NOvA: how appearance, disappearance and the Earth fit together

A longer route tests oscillations differently

NOvA uses Fermilab’s NuMI beam and a far detector near Ash River, Minnesota, about 810 kilometers away. Because the detector sits off axis, its events cluster around two GeV. Both the travel distance and the energy window differ from T2K’s.

That difference matters because oscillation phases depend on distance and energy, while matter effects also depend on the path and the medium. Two experiments can favor similar vacuum behavior yet respond differently to the Earth. How well they complement each other has to be calculated; it is more than having twice as many events.

The longer baseline makes NOvA more sensitive to the way matter makes neutrinos and antineutrinos behave differently. Mass ordering still cannot be seen in a single interaction. It enters the prediction for whole spectra, together with the mixing angles, the CP phase and the uncertainty parameters.

Sources: [1] · [2] · [3]

Tracks and showers in segmented scintillator

NOvA’s near and far detectors are built from long cells filled with liquid scintillator. The layers alternate in orientation, so the collected light gives several views of each interaction. A muon tends to leave a long track; an electron starts an electromagnetic shower with a different shape.

Events are classified with calibrated data and simulation, and backgrounds are checked against control data. Image-based selection still misidentifies some particles. Its efficiency and its mistakes must be built into the prediction, especially for neutral-current interactions that look like electron-flavor signal.

Energy reconstruction adds the charged lepton to the visible hadronic system. Some energy hides in nuclear breakup, in neutrons or in particles that leave the active region. A clean-looking image therefore need not give the exact incoming energy, and response uncertainties remain part of the measurement.

Sources: [4] · [1] · [5]

One fit must explain appearance and disappearance together

The muon-flavor spectrum tests disappearance; electron-flavor candidates test appearance. Running the beam in both polarities adds the matching antineutrino samples. The analysis must predict all these selections at once instead of picking whichever channel looks most striking.

A change in the atmospheric mixing angle can shift appearance and disappearance in different ways. The mass ordering and the CP phase can change the appearance rate in partly similar ways. Spectral shapes and the other channels help tell them apart, but with limited statistics some combinations stay hard to distinguish.

A near detector built with the same technology cancels several uncertainties when its data are extrapolated to the far detector. Some differences in acceptance and energy response remain. The final constraints come from a likelihood that includes flux, interaction and detector uncertainties, not from dividing two uncorrected counts.

Sources: [1] · [6] · [3]

The Earth treats neutrinos and antineutrinos differently

Ordinary matter contains electrons but no matching population of positrons. Electron-flavor neutrinos and antineutrinos therefore feel charged-current forward-scattering potentials of opposite sign. Their propagation can differ even if the intrinsic CP phase in vacuum conserves the symmetry.

This asymmetry in the surroundings helps ordering measurements but complicates a CP test. The analysis needs the path, a model of the matter density and their uncertainties. Reversing the horn polarity changes the beam; the Earth it crosses is still made of matter.

Comparing spectra and baselines helps pull these effects apart. T2K’s shorter route and NOvA’s longer one give different combinations of the same unknowns. A difference between counts has to be checked against this full prediction before it can be attributed to intrinsic CP violation.

Sources: [3] · [1] · [5]

The joint T2K–NOvA fit rests on shared assumptions

The first joint T2K–NOvA oscillation paper appeared in October 2025. It fitted the two experiments’ earlier published datasets together, used external oscillation constraints and studied how systematic errors correlate between the experiments. It was a combined analysis, not a new detector or a simple merger of event lists.

The paper found no statistically strong preference for either mass ordering, and its CP interpretation depended on which ordering was assumed. That conditional answer says more than a claim that the two experiments together proved why the Universe contains matter.

Published data products have limits too. A likelihood surface or binned prediction supports particular parameter comparisons; reanalysis at detector level needs more information on response and selection. Check the version and the paper attached to a release before treating it as the latest general result.

Sources: [5] · [6]

Try it in the Lab

Primary sources & revision

1. NOvA Collaboration · Bayesian constraints on oscillation parameters (2023/2024)
2. Adamson et al. · The NuMI Neutrino Beam (2015/2016)
3. Particle Data Group · Neutrino Masses, Mixing, and Oscillations (2024)
4. NOvA Collaboration · First electron-neutrino appearance measurement (2016)
5. T2K & NOvA · Joint neutrino oscillation analysis (22 October 2025)
6. NOvA Collaboration · Data releases

First published 2026-10-11; last revised 2026-10-13. 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.