{"id":"84a75050-1955-45b4-90d4-d45fdc6a6328","arxiv_id":"2505.06878","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Iterated tree-level QED scattering maps saturate helicity entanglement to maximally entangled Bell states for fermion-fermion processes, except in stated zero-initial-entropy-gain cases, and do so only partially when photons are involved.","lead":"Repeatedly filtering and re-scattering particles in QED, the paper argues, drives almost every initial spin state to a maximally entangled state after many rounds. It is a theoretical result connecting scattering amplitudes to entanglement dynamics, with clear exceptions and a partial story when photons are involved.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Saturation proof does not establish convergence to a Bell state: in the t<0 branch λ3 and λ4 have equal modulus, so normalized iterates oscillate without a limit, and ties with λ1 or λ2 are never ruled out.","rationale":"Reader's CONDITIONAL verdict is appropriate. The invariant-set and self-similarity observations in Sec. IV are structural and plausible, and the numerics are suggestive, but the paper contains no machine-checked proof or reproducible data, so the convergence claim must stand on the spectral argument. That argument has a genuine gap in the t<0 branch: the two complex eigenvalues have equal modulus by construction, so a single dominant eigenvector cannot select an asymptotic state; the normalized trajectory rotates in the maximally entangled subspace. Strict dominance over the other two eigenvalues is asserted in words but never established, and the closing remark of Sec. IV explicitly defers this to explicit amplitudes that the preprint does not provide. If such a degeneracy occurs for physical parameters, even concurrence saturation is not guaranteed. This does not disprove the phenomenon; it makes the central claim unproven as stated. The abstract should exclude fixed points and should say 'entanglement saturates' rather than 'any initial state is transformed into a maximally entangled state' unless state convergence is proven. I therefore keep CONDITIONAL.","tokens_in":8972,"tokens_out":9270,"duration_ms":95931,"concrete_test":"Using the explicit tree-level Bhabha and Møller amplitudes (from Ref. [24] or recomputed from QED), form matrix (3) and scan θ ∈ (0,π), μ ∈ (10^{-2},10^2). At each point compute t = (−A+D+E+F)^2 − 16B^2 and the eigenvalue moduli |λ1|, |λ2|, r = sqrt(s1^2+s2^2). Check whether (i) t<0 occurs on a positive-measure set; (ii) the largest modulus is strictly larger than the second largest for all points, with a quantified gap; and (iii) in the t<0, r-dominant case, the normalized iterates f_n converge (e.g. sup_n ||f_{n+1}−f_n|| → 0 for n up to 10^4), or at least concurrence tends to 1. If a positive-measure region has no strict spectral gap, the claimed saturation fails; if t<0 is absent, the oscillatory objection disappears.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim requires that M^n|i>/||M^n|i>|| tends to a maximally entangled state for every allowed initial state, with the limiting state selected by a dominant eigenvector of M. Section IV attempts to prove this from the eigen-decomposition of Eq. (3), but the t<0 branch is not handled. For t<0, Eqs. (12)-(14) give λ3 = r e^{iη} and λ4 = r e^{-iη}, so |λ3| = |λ4| = r identically in this whole region, not merely at a measure-zero boundary. If r dominates, Eq. (19) yields normalized states proportional to a_n Φ− + b_n Ψ+, with a_n and b_n oscillating as cos(nη) and sin(nη); this sequence generically has no limit, so there is no asymptotic state and no dominant eigenvector that survives. The sentence following Eq. (19) ('the asymptotic state will be ... a state of the form cosξ Φ− + sinξ Ψ+') does not follow: the angle ξ would depend on n and generically fail to converge. Moreover, if r coincides with |λ1| or |λ2|, the relative weights of Φ+ or Ψ− versus the oscillating pair remain nonzero and the concurrence need not saturate to 1; the paper never rules out such degeneracies. The authors concede at the end of Sec. IV that 'the identification of the dominant eigenvalue requires the explicit expressions of the scattering amplitudes,' but those expressions are not given in this preprint, so the generic dominance premise is unverified. Separately, the abstract's 'any initial state' is contradicted by the paper's own |RR> ultrarelativistic fixed point (Sec. III).","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies two-particle QED scattering processes (Bhabha, Møller, e−e+ → γγ, and Compton scattering) under arbitrarily sharp momentum filtering, modeled by 4×4 matrices M acting on the initial helicity state. It claims that iterating M, either at a fixed scattering angle or with the angle randomly changed at each step, sends any initial state to a maximally entangled Bell-type state for fermion-fermion scattering, with only partial saturation when photons are involved. Section III presents numerical concurrence plots for Bhabha and Møller scattering and a table of asymptotic states. Section IV attempts a proof of the saturation mechanism for the Bhabha matrix by expanding initial states in eigenvectors of M and separating the cases t ≥ 0 and t < 0. Section V discusses the photon-involving processes and argues that saturation is reduced or absent there. The paper's central claim is that the saturation is a structural consequence of the form of the scattering matrices, independent of the explicit scattering amplitudes.","tokens_in":9232,"tokens_out":6040,"duration_ms":60317,"significance":"If the central claim were fully established, the paper would identify a robust, amplitude-independent mechanism by which iterated momentum-filtered QED scattering generates maximum helicity entanglement from generic initial states. The reliance on map structures derived from tree-level QED amplitudes, rather than on fitted parameters, is a genuine strength, and the numerical evidence for Bhabha and Møller scattering is suggestive. However, the proof as written has load-bearing gaps: the abstract overstates the result, the t < 0 branch of the eigenvalue analysis does not yield a convergent asymptotic state, and the dominance and diagonalizability assumptions are not verified. The result may be salvageable with a more careful statement and a complete eigenvalue analysis, but in its current form the central claim is not established.","major_comments":[{"comment":"The abstract's claim that 'any initial state' is transformed into a maximally entangled state is contradicted by the paper's own analysis. For the initial state |RR⟩ in the ultrarelativistic limit, the text states that M(RR,rs) = M(rs,RR) = 0 for all r,s ≠ RR, so |RR⟩ is a fixed point of the map with zero concurrence; Figs. 2 and 4 show the concurrence decreasing to zero as μ grows. The theorem statement must exclude such states or be qualified, and Section V provides further counterexamples, such as e−e+ → γγ with initial |RR⟩ producing a non-maximally entangled state.","section":"Abstract; Section III, Figs. 2 and 4"},{"comment":"In the t < 0 branch, the eigenvalues λ3 and λ4 have exactly equal modulus r over the entire region, not merely on a measure-zero boundary. Consequently, there is no dominant eigenvector among them. Equation (19) contains terms a_n Φ− + b_n Ψ+ with a_n, b_n proportional to cos(nη) and sin(nη), so if r dominates the normalized state rotates in the span of Φ− and Ψ+ and generically has no limit as n → ∞. The sentence following Eq. (19), which asserts convergence to a state cosξ Φ− + sinξ Ψ+, does not follow. The argument also does not rule out degeneracies |λ1| = r or |λ2| = r, in which case the asymptotic state retains an admixture of Φ+ or Ψ− and need not have unit concurrence.","section":"Section IV, Eqs. (12)-(14), (18)-(19)"},{"comment":"The proof assumes without demonstration that M is diagonalizable with linearly independent eigenvectors and that a single eigenvalue strictly dominates after iteration. At t = 0, which is included in the t ≥ 0 case, λ3 = λ4 and the two eigenvectors in Eq. (5) coincide, so the expansion used in Eq. (11) is invalid there. Moreover, the last paragraph of Section IV concedes that 'the identification of the dominant eigenvalue requires the explicit expressions of the scattering amplitudes,' which are not provided; thus property 3 is not established from the map's form alone as claimed.","section":"Section IV, first paragraph and Eq. (5)"}],"minor_comments":[{"comment":"The text says the second eigenvector |λ2⟩ = Ψ− has eigenvalue λ1 = E − F; the eigenvalue should be λ2 = E − F.","section":"Section IV, paragraph after Eq. (5)"},{"comment":"The acronym 'CSTP' should be 'CPTP' (completely positive, trace-preserving).","section":"Footnote 4"},{"comment":"The figure numbering is confusing: the text refers to Fig. 6 both for the fixed-angle ultrarelativistic saturation over all scattering angles and for the random-angle procedure, while Fig. 7 is not explicitly referenced in the body.","section":"Section III, Figs. 5-7"},{"comment":"The definition η = arctan(s2/s1) requires a branch choice when s1 < 0; the subsequent normalization of Ξ3 and Ξ4 should specify the chosen branch to avoid ambiguities in the oscillatory terms cos(nη) and sin(nη).","section":"Section IV, Eq. (14)"},{"comment":"The qualification 'apart in the few cases in which there is no entropy gain at the first step' is stated only in the body, not in the abstract; the abstract should be aligned with this qualification.","section":"Section III, first paragraph"}],"recommendation":"major_revision","confidential_remarks":"The manuscript leans heavily on the authors' earlier Refs. [20,24] for the map structures and invariant-set results. This is not circular in the strong sense, but the new proof of saturation is incomplete without explicit scattering amplitudes and a rigorous dominance criterion. The discrepancy between the abstract and the body's own counterexamples should be resolved before the paper can be considered for publication."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The short version: the paper's central claim is not proved as stated. The iteration mechanism for entanglement saturation in momentum-filtered QED scattering is a real idea, and the paper does a decent job of showing how a dominant eigenvector of the scattering map could select a Bell state. But the proof breaks down in the t<0 branch, and the abstract overreaches with 'any initial state.'\n\nWhat's new: the authors extend their own earlier invariant-set results [20,24] to repeated applications of the map. The observation that M^n keeps the same block form is correct, and the numerical evidence for Bhabha and Møller scattering at selected parameters is suggestive. The idea that entanglement saturates because one eigenvector dominates under iteration is natural and worth checking.\n\nThe soft spots are real. For t<0, the two non-Bell eigenvalues are complex conjugates with equal modulus r. Equation (18) shows that the unnormalized state contains terms like r^n cos(nη) Φ− and r^n sin(nη) Ψ+. After normalization, these coefficients oscillate with n and do not converge to any state unless η is a multiple of π. The sentence after Eq. (19) says the asymptotic state will be 'a state of the form cosξ Φ− + sinξ Ψ+' with ξ obtained in the limit, but no such limit exists. The paper also never rules out ties between r and the other eigenvalue moduli, which would spoil the dominant-eigenvector argument. The t=0 boundary gives coincident eigenvectors, so the expansion basis is defective there. These aren't fatal to the whole research program, but they are fatal to the proof as written.\n\nThe abstract's 'any initial state' is contradicted by the paper's own ultrarelativistic |RR> fixed point, which the text acknowledges. That should be an explicit exception.\n\nThe reliance on self-cited prior work is legitimate but makes the preprint hard to check independently. The plots are not backed by code or data, which matters for a theory paper claiming numerical evidence.\n\nBottom line: this is a within-subfield idea that deserves a serious referee, but not acceptance as-is. I'd send it to peer review with a clear request to fix the t<0 convergence, handle the defective case, and restate the claim with exceptions.","headline":"Plausible mechanism, unproven convergence: the t<0 branch makes iterates oscillate rather than settle, and the abstract's 'any initial state' overreaches.","tokens_in":9843,"tokens_out":2713,"would_cite":false,"duration_ms":27582,"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":"Repeated momentum-filtered QED scattering drives two-fermion helicity states to maximally entangled Bell states.","keywords":["entanglement saturation","quantum electrodynamics","scattering amplitudes","dynamical quantum maps","Bell states","helicity entanglement","momentum filtering","concurrence"],"falsifier":"Evaluate the Bhabha or Møller scattering matrix at momenta and angles where $(-A+D+E+F)^2-16B^2=0$ and check whether the matrix is defective; if numerical iteration of $M$ on a generic initial state then fails to converge to a maximally entangled state, the claimed general saturation would be false.","tokens_in":8659,"feed_emoji":"🔗","tokens_out":9755,"duration_ms":89149,"temperature":0.7,"pith_summary":"The paper argues that when outgoing particles in a two-fermion QED scattering event are sharply filtered in momentum, the process becomes a fixed quantum map acting on helicity states (spin along the particle momentum), and iterating this map saturates entanglement: almost any initial state is driven, in the infinite-iteration limit, to a maximally entangled Bell state. The result is structural rather than numerical, because it follows from the general form of the 4×4 scattering-amplitude matrix and not from the explicit values of the amplitudes. This matters because it identifies a universal attractor for scattering-produced entanglement, independent of energy and angle, and explains why maximal entanglement is conserved in fermion-fermion processes. The paper explicitly concedes the exception of initial states that already gain no entropy at the first step, such as |RR⟩ in the ultrarelativistic limit, which remain fixed points with zero concurrence.","feed_headline":"Repeated QED scattering saturates entanglement to Bell states","feed_subtitle":"Why it matters: even noisy, angle-varying iterations send any fermion helicity state to a maximally entangled attractor.","key_machinery":"The central object is the 4×4 helicity scattering-amplitude matrix $M$, which acts as a quantum map on the post-measurement state once the outgoing particles are sharply filtered in momentum; for Bhabha scattering it has the block form with elements $A,\\dots,F$ shown in Eq. (3). Three structural properties carry the argument: the set of maximally entangled Bell states (the four two-particle spin states of maximal entanglement) is invariant under $M$; powers $M^n$ retain exactly the same form for every $n$; and every initial state can be expanded in eigenvectors of $M$, with the dominant eigenvector surviving normalization in the $n\\to\\infty$ limit. The two non-Bell eigenvectors of the Bhabha matrix appear as a conjugate pair with common modulus when $t<0$, so their oscillatory phases cancel and the long-time state is again a combination of $\\Phi^-$ and $\\Psi^+$, which is maximally entangled.","core_discovery":"According to the paper, iterated momentum-filtered scattering in Bhabha and Møller processes is governed by a scattering-amplitude matrix $M$ whose powers retain the same block form, so a normalized trajectory $N_n^{-1} M^n |i\\rangle$ is controlled entirely by the eigenvector of $M$ with the largest eigenvalue. In the ultrarelativistic Bhabha case the eigenvectors are the four Bell states, and for initial $|RL\\rangle$ the asymptotic state is $\\Psi^+$ for generic scattering angles; in the nonrelativistic case the non-Bell eigenvectors are either real combinations of $\\Phi^-$ and $\\Psi^+$ (when the discriminant $t\\ge0$) or conjugate pairs that rotate with $n$ but collapse onto a combination of $\\Phi^-$ and $\\Psi^+$ after normalization (when $t<0$). In every surviving case the dominant eigenvector is maximally entangled, so the concurrence saturates to 1. The same conclusion holds when the scattering angle is chosen randomly at each step, because each factor $1+\\cos^2\\theta_i$ exceeds the competing factor $2\\cos\\theta_i$. For photon-involving processes this structural saturation is reduced ($e^-e^+\\to\\gamma\\gamma$) or absent (Compton scattering).","pith_inferences":["One extension of this result is that the attractor should be robust to imperfect filtering: a noisy, angle-varying protocol still saturates according to the random-angle argument, which could be tested experimentally.","Because the mechanism depends only on the block form of the map, similar saturation may occur in other two-particle interactions whose scattering amplitudes satisfy the same relations, such as certain weak or strong processes.","The self-similarity of $M^n$ suggests an underlying discrete symmetry of the post-measurement dynamics; uncovering that symmetry could predict which Bell state dominates without computing amplitudes."],"forward_implications":["For two-fermion QED scattering, repeated momentum-filtered scattering drives the helicity state to a maximally entangled Bell state, so entanglement generated by scattering saturates instead of fluctuating indefinitely.","The asymptotic Bell state is selected by the dominant eigenvector of the scattering matrix, so identifying it only requires comparing the competing eigenvalues, not simulating the full dynamics.","Entanglement increase is not monotonic: entanglophobous states can temporarily reduce concurrence, but the trajectory still converges to unit concurrence.","When photons participate, saturation becomes process-dependent: $e^-e^+\\to\\gamma\\gamma$ saturates only for some initial states, while Compton scattering never saturates.","The mechanism is fixed by the form of the map, so the explicit energy dependence of the amplitudes decides which Bell attractor is reached, but not whether saturation occurs."],"supporting_citations":[{"why":"supplies the tree-level QED formalism and the POVM description of sharp momentum filtering that defines the post-measurement helicity state.","marker":"[18]"},{"why":"establishes the invariant sets of maximally entangled states for these maps and identifies the entanglophobous states that produce nonmonotonic entanglement.","marker":"[20]"},{"why":"introduces the dynamical quantum maps for QED processes and their invariant sets, which the present paper iterates and analyzes.","marker":"[24]"}],"fun_headline_variants":["Iterated QED scattering saturates to Bell states","Repeated fermion scattering forces maximal entanglement","Bhabha and Møller iterates hit Bell-state ceiling","Scattering maps send any state to entangled attractor","Quantum electrodynamics maps converge to Bell states"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The argument assumes the scattering matrix can be diagonalized and that a single eigenvalue strictly dominates every other in the infinite-iteration limit for every allowed initial state and regime; if the matrix becomes defective, as it does at the $t=0$ boundary of the Bhabha map, or if two eigenvalues tie in magnitude, the eigenvector expansion fails and the state need not settle on a single Bell state.","fun_headline_variants_meta":{"raw":{"variants":["Iterated QED scattering saturates to Bell states","Repeated fermion scattering forces maximal entanglement","Bhabha and Møller iterates hit Bell-state ceiling","Scattering maps send any state to entangled attractor","Quantum electrodynamics maps converge to Bell states"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000179,"raw_usage":{"total_tokens":1274,"prompt_tokens":891,"completion_tokens":383,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":507,"completion_tokens_details":{"reasoning_tokens":306}},"tokens_in":507,"tokens_out":383,"duration_ms":4120,"temperature":1.0,"reasoning_tokens":306,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T22:30:27.485703+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Evaluate the Bhabha or Møller scattering matrix at momenta and angles where $(-A+D+E+F)^2-16B^2=0$ and check whether the matrix is defective; if numerical iteration of $M$ on a generic initial state then fails to converge to a maximally entangled state, the claimed general saturation would be false.","supporting_citations":[{"cited_title":"Sinha and A","cited_arxiv_id":null,"evidence_quote":"supplies the tree-level QED formalism and the POVM description of sharp momentum filtering that defines the post-measurement helicity state."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"establishes the invariant sets of maximally entangled states for these maps and identifies the entanglophobous states that produce nonmonotonic entanglement."},{"cited_title":"Kowalska and E","cited_arxiv_id":null,"evidence_quote":"introduces the dynamical quantum maps for QED processes and their invariant sets, which the present paper iterates and analyzes."}],"review_version":1}