Neutrinos 101 — Part 3: Three flavors and oscillation

In Part 2 we saw how the neutrino went from a desperate hypothesis to a directly detected particle. By 1956, the field had one kind of neutrino: the antineutrino emitted in beta decay. Within a decade, that number went up to two. By 2000, it was three. And along the way, physicists discovered something even stranger about the neutrino: it does not have a fixed identity.

A second flavor (1962)

The first sign that something unusual was going on came from looking at how muons — heavier cousins of the electron — behave when produced by neutrino interactions. By the late 1950s, physicists had noticed that the neutrinos involved in muon physics seemed to behave differently from the neutrinos involved in electron physics. The two were either the same particle being weird, or two different particles entirely.

Leon Lederman, Melvin Schwartz, and Jack Steinberger tested the question at Brookhaven National Laboratory in 1962. They fired pions (a kind of unstable particle) at high speed, let them decay into muons and muon-associated neutrinos, blocked everything except the neutrinos with 13 meters of steel, and watched a downstream detector for interactions.

The result was clear: the neutrinos from muon-related decays produced only muons when they interacted, not electrons. They had to be a distinct kind of neutrino. The team had discovered the muon neutrino as separate from the electron neutrino. They shared the 1988 Nobel Prize for the result.

A third flavor — the tau neutrino — was inferred theoretically after Martin Perl discovered the tau lepton in 1975. It was directly observed in 2000 by the DONUT experiment at Fermilab. All three flavors had been confirmed.

Pontecorvo’s prediction (1957)

Before any of this experimental work, in 1957, an Italian-Soviet theorist named Bruno Pontecorvo was thinking about something even stranger. He wondered: if there are different kinds of neutrino, can one kind transform into another while traveling?

The idea sounded crazy. Particles in the Standard Model don’t change identity in flight. An electron stays an electron. A photon stays a photon. Pontecorvo’s question — phrased in terms of “mesonium” oscillation by analogy with the well-known K⁰-K̄⁰ system — was that neutrinos might be different.

The mathematical structure he sketched (and that Maki, Nakagawa, and Sakata extended in 1962) said: if neutrinos have small but non-zero masses, and if the “flavor” states (electron, muon, tau) are quantum-mechanical mixtures of underlying “mass” states (called ν₁, ν₂, ν₃), then a neutrino produced as one flavor will have a probability of being detected later as a different flavor. The probability depends on how far the neutrino has traveled and at what energy.

This is neutrino oscillation. The 1957 framework — slightly extended by the 1962 MNS paper — described it mathematically nearly four decades before any experimental confirmation. The framework is now called the PMNS mixing matrix, after Pontecorvo, Maki, Nakagawa, and Sakata.

The solar neutrino problem

The first hint that oscillation might actually be happening came from solar neutrino experiments. Raymond Davis Jr. started running his chlorine-37 detector at the Homestake gold mine in 1968, expecting to detect about 8 solar electron neutrinos per day according to John Bahcall’s standard solar model. He found only about 3 — a deficit by a factor of roughly three that persisted across years of running.

Either the standard solar model was wrong about how the Sun produces energy, or the missing neutrinos had somehow disappeared during their 8-minute journey from the Sun. Other experiments through the 1980s and 1990s — Kamiokande, GALLEX, SAGE — found similar but energy-dependent deficits. The pattern looked exactly like what oscillation would produce, but nobody could prove it definitively.

The 1998 discovery

The breakthrough came in 1998 with Super-Kamiokande’s atmospheric neutrino announcement. Super-K measured muon neutrinos produced by cosmic rays hitting the Earth’s upper atmosphere. The expected pattern was: roughly the same rate of muon neutrinos coming down from the sky (short travel distance through atmosphere) and going up through the Earth (long travel distance, passing through the entire planet).

Super-K found that upward-going muon neutrinos were missing. The deficit depended on the zenith angle in a specific way, perfectly matching what oscillation would produce if neutrinos had mass-squared differences around 10⁻³ eV². The result was at 5σ statistical significance — well above the discovery threshold.

Three years later, the SNO experiment in Canada closed the loop on the solar neutrino problem. SNO used heavy water as its target, which let it measure both the electron-neutrino flux from the Sun and the total all-flavor flux. The total matched Bahcall’s standard solar model exactly. The electron-neutrino flux was only about a third — the missing two-thirds had converted to other flavors during the journey.

The 2015 Nobel Prize went to Takaaki Kajita and Art McDonald for the discovery.

What oscillation requires

For neutrinos to oscillate between flavors, two things must be true:

1. The flavor states must be mixtures of mass states. The electron neutrino, for example, is not a particle with a definite mass — it is a quantum-mechanical superposition of three states with three different masses.
2. At least two of the three masses must be non-zero (and different from each other). If all three masses were zero or identical, the superposition would propagate without any change. Oscillation requires the components to evolve at different rates so that the interference between them produces flavor-changing probabilities.

The 1998 and 2001 discoveries therefore established not just that oscillation occurs, but also that neutrinos have mass. This was a foundational change to the Standard Model, which had been built on the assumption that neutrinos are exactly massless.

The parameters being measured

Modern neutrino oscillation is described by six numbers:

• Three mixing angles (θ₁₂, θ₂₃, θ₁₃) — describing the relative orientations of the three mass states in flavor space. All three are now well-measured.
• One CP-violating phase (δ_CP) — describing the asymmetry between neutrino and antineutrino oscillation. Currently being measured.
• Two mass-squared differences (Δm²₂₁ and Δm²₃₂) — describing how different the three mass states are from each other. The magnitudes are well-measured. The sign of Δm²₃₂ — the “mass ordering” — is still unknown.

Together, these six numbers form what is called the PMNS matrix, the central bookkeeping object of neutrino oscillation. For the deeper math, see our How neutrino oscillation works primer.

What comes next

In Part 4, we’ll look at where neutrinos come from in much more detail — from the Sun and supernovae to cosmic rays, reactors, the Earth’s interior, and the Big Bang. Each source has its own physics and has been used to probe specific oscillation parameters.

Neutrino Times Editorial Team. (2026, January 17). Neutrinos 101 — Part 3: Three flavors and oscillation. Neutrino Times. https://neutrino-times.com/articles/neutrinos-101-part-3-three-flavors-and-oscillation/

Neutrino Times Editorial Team. "Neutrinos 101 — Part 3: Three flavors and oscillation." Neutrino Times, January 17, 2026. https://neutrino-times.com/articles/neutrinos-101-part-3-three-flavors-and-oscillation/.

Neutrino Times Editorial Team. "Neutrinos 101 — Part 3: Three flavors and oscillation." Neutrino Times, 17 Jan. 2026, https://neutrino-times.com/articles/neutrinos-101-part-3-three-flavors-and-oscillation/.

@misc{neutrino-times-neutrinos-101-part-3-three-flavors-and-oscillation,
  author       = {Neutrino Times Editorial Team},
  title        = {Neutrinos 101 — Part 3: Three flavors and oscillation},
  howpublished = {Neutrino Times},
  year         = {2026},
  month        = {jan},
  url          = {https://neutrino-times.com/articles/neutrinos-101-part-3-three-flavors-and-oscillation/},
  note         = {Accessed: 2026-01-17}
}

TY  - GEN
TI  - Neutrinos 101 — Part 3: Three flavors and oscillation
AU  - Neutrino Times Editorial Team
PY  - 2026
DA  - 2026-01-17
PB  - Neutrino Times
UR  - https://neutrino-times.com/articles/neutrinos-101-part-3-three-flavors-and-oscillation/
ER  -