{"id":"7268fb49-6f91-4d43-bf7e-afe482e44a1f","arxiv_id":"1908.01818","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Two-excitation subradiant dimer states in a 1D waveguide QED chain can have longer lifetimes than the best single-excitation state.","lead":"Two close-together excited atoms in a chain can store light longer than any single excited atom can. This paper derives the exact math for these slow-decaying 'dimer' pairs and reveals a simple mapping to missing-site defects.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The type-II dimer construction is only valid for k1Dd<π/4, but the paper applies it up to k1Dd=0.4π; for cos(2k1Dd)<0 the printed qII and Eq. (4) fail, so infinite-chain perfect subradiance is not established in that region.","rationale":"The reader's stated weakest assumption is the Born-Markov/linear-dispersion reduction to Eq. (1). That is a standard model assumption and not an internal inconsistency, so I do not treat it as the most load-bearing concern. The most concrete threat to the central claim is that the analytical ansatz for the very states whose subradiance is claimed is not valid on the stated domain. The sign error is corrected in the SM, but the domain restriction is not: once cos(2k1Dd) becomes negative, Eq. (4) is not a probability and qII is undefined with the formulas given. Because the paper uses the infinite-chain analytical solution to identify and target type-II states in finite chains ('Knowing these asymptotic values allows efficient search...'), an undefined or mis-specified qII can lead to misidentification of eigenstates in the region k > π/4. The narrow benchmark near 0.1676π is safe, so this does not overturn the central claim, but it warrants the conditional verdict and explicit domain statements. The reader's rationale did flag the k1Dd > π/4 issue, though not as the weakest assumption, hence partial agreement.","tokens_in":12517,"tokens_out":15226,"duration_ms":168172,"concrete_test":"Re-derive the infinite-chain type-II bound state at k1Dd = 0.3π by solving the characteristic condition e^{2ik1Dd} = A_{-q}/A_q (SM Eq. (14)) with K = π/d and complex q, allowing both branches of the logarithm. Then compare the resulting wavefunction with the finite-N eigenstate identified as the most subradiant type-II dimer at N = 100 and N = 200. If no normalizable q exists, or the overlap with the finite-N eigenstate is not close to 1, the paper's Eq. (4) and the perfect-subradiance statement for k > π/4 are unsupported; if a normalizable branch exists, the paper must state its q and domain explicitly.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Type-II dimer states are central to the headline claim. In the main text they are written as ∑′ e^{iqII Δ/d}|K=π/d;Δ>, with qIId = [π + i ln cos(2k1Dd)]/2 and probability pII(Δ) ∝ (cos 2k1Dd)^{Δ/d} (Eq. (4)). As printed, the imaginary part of qII has the wrong sign: |e^{iqII Δ/d}| = (cos 2k1Dd)^{-Δ/(2d)}, which grows with Δ, so the state is not normalizable. The supplemental expression qII = π − (i/2) ln cos(2k1Dd) repairs the sign but still requires cos(2k1Dd) > 0. Since the paper's stated domain is 0 < k1Dd < π/2, and Figs. 2 and 3 show type-II data up to k1Dd = 0.4π, the analytical construction ceases to be defined for k1Dd > π/4. The claim that 'these dimers are perfectly subradiant with vanishing decay rates on infinite chains' is therefore not supported outside 0 < k1Dd < π/4, and the paper does not state or analyze this restriction. The specific benchmark k1Dd = 0.1676π lies inside the valid region, so the narrow demonstration of dimer subradiance is not destroyed, but the paper's unqualified infinite-chain claim and the portions of Figs. 2–3 beyond π/4 rest on an unstated branch assumption.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript studies two-excitation eigenstates of the effective non-Hermitian Hamiltonian (1) for a chain of two-level emitters coupled to a one-dimensional waveguide. It identifies two families of 'dimer' states, derives exponential separation distributions for their amplitudes, Eqs. (3) and (4), and shows that on infinite chains these states have real eigenvalues, i.e., vanishing decay rates. The authors also present a formal mapping between these confined dimer states and single-excitation states localized around an unoccupied defect site, and they extend the mapping to emitter chains coupled to the three-dimensional free-space field. For finite chains, numerical diagonalization of Heff shows that type-II dimer states can have decay rates below those of the most subradiant fermionic states and even below the most subradiant one-excitation states in a narrow parameter window around k1Dd ≈ 0.1676π, with unusual non-monotonic dependence on emitter number N.","tokens_in":12812,"tokens_out":4997,"duration_ms":48179,"significance":"If the central claims hold, this is a significant result: two-excitation subradiant states that outlive the best single-excitation states challenge a common expectation, and the confinement-localization mapping is elegant and potentially generalizable. The Letter is largely self-contained: the analytical expressions are derived from the stated Hamiltonian without fitted parameters and are compared with direct numerical diagonalization, and the finite-chain benchmark at k1Dd = 0.1676π is clearly identified. However, the infinite-chain claims currently rest on an incorrectly printed exponent and an unstated domain restriction, so the paper needs revision before the result can be fully assessed.","major_comments":[{"comment":"The main-text expression qIId = [π + i ln cos(2k1Dd)]/2 has the wrong sign for the imaginary part. With this sign, |e^{iqII Δ/d}| = (cos 2k1Dd)^{-Δ/(2d)}, which grows with Δ and makes the type-II state non-normalizable. The supplemental expression qII = π − (i/2) ln cos(2k1Dd) repairs the sign and is consistent with the stated probability pII(Δ) ∝ (cos 2k1Dd)^{Δ/d}, but the printed main-text formula contradicts its own Eq. (4). This error is load-bearing because the exponential localization of the type-II dimer is the basis for the infinite-chain eigenvalue ωII = 2Γ1D cot(2k1Dd).","section":"Subradiant dimer excited states, Eq. (4)"},{"comment":"The type-II ansatz with the real (cos 2k1Dd)^{Δ/d} weighting is defined only when cos(2k1Dd) > 0, i.e., 0 < k1Dd < π/4. The paper states the lattice domain as 0 < k1Dd < π/2 and plots type-II data up to k1Dd = 0.4π in Figs. 2 and 3, while the unqualified claim of perfectly subradiant dimers on infinite chains and the EPR discussion at k1Dd = 0.25π rely on the branch where cos(2k1Dd) < 0. The region k1Dd > π/4 must either be explicitly excluded from the infinite-chain analytical statements and the affected figure ranges, or a continuation analysis must be provided. As written, the statement that these dimers are perfectly subradiant on infinite chains is not supported in that region.","section":"Subradiant dimer excited states and Figs. 2–3"},{"comment":"The equivalence H^{π/d}_{def}(k1Dd) ≅ H^{0}_{def}(2k1Dd) used to derive the type-II states inherits the same restriction on 2k1Dd. The supplemental derivation of qII and ωII is therefore valid only for 0 < k1Dd < π/4. The paper should state this restriction explicitly where the type-II asymptotic eigenvalue is quoted, since the unqualified 'vanishing decay rates on infinite chains' statement appears in the main text before the defect mapping is introduced.","section":"Confinement-localization mapping and Supplemental Sec. H"}],"minor_comments":[{"comment":"In the sentence introducing Eq. (4), the prime on the summation is said to include only even values of Δ/d, but the notation is not defined until later; please define the prime explicitly at first use.","section":"Subradiant dimer excited states"},{"comment":"The caption reads 'emitter number' in the axis label; please write 'emitter number N' or simply 'N' to avoid ambiguity with the separation Δ.","section":"Fig. 2 caption"},{"comment":"Ref. [31] is cited as 'Supplemental Material' without an arXiv or journal anchor; please state where the supplement is available or include it as a clearly labelled appendix.","section":"References"},{"comment":"The phrase 'unwritten orthodoxy' is informal for a Letter; consider replacing it with 'common expectation' or 'common assumption'.","section":"Conclusions"},{"comment":"The right-hand side of Eq. (1) is an operator expression, but the ordering of σ_m^† and σ_n is conventional; please specify that this is the Born-Markov reduced Hamiltonian with normal ordering and state any sign convention explicitly.","section":"Spin Model, Eq. (1)"}],"recommendation":"major_revision","confidential_remarks":"The main concern is not circularity or fitted parameters—the analytical construction is genuine and the numerics are consistent—but the manuscript's unqualified infinite-chain claims and the plotted range in Figs. 2–3 exceed the domain in which the type-II ansatz is defined. The sign error in qII is a straightforward fix, but the domain restriction requires either a careful continuation analysis or an explicit restriction of the claims and figures. I would be comfortable with a revised version that addresses these points."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe paper is worth reading. It gives exact analytical eigenstates for two-excitation subradiant dimers in a 1D waveguide QED chain and shows numerically that type-II dimers can decay slower than the best single-excitation states. The confinement-localization mapping (dimer to defect) is a genuinely new technique, and the analytical eigenvalues match exact diagonalization. No free parameters, no circularity; the derivation is transparent.\n\nThe soft spots are real but localized. The printed qII in the main text has the wrong sign for the imaginary part: as written, the type-II ansatz grows with separation rather than decaying. The supplemental expression repairs the sign. Relatedly, the main-text derivation of type-II assumes cos(2k1Dd)>0, i.e., k1Dd<π/4, while the paper states the domain 0<k1Dd<π/2 and plots data up to 0.4π. The infinite-chain claim of perfect subradiance is therefore not established in the text for k>π/4. A careful reader can repair this by choosing the appropriate branch of ln cos, and the numeric data beyond π/4 is likely fine, but the paper should state the restriction or the branch choice explicitly.\n\nThe central result—the narrow subradiant region near k=0.1676π with lifetimes exceeding single-excitation states—lies safely inside the valid domain, and the numerics don't depend on the analytic expression. So the main claim holds. The Born-Markov assumption is standard and not a serious concern. The self-citation to Ref. [18] is appropriate: the numerical discovery came from there, but the analytical work here is new.\n\nThis is for people working on subradiance and quantum memory in waveguide QED; the mapping might also interest those studying quasiparticle confinement in other lattice models. Recommendation: send to peer review. A competent referee will ask for a corrected qII and a clarified domain of validity, but the paper deserves that round.","headline":"A solid waveguide-QED paper with a real analytical technique and a genuine subradiant-dimer claim; the printed qII sign is wrong and the k>π/4 domain needs clarification, but the central result holds.","tokens_in":13359,"tokens_out":8931,"would_cite":true,"duration_ms":83060,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.50.Nn"],"model":"deepseek-v4-flash","headline":"Dimer states of two nearby excitations in an emitter chain coupled to a 1D waveguide can decay more slowly than even the most subradiant single-excitation state, and become perfectly subradiant in the infinite-chain limit.","keywords":["subradiance","emitter chains","one-dimensional waveguide","two-excitation dimers","confinement-localization mapping","defect-localized states","collective decay","quantum optics"],"falsifier":"Prepare an $N=48$ emitter chain with spacing $k_{1\\mathrm{D}}d=0.1676\\pi$, initialize the type-II dimer (two excitations separated by $2d$ with center-of-mass quasimomentum $\\pi/d$), and measure the decay rate of the two-excitation population alongside the most subradiant one-excitation state; if the dimer decays faster, the central claim is wrong. The prediction is sharpest at $k_{1\\mathrm{D}}d=\\pi/6$, where the dimer rate should fall faster than $N^{-3}$ as $N$ grows.","tokens_in":12297,"feed_emoji":"⚛️","tokens_out":8311,"duration_ms":76770,"temperature":0.7,"pith_summary":"The paper claims that two nearby excited emitters in a chain coupled to a one-dimensional waveguide can form subradiant dimer states whose collective decay is slower than that of any state with a single excitation. This contradicts the common expectation that more excitations necessarily decay faster, and it shows that emitter saturation does not merely perturb few-excitation physics but can create a new protection mechanism. For infinite chains the dimers have exactly zero decay rate in the effective model, with analytical expressions for their energies and spatial profiles. The authors explain the mechanism through a mapping that identifies a dimer with a single excitation localized around an unoccupied site, and they extend the same argument to chains emitting into three-dimensional free space, where the dimers keep finite but strongly suppressed decay rates.","feed_headline":"Two-excitation dimers can outlive the longest-lived single excitations","feed_subtitle":"Two nearby excited emitters can decay slower than any single-excitation state in a 1D waveguide chain.","key_machinery":"The argument is carried by the effective non-Hermitian Hamiltonian $H_{\\mathrm{eff}}$ plus a confinement-localization mapping. Acting on dimer basis states $|K;\\Delta\\rangle$ with center-of-mass wave number $K$ and separation $\\Delta$, $H_{\\mathrm{eff}}$ reduces for infinite chains to a matrix $H^K_{\\Delta,\\Delta'}$ acting only on the relative coordinate. The paper proves that eigenstates of $H^K$ with positive separations are in one-to-one correspondence with even-parity eigenstates of a defect Hamiltonian $H^K_{\\mathrm{def}}$, which describes a single excitation localized around an empty site. Since an excited emitter blocks a second excitation on the same site, each excitation acts as a defect that localizes the other, explaining the dimer's stability. The mapping also shows that type-II dimers at $K=\\pi/d$ are equivalent to type-I dimers at $K=0$ after replacing $d$ by $2d$ and applying alternating sign flips, and it extends by linearity to Hamiltonians of the form $\\int d\\mu(k_{1\\mathrm{D}})H_{\\mathrm{eff}}(k_{1\\mathrm{D}})$, covering coupling to 3D free-space modes.","core_discovery":"The paper establishes that the non-Hermitian Hamiltonian $H_{\\mathrm{eff}} = -\\frac{i}{2}\\Gamma_{1\\mathrm{D}}\\sum_{m,n}e^{ik_{1\\mathrm{D}}|z_m-z_n|}\\sigma_m^\\dagger\\sigma_n$ has, in the two-excitation sector, two families of subradiant dimer eigenstates. Type-I dimers have center-of-mass wave number $K\\approx0$ and dominant separation $\\Delta=d$; type-II dimers have $K\\approx\\pi/d$ and dominant separation $\\Delta=2d$. For infinite chains, these eigenvalues are real: $\\omega_\\mathrm{I}=2\\Gamma_{1\\mathrm{D}}\\cot(k_{1\\mathrm{D}}d)$ and $\\omega_\\mathrm{II}=2\\Gamma_{1\\mathrm{D}}\\cot(2k_{1\\mathrm{D}}d)$, with separation distributions $p_\\mathrm{I}(\\Delta)\\propto(\\cos k_{1\\mathrm{D}}d)^{2\\Delta/d}$ and $p_\\mathrm{II}(\\Delta)\\propto(\\cos2k_{1\\mathrm{D}}d)^{\\Delta/d}$ on even separations. On finite chains the decay rates are small but nonzero, and exact diagonalization shows that for $k_{1\\mathrm{D}}d\\simeq0.1676\\pi$ a chain of $N=48$ emitters already supports a type-II dimer whose minimal decay rate is smaller than that of the most subradiant fermionic two-excitation state and of the most subradiant one-excitation state, whose rate at this spacing scales as $N^{-3}$.","pith_inferences":["The defect picture suggests a direct experimental test of the mechanism: compare the decay of a single excitation next to a missing site with that of a dimer with the same effective separation; the mapping predicts matching exponential suppression governed by the shorter side of the chain.","The sharp dip at $k_{1\\mathrm{D}}d=\\pi/6$ may reflect a commensurability between the dimer bond length and the resonant wavelength, so varying the chain length modulo four emitters (a half-wavelength) could be used to steer the decay rate in finite systems.","If the confinement-localization equivalence holds in other lattice models with long-range hopping, two-particle bound states in cold-atom, ion, or Rydberg arrays might be designed to inherit defect-mode lifetimes, turning 'anti-blockade' into a resource rather than a loss channel.","One could test the non-Markovian sensitivity by repeating the exact diagonalization with a finite-bandwidth waveguide dispersion; the dimer decay rates should generically acquire corrections, and the perfect subradiance at $k_{1\\mathrm{D}}d=\\pi/6$ would be the first quantity to move."],"forward_implications":["A finite chain of about 50 emitters with spacing $k_{1\\mathrm{D}}d\\simeq0.1676\\pi$ should already show a two-excitation state decaying more slowly than the best single-excitation state, giving an experimentally accessible signature of the effect.","At $k_{1\\mathrm{D}}d=\\pi/6$ the type-II dimer decay rate dips sharply and falls faster than $N^{-3}$, so increasing chain length can produce a qualitatively stronger lifetime boost than the standard one-excitation scaling.","At $k_{1\\mathrm{D}}d=\\pi/4$ the type-II dimer becomes an eigenstate of both total momentum and relative position, a travelling EPR-like two-excitation state with no amplitude for separations beyond $2d$.","Because the mapping is linear, chains coupled to 3D free-space vacuum fields also support subradiant dimers; their decay rates remain finite in the infinite-chain limit but are strongly suppressed, for both transverse and parallel emitter polarizations.","Proposed verification: excite emitters around a missing or suitably perturbed site to address the single-defect localized state, or use interactions to prepare correlated dimers; the long-lived state should survive as orthogonal modes decay."],"supporting_citations":[{"why":"supplies the Born-Markov elimination of the waveguide modes that produces the effective spin Hamiltonian Eq. (1) used throughout.","marker":"[33]"},{"why":"earlier work that identified fermionic multi-excitation subradiant states and the Tonks-Girardeau equivalence, the baseline these dimer states are compared against and extend.","marker":"[18]"},{"why":"defines the one-excitation subradiant states and fermionic-state decay behavior that the type-II dimers are shown to beat.","marker":"[15]"},{"why":"gives the infinite-range spin-spin interaction Hamiltonian form in Eq. (1), the model whose two-excitation spectrum is analyzed.","marker":"[34]"},{"why":"supplemental material containing the analytic derivations of the asymptotic dimer eigenvalues, the confinement-localization mapping, and the disorder simulations.","marker":"[31]"}],"fun_headline_variants":["Two-excitation dimers beat single-photon decay in 1D chains","Subradiant dimers in 1D waveguides outlive single excitations","Pairs of excited emitters beat single-photon decay in 1D chains","Close emitter pairs decay slower than any single excitation"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim rests on the effective model in which the waveguide is eliminated to give emitters instantaneous, distance-dependent interactions and decay (the Born-Markov approximation with a linear dispersion); if non-Markovian retardation or band curvature is significant, the predicted vanishing dimer decay rates are not guaranteed.","fun_headline_variants_meta":{"raw":{"variants":["Two-excitation dimers beat single-photon decay in 1D chains","Subradiant dimers in 1D waveguides outlive single excitations","Pairs of excited emitters beat single-photon decay in 1D chains","Close emitter pairs decay slower than any single excitation"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000663,"raw_usage":{"total_tokens":3043,"prompt_tokens":973,"completion_tokens":2070,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":589,"completion_tokens_details":{"reasoning_tokens":2001}},"tokens_in":589,"tokens_out":2070,"duration_ms":15803,"temperature":1.0,"reasoning_tokens":2001,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T15:03:40.810884+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Prepare an $N=48$ emitter chain with spacing $k_{1\\mathrm{D}}d=0.1676\\pi$, initialize the type-II dimer (two excitations separated by $2d$ with center-of-mass quasimomentum $\\pi/d$), and measure the decay rate of the two-excitation population alongside the most subradiant one-excitation state; if the dimer decays faster, the central claim is wrong. The prediction is sharpest at $k_{1\\mathrm{D}}d=\\pi/6$, where the dimer rate should fall faster than $N^{-3}$ as $N$ grows.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the Born-Markov elimination of the waveguide modes that produces the effective spin Hamiltonian Eq. (1) used throughout."},{"cited_title":"Asenjo-Garcia, M","cited_arxiv_id":null,"evidence_quote":"defines the one-excitation subradiant states and fermionic-state decay behavior that the type-II dimers are shown to beat."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"gives the infinite-range spin-spin interaction Hamiltonian form in Eq. (1), the model whose two-excitation spectrum is analyzed."}],"review_version":1}