A. Rubbia: Results from T2K
Andre’ Rubbia gave a nice talk today at NEUTEL11, providing an overview of the T2K experiment and the first results on theta_13. I offer a quick transcript of his presentation below.
He presented the first oscillation results from T2K in world premiere. Before discussing this, he mentioned that the J-Parc accelerator and the T2K experiments were not hit by tidal wave. There were no reports of casualties in the experimental environment, despite the location of the experimental setup near the zone hit by the earthquake on March 11th. He mentioned that the priorities for the lab are now to restore electricity, water supplies and gas.
He described the T2K experimental setup, and the data taking which started in January 2010 unti March 2011. The first oscillation results are only shown from the 2010a data, taken until June 2010.
T2K is a large international collaboration, counting 500 members, located in Japan. It was purposely built to measure the mixing angle theta_13. It relies on a 30 GeV proton beam with an intensity of 750 kW. Interactions are recorded at the site and in the Kamioka mine, 295 km away, where a high-intensity muon neutrino beam is sent.
30 GeV protons are shot on graphite, then there is a focusing system, followed by a decay volume, muon monitors, and an off-axis configuration: the beam is not pointing to the SuperKamiokande experiment, but 2.5 degree off it. On-axis they have a near detector. The idea is that the flux is maximized in the region of interest, with neutrinos with energy below 1 GeV, where the effect of an oscillation would be maximized.
The decay volume is 100 m long, filled with helium. The focusing system has three horns working at 250 kAmperes. A set of near detectors are installed. The primary beam is also monitored to know the precise profile of the beam.
The on-axis detector has 14 identical modules in a cross plus two off-cross. This provides a beam coverage area of 100 square meters, 10×10. Each module has a mass of 7 tons. 700 neutrino interactions are observed per day at 50 kW operation to monitor the steering of the beam, a critical parameter given that a 1 mrad difference in the pointing corresponds to a 2% change in the flux that SuperKamiokande will receive at peak energy: they do need to know this well.
Inside the off-axis near detector is located a neutral pion detector built with scintillator planes and water and lead-brass layers, optimized for gamma detection. A fine-grained detector is made of thin,wide scintillator planes, which also provide active mass.
As for the far detector, this is SuperKamiokande, which requires no introduction (for a post on SK and its results see below). In SK, Cherenkov rings discriminate electron and muon events. Backgrounds are from neutral pions from NC interactions, and from a residual nu_e contamination in the beam. Both are at the same scale.
Data shown today comes from run 2010a, Feb 24 to June 26 2010.
The new data have been accumulated in runs from Nov 16th 2010 to March 11th 2011. Results from these data are not yet ready to be shown.
Anyway, the 2010a run amounts by itself to an equivalent 3.23×10^19 protons on target.
Events are synchronized with GPS. All events within a window of +-500 microseconds are recorded. The synchronization is corrected for time of flight. The event rate at the near detector is well compatible with the expectation, the ratio being 1.073+-0.01+-0.04.
Andre’ then discussed the data analysis. The event reduction starts from all events collected in the 1 ms window around the spill. They first throw away events with hits in the outer detector, then they cut on the energy, reject “flashers”, and they remain with FC (fully contained) candidates.
They observed 33 events fully contained in the 2010a data. Two of them got rejected by additional cleanup. Accidental background is less than 0.001 events at this stage. Data reduction proceeds from here, with the 33 fully contained events. A fiducial volume cut reduces them, and a requirement on visible energy above 30 MeV leads to 23 events. Single ring muon-like are 8 with muon momentum above 200 MeV, while single-ring electron-like events with electron momentum above 100 MeV are 2. Multi ring events are 13.
So the focus is on the 8 mu-like and 2 e-like events for the measurements of T2K. Rubbia showed several nice event displays for this ring events set.
He explained that they are working on two parallel analyses at this stage. The first is a nu_mu disappearance analysis, where they predict the yields for both the null oscillation hypothesis and the oscillations hypothesis with delta m^2_23=2.4E-3 eV^2, and maximal angle. They can thus compare with expectations;
For nue appearance analysis, additional cuts are applied for background supression. The background-only case consists in oscillations coming only from muon neutrino disappearance, which implies theta_13=0. They compare with a signal plus background analysis where sin^2 2 theta_13 is set to 0.1. They then fit for this parameter.
At the moment, they are just doing a counting experiment: it does not make much sense yet in studying a spectrum at this stage. So they only need to predict the flux, which means they need proton beam data, and measurements in the near detector. They compare observed to data to normalize expected events. Need neutrino cross sections, tuned to external data. Then, for the far detector, once the data has been reduced, they correct the number of backgrounds and signal expected by the ratio measured in the near detector, and this is how they calculate the expected events in SK. Then they can extract the oscillation parameters. At the moment they do not use the measured near spectrum and the near/far ratio.
For the flux prediction, Rubbia explained that they are starting from a specific Monte Carlo model, tuned to experimental data. They use for historical reasons FLUKA 2008, where however pion production differential cross sections are corrected with dedicated CERN NA61 data. Propagation of secondaries in the focusing system and decay volume is performed with GEANT3, and for meson decay they calculate and measure the flux in the near and far detectors.
The tuning of fluxes is below the level of 5%. The total muon neutrino flux at SuperKamiokande has a fractional error of 20 to 40%, dominated by pion and kaon uncertainty.
Predictions for the neutrino fluxes in the nu_mu disappearance analsis shows what difference they expect for oscillation and no oscillation. The off-axis configuration is tuned to expose different fluxes depending on the delta_m^2, since they are sitting at the maximum of the predicted oscillation.
The data are 8 events, compared to 22.8+-3.2 for no oscillation, 6.3+-1.0 for oscillation at dm^2=2.4E-3 eV^2, for a maximal effect (sin^2 2theta_23=1.0). Data are therefore consistent with oscillation parameters measured by MINOS and other experiments. The fitting of parameters is underway.
For the electron neutrino appearance analysis, they have 2 events, with a prediction of 1.5 events +-0.7 for no oscillation, and 1.3+-0.6 for oscillation and theta_13=0. But this is before some additional cleanup. In the end, they have just one event surviving further cuts: one of the two event candidates is in fact killed by requiring no electrons from decay of muons. At the end of the day there’s only one candidate.
Rubbia explained that they expect 0.3+-0.07 events for no signal, 1.20+-0.23 for signal and background together, with sin^2 (2theta_13)=0.1. The p-value is about 29% that we should see >=1 event if that is the value of the ephemeral angle. This can be translated on a result for theta_13. They use two independent statistical procedures, FC and classical one-sided. From these they derive upper limits on the angle for normal and inverted hierarchy. Their FC result is at 0.50 or 0.59, 90% cL.
Of course, more data are on tape waiting to be analyzed. With 10^22 protons on target (which they might one day obtain) the sensitivity on the sin^2 2theta_13 should go down to 10^-2.
In conclusion, Rubbia explained that T2K searched for numu->nue oscillations, and determine the atmospheric sector parameters. They found one event surviving all cuts, when they expected 0.3 for the null hypothesis. For the disappearance analysis, the observed candidates are consistent with oscillation parameters measured by SK, K2K, MINOS.