Nobel Prize in Physics 2026: how IceCube catches neutrinos

The 2026 Nobel Prize in Physics went to Francis Halzen of the University of Wisconsin-Madison, "for decisive contributions to the IceCube Neutrino Observatory and the discovery of high-energy neutrinos of astrophysical origin." The Royal Swedish Academy of Sciences announced it on 6 October 2026. Halzen is the only laureate this year, so he receives the full 12 million Swedish kronor.

IceCube is a detector made from a cubic kilometre of ice at the South Pole. It catches neutrinos, particles that pass through almost anything, including the whole Earth. This post explains how the detector works, what it found and why physicists needed something this unusual.

Who is Francis Halzen?

Francis Halzen is a theoretical physicist who has led IceCube as its principal investigator since 2001. He earned his PhD at KU Leuven in Belgium in 1969, worked at CERN until 1971 and then moved to the University of Wisconsin-Madison. In 1988 he and John Learned of the University of Hawaii sketched a neutrino telescope in polar ice, in a symposium paper.

Halzen did not invent the basic idea. In the early 1960s, physicists including Moisey Markov proposed using large natural bodies of water or ice to catch neutrinos. Halzen's contribution was to push the ice version for decades, until it worked. "This was kind of an adventure where success was not guaranteed," he told Physics Today.

What is a neutrino, and why is it hard to catch?

A neutrino is a tiny particle with no electric charge and almost no mass. It feels only the weak nuclear force and gravity, so it very rarely hits anything. Trillions pass through your body every second without a trace.

That makes neutrinos hard to detect but very useful for astronomy. The other messengers from space each have a problem:

• Light and gamma rays: dust and gas can absorb them, so the centres of the most violent objects stay hidden.
• Cosmic rays: these charged particles are bent by magnetic fields in space, so by the time they reach Earth they no longer point back to their source.
• Neutrinos: they travel in straight lines and are not absorbed, so they keep their original direction and energy across the universe.

The catch is numbers. Because neutrinos so rarely interact, you need an enormous target to see even a few from deep space. That is why IceCube is a cubic kilometre in size.

How does the IceCube Neutrino Observatory work?

IceCube works by watching a huge block of very clear Antarctic ice for faint flashes of blue light. Once in a while, a neutrino hits an atomic nucleus in the ice. The collision creates charged particles such as muons. These move faster than light travels in ice, which is about three quarters of its speed in a vacuum. Like a boat making a bow wave, they leave a cone of blue light called Cherenkov radiation.

Here is the detector, from IceCube's own figures:

• Location: near the Amundsen-Scott South Pole Station in Antarctica.
• Size: one cubic kilometre of ice, about a billion tonnes.
• Sensors: 5,160 digital optical modules, each with a 10-inch light sensor, on 86 strings of 60.
• Depth: between 1,450 and 2,450 metres below the surface.
• Spacing: strings 125 metres apart, sensors 17 metres apart on each string.
• Construction: seven Antarctic summers. Holes were melted into the ice and the last string went in on 18 December 2010.
• Cost: $279 million, about $242 million of it from the US National Science Foundation.

Each sensor time-stamps the light it sees and sends the data to a lab on the surface. From which sensors lit up, and when, computers rebuild the direction and energy of the particle.

The light pattern takes two main shapes:

• Tracks: a muon neutrino makes a muon that can cross the whole detector in a line. Tracks point back to the sky well, so they are best for finding sources.
• Cascades: other neutrino interactions make a roughly round burst of light. Cascades measure energy well but point back poorly.

A denser core called DeepCore, eight strings in the centre, lowers the energy threshold to about 10 GeV so IceCube can also study how neutrinos change type. On the surface, 81 IceTop stations detect showers from cosmic rays and help reject background.

How does IceCube tell cosmic neutrinos from noise?

Most of what IceCube sees is not from deep space. Cosmic rays hitting the atmosphere make about a million times more muons than neutrinos do. IceCube uses the Earth itself as a filter. Muons from the atmosphere come down from above. A muon moving upward through the detector must come from a neutrino that crossed the planet, because nothing else can get through.

Even the upward events are mostly ordinary neutrinos, made by cosmic rays on the far side of the Earth. The full detector sees about 75 upward-going neutrinos a day. Neutrinos from space stand out mainly by their much higher energies, and a source stands out when several events come from the same spot. Finding them is a statistics problem: a few special events against a large, well-understood background.

A second trick is to keep only events that start inside the detector. A neutrino can appear out of nowhere in the middle of the ice, but a muon from the air cannot. IceCube's 2023 study of the Milky Way used such starting cascade events, which cut the background from atmospheric muons and neutrinos.

What did IceCube discover?

IceCube first proved that the universe sends us high-energy neutrinos, and then began to find where they come from. The main results:

• 2013: IceCube reported 28 high-energy neutrinos that likely came from beyond the solar system, including two at peta-electronvolt energies nicknamed Bert and Ernie. Physics World named it the Breakthrough of the Year.
• 2017: a single neutrino detected in September 2017 was traced to a blazar called TXS 0506+056, in the constellation Orion. A blazar is a galaxy whose central black hole fires a jet toward Earth. IceCube puts it about 4 billion light-years away. The results were published in 2018.
• 2022: about 80 neutrinos were traced to NGC 1068, also called Messier 77, an active galaxy 47 million light-years away. Dust around its black hole blocks most gamma rays, but the neutrinos get out.
• 2023: IceCube made the first picture of our own Milky Way in neutrinos. It used 60,000 neutrinos from 10 years of data, and machine learning methods built at TU Dortmund University.

Halzen's work also turned neutrino astronomy into a field with many detectors. KM3NeT in the Mediterranean Sea, Baikal-GVD in Lake Baikal, P-ONE in the Pacific and TRIDENT in the South China Sea all follow the same idea.

What comes next for neutrino astronomy?

IceCube finished its first major upgrade in early 2026, adding closely spaced strings to study lower-energy neutrinos in more detail. A much larger IceCube-Gen2 is in the US federal approval process. Asked about his plans after the prize, Halzen said: "I'm working on a proposal, and I hope that this prize will help getting it approved." He did not say which proposal.

The biggest open question is where most cosmic neutrinos come from. IceCube has tied only a few sources to specific objects so far. Blazars and dust-hidden black holes explain some events, but most of the high-energy flux still has no known source.

How students can work with IceCube data

You do not need to travel to the South Pole. IceCube shares data that anyone can use:

• Public data: IceCube's data releases page lists public datasets. They include a 14-year muon track release covering 2008 to 2022, alert catalogues and the Milky Way data.
• Machine learning: in 2023 IceCube ran a public Kaggle competition, "IceCube - Neutrinos in Deep Ice", to rebuild neutrino directions from simulated events. 812 teams took part, and the top solutions performed close to the best methods in the field. The top three solutions are published.
• A first project: download a public track dataset, plot the arrival directions on a sky map and check whether events cluster near NGC 1068.

Projects like these teach real skills: statistics, sky coordinates and machine learning on messy physics data. If you want a mentor, our guide to starting research as a student explains how to approach one. Our course on statistics for researchers covers the significance tests that results like these depend on.