Weirdly Cited

In the South Pole ice behind the 2026 physics Nobel, about twice as much light reaches sensors lined up with the glacier's flow as sensors across it. Why?

The leading explanation is the ice crystals. Ice splits light into two rays. The IceCube team thinks that light crossing one crystal after another gets nudged, a little at a time, toward the line the glacier flows along.

It is not settled. The team calls it "a possible explanation". The model needs small corrections, plus one extra absorption term with no known cause.

The claims, and the sentence behind each

The Short says five of these out loud: that the South Pole ice is a glacier, that about twice the light arrives along the flow as across it at about 125 m, that the ice crystals are the leading explanation, that the study calls them a possible explanation, and that it is not settled. The prize and the other entries below are not spoken in it.

The prize was for IceCube

The 2026 Nobel Prize in Physics went to Francis Halzen. It was announced on 6 October 2026, for his work on IceCube and for the discovery of high-energy neutrinos from astrophysical sources.

"The Royal Swedish Academy of Sciences has decided to award the Nobel Prize in Physics 2026 to Francis Halzen"

"for decisive contributions to the IceCube Neutrino Observatory and the discovery of high-energy neutrinos of astrophysical origin"

Source: Nobel Prize in Physics 2026 press release, nobelprize.org.

A cubic kilometre of ice, wired for light

IceCube is about one cubic kilometre of deep glacial ice at the South Pole, fitted with 5,160 light sensors. They watch for the light that fast charged particles give off as they pass through the ice.

"It uses 5160 photomultipliers to detect Cherenkov light emitted by charged relativistic particles."

Source: Abstract.

The detector flashes its own lights

The sensors sit between 1,450 m and 2,450 m down. Each one carries 12 LEDs that fire pulses of known brightness into the ice, so the team can work out how the ice handles light. Those flashes are the data behind this result. The paper calls each sensor a DOM.

"each DOM features 12 LEDs which can emit light pulses of known intensity and duration into the ice and which are used to calibrate the optical properties of the instrumented ice"

Source: The detector.

About twice the light along the flow

This is the measurement. At about 125 m, roughly twice as much LED light arrives at sensors lined up with the ice flow as at sensors at right angles to it. It is light counted at the sensors, at that distance. It does not mean light goes twice as far.

"About twice as much light is observed along the direction of the ice flow compared to the orthogonal ice tilt direction when measured at distances of ∼125 m"

Source: Experimental signature, and Fig. 4.

The direction matches the glacier

The direction of the effect is the same everywhere in the detector. It matches today's ice flow, measured on the surface with GPS stakes. The glacier there moves about 10 m a year.

"The anisotropy axis is seen to have constant direction throughout the entire detector"

"is in excellent agreement with present-day flow direction as measured using a GPS stake field"

Source: The anisotropy axis, and Layer undulation.

Dust in the ice could not explain it

Earlier models blamed dust and other impurities in the ice. One needed absorption to change by a factor of 11 with direction, which the authors call unphysical. None of them reproduced the data.

"A factor of 11 modulation of the absorption coefficient was required to fit the data, which seems unphysical."

"directional modifications to Mie scattering or absorption cannot reproduce observations nor are such modifications well motivated on first principles."

Source: Early empirical modeling.

Ice splits light in two

Ice crystals are birefringent: they split incoming light into two rays. IceCube's light, at about 400 nm, is significantly smaller than the ice grains, which are expected to be millimetre-sized.

"Ice crystals also are a birefringent material such that any incoming electromagnetic radiation is separated into an ordinary and extraordinary ray"

"The wavelength of ∼400 nm employed in the IceCube calibration studies is significantly smaller than the average grain size, which is expected to be on the millimeter scale."

Source: Introduction.

The team thinks light is nudged toward the flow line

In the authors' model, light refracts and reflects at grain boundary after grain boundary, and on average is slowly deflected toward the flow axis. An axis is a line. The pull is toward the line the glacier flows along, from either side, not toward where the glacier is heading. To the authors' knowledge, it is a newly discovered optical effect. It is the model's explanation of the measurement. The paper did not watch light paths curve.

"With birefringence rays get slowly deflected towards the flow axis."

"To our knowledge this is a newly discovered optical effect not described in the literature before."

Source: Fig. 10 caption, and Comparison to fabric-induced anisotropies in radar measurements.

"A possible explanation"

The paper offers birefringence as "a possible explanation". Its model, built from first principles, matches the main features of the data in kind. It fits them closely only after small empirical corrections.

"We examine birefringent light propagation through the polycrystalline ice microstructure as a possible explanation for this effect."

"provide a qualitatively good match to the main features of the data"

"By adding small empirical corrections to this first-principles model, a quantitatively accurate description of the optical properties of the IceCube glacial ice is obtained."

Source: Abstract.

Not settled

Birefringence alone could not fully describe the data. To fit it, the authors added an absorption that changes with direction, by a factor of 2.45, with no known physical cause. They say it is unclear whether that absorption is real or an artifact of incomplete modelling.

"The first-principles birefringence explanation was not able to fully describe the experimental data."

"a directional modulation of the absorption coefficients by a factor of 2.45"

"it remains unclear whether the anisotropic Mie absorption is real or if it is an artifact from incomplete modeling of birefringence effects"

Source: Fitting to flasher data, Resulting ice model, and Conclusions.

What else the paper found

In the model, the crystal blurring on its own is strongest along the flow. That would dim light in that direction, the opposite of what is seen. It is the slow deflection that moves light away from the cross-flow direction and piles it up along the flow.

"The diffusion being largest along the flow axis results in a reduction of intensity in this direction, which is contrary to observations."

"The deflection, however, slowly diverts intensity from the tilt direction and overpopulates the flow"

Source: Simulating diffusion patterns, and Fig. 9.

The paper calls the effect unexpected, and it is not new. It was first discussed in 2013.

"An unexpected light propagation effect observed by the experiment is an anisotropic attenuation, which is aligned with the local flow direction of the ice."

"This ice optical anisotropy was first discussed in 2013"

Source: Abstract, and Experimental signature.

What the paper doesn't say

It does not show light bending. What was measured is how much light reached the sensors. The bending is the model's explanation. The paper names a test for it: with aimable lights in the IceCube Upgrade, the brightest direction should sit off the straight line to the receiver. This page does not say whether that test has been run.

"The birefringence-induced deflection yields a unique signature, where the emission direction of maximum received intensity is offset from the geometric direction to the receiver."

Source: IceCube Upgrade.

The prize was not for this result. Francis Halzen is one of the paper's many authors, writing as the IceCube Collaboration, and the collaboration credits the substantial contributions to this paper to two other members.

"The IceCube collaboration acknowledges the substantial contributions to this paper from MR and DCh."

Source: Author contributions.

The factor of 11 and the factor of 2.45 come from different models. The 11 is the old dust-based fit. The 2.45 is the extra term in the new one.

The source

IceCube Collaboration (R. Abbasi et al.) (2024). In situ estimation of ice crystal properties at the South Pole using LED calibration data from the IceCube Neutrino Observatory. The Cryosphere 18, 75-102. DOI: 10.5194/tc-18-75-2024. Full text: tc.copernicus.org/articles/18/75/2024

Prize: Nobel Prize in Physics 2026 press release, nobelprize.org, 6 October 2026.

The article is licensed CC BY 4.0. This page quotes and links to it.

The Short