Neutrino oscillation
Neutrinos change their identity while they travel. This single fact proved that neutrinos have mass, and that the Standard Model of particle physics is incomplete.
Try it yourself: oscillation calculator →
Beginner
The idea, simply put
There are three kinds of neutrino: electron, muon and tau neutrinos. The Sun produces only electron neutrinos. On the way to Earth, however, some of them turn into the other two kinds, back and forth. Which kind you measure depends on how far the neutrino has travelled.
For decades, detectors counted only about a third of the solar neutrinos the theory predicted. Physicists called it the solar neutrino problem. Oscillation solved it: the missing neutrinos had never disappeared, they had simply changed kind.
How we know
In 1998 the Super-Kamiokande detector in Japan saw fewer muon neutrinos arriving from below, after crossing the Earth, than from above.1 In 2001 and 2002 the SNO experiment in Canada counted all three kinds of solar neutrinos together and found exactly the total the Sun produces.2
A picture that helps: three clocks
Imagine that every neutrino carries three tiny clocks that tick at very slightly different rates. Which kind of neutrino a detector sees depends on how far the clocks have drifted apart. At the source they are in step, and the neutrino is, say, an electron neutrino. After a long enough flight they are out of step, and the same neutrino is now more likely to show up as a muon or tau neutrino. Fly further and they come back into step again. This back-and-forth is why the effect is called an oscillation.
Two things decide how quickly the clocks drift apart: the distance travelled and the neutrino's energy. Low-energy neutrinos change kind over short distances, high-energy ones need much longer flights. That is why experiments place their detectors at carefully chosen distances from a reactor or an accelerator, from about one kilometre to more than a thousand.
Why it matters
The clocks can only tick at different rates if neutrinos have mass, and if the three masses are not all the same. The Standard Model of particle physics had assumed that neutrinos are massless. Oscillation therefore showed that the Standard Model is incomplete. For this discovery Takaaki Kajita and Arthur B. McDonald received the 2015 Nobel Prize in Physics.
The masses involved are tiny: every neutrino is at least several hundred thousand times lighter than the electron, the lightest particle of ordinary matter. Oscillation tells us that the masses exist and how far apart they are, but not how large they are. Measuring the actual mass needs other experiments, such as KATRIN.
Where it is measured today
Oscillations are now studied with neutrinos from every kind of source: from the Sun, from cosmic rays hitting the atmosphere, from nuclear reactors and from particle accelerators. Large detectors such as JUNO in China, the planned Hyper-Kamiokande in Japan and DUNE in the United States are designed to answer the remaining questions, for example which of the three neutrinos is the heaviest.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Intermediate
Flavour states and mass states
The neutrinos that take part in weak interactions, the flavour states νe, νμ and ντ, are not the states with a definite mass. Each flavour state is a quantum superposition of three mass states ν1, ν2 and ν3. Because the mass states travel with slightly different phases, the mixture changes along the way, and so does the probability of detecting each flavour.
The idea goes back to Bruno Pontecorvo, who proposed neutrino–antineutrino oscillations in 1957.3 Maki, Nakagawa and Sakata extended it to flavour mixing in 1962.
The decisive experiments
In 1998 Super-Kamiokande measured a zenith-angle dependent deficit of atmospheric muon neutrinos. SNO used heavy water to measure the electron-neutrino flux and the total flux of all flavours separately; the total matched the solar model. Takaaki Kajita and Arthur B. McDonald received the 2015 Nobel Prize in Physics for these discoveries.
Amplitude and wavelength
In the simplest case of two flavours, two numbers describe the whole effect. The mixing angle θ sets the amplitude: how large a fraction of the neutrinos can change flavour at most. The mass-squared difference Δm² sets the wavelength: how quickly the change happens as a function of distance L and energy E. The oscillation depends only on the ratio L/E, which is why experiments are often described by this ratio.
Nature has given us two very different wavelengths. The atmospheric splitting |Δm²₃₁| ≈ 2.5 × 10⁻³ eV² puts the first oscillation maximum for a 1 GeV neutrino at about 500 km. That is why accelerator experiments such as T2K (295 km, about 0.6 GeV), NOvA (810 km, about 2 GeV) and DUNE (1,300 km) are built over these distances. The solar splitting Δm²₂₁ ≈ 7.5 × 10⁻⁵ eV² is about thirty times smaller; for reactor antineutrinos of a few MeV it leads to oscillations over roughly 50 to 200 km, as seen by KamLAND and JUNO.
Disappearance and appearance
Experiments look for oscillation in two ways. In a disappearance measurement they count how many neutrinos of the original flavour arrive and compare with the expectation without oscillation. Reactor experiments such as Daya Bay, RENO and Double Chooz measured the third mixing angle θ₁₃ this way in 2012, at only 1 to 2 km from the reactors. In an appearance measurement they look for a flavour that the source does not produce: T2K observed electron neutrinos appearing in a muon-neutrino beam, and OPERA identified individual tau neutrinos in a beam from CERN.
Matter changes the picture
Inside matter, electron neutrinos feel an additional push: they can scatter coherently on electrons through the exchange of a W boson, which muon and tau neutrinos cannot. This changes their effective mixing. In the Sun the effect becomes resonant, the MSW effect: high-energy ⁸B neutrinos leave the solar core almost entirely in the second mass state and arrive as electron neutrinos only about a third of the time, while low-energy pp neutrinos follow ordinary vacuum oscillation with a survival probability of about 0.55. In the Earth the same effect differs for neutrinos and antineutrinos and for the two possible mass orderings, which is how long-baseline experiments hope to determine the ordering.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Expert
Mixing in the three-flavour framework
Flavour and mass eigenstates are related by the unitary Pontecorvo–Maki–Nakagawa–Sakata matrix U, parameterised by three mixing angles θ12, θ23, θ13 and a CP-violating phase δCP, plus two Majorana phases if neutrinos are their own antiparticles.
Two-flavour approximation
In vacuum and for two flavours, the appearance probability depends on one mixing angle and one mass-squared splitting:
Global fits (NuFIT 6.0) give Δm²21 ≈ 7.5 × 10⁻⁵ eV² and |Δm²3ℓ| ≈ 2.5 × 10⁻³ eV², with sin²θ12 ≈ 0.31, sin²θ13 ≈ 0.022 and θ23 close to 45°.4 In matter, coherent forward scattering on electrons modifies the effective mixing: the MSW effect, essential for solar neutrinos.
Matter effects and the mass ordering
In matter of constant electron density ne the electron-flavour potential V = √2 GF ne enters the Hamiltonian. For two flavours the effective mixing becomes sin²2θm = sin²2θ / [(cos 2θ − A)² + sin²2θ] with A = 2√2 GF ne E / Δm². The resonance at A = cos 2θ exists for neutrinos if Δm² > 0 and for antineutrinos if Δm² < 0. In long-baseline accelerator experiments and in atmospheric neutrinos crossing the Earth's mantle and core this asymmetry is the main handle on the mass ordering; JUNO instead uses the interference of the fast Δm²₃₁ and Δm²₃₂ oscillations on top of the slow Δm²₂₁ wave in the reactor spectrum at 52.5 km, a measurement independent of matter effects.
CP violation
The phase δCP in the PMNS matrix makes P(νμ → νe) differ from P(ν̄μ → ν̄e) in vacuum. The leading CP-odd term is proportional to the Jarlskog invariant J = c₁₂s₁₂c₂₃s₂₃c₁₃²s₁₃ sin δCP, whose maximum possible size is about 0.033, comparatively large because θ₁₃ is not small. Separating genuine CP violation from the fake asymmetry produced by matter is the central analysis task of T2K, NOvA, DUNE and Hyper-Kamiokande. Current data do not yet establish CP violation; see δCP and the octant of θ₂₃ for the state of the fits.
Coherence and averaging
The plane-wave formula assumes that the mass eigenstates stay coherent. In a wave-packet description they separate after a coherence length that grows with E²/Δm²; beyond it, and whenever the energy or baseline resolution is coarser than the oscillation length, only the averaged probability P = Σi |Uαi|²|Uβi|² survives. This is the regime of solar neutrinos at Earth and of high-energy astrophysical neutrinos: for a source dominated by pion decay, with initial flavour ratio about 1 : 2 : 0, averaging yields roughly 1 : 1 : 1 at Earth, a prediction that IceCube and KM3NeT test with their flavour measurements.
Quiz
Test your knowledge
4 questions on this level · Print as worksheet
Open questions
Oscillations measure only differences of squared masses. The absolute mass scale, the ordering of the three masses and whether CP is violated in the lepton sector remain open.
Terms to explore
People
On the timeline
- 1957Pontecorvo proposes neutrino oscillations
- 1998Super-Kamiokande discovers neutrino oscillation
- 2002SNO measures the total solar neutrino flux
References
- Y. Fukuda et al. (Super-Kamiokande), “Evidence for oscillation of atmospheric neutrinos”, Phys. Rev. Lett. 81, 1562 (1998), doi:10.1103/PhysRevLett.81.1562
- Q. R. Ahmad et al. (SNO), “Direct evidence for neutrino flavor transformation from neutral-current interactions in SNO”, Phys. Rev. Lett. 89, 011301 (2002), doi:10.1103/PhysRevLett.89.011301
- B. Pontecorvo, “Mesonium and anti-mesonium”, Zh. Eksp. Teor. Fiz. 33, 549 (1957); Sov. Phys. JETP 6, 429 (1958), jetp.ras.ru/cgi-bin/e/index/e/6/2/p429?a=list
- I. Esteban et al., NuFIT 6.0, JHEP 12, 216 (2024), arXiv:2410.05380
Last updated: · Editorial standard
Cite this page
Text
Neutrino Atlas (2026). “Neutrino oscillation”. Retrieved 22 November 2026, from https://neutrino-atlas.com/learn/neutrino-oscillation/BibTeX
@misc{neutrinoatlas2026learnneutrinooscillation,
title = {Neutrino oscillation},
author = {{Neutrino Atlas}},
year = {2026},
url = {https://neutrino-atlas.com/learn/neutrino-oscillation/},
note = {Accessed 2026-11-22},
language = {english}
}