Neutrino Atlas
EN
  1. Terms
  2. Experiments & detectors
  3. Atmospheric & Large water Cherenkov experiments
  4. Fréjus detector

Fréjus detector

ConfirmedLevel 3 of 4Also called: Fréjus experiment, Fréjus proton-decay detector

900-tonne iron tracking calorimeter in the Modane underground laboratory that saw no atmospheric neutrino anomaly in the late 1980s.

French entrance of the Fréjus road tunnel
The Fréjus road tunnel, beside which the Modane laboratory housed the 900-tonne Fréjus detector from 1984 to 1988.Image: Florian Pépellin · CC BY-SA 3.0 · commons.wikimedia.org

The Fréjus detector ran from February 1984 to September 1988 in the Laboratoire Souterrain de Modane. It consisted of 912 layers of 3 mm iron plates interleaved with flash tubes and Geiger tubes, reaching its full 912 t in June 1985. Built to search for proton decay, it found an electron-to-muon ratio of atmospheric neutrinos consistent with no oscillation, in apparent conflict with Kamiokande and IMB.

In plain words

In the 1980s several large detectors went underground to look for proton decay, and all of them also recorded atmospheric neutrinos. The water Cherenkov detectors Kamiokande and IMB found fewer muon neutrinos than expected compared with electron neutrinos. Fréjus, built from iron plates rather than water, did not. For several years this disagreement was a real argument against the idea that neutrinos oscillate.

How the detector works

The detector was installed in the Modane laboratory, dug beside the Fréjus road tunnel between France and Italy under about 4,800 metres water equivalent of rock. It was a fine-grained tracking calorimeter of 6 m × 6 m × 12.3 m: 912 vertical iron plates, each 3 mm thick, alternating with planes of 5 mm plastic flash tubes that recorded tracks in two orthogonal views, with Geiger tubes providing the trigger. Construction proceeded in stages, from 240 t when data taking began on 19 February 1984 to the full 912 t in June 1985. The collaboration included groups from Aachen, Wuppertal, Orsay, Palaiseau and Saclay.

Key results

With a fiducial exposure of 1.56 kt·yr, the 1990 analysis found electron-to-muon ratios of 0.53 to 0.57 in three independent methods, against Monte Carlo expectations of 0.56 to 0.61 without oscillation. The collaboration therefore saw no muon-neutrino deficit and set oscillation limits, in agreement with NUSEX but not with the water detectors. A 1995 paper derived absolute atmospheric νe and νμ spectra between 0.32 and 30 GeV. The proton-decay search, with 2.0 kt·yr, found no signal and set partial lifetime limits up to 10³² years.

Why it matters for neutrino physics

The Fréjus result raised the question of whether the anomaly came from the water Cherenkov technique itself. This motivated the iron calorimeter Soudan 2, which later confirmed a low ratio. The 1998 Super-Kamiokande zenith-angle result settled the issue; the small Fréjus data set with its systematic uncertainties is now regarded as a statistical fluctuation rather than a contradiction.

Frequently asked questions

Why did Fréjus disagree with Kamiokande?

The reason was never fully identified. Low statistics and systematic uncertainties in the iron calorimeter, about 0.08 on the ratio, are the usual explanation.

Was Fréjus a neutrino experiment?

It was built for proton decay, but atmospheric neutrinos were its main background and became an important measurement in their own right.

Original papers

  1. Final resultK. Daum et al. (Fréjus Collaboration), Determination of the atmospheric neutrino spectra with the Fréjus detector, Z. Phys. C 66, 417 (1995)DOI 10.1007/BF01556368

Related terms

Mentioned in 3 pages

Last updated: · Editorial standard

Cite this page

Text

Neutrino Atlas (2026). “Fréjus detector”. Retrieved 22 November 2026, from https://neutrino-atlas.com/terms/frejus-detector/

BibTeX

@misc{neutrinoatlas2026termsfrejusdetector,
  title        = {Fréjus detector},
  author       = {{Neutrino Atlas}},
  year         = {2026},
  url          = {https://neutrino-atlas.com/terms/frejus-detector/},
  note         = {Accessed 2026-11-22},
  language     = {english}
}