Random Mode Coupling Assists Kerr Beam Self-Cleaning in a Graded-Index Multimode Optical Fiber

Evgeniy V. Podivilov; Mikhail P. Fedoruk; Oleg S. Sidelnikov; Stefan Wabnitz

REVIEW 3 major objections 4 minor 14 references

Random Mode Coupling Assists Kerr Beam Self-Cleaning in a Graded-Index Multimode Optical Fiber

T0 review · 3 major / 4 minor · reviewed 2026-08-14 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Random coupling among modes with equal mode number, added to a coupled-mode model, reproduces both the linear speckle and the nonlinear beam self-cleaning observed in a graded-index multimode fiber.

desk verdict A numerically useful model showing that degenerate-mode random coupling explains Kerr self-cleaning, but a sign error in the central equation and missing quantitative support make the current claims unverifiable. read the letter →

arxiv 1908.07745 v1 pith:FTGCLRJE submitted 2019-08-21 physics.optics

Oleg S. Sidelnikov , Evgeniy V. Podivilov , Mikhail P. Fedoruk , Stefan Wabnitz This is my paper · ORCID

classification physics.optics

keywords beamself-cleaningmultimodeopticalfibergraded-indexKerreffectrandommodecouplingcoupled-modemodeldegeneratemodesspecklepattern

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper argues that Kerr beam self-cleaning in a graded-index multimode optical fiber can be understood within a coupled-mode model once random linear coupling between degenerate spatial modes is added. Without such coupling, the model produces no linear speckle and the fundamental mode power oscillates instead of stabilizing. With coupling restricted to modes sharing the same mode number n, and with a slightly off-center input beam, the model reproduces both the speckled output in the linear regime and the collapse to a near-fundamental-mode beam at high power. The authors report complete agreement with their experimental data, making degenerate-mode coupling the operative ingredient rather than coupling among all modes or among neighboring modes only.

What carries the argument

The central object is a random Hermitian coupling matrix C inserted into the coupled-mode equations for a graded-index fiber. Each off-diagonal element is normally distributed with zero mean and changes randomly along propagation with a 10 cm correlation length; the particular variant that works restricts coupling to modes with equal mode number n, i.e., degenerate modes. This matrix models fiber imperfections such as bending, tilting, and fabrication irregularities. It randomizes the phases and energy distribution among higher-order modes to create linear speckle, while leaving the fundamental mode energetically isolated, which the Kerr term then exploits to condense the beam into a stable bell-shaped output.

What would settle it

Launch a single higher-order mode into a real 62.5-micrometer graded-index fiber and measure how its power leaks into modes with different mode numbers over sub-meter scales; if significant non-degenerate coupling appears at the 10 cm scale, or if the coupling strength required to fit the speckle statistics is far outside measured fiber perturbations, the degenerate-mode coupling model would be ruled out.

Watch

Extended reading notes

Core claim

The central claim is that random linear coupling between degenerate modes, meaning modes with equal mode number n = 2p + |m|, is the key mechanism that lets Kerr nonlinearity clean a noisy multimode beam. In the linear regime this coupling scrambles energy among modes of the same n and, for an input beam slightly displaced from the fiber axis, produces a realistic speckle pattern; in the nonlinear regime the fundamental mode, being the only n = 0 mode, cannot lose energy through this coupling, so its power stabilizes and dominates, giving an output beam close to the fundamental mode. Other coupling prescriptions, such as coupling all modes or only neighboring modes, fail to reproduce the observed speckle and self-cleaning. The paper concludes that the degenerate-mode coupling model is in complete agreement with the available experimental data.

Load-bearing premise

The model's agreement depends on real fiber imperfections behaving like random coupling that mixes only modes with the same mode number, reshuffled along a 10 cm length, with a strength that the paper does not report.

Editorial extensions

If this is right

The coupled-mode model with degenerate-mode random coupling reproduces the linear speckled output observed in experiments, provided the input beam is slightly off-center.
In the nonlinear regime, the same model yields robust beam self-cleaning to a near-fundamental-mode output at 10 kW input power.
Coupling all modes, or only neighboring modes with n1 = n ± 1 and m1 = m ± 1, does not reproduce the experimental speckle or self-cleaning; only equal-n coupling works.
Because the fundamental mode is the sole n = 0 mode, its power is insensitive to degenerate coupling and can stabilize under Kerr nonlinearity.
Using the coupled-mode model with a large integration step reduces computation time compared with full three-dimensional nonlinear Schrödinger propagation.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If degenerate-mode coupling is indeed the operative perturbation, fiber designs that minimize non-degenerate perturbations, such as precisely controlled micro-bending, should show a sharper self-cleaning threshold, which could be tested experimentally.
The paper does not report the variance or full statistics of the coupling matrix; a natural next step is to extract these statistics from single-mode-launch experiments and test whether the model's quantitative predictions survive without tuning.
The equal-n coupling rule suggests a connection to mode condensation: the Kerr nonlinearity acts on a beam whose higher-order modes have already been randomized by disorder, which may link this work to hydrodynamic descriptions of multimode fibers.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper you asked about is a short numerical study of Kerr beam self-cleaning in graded-index multimode fiber. The thing worth knowing: the authors test three random-coupling models and find that only coupling among degenerate modes (equal mode number n) reproduces both the linear speckle and the nonlinear self-cleaning observed in their earlier experiment. That negative result is a useful contribution to model selection.

What the paper does well is the systematic comparison of coupling rules. The all-modes coupling gives rapid energy oscillations and no stable speckle; the n±1 coupling distorts but doesn't speckle; the equal-n coupling, with a slightly offset input, gives a realistic speckle and then self-cleaning at high power. The narrative is clear and the logic is honest about why the centered input fails (symmetry selection).

The soft spots are real, and one is load-bearing. Equation (6), the matrix form actually solved, has -1/2 C where Eq. (4) gives -i/2 C. With C Hermitian, -1/2 C is Hermitian, so the linear evolution is not unitary: total power would grow or decay, contradicting the paper's own requirement that C be Hermitian to preserve power. If the code solved Eq. (6) literally, the speckle could be a gain/loss artifact. If it solved Eq. (4), then Eq. (6) is a typo. Either way the manuscript can't be verified as written. This is not a nitpick; it's the central model.

The other issues are more standard. The "complete agreement" with experiment is supported by no quantitative comparison: no error bars, no side-by-side figure, no metric. The variance of the random coupling is never reported for the equal-n case, only a correlation length (10 cm) that appears in the nearest-neighbor model. No code or data is released. So the reader cannot reproduce the figures or judge whether the coupling strength was tuned.

The citation pattern is fine; the benchmark is the authors' own prior experiment, which is legitimate but means independent confirmation is still open.

This paper is for people actively modeling multimode fiber nonlinearities. It doesn't open a new direction, but it gives a concrete, computationally cheap model that explains a known effect. I'd send it to peer review, but a serious referee should demand a corrected Eq. (6), full parameter reporting, and a quantitative comparison before it can be considered verified.

Referee Report

3 major / 4 minor

Summary. The manuscript reports numerical simulations of Kerr beam self-cleaning in a graded-index multimode fiber using a coupled-mode model. Random linear coupling between modes is added to the base model of Ref. [14], and three coupling scenarios are compared: coupling between all modes, coupling between neighboring modes satisfying n1=n±1 and m1=m±1, and coupling only among degenerate modes with equal mode number n. The authors find that only the degenerate-mode coupling reproduces a speckled output in the linear regime and a stable self-cleaned beam in the nonlinear regime, and they claim that the results are in complete agreement with their own experimental data from Ref. [14].

Significance. If the model were fully specified and the agreement with experiment quantified, this would be a useful contribution: it offers a computationally efficient coupled-mode description and identifies degenerate-mode random coupling as the minimal ingredient that preserves the self-cleaning mechanism while generating realistic speckle. The paper's strengths are the systematic comparison of three coupling models, the direct calculation of modal overlap integrals, and the clear separation of linear and nonlinear regimes. The significance is currently limited, however, because the central claim of complete agreement is not backed by a single quantitative comparison, the random-coupling parameters are not reported, and the printed evolution equation appears to contain an error that breaks power conservation.

major comments (3)

[Section 2, Eqs. (4) and (6)] The matrix form in Eq. (6) reads dA/dζ = M A with M = -iD L/2 - (1/2) C - i p NL/2, but substituting Eq. (4) gives a coupling term of -(i/2) C A. Because C is declared Hermitian in order to preserve total power, the printed -C/2 term is Hermitian with the wrong sign of the imaginary unit, making the linear evolution non-unitary and capable of introducing spurious exponential growth or decay. If the leapfrog scheme in Eq. (9) solves Eq. (6) literally, the reported speckle and self-cleaning could be numerical artifacts of this gain or loss; if it solves Eq. (4), then Eq. (6) is a typo that must be corrected. In either case the text needs a clear statement of which equation is actually integrated.
[Section 3, random-coupling models] The parameters of the random coupling matrix C are never fully specified. The text states that each element is normally distributed with zero mean, but it does not give the standard deviation or variance of that distribution, and it gives a correlation length of 10 cm only for the n±1 model, not for the equal-mode-number model that is claimed to reproduce the experiments. The all-modes model is said to have coefficients that change at every integration step, which would make the correlation length equal to the step size rather than a physical value. Since the speckle contrast and the self-cleaning dynamics depend on the coupling strength, the model is under-specified and the reported agreement cannot be reproduced or checked.
[Section 3 and Conclusion] The statement that the numerical results are in complete agreement with the experimental data of Ref. [14] is not supported by any quantitative comparison in the manuscript. There is no side-by-side plot of simulated and measured output intensity profiles, no metric such as speckle contrast or output beam radius versus power, and no error bars. The conclusion repeats the claim without evidence. Please add at least one quantitative comparison, for example a plot of fundamental-mode fraction or output beam radius versus input power for the same fiber parameters and launch conditions as in Ref. [14].

minor comments (4)

[Section 3, first paragraph of Numerical results] The sentence 'the beam will not change during propagation in the linear regime' is not correct for Eq. (4) with C=0: the D(n+1)^2 term produces power-conserving phase evolution among modes and hence self-imaging of the beam. Please either account for this phase evolution in the discussion or state explicitly that D is neglected in that argument.
[Throughout] The symbol p is used both for normalized power (p = P/Psf) and as a mode index in A_{p,m} and in the subscripts of the overlap integrals, which makes equations such as (4) and (8) unnecessarily hard to read. Please use separate notation, for example P0 for normalized power.
[Throughout] There are several typographical and grammatical errors that should be corrected in a revision: 'Intoduction' in the section heading, 'proﬂe' instead of 'profile', 'we start out investigation', and 'completely agreement' in the abstract-like phrasing.
[Section 3, Eq. (9)] The numerical scheme is presented as an explicit leapfrog finite-difference method, but the text does not discuss the chosen integration step size, numerical stability, or convergence checks. A sentence reporting these details would be useful, especially because the random coefficients vary along the propagation distance.

Circularity Check

2 steps flagged · score 6.0 of 10

Validation reduces to self-citation and in-sample model selection: random-coupling parameters are unspecified and chosen to reproduce the authors' own [14] speckle data.

fitted input called prediction [Sec. 3, random-coupling model selection; Sec. 4 conclusion]
"Finally, we considered a model with random linear coupling between spatial modes with equal mode numbers n only."

The equal-n coupling rule was not derived from first principles; it was selected after trying all-mode and nearest-neighbour couplings, using the need to reproduce the experimentally observed speckle as the selection criterion ("To obtain a speckled intensity pattern in the linear regime, in further research we used the full system of equations (4) with a non-zero matrix C"). The variance of C is never reported, so the model has a free parameter that can be adjusted to match the output. Calling the resulting simulations "in complete agreement with our experimental data" is thus an in-sample fit presented as a prediction, not a parameter-free derivation.
self citation load bearing [Introduction and Conclusion; Ref. [14]]
"It was found that the results obtained using the model with a random linear coupling between modes with equal mode numbers are in complete agreement with available experimental data."

Ref. [14] (Podivilov et al., PRL 2018) supplies both the coupled-mode model ("in this work we use the coupled-mode model [14]") and the experimental data against which the agreement is asserted; several authors overlap with the present paper. No independent dataset, code reproduction, or machine-checked theorem is provided. Thus the validation chain that supports the paper's main claim terminates in the authors' own prior work, making the self-citation load-bearing.

full rationale

Score 6: partial circularity. The paper's own derivation is self-contained at the level of writing down the coupled-mode equations (Eq. (4)) and solving them numerically; no 'prediction' is mathematically equivalent to an input by construction. The circularity enters in the validation claim. The random-coupling matrix is specified only as Hermitian, zero-mean and normal, with no variance; the degenerate-mode coupling rule and 10 cm correlation length are chosen, after trying other rules, because they produce the observed speckle. The experiments and the base model are both from Ref. [14], whose authors overlap with the present paper. Therefore the 'complete agreement with our experimental data' is in-sample model selection plus self-citation rather than an independent confirmation. Separately, Eq. (6) writes the coupling term as -(1/2)C instead of -(i/2)C as required by Eq. (4); taken literally this makes the linear evolution non-unitary. That is a correctness/reproducibility issue, not itself a circular step.

Assumptions & free parameters 3 free parameters · 3 assumptions · 0 invented entities

The central claim rests on the choice of the random coupling model and its unstated strength; no new physical entities are introduced.

free parameters (3)

standard deviation of random coupling matrix C = not reported
The random coefficients are said to be normally distributed with zero mean, but the variance is never given; this controls the effective coupling strength and is likely chosen to make the output match experiments.
correlation length of random coupling = 10 cm
For the degenerate-mode model, the coupling is said to vary with a correlation length of 10 cm. This length is chosen without justification and affects the dynamics.
mode truncation order = n ≤ 16 (153 modes)
The simulation includes modes with mode number up to 16. This truncation is a numerical choice whose impact on the self-cleaning result is not studied.

assumptions (3)

standard math Rapidly oscillating four-wave mixing terms can be neglected (rotating wave approximation)
Used to derive the coupled-mode equations (4) from the NLSE (1); standard in mode-coupling theory but an approximation that may exclude some nonlinear dynamics.
domain assumption Fiber imperfections can be modeled as a Hermitian random linear coupling with zero-mean normal distribution, random along propagation
Introduced in Section 2 to produce speckle; the physical connection to stress, bending, or core fluctuations is asserted but not derived.
ad hoc to paper Coupling only among modes with equal mode number n models the relevant imperfections
Motivated by the failure of the other coupling models to reproduce experiments; this selection is made post hoc rather than derived from fiber perturbation theory.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Random Mode Coupling Assists Kerr Beam Self-Cleaning in a Graded-Index Multimode Optical Fiber." pith.science (2026). https://pith.science/paper/FTGCLRJE

@misc{pith2026190807745,
  author       = {Pith},
  title        = {Pith review of: Random Mode Coupling Assists Kerr Beam Self-Cleaning in a Graded-Index Multimode Optical Fiber},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FTGCLRJE}},
  note         = {Machine review of arXiv:1908.07745}
}

read the original abstract

In this paper, we numerically investigate the process of beam self-cleaning in a graded-index multimode optical fiber, by using the coupled-mode model. We introduce various models of random linear coupling between spatial modes, including coupling between all modes, or only between degenerate ones, and investigate the effects of random mode coupling on the beam self-cleaning process. The results of numerical investigations are in complete agreement with our experimental data.

Figures

Figures reproduced from arXiv: 1908.07745 by the authors.

**Figure 2.** Dynamics of energy distribution by modes (a) and the output field (b) for the [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗

**Figure 3.** Dynamics of energy distribution by modes for the model with random linear [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: Output field for the model with random linear coupling between modes with [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]

**Figure 5.** Figure 5: Dynamics of energy distribution by modes (a) and the output field (b) for the [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]

**Figure 6.** Figure 6: Output field for the model with random linear coupling between modes with [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

**Figure 7.** Figure 7: Dynamics of energy distribution by modes (a) and the output field (b) for [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[14]

E. V. Podivilov, D. S. Kharenko, V. A. Gonta, K. Krupa, O. S. Sidel- nikov, S. Turitsyn, M. P. Fedoruk, S. A. Babin, S. Wabnitz, Hydro- dynamic 2D turbulence and spatial beam condensation in multimode optical ﬁbers, Phys. Rev. Lett. 122 (2018) 103902. 12

work page 2018
[1]

D. J. Richardson, J. M. Fini, L. E. Nelson, Space-division multiplexing in optical ﬁbres, Nat. Photonics 7 (2013) 354–362

work page 2013
[2]

D. J. Richardson, J. Nilsson, W. A. Clarkson, High power ﬁber lasers: current status and future perspectives, J. Opt. Soc. Am. B 27 (2010) B63

work page 2010
[3]

L. G. Wright, D. N. Christodoulides, F. W. Wise, Controllable spa- tiotemporal nonlinear eﬀects in multimode ﬁbres, Nature Photonics (2015) 1–5

work page 2015
[4]

Krupa, A

K. Krupa, A. Tonello, A. Barth´ el´ emy, V. Couderc, B. M. Shalaby, A. Bendahmane, G. Millot, S. Wabnitz, Observation of Geometric Para- metric Instability Induced by the Periodic Spatial Self-Imaging of Mul- timode Waves, Phys. Rev. Lett. 116 (2016) 183901

work page 2016
[5]

W. H. Renninger, F. W. Wise, Optical solitons in graded-index multi- mode ﬁbres, Nat. Commun. 4 (2013)

work page 2013
[6]

L. G. Wright, W. H. Renninger, D. N. Christodoulides, F. W. Wise, Spatiotemporal dynamics of multimode optical solitons, Opt. Express 23 (2015) 3492–3506

work page 2015
[7]

L. G. Wright, S. Wabnitz, D. N. Christodoulides, F. W. Wise, Ultra- broadband Dispersive Radiation by Spatiotemporal Oscillation of Mul- timode Waves, Phys. Rev. Lett. 115 (2015) 223902

work page 2015

Show all 14 references

[8]

Longhi, Modulational instability and space time dynamics in nonlin- ear parabolic-index optical ﬁbers, Opt

S. Longhi, Modulational instability and space time dynamics in nonlin- ear parabolic-index optical ﬁbers, Opt. Lett. 28 (2007) 2363

2007
[9]

Lopez-Galmiche, Z

G. Lopez-Galmiche, Z. Sanjabi Eznaveh, M. A. Eftekhar, J. Antonio Lopez, L. G. Wright, F. Wise, D. Christodoulides, R. Amezcua Correa, Visible supercontinuum generation in a graded index multimode ﬁber pumped at 1064 nm, Opt. Lett. 41 (2016) 2553

2016
[10]

Krupa, C

K. Krupa, C. Louot, V. Couderc, M. Fabert, R. Guenard, B. M. Shal- aby, A. Tonello, D. Pagnoux, P. Leproux, A. Bendahmane, R. Dupiol, G. Millot, S. Wabnitz, Spatiotemporal characterization of supercon- tinuum extending from the visible to the mid-infrared in a multimode graded...

2016
[11]

Krupa, A

K. Krupa, A. Tonello, B. M. Shalaby, M. Fabert, A. Barth´ el´ emy, G. Mil- lot, S. Wabnitz, V. Couderc, Spatial beam self-cleaning in multimode ﬁbres, Nat. Photonics 11 (2017) 237–241

2017
[12]

Z. Liu, L. G. Wright, D. N. Christodoulides, F. W. Wise, Kerr self- cleaning of femtosecond-pulsed beams in graded-index multimode ﬁber, Opt. Lett. 41 (2016) 3675

2016
[13]

L. G. Wright, Z. Liu, D. A. Nolan, M. J. Li, D. N. Christodoulides, F. W. Wise, Self-organized instability in graded-index multimode ﬁbres, Nat. Photonics 10 (2016) 771–776

2016

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [14]

E. V. Podivilov, D. S. Kharenko, V. A. Gonta, K. Krupa, O. S. Sidel- nikov, S. Turitsyn, M. P. Fedoruk, S. A. Babin, S. Wabnitz, Hydro- dynamic 2D turbulence and spatial beam condensation in multimode optical ﬁbers, Phys. Rev. Lett. 122 (2018) 103902. 12

work page 2018

[2] [1]

D. J. Richardson, J. M. Fini, L. E. Nelson, Space-division multiplexing in optical ﬁbres, Nat. Photonics 7 (2013) 354–362

work page 2013

[3] [2]

D. J. Richardson, J. Nilsson, W. A. Clarkson, High power ﬁber lasers: current status and future perspectives, J. Opt. Soc. Am. B 27 (2010) B63

work page 2010

[4] [3]

L. G. Wright, D. N. Christodoulides, F. W. Wise, Controllable spa- tiotemporal nonlinear eﬀects in multimode ﬁbres, Nature Photonics (2015) 1–5

work page 2015

[5] [4]

Krupa, A

K. Krupa, A. Tonello, A. Barth´ el´ emy, V. Couderc, B. M. Shalaby, A. Bendahmane, G. Millot, S. Wabnitz, Observation of Geometric Para- metric Instability Induced by the Periodic Spatial Self-Imaging of Mul- timode Waves, Phys. Rev. Lett. 116 (2016) 183901

work page 2016

[6] [5]

W. H. Renninger, F. W. Wise, Optical solitons in graded-index multi- mode ﬁbres, Nat. Commun. 4 (2013)

work page 2013

[7] [6]

L. G. Wright, W. H. Renninger, D. N. Christodoulides, F. W. Wise, Spatiotemporal dynamics of multimode optical solitons, Opt. Express 23 (2015) 3492–3506

work page 2015

[8] [7]

L. G. Wright, S. Wabnitz, D. N. Christodoulides, F. W. Wise, Ultra- broadband Dispersive Radiation by Spatiotemporal Oscillation of Mul- timode Waves, Phys. Rev. Lett. 115 (2015) 223902

work page 2015