Francis Halzen turned a cubic kilometre of Antarctic ice into a detector for the universe's most elusive particle, and used it to see parts of the cosmos that no telescope can.
In one line
Space has monster particle accelerators we can't see directly. They shoot out “ghost particles” that fly straight to Earth. Halzen built a giant ice trap at the South Pole that catches a few of them, so we can tell where they came from.
If you only look at one thing on this page, look at this.
1 · THE SOURCE
Far away, objects like giant black holes act as natural particle accelerators, up to millions of times stronger than anything humans have built.
2 · THE MESSENGER
They spit out neutrinos. Other particles get bent or blocked on the way, but neutrinos fly dead straight, so they point back home.
3 · THE TRAP
Neutrinos almost never hit anything, so you need a gigantic detector. IceCube is a cubic kilometre of ice with about 5,000 light sensors frozen inside.
4 · THE CATCH
When one does hit an atom, it makes a flash of blue light. The sensors record when and how bright, and scientists trace the path back to the sky.
Chapter 1
Meet the neutrino, the ghost particle
Everything around you is made of tiny particles. Most of them bump into things all the time, which is why you can't walk through a wall. The neutrino is different. Two things make it a ghost. It weighs almost nothing. And it has no electric charge, the property that makes particles push and pull on each other, a bit like magnets. With no charge, a neutrino has almost no way to “grab” anything, so it hardly notices ordinary matter. It passes through walls, mountains, your body and the entire Earth as if they were empty space.
That makes neutrinos wonderful messengers, because almost nothing stops them, and terrible to catch, for the same reason.
Neutrinos that have passed through your body since you opened this page:0You felt none of them.
~100 trillion
neutrinos from the Sun pass through your body every second. Day and night: at night they come up through the Earth.
light-years of lead
You would need a wall of lead about a light-year thick to have a good chance of stopping one typical neutrino from the Sun.
3 types
Physicists call them “flavours”: electron, muon and tau neutrinos. When one hits something, each type leaves a different-shaped splash of light. You'll see two of those shapes in Chapter 5.
Neutrinos were predicted by Wolfgang Pauli in 1930, but nobody detected one until 1956. Neutrino research has already won Nobel Prizes in 1988, 1995, 2002 and 2015. This year's prize is different: it's about neutrinos from outside our galaxy, with energies far beyond the Sun's.
Key idea
Neutrinos are tiny “ghost” particles that pass through almost anything, even a whole planet, without touching it.
Quick check
A neutrino heads straight for Earth. What happens?
Chapter 2
Why chase neutrinos? A mystery from 1912
To see why scientists wanted neutrinos so badly, we need a little backstory. It begins with something called cosmic rays. We'll go one question at a time.
Question 1
What are cosmic rays?
Despite the name, they are not rays of light. They are tiny bits of atoms flying through space at almost the speed of light. Most of them are protons, the tiny particles packed into the centre of every atom.
They hit the top of Earth's air all the time, day and night. When they do, they shatter into showers of smaller particles. The air soaks up almost all of it, which is why you never notice.
The name came in the 1920s, from physicist Robert Millikan, who thought they were a kind of light, like X-rays.
Question 2
How were they discovered?
Around 1900, scientists noticed that the air was always slightly “charged” with radiation, even in a closed room. Everyone assumed it came from radioactive rocks in the ground.
If that were true, the radiation should get weaker the higher you go, away from the rocks. In 1912 the Austrian physicist Victor Hess tested this by taking a radiation meter up in a hydrogen balloon, more than 5 km high.
The radiation got stronger, not weaker. So it wasn't coming from the ground. It was raining down from space. Hess won the 1936 Nobel Prize for this.
Question 3
What does “most energetic” mean?
“Energy” here just means how hard a particle hits. That depends on how fast it is moving.
Think of a hailstorm: millions of tiny pellets and, very rarely, one giant hailstone. Cosmic rays are like that. Most are gentle, but a tiny few hit millions of times harder than anything our biggest human-made machines can produce. The record holder, in 1991, hit as hard as a thrown baseball, from one particle smaller than an atom.
Something out there is speeding these up. Scientists call those somethings cosmic accelerators, or “factories”. Giant black holes are a leading suspect.
Question 4
Why find the factories, and why is it so hard?
Finding these factories would answer a 100-year-old mystery: what are the most violent objects in the universe, and how do they fling particles so hard? To track one down, you need a messenger that points straight back to it, like following footprints to find who made them. So why not just follow the cosmic rays?
Because they get pushed off course. Protons carry electric charge, and space is full of weak magnetic fields. A magnet pushes on anything with charge, so over the huge distances between stars their paths get bent and scrambled. By the time one reaches Earth, the direction it arrives from tells you nothing about where it started. That's why the mystery stayed unsolved for over a century.
Below, a factory sends three kinds of particle towards Earth. Watch what happens to each.
cosmic ray (proton)lightneutrino
Cosmic ray: bent by magnetic fields, so it arrives from the wrong direction. Light: can be swallowed by clouds of gas and dust. Neutrino: no charge and almost never hits anything, so it flies straight to us.
Question 5
But why would a factory of protons also make neutrinos?
Because neutrinos are a by-product, like exhaust from an engine. Step through it:
So the two are chained together: no factory, no fast protons; no fast protons, no crashes; no crashes, no high-energy neutrinos. If a high-energy neutrino arrives from a spot in the sky, there must be a factory there. Where there's exhaust, there's an engine.
Why protons? What about neutrons and electrons?
Protons
They have electric charge, so the factory's magnetic fields can grab and push them. And they're everywhere: a hydrogen atom is just one proton plus one electron, and hydrogen is the most common stuff in the universe. Heavier atom centres (helium, iron) get sped up too, but there are far fewer of them.
Neutrons
No electric charge, so magnetic fields can't push them. There's nothing to grab. On their own they also fall apart in about 15 minutes. So they don't get sped up directly.
Electrons
They do get sped up! But they're about 2,000 times lighter than protons, so they lose their energy fast by giving off light. That glow is part of what telescopes see. And when an electron crashes into something, it almost never makes neutrinos. Only the heavy proton-type crashes do.
Do we know why the crash makes neutrinos?
Yes, this part is well understood and tested on Earth. Labs like Fermilab and CERN make neutrino beams on purpose, exactly this way: smash fast protons into a target. The chain goes:
Proton crashesinto another proton or atom centre
Pions are bornlike sparks from two rocks, except these are brand new: made out of the crash's energy, not chipped off the protons
Pion falls apartin ~26 billionths of a second → a muon + a neutrino
Muon falls apartan instant later → an electron + 2 more neutrinos
Wait, energy can turn into matter?
Yes. This is what Einstein's E = mc² means: matter and energy are two forms of the same thing, like cash and money in a bank account. You can exchange one for the other. The exchange rate is steep: it takes a huge amount of energy to make a tiny bit of matter. That's why only extremely fast crashes can “buy” new particles. The fast proton's speed is the money spent. After the crash the pieces move slower, and that lost speed became the new pions.
Think of a pond. Physicists now picture all of space as filled with invisible “ponds” (they call them fields), one for each kind of particle. A particle is simply a ripple on its pond. Drop a stone in: the ripples weren't hiding inside the stone. They are the water itself, set moving by the stone's energy. A high-speed crash is a stone hitting the pond so hard that new ripples, new particles, appear. Small stone, small splash, nothing new. Huge stone, big splash, new particles. And ripples fade, which is why pions fall apart.
And yes, energy really does weigh something: a hot cup of coffee weighs a tiny bit more than the same cup cold, far too little to measure on a kitchen scale. The exchange also runs the other way. The Sun shines by turning about 4 million tonnes of its mass into light every second.
Why must neutrinos come out? When a pion or muon falls apart, nature's bookkeeping rules (what goes in must balance what comes out) can only be satisfied if a neutrino is produced. It's the same reason Pauli predicted neutrinos in 1930. Some pions instead turn into pure light (gamma rays), which is why these factories also glow.
The open questions are elsewhere: which objects in space are the factories, and how they speed protons up so much.
So if a high-energy neutrino shows up, it tells us the factory is speeding up protons, not just electrons. Light alone can't tell those two apart. That's one of the big things IceCube adds.
Question 6
What is “multi-messenger astronomy”?
Picture a thunderstorm at night. You can see the lightning, hear the thunder and feel the wind. Each sense tells you something different. Put together, you understand the storm far better than with your eyes alone.
For most of its history, astronomy relied almost entirely on light: first what our eyes and telescopes could see, later radio waves, X-rays and more, which are all forms of light. In the last century new senses joined: cosmic rays (1912), neutrinos from the Sun (1960s) and gravitational waves (2015). Neutrinos from deep space are the newest. Scientists call anything that carries information to us from space a messenger. Using several of them on the same object is “multi-messenger astronomy”: astronomy with more than one sense.
LightThe oldest sense: eyes for thousands of years, telescopes since 1609, and today every kind of light from radio to gamma rays. Can be blocked by dust.the oldest
Gravitational wavesTiny ripples in space itself, made when black holes smash together. Like hearing the universe.since 2015
NeutrinosCome straight from inside the most violent places. Before IceCube, we had only caught them from the Sun and one exploding star.IceCube opened this sense
The first time it worked on a cosmic accelerator: 22 September 2017
IceCube catches oneA single very high-energy neutrino lights up the ice.
Alarm goes outWithin a minute, an automatic alert tells telescopes worldwide where in the sky it came from.
Telescopes lookThey find a distant galaxy, TXS 0506+056, blazing much brighter than usual, right at that spot.
Strong evidence of a factoryTwo different senses, neutrino and light, point at the same object: the first strong evidence of a cosmic-ray factory.
Key idea
Black holes are thought to be giant particle factories. Their neutrinos fly in straight lines, so catching one points you back to its factory.
Quick check
Why can't cosmic rays tell us where their factory is?
Side trip
The particle zoo: what's actually flying around out there?
Protons, neutrons and electrons are what you are made of. But they're only a small part of the crowd in space. Here's who else is out there, from most familiar to most mysterious.
Take one sugar cube of “empty” space
Pick a spot far from any star and draw a box the size of a sugar cube. Inside you'd find about:
400 particles of light, left over from the Big Bang
340 neutrinos, also left over from the Big Bang
almost never an atom. On average there's about one proton for every 4 cubic metres, a box the size of a wardrobe.
So light and neutrinos outnumber the stuff we're made of by about a billion to one.
Light
Photons
Particles of light. Every kind counts: radio, the light you see, X-rays, gamma rays. Telescopes catch these.
The most common particle in the universe
Ghost
Neutrinos
The star of this page. Almost no weight, no charge, passes through planets. Three types.
Almost as common as light
Matter
Protons
The centre of a hydrogen atom. Carries a positive charge, so magnetic fields can push it around.
~90% of cosmic rays
Matter
Atom centres
Protons and neutrons stuck together into a bundle: helium, carbon, all the way up to iron and beyond.
~10% of cosmic rays
Matter
Electrons
The tiny particles that orbit atoms, about 2,000 times lighter than a proton. Negative charge.
A small share of cosmic rays
Matter
Neutrons
No charge. Safe inside an atom's centre, but alone it falls apart in about 15 minutes.
Rare on their own in space
Antimatter
Antiparticles
Mirror-image twins of normal particles, like the anti-electron. When a twin meets its normal partner, both vanish in a flash of light.
Small amounts, caught by detectors in space
Short-lived
Muons
Heavier cousins of the electron, made when cosmic rays crash into our air. They live about two millionths of a second.
~1 through your hand every second
Short-lived
Pions and friends
The “sparks” from high-speed crashes. Gone in a fraction of a second, mostly turning into muons and neutrinos.
Only right where crashes happen
Unknown
Dark matter
About five times more of it than all normal matter. We know it's there because its gravity holds galaxies together, but nobody has ever caught one of its particles.
Biggest open mystery
Key idea
Space is mostly filled with light and neutrinos. The stuff we're made of is surprisingly rare.
Quick check
In a sugar-cube-sized box of “empty” space, what would you find most of?
Chapter 3
How powerful are these cosmic accelerators?
Physicists measure particle energy in electronvolts (eV). Each step on the ruler below is ten times the one before. Hover or tap a marker.
Tap a marker to compare.
~7 mosquitoes
One proton at the Large Hadron Collider, the world's biggest particle accelerator, carries about the energy of seven flying mosquito, packed into a single particle.
×150
IceCube's “Bert” and “Ernie” neutrinos each carried about 150 times the energy of an LHC proton, or a thousand mosquitoes' worth in one particle.
a thrown ball
The most energetic cosmic ray ever seen (1991) carried about the energy of a baseball pitch, all in one subatomic particle.
Key idea
Nature's particle factories hit millions of times harder than the most powerful machine humans have built.
Quick check
How hard do nature's strongest particles hit, compared with our most powerful machine?
Chapter 4
Building a trap for ghosts
We now know why scientists wanted to catch high-energy neutrinos from space. This chapter is about how, and why the answer turned out to be a block of ice at the bottom of the world.
Quick note: IceCube is not under the sea. It's inside the Antarctic ice sheet, a layer of ice about 2.8 km thick that sits on land at the South Pole.
Problem 1
Why does the detector have to be so big?
Catching a neutrino is a lottery. A neutrino flies through the detector, and for every atom it passes there's a ridiculously tiny chance it hits it. Almost always, it hits nothing and flies on.
Each atom in the detector is like a lottery ticket. More material means more tickets, which means more winners. And the neutrinos we want, the high-energy ones from deep space, are rare to begin with. It's like trying to collect rain during a drought: a cup catches nothing, so you need a bucket the size of a lake.
Try it: pick a detector size and watch how often a cosmic neutrino gets caught.
Problem 2
To catch the flash, the ice must pass five tests
Size gives you more lottery tickets, but it's only the start. When a neutrino does hit an atom, all you get is a tiny, faint flash of blue light, sometimes just a handful of light particles. To catch that flash and work out where it came from, the detector has to pass five tests.
1Huge
Billions of tonnes of material, for the lottery reason above.
2Crystal clear
The flash has to travel to a sensor. In murky water it fades within a metre, so you'd need millions of sensors. In clear ice it carries 100+ m, so a few thousand, spread far apart, can watch the whole block.
3Pitch black
The flash is extremely faint. Any daylight, even a glimmer, would drown it out completely.
4Shielded & quiet
Millions of particles from cosmic rays hitting our air rain down every day. They make flashes too. They have to be blocked, or they'd swamp the real signal.
5Perfectly still
The direction comes from tiny differences in when each sensor sees the flash (billionths of a second). That only works if every sensor stays exactly where we put it.
Why “clear” matters so much
Imagine the flash as a candle. In murky water, its light fades within a metre or two, so you'd need a sensor every metre in every direction: hundreds of millions of sensors for a cubic kilometre. Unaffordable. In clear deep ice, the light carries over 100 metres, so sensors can be spread out: 17 m apart on each cable, cables 125 m apart. That's why only 5,160 are needed.
Who passes all five tests?
Here's how the three options compare.
Human-built tank
Deep ocean
Antarctic ice
Huge
✕ The biggest (Super-Kamiokande, Japan) holds 50,000 tonnes, about 20,000 times too small
✓
✓ A whole continent of it
Clear
✓ Ultra-pure water
~ Good, but less clear
✓ Among the clearest natural solids on Earth
Dark
✓
✓
✓ 1.5 km of ice above
Quiet
✓
✕ Glowing sea creatures and natural radioactivity in seawater make false flashes
✓ Nothing lives or glows down there
Still
✓
✕ Currents sway the cables
✓ Frozen solid, no earthquakes
Halzen's insight (1988)
Building a clear, dark, quiet tank a kilometre wide would cost far more than anyone could pay. But nature had already built one, for free, over hundreds of thousands of years of snowfall: the Antarctic ice sheet. All you have to do is put sensors inside it.
Problem 3
Why so deep?
The ice near the surface looks solid, but it's full of tiny air bubbles, trapped when snow was squashed into ice. Bubbles scatter light in every direction, like fog scattering car headlights. You can't tell where a flash came from.
Deeper down, the crushing weight of the ice above squeezes the bubbles away. Below about 1.4 km the ice becomes astonishingly clear. The first prototype in the 1990s, called AMANDA, found this out the hard way: its first sensors were placed too shallow, in the bubbly ice, and the team had to go deeper. In the drawing below, see where the bubbles stop and the sensors start.
Problem 4
How do you get 5,000 sensors 2.5 km inside solid ice?
This was the engineering puzzle. Ordinary drills grind through ice slowly: drilling one hole this deep for ice-core research can take years. IceCube needed 86 of them. Step through the trick:
Why this was genius
It turned an impossible construction project (build a tank a kilometre wide) into a plumbing project: melt holes, drop in cables, let nature refreeze them. The water freezes into the same clear ice, so each sensor ends up perfectly embedded, with nothing between it and the light. The whole detector cost about $279 million, far less than any artificial tank of that size could. And because nothing can ever be repaired, the sensors were built to last: around 98% are still working today, more than 15 years later.
1 km³
of watched ice, about a billion tonnes. Big enough to hold the Great Pyramid hundreds of times.
5,160
sensors: basketball-sized glass balls, each able to sense a single particle of light.
7 summers
to build it (2004–2010). Work is only possible in the Antarctic summer, November to February.
Key idea
To catch a neutrino's faint flash you need a detector that's huge, clear, dark, shielded and still. Antarctica's ice is all five, for free.
Quick check
Why does the ice need to be crystal clear?
Chapter 5 · interactive
Catch a neutrino yourself
Below is IceCube: 86 strings, 5,160 sensors. Press Fire a neutrino. Here is what happens, step by step:
Very rarely, a neutrino bumps into the centre of an atom in the ice.
The bump knocks out new particles, which race off almost as fast as light in empty space.
Light slows down inside ice to about three-quarters of its normal speed. So these particles are actually outrunning light inside the ice. (Nothing beats light in empty space, but inside a material it can happen.)
When something outruns light, it leaves a cone of faint blue glow behind it, like a speedboat that outruns its own waves leaves a V-shaped wake. Physicists call this glow Cherenkov light.
The sensors catch that glow.
In the model below, sensors are coloured by when they saw light (rose = first, teal = last) and sized by how much. Drag to rotate.
earlylate
Drag to rotate the detector. Then fire a neutrino.
Try both buttons. Hits come in two main shapes:
Lines (“tracks”). One type of neutrino makes a muon, a heavier cousin of the electron, that keeps flying straight through a kilometre or more of ice, glowing as it goes. A long straight line is easy to aim, like an arrow, so the direction can be known to about one degree. That's how you name a spot on the sky.
Balls (“cascades”). The other types dump all their energy in one place, making a round ball of light. Great for measuring how much energy arrived, but a ball doesn't point anywhere, so the direction is fuzzy.
One problem: remember those cosmic rays smashing into the top of our air (Chapter 2)? Those crashes make neutrinos too, right here above Earth. So most of what IceCube catches is “local”, not from deep space. Scientists use two tricks to find the cosmic ones. First, they look downward: a particle coming up from below has passed through the whole planet, and only neutrinos can do that, so the Earth acts as a giant filter. Second, they look for energies far higher than our air's crashes can make.
Key idea
A neutrino hit makes a blue flash. The order in which sensors light up shows which way it came from.
Quick check
How does IceCube tell which way a neutrino came from?
Chapter 6
From a crazy idea to a Nobel Prize
Key idea
From idea (1988) to finished detector (2010) took 22 years, and the discoveries that followed earned the 2026 Nobel Prize.
Quick check
What happened after IceCube's alert on 22 September 2017?
Chapter 7
The laureate
The official reason for the prize:
“for decisive contributions to the IceCube Neutrino Observatory and the discovery of high-energy neutrinos of astrophysical origin”
Name
Francis Halzen
Born
1944, Tienen, Belgium
PhD
1969, KU Leuven
Now
Professor, University of Wisconsin–Madison
Role
Scientific leader of IceCube
Prize
12 million Swedish kronor
Halzen began as a theoretical physicist, someone who works with equations rather than equipment. He worked out what cosmic neutrinos should look like, then decided someone had to actually build the instrument to catch them. He pitched the South Pole idea in 1988, led the early prototype AMANDA in the 1990s, and has led IceCube as its scientific leader since 2001. Today the collaboration has about 450 scientists from 58 institutions in 14 countries. The detector cost about $279 million, mostly paid for by the US National Science Foundation.
“His tenacity and scientific vision has paved the way for a new kind of astronomy.” Mark Pearce, Chair of the Nobel Committee for Physics
The prize goes to one person, but the instrument was built by a big team. The citation honours his “decisive contributions” in imagining it, championing it for decades and leading the team that made it work.
Key idea
One stubborn idea, plus hundreds of people to build it, created a brand-new kind of telescope.
Quick check
Who built IceCube?
Recap
Remember these 5 things
If you remember nothing else from this lesson, remember this.
Ghost particles. Neutrinos pass through almost anything, even whole planets.
Cosmic factories. Giant black holes are thought to fling particles millions of times harder than our best machines, and the factories give off neutrinos.
Straight arrows. Light gets blocked and cosmic rays get bent, but neutrinos fly straight, so they point back home.
A trap made of ice. Catching one needs a giant, clear, dark detector. Halzen's team turned a cubic kilometre of Antarctic ice into one: IceCube.
A new way to see. IceCube found neutrinos from deep space and traced some to their sources, which won the 2026 Nobel Prize.