Light's Topology Shrugs Off Turbulence

Scientists Transmit Data via Magnetic Skyrmions in Laser Beams with 98% Fidelity Over Hundreds of Meters

A laser beam scrambled beyond recognition still carried its data intact across a 270-meter open-air link. The trick was to encode information not in the beam's shape, but in a topological invariant no amount of atmospheric churn can erase.

---

BLUF (Bottom Line Up Front): A team from the University of the Witwatersrand (Wits) in Johannesburg and the University of Bordeaux has sent optical skyrmions—particle-like topological textures woven into a laser beam's polarization field—across a 270-meter free-space link on the Wits campus and recovered the encoded topological number with roughly 98 percent fidelity under most conditions, degrading to about 86 percent only in the harshest midday turbulence. Crucially, the recovery used no adaptive optics and no turbulence pre-measurement: a single-shot polarization-camera measurement reads the invariant directly. The result, reported in Science Advances, is a clean experimental confirmation that topological encoding survives real-world atmospheric distortion. Two cautions temper the excitement. First, this is a robustness proof-of-principle—it demonstrates that an encoded topological number survives the channel, not a high-throughput data link; the gigabit-to-terabit rates circulating in some secondary coverage are not part of the experiment. Second, and contrary to several popular write-ups, these are optical skyrmions (topological structures in light's electromagnetic field), not the magnetic skyrmions found in thin-film spintronics. The physics is real; the "magnetic laser beam" framing is not.

---

The problem: turbulence eats structured light

Free-space optical (FSO) communication—laser links through open air rather than glass fiber—has an enduring appeal for last-mile connectivity, campus and rooftop hops, drone-to-ground links, and satellite-to-ground terminals. It offers fiber-class bandwidth with no trenching and no spectrum licensing. Its enemy is the atmosphere. Temperature gradients and turbulent eddies scramble the refractive index on millisecond timescales, distorting a beam's amplitude, phase, and polarization.

For engineers who came up through radar and coherent RF, the failure mode is familiar in a new guise. Conventional structured-light schemes buy extra channel capacity by multiplexing spatial modes—orbital angular momentum (OAM) states, for instance—but turbulence induces mode crosstalk that mixes those channels together, exactly as multipath and scintillation corrupt a spatial waveform. The usual countermeasure is adaptive optics: sense the wavefront error with a wavefront sensor, then correct it with a deformable mirror in a fast feedback loop. It works, but it adds cost, latency, and mechanical complexity, and it must keep up with the turbulence in real time.

The Wits–Bordeaux approach sidesteps that loop entirely. Instead of correcting the distortion, it encodes information in a quantity that the distortion cannot change.

What an optical skyrmion actually is

A skyrmion, borrowed originally from nuclear and condensed-matter physics, is a field configuration whose winding—how many times a field vector wraps around a sphere as you traverse the plane—is a topological invariant. That integer wrapping number can't change under smooth deformation; altering it requires a discontinuity, which is energetically forbidden for small perturbations.

An optical skyrmion realizes this in the electromagnetic field of a structured light beam rather than in the magnetization of a material. The Wits team builds one as a vector beam: a coherent superposition of spatial modes (Laguerre–Gaussian, carrying OAM) placed in orthogonal polarization states, so that the beam's polarization vector traces out a full skyrmionic texture across its cross-section. The encoded value is the skyrmion, or wrapping, number—an integer. Andrew Forbes, the Wits group leader and corresponding author, has offered the standard topologist's analogy: a coffee mug and a doughnut are "the same" because each has exactly one hole, and you can knead either shape without changing that count. Turbulence kneads the beam; the hole count survives.

That is the conceptual departure from OAM multiplexing. OAM encodes data in a mode index that turbulence smears through crosstalk. Skyrmion encoding places the data in a global topological property of the whole vector field, which is invariant to the local, continuous distortions turbulence imposes—provided the channel is (approximately) unitary, i.e., it reshuffles the field without destroying its coherence.

The experiment

The team generated skyrmion-bearing vector beams with standard structured-light optics—spatial light modulators and an interferometric combiner to co-align orthogonally polarized spatial modes—and launched them across a 270-meter horizontal path between two buildings on the Wits campus in central Johannesburg. The link was deliberately run across a full daily cycle, from calm, cool morning air to the violent seeing of midday heat, sampling a wide range of turbulence strengths in the wild rather than in a lab-simulated phase screen.

Recovery is where the engineering is cleanest. At the receiver, the beam—by then visibly mangled—passes through a 50:50 beam splitter and a quarter-wave plate onto a polarization-sensitive ("division-of-focal-plane") camera, which captures all the projections needed to reconstruct the four Stokes parameters across the beam in a single shot. From that spatially resolved Stokes map, the wrapping number is computed directly. There is no wavefront sensor, no deformable mirror, no iterative optimization, and—contrary to the uploaded secondary account—no machine-learning classifier trained on distorted skyrmions. The topology is simply measured.

The headline numbers: the wrapping number was recovered correctly with about 98 percent fidelity across most of the day's conditions, falling to roughly 86 percent only under the most extreme turbulence. The authors also probed the regime beyond the atmospheric coherence time τ₀—the interval over which the channel looks momentarily "frozen" to the beam—and showed that while the beam's degree of polarization decays as the medium fluctuates (a decoherence-like effect), the topological number remains stable. As a visual proof of concept, they encoded a small multi-value message by mapping symbols to distinct wrapping numbers and read it back across the link.

What it is—and isn't

For a working communications engineer, the honest framing matters. This experiment establishes channel robustness of a topological alphabet, not a demonstrated high-bit-rate link. The recovered quantity is a small integer per beam; the achievable throughput depends on how many distinguishable skyrmion states you can pack, how fast you can generate and measure them, and how they multiplex—none of which this paper set out to maximize. Claims of "several gigabits per second" scaling to "terabit" capacity, seen in some coverage, are extrapolation, not result.

What the work does deliver is a compelling systems argument: a passive, compute-light receiver that skips the adaptive-optics feedback loop entirely. In link-budget and latency terms, removing the deformable-mirror control loop is attractive for compact, pole-mounted, or airborne terminals where size, weight, and power are at a premium. It is also a natural fit for links to and from space, where the turbulence is concentrated in the ground-side boundary layer and a correction-free decoder is appealing.

The obvious limitations are the ones physics imposes. Topological protection assumes a roughly unitary, coherence-preserving channel. Fog, heavy rain, and snow scatter and absorb the light—non-unitary loss that no winding number can survive—so this is a clear-air technique. Generation still relies on bulky bench optics today, though integrated photonic skyrmion sources (metasurface fibers and on-chip generators) are advancing quickly. And the alphabet size and symbol rate that make this competitive with mature intensity/phase modulation remain to be demonstrated.

Where it sits in the field

This is not an isolated flash. The same Wits group had already shown, in 2023, that vectorial structured light retains recoverable structure across a real outdoor link, and in early 2026 a companion effort with Yijie Shen's group at Nanyang Technological University demonstrated topological resilience of both classical and quantum optical skyrmions through lab-controlled turbulence—finding that even as entanglement degrades, the topological character of quantum skyrmion states persists. Related lines of work include quantum "nonlocal" skyrmions as noise-resilient entangled states, topological protection of skyrmions through scattering media, the stability of optical knots in turbulence, and a decade of turbulence-resilient vector-beam communication going back to compensation-free SPDPSK links. The foundational demonstration of optical skyrmion lattices in evanescent fields dates only to 2018, which makes the jump to an outdoor communications link in under a decade a notably fast translation from exotic physics to field experiment.

On the intellectual-property and legal front: a targeted search of patent databases and dockets turned up an active FSO patent landscape—turbulence-mitigation methods, patterned-light demultiplexing with convolutional neural networks, non-mechanical beam steering for satellite links, and ultra-wideband FSO apparatus among them—but no litigation or court filings specific to optical-skyrmion communication, which remains squarely in the pre-commercial research phase. The primary "official release" is the Wits University news announcement accompanying the Science Advances publication; there is no standards activity yet defining interoperable skyrmion formats.

The takeaway

The elegant part is philosophical as much as technical: rather than fighting the channel, encode in a quantity the channel can't touch. That the invariant survived a Johannesburg summer afternoon, measured by a single camera frame with no correction loop, is a genuine milestone for turbulence-resilient FSO. Whether it becomes a deployed modulation format will hinge on the unglamorous engineering the paper leaves open—alphabet size, symbol rate, multiplexing, integrated sources, and honest link budgets against clear-air scatter. The physics has cleared its bar. The systems case is the next experiment.

---

Verified sources

Primary research

1. C. Peters, V. Hakobyan, A. Drozdov, E. Brasselet, M. Cox, and A. Forbes, "Topological robustness of optical skyrmions through a real-world free-space link," Science Advances (2026). DOI: 10.1126/sciadv.aee2671 — https://doi.org/10.1126/sciadv.aee2671
2. Preprint of the above: arXiv:2602.04446 — https://arxiv.org/abs/2602.04446
3. Z. Guo, C. Peters, N. Mata-Cervera, A. N. Vetlugin, R. Guo, et al., "Topological robustness of classical and quantum optical skyrmions in atmospheric turbulence," Nature Communications (2026). DOI: 10.1038/s41467-026-68751-3 — https://doi.org/10.1038/s41467-026-68751-3

Official release

4. University of the Witwatersrand, "Wits researchers use light's topology to beat atmospheric distortion," Wits News, August 2026 — https://www.wits.ac.za/news/latest-news/research-news/2026/--2026-08/wits-researchers-use-lights-topology-to-beat-atmospheric-distortion.html

Key related literature

5. C. Peters, M. Cox, A. Drozdov, and A. Forbes, "The invariance and distortion of vectorial light across a real-world free-space link," Applied Physics Letters 123(2) (2023). DOI: 10.1063/5.0155203 — https://doi.org/10.1063/5.0155203
6. Y. Shen, et al., "Optical skyrmions and other topological quasiparticles of light," Nature Photonics 18, 15–25 (2024). DOI: 10.1038/s41566-023-01325-7 — https://doi.org/10.1038/s41566-023-01325-7
7. S. Tsesses, et al., "Optical skyrmion lattice in evanescent electromagnetic fields," Science 361, 993–996 (2018). DOI: 10.1126/science.aau0227 — https://doi.org/10.1126/science.aau0227
8. I. Nape, et al., "Revealing the invariance of vectorial structured light in complex media," Nature Photonics 16, 538–546 (2022). DOI: 10.1038/s41566-022-01023-w — https://doi.org/10.1038/s41566-022-01023-w
9. P. Ornelas, I. Nape, R. de Mello Koch, and A. Forbes, "Nonlocal skyrmions as topologically resilient quantum entangled states of light," Nature Photonics 18, 258–266 (2024). https://doi.org/10.1038/s41566-023-01360-4
10. P. Ornelas, I. Nape, R. de Mello Koch, and A. Forbes, "Topological rejection of noise by quantum skyrmions," Nature Communications 16, 2934 (2025). https://doi.org/10.1038/s41467-025-58208-4
11. A. A. Wang, et al., "Topological protection of optical skyrmions through complex media," Light: Science & Applications 13, 314 (2024). https://doi.org/10.1038/s41377-024-01679-9
12. L. Wang, et al., "The robustness of skyrmion numbers of structured optical fields in atmospheric turbulence," Optics Communications 579, 131568 (2025). arXiv:2410.05999 — https://arxiv.org/abs/2410.05999
13. G. Xie, et al. (Willner group), "Compensation-free high-dimensional free-space optical communication using turbulence-resilient vector beams," Nature Communications 12, 1666 (2021). https://doi.org/10.1038/s41467-021-21793-1
14. D. Pires, et al. (Litchinitser group), "Stability of optical knots in atmospheric turbulence," Nature Communications 16, 3001 (2025). https://doi.org/10.1038/s41467-025-58245-z
15. Optical skyrmion generation via integrated metafiber, Nature Communications (2024). DOI: 10.1038/s41467-024-54207-z — https://doi.org/10.1038/s41467-024-54207-z

Selected press coverage

16. Tech Xplore / Phys.org, "Information encoded in a beam of light can survive real-world atmospheric turbulence," August 2026 — https://techxplore.com/news/2026-08-encoded-survive-real-world-atmospheric.html
17. Interesting Engineering, "World-first test sends information through light beams without loss," August 2026 — https://interestingengineering.com/innovation/world-first-experiment-light-signals-chaotic-air
18. Photonics Online, "Wits researchers use light's topology to beat atmospheric distortion," August 2026 — https://www.photonicsonline.com/doc/wits-researchers-use-light-s-topology-beat-atmospheric-distortion-0001

Patent-landscape context (no skyrmion-specific litigation found)

19. T. Doster, et al., "Method for free space optical communication utilizing patterned light and convolutional neural networks," U.S. Patent 10,187,171 — https://patents.google.com/patent/US10187171
20. "System and method for transmitting signals in a free space optics communication system" (turbulence/interception mitigation), U.S. Prov. App. 63/813,253, filed May 2025 — https://patents.google.com/patent/US12457038

---

Editor's note on the source article: the uploaded WebProNews/TechRadar write-up describes this work as encoding data in "magnetic skyrmions" generated by "magnetic thin-film emitters" and decoded by a machine-learning algorithm over 300 meters. Those specifics are incorrect. The experiment uses optical (electromagnetic-field) skyrmions in a vector laser beam, generated with spatial light modulators and an interferometer, decoded by direct single-shot Stokes polarimetry, over a 270-meter link. The ~98% / ~86% fidelity figures are approximately right; the gigabit/terabit data-rate and multiplexing claims are extrapolations not demonstrated in the paper.