Università degli Studi di Bologna

1. Università degli Studi di Bologna Neutrino Oscillation Studies with a Massive Magnetized Calorimeter Marco Selvi

2. Neutrino Oscillations Status of the experimental scenario and need for new detectors Magnetized calorimeter performances Atmospheric neutrino physics CNGS beam physics n-factory physics Summary

3. Neutrino Oscillations

4. Neutrino Oscillations : 2 flavors If neutrinos have mass the flavour eigenstatescould not coincide with mass eigenstate: |nm (0)> = cosq|n1> + sinq |n2> |nt (0)> = -sinq|n1> + cosq |n2> After time evolution: |nm(t)> = cosqexp(-iE1t)|n1> + sinq exp(-iE2t)|n2> |nt(t)> = -sinqexp(-iE1t)|n1> + cosq exp(-iE2t)|n2>

5. Neutrino Oscillations : 2 flavors Oscillation Probability:P(nm nt ) = sin2(2q) sin2(1.27 Dm2L/E) Dm2 = m22- m12 Survival Probability: P(nm nm ) = 1 - P(nm nt ) sin2(2q) = 1. Dm2=0.003 eV2

6. Atmospheric neutrinos

7. P + air --> (p, K) Atmospheric neutrinos m + nm e + nm +ne • < h > = 10 km • Fn ~ E-3.7

8. The L/E range the LBL nightmare q p-q L Updated NuMI beam q From Battistoni & Lipari (1998) L (down-going) ~ 10 km L (up-going) ~ 104 km E from 100 MeV up to 100 GeV

9. Status of neutrino studies with Atmospherics

10. Status of atmospheric neutrino data Superkamiokande: 79.3 kty (Y. Totsuka, TAUP2001) • Up/down asymmetry: robust indication of nm disappearance (10s) • (fixes the mixing in a model independent way) • Disappearance occurs near the horizon • + upgoing, througoing, multiring muon-like and NC-like, indication of nt appearance • + MACRO, Soudan 2 nmdeficit increasing with L no anomalyfor ne

11. Interpretation of atmospheric neutrino data • Best fit: • Dm2 = 2.5 x 10-3 eV2 sin22Q = 1 • SK data interpreted as 2n oscillations in the nm - nt channel • (Supported by MACRO, SOUDAN2,CHOOZ) • Pure nm– ns oscillations excluded • Pure nm– ne oscillations excluded • Dinamycs of disappearance fit an L/E law (FCNC, VLI, VEP excluded) • Is pure nm- nt oscillation the end of atmospheric neutrino history?

12. L/E resolution of SuperKamiokande not sufficient to detectoscillations explicitely Limited precision onDm2 There are viable alternative hypotheses with L/E law: Decay Decoherence At least one full oscillation cycle has to be detected to prove oscillations(disprove alternative hypotheses). Explicit detection of oscillation? The oscillation is damped by finite detector L/E resolution !!

13. Damped oscillation sin2(2q) = 1. Dm2=0.003 eV2 Perfect resolution Damped Oscillation Critical damping

14. Physics with a Massive Magnetized Spectrometer on Atmosphericn

15. New detector concepts • Overcome limitations of current atmospheric neutrino detectors: • High L/E resolution • Fully exploit far/near source method fornm disappearance • Systematic-free analysis of the oscillation pattern

16. Atmospheric: comparison up/down It is a good experimental rule that precise measurements are obtained by comparison with a reference • For E >2 GeV • The atmospheric neutrino flux is up/down symmetric at the source • The downward is not affected by oscillations (Dm2 < 10-2 eV2) reference near source • Upward flux is affected by oscillations: L/E goes up to 6·104 km/GeVfar source

17. Measurement of disappearance N up(L/E) N down(L’/E) = P(nm nm; L/E) The disappearance probability can be measured with a single detector and two equal sources: L’ q L(qup) = 2Rcos(qup)L’(qdown) = L(p–qdown) L = 1 - sin2 (2Q) sin2 (1.27 Dm2L/E) An oscillation pattern should appear in the experimental ratio of up to down fluxes (*) q *) method first suggested by P.Picchi and F.Pietropaolo

18. The L/E resolution isdetermined bythe capability of the experiment to reconstruct the neutrino energy and the neutrino direction of flight (L ~2R cosqn): What affects the L/E resolution • But qn is not measured; justqm ... so events near the horizon are of no use: resolution is spoiled by the tan2q term • Low L/E values must be obtained with high E • A detector with a modest hadronic energy resolution, but a good muon momentum measurement can be effectively used provided that low-y events are selected • Limitation of SK: due to the limited acceptance at high energies, oscillations occur near the horizon

19. Magnetized tracking calorimeter Emby range measurement for fully contained events Emby tracking in magnetic field for partly-contained events qmby tracking Up/Downby time of flight (plus vertex identification) high time resolution (< 2 ns) is also required Detector choice

20. The Monolith Detector 14.5 m B B 13 m 29.5 m Large mass34kton Magnetized Fe spectrometer B = 1.3 Tesla Time resolution ~ 1 ns(for up/down discrimination) Space resolution ~1 cm (rms on X-Y coordinates) Momentum resolution sp/p ~ 20% from track curvature for outgoing m ~ 6% from range for stopping m Hadron E resolution sEh/Eh ~ 90%/Eh 30% ~52000 m2 of detector : Glass Spark Counters Fe 2.2 cm Fe 8 cm

21. Event selection Event selection developed to optimise the observation of the oscillation pattern (keep under control the relative L/E resolution) • Em > 1.5 GeV • Fiducial selection of 40 cm on each side • FC events: inside fiducial volume • PC events: one single outgoing track with range > 4 m • Nb. of fired layers > 6 • Selection on combination of the observables Em, qm, Eh to ensure the required L/E resolution

22. L/E resolution in MONOLITH Contributions to L/E resolution • Angular spread contribution of track fit error • Energy measurement • Final L/E resolution

23. Efficiencies and resolutions • Selected nm CC (downgoing only!) after 4 y of data taking: • Fully contained: 931 • Partially contained: 259 • Total: 1190

24. Effect of the Magnetic Field • Higher efficiency in the low L/E region • Higher efficiency in the L/E region of physical interest (102-103) • Slightly higher cost and complexity • (anti-seismic rules for LNGS impose expensive mechanics anyway)

25. Expected L/E distributions (1) Dm2 = 710-4 eV2 99% C.L 90% C.L. 68% C.L. Dm2 = 210-3 eV2 Central value in each bin is obtained with a 26 years statistics. Event rates, error bars and contour lines correspond to 4 years.

26. Expected L/E distributions (2) Dm2 = 510-3 eV2 Dm2 = 810-3 eV2

27. Monolith sensitivity – 4 y • Comparison of MONOLITH sensitivity to oscillations with Kamiokande and SuperKamiokande • 90% C.L. allowed regions after 4 years for differentDm2(left) • Exclusion regions if no effect is found (right)

28. Detection of the oscillation pattern • best fit to oscillation • best fit to decay • best parametric fit Four simulated experiments of 4 years with Dm2 = 0.003 eV2

29. A staged approach 14.5 m B 1 module = 17 kt 13.1 m 15 m Maximum size that fits in Gran Sasso Hall A (between LVD and GNO) 12 kt 13.5 m B 8.5 m 16 m

30. Efficiencies and resolution in a 12 kt module Ln  2REarthcosq En=Em+Eh qn =qm 34 kt 17 kt Resolution comparable to the full detector (34 kt) Efficiency loss < 20% w.r.t. the full detector (fiducial cut against cosmic muon background)

31. A 12 kt detector (4 years) 90% C.L. allowed regions Kamiokande RMS Precision on sin22Q RMS Precision on Dm2 SK 90% C.L. region SK 0.007 eV2 0.003 eV2 0.001 eV2 10kt Efficiency for decay model rejection at 95% C.L.

32. 12 kt detector

33. 34 kt detector

34. 100 kt detector 3.0 10-3 1-3% precision in the oscillation parameters is achievable

35. Vertical vs horizontal layers for atmospheric neutrinos (FAQ) • Lower reconstruction efficiency along the vertical direction with vertical plates • About the same efficiency at small L/E (where the 1st minimum is expected): • Events near the horizon filtered by resolution requirements! • Need for an external VETO Selected atm. n’s events for fixed L/E resolution Pay on mixing, but marginally on Dm2

36. Physics with the CNGS beam

37. CNGS beam • nm from p, K • < E > ~ 20 GeV • L = 732 km • Optimized for tau appearence • Rate CC ~ 2600/kt y

38. Detector layout n n

39. CNGS event example n n

40. Efficiencies and resolutions Almost flat around 50% for E>10 GeV

41. L/E Range AtmosphericsThe L/E distribution, resulting after selections, is populated up to 5·103km/GeV High sensitivity to Dm2 values down to a few 10-4 eV2 The Log(l/E) distribution is more populated at high L/E The sensitivity of the experiment decreases for increasing values ofDm2 • Can the beam help at highDm2 ? • atmospheric  no systematic • beam  systematic to be understood • L/E distributions after selections • 4 y atmospheric (shaded) • 1 y CNGS beam

42. CNGS beam will cover with very high statistics the region L/E < 100 km/GeV: ~ 40,000 events/year nm CC after selections vs. ~ 200 events/year from up-going atmospheric. Monolith on CNGS beam • Systematic effects: a tough job! 10% bin per bin systematics assumed Accordingly with BMPT .

43. Impact of CNGS beam Dm2=0.007 eV2 Atmo’s alone Atmo’s + Beam

44. CNGS beam: CC/NC ratio Atm. full MONOLITH (CC/NC)obs /(CC/NC)no-osc CC/NC 90% allowed regions (includes uncertainties of beam shape and composition, detector effects, …) Dm2 = 0.003 eV2 MONOLITH 12ktx5y Visible hadron energy (GeV) CC/NC ratio can supplement atmospheric data constraints on sterile neutrinos

45. Physics at the n-factory with a Monolith-like Detector

46. Neutrino Oscillations : 3 flavors If neutrinos have mass the flavour eigenstates could not coincide with mass eigenstate: 3 mixing angles: q12, q23, q13 2 mass differences: Dm212 Dm223 1 CP-violation fase: d

47. One-mass scale dominance << at terrestial distances Pne-nm= sin2(q23)sin2(2q13) sin2(1.27 Dm223 L/E) q13bounded by CHOOZ exp. to be small sin2(2q13) < 0.1 (90% C.L.) Very high intensity beam needed

48. n-factory • Features: • High intensity • Well-known beam • Both flavors • Different helicity

49. Physics at a n-Factory • Circulating 50 GeVm+ in a NuFactory(1021 decays in 5 years) • Beam made by nm and ne (m+e+ ne nm) • Search at LBL for wrong sign muons (m-) coming from • ne oscillated into nm q13, sign of Dm2, study of matter effects, CP violation

50. Golden channel: wrong sign muons