Neutrinos and Cherenkov light: the road to the 2026 Physics Nobel

Why physicists invented an invisible particle, how they finally caught it, and why a cubic kilometre of Antarctic ice became a telescope for it. With animations you can play and numbers you can check yourself.

In October 2026 the Nobel Prize in Physics went to Francis Halzen for IceCube, a telescope built from a cubic kilometre of ice at the South Pole, and for its discovery of neutrinos coming from far outside our solar system. This first part of the series covers the ninety years of ideas that made such a strange telescope sensible: a particle invented to save the law of energy conservation, the 26 years it took to catch one, and the faint blue light that lets ice “see” it. You can test each step yourself with the relativistic kinematics calculator, the Counting statistics helper and the EM spectrum explorer.

In short

  • A neutrino is a tiny, neutral particle that hardly ever hits anything: about 65 billion from the Sun pass through each square centimetre of you every second.
  • It was invented on paper in 1930 to explain missing energy, and caught in 1956 next to a nuclear reactor.
  • When a neutrino does hit something in water or ice, it can make a fast charged particle that glows with Cherenkov light. In 1988 Halzen and John Learned proposed using the clear, deep ice under the South Pole to catch that glow.

A crisis in 1930: energy goes missing

Some atomic nuclei change by spitting out an electron. This is beta decay. In the 1920s careful measurements showed something shocking: the electrons came out with any energy from zero up to a maximum, not with one fixed energy.

Why is that shocking? If a nucleus at rest breaks into just two pieces, the laws of energy and momentum fix exactly how the energy is shared, so every electron should leave with the same energy. Some physicists, Niels Bohr among them, were ready to give up exact energy conservation.

Wolfgang Pauli did not want to. In a famous letter of December 1930 to a physics meeting he proposed a “desperate remedy”: a third particle, neutral and very light, that leaves the nucleus too and takes the missing share of the energy. Nobody could see it, so the electron seemed to lose energy at random. In 1934 Enrico Fermi built a full theory of beta decay around it and gave it its name, the neutrino (“little neutral one”).

Try the decay below. Each press records one electron’s energy. In the real world the bars pile up into a broad hump; switch to “if no neutrino existed” and every electron lands in the last bar.

Interactive animation: beta decay electron energies (needs JavaScript).

Why the hump has that shape

The shape comes from counting the ways the energy can be shared. For an “allowed” decay the number of electrons with kinetic energy T goes as p·E·(Q − T)², where p and E are the electron’s momentum and total energy and Q is the decay energy. The (Q − T)² factor is the room left for the neutrino. Near T = Q that room vanishes, so the spectrum falls smoothly to zero there. If the neutrino has a small mass m, the end point moves down by m and the very last part of the curve changes shape: that is exactly how the KATRIN experiment in Germany has been squeezing the neutrino mass below 0.5 eV. A small Coulomb correction (the Fermi function) bends the low-energy end; the animation leaves it out.

How do you catch a ghost? (1956)

Fermi’s theory also predicted how rarely a neutrino would interact. The numbers were sobering: a typical neutrino from a reactor could pass through light-years of lead without stopping. Pauli is said to have worried that he had invented a particle that could never be detected.

The way out is numbers. A nuclear reactor makes an enormous flood of antineutrinos (about 10²⁰ per second for a large one). In 1956 Clyde Cowan and Frederick Reines put tanks of water with dissolved cadmium chloride next to the Savannah River reactor in South Carolina. Once in a while an antineutrino hit a proton in the water and turned it into a neutron and a positron:

ν̄ₑ + p → n + e⁺

The positron annihilated at once, giving a flash; a few microseconds later the neutron was captured by cadmium, giving a second flash. That delayed pair of flashes was the fingerprint. They counted a few such events per hour more when the reactor was on than when it was off, and sent Pauli a telegram. Reines shared the 1995 Nobel Prize for it.

This reaction needs a minimum neutrino energy, because a neutron and a positron weigh more than a proton. With the relativistic kinematics calculator you can work it out: the threshold is 1.806 MeV. Reactor antineutrinos below that energy pass unseen.

Neutrinos everywhere

After 1956 neutrinos turned up wherever physicists looked:

  • The Sun. Fusion in the Sun’s core turns hydrogen into helium and releases two neutrinos each time. From the Sun’s power alone you can estimate the flux at Earth (we do it below): about 6.5 × 10¹⁰ per square centimetre per second. Ray Davis’s chlorine tank in a South Dakota gold mine caught them from the late 1960s, and found only about a third of the expected number: the “solar neutrino problem”.
  • A supernova. On 23 February 1987 detectors in Japan, the USA and Russia caught about two dozen neutrinos from a star that exploded in the Large Magellanic Cloud, 160,000 light-years away. Davis and Masatoshi Koshiba shared the 2002 Nobel Prize for opening neutrino astronomy with the Sun and this supernova.
  • The atmosphere. Cosmic rays smashing into air make showers of pions and muons, which decay into neutrinos. In 1998 the Super-Kamiokande detector showed these change type (“oscillate”) on the way through the Earth, which means neutrinos have mass. That, and the solution of the solar puzzle at the SNO detector, won the 2015 Nobel Prize.
  • Three kinds. There are electron, muon and tau neutrinos, each paired with its charged partner. The muon neutrino was found in 1962 (Nobel Prize 1988), the tau neutrino in 2000.

By the 1980s one big source was still missing: the most violent objects in the universe. Cosmic rays, mostly protons, reach Earth with energies up to a hundred million times what the Large Hadron Collider can give. Something out there accelerates them, and nobody knew what. When such protons hit gas or light near their source they must make pions, and pions decay into neutrinos. So a neutrino telescope could find the accelerators. The catch: such high-energy neutrinos are rare, so the detector would have to be about a cubic kilometre in size.

Blue light faster than light

How do you watch a cubic kilometre? The answer is a kind of light discovered in 1934 by Pavel Cherenkov in Moscow.

Nothing beats the speed of light in a vacuum. But light slows down inside a material: in ice it travels at c/n, where n ≈ 1.31 is the refractive index, so about 229,000 km/s. A charged particle with a lot of energy can go faster than light in ice. When it does, it leaves behind a cone of faint blue light, much as a boat faster than its own waves leaves a V-shaped wake, or a jet faster than sound leaves a sonic boom. Ilya Frank and Igor Tamm explained it in 1937, and the three shared the 1958 Nobel Prize.

Play with it. Set the particle’s speed and the material: below the threshold, the light rings never overlap; above it, they pile up into a bright cone.

Interactive animation: Cherenkov light cone (needs JavaScript).

The cone’s angle obeys a simple rule:

cos θ = 1 / (n β)

where β is the particle’s speed as a fraction of c. For a muon at nearly the speed of light in ice, θ ≈ 40°. Because the angle is fixed, measuring when the light reaches many sensors tells you the direction of the particle. And since a muon made by a neutrino flies almost in the neutrino’s direction, you learn where in the sky the neutrino came from.

How much light, and what colour

The Frank–Tamm formula gives the number of photons per unit length per unit wavelength as proportional to (1 − 1/(n²β²))/λ². The 1/λ² factor makes Cherenkov light strongest at short wavelengths, which is why it looks blue (and why the sensors are tuned to blue and near-ultraviolet). A relativistic muon makes roughly 250 photons per centimetre in the 300 to 500 nm band in ice. The EM spectrum explorer gives each photon’s energy: about 3 eV at 400 nm.

Water, then ice: the idea Halzen pushed

In 1960 Moisey Markov suggested putting detectors deep in a lake or the sea, where the water above blocks most of the cosmic-ray muons raining down from the sky. The American DUMAND project tried this off Hawaii from the late 1970s; the ocean proved brutal on cables and electronics and the project ended in 1995. The Baikal telescope in Siberia, in the deep clear water of Lake Baikal, ran from the 1990s and is still growing today.

In 1988 Francis Halzen and John Learned proposed a different medium: the Antarctic ice sheet. Deep polar ice is about three kilometres thick at the South Pole, extremely clear, perfectly dark, free of the light glowing from radioactive potassium and from sea creatures in seawater, and it does not move around. Holes can be melted with hot water and sensors frozen in for good. The US science base at the Pole already had the flights, power and people.

The first tests began in the early 1990s as AMANDA (Antarctic Muon And Neutrino Detector Array). They revealed a surprise: at shallow depths the ice was full of tiny air bubbles that scattered the light hopelessly. Deeper, below about 1.4 km, the pressure had squeezed the bubbles into crystals and the ice turned glass-clear. AMANDA moved deeper and was complete by 2000. It saw atmospheric neutrinos and proved the method worked, but at about 1/50 of a cubic kilometre it was far too small to find cosmic sources. The next step had to be fifty times bigger. That is part 2.

Checked with the tools

We checked three of the numbers in this story with the maths behind fibub’s own tools. Everything below is our own recomputation or simulation; nothing is copied from a paper.

1. The reaction Cowan and Reines used needs 1.806 MeV. In the Reaction threshold part of the relativistic kinematics calculator, a massless beam on a proton at rest making a neutron (939.565 MeV) and a positron (0.511 MeV) needs a beam energy of 1.806 MeV. That matches the textbook value.

2. The solar neutrino flux follows from the Sun’s power. Each fusion of four protons into helium releases 26.73 MeV and two neutrinos; about 0.6 MeV of that leaves with the neutrinos, so about 26.1 MeV heats the Sun. With the Sun’s luminosity (3.828 × 10²⁶ W) spread over a sphere one astronomical unit across:

Φ = 2 L ÷ (26.1 MeV × 4π d²) ≈ 6.5 × 10¹⁰ neutrinos per cm² per second

This is the number quoted by solar models, from nothing more than the Sun’s brightness and its distance.

3. How long you must count. Cowan and Reines saw about three extra events per hour. With round numbers of that size (3 per hour of signal, 1 per hour of background, the reactor on and off for equal times), the net rate grows clearer like the square root of the counting time:

Reactor on minus reactor off: how sure after t hours?Reactor on minus reactor off: how sure after t hours?05101520406080100Counting time with the reactor on (and as long off), hoursNet rate ÷ its uncertainty (σ)5σ3σfibub.com
Made with the Counting statistics tool's maths (net rate and its uncertainty from gross and background counts). Rates are our own round numbers of the same size as the 1956 experiment's, not its data.

4. The Cherenkov angle settles near 40° in ice. From the muon’s energy we got its speed with the relativistic maths, then the angle with cos θ = 1/(nβ):

Cherenkov angle of a muonCherenkov angle of a muon0102030402004006008001e+3Muon total energy (MeV)Angle to the path (degrees)ice limit 40.2°Ice, n = 1.31Water, n = 1.33fibub.com
Speeds from the Relativistic kinematics tool's maths (muon mass 105.66 MeV), angle from cos θ = 1/(nβ). Real ice and water refractive indices depend on wavelength; 1.31 and 1.33 are round values.
Quantity Standard value What the tool maths gives
Inverse beta decay threshold 1.806 MeV 1.806 MeV
Solar neutrino flux at Earth about 6.5 × 10¹⁰ cm⁻² s⁻¹ 6.5 × 10¹⁰ cm⁻² s⁻¹
Muon Cherenkov threshold in ice (total energy) about 160 MeV 164 MeV
Largest Cherenkov angle in ice (n = 1.31) about 40° 40.2°

How we made the graphs

The charts come from one script, scripts/blog-figures/nobel-physics-2026.mjs, that imports the same TypeScript the tools run in your browser: threshold(), fourFrom() and describe() from src/lib/sci/phrel.ts (the Relativistic kinematics calculator) and netRate() from src/lib/sci/phcount.ts (the Counting statistics helper). Particle masses are the Particle Data Group’s values, rounded. The counting rates are our own round numbers of the same size as the 1956 experiment’s, not its data. The animations on this page run small simulations written for this series (the beta spectrum shape and the Cherenkov wavefronts) and draw a fibub.com mark in the corner like every plot here.

How we test the maths

Every change to the site runs the checks in tests/sci-tools/relativistic-kinematics.mjs and tests/sci-tools/counting-statistics.mjs. They pin the relativistic maths to textbook results: the antiproton threshold in p + p → p p p p̄ (kinetic energy 6 m_p = 5,629.6 MeV, the number that set the Bevatron’s design in the 1950s), π⁰ photoproduction on a proton (144.69 MeV), and the muon’s momentum in π⁺ → μ⁺ν (29.79 MeV/c). The counting maths is pinned to Currie’s 1968 formulas (L_C = 2.326 √B for a paired blank) and to exact Poisson confidence limits computed with SciPy. If a change broke any of them, the tests would fail and the change would not go live.

Check it yourself

  1. Open the Relativistic kinematics calculator and set the unit to MeV.
  2. In 4. Reaction threshold, type nu (or a mass of 0) as the Beam particle, p as the Target particle (at rest) and n + e+ as the Products. The threshold beam energy should read about 1.806 MeV.
  3. Change the products to n + mu. The threshold jumps to about 113 MeV: low-energy reactor antineutrinos can never make a muon.
  4. Open the Counting statistics helper. In 1. Net counts and detection limit, enter Gross counts 48 in Gross counting time 12 (hours), and Background counts 12 in 12. The net rate is 3 per hour with an uncertainty of about 0.65 per hour: about 4.7σ after 12 hours, as on the chart.
  5. Open the EM spectrum explorer, type 400 nm and read the photon energy (about 3.1 eV): the colour Cherenkov light is richest in.

What you can use it for

  • Students: the threshold and two-body decay sections of the relativistic kinematics calculator work through any reaction in a textbook problem, step by step.
  • Lab work: the Counting statistics helper tells you how long to count a weak source against background, or whether a small excess is real.
  • Teaching: the Cherenkov animation shows why there is a speed threshold and why the angle is fixed, which a static picture rarely gets across.
  • Next in this series: how a cubic kilometre of sensors was frozen into the ice, and how it found the first neutrinos from beyond our galaxy, in part 2.

Sources

  • Nobel Prize in Physics 2026, announced 6 October 2026: Francis Halzen, University of Wisconsin–Madison, 'for decisive contributions to the IceCube Neutrino Observatory and the discovery of high-energy neutrinos of astrophysical origin'. NobelPrize.org (nobelprize.org/prizes/physics/2026/) and the Royal Swedish Academy of Sciences press release (kva.se).
  • Physics World (2026), 'Francis Halzen wins Nobel Prize in Physics': timeline of AMANDA and IceCube.
  • Pauli W (1930), open letter to the Tübingen meeting ('Dear radioactive ladies and gentlemen'), 4 December 1930.
  • Fermi E (1934), Zeitschrift für Physik 88: 161 (theory of beta decay).
  • Cowan CL, Reines F, Harrison FB, Kruse HW, McGuire AD (1956), Science 124: 103 ('Detection of the free neutrino: a confirmation').
  • Cherenkov PA (1934), Doklady Akad. Nauk SSSR 2: 451; Frank IM, Tamm IE (1937), Doklady 14: 109. Nobel Prize in Physics 1958.
  • Markov MA (1960), Proc. 10th Int. Conf. on High-Energy Physics, Rochester (deep underwater or underground neutrino detection).
  • Halzen F, Learned JG (1988), 'High energy neutrino detection in deep polar ice', Proc. 5th Int. Symp. on Very High Energy Cosmic Ray Interactions, Łódź.
  • Particle masses: Particle Data Group, Review of Particle Physics (2024), rounded. Solar luminosity 3.828 × 10²⁶ W (IAU 2015 nominal value).

If you are an author or publisher and want a credit or a passage changed or removed, tell us and we will act quickly.

Tools in this article

Read next

  • Channelrhodopsin: the protein that is both a light sensor and a gate Nobel series ·

    How Georg Nagel, Peter Hegemann and Ernst Bamberg put two algal genes into frog eggs and human kidney cells and found a channel opened by light: channelrhodopsin-1 in 2002 and channelrhodopsin-2 in 2003. The evidence, the physics of the photon and the pore, rebuilt with fibub tools.

  • How IceCube was built and found neutrinos from deep space Nobel series ·

    5,160 light sensors frozen into the South Pole ice, a flood of background a million times larger than the signal, and the 2013 discovery of neutrinos from beyond the solar system. Play with an event display and check the statistics yourself.

  • Where cosmic neutrinos come from: blazars, NGC 1068 and the Milky Way Nobel series ·

    A flaring black hole 4 billion light-years away, a single 6-PeV event that confirmed a 1960 prediction, a dusty nearby galaxy and the glow of our own Milky Way: the evidence behind the 2026 Physics Nobel, and how to check it yourself.

  • Calibrating a gamma-ray detector with natural rock standards Discovery ·

    A 2025 study calibrates a germanium detector with cheap reference powders of potassium, uranium and thorium instead of a costly mixed source. We refit its efficiency curve and check the source's decay with fibub's tools.

  • How to draw a Feynman diagram and export it to LaTeX (TikZ-Feynman) Guide ·

    Describe vertices and lines in plain text, get a clean Feynman diagram as SVG or PNG plus TikZ-Feynman code. The line types, labels, loops and a worked e⁺e⁻ → γγ example.