How many protons does it take to make a neutrino?

The NuMI target is mounted in Fermilab's MIPP experimental hall.

The NuMI target is mounted in Fermilab’s MIPP experimental hall.

The Neutrinos at the Main Injector beam facility, known as NuMI, was built at Fermilab so that scientists could have a tunable source of neutrinos with which to conduct experiments. The NuMI beam has been a fantastic success, leading to important measurements made by the MINOS, MINERvA and other experiments. The beam has now been “retuned” to optimize the sensitivity of the NOvA experiment, which expects to have first neutrino oscillation results in 2015.

The NuMI beam is produced by smashing more than 10 million million high-energy protons extracted from the Main Injector accelerator into a specially designed graphite target every couple of seconds. The protons break apart the nuclei of the carbon atoms in the graphite, producing short-lived pions and kaons. Some of these newly produced particles escape the target and decay into muons and neutrinos. The muons are absorbed in the underground rock, but the neutrinos stream right through the rock and form the NuMI beam.

So just how many neutrinos does Fermilab produce at the NuMI beam facility? This seemingly simple question has not been so easy for scientists to answer. It is also an important question, since some measurements using the NuMI beam require reliable, precise knowledge of the number of neutrinos passing through their detectors. This number is known as the neutrino flux. Attempts to measure the neutrino flux directly are complicated by the fact that the neutrino properties known as cross sections needed to determine the flux have large uncertainties.

Pion yields per proton on target as a function of longitudinal momentum (pz) in steps of momentum in the transverse direction (pT). Different colors and markers represent steps of pT, and the yields are scaled such that the points in different pT steps do not overlap.

Instead we rely on computer simulations of many millions of proton interactions in the NuMI target to predict the neutrino flux. But these simulations require precise knowledge of how proton, pion and kaon particles — the neutrino parent particles — are produced when they interact with the target material, and there are rather large uncertainties on particle production. As a result, the neutrino flux uncertainties are three to five times larger than any other uncertainty in neutrino cross section measurements.

In order to improve the neutrino flux prediction, the MIPP (Main Injector Particle Production) collaboration counted the number of neutrino parents produced using an actual NuMI target (shown installed in the MIPP experimental hall in the above picture) and recently released results of charged-pion production yields as a function of pion momentum. The results, shown below, have been published in Physical Review D.

These data may be directly compared to the results of the computer simulations used by all NuMI experiments, and preliminary studies indicate significant differences between the simulation and the data. More importantly, the newly released MIPP data provide an opportunity to directly correct the simulation of pion production off the NuMI target with data collected in a matching configuration. Several NuMI-based experiments, including MINERvA, MINOS and NOvA, are investigating ways in which to make best use of these important data from MIPP.

In the near future it is expected that the MIPP data will help significantly reduce NuMI flux systematic uncertainties, making NuMI the world’s best understood accelerator neutrino beam and making Fermilab the place to do neutrino physics.