LEP and the Counting of Neutrino Species

By 1989, the Standard Model of particle physics had been validated in countless precision measurements at lower energies. The discovery of the electroweak gauge bosons W and Z in 1983 had established the framework. What remained, at the precision frontier, was a question that had concerned physicists since the early 1960s: how many neutrino species exist?

The answer to this question was not just academic. The minimal Standard Model predicted three families of fermions (quarks and leptons), with three neutrinos (one per family). But there was no a priori reason that a fourth, fifth, or higher family should not exist. Direct accelerator searches for charged leptons heavier than the tau had set lower bounds, but these bounds were not stringent enough to definitively close the question.

The Large Electron-Positron Collider — LEP — at CERN, operational from 1989 to 2000, provided the answer. By measuring the precise properties of the Z boson, the four LEP detectors (ALEPH, DELPHI, L3, OPAL) determined the number of light neutrino species to be: $N_{\nu }=2.984\pm 0.008$ or, in plain language: there are exactly three of them. A fourth would have been visible, and it was not.

This single number, more than any other LEP result, is the experimental foundation for the three-flavor Standard Model neutrino sector. This post explains what was measured, how, and why the conclusion is so robust.

The Z resonance

LEP collided electrons and positrons at the Z-pole — a center-of-mass energy of approximately 91.2 GeV, exactly the mass of the Z boson. At this energy, the cross-section is dominated by resonant Z production: $e^{+}+e^{-}\to Z\to \text{anything}$ The Z then decays to all kinematically allowed final states. The branching fractions are:

• Hadrons (quark pairs that hadronise to jets): ~70%
• Charged leptons (e⁺e⁻, μ⁺μ⁻, τ⁺τ⁻): ~10% combined
• Neutrino pairs (ν̄_eν_e, ν̄_μν_μ, ν̄_τν_τ): ~20% combined

The neutrino decays are invisible to the detector — the neutrinos escape with the missing energy and momentum. They are seen only by their absence in event reconstruction.

The total decay rate of the Z (its “total width”, ${\Gamma }_Z$) is the sum of all decay channels. The width is proportional to the inverse of the resonance lifetime and is measured by fitting the cross-section as a function of center-of-mass energy near the peak.

For each visible decay channel, the partial width can be measured directly by counting events of that type. The “invisible width”, ${\Gamma }_{inv}$, is then ${\Gamma }_{inv}={\Gamma }_{Z,\text{total}}-{\Gamma }_{ee}-{\Gamma }_{\mu \mu }-{\Gamma }_{\tau \tau }-{\Gamma }_{hadrons}$ i.e., the difference between the measured total and the sum of measured visible channels.

The Standard Model predicts that each light neutrino contributes a partial width of ${\Gamma }_{\nu \bar{\nu }}\approx 167$ MeV. Therefore the number of light neutrino species is $N_{\nu }={\Gamma }_{inv}/{\Gamma }_{\nu \bar{\nu }}$

What was measured

The four LEP experiments measured the cross-section as a function of center-of-mass energy across the Z resonance, accumulating about $10^7$ Z events between 1989 and 1995. Each experiment independently extracted:

• The total width ${\Gamma }_Z=2.4952\pm 0.0023$ GeV
• The visible widths from electron, muon, tau, and hadronic channels
• The invisible width by subtraction

The combined result from all four experiments, after careful averaging and consistency checks: $N_{\nu }=2.984\pm 0.008$

The systematic error is dominated by the luminosity calibration (knowing exactly how many e⁺e⁻ collisions occurred — needed to convert event counts to cross-sections). The small deficit from the integer 3 has been investigated extensively; modern analyses with improved theoretical inputs reduce it further toward 3.00, well within statistical uncertainty.

What it means

The number 2.984 is the count of light active neutrino species. “Light” means with mass below $M_Z/2\approx 45$ GeV. “Active” means coupling to the Z boson via the standard weak interaction. The LEP measurement therefore:

Definitively rules out a fourth active neutrino family with mass below 45 GeV. A fourth charged-lepton family with mass below 90 GeV would have been seen as direct production at LEP. A heavy fourth family with the same neutrino flavor structure (charged lepton plus active neutrino) is excluded.

Does not rule out sterile neutrinos. A neutrino that doesn’t couple to the Z would not contribute to ${\Gamma }_{inv}$ and would be invisible to LEP. The LSND/MiniBooNE/gallium anomalies, if they are sterile-neutrino phenomena, are entirely consistent with $N_{\nu }=3.0$ at LEP.

Does not rule out heavy neutrinos. Right-handed Majorana neutrinos with TeV-scale masses, as in the seesaw mechanism, would not contribute to LEP’s measurement. Such neutrinos remain a viable beyond-Standard-Model possibility.

A subtle but important systematic

One detail of the LEP measurement deserves mention because it has been the subject of repeated reanalysis.

The luminosity at LEP was measured using small-angle Bhabha scattering ($e^{+}{e}^{-}\to e^{+}{e}^{-}$ at small angles), which has a precisely calculable QED cross-section. The luminosity calibration was the dominant systematic on the cross-section measurements at the time of LEP’s running.

In 2019, an updated theoretical evaluation of the Bhabha-scattering cross-section was published, including next-to-next-to-leading-order QED corrections that had not been fully accounted for. Re-applied to the LEP data, this correction shifts the extracted $N_{\nu }$ slightly upward, from 2.984 to about 2.996 — within $1\sigma$ of 3. The shift is small, but it illustrates the principle that “definitive” measurements continue to be refined as theoretical inputs improve.

The currently recommended value, with the updated luminosity inputs, is approximately: $N_{\nu }=2.996\pm 0.007$ again consistent with 3.

Why this measurement matters

The LEP $N_{\nu }$ measurement is foundational for several reasons.

It establishes the three-flavor Standard Model. Together with subsequent direct observation of the tau neutrino at DONUT (2000), the LEP result confirms that the Standard Model has exactly three lepton families. All subsequent experimental work in neutrino physics builds on this assumption.

It was the first 5σ-class precision measurement of the lepton sector. Pre-LEP, lepton-sector quantities were known only to a few percent. LEP brought relative precision to the per-mille level, comparable to the best quark-sector measurements.

It complements cosmology. Cosmological observations of Big Bang Nucleosynthesis and the cosmic microwave background also count light, relativistic species at the time of decoupling — the parameter $N_{\text{eff}}\approx 3.04$, with the ~0.04 excess from incomplete decoupling and finite-temperature QED effects. The agreement between LEP and cosmology is one of the most stringent independent cross-checks of both the Standard Model and the standard cosmological model.

It anchors the count of cosmic neutrinos. The relic neutrino density of about 336 per cubic centimetre relies on knowing that there are exactly three species. If LEP had found four or five, the cosmological neutrino picture would be qualitatively different.

Legacy of LEP

The Large Electron-Positron Collider operated for 11 years and was decommissioned in 2000 to make room for the Large Hadron Collider in the same tunnel. Its legacy in neutrino physics is concentrated in this single number — three.

Three neutrino flavors. Three mass eigenstates. Three mixing angles in the PMNS matrix. The entire framework of modern neutrino oscillation physics is built on the three-flavor structure that LEP established.

The number 2.984 (now 2.996) is, in the literal sense, the most precisely measured property of any neutrino. It is also one of the most thoroughly cross-checked, with direct LEP, indirect cosmology, and direct accelerator (DONUT, OPERA) measurements all converging on the same answer.

When the next generation of $e^{+}{e}^{-}$ Higgs factories — FCC-ee at CERN, ILC in Japan, the Chinese CEPC — comes online, they will repeat the LEP measurement at higher precision, possibly reaching $N_{\nu }=3.000\pm 0.001$. But the qualitative answer — exactly three light active neutrino species — is unlikely to change.