The Three Neutrino Flavors and Their Interactions

The neutrino comes in three kinds. We call them flavors, and we label them by the charged lepton each one partners with in a weak interaction: the electron neutrino ${\nu }_e$ pairs with the electron, the muon neutrino ${\nu }_{\mu }$ with the muon, the tau neutrino ${\nu }_{\tau }$ with the tau lepton. No fourth active flavor exists at observable mass, and the Standard Model contains exactly these three.

How did we establish that there are three and only three? How is flavor defined operationally — what is it, concretely, that distinguishes ${\nu }_e$ from ${\nu }_{\mu }$? And why, given that flavor is carefully conserved at every vertex, do neutrinos still end up changing from one flavor to another during flight?

Flavor is defined at the vertex

The weak charged-current vertex has a strict rule: a $W$ boson decaying to a lepton must produce a lepton and its partner neutrino together. $W^{+}\to e^{+}{\nu }_e$, $W^{+}\to {\mu }^{+}{\nu }_{\mu }$, $W^{+}\to {\tau }^{+}{\nu }_{\tau }$. The neutrino is labelled by the charged lepton it emerges alongside.

The converse holds at absorption: when a neutrino hits a target and interacts via charged current, the outgoing charged lepton is determined by the neutrino’s flavor. A ${\nu }_{\mu }$ produces a muon; a ${\nu }_e$ produces an electron. This is how experiments identify flavor: count the muons, count the electrons, compare against expectation.

There is no “intrinsic” measurement of flavor — no internal quantum number like spin or electric charge that can be read off the particle directly. Flavor lives entirely in the production and detection events.

Discovering the second flavor

Until the early 1960s, the obvious but unverified question was whether the neutrino emitted alongside a muon was the same particle as the one emitted alongside an electron. Pion decays ${\pi }^{+}\to {\mu }^{+}{\nu }_{?}$ produced copious neutrinos, but nobody had yet tested whether those neutrinos would produce electrons or muons when they interacted.

In 1962, Leon Lederman, Melvin Schwartz, and Jack Steinberger at Brookhaven built what amounted to the first accelerator-neutrino experiment. A proton beam hit a target, producing pions whose decays in flight released neutrinos. A massive steel shield absorbed everything except the neutrinos, which then entered a spark chamber. The chamber recorded muon tracks from the neutrino interactions — but no electron tracks. The neutrino partner of the muon is a different particle from the neutrino partner of the electron.

The experiment earned the 1988 Nobel Prize and established the two-neutrino picture. The muon neutrino joined the electron neutrino in the lepton taxonomy.

The third flavor

The tau lepton was discovered by Martin Perl at SLAC in 1975. By lepton universality, it should have a neutrino partner. But directly observing the tau neutrino turned out to be very hard: tau leptons are short-lived (about $3\times 10^{-13}$ s) and leave only a tiny track before decaying, requiring very high spatial resolution in the detector.

The confirmation came in 2000, by the DONUT (Direct Observation of the NU Tau) collaboration at Fermilab. An 800 GeV proton beam hit a tungsten target, producing charmed mesons whose prompt decays included $D_s\to \tau +{\nu }_{\tau }$. The beam was then filtered down to neutrinos and directed at nuclear emulsion — a high-resolution photographic medium that records charged-particle tracks in three dimensions. A handful of tau leptons were identified by their characteristic “kink” decay signature in the emulsion, completing the three-flavor picture after 38 years of waiting.

Three flavors, not four

How do we know there are only three? The Large Electron-Positron collider at CERN measured the total width of the $Z$ boson through the 1990s and extracted the partial width for decay into neutrinos (which contribute invisibly to the total). The result, averaged over four LEP experiments, gives $N_{\nu }=2.9840\pm 0.0082$ The number of active neutrino species with mass below $m_Z/2\approx 45$ GeV is exactly three. A fourth active flavor with mass in that range is ruled out.

This leaves room for sterile neutrinos — hypothetical states that do not couple to the $Z$ and would therefore not register in the LEP measurement. Short-baseline oscillation anomalies have kept sterile searches active, but no compelling evidence has emerged.

The atmospheric 2:1 ratio

One of the most important tests of flavor universality comes from atmospheric neutrinos. When a high-energy cosmic ray hits the upper atmosphere, it produces pions. Each pion decays through a two-step chain: ${\pi }^{+}\to {\mu }^{+}+{\nu }_{\mu },{\mu }^{+}\to e^{+}+{\nu }_e+{\bar{\nu }}_{\mu }$ A single pion therefore yields two muon-flavor neutrinos (one from the pion, one from the muon) and one electron neutrino. The expected ratio at detection is 2:1 muon to electron.

Super-Kamiokande, in the late 1990s, measured this ratio and found a significant muon deficit for neutrinos that had travelled through the Earth — pointing to oscillation from ${\nu }_{\mu }$ into something else. The discovery was the first direct evidence of neutrino mass, and Takaaki Kajita shared the 2015 Nobel Prize for the result.

Flavor is not conserved in propagation

Here is the puzzle. Every weak-interaction vertex respects flavor individually. Electron number, muon number, tau number — each should be conserved at production and at detection. Yet oscillation experiments see ${\nu }_{\mu }$ produced at one place and ${\nu }_e$ or ${\nu }_{\tau }$ detected at another.

The resolution is that neutrinos do not propagate as flavor eigenstates. They propagate as mass eigenstates — the three states ${\nu }_1$, ${\nu }_2$, ${\nu }_3$ that have definite energy-momentum relations $E_i=\sqrt{p^2+m_i^2}$. A flavor eigenstate is a specific superposition of mass eigenstates, and the superposition evolves as each component picks up a different phase during propagation. After a distance $L$ and for a neutrino of energy $E$, the flavor content has changed in a way that depends on $L/E$ and on the mass-squared differences $\Delta m_{ij}^2=m_i^2-m_j^2$.

The relation between the two bases is captured by the PMNS matrix, the subject of its own concept page. Its three mixing angles and Dirac CP phase are now the best-measured parameters in the neutrino sector.

Why the three-flavor framework matters

The three-flavor structure, simple as it looks, took the field seven decades to establish. And every oscillation measurement, every CP-violation search, every mass-hierarchy test depends on it. It is the platform on which the rest of neutrino physics is built.

It is also the launching point for beyond-Standard-Model questions: why three flavors and not one or five? Why these particular mixing angles? Is there a fourth, sterile, species hiding at eV or keV scales? Does lepton flavor conserve globally the way it seems to locally? These questions are where the next generation of experiments — IceCube-Upgrade, DUNE, Hyper-Kamiokande, KATRIN, neutrinoless double-beta-decay searches — will take the field through the 2030s.