Direct observation of the spin-orbit coupling interaction in ring-core optical fibers

Miles Padgett; Siddharth Ramachandran; Steven D. Johnson; Zelin Ma

REVIEW 4 major objections 5 minor 20 references

Direct observation of the spin-orbit coupling interaction in ring-core optical fibers

T0 review · 4 major / 5 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 A single-pixel SWIR camera resolves OAM modes spaced 30 ps apart, directly observing spin-orbit coupling in a ring-core fiber and confirming that SOa/SOaa delays scale with the square of the mode order.

desk verdict Useful methods paper: 30 ps-resolved SWIR mode imaging in ring-core fiber; L² scaling claim is plausible but over-fitted to four selected points without error bars. read the letter →

arxiv 1908.08252 v1 pith:IHAPINCM submitted 2019-08-22 physics.optics

Steven D. Johnson , Zelin Ma , Miles Padgett , Siddharth Ramachandran This is my paper · ORCID

classification physics.optics

keywords spin-orbitcouplingoflightring-coreopticalfiberorbitalangularmomentummodessingle-pixelimagingshort-waveinfraredHadamardtransformmodegroupdelaydispersion

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 demonstrates that a single-pixel camera operating at 1550 nm can image orbital-angular-momentum modes at the output of a ring-core fiber with enough temporal resolution to separate modes whose arrival times differ by only 30 ps. Using Fourier deconvolution to remove the detector's 18.5 ns impulse response, the authors identify individual OAM modes and resolve each mode's spin-orbit-aligned (SOa) and anti-aligned (SOaa) components. The central result is that the measured SOa/SOaa arrival-time difference grows with the square of the OAM order $L$, matching the spin-orbit interaction picture with a polarization correction. This matters because mode dispersion limits multimode fiber communication, and direct time-resolved imaging of modes is a route to understanding and potentially controlling that dispersion.

What carries the argument

The central object is the single-pixel imaging system: a DMD displaying Hadamard masks, a high-speed photodetector, and a 50 GHz oscilloscope reconstruct the fiber output as a 32-by-32-pixel movie, while Fourier deconvolution with the measured impulse response sharpens the temporal signal. The physical identity carrying the quantitative claim is the spin-orbit splitting formula $\delta\beta = \frac{L}{k a 2 n_{\mathrm{co}}^2} \int_0^\infty |E(r)|^2 \frac{\partial \Delta n(r)}{\partial r}\,dr$, whose polarization-correction integral accounts for the index-gradient interaction. Because that integral is of order $L$ in small-core fibers, the arrival-time difference between SOa and SOaa modes is predicted to scale as $L^2$, and the deconvolved temporal peaks are what tie the measured delays to this scaling.

What would settle it

Inject a known OAM order with controlled circular polarization into the same fiber and measure the SOa/SOaa delay directly without relying on mode labelling from reconstructed images; if the extracted arrival-time differences do not scale with the square of the mode order, or do not grow linearly with fiber length for fixed mode order, the claimed spin-orbit scaling would be contradicted.

Watch

Extended reading notes

Core claim

The paper's claim is that the spin-orbit coupling of light inside a ring-core fiber, namely the splitting of each propagation constant into aligned and anti-aligned polarization/OAM states, can be seen directly in the time domain. The authors reconstruct movies of the fiber output from Hadamard-coded single-pixel measurements and show separate temporal peaks for SOa and SOaa modes. From a full 15 ns acquisition they extract the arrival-time difference between SOa and SOaa for the modes whose splitting exceeds the instrument resolution, and find that this difference plots linearly against $L^2$ with $R^2 = 0.999$. They take this as confirmation of the standard spin-orbit splitting relation, in which the propagation-constant splitting $\delta\beta$ is proportional to $L$ times a polarization-correction integral that itself scales with $L$, so that group-delay separations scale as $L^2$.

Load-bearing premise

The load-bearing premise is that the deconvolution and smoothing do not shift or invent temporal peaks, so that each extracted delay is the true group-delay separation of a correctly identified SOa/SOaa pair; if that fails for any of the four fitted mode orders, the reported $L^2$ scaling is not established.

Editorial extensions

If this is right

The technique resolves mode separations of 30 ps, an order-of-magnitude improvement in temporal resolution over earlier single-pixel light-in-flight work at this wavelength.
The observed $L^2$ scaling gives a practical rule for when SOa/SOaa modes in ring-core fibers will be separable: higher-order OAM modes separate faster.
Deconvolution of the instrument response is the step that makes picosecond-level mode identification possible with a nanosecond-response detector, so the same approach transfers to other slow detectors.
Adding polarization control and detection could isolate pure SOa/SOaa pairs and measure their interference directly, which the authors note as a next step.
The method applies beyond ring-core fibers to other multimode fibers where group-velocity differences between modes need to be characterized.

Reading between the lines

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

If the $L^2$ scaling holds for all supported orders, then in longer fibers the SOa/SOaa delay grows quadratically with mode order; this could make time-gated detection a mode-sorting mechanism without spatial mode filtering.
A decisive extension would vary fiber length and verify that measured delays scale linearly with length for fixed $L$, separating the material and waveguide contributions from the spin-orbit term.
The same Hadamard temporal-imaging scheme, with a faster detector, could push below 10 ps and might resolve the near-degenerate groups (e.g., SOa0,2/TE0,2/SOa1,2/TM0,2) that currently appear as a broad feature.
Because the measurement uses no polarization control, the petal interference between SOa and SOaa components complicates mode labelling; repeating with controlled input polarization would test whether the extracted delays are independent of coupling conditions.

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

Colleague, what you should know: this is a methods paper first and a physics claim second. The authors reuse their own single-pixel camera to image OAM modes in a ring-core fiber and resolve, after deconvolution, SOa/SOaa pairs separated by 30 ps. That is a genuine, useful capability – the best temporal resolution I've seen for imaging fiber modes at 1550 nm, and an order of magnitude better than their previous light-in-flight work. The images are clean and the mode assignments look reasonable. The deconvolution step is sensible, though parameters are under-reported.

The soft spot is the L² scaling argument in Section 3 and Fig. 4. They fit arrival-time differences vs L for L=±4 to ±7 and report R²=0.999. But the inclusion rule – only modes whose separation exceeds temporal resolution – is correlated with the measured values; lower L modes are left out precisely because they're smaller. With four points and a two-parameter fit, that selection can inflate R². No error bars on arrival times, and the Gaussian smoothing width after deconvolution is not stated. These are exactly the kinds of things a referee should ask for. The qualitative observation of 30 ps separation and the mode images stand; the quantitative confirmation of the L² law is weaker than the text suggests.

I don't think the paper is wrong in its central claim. The trend is consistent with theory, and the authors aren't fitting model constants to claim a prediction – they compare to a known polarization-correction model. But the overconfident phrasing ('clearly indicates') should be tempered, and the analysis should be made more transparent. The citation pattern is fine: self-citations to the fiber and prior camera papers are appropriate given the lineage.

Bottom line: this deserves peer review. It's not a major physics breakthrough, but it's a solid experimental technique paper with a clear niche and a new measurement. I'd send it to an editor for Optics Express or similar, with a request for error bars, a statement of the smoothing parameters, and either all data points or a discussion of the selection effect. A good referee can turn this into a more honest paper.

Referee Report

4 major / 5 minor

Summary. The paper reports time-resolved single-pixel imaging of orbital-angular-momentum (OAM) modes emerging from a ring-core optical fiber at 1550 nm. A 32x32 Hadamard single-pixel camera with a fast photodetector and oscilloscope is used to record temporal traces for each mask, and Fourier deconvolution with a Gaussian smoothing step is applied to remove the 18.5 ns detector impulse response. The authors identify SOa/SOaa mode pairs separated by as little as 30 ps, present images of the modes, and claim in Section 3 that the measured arrival-time differences between SOa and SOaa modes follow an L^2 scaling, supported by a linear regression with R^2=0.999 over the four mode orders L=±4 to ±7 for which the separation was measurable.

Significance. If the result holds, the paper demonstrates a practical route to picosecond-level time-resolved imaging of fiber modes at telecommunications wavelengths using a single-pixel camera, and provides the first direct time-domain observation of spin-orbit-coupling-induced mode splitting in a ring-core fiber. The strengths of the work are the clear experimental design, the use of deconvolution to overcome a slow detector response, and the explicit comparison of the measured separations with a predicted L^2 scaling. The paper does not ship code or data, and it does not provide machine-checked derivations; its reproducibility rests on the description of the apparatus. The central quantitative claim, however, is currently supported by a small and selectively chosen dataset without uncertainty estimates, so the significance is real but the quantitative conclusion is not yet secure.

major comments (4)

[§3, Fig. 4] The L^2-scaling claim rests on a linear regression of only four data points (L = ±4 to ±7), with no error bars, no reported slope or intercept, and no statement of the fit procedure. The caption states that points were included only when the SOa/SOaa separation was 'larger than the temporal resolution of our experiment,' so the inclusion criterion is explicitly correlated with the measured response variable. This truncation can bias the fit and inflate R^2. The manuscript should report all extracted separations, including unresolved modes as upper limits, justify or prespecify the inclusion rule, and provide uncertainties for the extracted arrival-time differences.
[§2, Fig. 2] Arrival times are extracted from signals after Fourier deconvolution and Gaussian smoothing, but the smoothing width is not specified and no sensitivity analysis is presented. Because the detector has an 18.5 ns ringing impulse response and the signals are sampled over a 15 ns window (1 ns for the higher-resolution acquisitions), deconvolution can shift or create temporal peaks. The claimed 30 ps separation and the Fig. 4 data are therefore not yet quantitatively established. The authors should quantify the uncertainty in peak positions, for example by varying the regularization or smoothing parameters and by validating the processing on synthetic signals with known mode separations.
[§3, Fig. 3] The assignment of measured peaks to SOa/SOaa pairs is made from reconstructed 32x32 images, but several modes are near-degenerate (for example, the SOa0,2/TE0,2/SOa1,2/TM0,2 group is described as a 'low haystack'). Without an explicit labeling criterion, such as petal count, ring radius, or comparison with a simulated mode set, and without an uncertainty analysis, a misidentification of the peaks used in Fig. 4 would change the reported scaling. Please specify the labeling rule and test its robustness.
[§3, Eq. (2)] The prediction that the separation scales as L^2 relies on two assertions: that the polarization correction integral is 'typically of order L' and that the group-velocity splitting scales in the same way as the propagation-constant splitting. Neither step is derived in the manuscript or tied to a specific equation in the cited references. If these assumptions do not hold for the ring-core fiber used here, the L^2 comparison is not a valid test of the model. Please add the derivation or a precise reference and state the conditions under which the L^2 scaling is expected.

minor comments (5)

[§1 and §3] There are typographical errors: 'expresses as an integer value L' should be 'expressed as an integer value L', and 'direction observation' should be 'direct observation'.
[§2] The statement that 'there was little difference in the results when either method was used' should be quantified or removed, since the choice of deconvolution domain is part of the data-processing pipeline.
[§3] The claim that an 18 ps detector response would allow a 10 ps difference to be detected is unsupported; a quantitative criterion, such as separation relative to the deconvolved pulse width and the signal-to-noise ratio, is needed.
[§3, Eq. (2)] The notation 'ka 2n2co' is ambiguous; it should be written as k a^2 2 n_co^2 (or equivalent) so that the denominator is clear.
[§2, §3] The specific models of the pulsed laser, the photodetector, and the Goldeye comparison camera are named but their key specifications are not all given in the text; adding the relevant data-sheet parameters would improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity found: the reported L^2 scaling is a measured comparison against an independent theoretical model, not a fitted parameter renamed as a prediction.

full rationale

The paper's derivation chain is: (1) a single-pixel camera records time-resolved images of OAM modes via Hadamard-mask measurements; (2) Fourier deconvolution and Gaussian smoothing separate temporally close modes; (3) arrival-time differences between SOa and SOaa modes are extracted; (4) these differences are plotted against mode order L and compared with Eq. (2), the standard polarization-correction expression from Refs. [17,19,20], which predicts an L^2 scaling. The model constants and the polarization-correction integral are not fitted to the measured arrival times; the paper tests only the predicted functional form. The linear regression in Fig. 4 is a data-analysis fit, not a model parameter fit, so the R^2 value does not reduce to the model being used as an input. The self-citations to Ref. [7] (the fiber) and Ref. [15] (the prior single-pixel technique) are contextual: they identify the fiber and the instrumentation, but the central observation of 30 ps-separated SOa/SOaa modes and their arrival-time scaling is new, directly measured data. The concern that the Fig. 4 point selection is correlated with temporal resolution is a statistical/robustness issue, not a circularity: it does not make the measured L^2 dependence equivalent by construction to the input model. No load-bearing circular step is identifiable in the paper's argument.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper contributes a measurement rather than a new theory. It introduces no new particles, forces, or conserved quantities. Its quantitative conclusion depends on a known formula, Eq. (2), and on two analysis choices: the Gaussian smoothing width and the linear fit to four selected data points. No raw data or code are provided.

free parameters (2)

Gaussian smoothing width = not stated
Applied after Fourier deconvolution to produce the final temporal signals in Fig. 2; the width is chosen by the authors and can shift or broaden temporal peaks, yet no value or uncertainty is reported.
Linear regression slope and intercept in Fig. 4 = not reported
The line relating arrival-time difference to L² is fit to four measured separations; only R²=0.999 is reported, and the slope is not compared with the polarization-correction integral prediction.

assumptions (5)

standard math The Hadamard transform reconstruction in Eq. (1) correctly inverts the mask measurements to form the image at each time bin.
Invoked in Section 2; standard linear algebra for orthogonal binary patterns.
standard math The measured signal is the convolution of the true optical signal with a stationary impulse response, so Fourier deconvolution is valid.
Invoked in Section 2; the impulse response is measured separately from the laser, and deconvolution assumes it applies unchanged to all mode measurements.
domain assumption The polarization-correction integral in Eq. (2), taken from [7,19,20], correctly describes spin-orbit splitting of propagation constants in this fiber.
Invoked in Section 3; the paper does not derive or independently verify this theory.
domain assumption The fiber output mode pattern is stable over the roughly 20 minute acquisition across 2048 DMD masks.
Invoked in Section 3; slow drift or mode mixing during acquisition would corrupt the reconstructed time-resolved images.
domain assumption Each temporal peak corresponds to one mode or near-degenerate mode group, and the L labels assigned from the reconstructed images are correct.
Invoked in Section 3 and Fig. 3; near-degenerate groups such as SOa0,2/TE0,2/SOa1,2/TM0,2 are treated as single arrival-time features.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Direct observation of the spin-orbit coupling interaction in ring-core optical fibers." pith.science (2026). https://pith.science/paper/IHAPINCM

@misc{pith2026190808252,
  author       = {Pith},
  title        = {Pith review of: Direct observation of the spin-orbit coupling interaction in ring-core optical fibers},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/IHAPINCM}},
  note         = {Machine review of arXiv:1908.08252}
}

read the original abstract

Ring-core optical fibers have been designed to carry orbital angular momentum modes. We demonstrate the imaging of these modes, individually identifying modes separated temporally by only 30~ps. A single-pixel camera operating in the short-wave infrared detection range is used to image the 1550~nm wavelength optical modes. With this technique, examination of these optical modes can be performed with significantly higher temporal resolution than is possible with conventional imaging systems, such that the imaging of modes separated by spin-orbit coupling is achieved and evaluated. Deconvolution is required to separate the instrument response from the optical mode signal, increasing the clarity and temporal resolution of the measurement system.

Figures

Figures reproduced from arXiv: 1908.08252 by the authors.

**Figure 1.** The layout of the optical fiber and single-pixel camera. The pulsed laser is incident on a spatial light modulator (SLM) to precisely adjust the coupling into the optical fiber. The light is coupled into and travels through the ring-core fiber, and the collimated output is incident on a digital micromirror device (DMD). The signal reflected by each mask displayed on the DMD is measured by a fiber-coupled high-speed … view at source ↗

**Figure 2.** Example of the deconvolution process. Using the data set presented later in figure 3c) an example of the deconvolution is shown. The raw data a) is deconvolved with the measured impulse response b), producing the resulting signal c). The signal is then Gaussian smoothed to produce the final signal d). pattern displayed on the DMD, with respect to a trigger signal from the picosecond laser. The measurement was transf… view at source ↗

**Figure 3.** Measurement of the optical modes. Each data set is shown with the total signal measured over time and the modes present for the peaks indicated by the arrows. a) shows the full measurement of all present modes. The modes associated with single OAM are shown below the time series (modes L = ±1 to L = ±7), other modes are shown above. The image b) is the summed signal for all time bins, showing similar signal to c), a… view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Difference in arrival times as a function of mode order L. Through the analysis of the propagation times measured in figure 3a) for the modes L = ±4 to L = ±7, where the time delay between the SOa and SOaa modes is larger than the temporal resolution of our experiment,…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

20 extracted references · 20 canonical work pages

[1]

Capacity trends and limits of optical communication networks,

R.-J. Essiambre and R. W. Tkach, “Capacity trends and limits of optical communication networks,” Proc. IEEE100, 1035–1055 (2012)

work page 2012
[2]

Mode-multiplexed transmission over conventional graded-index multimode ﬁbers,

R. Ryf, N. K. Fontaine, H. Chen, B. Guan, B. Huang, M. Esmaeelpour, A. H. Gnauck, S. Randel, S. Yoo, A. Koonen, R. Shubochkin, Y. Sun, and R. Lingle, “Mode-multiplexed transmission over conventional graded-index multimode ﬁbers,” Opt. Express23, 235–246 (2015)

work page 2015
[3]

4x10 Gbit/s bidirectional transmissionover2kmofconventionalgraded-indexOM1multimodeﬁberusingmodegroupdivisionmultiplexing,

K. Lenglé, X. Insou, P. Jian, N. Barré, B. Denolle, L. Bramerie, and G. Labroille, “4x10 Gbit/s bidirectional transmissionover2kmofconventionalgraded-indexOM1multimodeﬁberusingmodegroupdivisionmultiplexing,” Opt. Express24, 28594–28605 (2016)

work page 2016
[4]

Optical vortices in ﬁber,

S. Ramachandran and P. Kristensen, “Optical vortices in ﬁber,” Nanophotonics2, 455–474 (2013)

work page 2013
[5]

Terabit-scale orbital angular momentum mode division multiplexing in ﬁbers,

N. Bozinovic, Y. Yue, Y. Ren, M. Tur, P. Kristensen, H. Huang, A. E. Willner, and S. Ramachandran, “Terabit-scale orbital angular momentum mode division multiplexing in ﬁbers,” Science340, 1545–1548 (2013)

work page 2013
[6]

Orbital angular momentum of light and the transformation of Laguerre-Gaussian laser modes,

L. Allen, M. W. Beijersbergen, R. J. C. Spreeuw, and J. P. Woerdman, “Orbital angular momentum of light and the transformation of Laguerre-Gaussian laser modes,” Phys. Rev. A45, 8185–8189 (1992)

work page 1992
[7]

Conservationoforbital angularmomentumin air-coreoptical ﬁbers,

P.Gregg, P. Kristensen, andS.Ramachandran, “Conservationoforbital angularmomentumin air-coreoptical ﬁbers,” Optica 2, 267–270 (2015)

work page 2015
[8]

Atrillionframespersecond: thetechniquesandapplicationsoflight-in-ﬂightphotography

D.FaccioandA.Velten,“Atrillionframespersecond: thetechniquesandapplicationsoflight-in-ﬂightphotography.” Reports on Prog. Phys.81, 105901 (2018)

work page 2018

Show all 20 references

[9]

3D imaging with 128× 128 eye safe InGaAs p-i-n lidar camera,

C. P. Bradley, S. S. Mukherjee, A. D. Reinhardt, P. F. McManamon, A. O. Lee, and V. Dhulla, “3D imaging with 128× 128 eye safe InGaAs p-i-n lidar camera,” Proc. SPIE 11005, Laser Radar Technol. Appl. XXIV (2019)

2019
[10]

Cross-correlated (C2) imaging of ﬁber and waveguide modes,

D. N. Schimpf, R. A. Barankov, and S. Ramachandran, “Cross-correlated (C2) imaging of ﬁber and waveguide modes,” Opt. Express19, 13008–13019 (2011)

2011
[11]

Sub-second mode measurement of ﬁbers using C2 imaging,

J. Demas and S. Ramachandran, “Sub-second mode measurement of ﬁbers using C2 imaging,” Opt. Express22, 23043–23056 (2014)

2014
[12]

Spatially and spectrally resolved imaging of modal content in large-mode-area ﬁbers,

J. W. Nicholson, A. D. Yablon, S. Ramachandran, and S. Ghalmi, “Spatially and spectrally resolved imaging of modal content in large-mode-area ﬁbers,” Opt. Express16, 7233–7243 (2008)

2008
[13]

Principles and prospects for single-pixel imaging,

M. P. Edgar, G. M. Gibson, and M. J. Padgett, “Principles and prospects for single-pixel imaging,” Nat. Photonics13, 13–20 (2019)

2019
[14]

Single-photon sensitive light-in-ﬁght imaging,

G. Gariepy, N. Krstajić, R. Henderson, C. Li, R. R. Thomson, G. S. Buller, B. Heshmat, R. Raskar, J. Leach, and D. Faccio, “Single-photon sensitive light-in-ﬁght imaging,” Nat. Commun.6, 6021 (2015)

2015
[15]

A light-in-ﬂight single-pixel camera for use in the visible and short-wave infrared,

S. D. Johnson, D. B. Phillips, Z. Ma, S. Ramachandran, and M. J. Padgett, “A light-in-ﬂight single-pixel camera for use in the visible and short-wave infrared,” Opt. Express27, 9829–9837 (2019)

2019
[16]

Hadamard transform image coding,

W. K. Pratt, J. Kane, and H. C. Andrews, “Hadamard transform image coding,” Proc. IEEE57, 58–68 (1969)

1969
[17]

Spin and orbital rotation of electrons and photons via spin-orbit interaction,

C. C. Leary, M. G. Raymer, and S. J. van Enk, “Spin and orbital rotation of electrons and photons via spin-orbit interaction,” Phys. Rev. A80, 061804 (2009)

2009
[18]

Generation and propagation of radially polarized beams in optical ﬁbers,

S. Ramachandran, P. Kristensen, and M. F. Yan, “Generation and propagation of radially polarized beams in optical ﬁbers,” Opt. Lett.34, 2525–2527 (2009)

2009
[19]

Bures,Guided Optics: Optical Fibers and All-ﬁber Components(Wiley, 2008)

J. Bures,Guided Optics: Optical Fibers and All-ﬁber Components(Wiley, 2008)

2008
[20]

Snyder and J

A. Snyder and J. Love,Optical Waveguide Theory(Springer, 1983)

1983

Pith tools

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

[1] [1]

Capacity trends and limits of optical communication networks,

R.-J. Essiambre and R. W. Tkach, “Capacity trends and limits of optical communication networks,” Proc. IEEE100, 1035–1055 (2012)

work page 2012

[2] [2]

Mode-multiplexed transmission over conventional graded-index multimode ﬁbers,

R. Ryf, N. K. Fontaine, H. Chen, B. Guan, B. Huang, M. Esmaeelpour, A. H. Gnauck, S. Randel, S. Yoo, A. Koonen, R. Shubochkin, Y. Sun, and R. Lingle, “Mode-multiplexed transmission over conventional graded-index multimode ﬁbers,” Opt. Express23, 235–246 (2015)

work page 2015

[3] [3]

4x10 Gbit/s bidirectional transmissionover2kmofconventionalgraded-indexOM1multimodeﬁberusingmodegroupdivisionmultiplexing,

K. Lenglé, X. Insou, P. Jian, N. Barré, B. Denolle, L. Bramerie, and G. Labroille, “4x10 Gbit/s bidirectional transmissionover2kmofconventionalgraded-indexOM1multimodeﬁberusingmodegroupdivisionmultiplexing,” Opt. Express24, 28594–28605 (2016)

work page 2016

[4] [4]

Optical vortices in ﬁber,

S. Ramachandran and P. Kristensen, “Optical vortices in ﬁber,” Nanophotonics2, 455–474 (2013)

work page 2013

[5] [5]

Terabit-scale orbital angular momentum mode division multiplexing in ﬁbers,

N. Bozinovic, Y. Yue, Y. Ren, M. Tur, P. Kristensen, H. Huang, A. E. Willner, and S. Ramachandran, “Terabit-scale orbital angular momentum mode division multiplexing in ﬁbers,” Science340, 1545–1548 (2013)

work page 2013

[6] [6]

Orbital angular momentum of light and the transformation of Laguerre-Gaussian laser modes,

L. Allen, M. W. Beijersbergen, R. J. C. Spreeuw, and J. P. Woerdman, “Orbital angular momentum of light and the transformation of Laguerre-Gaussian laser modes,” Phys. Rev. A45, 8185–8189 (1992)

work page 1992

[7] [7]

Conservationoforbital angularmomentumin air-coreoptical ﬁbers,

P.Gregg, P. Kristensen, andS.Ramachandran, “Conservationoforbital angularmomentumin air-coreoptical ﬁbers,” Optica 2, 267–270 (2015)

work page 2015

[8] [8]

Atrillionframespersecond: thetechniquesandapplicationsoflight-in-ﬂightphotography

D.FaccioandA.Velten,“Atrillionframespersecond: thetechniquesandapplicationsoflight-in-ﬂightphotography.” Reports on Prog. Phys.81, 105901 (2018)

work page 2018