{"id":"9cfb4ca9-83b4-47e1-948f-953433f83e36","arxiv_id":"2505.09557","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Using a modified multimode nonlinear Schrödinger equation, the authors show that random birefringence generally reduces nonlinear effects in multimode fibers, but the Raman-induced soliton self-frequency shift varies non-monotonically with correlation length, and beam self-cleaning survives at…","lead":"This paper simulates how random twisting of a fiber's polarization axes changes nonlinear light pulses in multimode fibers. It finds that random birefringence usually weakens nonlinear effects, but some effects, like soliton frequency shifting, respond non-monotonically to the twist rate.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The central LC-dependence of the Raman-induced frequency shift is reported without ensemble averaging or error bars, so the claimed nonmonotonic behavior and the resilience of beam self-cleaning are not statistically established.","rationale":"The reader's weakest assumption focuses on the discretized random-birefringence model and neglected inter-mode-group coupling. That is a legitimate model-level concern, but the more load-bearing issue is statistical: the paper's central quantitative result—the LC-dependence of the Raman shift—is generated from a stochastic process without any ensemble averaging or error characterization. Even if the model is exactly the right one, a single realization cannot establish a nonmonotonic trend, particularly when LC is comparable to the propagation length. This concern is distinct from, though related to, the reader's model-approximation worry, and it is also flagged in the reader's rationale as a lack of statistical characterization. Because the reader already assigned a conditional verdict and identified the stochastic-evidence issue, the appropriate final recommendation remains CONDITIONAL, which corresponds to 'UNCHANGED' relative to the reader's verdict. The proposed ensemble test would settle whether the claimed nonmonotonicity is real or a realization artifact; if the ensemble test fails, the central claim would need to be substantially weakened.","tokens_in":10523,"tokens_out":3628,"duration_ms":41267,"concrete_test":"Rerun Example 2 with at least 100 independent random realizations for each LC value in Fig. 5 (e.g., LC = 16.6 m, 50 m, 166 m, ∞) using a fixed, documented random-seed and section-generation protocol, and report the mean and 95% confidence interval of the Raman-induced wavelength shift for both input SOPs. If the nonmonotonic dependence persists in the ensemble mean, the central claim is supported; if not, it is an artifact of a single realization. Similarly, rerun Example 3 at 40 nJ and 70 nJ for LC = 10 m over 100 realizations and report the mean and spread of the LP01 fraction to test the beam self-cleaning resilience claim.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline quantitative claim—that the Raman-induced soliton self-frequency shift first decreases then increases as the correlation length LC decreases (Section IV-B, Fig. 5)—rests on a stochastic model, yet no ensemble averaging, error bars, seed information, or realization count is reported. As presented, each curve appears to be a single random realization. This matters especially because the smallest LC value used, LC = 16.6 m, is comparable to the 15 m fiber length, so the 'small LC' regime may involve only a handful of random SOP rotations. The observed nonmonotonicity could therefore be realization-specific rather than a robust physical trend. The same issue affects the beam self-cleaning resilience claim in Section IV-C and Fig. 6, where the LP01 fraction is reported without statistical spread. Additionally, Appendix B does not fully specify the random process: it mentions section lengths obeying a Gaussian distribution and defines LC through a correlation function (Eq. A4), but it does not give the parameters of that Gaussian, the statistics of the angular increments Δα, or the number of sections used. Hence the simulation is not reproducible from the manuscript alone and the stochastic evidence for the central claim is incomplete.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies nonlinear ultrashort pulse propagation in multimode fibers (MMFs) with random birefringence, using a vector generalized multimode nonlinear Schrödinger equation (GMMNLSE) in which the fiber is divided into sections with constant birefringence axes and random polarization rotations at the interfaces. The authors apply this model to three numerical examples: a basic spectral-broadening case, multimode soliton propagation with Raman-induced self-frequency shift, and Kerr spatial beam self-cleaning. The central claims are that random birefringence usually weakens nonlinearity but that the Raman-induced soliton self-frequency shift first decreases and then increases as the correlation length LC decreases, and that spatial beam self-cleaning can withstand random birefringence at high input peak powers. The model and simulation approach are inherited from earlier work, and the present contribution is the application to ultrashort-pulse regimes and the two specific nonlinear phenomena.","tokens_in":10787,"tokens_out":2997,"duration_ms":32100,"significance":"If the reported trends are robust, the paper addresses a timely and practically relevant question: whether random birefringence, unavoidable in real MMFs, prevents the formation of multimode solitons and spatial beam self-cleaning. The non-monotonic dependence of the Raman-induced self-frequency shift on LC would be an interesting and nontrivial result, and the resilience of beam self-cleaning at high power would offer practical guidance. The work is based on an established propagation equation and uses realistic fiber parameters computed from the fiber specification. However, the paper does not provide machine-checked proofs or public code, and the central quantitative claims rest entirely on stochastic simulations for which no ensemble averaging, error bars, or realization counts are reported. As presented, the evidence for the headline non-monotonic behavior and for the self-cleaning resilience is statistically incomplete.","major_comments":[{"comment":"","section":"Section IV-B, Fig. 5"},{"comment":"","section":"Appendix B, Eqs. (A3)-(A4)"},{"comment":"","section":"Section IV-C, Fig. 6"},{"comment":"","section":"Appendix B, mode-coupling matrix"}],"minor_comments":[{"comment":"","section":"Section IV-B, paragraph after Fig. 4"},{"comment":"","section":"Section I, Introduction"},{"comment":"","section":"Table I"},{"comment":"","section":"Appendix A, text after Eq. (A2)"},{"comment":"","section":"Section III"},{"comment":"","section":"Section IV-B, explanation of non-monotonicity"}],"recommendation":"major_revision","confidential_remarks":"The topic is suitable for the journal and the underlying propagation model is sound, but the central numerical claims currently lack statistical support. The authors should be asked to provide ensemble-averaged results with error bars or a clear convergence demonstration, and to complete the specification of the random process in Appendix B. If these issues are addressed, the paper could be a useful contribution; if the statistical evidence is not provided, the headline claims cannot be considered established."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's my take. The paper is a numerical study of random birefringence in multimode fibers, extending a vector GMMNLSE to include Raman and self-steepening. The new physics claim is that the soliton self-frequency shift depends non-monotonically on the correlation length LC, and that beam self-cleaning survives at high peak powers. Both are worth checking.\n\nThe model itself is on solid ground: the propagation equation is the standard vector GMMNLSE, and the random birefringence is modeled as in Guasoni et al., with mode coupling at section interfaces. The authors use realistic fiber parameters and up to fourth-order dispersion. Their qualitative explanation for the non-monotonic Raman shift—birefringence splitting the pulse, then rapid SOP scrambling restoring it—is plausible.\n\nBut the central evidence is not statistically established. The paper reports no ensemble averaging, no error bars, and no realization count. Random birefringence is a stochastic process; the Raman shift and LP01 fraction are random variables. As presented, Fig. 5 could be a single random draw. This matters a lot: the smallest LC used is 16.6 m in a 15-m fiber, so the 'strong scrambling' regime involves maybe one or two section interfaces. The non-monotonic behavior might be realization-specific. The same criticism applies to Fig. 6. Appendix B also leaves out the parameters of the Gaussian section lengths and the statistics of the angular increments, and no code or data is provided, so the simulation is not reproducible from the manuscript.\n\nA second soft spot is the neglect of linear coupling between mode groups. The paper states this explicitly, but for a 50-um core with several mode groups, that coupling might not be negligible. I'd want a justification or a sensitivity test.\n\nNone of this is fatal to the approach. The topic is timely—ultrashort pulses in MMFs are used in lasers and short-reach links, and random birefringence is unavoidable. The paper deserves peer review, but a serious referee should require ensemble-averaged results with error bars, a complete description of the random process, and ideally a test against a continuous-birefringence model. If the non-monotonicity survives that, it's a real result. As it stands, I'd treat it as an interesting but unverified numerical observation.","headline":"Interesting numerical study, but the headline non-monotonicity rests on single realizations—needs ensemble statistics before I'd trust it.","tokens_in":11310,"tokens_out":4011,"would_cite":false,"duration_ms":38433,"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":"Random birefringence in multimode fibers normally weakens nonlinearity, yet multimode solitons and high-power beam self-cleaning survive; the Raman soliton shift first drops then recovers as correlation length shrinks.","keywords":["random birefringence","multimode fibers","multimode solitons","soliton self-frequency shift","beam self-cleaning","generalized multimode nonlinear Schrödinger equation","Raman effect","correlation length"],"falsifier":"Measure the output center wavelength of a 50 fs, 1550 nm pulse after 15 m of graded-index multimode fiber while varying the correlation length of birefringence, for example by controlled twisting or bending: the paper predicts the Raman-induced shift is smallest at an intermediate $L_C$ and returns toward the no-birefringence value for small $L_C$. A monotonic decrease of the shift with decreasing $L_C$, or a collapse of beam self-cleaning at 70 nJ input energy, would contradict the central claim.","tokens_in":10345,"feed_emoji":"⚡","tokens_out":9279,"duration_ms":86985,"temperature":0.7,"pith_summary":"Random birefringence is the unavoidable polarization disorder in real multimode fibers, and this paper asks whether it spoils the nonlinear ultrashort-pulse effects that make short multimode fiber segments attractive. Simulating a vectorized generalized multimode nonlinear Schrödinger equation, the paper finds that birefringence generally reduces Kerr nonlinearity and spectral broadening. It then shows two headline effects survive in practical regimes: multimode solitons still form, and spatial beam self-cleaning withstands random birefringence when the input peak power is high. It also reports a nonmonotonic dependence of the Raman-induced soliton self-frequency shift on the correlation length, first falling and then rising as polarization scrambling intensifies. The practical point is that random birefringence in ordinary fibers need not block multimode soliton formation or beam self-cleaning, which matters for multimode-fiber lasers and high-energy pulse transmission.","feed_headline":"Polarization disorder weakens, then spares, multimode nonlinear effects","feed_subtitle":"Multimode solitons still form and beam self-cleaning holds at high peak power despite random birefringence.","key_machinery":"The load-bearing object is the vector generalized multimode nonlinear Schrödinger equation (GMMNLSE), in which each LP spatial mode is split into $x$- and $y$-polarized Jones-vector components, doubling $N$ modes to $2N$ polarized components. Random birefringence is inserted by dividing the fiber into short sections with fixed birefringence axes and applying a projection matrix $P$ at each interface to rotate the state of polarization by a random angle. The correlation length $L_C$ sets how fast these rotations decorrelate, while the beat length $L_B$ sets the birefringence strength inside a section; the nonlinear coupling coefficients $S^K$ and $S^R$ carry the mode-overlap dependence of the Kerr and Raman terms. This machinery converts random birefringence into two tunable parameters whose effects on soliton frequency shift and beam cleaning can be mapped quantitatively.","core_discovery":"The paper's central claim is that random birefringence in multimode fibers acts as a general suppressor of Kerr nonlinearity while leaving the signature spatiotemporal phenomena partially intact. In the multimode-soliton scenario, the pulse still binds its modes together: modal walk-off does not change significantly compared with the no-birefringence case, even for correlation lengths down to 16.6 m. The Raman-induced soliton self-frequency shift, however, is not monotonically suppressed: starting from no birefringence, the shift first decreases as $L_C$ drops to about 166 m, then increases as $L_C$ drops further, so that at $L_C = 16.6$ m it is nearly the same as with no birefringence. In the beam self-cleaning scenario, random birefringence degrades cleaning at 40 nJ input energy, but at 70 nJ the fundamental-mode energy fraction and output beam quality improve substantially, showing that high peak power overcomes the polarization disorder. These findings are presented as numerical results from a modified vector GMMNLSE with parameters of realistic graded-index fibers.","pith_inferences":["The paper leaves implicit that the correlation length of a fiber is a control knob for the Raman soliton self-frequency shift in multimode systems; one could test this by straining or twisting fibers to vary $L_C$ and observing whether the shift dips and recovers.","Because the random-birefringence model is piecewise constant and neglects linear coupling between different mode groups, real fibers with smoothly varying birefringence may show a smoother or shifted $L_C$ curve; comparing against a continuously varying birefringence model would sharpen the prediction.","The high-power resilience of beam self-cleaning suggests that multimode-fiber lasers and high-energy ultrashort-pulse delivery systems may not require stringent polarization control at high energies, an implication the paper does not state.","The nonmonotonic $L_C$ dependence of the Raman shift may also affect supercontinuum generation and intermodal four-wave mixing in multimode fibers, where Raman and Kerr terms compete; these processes are not examined here."],"forward_implications":["In a 15 m graded-index multimode fiber, a 50 fs, 1550 nm multimode pulse can still form a multimode soliton even when random birefringence is present; the output's modal walk-off is not significantly changed.","The Raman-induced soliton self-frequency shift is smallest for an intermediate correlation length, not for the strongest birefringence; as $L_C$ becomes very small the shift approaches the no-birefringence value.","Spatial beam self-cleaning, degraded by random birefringence at 40 nJ input energy, is restored at 70 nJ, so higher input peak power gives stronger resistance to polarization disorder.","Overall spectral broadening from the Kerr effect is reduced by random birefringence, because polarization components separate in time and SOP-dependent nonlinear coupling weakens.","The modified vector GMMNLSE can be reused to study input state of polarization, polarization mode dispersion, and nonlinear polarization dynamics in multimode fibers."],"supporting_citations":[{"why":"Defines the discrete-section random-birefringence model, the section-interface projection matrix, and the correlation length $L_C$ used throughout.","marker":"[24]"},{"why":"Supplies the vector GMMNLSE solver and the baseline multimode-soliton and beam-self-cleaning dynamics that this paper extends to random birefringence.","marker":"[21]"},{"why":"Derives the generalized multimode nonlinear Schrödinger equation that the paper modifies to include polarization.","marker":"[20]"},{"why":"Shows single-mode solitons resist random birefringence, the behavior the paper finds persists for multimode solitons.","marker":"[2]"},{"why":"Provides the wave-condensation picture of beam self-cleaning under structural disorder that motivates the high-power resilience test.","marker":"[28]"},{"why":"Preliminary study of polarization effects on nonlinear pulse propagation in multimode fibers that this work extends.","marker":"[25]"},{"why":"Preliminary study of random birefringence on nonlinear beam self-cleaning, extended here to multimode solitons and varying $L_C$.","marker":"[26]"}],"fun_headline_variants":["Random birefringence tempers, not kills, multimode nonlinearity","High power lets self-cleaning beat random birefringence","Soliton shift does a U-turn under fiber birefringence","Polarization disorder weakens but spares core multimode effects","Random twist in fibers: nonlinearity dips, then recovers"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central premise is that a real multimode fiber can be represented by short straight sections with fixed birefringence axes and random polarization rotations only at the interfaces between sections, with no coupling between different mode groups; if that representation fails for a real fiber, the predicted dependence on correlation length and the resilience of beam self-cleaning may not appear.","fun_headline_variants_meta":{"raw":{"variants":["Random birefringence tempers, not kills, multimode nonlinearity","High power lets self-cleaning beat random birefringence","Soliton shift does a U-turn under fiber birefringence","Polarization disorder weakens but spares core multimode effects","Random twist in fibers: nonlinearity dips, then recovers"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000508,"raw_usage":{"total_tokens":2490,"prompt_tokens":972,"completion_tokens":1518,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":588,"completion_tokens_details":{"reasoning_tokens":1423}},"tokens_in":588,"tokens_out":1518,"duration_ms":10893,"temperature":1.0,"reasoning_tokens":1423,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T21:27:53.628670+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the output center wavelength of a 50 fs, 1550 nm pulse after 15 m of graded-index multimode fiber while varying the correlation length of birefringence, for example by controlled twisting or bending: the paper predicts the Raman-induced shift is smallest at an intermediate $L_C$ and returns toward the no-birefringence value for small $L_C$. A monotonic decrease of the shift with decreasing $L_C$, or a collapse of beam self-cleaning at 70 nJ input energy, would contradict the central claim.","supporting_citations":[{"cited_title":"Intermodal Four -Wave Mixing and Parametric Amplification in Kilometer-Long Multimode Fibers,","cited_arxiv_id":null,"evidence_quote":"Defines the discrete-section random-birefringence model, the section-interface projection matrix, and the correlation length $L_C$ used throughout."},{"cited_title":"Multimode Nonlinear Fiber Optics: Massively Parallel Numerical Solver, Tutorial, and Outlook,","cited_arxiv_id":null,"evidence_quote":"Supplies the vector GMMNLSE solver and the baseline multimode-soliton and beam-self-cleaning dynamics that this paper extends to random birefringence."},{"cited_title":"Description of ultrashort pulse propagation in multimode optical fibers,","cited_arxiv_id":null,"evidence_quote":"Derives the generalized multimode nonlinear Schrödinger equation that the paper modifies to include polarization."},{"cited_title":"Stability of solitons in randomly varying birefringent fibers,","cited_arxiv_id":null,"evidence_quote":"Shows single-mode solitons resist random birefringence, the behavior the paper finds persists for multimode solitons."},{"cited_title":"Wave condensation with weak disorder versus beam self- cleaning in multimode fibers,","cited_arxiv_id":null,"evidence_quote":"Provides the wave-condensation picture of beam self-cleaning under structural disorder that motivates the high-power resilience test."},{"cited_title":"Effects of Polarization on the Nonlinear Pulse Propagation in Multimode Fibers,","cited_arxiv_id":null,"evidence_quote":"Preliminary study of polarization effects on nonlinear pulse propagation in multimode fibers that this work extends."},{"cited_title":"Effects of Random Birefringence in Multimode Fibers on Nonlinear Beam Self -cleaning,","cited_arxiv_id":null,"evidence_quote":"Preliminary study of random birefringence on nonlinear beam self-cleaning, extended here to multimode solitons and varying $L_C$."}],"review_version":1}