1. NuMI MINOS Neutrinos Change Flavor Under Wisconsin Current NuMI/MINOS Oscillation Results Alec Habig, for the MINOS Collaboration XLIII Rencontres de Moriond EW 2008 27 institutions 147 physicists Argonne Arkansas Tech Athens Benedictine Brookhaven Caltech Cambridge Campinas Fermilab Harvard IIT Indiana Minnesota-Twin Cities Minnesota-Duluth Oxford Pittsburgh Rutherford Sao Paulo South Carolina Stanford Sussex Texas A&M Texas-Austin Tufts UCL Warsaw William & Mary

2. NuMI MINOS Neutrinos Change Flavor Under Wisconsin Current NuMI/MINOS Oscillation Results Alec Habig, for the MINOS Collaboration XLIII Rencontres de Moriond EW 2008 27 institutions 147 physicists Argonne Arkansas Tech Athens Benedictine Brookhaven Caltech Cambridge Campinas Fermilab Harvard IIT Indiana Minnesota-Twin Cities Minnesota-Duluth Oxford Pittsburgh Rutherford Sao Paulo South Carolina Stanford Sussex Texas A&M Texas-Austin Tufts UCL Warsaw William & Mary

3. NuMI MINOS MINOS Main Injector Neutrino Oscillation Search Investigate atmospheric sector  oscillations using intense, well-understood NuMI beam Two similar magnetized iron- scintillator calorimeters –Near Detector 980 tons, 1 km from target, 90 m deep –Far Detector 5400 tons, 735 km away, 700 m deep 735 km

4. NuMI MINOS Physics Goals Precise (~10%) measurement of  m 2 23 Confirm  ↔  flavor oscillations –Provide high statistics discrimination against alternatives such as sterile, decoherence, decay, etc Search for subdominant  ↔ e oscillations –a shot at measuring  13 Directly compare vs oscillations (a test of CPT) –MINOS is first large underground detector with a magnetic field for  + /  - tagging Study interactions and cross sections using the very high statistics Near Detector data set Cosmic Ray Physics This talkPublishedComing later this year Atmospherics Beam

5. NuMI MINOS  Disappearance Sensitivity Measure  flux at Near Det, see what’s left at Far Det Simulated results plotted as F/N ratio –Position of dip gives  m 2 –Depth of dip gives sin 2 2  Spectral ratio shapes would differ in alternative models Unoscillated Oscillated  spectrum Monte Carlo Spectrum ratio

6. NuMI MINOS Far Detector M16 PMT 16 mm A module of 20 strips …on a plane 8 fibers on a pixel 486 planes, 5400 tons total –Each 1” steel + 1 cm plastic scintillator thick –8 m diameter with torodial ~1.5 T B-field –31 m long total, in two 15 m sections –192 scintillator strips across Alternating planes orthogonal for stereo readout –Scint. CR veto shield on top/sides Light extracted from scint. strips by wavelength shifting optical fiber –Both strip ends read out with Hamamatsu M16 PMTs –8x multiplexed

7. NuMI MINOS 3D Reconstruction Take all the “U” view lit-up strips –Cross with all the “V” view lit-up strips –X marks the spot(s) See live events at http://www.soudan.umn.edu

8. NuMI MINOS 3D Reconstruction This is a real  interaction from the beam –   appears inside detector, –cruises along through many planes, –curving in the magnetic field, Curvature tells us momentum… –stops. …so does range See live events at http://www.soudan.umn.edu

9. NuMI MINOS Near Detector 282 planes, 980 tons total –Same 1” steel,1 cm plastic scintillator planar construction, B-field –3.8x4.5 m, some planes partially instrumented, some fully, some steel only –16.6 m long total Light extracted from scint. strips by wavelength shifting optical fiber –One strip ended read out with Hamamatsu M64 PMTs, fast QIE electronics –No multiplexing upstream, 4x multiplexed in spectrometer region 3.8 m 4.8 m

10. NuMI MINOS Hadronic: 56±2%/  E EM: 21±4%/  E 60-plane ‘micro-MINOS’ –has taken data at T7 & T11 test beam lines at CERN during 2001, 2002, 2003 Instrumented with both Near and Far Detector electronics –To provide cross-calibrations –Energy uncertainties: 3% relative, and 1.9% (ND) & 3.5% (FD) absolute Caldet

11. NuMI MINOS NuMI Spectrum Three standard beam configurations –Obtained by changing target and horn 2 locations –Target easy to move, used to approximate ME, HE with “pME”, “pHE” beams –Horn 2 harder to move, has not yet been adjusted –Can also adjust horn currents for more variations At L of 735 km and  m 2 indicated by Super-K, the “Low” energy beam is closest to the first oscillation minima –Unfortunately, also the least intense (still 10 7 per 10 20 pot at the Near Detector!) Beam Target z position (cm) FD Events per 1e20 pot LE-10-10390 pME-100970 pHE-2501340

12. NuMI MINOS Beam Data Analyzed up to 3x10 13 protons every 2.2 seconds (~250 kW) (10 4 events/day in ND!) RUN I - 1.27x10 20 POT (published in PRL) Higher energy beam RUN IIa 1.23x10 20 POT (NEW) RUN IIb ~0.75x10 20 POT (in can) This talk: Run I + Run IIa - 2.497x10 20 POT RUN III >0.63x10 20 POT (current) Far Det >99% live!

13. NuMI MINOS Near Detector Data How do data look in the Near Detector, where we have ~unlimited statistics? If we understand things there, we can then look at the Far Detector data where the physics is happening, so: –Examine ND closely –Compare ND data/MC –“Blind” analysis done ? ? ?

14. NuMI MINOS Lots of in the Near Detector A mean of 3 interactions per spill (in 8 or 10  s), up to 10 Near Detector Electronics gates for 19  s during the entire spill –Digitizes continuously every 19 ns, no dead time Separate events using timing and topology Below: ~35 x 10 6 events for 1.27 x 10 20 POT image the ND’s internal structure with ! A typical 6-event spill, colored by time

15. NuMI MINOS   Z  N (+X)   W N X e e W N X What sort of Interaction?  CC Event NC Event e CC Event Long  track + hadronic activity at vertex 3.5m Short event, often diffuse 1.8m Monte Carlo E = E shower +E  55%/  E 5% range, 10% curvature (hadronic) 22%/  E (leptonic) E = E shower Short, with typical EM shower profile 2.3m 7 variable log-likelihood PID formed, improves over PRL analysis

16. NuMI MINOS Reconstructed Beam Spectrum LE-10pME pHE LE-10 events Weights applied as a function of hadronic x F and p T. MIPP data on MINOS target will be used to refine this in the future, NA49 and Harp results also used Discrepancies between data and Fluka05 Beam MC vary with beam setting: so source is due to beam modeling uncertainties rather than cross-section uncertainties MC tuned by fitting to hadronic x F and p T over 7 beam configurations (3 shown here)

17. NuMI MINOS What is Expected in Soudan? Measure Near Detector E spectrum To first order the beam spectra at Soudan is the same as at Fermilab, but: –Small but systematic differences between Near and Far –Use Monte Carlo to correct for energy smearing and acceptance –Use our knowledge of pion decay kinematics and the geometry of our beamline to predict the FD energy spectrum from the measured ND spectrum ff to far Detector Decay Pipe   (soft) (stiff) nn target ND

18. NuMI MINOS On to the Far Detector… “Blind” analysis –Only after understanding the Near Detector, reconstruction, selected non-oscillation Far Detector parameters, and early pHE (ie, non-oscillating) beam data did we “open the box” –Data “re-blinded” when developing analysis improvements and adding new data

19. NuMI MINOS Spectrum |  m 2 32 | = 2.38 +0.20 (stat + syst) x 10 -3 eV 2 sin 2 2  23 = 1.00 -0.08 (stat + syst)  2 /ndf=41.2/34 ( 18 bins x 2 spectra (Run I, Run IIa) – 2) − 0.16 Measurement errors are 1σ, 1 DOF

20. NuMI MINOS Allowed Region Fit includes systematic penalty terms Fit is constrained to physical region: sin 2 (2  23 )≤1 Consistent with previous experiments Already best measurement of |  m 2 32 |

21. NuMI MINOS MC MINOS MC The Future  disappearance –more data, looser cuts as systematics are better understood –add rock muons Anti-neutrino oscillations –in neutrino beam –anti-neutrino running? Search for/rule out exotics –Sterile neutrinos –Neutrino decay –Neutrino de-coherence e appearance –New sensitivities! Simulated data based on  m 2 =2.7x10 -3 eV 2 and sin 2 2  =1.0 (Old values! For Comparison Purposes Only,  m 2 really does = 2.38x10 -3 eV 2 these days, but this plot has not been remade)

22. NuMI MINOS Progress Towards the Future This year should see the first looks at: –A NC analysis to directly test for participation of s in the  disappearance –A e analysis to search for  13 via e appearance As with the  disappearance analysis, the Near Detector data and backgrounds must be well understood before the Far Detector data (and its potential new physics) is examined Here follow two important steps towards this goal:

23. NuMI MINOS Selecting Neutral Current Events Simple NC event selection: –< 60 planes; –no track or –no track beyond 5 planes from shower

24. NuMI MINOS NC Spectrum NC events can be used to search for sterile neutrino component in FD –via disappearance of NC events at FD –FD data coming soon Real ND NC Data Projected s Limits when FD box is opened

25. NuMI MINOS e Appearance Sensitivity Working towards a  13 measurement by using Near Detector data-driven methods to estimate e appearance backgrounds –At Near Detector, all e events are background rather than due to e appearance. –Apply the NN-based e selection to the ND data to get an all- background sample. –Find what fraction of those background events are NC showers, mis-ID’d CC events, or real e from the beam Use these new background estimates to correct Far Detector sensitivity projections for unknowns in hadronic shower modeling etc. Two independent methods agree Estimated systematic errors based on these Near Detector studies are 10% –Far Detector sideband studies are next

26. NuMI MINOS Muon Removal “MRCC” sample: –shower only events obtained by removing the muon from CC selected events and reconstructing the shower remnant. Comes from a well understood spectra, with well known efficiency and purity. Apply the e selection to these non- e events –Results in a background sample of electromagnetic dominated hadronic showers.

27. NuMI MINOS Muon Removal >20% Data/MC discrepancy in both the standard e and the muon removed CC samples Comparisons of standard Data and MC shower topological distributions disagree in the same way as does MRCC data with MRCC MC –So MC hadronic shower production/modeling is a major contribution to the disagreement. –Kinematic phase space of MRCC and selected NC events matches well, but MRCC and selected CC events do not. The MRCC sample is thus used to make and ad-hoc correction to the model to NC events per bin –Beam e from MC, CC events are the remainder

28. NuMI MINOS Horn on/off Method After applying e selection cuts to Near Detector data, the composition of the selected events is quite different with the NuMI focusing horns on or off. Get horn on/off ratios from MC, then solve for NC and CC backgrounds in bins of energy, get beam e from the beam MC (a well understood number) –Independent of hadronic modeling N on = N NC + N CC + N e (1) N off = r NC *N NC + r CC *N CC + r e *N e (2) from MC: r NC(CC,e) = N NC(CC,e) off /N NC(CC,e) on

29. NuMI MINOS Horn on/off Method r CC r NC Solve eqs. (1), (2) On/off ratios match well between data and MC (after fiducial volume cuts). Apply these fractions to the total data to deconvolute it into its components. The MC predicts 24% too much CC and 28% too much NC background! Also checked using 3 rd beam spectrum, pHE Agrees with Muon Removal method!

30. NuMI MINOS Projected  13 Limits Extrapolating these data- driven backgrounds to the Far Detector, e appearance limit projections are calculated using the horn on/off method Shown as a function of CP- violating phase , since over the MINOS allowed  m 2 range the limits vary little This plot is for our current exposure, with estimated 10% systematics Blind analysis in progress

31. NuMI MINOS Projected  13 Limits Projected limits shown for expected MINOS exposures Data-driven systematics are hoped to drop to 5% with more work plus more horn off data in future years Inverted hierarchy shown only for lowest exposure for simplicity, but mostly marches along behind the normal limits

32. NuMI MINOS The first 2.5  10 20 POT of NuMI beam data have been analyzed for   disappearance oscillations and are consistent with standard neutrino oscillations with the following parameters: A Neutral Current data spectrum has been measured at the Near Detector. When extrapolated to the Far Detector and compared to data, direct limits on sterile neutrino participation will be set Studies of Near Detector data are used to establish data-drive backgrounds to e appearance at the Far Detector. The current MINOS exposure will be sensitive to  13 at the Chooz limit Summary This work was supported by the U.S. Department of Energy, the U..K. Science and Technology Facilities Council, and the State and University of Minnesota. We gratefully acknowledge the Minnesota Department of Natural Resources for allowing us to use the facilities of the Soudan Underground Mine State Park. This researcher was directly supported by NSF RUI grant #0653016.

33. NuMI MINOS Extra Slides

34. NuMI MINOS New Improved Analysis Several improvements over previous analysis –better reconstruction –improved event separation –improved shower modelling –new intra-nuclear modelling CC event identification based on –track properties –event length –event kinematics Derive PID based on PDFs of the above

35. NuMI MINOS Improvements Larger fiducial volume(3%) improved reconstruction (4%) improved PID –higher efficiency –higher purity

36. NuMI MINOS Unconstrained Fit

37. NuMI MINOS Parameter Spaces

38. NuMI MINOS What Changed?

39. NuMI MINOS What Changed? Improvements: reco & selection shower modelling

40. NuMI MINOS What Changed? Data sets: Pre-shutdown Post-shutdown Improvements: reco & selection shower modelling

41. NuMI MINOS What Changed? Data sets: Pre-shutdown Post-shutdown Improvements: reco & selection shower modelling