R.Sulej: Sterile Neutrino Search With ICARUS
[Before I report on this presentation, I need to point out that due to the unavailability of slides at the time of the talk, I cannot complement these writeups with illustrations. I refer readers to the web site of the conference, where all presentations are uploaded with a few hours of delay. Apologies – TD]
Sterile neutrinos were hypothesized in a paper by Pontecorvo in 1957, as particles not interacting in any way except gravitationally. They of course are extremely hard to detect. They could be contributing to dark matter in the universe. Sterile neutrinos may also mix with ordinary neutrinos with a mass term. Evidence may be building up by anomalies observed by several experiments. The speaker showed a table of several resumts from LSND, MiniBoone, Gallium experiments, and ones from reactors, which are all in the 1 to 3 standard deviation ballpark.
So that is motivation for ICARUS, which searched for electron neutrinos in the 10-30 GeV region from the CNGS muon neutrino beam. ICARUS consists in two identical T300 modules, two TPC per module. The active mass of liquid argon is of 476 tons. This is equipped with three readout wire planes with different angular orientations, with a total of 53000 wires. They can therefore perform a 3D spatial reconstruction from stereoscopic 2D projections (one for each of the three wire planes per TPC).
The key feature of ICARUS is the purity of the liquid argon. Impurities from electro-negative molecules (oxygen water, C02) must be reduced. A continuous purification is performed by recirculation in liquid and gaseous phases. During data taking one accepts a maximum of 17% signal attenuation after 1.5 meters for the longest electron drift length.
Data taking went on in the period October 2010- December 2012. So far with those data they produced three papers on superluminal neutrino searches, but of course they are studying neutrino oscillations and have in press a paper on that topic.
In the present sample they have so far collected 1091 events, and the rate is consistent within 6% with Monte Carlo expectations. The total data corresponds to a total of 0.86E20 POT. The signature of electron neutrinos can be observed visually. A fiducial volume is defined such that the primary vertex is at least 5 cm from each side, at least 50cm from the downstream wall (to allow shower identification). The energy cut is E<30 GeV to optimize signal to noise ratio, since the electron neutrino contamination extends to higher energies.
The visibility cuts for electron neutrino events is based on the observation of a charged track from the vertex which is compatible with being a minimum-ionizing particle (based on the dE/dx measurement), and that the track has good spatial separation from other tracks at the vertex, at least in one of the transverse views.
In the analyzed data they expected from normal beam contaminations 3.0+-0.4 events, plus 1.3+-0.3 from theta_13 oscillations (for sin^2(theta_13)=0.0242+-0.0026), and 0.7 from nu_mu->nu_tau oscillations. So the total is of 5.0+-0.6 events. Backgrounds are expected to amount to 3.7+-0.6 events. As for the signal selection efficiency, they estimate it at 0.74+-0.05, and confirm this measurement with alternative methods.
In the data, two candidates are observed. This observation is presently compatible with the absence of a LSND anomaly. The limits on number of events are 3.41 and 7.13 at 90% and 99% CL respectively. ICARUS should have observed as many as 30 events if the LSND signal were true. The surviving area in the plane of tan^2(theta) vs dm^2 is a tiny dot located at values of (dm^2, sin^2(2theta)) = (0.5 eV^2, 0.005). In that spot there is overall agreement within 90% CL between the limit of ICARUS, the limits of KARMEN, and the positive signals of LSND and MiniBooNE. The relevant graph is shown below.