It is no secret that neutrinos change flavor or oscillate as they travel from one place to another, and that the amount they change depends on how much time they have to change. This time is directly related to the distance the neutrino traveled and the energy of the neutrino itself. Measuring the distance is easy. The hard part is measuring the neutrino energy.
Experiments do this by measuring the energies of particles that get produced by the neutrino when it interacts in their detectors. But what happens if one of the produced particles, for example a neutron, leaves barely any of its energy in the detector?
Oscillation experiments have to predict how much energy is lost and then correct for that loss. These predictions depend on accurate models of how neutrinos interact, and those models have to be right not only for neutrinos but also for antineutrinos, which are particularly good at making neutrons.
The MINERvA collaboration analyzed data from interactions of antineutrinos that produced positively charged muons. Scientists looked at both the momentum and energy that was transferred to the nucleus in those interactions. By focusing on the kinematic region where only a neutron should be knocked out, they looked at the worst-case situation: Most of the hadronic energy will go missing. In this way, scientists directly measured the effects of an imperfect model for missing energy.
In order to appreciate why this new analysis of antineutrino interactions is exciting, you should know that over a year ago, MINERvA published a similar measurement with neutrino interactions producing negatively charged muons, where a proton is much more likely to be produced than a neutron. A proton’s energy is much easier to measure than a neutron’s in a detector such as MINERvA. This analysis found that, for neutrino interactions on a proton-neutron pair (rather than on only one of those two particles), scientists observed a much larger number of events than the state-of-the-art models predicted. Neutrino cross section enthusiasts are never surprised when models don’t describe data. So here is the surprise: When they used the neutrino results to change the antineutrino model to predict the antineutrino data described above, it worked. You can see the improvement in the middle of the plots above.
This is interesting, because this is new information about how well models do and where they fall short. Searches for CP violation or “what makes matter special compared to antimatter” depend on comparing neutrino and antineutrino samples and looking for small differences. Large, unknown differences between neutrino and antineutrino reaction rates would hide the presence or absence of CP signatures. We are converging on better models that describe both neutrino and antineutrino data.
Miranda Elkins and Rik Gran, University of Minnesota-Duluth, were two of the scientists who analyzed this result.