RENO and Double Chooz: why repeat the mixing-angle measurement?

One mixing angle, several ways to measure it

The third mixing angle sets part of the pattern of electron-antineutrino disappearance at distances of about a kilometer from a reactor. A nonzero value also lets accelerator experiments see electron-flavor appearance. The angle therefore links the two kinds of experiment, though their signals and systematic uncertainties stay different.

Reactor disappearance is especially useful because, to leading order in vacuum, the survival probability does not depend on the CP phase. Reactors can pin down the angle without having to extract that phase at the same time. This holds within the standard oscillation model; reactor experiments still have their own sources of error.

Daya Bay’s 2012 observation was followed by RENO’s independent result, and Double Chooz had already reported an indication in a preprint submitted in December 2011. Each step has its own date and strength of evidence. Calling all three experiments interchangeable discoveries flattens that history.

Sources: [1] · [2] · [3] · [4]

RENO watched six cores from two distances

RENO used two similar detectors near the six-core Yonggwang plant in Korea, now called Hanbit. The 2012 paper put them 294 and 1,383 meters from the center of the reactor array. These are reference distances; each core sits at its own distance from each detector, and the prediction has to use those individual distances.

The near sample tells the analysis what the far sample should look like without the extra oscillation loss. Each core contributes with a different weight, so simply dividing the total counts is not enough. Target size, efficiency, running time and backgrounds also enter the comparison.

The first rate analysis reported disappearance at 4.9 standard deviations. Later spectral analyses used how the far-to-near deficit changes with energy to constrain the oscillation frequency as well as its amplitude. Two similar detectors can therefore test more than one feature of the oscillation model.

Sources: [2] · [5]

Double Chooz added its checks step by step

Double Chooz studied antineutrinos from two reactor cores in France. Its first analysis used only a far detector, about 1,050 meters away, and tied its flux prediction to Bugey-4 data. That early result was not yet a near–far comparison.

Changes in reactor power gave another check. Periods with different core powers change the expected signal, and periods with both cores off help constrain the backgrounds. Even with the reactors off, leftover decays and selection details mean reactor-related activity is not guaranteed to be exactly zero.

The later two-detector analysis, published in Nature Physics in 2020, also accepted a wider range of neutron captures. Captures on hydrogen and on gadolinium give different delayed signals and are sensitive to different backgrounds. Using both checks the detection efficiency, while the prompt signal comes from the same inverse beta decay in each case.

Sources: [3] · [6] · [7]

Compare ranges and methods, not just central values

The values below come from two analyses from different years, with different exposures and event selections. They measure the same standard mixing quantity with different uncertainties. The question is whether the reported ranges are compatible; the central values need not match digit for digit.

Statistical uncertainty comes from the finite number of events. Systematic uncertainty covers the relative detector response, backgrounds and reactor weights. A combined result must keep track of correlations: if three collaborations use the same nuclear prediction, that is not three independent tests of it.

Repeating a measurement helps because it changes the details most likely to go wrong. Different sites, power histories, calibrations and background methods make a local instrumental explanation hard to sustain. Agreement across those changes says more than multiplying headline significances or treating an average as free of error.

Sources: [2] · [7] · [8]

The angle is reliable even though the reactor spectrum is puzzling

Several reactor experiments also saw more events than their adopted predictions near five MeV in the prompt spectrum. A feature present in both detectors can largely cancel when the near and far spectra are compared. A disagreement with the source model therefore need not erase the near–far evidence for flavor change.

How well it cancels depends on how the two samples weight the cores and respond to energy, so analyses still check what is left over. The lesson is to separate the robust measurement from the error it suppresses, then quantify what remains instead of using cancellation as a slogan.

These measurements now feed into more complex fits of appearance and mass ordering. The Korean and French results show how knowledge builds up: a parameter becomes useful elsewhere because its measurement held up in different instruments, not because one experiment’s number was repeated often enough.

Sources: [5] · [7] · [9] · [4]

Try it in the Lab

Primary sources & revision

1. Daya Bay Collaboration · Observation of electron-antineutrino disappearance (2012)
2. RENO Collaboration · Observation of Reactor Electron Antineutrino Disappearance (2012; v2)
3. Double Chooz Collaboration · Indication for reactor electron antineutrino disappearance (2011/2012)
4. Particle Data Group · Neutrino Masses, Mixing, and Oscillations (2024)
5. RENO Collaboration · Observation of Energy and Baseline Dependent Reactor Antineutrino Disappearance (2015/2016)
6. Double Chooz Collaboration · Hydrogen-capture measurement with novel background rejection (2015/2016)
7. Double Chooz Collaboration · Total neutron capture measurement (2019 preprint; Nature Physics, 2020)
8. Particle Data Group · Statistics (2025)
9. Daya Bay Collaboration · Measurement of the Reactor Antineutrino Flux and Spectrum (2015/2016)

First published 2026-10-10; last revised 2026-10-12. Original explanatory text and diagrams by Sandbox Physics. Illustrations are schematic; no experimental event records are reproduced here. This is an independent educational publication, not an official Nobel or experiment collaboration publication.