{"id":"0bc12750-40bf-4743-8159-764afe289364","arxiv_id":"2411.09039","paper_version":2,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Linear polaritonic spectra contain finite-size (1/N) vacuum-induced Raman sidebands, observable when the cavity decay rate is comparable to the single-molecule coupling.","lead":"The paper shows that linear absorption spectra of molecules strongly coupled to a cavity harbor tiny nonlinear Raman sidebands, invisible in typical experiments but growing in very-low-loss single-mode cavities with few molecules. This gives experimentalists a concrete place to look for quantum cavity effects that classical optics cannot reproduce.","discovery_kind":"first_principles","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. (8), the central first-order correction, is not verified against the exact continued fraction, and the printed Eq. (6)/S18 has a dagger-ordering inconsistency that makes the derivation as printed non-reproducible.","rationale":"The reader's CONDITIONAL verdict is reasonable: the physical picture and the algebraic structure of Eq. (8) are plausible, but the proof is not fully rigorous and no numerical check is shown. I agree with the reader that the idealizations (zero temperature, no disorder, single mode) are an important caveat for the observability claim. However, my most load-bearing concern is more basic: the central derivation is not reproducible as printed because of the dagger-ordering inconsistency between Eq. (6)/S18 and the derivation of Eq. (8) in Sec. S6b. If Eq. (8) does not actually reproduce the exact Green's function to O(1/N), the predicted 1/N Raman sidebands would not be established. This is a concrete, checkable issue: a direct numerical comparison for small N would settle it. If the check passes, the paper's core theoretical claim stands and the remaining issues are presentation and experimental idealization; if it fails, the central claim would need to be rejected or substantially revised. Therefore I do not move the verdict away from CONDITIONAL; I keep it unchanged while adding a sharper condition.","tokens_in":17830,"tokens_out":40834,"duration_ms":397094,"concrete_test":"For a small ensemble (e.g., N=3) of three-level molecules with the parameters of Fig. 2(a), compute the exact retarded photon Green's function DR_N(omega) by numerically solving (omega - H1 + iGamma) G = 1 for the block-tridiagonal Hamiltonian in Eq. (S3), with H1 constructed from Eq. (1). Then compute dN,0(omega) + dN,1(omega) from Eqs. (7) and (8). If the difference is not O(N^-2) at frequencies away from degenerate poles, Eq. (8) is incorrect. Also recompute the continued fraction exactly as printed in Eq. (6): if it disagrees with the direct matrix inversion, the printed equation contains a dagger-ordering error that must be corrected before the derivation can be reproduced.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central result of the paper, Eq. (8) for dN,1, is derived from the continued fraction in Eq. (6) and Supplemental Eq. (S18). But as printed, that continued fraction is internally inconsistent. In the block matrix of Eq. (S3), v0 is the (He,0 ; Hph,1) block, so v0 maps Hph,1 to He,0. The correct self-energy of He,0 due to Hph,1 is then v0 (omega - Hph,1 - ...)^-1 v0-dagger, exactly as used in Eq. (8). Yet Eq. (6) and S18 place the daggers on the opposite factors, writing v0-dagger (...) v0 inside the denominator for He,0, and similarly V1-dagger (...) V1 inside the denominator for Hph,1. With the stated definitions, those expressions are not valid operators on the subspaces where they appear. The derivation in Supplemental Sec. S6b silently switches to the correct ordering when it writes v0 Gph,1 (...) v0-dagger Ge,0, so Eq. (8) may be correct, but it does not follow from the displayed Eq. (6). Because the existence and line shape of the predicted Raman sidebands rest entirely on Eq. (8), the paper needs an independent check that this formula is the true O(1/N) term of the exact Green's function; no such numerical comparison is currently provided.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript asks why molecular polaritons' linear spectra are so well described by classical linear optics, despite the expected nonlinear character of strong coupling. Using a bosonic (Holstein–Tavis–Cummings) model restricted to the first excitation manifold, the authors express the photon Green's function as a matrix continued fraction (Eq. 6) over the block-tridiagonal 'CUT-E' chain, in which V-type couplings are collective (∝ λ√N) and v-type couplings are single-molecule Raman processes (∝ λ). They then develop a 1/N expansion of the Green's function. The zeroth order reproduces the classical linear-optics formula in terms of the ensemble linear susceptibility; the first-order term, dN,1 (Eq. 8), is interpreted as a vacuum-mediated Stokes–anti-Stokes process and predicts Raman sidebands at ±ωv around the polaritons, with height ~1/N and reduced first-order Rabi splitting 2λ√(N−1). The authors argue these sidebands are washed out in low-Q and multimode cavities but should be observable in high-Q single-mode cavities with κ ~ λ. They also classify the Dyson diagrams into reducible and irreducible contributions and claim (Eq. 9) that the photon self-energy equals the sum of irreducible odd-order nonlinear susceptibilities of the molecular ensemble, with O(N^{-2}) terms generating two-molecule Raman sidebands.","tokens_in":18112,"tokens_out":30742,"duration_ms":237749,"significance":"The central claim is significant and cleanly falsifiable: linear spectroscopy of high-Q single-mode molecular polaritons should reveal finite-size Raman sidebands with a specific N-scaling, providing a molecule-specific signature of nonlinear response mediated by vacuum fluctuations. If Eq. (8) holds, the paper resolves an active conceptual puzzle (polaritons as 'optical filters') by showing where the nonlinearity goes in the N→∞ limit. Strengths: the derivation is self-contained and parameter-free; the N-counting is consistent throughout; and the reduction to classical optics in the thermodynamic limit is a genuine consistency check rather than a circular assumption. The weaknesses are local but load-bearing: (i) the printed continued fraction (Eq. 6 and SI Eqs. S18, S22) uses a dagger ordering opposite to that of Eq. (5b), so Eq. (8) does not follow from the displayed equations as written; (ii) no numerical comparison of the truncated 1/N expansion with the exact Green's function is provided; and (iii) the proof that the self-energy equals the sum of irreducible susceptibilities (Eq. 9) is delegated to the reader. These issues are repairable within the scope of the manuscript.","major_comments":[{"comment":"The continued fraction printed in Eq. (6), and repeated in SI Eqs. (S18) and (S22), is not the iteration of the recursion in Eq. (5b). Since v0 is the (He,0; Hph,1) block in Eq. (S3), the self-energy on He,0 from Hph,1 is v0(ω − Hph,1 + ...)^{-1}v†0, and the self-energy on Hph,1 from He,1 is V1(ω − He,1 + ...)^{-1}V†1; Eq. (5b) states exactly this. Eq. (6) instead writes v†0(...)v0 and V†1(...)V1 in those denominators. Because Eq. (8) and the SI's Dyson terms (e.g., V0Ge,0v0Gph,1v†0Ge,0V†0 in the fourth-order table of Sec. S5) use the opposite ordering, Eq. (8) cannot be obtained from Eq. (6) as printed; the conversion in Sec. S6b silently switches to the correct ordering when it introduces X = v0Gph,1(I − V1Ge,1V†1Gph,1)^{-1}v†0Ge,0. The ordering is immaterial in the fully scalar 3-level model used in Fig. 2, which may explain the oversight, but the manuscript claims validity for arbitrary vibronic manifolds in which the blocks are non-commuting matrices. Please correct the dagger placement in Eq. (6), Eq. (S18), and Eq. (S22), or state the alternative convention explicitly and make Eq. (5b), the diagram rules, and Eq. (8) consistent with it.","section":"Eq. (6); SI Eqs. (S18), (S22)"},{"comment":"No comparison is made between the 1/N-truncated Green's function and the exact DR_N. The exact result is readily available either by numerically solving the finite tridiagonal system defined by Eq. (5) (or the continued fraction once its ordering is fixed) or by exact diagonalization of Eq. (2) for small N. Such a comparison would serve two purposes: it would independently confirm that Eq. (8) is indeed the O(1/N) term of the exact photon Green's function — essential given the ordering inconsistency in Eq. (6) — and it would demonstrate that the 1/N expansion actually converges for the N values plotted in Fig. 2. I recommend adding a panel overlaying the exact and truncated absorption spectra for a few values of N (e.g., N=5–50) at fixed λ√N, including an error estimate for the truncated series.","section":"Figs. 2a–2b; SI Sec. S6"},{"comment":"The paper's title-level claim, that linear polaritonic spectra are governed by hidden nonlinear molecular susceptibilities, is carried by Eq. (9). The supporting argument equates the continued fraction (S18) with the nested matrix-geometric series (S19), but the SI only says that 'this result can be explicitly checked by the reader'; no induction, convergence statement, or justification of the order of non-commuting factors in the nested powers of Eq. (S19) is given. Furthermore, with Eq. (S18) exactly as printed, the two expressions are not equal as operator identities in the general multi-level case (see major comment 1 on the dagger ordering), so the equality cannot simply be taken on trust. I ask the authors to replace the sketch with a direct derivation of Eq. (9), e.g., an induction on the truncation level of the continued fraction, together with a statement of the sense in which the infinite sums converge.","section":"Eq. (9); SI Sec. S5A, Eqs. (S18)–(S19)"}],"minor_comments":[{"comment":"In the opening sentence of Sec. S6c, 'To obtain dN,2(ω) ∝ N^{−k}' should read 'dN,2(ω) ∝ N^{−2}'.","section":"SI Sec. S6c"},{"comment":"The caption to Fig. 2 does not identify which colors in panel (b) correspond to the zeroth- and first-order spectra, and it is unclear whether the blue curve shows the O(N^{-2}) increment alone or the sum of all included orders; please clarify.","section":"Fig. 2 caption"},{"comment":"The statement that 'we cannot provide an expression for dN,k when k ≥ 2' is stronger than what the Supplemental Material achieves, since Sec. S6c gives d^{(1)}_{N,2} in closed form; the wording should be changed to 'no complete closed-form expression for dN,k is obtained for k ≥ 2'.","section":"Main text, '1/N expansion'"},{"comment":"The observability prediction assumes a homogeneous ensemble at zero temperature with Markovian losses and a single cavity mode; the text acknowledges the multimode-cavity blurring but not the possible suppression of the 1/N sidebands by static disorder or finite-temperature phonon populations. A brief quantitative estimate (or an explicit statement that these effects are beyond the present scope) would make the experimental claim more complete.","section":"Experimental considerations"},{"comment":"The index placement in Eqs. (S17) is inconsistent (e.g., µ_{0g}^{me} versus µ_{me 0g}), which makes the susceptibility expressions difficult to parse; a uniform tensor notation consistent with Eq. (S.2) is recommended.","section":"SI Eqs. (S17a)–(S17c)"},{"comment":"The equivalence of the CUT-E diagram rules with the ket-only Dyson diagrams is asserted with the parenthetical promise of an induction proof that is not given; please either supply the induction or state the induction hypothesis explicitly.","section":"SI Sec. S5"},{"comment":"Since the block-tridiagonal (CUT-E) structure and the 1/N expansion were introduced in Ref. [23] by the same group, the authors should state explicitly which elements (the linear-spectroscopy application, the Raman-sideband prediction, and the susceptibility identification) are new in the present work.","section":"Introduction / Ref. [23]"}],"recommendation":"major_revision","confidential_remarks":"The ordering inconsistency in Eq. (6) appears to be a transcription/typesetting error rather than a conceptual one, because Eq. (5b) and the diagrammatic expansion are mutually consistent and produce Eq. (8) with the ordering printed there; the figures use the scalar 3-level model where the ordering cannot matter. I nevertheless regard the missing numerical check of Eq. (8) as a hard requirement for acceptance, since it is the only way to certify the central formula independently of the displayed derivation. The paper shares the CUT-E framework and the 1/N idea with the companion Ref. [23] (arXiv:2410.14175); the novelty boundary should be drawn explicitly. The self-citation rate is high but each citation is directly relevant to the method or the problem."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The headline: this paper likely contains a real and interesting result—the linear response of molecular polaritons has finite-size 1/N corrections that show up as vacuum-mediated Raman sidebands, becoming visible when the cavity linewidth κ is comparable to the single-molecule coupling λ. That would explain why classical optics works in low-Q cavities while still leaving room for nonlinear effects at strong coupling.\n\nWhat's new: the 1/N expansion of the exact polaritonic Green's function, the identification of the O(1/N) term as a sequential Stokes-anti-Stokes Raman process mediated by the vacuum field, and the diagrammatic classification in terms of irreducible nonlinear susceptibilities. The derivation of dN,0 and dN,1 from the continued fraction is clean, and the thermodynamic limit correctly reduces to the standard linear susceptibility—a good consistency check. The physical picture of the timescale separation (collective O(λ√N) Rayleigh vs single-molecule O(λ) Raman transitions) is persuasive.\n\nThe soft spots are real but addressable. First, Eq. (6) as printed has a dagger-ordering inconsistency: it writes v†0 (...) v0 inside the denominator of H_e,0, but the block structure of the Hamiltonian requires v0 (...) v†0. The supplemental derivation in Sec. S6b silently uses the correct ordering when it derives Eq. (8), so Eq. (8) is probably right, but it does not follow from the displayed Eq. (6). That typo needs fixing before the derivation is reproducible from the text. Second, the proof that the self-energy equals the sum of irreducible nonlinear susceptibilities (Eq. 9) is sketched rather than rigorous, with a \"reader can check\" signpost. Third, there is no numerical check of Eq. (8) against the exact continued fraction, even for a few molecules. Since the whole prediction of observable sidebands rests on that formula, an independent numerical comparison is the most important missing piece. Finally, the text contradicts itself about whether closed-form expressions exist for k ≥ 2, and the observable prediction assumes zero temperature, no disorder, and a single cavity mode; the authors acknowledge multimode blurring but do not quantify the suppression.\n\nBottom line: this is a solid theoretical letter with a novel, credible claim that deserves a serious referee. I would ask for a corrected Eq. (6), a numerical verification of Eq. (8), and a clearer statement about higher-order terms. With those, it would be a nice contribution for the molecular polaritonics community.","headline":"A credible 1/N expansion showing vacuum-mediated Raman sidebands as finite-size corrections to polaritonic linear spectra, but the printed continued fraction has a dagger-ordering typo and the central formula lacks a numerical check.","tokens_in":18643,"tokens_out":8243,"would_cite":true,"duration_ms":62702,"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":"Molecular polariton linear spectra carry a hierarchy of finite-size 1/N quantum corrections from vacuum-mediated Raman processes, visible when cavity loss is comparable to the single-molecule coupling.","keywords":["molecular polaritons","strong light-matter coupling","1/N expansion","vacuum-mediated Raman scattering","linear response","photon Green's function","high-Q cavities","nonlinear susceptibilities"],"falsifier":"Measure the linear absorption or transmission of a high-Q single-mode microcavity containing a known number N of identical molecules at low temperature, with κ ≈ λ and the collective coupling λ√N kept fixed while N is varied. The paper predicts sidebands offset from each zeroth-order polariton by the ground-state vibrational frequency, with peak height ∝1/N and blue shift Δ = 2λ(√N − √(N−1)). If those sidebands are absent, do not scale with N, or persist unchanged in a multimode cavity, the central claim fails.","tokens_in":17618,"feed_emoji":"🔬","tokens_out":6100,"duration_ms":52562,"temperature":0.7,"pith_summary":"Molecular polaritons — N molecules strongly coupled to one cavity photon mode — have linear spectra that are usually captured by classical linear optics. This paper argues that this classical description is only the thermodynamic-limit term of an exact 1/N expansion, and that the first quantum correction contains vacuum-mediated molecular Raman processes that produce weak sidebands around every polariton peak. Those sidebands are hidden in ordinary low-Q cavities because the cavity linewidth κ vastly exceeds the single-molecule coupling λ, but they should become visible in high-Q single-mode cavities with κ ≈ λ. If true, this explains where the expected nonlinear many-body effects of strong coupling live in linear spectroscopy and gives a concrete experimental regime in which to observe them.","feed_headline":"High-Q cavities reveal hidden Raman sidebands in polariton spectra","feed_subtitle":"Finite-size 1/N corrections from vacuum-mediated Raman processes become visible when cavity loss matches single-molecule coupling.","key_machinery":"The central object is the block-tridiagonal representation of the single-excitation Hamiltonian, a nearest-neighbor chain of photonic and molecular blocks that the paper calls the CUT-E diagram. This chain separates fast collective couplings $V_n\\propto\\lambda\\sqrt{N}$ — Rayleigh processes that conserve the number of ground-state phonons — from slow single-molecule couplings $v_n\\propto\\lambda$ — Raman processes that create or destroy one vibrational quantum. Matrix continued-fraction techniques applied to this chain produce the photon Green's function, and the timescale separation organizes the $1/N$ expansion that isolates the hidden sidebands.","core_discovery":"The paper derives the exact photon Green's function $D^R_N(\\omega)$ for a single-mode cavity coupled to N vibronic molecules as a continued fraction (Eq. 6) and expands it in powers of $1/N$. The zeroth-order term $d_{N,0}$ reproduces classical linear optics: the self-energy is just the molecular linear susceptibility $\\chi^{(1)}$, and in the limit $N\\to\\infty$ nothing else survives. The first-order term $d_{N,1}$ (Eq. 8) is the new claim: it factorizes into $d_{N,0}\\cdot V_0G_{e,0}v_0\\bigl(\\omega-H_{\\mathrm{ph},1}+i\\kappa/2-V_1G_{e,1}V_1^\\dagger\\bigr)^{-1} v_0^\\dagger G_{e,0} V_0^\\dagger\\cdot d_{N,0}$, which describes a Stokes Raman transition, propagation through first-order polaritons formed by the remaining $N-1$ molecules, and an anti-Stokes transition back. These vacuum-mediated processes are each penalized by $1/\\sqrt{N}$, so the sidebands have height $\\propto 1/N$ and blue shift $\\Delta = 2\\lambda(\\sqrt{N}-\\sqrt{N-1})$. The same continued fraction is rewritten (Eq. 9) as the photon self-energy equal to a sum of irreducible Rayleigh and Raman nonlinear susceptibilities of the molecular ensemble, making the hidden nonlinear character of the linear spectrum explicit.","pith_inferences":["If the sidebands survive at the predicted $1/N$ height, a careful measurement of sideband intensity versus $N$ would give a direct estimate of the single-molecule coupling $\\lambda$, which is otherwise hard to isolate in the collective strong-coupling regime.","The two-species sum-frequency sideband at $\\omega_{v,A}+\\omega_{v,B}$ suggests using linear transmission as a probe of cooperative Raman events between distinct molecular species, a measurement that does not require a pulsed nonlinear experiment.","The paper's own remark that multimode cavities blur the sidebands implies the observable doubles as a diagnostic: the presence of sharp sidebands certifies effective single-mode behavior, while their absence could indicate multimode or disordered broadening.","Because the first-order sidebands are genuine Raman transitions driven by vacuum fluctuations, one might expect weak nonclassical photon statistics or entanglement in the transmitted field at those frequencies; the paper leaves this as future work."],"forward_implications":["In the thermodynamic limit ($N\\to\\infty$) the polariton linear response becomes exactly the classical linear-optics result built from the molecular linear susceptibility, so no nonlinearity is visible.","In low-$Q$ cavities ($\\kappa\\gg\\lambda$) the $1/N$ terms are suppressed both by the timescale separation and by limited spectral resolution, which is why transfer-matrix and effective-medium models work so well.","In high-$Q$ single-mode cavities with $\\kappa\\sim\\lambda$, each zeroth-order polariton acquires sidebands shifted by the ground-state vibrational frequency $\\omega_v$; their height scales as $1/N$ and their Rabi splitting is $2\\lambda\\sqrt{N-1}$.","The photon self-energy can be written as a sum over irreducible Rayleigh and Raman nonlinear susceptibilities, so a linear measurement in this regime directly carries nonlinear molecular response information.","For two molecular species, the $O(N^{-2})$ terms produce sidebands at $2\\omega_{v,A}$, $2\\omega_{v,B}$, and $\\omega_{v,A}+\\omega_{v,B}$, revealing collective Raman processes involving different molecules."],"supporting_citations":[{"why":"Supplies the linear-response formalism and absorption/transmission expressions that the paper builds on.","marker":"[16]"},{"why":"Establishes the classical linear-optics description of strongly coupled organic polaritons that the 1/N expansion refines.","marker":"[14]"},{"why":"Raises the motivating question by showing molecular polaritons can behave like optical filters.","marker":"[9]"},{"why":"Provides the block-tridiagonal CUT-E representation and the timescale separation underlying the 1/N expansion.","marker":"[23]"},{"why":"Supplies the vibronic polariton spectroscopy framework with dark and bright vibronic states.","marker":"[25]"},{"why":"Defines the nonlinear susceptibilities and irreducible diagrammatic concepts used in Eq. 9.","marker":"[30]"},{"why":"Reports previous small-N spectra showing the first-order polariton features that the paper explains as 1/N corrections.","marker":"[31]"},{"why":"Supports the statement that multimode cavities blur the predicted sidebands into continua.","marker":"[37]"}],"fun_headline_variants":["Cavity loss unlocks Raman sidebands in polariton spectra","1/N corrections expose Raman fingerprints in polariton linear spectra","Vacuum-mediated Raman sidebands emerge when loss equals coupling","High-Q cavities turn linear polariton spectra into Raman maps"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The prediction that the sidebands are visible assumes a perfectly identical set of molecules at zero temperature with only simple Markovian loss and a single cavity mode; if static disorder, thermal vibrations, or extra cavity modes blur the resonances, the 1/N features could be washed out.","fun_headline_variants_meta":{"raw":{"variants":["Cavity loss unlocks Raman sidebands in polariton spectra","1/N corrections expose Raman fingerprints in polariton linear spectra","Vacuum-mediated Raman sidebands emerge when loss equals coupling","High-Q cavities turn linear polariton spectra into Raman maps"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000255,"raw_usage":{"total_tokens":1585,"prompt_tokens":969,"completion_tokens":616,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":585,"completion_tokens_details":{"reasoning_tokens":548}},"tokens_in":585,"tokens_out":616,"duration_ms":6849,"temperature":1.0,"reasoning_tokens":548,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T21:08:29.894479+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the linear absorption or transmission of a high-Q single-mode microcavity containing a known number N of identical molecules at low temperature, with κ ≈ λ and the collective coupling λ√N kept fixed while N is varied. The paper predicts sidebands offset from each zeroth-order polariton by the ground-state vibrational frequency, with peak height ∝1/N and blue shift Δ = 2λ(√N − √(N−1)). If those sidebands are absent, do not scale with N, or persist unchanged in a multimode cavity, the central claim fails.","supporting_citations":[{"cited_title":"Linear response of molec- ular polaritons,","cited_arxiv_id":null,"evidence_quote":"Supplies the linear-response formalism and absorption/transmission expressions that the paper builds on."},{"cited_title":"Excitonic spectral features in strongly coupled organic polaritons,","cited_arxiv_id":null,"evidence_quote":"Establishes the classical linear-optics description of strongly coupled organic polaritons that the 1/N expansion refines."},{"cited_title":"Dark vibronic polaritons and the spectroscopy of organic microcavities,","cited_arxiv_id":null,"evidence_quote":"Supplies the vibronic polariton spectroscopy framework with dark and bright vibronic states."},{"cited_title":"Mukamel, Principles of nonlinear optical spectroscopy (Oxford University Press, 1995)","cited_arxiv_id":null,"evidence_quote":"Defines the nonlinear susceptibilities and irreducible diagrammatic concepts used in Eq. 9."},{"cited_title":"Exact states and spectra of vibrationally dressed polaritons,","cited_arxiv_id":null,"evidence_quote":"Reports previous small-N spectra showing the first-order polariton features that the paper explains as 1/N corrections."},{"cited_title":"Polariton localization and dispersion properties of disordered quantum emitters in multimode microcavities,","cited_arxiv_id":null,"evidence_quote":"Supports the statement that multimode cavities blur the predicted sidebands into continua."}],"review_version":1}