHB without phonons and multi-trajectory Ehrenfest at N_ex=1), with nuclear forces taken from the multiconfigurational one-body density ρ⁽¹⁾.","core_discovery":"A multiconfigurational mixed quantum-classical ansatz—excitons as a sum of permanents over M time-dependent bosonic modes, phonons quasi-classical with forces from the multiconfigurational one-body density—captures phonon-coupled multi-exciton spatial pair correlations that the mean-field (M=1) limit misses, while mean-field still yields semi-quantitatively accurate diffusive transport; the mean-field path approximation reproduces early-time spatial correlations to good accuracy.","pith_inferences":["If MFPA holds more generally, screening multi-exciton or polariton materials for short-time quantum correlations becomes feasible on device-scale lattices without multi-mode nuclear forces on every trajectory.","The reported crossover—from diffusion falling with N_ex at weak U to rising with N_ex at strong U—suggests a practical knob (density versus interaction) for tuning whether phonon disorder or repulsion dominates transport.","Failure of classical-path and mean-field-force approximations on MSD implies that any polariton extension will still need some form of excited-state nuclear back-reaction for long-time transport."],"forward_implications":["Mean-field mixed quantum-classical runs remain usable for excitation-dependent diffusion constants when only transport, not correlations, is required.","Early-time spatial pair correlations in multi-exciton systems can be computed by reusing precomputed mean-field nuclear paths (MFPA) instead of full multi-mode back-reaction.","Increasing the number of modes systematically improves pair-density maps and modestly corrects diffusion when repulsion competes with phonon disorder.","The same multiconfigurational mixed framework is positioned to treat correlated exciton-polariton dynamics under phonon disorder."],"fun_headline_variants":["Mean-field misses exciton pair correlations; multiconfig paths recover them","Multiconfigurational mixed QC captures multi-exciton spatial correlations","Mean-field transport is fine; pair maps need multiconfigurational ansatz","Precomputed mean-field paths restore early exciton spatial correlations","Exciton-exciton interactions plus phonons demand more than mean-field"],"cache_read_input_tokens":128,"weakest_assumption_plain":"That treating the phonons as classical harmonic trajectories driven by the instantaneous multiconfigurational density is accurate enough for the finite-temperature transport and correlation claims on large lattices, even though exact checks exist only for pure excitonic dynamics on small systems.","fun_headline_variants_meta":{"raw":{"variants":["Mean-field misses exciton pair correlations; multiconfig paths recover them","Multiconfigurational mixed QC captures multi-exciton spatial correlations","Mean-field transport is fine; pair maps need multiconfigurational ansatz","Precomputed mean-field paths restore early exciton spatial correlations","Exciton-exciton interactions plus phonons demand more than mean-field"]},"model":"grok-4.5","effort":"low","cost_usd":0.004246,"raw_usage":{"total_tokens":1264,"prompt_tokens":780,"num_sources_used":0,"completion_tokens":80,"cost_in_usd_ticks":42464000,"prompt_tokens_details":{"text_tokens":780,"audio_tokens":0,"image_tokens":0,"cached_tokens":128},"completion_tokens_details":{"audio_tokens":0,"reasoning_tokens":404,"accepted_prediction_tokens":0,"rejected_prediction_tokens":0}},"tokens_in":780,"tokens_out":80,"duration_ms":8215,"temperature":1.0,"reasoning_tokens":404,"cache_read_input_tokens":128,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-07-30T22:53:54.609256+00:00","model_set":{"reader":"grok-4.5"},"falsifier":"On a phonon-coupled multi-exciton chain small enough for a fully quantum or higher-level nuclear treatment, compare multi-mode multiconfigurational pair densities and MSD against that benchmark; a large mismatch would falsify the claim that the quasi-classical multiconfigurational forces and MFPA are reliable.","supporting_citations":[],"review_version":1}