{"id":"2b9a1a7a-c48f-484d-8eb7-16f29db42463","arxiv_id":"2505.11299","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":3.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Random index fluctuations speed up the thermalization and condensation of light in multimode fibers, enabling beam self-cleaning, but modal phase-locking shows the effect is not purely statistical.","lead":"This review derives wave-turbulence kinetic equations showing that random refractive-index disorder accelerates optical thermalization and condensation in multimode fibers, explaining beam self-cleaning. It also surveys experiments that confirm Rayleigh-Jeans thermalization, while noting that observed modal phase-locking goes beyond the statistical theory.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Weak-disorder acceleration scaling (Eq. 30) is derived for Ld<<Lnl and for artificial disorder models, while BSC experiments operate at Ld~Lnl; the paper's central quantitative claim is unverified in its target regime.","rationale":"The paper is a review that consolidates a substantial body of prior work. The equilibrium Rayleigh-Jeans prediction for mode occupancies is independently supported by experiments such as Baudin et al. (2020), Pourbeyram et al. (2022), and Mangini et al. (2022), and the kinetic equations are validated against NLS simulations without adjustable parameters within the assumptions of the theoretical regime. Credit is due for the detailed derivations and the parameter-free numerical comparisons. However, the distinctive new claim of this review is the disorder-mediated acceleration of thermalization, quantified by Eq. (30). This scaling is derived under Ld << Lnl and for specifically chosen disorder correlation models, yet the paper itself admits in Sec. 2.5 that BSC experiments lie at Ld ~ Lnl and in a mixed coherent-incoherent regime. This is not a disagreement with consensus; it is an internal gap between the regime in which the derivation is controlled and the regime to which the abstract's practical claim refers. The artificial status of the partially correlated disorder model, flagged in Sec. 2.3.9, further weakens the link between Eq. (30) and real fiber imperfections. Phase-locking is also a serious limitation, but it mainly affects the completeness of the statistical explanation rather than the quantitative rate claim, and the paper already frames it as a separate mechanism in Sec. 3.4. The conditional verdict remains appropriate: the central equilibrium claim is plausible and well supported, but the quantitative disorder-acceleration claim should be either demonstrated in the relevant regime or explicitly labeled as a model-dependent prediction. These concerns are addressable through additional simulations at Ld ~ Lnl with experimentally motivated disorder statistics, plus softening of the abstract; they do not require rejection of the paper.","tokens_in":61784,"tokens_out":9817,"duration_ms":111287,"concrete_test":"Re-run the NLS simulations behind Fig. 3 with disorder parameters chosen so that Ld/Lnl = 1, the ratio the authors state is relevant to BSC experiments, and extract the thermalization length from n0(z) for Δβ/β0 = 10^-3, 10^-2, and 10^-1. Then compare the observed scaling exponent with Eq. (30). If the acceleration factor deviates from Δβ/β0 by more than the numerical uncertainty, the central quantitative claim fails in its stated experimental regime.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claim is Eq. (30): weak disorder accelerates RJ thermalization and condensation by a factor L_disord_kin / L_ord_kin ~ Δβ/β0. The underlying discrete kinetic equation (26) is derived under the disorder-dominated assumption Ld << Lnl (Eq. 9, Sec. 2.3.1), which is what justifies the Furutsu-Novikov damping 8Δβ in Eq. (13) and the adiabatic elimination of fourth-order moments used in Eqs. (14)-(18) and (A.10)-(A.12). In Sec. 2.5 the authors explicitly concede that in beam-cleaning experiments polarization disorder and nonlinearity are of the same order (Ld ≳ Lnl), so the system enters a 'mixed coherent-incoherent regime' that the kinetic equation does not describe. Moreover, the numerical validation of Eq. (30) is performed only for the mode-decorrelated model and for the partially mode-correlated model, which Sec. 2.3.9 itself calls an 'artificial model'; no experimentally measured disorder correlation statistics or values of Ld are provided. Thus the abstract's statement that a weak disorder accelerates thermalization in a practical experimental setting is a model-dependent prediction derived outside the regime in which the BSC experiments actually operate, and it is not quantitatively supported for real GRIN fibers.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This review-style paper argues that beam self-cleaning (BSC) in graded-index multimode fibers can be understood as classical wave thermalization toward the Rayleigh-Jeans equilibrium, and that refractive-index disorder changes only the rate of this thermalization, not the final state. The authors derive wave-turbulence kinetic equations for both weak polarization-type disorder and strong random mode coupling, obtaining closed kinetic equations (Eqs. 21, 26, 37) whose stationary solutions are the RJ distribution or the equipartition distribution depending on the regime. The central quantitative claim is Eq. (30), namely that weak disorder accelerates thermalization and condensation by the factor L_disorder_kin / L_ord_kin ~ Δβ/β0. The kinetic equations are validated against numerical simulations of the NLS equation without adjustable parameters (Figs. 3, 4, 7-10), and the experimental part reviews recent measurements of entropy growth, RJ condensation, negative-temperature equilibria, and modal phase-locking. The paper concludes that the observed phase-locking accompanying BSC is outside the current statistical-mechanics description and motivates future extensions.","tokens_in":62049,"tokens_out":3376,"duration_ms":39578,"significance":"If the central kinetic-theory result holds as stated, this paper provides a parameter-free, first-principles description of thermalization rates in disordered multimode fibers, which would be a substantial step beyond the equilibrium thermodynamic theories of Refs. [11,13-15]. The appendices contain self-contained derivations of the discrete kinetic equations, and the NLS-to-kinetic-equation comparisons in Figs. 3, 4, 8-10 are quantitative and do not use adjustable parameters. The review also usefully collects recent experiments on entropy growth, calorimetry, and negative-temperature equilibria that support the equilibrium part of the statistical-mechanics picture, while honestly highlighting the phase-locking phenomenon that lies beyond the incoherent-wave assumption. The main weakness is that the central quantitative claim, Eq. (30), is derived in the disorder-dominated regime Ld << Lnl, whereas the paper itself concedes that BSC experiments operate near Ld ~ Lnl; the numerical support for Eq. (30) covers only idealized disorder models, not experimentally characterized GRIN fibers.","major_comments":[{"comment":"Equation (30) is the central quantitative claim of the paper, but it is derived from the discrete kinetic equation (26) under the disorder-dominated hypothesis Ld << Lnl stated in Eq. (9). In Sec. 2.5 the authors explicitly write that in beam-cleaning experiments the impact of weak polarization disorder is expected to be of the same order as nonlinear effects and that for Ld ≥ Lnl the system enters a mixed coherent-incoherent regime that the kinetic equation does not describe. Since the abstract states that weak disorder accelerates thermalization in a practical experimental setting, the paper needs either a quantitative argument that the Ld/Lnl ratio in the actual BSC experiments is small enough for Eq. (30) to apply, or a clear reformulation of the claim as a model prediction that remains to be tested experimentally in the BSC regime.","section":"Sec. 2.3.6, Eq. (30), and Sec. 2.5"},{"comment":"The numerical validation of Eq. (30) is performed only for the mode-decorrelated model and for the partially mode-correlated model, and the latter is explicitly called an artificial model in Sec. 2.3.9. No experimentally measured disorder correlation statistics, values of lβ, or values of Ld for real GRIN fibers are provided. The claim that disorder accelerates thermalization in the experimental setting of BSC therefore rests on model disorder statistics rather than on measured fiber imperfections; this should be stated explicitly when the scaling is promoted in the abstract and conclusions.","section":"Sec. 2.3.9, Figs. 3 and 7"},{"comment":"The discussion of dispersion-relation corrections shows that the acceleration scaling can be weakened or modified when the perturbation b0 is comparable to Δβ. Since typical beam-cleaning parameters give b0/β0 ~ 5×10^-4 to 5×10^-3 and no experimental estimate of Δβ for real GRIN fibers is given, the paper does not establish that the condition Δβ >> b0 required for Eq. (30) holds in the experiments it discusses. This is a load-bearing caveat for the quantitative validity of the central claim and should be moved from the end of Sec. 2.3 into the statement of Eq. (30).","section":"Sec. 2.3.10, Eq. (33)"}],"minor_comments":[{"comment":"There is a typo in the introduction: the text reads \"back in 1978 by by Lin et al.\"; the duplicated \"by\" should be removed.","section":"Sec. 1"},{"comment":"The notation for the continuous extension is unclear: the text defines κ = β0(px,py) and then writes β̃_k = κx + κy + β0, which mixes discrete indices and continuous variables; please define k and β̃_k unambiguously.","section":"Eq. (20)"},{"comment":"The sentence defining M_lmnp(n) contains a garbled parenthetical \"with ’ nm’ for ’ nm(z)’\"; the definition should be written in standard notation with all mode indices explicitly time-dependent.","section":"Eq. (26)"},{"comment":"The reference titles contain typographical errors: \"Wave Trubulence\" appears in [41] and \"Physics of Wave Trubulence\" in [43]; these should be corrected.","section":"References [41,43]"},{"comment":"The minimization problem in Eq. (47) would be easier to follow if the functional ε were explicitly defined as a function of the modal phases and if the constraint that the total power is fixed were written out.","section":"Sec. 3.4, Eq. (47)"}],"recommendation":"major_revision","confidential_remarks":"The paper is largely a review of the authors' own prior contributions, which is appropriate for this venue, but the editorial process may wish to ensure that the genuinely new synthesis—the disorder-acceleration scaling in the BSC regime—is not over-stated relative to its demonstrated domain of validity. The experimental sections are competent and well matched to the equilibrium theory; the main risk is that the central quantitative claim, Eq. (30), is presented in the abstract as a practical experimental result while the paper's own Sec. 2.5 acknowledges the regime mismatch."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about arXiv:2505.11299. First, it is a review, not a new-results paper: the kinetic equations with weak and strong disorder, and the experimental survey, reproduce the authors' own prior work (Refs. 25, 103, 104, 38, 109–111). Second, it is a genuinely good review: the derivations in Appendices A–C are detailed and self-contained, the NLS-versus-kinetic-equation comparisons are parameter-free and convincing, and the paper is unusually candid about where the theory does not apply.\n\nThe main quantitative claim is Eq. (30): weak disorder accelerates thermalization/condensation by a factor Δβ/β0. The stress-test note worries that this is derived under Ld << Lnl while the beam-cleaning experiments operate near Ld ~ Lnl. That concern is correct but not fatal, because the paper itself says as much in Sec. 2.5. The scaling is presented as a theoretical prediction from a specific model, validated numerically in that model's regime; the experiments are cited as qualitative support. The abstract does overreach by saying the theory 'explains' BSC, when Sec. 3.4 shows that phase-locking is a separate mechanism outside the statistical theory. That is a real soft spot—an editor should ask for a softer abstract—but it is not a load-bearing flaw.\n\nWhat the paper does well: it gives a coherent synthesis of the wave-turbulence approach, including the discrete regime relevant to graded-index fibers, the role of disorder correlations, and the thermodynamic constraints on mode populations. The simulations in Figs. 3, 4, 8–10 match the kinetic equations without adjustable parameters, and the theoretical derivations are reproducible from the text. The paper also flags its own limitations honestly, including the mixed coherent-incoherent regime, the artificial partially-correlated disorder model, and the unresolved tension with modal phase-locking.\n\nWeaker spots: no public code or data, key experimental figures lack error bars, and the self-citation density is high (though the cited prior results are real and formally derived). None of these are disqualifying for a review, but they do limit how strongly a reader can independently check the experimental claims.\n\nWho this is for: anyone working on multimode fiber beam propagation, classical wave thermalization, or wave turbulence in disordered media. A serious referee can add value by pushing on the abstract's overclaim and asking for a clearer statement of the regime of validity of Eq. (30). I would not desk-reject it. I would send it out, and recommend the authors soften the abstract and release simulation data.","headline":"Thorough and honest review of wave turbulence in multimode fibers; the disorder-acceleration scaling is a reproduced result with a regime caveat the authors themselves flag, so treat the abstract's 'explains' with a grain of salt.","tokens_in":62689,"tokens_out":1079,"would_cite":true,"duration_ms":14092,"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":"Beam self-cleaning in multimode fibers is wave thermalization, and weak disorder accelerates it.","keywords":["Nonlinear optics","Statistical mechanics","Wave turbulence","Multimode fibers","Beam self-cleaning","Wave thermalization","Wave condensation","Rayleigh-Jeans distribution"],"falsifier":"Measure, in a GRIN fiber with controllable distributed stress or index fluctuations, the propagation length at which the fundamental-mode fraction reaches half its Rayleigh-Jeans equilibrium value for several disorder strengths at fixed injected power and energy; if the central claim is wrong, the measured thermalization length will not scale inversely with the disorder parameter $\\Delta\\beta$ and will not converge to the same RJ endpoint. A second check: in a strongly disordered fiber with $L^{\\rm RJ}_{\\rm kin}\\ll L^{\\rm eq}_{\\rm kin}$, observe the modal populations as a function of length; the transient condensate predicted by the kinetic equation should rise and then decay to equipartition, and its absence would falsify the strong-disorder kinetic closure.","tokens_in":61528,"feed_emoji":"💡","tokens_out":6383,"duration_ms":60770,"temperature":0.7,"pith_summary":"This review argues that beam self-cleaning in graded-index multimode fibers is a classical wave thermalization process: the power in each guided mode relaxes toward the Rayleigh-Jeans distribution that maximizes optical entropy, and the final cleaned beam is the condensate state set by the conserved power and energy. On the wave-turbulence side, the paper derives kinetic equations showing that weak refractive-index disorder accelerates this relaxation, with the thermalization length reduced by the ratio of disorder strength to mode spacing. This matters because it turns beam cleaning from an empirical observation into a prediction from two conserved quantities, and it explains why fabrication imperfections help rather than hinder self-cleaning. The paper also reports experiments measuring entropy growth and negative-temperature equilibria, and identifies a modal phase-locking that accompanies cleaning but lies beyond the incoherent statistical description.","feed_headline":"Weak disorder accelerates beam self-cleaning in multimode fibers","feed_subtitle":"Mode occupancies relax to the Rayleigh-Jeans law; entropy growth experiments confirm the thermal picture.","key_machinery":"The load-bearing object is the discrete wave-turbulence kinetic equation for the averaged modal powers $n_p(z)$, obtained from the modal NLS equation by closing the moment hierarchy with the Gaussian factorization that disorder-induced random phases justify. In the parabolic GRIN potential the mode spectrum is equally spaced, $\\beta_p=\\beta_0(p_x+p_y+1)$, so four-wave resonances are discrete and only exact ones, $\\Delta\\omega_{lmnp}=0$, contribute; the disorder term supplies the damping rate $8\\Delta\\beta$ that replaces the Dirac resonance condition with a finite bandwidth. The equation conserves power and energy, obeys an $H$-theorem for $S=\\sum_p \\log n_p$, and relaxes to the RJ distribution, with the thermalization length $L^{\\rm disor}_{\\rm kin}\\sim \\Delta\\beta L_{\\rm nl}^2/\\bar S^2_{lmnp}$ controlling the rate. In the strong-disorder case the same closing procedure yields a kinetic equation whose collision term gives RJ thermalization and whose linear term drives power equipartition, so their competition fixes whether condensation or equipartition wins.","core_discovery":"On the paper's own terms, the central claim is that the nonequilibrium evolution of a speckle beam in a GRIN multimode fiber is described by a discrete wave-turbulence kinetic equation whose equilibrium is the Rayleigh-Jeans law $n_p^{\\rm eq}=T/(\\beta_p-\\mu)$, where $\\beta_p$ is the propagation constant of mode $p$. Beam self-cleaning is the corresponding condensation: most power accumulates in the fundamental mode while higher-order modes keep a small population required by energy conservation. The new quantitative result is that weak, mode-decorrelated disorder introduces an effective dissipation $\\Delta\\beta$ that broadens the four-wave resonances; in the discrete regime only exact resonances survive, and the thermalization length obeys $L^{\\rm disor}_{\\rm kin}/L^{\\rm ord}_{\\rm kin}\\sim \\Delta\\beta/\\beta_0$, so disorder speeds up thermalization without changing the final RJ state. For strong disorder, the kinetic equation predicts that, although the long-distance equilibrium is power equipartition, a transient RJ condensation can still occur when the RJ thermalization length is shorter than the disorder-induced equipartition length. These predictions are validated by parameter-free comparisons with NLS simulations and by experiments on entropy growth, RJ condensation, negative temperatures, and modal phase-locking.","pith_inferences":["A testable extension suggested by the rate scaling: in fibers with engineered disorder strength, measuring the propagation length needed to reach a fixed fraction of the equilibrium condensate should show $L_{\\rm kin}\\propto 1/\\Delta\\beta$ at fixed power and energy, a clean way to isolate the disorder contribution from nonlinearity.","If disorder only sets the rate, the same RJ endpoint should be reachable in fibers of very different imperfection levels; that invites comparing step-index and GRIN fibers at matched mode counts to test whether the absence of beam cleaning in step-index fibers comes from the quasi-degenerate resonance structure rather than from a different equilibrium.","The modal phase-locking that accompanies the amplitude thermalization suggests that the true attracting state may be a phase-coherent condensate dressed by a thermal tail; a statistical theory retaining phase information could predict the observed locked phases from the same conserved quantities, extending the current incoherent description.","Because strong disorder can produce a transient condensate that later decays, single-shot measurements along the fiber might reveal a non-monotonic fundamental-mode population; time-averaged output measurements would miss this unless the transient length is resolved."],"forward_implications":["If the central claim is right, the output mode distribution of a self-cleaned beam is fixed by the injected power and kinetic energy alone, so fiber output profiles can be predicted from input statistics without solving the full nonlinear wave equation.","Weak disorder accelerates the approach to that equilibrium, implying that small fabrication imperfections or bending noise make beam cleaning easier to observe in short fibers; the final cleaned state, however, is not changed by the disorder.","In the strong-disorder regime, transient condensation can appear before equipartition, so experiments on stressed fibers should see a rise followed by a decay of the fundamental-mode population before the beam settles to a uniform mode distribution.","Because the kinetic theory conserves power and energy, higher-order modes are never fully emptied, so a cleaned beam is expected to carry a measurable residual population in higher-order modes rather than a pure single-mode output.","The measured entropy growth during thermalization, including the two-beam calorimetry results, follows from the same $H$-theorem, giving a direct experimental handle on the optical temperature and chemical potential."],"supporting_citations":[{"why":"Supplies the wave-turbulence kinetic equation and the RJ condensation equilibrium for waveguides in the absence of disorder, the baseline for the ordered thermalization length.","marker":"[11]"},{"why":"Provides the experimental observation of Rayleigh-Jeans condensation in a 12-m GRIN fiber with speckle ensembles, anchoring the equilibrium predictions.","marker":"[12]"},{"why":"Derives wave condensation with weak disorder versus beam self-cleaning and gives the NLS and kinetic simulations behind Eq. (30).","marker":"[25]"},{"why":"Establishes the dramatic disorder-mediated acceleration of wave condensation, supporting the Ldisor scaling.","marker":"[103]"},{"why":"Derives the strong-disorder kinetic equation and reports NLS simulations and clamped-fiber experiments showing transient RJ condensation.","marker":"[104]"},{"why":"Measures configuration entropy growth during thermalization of a single beam at fixed length and in cutback experiments.","marker":"[109]"},{"why":"Demonstrates two-beam optical calorimetry, showing equilibration of temperature and chemical potential through heat exchange.","marker":"[110]"},{"why":"Reports the experimental modal phase-locking that accompanies beam self-cleaning, motivating the extension beyond thermodynamics.","marker":"[111]"}],"fun_headline_variants":["Disorder speeds up thermalization in multimode fibers","Weak disorder accelerates beam self-cleaning","Disorder accelerates Rayleigh-Jeans thermalization in fibers","Weak disorder boosts beam self-cleaning in GRIN fibers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that refractive-index disorder acts much more strongly than the Kerr nonlinearity, so the modal phases are thoroughly randomized and the mode statistics factorize like Gaussian variables; the paper itself notes that in real beam-cleaning fibers the disorder and nonlinear lengths are comparable, so the quantitative acceleration scaling is derived in a regime experiments only partially satisfy.","fun_headline_variants_meta":{"raw":{"variants":["Disorder speeds up thermalization in multimode fibers","Weak disorder accelerates beam self-cleaning","Disorder accelerates Rayleigh-Jeans thermalization in fibers","Weak disorder boosts beam self-cleaning in GRIN fibers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000508,"raw_usage":{"total_tokens":2559,"prompt_tokens":1110,"completion_tokens":1449,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":726,"completion_tokens_details":{"reasoning_tokens":1387}},"tokens_in":726,"tokens_out":1449,"duration_ms":9279,"temperature":1.0,"reasoning_tokens":1387,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T20:53:53.238963+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure, in a GRIN fiber with controllable distributed stress or index fluctuations, the propagation length at which the fundamental-mode fraction reaches half its Rayleigh-Jeans equilibrium value for several disorder strengths at fixed injected power and energy; if the central claim is wrong, the measured thermalization length will not scale inversely with the disorder parameter $\\Delta\\beta$ and will not converge to the same RJ endpoint. A second check: in a strongly disordered fiber with $L^{\\rm RJ}_{\\rm kin}\\ll L^{\\rm eq}_{\\rm kin}$, observe the modal populations as a function of length; the transient condensate predicted by the kinetic equation should rise and then decay to equipartition, and its absence would falsify the strong-disorder kinetic closure.","supporting_citations":[{"cited_title":"Baudin, A","cited_arxiv_id":null,"evidence_quote":"Provides the experimental observation of Rayleigh-Jeans condensation in a 12-m GRIN fiber with speckle ensembles, anchoring the equilibrium predictions."},{"cited_title":"Garnier, A","cited_arxiv_id":null,"evidence_quote":"Derives wave condensation with weak disorder versus beam self-cleaning and gives the NLS and kinetic simulations behind Eq. (30)."},{"cited_title":"129 (2022) 063901.doi:10.1103/PhysRevLett.129.063901","cited_arxiv_id":null,"evidence_quote":"Derives the strong-disorder kinetic equation and reports NLS simulations and clamped-fiber experiments showing transient RJ condensation."}],"review_version":1}