
Physicists at CERN have nailed down one of the rarest things ever measured in a lab: a charged kaon falling apart into a pion and two neutrinos. It happens about 13 times in every 100 billion kaon decays, or roughly once in 7.7 billion, and the NA62 experiment near Geneva needed seven years of data to see it with confidence.
The result, reported in the collaboration’s paper in the Journal of High Energy Physics, puts the decay’s probability, or branching ratio, at 13.0 × 10^−11 (0.00000000013), with an uncertainty of roughly plus 3.3 or minus 3.0 in those units. The team says it is the smallest branching ratio ever measured with a statistical significance above five sigma, the bar particle physicists use before calling something an observation rather than a hint.
A needle in a 270-meter haystack
Kaons are short-lived particles made of a strange quark paired with an up or down quark. Wikipedia’s history of the kaon traces their discovery to 1947, when George Rochester and Clifford Butler at the University of Manchester spotted strange tracks in cosmic-ray photographs. Kaons went on to reveal CP violation in 1964, a clue to why the universe holds more matter than antimatter.
To catch the ultra-rare version, NA62 fires 400 GeV protons from CERN’s Super Proton Synchrotron at a target, producing a mixed beam in which only about 6% of particles are kaons. According to the Wikipedia entry on the NA62 experiment, the detector chain stretches about 270 meters and sees about 4 million kaon decays per second inside its decay zone. The collaboration lists 308 participants from 33 institutions in 16 countries.
The hard part is that both neutrinos escape unseen. The detector has to prove that a lone pion came from this exact decay and not from one of the far more common ones, using silicon trackers, a liquid krypton calorimeter and rings of photon vetoes to rule everything else out.
Badiyaa’s math: what 51 events really means
Combining runs from 2016 to 2022, the team counted 51 candidate events against an expected background of about 18, per the arXiv version of the NA62 paper. That leaves an excess of roughly 33 events that the background alone can’t explain, and that excess is what carries the measurement.
Put another way, at about 4 million decays a second, a one-in-7.7-billion process would in principle occur around once every 32 minutes inside the decay zone. Most of those are lost to tight selection cuts, which is why seven years of running produced only a few dozen clean candidates.
Why physicists care
This decay is one of the cleanest tests of the Standard Model, because theorists can predict its rate unusually precisely. NA62’s main design goal, per its Wikipedia entry, is to measure it to about 10% precision, which would in turn pin down a key quark-mixing number known as Vtd. A rate that drifts away from the prediction could point to new particles beyond the Standard Model. Data-taking continues until CERN’s next long shutdown, which should sharpen the number further.
CERN’s accelerator complex is getting bigger magnets too: read how the LHC upgrade is getting 11.3-tesla niobium-tin magnets, and our look back at Einstein’s field equations of 1915.
Sources
- Observation of the K+ → π+νν̄ decay and measurement of its branching ratio, Journal of High Energy Physics (NA62 Collaboration), 2025
- NA62 Collaboration preprint, arXiv:2412.12015, arXiv, 2024
- NA62 experiment, Wikipedia, 2026
- Kaon, Wikipedia, 2026
Image: Sophia Elizabeth Bennett, CC BY 4.0 via Wikimedia Commons
This article was produced with AI assistance and reviewed by our Editorial Desk. How we work.


