How do you make a beam of particles you cannot steer?

Accelerate protons, then make the parent particles

The machine never accelerates neutrinos. It gives charged protons energy and guides them onto a target. The collisions produce secondary particles, including charged pions and kaons, and these unstable particles are the parents of much of the neutrino beam.

The target helps shape the beam. Its material and geometry affect which particles come out and whether they interact again before leaving. It must also survive repeated heating, radiation damage and mechanical stress. If the target’s condition changes, the parent particles change too, even though nothing has changed in how neutrinos travel.

Beam predictions therefore rely on measurements of hadron production and on transport calculations, each with uncertainties. Counting the incoming protons gives the exposure, called protons on target. That count is not the number of neutrinos reaching a detector, because production, focusing, decay and geometry all lie in between.

Sources: [1] · [2] · [3]

Magnetic horns act before the neutrinos exist

A magnetic horn is a lens for charged parent particles. A pulsed current runs through shaped conductors and creates a magnetic field that bends the charged secondaries. The shape is chosen to send parents in a useful momentum range toward the decay region, so which particles the horn accepts depends on their momentum.

Reversing the current swaps which charge is focused best. Positive pions decay mostly into muon neutrinos, and negative pions mostly into muon antineutrinos. The beam can therefore run in a neutrino-rich or an antineutrino-rich mode, but neither mode gives a completely pure sample.

A neutrino has no electric charge, so once it exists the horn cannot bend its path. The beam direction comes from focusing the parents, aligning the beamline and the kinematics of the decay. A drawing of a neutrino curving through the horn puts the steering at the wrong stage.

Sources: [1] · [4] · [5]

In the decay region, parents decay into a spread of neutrinos

The focused parents fly into a decay region. Its length sets how much chance parents with different lifetimes and boosts have to decay. Absorbers and shielding at the far end stop the remaining hadrons and most charged decay products; the neutrinos pass through the material and continue toward the experiment.

Even pions with identical momentum give neutrinos of different energies in different directions. The two-body relation below links the neutrino energy to its angle from that pion’s direction. A real beam adds up many momenta, angles, production points and decay points.

Placing a detector at an angle to the nominal beam axis reshapes the energy spectrum it sees; this is the idea behind off-axis experiments. The angle does not pick out one energy exactly. The beam’s width and energy range are physical, and the predicted detector spectrum must include them.

Sources: [1] · [3] · [5]

Monitor the beam, then measure its interactions nearby

Monitors on the proton beam check its intensity, position and alignment. Hadron and muon monitors farther along report on conditions downstream. Their readings flag changes that could shift the neutrino flux, but a muon monitor does not count every neutrino emitted.

A near detector measures neutrino interactions before the large flavor change that builds up over the long baseline. Its events pin down flux, cross section and detector response in combination. That beats trusting the production simulation alone, but without other information the events cannot separate those ingredients uniquely.

The near and far detectors also see different spreads of angle and decay position. Because the source is extended, the far spectrum is not simply the near spectrum scaled down by the inverse square of distance. Turning the near measurement into a far prediction needs the beam model and its correlated uncertainties.

Sources: [1] · [3] · [6] · [7]

Known timing and geometry sharpen the test

Short accelerator spills give each event a time tag. A far event that falls in the expected arrival window is easier to tie to the beam than an atmospheric interaction with no tag. Timing cuts background, but on its own it says nothing about how the neutrino’s flavor evolved.

Known positions of source and detector fix the baseline. Reversing the horn polarity, changing the range of parents accepted or viewing off axis changes the source in ways that can be monitored. None of these choices sets the starting energy of each detected neutrino exactly, so event reconstruction and detector response remain essential.

A manufactured beam is powerful because key source settings can be chosen deliberately and checked independently. K2K and MINOS used this control to compare near and far measurements. They tested whether the explanation drawn from atmospheric neutrinos also worked for a human-made source.

Sources: [6] · [7] · [1] · [3]

Try it in the Lab

Primary sources & revision

1. Adamson et al. · The NuMI Neutrino Beam (2015/2016)
2. Fermilab · Designing resilient targets for high-energy particle accelerators (2024)
3. T2K Collaboration · The T2K Experiment (2011)
4. Fermilab · Funneling fundamental particles (2016)
5. Fermilab · Accelerator neutrinos
6. K2K Collaboration · Measurement of Neutrino Oscillation by the K2K Experiment (2006; v3)
7. MINOS Collaboration · Muon-neutrino disappearance with MINOS and NuMI (2006)

First published 2026-10-10; last revised 2026-10-13. Original explanatory text and diagrams by Sandbox Physics. Illustrations are schematic; no experimental event records are reproduced here. This is an independent educational publication, not an official Nobel or experiment collaboration publication.