{"id":"d79a6772-f1ec-45af-9c49-3a12d428909a","arxiv_id":"2509.02383","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A single strong ZZ interaction in the 1D XY/Anderson model yields Lieb-Robinson bounds with a logarithmic lightcone and amplitude suppressed as 1/Δ.","lead":"This paper proves that adding a strong, rare interaction to a disordered quantum chain slows down the spread of information, producing a logarithmic lightcone whose size shrinks as the interaction strength grows. The result gives a rigorous handle on how interactions destabilize Anderson localization in one dimension.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Corollary 3 drops right-side perturbation terms without a bound; they can overlap the kept term via sites 0 and 1, so (104) is not proven as written.","rationale":"The reader's weakest_assumption identifies exactly the same gap: the right-side parts of C1 and C2 are dropped without an explicit bound, despite their evolved supports overlapping the kept terms at sites 0 and 1. My analysis confirms that the overlap is real: the Lieb-Robinson propagation under E1 spreads C1,[1,2] left to site 0 and C1,[-1,0] right to site 1, so the double commutator in the derivative (124) need not vanish. This is the most load-bearing concern because Corollary 3 is the step that converts the general theorem into the claimed 1/Δ logarithmic lightcone for the single-impurity Anderson model; if the dropped terms are not controlled, (104) and hence (63) are unproven. However, the omitted contributions appear to be of the same order (J^2/Δ times t e^{-d/ξ}) as terms already present, so the qualitative claim is likely salvageable by extending the norm bound. This supports the reader's CONDITIONAL verdict: the paper's central idea is plausible, but the proof of the key corollary needs a corrected or expanded estimate. No ad hominem or theatrical framing is needed; the issue is a concrete algebraic omission in the derivation.","tokens_in":21418,"tokens_out":17170,"duration_ms":173019,"concrete_test":"On a small chain (e.g. L=4), take A=σ^z_{-3}, B=σ^z_4, J=1, Δ=1, ω=0, and s=t=1. Numerically or symbolically compute the double commutator [e^{i(E1+E3)s}C1,[1,2]e^{-i(E1+E3)s}, [e^{i(E1+E2)s}C1,[-1,0]e^{-i(E1+E2)s}, e^{itE2}σ^z_{-3}e^{-itE2}]]. If its norm is nonzero, the drop in (113)-(118) and (124) is unjustified. Then check whether this norm is bounded by an expression of the form C(J^2/Δ)t e^{-d/ξ} with a reasonable constant; if so, (104) can be repaired by adding the omitted term, while if not, the logarithmic lightcone claim is in jeopardy.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"In Corollary 3, equations (113) and (117) discard C1,[1,2] and C2,[0,2] from the inner commutator, and (124) asserts 'only the terms of I(s) supported on [-L,0] contribute.' This is not justified. Although A is supported in [-L,-3] and e^{itE2}Ae^{-itE2} stays in [-L,-1], the inner commutator [e^{i(E1+E2)s}C1,[-1,0]e^{-i(E1+E2)s}, e^{itE2}Ae^{-itE2}] has support reaching site 1 through [E1,C1,[-1,0]] (e.g. 2iJ^2 σ^y_{-1}σ^z_0σ^x_1). Meanwhile e^{i(E1+E3)s}C1,[1,2]e^{-i(E1+E3)s} acquires support on site 0 via [E1,C1,[1,2]] = 2iJ^2(σ^x_0σ^z_1σ^y_2 - σ^y_0σ^z_1σ^x_2). Thus the dropped operators and the kept inner commutator have overlapping support at sites 0 and 1, so the double commutator in (124) can be nonzero. The proof gives no estimate for these contributions; the bound (125)-(127) therefore upper-bounds a strictly smaller quantity than the derivative it claims to bound. Consequently the displayed bound (104), and the stronger form (63) relying on the same argument, are not established as written. This is a genuine proof gap, though likely fixable: the missing contributions are of similar order (J^2/Δ times t e^{-d/ξ}), so the qualitative logarithmic lightcone with 1/Δ amplitude may survive with modified constants.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper proves Lieb-Robinson-type bounds with logarithmic lightcones for one-dimensional XY/Anderson spin chains perturbed by a large ZZ coupling Δ. The main mechanism is a general theorem (Theorem 1) bounding the effect of a high-frequency local perturbation in the interaction picture. For the Anderson (random field) case, Lemma 2 gives a single-ZZ bound whose correction is ∝ Δ^{-1} t e^{-d/ξ} for large Δ, and Corollary 3 claims an improved bound when the observed operators are on opposite sides of the impurity. Corollary 4 extends the same mechanism to the clean XY model. The paper also sketches an extension to a sparse set of ZZ terms in Section IV. The central idea is attractive, and the derivation of Theorem 1 is careful and explicit.","tokens_in":21874,"tokens_out":9316,"duration_ms":111085,"significance":"If the main results hold, the paper makes a useful contribution to the rigorous study of interaction-induced delocalization in disordered spin chains. The explicit 1/Δ suppression of the lightcone amplitude is a concrete, falsifiable prediction, and the comparison with the existing result by Gebert, Moon, and Nachtergaele is valuable. The paper also supplies a clean general theorem for high-frequency local perturbations and a remarkably short proof of the L-R bound for nearest-neighbour time-dependent Hamiltonians. However, the paper currently overclaims in two directions: Corollaries 3 and 4 contain a genuine proof gap, and the advertised 'sparse set of ZZ terms' result is not proved as a theorem. The single-ZZ bound of Lemma 2 appears defensible, so the gaps are likely fixable, but the manuscript as written needs revision.","major_comments":[{"comment":"The proof drops the right-side components C1,[1,2] and C2,[0,2] from the inner commutator, and in Eq. (124) it asserts that 'only the terms of I(s) supported on [-L,0] contribute'. This is not justified. The retained inner commutator [e^{is(E1+E2)}C1,[-1,0]e^{-is(E1+E2)}, e^{itE2}Ae^{-itE2}] contains terms whose support extends to site 1, e.g. through [E1, C1,[-1,0]] ∝ J^2 σ^y_{-1} σ^z_0 σ^x_1. The dropped operator e^{is(E1+E3)}C1,[1,2]e^{-is(E1+E3)} acquires support at site 0 through [E1, C1,[1,2]]. These operators therefore have overlapping support at sites 0 and/or 1, so the double commutator in (124) can be nonzero. The bound (125)-(130) estimates a strictly smaller expression than the derivative it claims to bound. Hence Eq. (104) is not proven as written. Corollary 4, whose proof is declared identical, inherits this gap. The missing contributions appear to be of the same order J^2/","section":"Corollary 3, Eqs. (111)-(117), (124)"},{"comment":"The abstract and title promise results for a 'sparse set' of ZZ terms, but the rigorous analysis in Section III concerns a single ZZ term. Section IV is a sketch: no theorem states the sparse L-R bound, the claimed convergence of the sum in Eq. (147) is asserted without a quantitative density/locality condition, and the two-step procedure for separated clusters in Fig. 3 would require an intermediate L-R bound that is not derived. As written, the sparse-interaction claim is not established. The authors should either add a rigorous statement and proof for sparse perturbations or narrow the abstract/title to the single-perturbation results of Lemma 2.","section":"Title, Abstract, Section IV"}],"minor_comments":[{"comment":"The inequality ∥[E,C1]∥≤8J(J+2Ω) is said to 'be checked numerically'. This is a finite computation and should be replaced by an explicit algebraic verification or a clear derivation.","section":"Eq. (99)"},{"comment":"The sentence 'we consider n ∈ N such that . Then' is missing the condition on n. It should read 'such that (n-1)π/(4Δ) < t ≤ nπ/(4Δ)'.","section":"Eq. (94)"},{"comment":"The reference to 'lemma 1' in the proof of Lemma 2 should be 'Theorem 1'. The manuscript uses both names; please make the numbering consistent.","section":"Throughout Section III"},{"comment":"Minor typographical issues: the first displayed line of (62) is missing the factor 2 in front of K in some occurrences, and Eq. (125)-(127) contain stray double brackets such as ∥C1,[−1,0]]∥.","section":"Eq. (62)"},{"comment":"The condition 'dist(suppA, [0,1]) > dist(suppB, [0,1])}' has an extra brace and should be cleaned up.","section":"Corollary 3 statement"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern raised by the reader is valid and localized. I agree that Eq. (124) of Corollary 3 is not justified as written, and the fix is likely to require bounding the dropped right-side contributions rather than claiming they disappear. The sparse-set claim in the title/abstract is also not backed by a proof in Section IV. The single-ZZ bound in Lemma 2 looks sound and is the strongest defensible result of the paper; if the authors repair Corollaries 3/4 and either prove or soften the sparse claim, the paper could be acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The core of this paper is Theorem 1, a general Lieb-Robinson bound for local time-dependent perturbations, and its application to the XY/Anderson chain with a single large ZZ term. The theorem is genuinely new, and the derivation is careful. The resulting bound (63), with amplitude ~ 1/Δ and a logarithmic lightcone, looks correct: it comes from a direct application of the interaction-picture bound (6), using the full C1 and C2, with no terms dropped. That part deserves a serious referee.\n\nWhere the paper gets into trouble is Corollary 3. There the authors try to improve the constant by decomposing the perturbation into left and right pieces and asserting that only the [-L,0] part contributes to the commutator. The stress test is right: e^{i(E1+E3)s} C1,[1,2] e^{-i(E1+E3)s} acquires support on site 0 through [E1, C1,[1,2]], and the kept inner commutator has support reaching site 1, so the dropped operators overlap. No bound is given. So (104) is not proven as written. This is a real gap, though probably patchable: the missing terms are the same order as the kept ones, so the qualitative conclusion survives with adjusted constants.\n\nThe abstract promises 'sparse set of ZZ terms', but Section IV only sketches the many-perturbation extension; there is no theorem for multiple impurities. That is an overclaim in the abstract.\n\nThe paper is worth citing for Theorem 1 and (63). The comparison with Gebert-Moon-Nachtergaele is reasonable and the constants may indeed be better, but I would not take that as established until Corollary 3 is fixed.\n\nI would send it to peer review; the main result is substantial and likely correct, but the referee should ask for a fix of Corollary 3 and a toned-down abstract.","headline":"Solid new theorem and a plausible 1/Δ lightcone result for a single ZZ impurity; the improved Corollary 3 has an unjustified term-dropping step, and the sparse-set claim is only sketched.","tokens_in":22286,"tokens_out":2650,"would_cite":true,"duration_ms":29591,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"A strong local coupling slows quantum spread to a logarithmic lightcone.","keywords":["Lieb-Robinson bounds","Anderson localization","XY model","sparse interactions","logarithmic lightcone","interaction picture","disordered spin chains","quantum dynamics"],"falsifier":"Compute numerically the exact commutator norm for the Hamiltonian (53) on a chain of, say, 30 sites with A on the far left and B on the far right, scanning Δ/J = 1, 2, 4, 8 at fixed disorder samples; the bound predicts the norm decays at least as J/Δ at fixed t and d. If the norm instead grows with Δ, or if the omitted right-side terms such as [e^{isE} C_{1,[1,2]} e^{-isE}, e^{itE_2} A e^{-itE_2}] are not exponentially small in the distance from supp A to the impurity, the central claim fails.","tokens_in":21366,"feed_emoji":"⚛️","tokens_out":7055,"duration_ms":84404,"temperature":0.7,"pith_summary":"The paper proves that adding one very strong ZZ interaction to a one-dimensional XY spin chain—whether or not a random magnetic field is present—does not destroy locality. Instead, it slows the spread of quantum information down to a logarithmic lightcone, with the spreading amplitude suppressed as the inverse of the coupling strength. The result comes from a general theorem for time-dependent local perturbations, applied after moving to the interaction picture, where the large static ZZ term becomes a high-frequency periodic perturbation. In the disordered case, the bound combines the Anderson-localized dynamics of the unperturbed chain with a 1/Δ suppression, so distant regions remain effectively uncommunicating for an exponentially long time in their separation. In the clean XY case the linear lightcone is only logarithmically modified by the large ZZ term, and all bounds are non-perturbative.","feed_headline":"A strong local coupling slows quantum spread to a logarithmic lightcone","feed_subtitle":"For spin chains with or without disorder, one large local coupling shrinks the Lieb-Robinson amplitude to 1/Δ.","key_machinery":"The central mechanism is Theorem 1, a general Lieb-Robinson bound for a local time-dependent one-dimensional spin Hamiltonian E(t) plus a bounded local perturbation λ(at)C. In the fast-perturbation regime the proof moves to the interaction picture with E as the reference dynamics, writes the full evolution as a time-ordered exponential of oscillating terms, and integrates by parts: each integration converts the primitive of the oscillation into a factor 1/a, and the derivative of the interaction-picture operator reintroduces a factor a, leaving an overall 1/Δ suppression. Applied to (53), the interaction picture with respect to the ZZ term turns the neighboring XY bonds into cos(2Δs)C1 + sin","core_discovery":"The central claim is the bound (63) for the Hamiltonian (53): a one-dimensional XY chain with random field ω_j and one added ZZ term of strength Δ on sites 0 and 1. For any operators A and B with fixed bounded supports separated by l and at distance d from the impurity region, the disorder-averaged commutator satisfies E_ω ∥[e^{itH_ω}A e^{-itH_ω}, B]∥ ≤ 2K∥A∥∥B∥ e^{-l/ξ} + 16K(J/Δ + 4J²/Δ²)∥A∥∥B∥ e^{-d/ξ} + 64K∥A∥∥B∥ (J/Δ)(4J/π + Ω) t e^{-d/ξ}. Because the last term is linear in t with amplitude of order J/Δ, it describes a logarithmic lightcone: operators on opposite sides of the impurity remain effectively commuting up to times exponentially large in d, and the amplitude of the cone is sup","pith_inferences":["Inference: the same integration-by-parts mechanism should apply to any large commuting local term whose oscillation frequency is set by its own strength—for instance a strong σ^z field—yielding a similar 1/strength suppression for operators separated across it, though the paper only proves the ZZ case.","Inference: a direct numerical check of the bound would look for the time needed for a left-side operator to develop O(1) commutator with a right-side detector to grow linearly with Δ at fixed separation; observing that scaling would confirm the 1/Δ amplitude, while a growth with Δ would rule it out.","Inference: because corollary 3 drops the right-side-supported pieces of the perturbation when commuting with an operator on the left, an explicit estimate of those omitted terms would either complete the proof with a slightly adjusted constant or show that the bound holds with a different distance factor."],"forward_implications":["If the main bound (63) is correct, operators on opposite sides of the ZZ impurity stay exponentially close to commuting until a time t_max ~ (Δ/J) e^{d/(2ξ)}, which is the defining signature of a logarithmic lightcone.","For large Δ the lightcone amplitude is proportional to J/Δ, so a strong impurity suppresses the disturbance crossing it linearly in 1/Δ even in the absence of disorder.","The same theorem generalizes to sparse sets of well-separated ZZ terms: the overall bound is controlled by the closest perturbation, and the exponential decay in distance degrades to roughly e^{-l/(2ξ)}.","The paper notes that a logarithmic lightcone has been shown to give at most logarithmic growth of dynamical entanglement entropy; the new bound therefore places the sparse-interaction disordered chain in that class.","For the clean XY model, the large ZZ interaction modifies the linear lightcone only by a correction logarithmic in Δ/J rather than replacing it with a slower cone."],"supporting_citations":[{"why":"Supplies the Anderson-localization Lieb-Robinson bound for the unperturbed disordered XY chain, giving the constant K and localization length ξ used throughout.","marker":"[8]"},{"why":"Prior work by the authors provides the stability-of-slow-dynamics theorem that theorem 1 extends, and it yields the Δ ≤ J regime in the informal lemma.","marker":"[5]"},{"why":"Comparison target: its theorem 3.1 and lemma 4.4 give a 1/Δ Lieb-Robinson bound for strong impurities, and the present corollary 3 claims a smaller constant and more general hypotheses.","marker":"[6]"},{"why":"Supplies the linear Lieb-Robinson velocity v_LR used in corollary 4 for the clean XY model.","marker":"[17]"},{"why":"Provides the nearest-neighbour Lieb-Robinson proof method and the velocity bound used in the appendix and in corollary 4.","marker":"[18]"}],"fun_headline_variants":["Logarithmic lightcone from sparse strong interactions","Sparse large coupling slows quantum spread to log cone","One strong sparse bond yields logarithmic Lieb-Robinson cone","Strong sparse coupling shrinks lightcone to logarithmic shape"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The 1/Δ suppression in the main corollary rests on the assumption that, inside the integrated commutator, only the part of the perturbation supported on the left of the impurity contributes to the commutator with an operator A on the left; the right-side pieces are dropped without an explicit bound.","fun_headline_variants_meta":{"raw":{"variants":["Logarithmic lightcone from sparse strong interactions","Sparse large coupling slows quantum spread to log cone","One strong sparse bond yields logarithmic Lieb-Robinson cone","Strong sparse coupling shrinks lightcone to logarithmic shape"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000555,"raw_usage":{"total_tokens":2498,"prompt_tokens":778,"completion_tokens":1720,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":522,"completion_tokens_details":{"reasoning_tokens":1656}},"tokens_in":522,"tokens_out":1720,"duration_ms":15570,"temperature":1.0,"reasoning_tokens":1656,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T11:36:29.738539+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute numerically the exact commutator norm for the Hamiltonian (53) on a chain of, say, 30 sites with A on the far left and B on the far right, scanning Δ/J = 1, 2, 4, 8 at fixed disorder samples; the bound predicts the norm decays at least as J/Δ at fixed t and d. If the norm instead grows with Δ, or if the omitted right-side terms such as [e^{isE} C_{1,[1,2]} e^{-isE}, e^{itE_2} A e^{-itE_2}] are not exponentially small in the distance from supp A to the impurity, the central claim fails.","supporting_citations":[{"cited_title":"Bounds in Nonequilibrium Quantum Dynamics","cited_arxiv_id":"2202.02011","evidence_quote":"Supplies the Anderson-localization Lieb-Robinson bound for the unperturbed disordered XY chain, giving the constant K and localization length ξ used throughout."},{"cited_title":"Our fist goal is to evaluate the first term at the exponent inside the time-ordered operator","cited_arxiv_id":null,"evidence_quote":"Prior work by the authors provides the stability-of-slow-dynamics theorem that theorem 1 extends, and it yields the Δ ≤ J regime in the informal lemma."},{"cited_title":"The overall decay in distance will be of O(e− l ξ )","cited_arxiv_id":null,"evidence_quote":"Comparison target: its theorem 3.1 and lemma 4.4 give a 1/Δ Lieb-Robinson bound for strong impurities, and the present corollary 3 claims a smaller constant and more general hypotheses."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the linear Lieb-Robinson velocity v_LR used in corollary 4 for the clean XY model."},{"cited_title":"Dynamics of many-body localized systems: logarithmic lightcones and $\\log \\, t$-law of $\\alpha$-R\\'enyi entropies","cited_arxiv_id":"2408.02016","evidence_quote":"Provides the nearest-neighbour Lieb-Robinson proof method and the velocity bound used in the appendix and in corollary 4."}],"review_version":1}