1 / 23

n m and n e Physics in MINOS

n m and n e Physics in MINOS. Alex Himmel, Pedro Ochoa. Antineutrinos Overview Oscillations Systematics n e Analysis Nearest neighbors selection Background estimations Summary. 1. 0.5. with SK parameters. 0. E (GeV). 0. 15. 30. Antineutrinos in MINOS.

lola
Télécharger la présentation

n m and n e Physics in MINOS

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. nm and ne Physics in MINOS Alex Himmel, Pedro Ochoa • Antineutrinos • Overview • Oscillations • Systematics • ne Analysis • Nearest neighbors selection • Background estimations • Summary

  2. 1 0.5 with SK parameters 0 E (GeV) 0 15 30 Antineutrinos in MINOS • Approx. 6% of our beam is made of muon antineutrinos. Amplified spectrum MC 1x1020 POT Difficulty: not many events in osc. peak region Difficulty: not many events in osc. peak region • Unique advantage: both MINOS detectors are magnetized. Allows us toseparate neutrinos and antineutrinos on an event-by-event basis.

  3. Antineutrino physics • Very interesting physics can be done with antineutrinos: 1)noscillation analysis: A large CPT-violating region still unexplored 90%, 95%, 99% and 3σ CPT violating regions still allowed by global fit (except LSND) M.C. Gonzalez-Garcia, M. Maltoni and T. Schwetz (hep-ph/0306226) 2)n→n transitions: have never been looked for beforein atmos sector. • Some models beyond the SM predict them (i.e. Langacker and Wang, Phys. Rev. D 58 093004). • Could fully explain the atmospheric neutrino results (Alexeyev and Volkova, hep ex/0504282) • If 10% or more of neutrinos that disappear transition to antineutrinos then we will see them. 3) Measurement ofBeam ne’s: important for ne analysis • Very strong involvement of Caltech group in these areas.

  4. Antineutrino oscillations • MINOS could distinguish between Dm223 and Dm223 at 90% C.L. if Dm223 > 0.004 eV2 in ~1 more year, during normal “neutrino” running • But if CPT is conserved, the reach of antineutrino oscillation analysis would be relatively small: 90% Tentative exclusion limit only valid for high mixing angle Preliminary MC 90%95%99% 3σ ~3 first years of data (6x1020 POT) (no systematics included)

  5. Antineutrino running • These difficulties can be overcome with a small amount of reversed horn current running (RHC). • In this case negative particles from the target are focused thus yielding an antineutrino beam: 1x1020 POT 1x1020 POT Forward horn current (FHC) Reversed horn current (RHC) Peak reduction due primarily to cross-section difference (sn < sn)

  6. 90% Tentative exclusion limit 90%95%99% 3σ Antineutrino running • Combining FHC with RHC can obtain ameasurement of Dm223 that rivals the first MINOS measurement of Dm223: 6x1020 POT (FHC) + 1x1020 POT (RHC) (no systematics) 90% C.L. 68.3% C.L. Preliminary MC Only ~4 months of antineutrino running (plus ~1 more year of normal running) are needed ! • This data would considerably reduce our best current limits on neutrino CPT • Effort led by Caltech

  7. Antineutrino systematics • Systematic errors are a crucial question in combining FHC and RHC data. • ~30% of antineutrinos produced outside of the target region • While neutrinos are also produced outside the target, they are negligible compared to those focused from the target. • A large fraction of the difference between the near and far detectors comes from decay pipe antineutrinos.

  8. Antineutrino systematics • Uncertainties in the decay pipe modeling could affect the far/near ratio creating a false signal. • Toy systematic model: • 50% Scaling of the decay pipe component • The other components of the flux unchanged • Preliminary results suggest that the effect is small compared to the expected statistical error at 1x1021 POT.

  9. Beam systematics Old Monte Carlo New Monte Carlo Flugg Geant 4 Geometry Geant 3 Geometry Flugg Fluka Geometry Geant-Fluka Physics Geant 4 Physics Fluka Physics • Working to update the beam Monte Carlo from Geant3 to Geant4. • Use Flugg to run the new geometry in Fluka, a more trusted physics simulation

  10. Muon Scattering • How well does Geant4 model multiple scattering?  Compared Geant4 and some IHEP data (1986) of muons on a Cu target. • Study shows Geant4 greatly overestimates the data, especially at lower muon momenta.

  11. Muon Chopper • Another technique for assessing systematics associated with charge separation was developed at Caltech: 1) Identify stopping muons at the ND: 2) Remove everything but the last x GeV’s of energy: 3) Run reconstruction over muon chopped data and MC 4) Calculate ID efficiency & purity in data & MC for different values of x.

  12. < 10% systematic in n purity ! Muon Chopper • The following sources of systematics are addressed by the Muon Chopper: Magnetic field Multiple scattering Reconstruction / Backgrounds • Preliminary results indicate charge separation is reasonably well modeled by the MC:

  13. ne Appearance • At MINOS’ baseline of 735 km, • Expect ~14 ne CC events (E<10 GeV) appearing in the MINOS Far Detector for every 1x1020 POT of data if q13 is at CHOOZ limit • Main challenge at MINOS consists in distinguishing between EM and hadronic showers. • At Caltech concentrating on: • developing the best possible ne selection • measuring two of the main backgrounds.

  14. Nearest Neighbors Selection • For analysis need to have as good neselection as possible to maximize signal. • Most available selections use multivariate techniques that rely on reconstructed quantities. • But this analysis is a special case: Number of reco variables ~ number of strips in event • Why not perform event ID using strip information alone? • Have been working on a nearest neighbors selection in collaboration with Cambridge University. • Basic idea: • Compare each input event to large libraries of simulated ne CC and NC events. • Select N best matches • Construct discriminant from N best matches information (e.g. fracCC=fraction of N best matches which are neCC)

  15. Nearest Neighbors Selection • Advantages: • Approach is in principle optimal. No loss of info from raw→reconstructed quantities • Largely reconstruction-free. • But only optimal if fully sample phase space • Need large libraries (~50-100 Million events of each type). • So far have generated ~50 Million events at Caltech. • Determine how well two events match by asking: Strip # Strip # “what is probability the two events come from same hit pattern at PMTs?” plane # plane # Strip # Strip # Poisson plane # plane #

  16. Nearest Neighbors Selection • Example of discriminant: fracCC(y<0.5)=fraction of 20 best matches that were ne with y<0.5 Library size: ~1M ne ~1.5M NC NC CC ne • Already provides the best significance ! • Information of N best matches is very rich: • Plenty of room for further improvement ! Good separation • Currently working hard to get selection fully operational in the Near Detector: data mc nc νe νμcc • Using different background estimating techniques to understand data-MC discrepancy.

  17. sin2(2q13) = 0.1, |m31|2 =2.710-3eV2, sin2(2q23) = 1, POT=4x1020 FD Performance • Selected events: ~end of 2007 • Sensitivity limited by statistical fluctuations of background. Define figure of merit FOM=Signal/√Background Note: preselection included Preliminary MC Library size: 5M ne 10M NC cut • Our selection already has a FOM at least 15% higher than all the other selection methods. Selected events: For 0 < Ereco < 6 GeV: FOM=2.29 2 methods for addressing background have been developed at Caltech. antineutrinos muon removal

  18. NC Background • Use Muon Removal (MR) to assess the NC Background: • Apply muon removal (MR) to both data and MC • Apply ne selection on both. • Use differences in both samples to reweight the NC expectation in the ne analysis. # of ne candidates in MR data # of NC events in ne analysis ND data before MR after MR (NN selected events) # of ne candidates that are NC in MC # of ne candidates in MR MC • MR reweighting removed the ~60% overall normalization discrepancy

  19. Beam ne’s from antineutrinos • Irreducible background in ne analysis: intrinsic beam ne‘s Nearly all come from m+→ e+ + ne + nm • Need to tag antineutrinos coming from m+ decay. Use fact that antineutrino spectrum is practically the same independently of the beam configuration: Most antineutrino parents just go through the center of both horns pseudo-high energy (pHE) pseudo-medium energy (pME) Low energy (LE) MC MC MC • Work led by Caltech, in collaboration with BNL

  20. n from m+ nfromm+ Beam ne’s from antineutrinos • Only m+ component changes significantly when running in pME or pHE ! The Technique: (pME-LE)TRUE at 1e18 POT • Scale pME (or pHE) and LE data to same POT and take the difference • Fit with using shapes from the MC: LE Corrections due to differences in the antineutrinos from p- and K- pME • Expected sensitivity: pHE data already taken!

  21. Summary & Ongoing Work • Antineutrinos: • Only ~4 months of antineutrino running are needed to make a measurement of Dm223 with a precision that rivals the first MINOS Dm223 result. • Will search for nm→nm transitions for the first time. • Developing tools for beamline simulation. • ne appearance:  The CHOOZ limit will be reached by end of 2007 Expect 1st MINOS neappearance result by next year. • Very positive outlook. Working hard to: • Further improve selection • Assess systematics. • Critical role played by Caltech group in these two areas.

  22. Backup

  23. 3.5m 1.8m 2.3m short event, often diffuse short, with typical EM shower profile neAppearance in MINOS • Challenge: At MINOS, we lack the granularity to fully resolve hadronic vs. EM showers: steel thickness: 2.54cm| strip width: 4.12cm (Molière radius ~3.7cm) νμ CC Event νeCC Event NC Event ne e- nm m- nm nm W W Z n p p n p p (MC) long μ track & hadronic activity at vertex

More Related