{"id":"bc02085e-52a8-4c40-9d64-fdf564f7550a","arxiv_id":"2412.20648","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"For N particles in a common boson bath, the critical coupling for the excitation-trapping transition is non-monotonic in N and is reduced by ferromagnetic Ising coupling.","lead":"Two-level particles coupled to a shared boson bath show a sharp transition between total decay and partial trapping of excitation as the coupling increases. This paper uses numerical renormalization group simulations to show that the critical coupling peaks at an intermediate particle number and is lowered by ferromagnetic interactions between the particles.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The extraction of α_c for N≥2 rests on an unbenchmarked NRG truncation and an assumed linear vanishing of the gap; a systematic N-dependent bias could reshape the claimed non-monotonic α_c(N).","rationale":"The paper's central claim is that α_c(N) is non-monotonic and that the sign of g shifts α_c. For N=1 the numerical method is checked against an analytic solution (Appendix C), and the α→∞ limit for N=2 is reproduced (Eqs. 6–11). These are real supporting checks. However, the entire N≥2 α_c curve in Fig. 4 is produced by one procedure: linear extrapolation of ΔE=E_0−E_1 to zero in the C=N subspace. The procedure has two unvalidated components — convergence of the NRG truncation for N≥2, and the assumed linear vanishing of the gap — and neither is benchmarked for N>1. Near the threshold, the gap is comparable to the discretization scale, so errors in either component translate directly into errors in α_c. Because the non-monotonic feature is only a few times 10^{-3} wide in α_c, even a moderate N-dependent bias could alter the conclusion. This is the same weakest assumption identified by the reader. The recommendation remains conditional rather than reject because the internal consistency checks and the N=1 control make a total failure unlikely, but the central quantitative claim should not be accepted without a multi-N convergence study.","tokens_in":16790,"tokens_out":10393,"duration_ms":120484,"concrete_test":"Repeat the α_c determination for N=2, 4, 6, s=1, g=0 and for the g=±0.01 cases that set the sign claim, using NS=2000 or 4000 kept states, Λ=1.05, and M=120, and additionally compare a linear fit of ΔE versus α with a power-law fit ΔE = A(α−α_c)^ν; as a control, require the same procedure to reproduce the analytic N=1 value α_c=0.025. If the inferred α_c(N) curve changes by more than the typical point spacing in Fig. 4, or if the sign of the g-shift flips for any N, the central qualitative claim is not robust.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The load-bearing step is the extraction of α_c from the crossing of the linearly fitted gap ΔE = E_0 − E_1 in the conserved C=N subspace (Section III, Figs. 2 and 4, Appendix B). For N=1 this is benchmarked against the analytic α_c = sΔ/2ω_c (Appendix C), but for N≥2 no analogous benchmark exists, and two uncontrolled approximations enter. First, the NRG truncation (Λ=1.1, NS=1000, M=100) is asserted converged only for N=1 (Fig. 13); the C=N sector for N≥2 contains up to N bosons and a much larger low-energy state space, so discarded states can shift E_0 and E_1. Second, the gap is assumed to vanish linearly in α−α_c so a linear fit through points above threshold can be extrapolated to zero, but the near-threshold scaling of a multi-particle bound state is not derived. Because the central claim in Fig. 4 is read off from these extrapolated crossings, a systematic N-dependent bias of even a few times 10^{-3} in α_c for s=1 — comparable to the spacing between adjacent N points — would rearrange the claimed maximum and its dependence on g. The paper provides no estimate of this bias.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies the generalized Jaynes-Cummings model with N two-level particles coupled to a common bosonic continuum with spectral function J(ω) ~ ω^s, including an all-to-all Ising interaction. Using the numerical renormalization group (NRG), the authors compute eigenstates in the conserved C=N subspace, identify a dynamical transition in the excited-state probability Pe(t) as the emergence of a discrete ground state below the continuum, and extract the critical coupling α_c from a linear extrapolation of the gap ΔE=E0-E1. For N=1 they reproduce the analytic result α_c = sΔ/(2ω_c). For N≥2 they find that α_c first increases with N, reaches a maximum at a finite N, and then decreases; ferromagnetic (antiferromagnetic) Ising coupling reduces (increases) α_c. The α→∞ limit for N=2 is treated analytically and agrees with the NRG results for large α.","tokens_in":16997,"tokens_out":8173,"duration_ms":72286,"significance":"If the numerical extraction is reliable, the result provides a nontrivial extension of the single-particle dynamical transition to multi-particle and interacting systems, with practical implications for controlling decoherence and for CNOT gate operations. The paper includes a clear benchmark against the analytic N=1 solution, a physically motivated transition criterion, and an analytic large-coupling analysis for N=2. However, the central quantitative claim depends on the NRG results for N≥2, for which the convergence evidence is incomplete and the method of extracting α_c is not fully validated beyond N=1.","major_comments":[{"comment":"The confirmation that the ground state becomes discrete for N=2 is asserted with reference to Fig. 11, but Fig. 11 shows the overlap convergence for a single particle (N=1), not for N=2. The authors should provide analogous data for N≥2, for example |⟨ℓ=0|ψ_2(0)⟩|^2 versus iteration M and ΔE versus M for α just above α_c, to substantiate the discrete-state criterion used to extract α_c.","section":"Section III, paragraph starting \"The calculated eigen-energies provide a way...\" and Fig. 2"},{"comment":"Convergence of the NRG parameters (Λ=1.1, NS=1000, M=100) is demonstrated only for N=1. For N≥2, the C=N subspace contains up to N bosons and the truncation may systematically shift E0 and E1 near threshold, where ΔE is small. No convergence study (e.g., α_c versus M, NS, or Λ for fixed N) is provided for N≥2. This is load-bearing because Fig. 4 and Table I are based on these extrapolated crossings; a systematic N-dependent bias of even a few times 10^{-3} in α_c, comparable to the spacing between adjacent points in Fig. 4 for s=1, could alter the claimed non-monotonic behavior and its dependence on g. Please provide convergence tests for at least N=2 and N=4 and report the resulting uncertainty in α_c.","section":"Section III and Appendix C (Fig. 13)"},{"comment":"The linear extrapolation of ΔE in α to locate α_c is assumed to hold for all N, but the near-threshold scaling of the gap for multi-particle bound states is not derived. For N=1 the method is benchmarked against the analytic result, but for N≥2 the linearity is not demonstrated. The authors should report the fit ranges, the number of points used, and a fit-quality measure for the crossings underlying Fig. 4, or alternatively test a nonlinear (e.g., square-root) form for ΔE near α_c. Without this, the extrapolation error in α_c is uncontrolled and could be comparable to the spacing between data points in Fig. 4.","section":"Section III, paragraph on linear fit, and Fig. 12"}],"minor_comments":[{"comment":"The phrase \"we access the effects\" should read \"we assess the effects\".","section":"Introduction, first paragraph of Section I"},{"comment":"The construction \"multi-particles\" is nonstandard; \"multi-particle\" is the conventional adjective form (as in \"multi-qubit\" used later).","section":"Title and text throughout"},{"comment":"The Wigner-Weisskopf argument that Pe(t,g1)<Pe(t,g2) for g1>g2 and the subsequent conclusion about α_c are heuristic; a more direct derivation or numerical test of the ordering of Pe(t) curves would strengthen the explanation.","section":"Section III, paragraph on the g-dependence of α_c"},{"comment":"The explanation of the non-monotonic dependence of α_c on N (competition between NΔ and the number of matrix elements ~e^{N ln M}) is qualitative. While not a blocking issue, the authors might quantify this argument or support it with additional numerical checks.","section":"Section III, paragraph on the maximum of α_c"}],"recommendation":"major_revision","confidential_remarks":"The manuscript addresses a timely topic and the NRG implementation is benchmarked for N=1, but the central quantitative claims for N≥2 rest on convergence and extrapolation procedures that are not yet demonstrated. The requested convergence tests and error estimates are feasible within the scope of the paper and would significantly strengthen the reliability of the reported α_c(N,g)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the N-particle extension of the known single-particle generalized Jaynes-Cummings transition, with an all-to-all Ising interaction. The finding that α_c is non-monotonic in N, peaking at a finite N, and that ferromagnetic coupling suppresses while antiferromagnetic coupling enhances α_c, is not in the cited literature and is physically plausible. The NRG implementation is careful: the N=1 benchmark against the analytic solution works, the α→0 and α→∞ limits are checked analytically for N=2, and the transition criterion via the gap between the ground state and the continuum is well motivated. The paper is coherent and honestly argued.\n\nThe soft spots are real but addressable. First, there are no error bars on α_c. The extraction is a linear fit of the gap ΔE versus α extrapolated to zero, and for N≥2 there is no independent check that this extrapolation is unbiased. The near-threshold scaling of a multi-particle bound state is not derived. Since the spacing between adjacent N points in Fig. 4 is small, a systematic NRG bias of even a few times 10^-3 in α_c could rearrange the claimed maximum. Second, convergence of Λ, NS, and M is demonstrated only for N=1; the C=N sector for N≥2 has a larger state space, and the authors simply assert the same truncation works. A few NRG runs at N=2 or 3 with smaller Λ or larger NS would settle this. Third, the abstract I saw has the sign of the g dependence reversed relative to the main text and Table I: main text says ferromagnetic (g<0) suppresses α_c, but the abstract as written says antiferromagnetic suppresses. That is a straightforward slip that must be fixed.\n\nNone of this invalidates the central qualitative picture. The small-α Wigner-Weisskopf argument for the g dependence is sensible, and the α→∞ limit confirms a discrete state exists. I think the claimed non-monotonicity is probably right, but the paper as it stands does not quantify its uncertainty well enough for me to cite the specific α_c values.\n\nThis paper deserves a serious referee, not a desk reject. I would send it to review, with the request that the referee ask for explicit convergence checks at N≥2, error estimates on α_c, and a corrected abstract. The qualitative result is worth publishing; the quantitative claim needs support.","headline":"Plausible new result on the N-dependence of the dynamical transition in the generalized Jaynes-Cummings model, but the quantitative claim needs error bars and a convergence check for N≥2 before I'd trust the maximum.","tokens_in":17563,"tokens_out":6445,"would_cite":false,"duration_ms":66690,"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":"For N particles sharing a continuum boson bath, the critical coupling for the dynamical transition peaks at a finite N and then declines with further increases in N, and a ferromagnetic Ising interaction lowers this threshold while an…","keywords":["generalized Jaynes-Cummings model","dynamical transition","numerical renormalization group","boson bath","spectral function","Ising interaction","decoherence","multi-qubit dynamics"],"falsifier":"Take the $N=2$, $s=1$, $g=0$ case with $\\Delta=0.05$ and $\\omega_c=1$, for which the paper reports $\\alpha_c=0.0443$. Recompute the gap $\\Delta E=E_0-E_1$ with a substantially finer discretization ($\\Lambda\\to1.01$ or smaller), more kept states, and longer chains, or with an independent non-perturbative method; if the linear fit to $\\Delta E$ crosses zero at a value that differs from $0.0443$ by more than a few percent, or if $\\Delta E$ remains finite and positive for $\\alpha$ slightly below that value, the reported non-monotonic $\\alpha_c(N)$ curve fails. A minimal version is to check whether the maximum in $\\alpha_c(N)$ persists under this convergence test for $s=1/2$, $1$, and $2$.","tokens_in":16550,"feed_emoji":"⚛️","tokens_out":12377,"duration_ms":109171,"temperature":0.7,"pith_summary":"This paper asks how the number of excited two-level particles and an all-to-all Ising interaction between them change the dynamical transition of the generalized Jaynes-Cummings model, where the particles share a common continuum of boson modes. Using a non-perturbative numerical renormalization group calculation benchmarked against the known single-particle solution, the authors find that the critical system-bath coupling $\\alpha_c$ — the value below which the excited-state probability decays monotonically to zero and above which a finite fraction survives in the long-time limit — is non-monotonic in the particle number $N$, reaching a maximum at some finite $N$ and decreasing for larger $N$. They also find that a ferromagnetic Ising coupling between particles ($g<0$) suppresses $\\alpha_c$, making persistent excitation easier, while an anti-ferromagnetic coupling ($g>0$) enlarges it. If correct, these results give a concrete principle for protecting multi-qubit states from bath-induced decay: engineering ferromagnetic interactions among qubits sharing a bath lowers the coupling threshold for dynamical localization.","feed_headline":"Critical coupling peaks at finite N, then drops","feed_subtitle":"The threshold first rises with particle number, then falls; ferromagnetic coupling lowers it throughout.","key_machinery":"The central object is the discrete dressed ground state $|D\\rangle$ that emerges below the continuum band bottom in the $C=N$ subspace; the transition is defined by the vanishing of the gap $\\Delta E = E_0 - E_1$ between the ground state and the lowest continuum state. The numerical machinery is the bosonic numerical renormalization group: the continuum bath with spectral function $J(\\omega) \\sim \\omega^s$ is logarithmically discretized (parameter $\\Lambda=1.1$) and mapped to a semi-infinite Wilson chain whose hopping amplitudes decay exponentially, and the low-energy states are obtained by iteratively adding chain sites while keeping $N_S=1000$ states per iteration up to $M=100$ sites. The discrete nature of a state is confirmed by checking that its overlap with the initial state converges to a nonzero value as the iteration number $M$ increases, while overlaps of continuum states shrink toward zero. The inter-particle term is the all-to-all Ising coupling $g\\sum_{j<k}\\sigma_j^z\\sigma_k^z$, with $g<0$ ferromagnetic and $g>0$ anti-ferromagnetic.","core_discovery":"The paper establishes that the dressed-discrete-ground-state mechanism known for a single particle carries over to $N$ particles. Working in the conserved subspace $C=N$, where $C$ counts excited particles plus boson occupation, the authors show numerically that once $\\alpha$ exceeds $\\alpha_c$ the ground state separates from the bottom of the continuum band and becomes a discrete state; this discrete state is what makes $P_e(t)$ converge to a nonzero value at long times. They locate $\\alpha_c$ by computing the gap $\\Delta E = E_0 - E_1$ and linearly fitting it to zero, and find: for fixed $N$ and $g$, $\\alpha_c$ increases with the spectral exponent $s$; for fixed $s$ and $g$, $\\alpha_c$ exhibits a maximum at a finite $N$ and decreases with sufficiently large $N$; and $\\alpha_c$ is suppressed by ferromagnetic ($g<0$) and enlarged by anti-ferromagnetic ($g>0$) Ising interaction. For $N=2$ an analytic strong-coupling limit is also derived, giving a three-state description with beating oscillations and a nonzero long-time average that agrees with the numerics at large $\\alpha$.","pith_inferences":["An implication the authors leave implicit is that the apparent finite limit of $\\alpha_c$ for large $N$ could be described by an effective collective or giant-spin degree of freedom coupled to the bath; computing the long-time plateau of $P_e(t)$ directly for $N\\gg 1$ would test this picture.","The competition argument in the paper (energy cost $N\\Delta$ versus bath-dressing matrix elements growing like $e^{N\\ln M}$) suggests a quantitative scaling prediction: the location of the finite-$N$ maximum in $\\alpha_c$ should move to larger $N$ as $s$ increases, consistent with the paper's Fig. 4; a dedicated scaling collapse over a wider range of $s$ and $N$ would make this precise.","Because the transition is a statement about $t\\to\\infty$, finite-time measurements can miss the plateau; a practical experimental marker would be the appearance of a non-decaying tail in $P_e(t)$ at times beyond the slowest bath timescale, rather than short-time exponential fits.","The sign dependence of $g$ suggests a design rule for decoherence protection: ferromagnetic Ising couplings among qubits sharing a common bath should lower the threshold for persistent excitation, while anti-ferromagnetic couplings should be avoided; this is testable in current multi-qubit quantum simulators by measuring survival probability versus interaction sign."],"forward_implications":["For every $N$ studied (up to $N=8$), a dynamical transition exists: below $\\alpha_c$, $P_e(t)$ decays to zero; above it, a finite long-time value remains, caused by a discrete ground state.","The critical coupling grows with the spectral exponent $s$ for fixed $N$ and $g$, so sub-Ohmic baths ($s=1/2$) localize the excitation at weaker coupling than Ohmic ($s=1$) or super-Ohmic ($s=2$) baths.","For sufficiently large $N$, $\\alpha_c$ decreases with $N$, and the paper's results suggest it approaches a finite limit as $N\\to\\infty$, implying a nonzero threshold even in the thermodynamic limit.","A ferromagnetic Ising interaction reduces $\\alpha_c$ for each $s$ studied, while an anti-ferromagnetic one increases it; in the CNOT-gate encoding where $\\left|\\uparrow\\uparrow\\right\\rangle$ represents a logical state, a ferromagnetic interaction would make that state less dissipative.","In the strong-coupling limit for $N=2$, the dynamics reduces to a three-state problem whose analytic solution shows beating oscillations and a nonzero long-time average, matching numerical results for large $\\alpha$."],"supporting_citations":[{"why":"Supplies the analytic single-particle solution showing the dynamical transition and the exact $\\alpha_c=s\\Delta/2\\omega_c$, used as the $N=1$ benchmark.","marker":"[46]"},{"why":"Provides the Green's-function treatment from which the authors take $U_e(\\omega)$, the discrete-state criterion, and the power-law long-time decay for $N=1$.","marker":"[47]"},{"why":"Supplies the bosonic numerical renormalization group algorithm (logarithmic discretization, Wilson-chain mapping, truncation) that the authors adapt for the multi-particle model.","marker":"[55]"},{"why":"Supplies the time-dependent numerical renormalization group framework used to compute the excited-state probability $P_e(t)$.","marker":"[54]"},{"why":"Establishes the spectral-function characterization of bath effects and the spin-boson transition phenomenology that the generalized Jaynes-Cummings model extends.","marker":"[11]"},{"why":"Provides the standard continuum-boson-bath formalism and classification of Ohmic, sub-Ohmic, and super-Ohmic spectral functions used throughout.","marker":"[12]"}],"fun_headline_variants":["N particles: critical coupling peaks then falls","Interactions reshape critical coupling in Jaynes-Cummings","Finite N maximizes critical coupling threshold","Ferromagnetic coupling lowers critical coupling","Critical coupling peak shifts with N and interaction"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The results depend on the truncated numerical renormalization group calculation with logarithmic discretization $\\Lambda=1.1$, $N_S=1000$ kept states, and $M=100$ chain sites locating the vanishing of the gap $\\Delta E$ exactly at the true $\\alpha_c$ for every $N$ and $g$, even though the convergence is benchmarked against the analytic solution only for $N=1$.","fun_headline_variants_meta":{"raw":{"variants":["N particles: critical coupling peaks then falls","Interactions reshape critical coupling in Jaynes-Cummings","Finite N maximizes critical coupling threshold","Ferromagnetic coupling lowers critical coupling","Critical coupling peak shifts with N and interaction"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000998,"raw_usage":{"total_tokens":4267,"prompt_tokens":1029,"completion_tokens":3238,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":645,"completion_tokens_details":{"reasoning_tokens":3172}},"tokens_in":645,"tokens_out":3238,"duration_ms":23998,"temperature":1.0,"reasoning_tokens":3172,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T23:14:54.474477+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the $N=2$, $s=1$, $g=0$ case with $\\Delta=0.05$ and $\\omega_c=1$, for which the paper reports $\\alpha_c=0.0443$. Recompute the gap $\\Delta E=E_0-E_1$ with a substantially finer discretization ($\\Lambda\\to1.01$ or smaller), more kept states, and longer chains, or with an independent non-perturbative method; if the linear fit to $\\Delta E$ crosses zero at a value that differs from $0.0443$ by more than a few percent, or if $\\Delta E$ remains finite and positive for $\\alpha$ slightly below that value, the reported non-monotonic $\\alpha_c(N)$ curve fails. A minimal version is to check whether the maximum in $\\alpha_c(N)$ persists under this convergence test for $s=1/2$, $1$, and $2$.","supporting_citations":[{"cited_title":"De Filippis, A","cited_arxiv_id":null,"evidence_quote":"Supplies the analytic single-particle solution showing the dynamical transition and the exact $\\alpha_c=s\\Delta/2\\omega_c$, used as the $N=1$ benchmark."},{"cited_title":"Zhang and Z","cited_arxiv_id":null,"evidence_quote":"Provides the Green's-function treatment from which the authors take $U_e(\\omega)$, the discrete-state criterion, and the power-law long-time decay for $N=1$."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the bosonic numerical renormalization group algorithm (logarithmic discretization, Wilson-chain mapping, truncation) that the authors adapt for the multi-particle model."},{"cited_title":"Burgess and M","cited_arxiv_id":null,"evidence_quote":"Supplies the time-dependent numerical renormalization group framework used to compute the excited-state probability $P_e(t)$."},{"cited_title":"Strathearn, P","cited_arxiv_id":null,"evidence_quote":"Establishes the spectral-function characterization of bath effects and the spin-boson transition phenomenology that the generalized Jaynes-Cummings model extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the standard continuum-boson-bath formalism and classification of Ohmic, sub-Ohmic, and super-Ohmic spectral functions used throughout."}],"review_version":1}