Sterile neutrinos Michele Maltoni Instituto de F´ ısica Te´ orica UAM/CSIC What is ν ? Invisibles 2012 and Alexei Smirnov Fest GGI, Firenze, Italy – June 28th, 2012 I. The LSND experiment and four-neutrino models II. MiniBooNE and models with two sterile neutrinos III. A word on MiniBooNE data after Neutrino 2012 Summary
I. The LSND experiment and four-neutrino models 2 The LSND problem 10 2 • LSND observed ¯ ν e appearance in a ¯ ν µ beam ( E ν ∼ ∆ m 2 (eV 2 /c 4 ) 30 MeV, L ≃ 35 m); 10 • Karmen did not confirm the claim, but couldn’t fully Karmen CCFR exclude it either; Bugey 1 • the signal is compatible with ¯ ν µ → ¯ ν e oscillations NOMAD provided that ∆ m 2 � 0 . 1 eV 2 ; -1 10 • on the other hand, other data give (at 3 σ ): 90% (L max -L < 2.3) 99% (L max -L < 4.6) ≃ 7 . 5 ± 0 . 6 × 10 − 5 eV 2 , ∆ m 2 -2 10 � ≃ 2 . 4 ± 0 . 3 × 10 − 3 eV 2 ; � � -3 -2 -1 � ∆ m 2 10 10 10 1 � � sin 2 2 θ • in order to explain LSND with mass-induced neutrino oscillations one needs at least one more neutrino mass eigenstate; • WARNING: having enough ∆ m 2 is not enough. To make sure that the model works, one has to check explicitly that all the experiments can be fitted simultaneously. Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 3 Four neutrino mass models • Approximation: ∆ m 2 ≪ ∆ m 2 ≪ ∆ m 2 ⇒ 6 different mass schemes: sol sol atm atm atm sol SBL SBL SBL SBL SBL SBL sol atm atm atm sol sol (a) (b) (c) (d) (A) (B) (3+1) (2+2) • Total: 3 ∆ m 2 , 6 angles, 3 phases. Different set of experimental data partially decouple : ∆ m 2 ∆ m 2 ∆ m 2 ϕ 13 SOL ATM LSND SOL η s ATM d µ NEV LSND ϕ 34 η e ϕ 12 θ SOL θ ATM θ LSND Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 4 (2+2): ruled out by solar and atmospheric data 40 D D N N A A L L solar m solar (pre SNO salt) m global 2 χ 30 a a PG K K + + solar solar 2 χ Real 2 20 PC ∆χ atm Restricted + atm atm LBL 10 3 σ 3 σ 3 σ + + LBL + LBL SBL 0 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 η s η s = d s d s • in (2+2) models, fractions of ν s in solar ( η s ) and atmos ( 1 − d s ) add to one ⇒ η s = d s ; • 3 σ allowed regions η s ≤ 0 . 31 (solar) and d s ≥ 0 . 63 (atmos) do not overlap; superposition occurs only above 4 . 5 σ ( χ 2 = 19 . 9 ); • the χ 2 increase from the combination of solar and atmos data is χ 2 = 28 . 6 (1 dof), corresponding to a PG = 9 × 10 − 8 [1]. [1] M. Maltoni, T. Schwetz, M.A. Tortola, J.W.F . Valle, Nucl. Phys. B643 (2002) 321 [ hep-ph/0207157 ]. Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 5 (3+1): tension between LSND and short-baseline data • In (3+1) schemes the SBL appearance probability is effectively 2 ν oscillations: 1 10 P µ e = sin 2 2 θ sin 2 ∆ m 2 41 L sin 2 2 θ = 4 | U e 4 | 2 | U µ 4 | 2 ; , ★ 2 ] 4 E 41 [eV LSND DAR (90%, 99%) • disappearance experiments bound | U e 4 | 2 and | U µ 4 | 2 ; 0 2 10 ∆ m 2 10 99% CL 9 0 % ★ -1 10 C atm + LBL + NEV L 1 10 CDHS -4 -3 -2 -1 0 10 10 10 10 10 2 ] 41 [eV 2 2 θ = 4 |U e4 | 2 |U µ 4 | 2 sin ★ Bugey 0 10 2 ∆ m • LSND is in conflict [1]: -1 10 atmospheric − with other appearance experi- Chooz ments (Karmen & Nomad); -2 10 -3 -2 -1 -3 -2 -1 10 10 10 10 10 10 2 2 − with all disappearance exp’s. |U e4 | |U µ 4 | [1] M. Maltoni, T. Schwetz, M.A. Tortola, J.W.F . Valle, Nucl. Phys. B643 (2002) 321 [ hep-ph/0207157 ]. Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 6 The MiniBooNE experiment ( ≤ 5/2012) [MB- ¯ ν e ] • E ν and L very different from LSND (but similar L / E ν ) ⇒ can check the oscillation solution of the LSND problem, not the signal itself; • very peculiar results: − strong low-energy excess in ν e , mild in ¯ ν e ; − mild mid-energy excess in ¯ ν e , but not in ν e . 3 Data Events / MeV Predicted Spectrum ν Data 80 e from [MB- ν e ] ν µ 2.5 Uncertainty in Prediction e Events/(100 MeV) + ν from K ± Neutrinos from K ’s e 0 from K ν 0 Neutrinos from K ’s 60 2 e L 0 misid π Neutrinos from ’s µ ∆ → N γ Neutrinos from π ’s dirt 1.5 40 other Total Background 1 20 [NuMI- ν e ] 0.5 0 0.2 0.6 1 1.4 1.8 2.2 2.6 3 0.2 0.4 0.6 0.8 1 1.2 1.4 1.5 1.6 3. Reconstructed E [GeV] QE E (GeV) ν ν Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 7 LSND vs MiniBooNE in (3+1) [ ¯ ν e ] • ν e : no signal ⇒ excludes LSND; ν e : signal ⇒ mildly confirms LSND. • ¯ 2 10 [ ν e ] 2 sin (2 θ ) upper limit MiniBooNE 90% C.L. y MiniBooNE 90% C.L. sensitivity 10 BDT analysis 90% C.L. ) 4 /c 2 | (eV 1 2 m ∆ | -1 10 LSND 90% C.L. LSND 99% C.L. -2 10 -3 -2 -1 10 10 10 1 2 sin (2 ) θ Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
I. The LSND experiment and four-neutrino models 8 2 10 Status of (3+1) models after MiniBooNE (3+1) [2] 10 • (3+1) four-neutrino schemes fail because: ) 2 (eV 1 − can’t reconcile appearance and disappearance data; 41 2 m ∆ -1 10 − can’t explain the different ν e (MB) and ¯ ν e (LSND) results; null SBL 90% CL null SBL 99% CL − can’t account for the low-energy ν e event excess in MB. LSND + BNB-MB( ν ) + BNB-MB( ν ) 90% CL -2 10 LSND + BNB-MB( ) + BNB-MB( ) 99% CL ν ν -4 -3 -2 -1 10 10 10 10 1 ⇒ (3+1) models are ruled out as explanation of SBL data. 2 sin (2 ) θ µ e BNB-MB( ) BNB-MB( ) LSND ν ν [OLD] 2 2 2 10 10 10 [2] 90% CL 90% CL 90% CL 99% CL 99% CL 99% CL 10 10 10 ) ) ) 2 2 2 (eV (eV (eV 41 41 41 2 2 2 m 1 m 1 m 1 ∆ ∆ ∆ -1 -1 -1 10 10 10 -3 -2 -1 -3 -2 -1 -3 -2 -1 10 10 10 10 10 10 10 10 10 2 2 2 sin (2 ) sin (2 ) sin (2 ) θ θ θ µ e µ e µ e [2] G. Karagiorgi et al. , Phys. Rev. D80 (2009) 073001 [ arXiv:0906.1997 ]. Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
II. MiniBooNE and models with two sterile neutrinos 9 3 The MiniBooNE excess Events / MeV Data ν from µ 2.5 e + from K ν e With the analysis cuts set, a signal-blind test of data- 0 from K ν 2 e MC agreement in the signal region was performed. The π 0 misid full two-neutrino oscillation fit was done in the range N ∆ → γ 300 < E QE < 3000 MeV and, with no information on dirt 1.5 ν the fit parameters revealed, the sum of predicted back- other ground and simulated best-fit signal was compared to Total Background 1 data in several variables, returning only the χ 2 . While agreement was good in most of the comparisons, the E vis spectrum had a χ 2 probability of only 1%. This triggered 0.5 further investigation of the backgrounds, focusing on the lowest energies where ν µ -induced backgrounds, some of which are difficult to model, are large. As part of this 0.2 0.4 0.6 0.8 1 1.2 1.4 1.5 1.6 3. study, one more piece of information from the signal re- QE E (GeV) gion was released: unsigned bin-by-bin fractional discrep- ν ancies in the E vis spectrum. While ambiguous, these re- inforced suspicions about the low-energy region. Though • MiniBooNE observed a 3 . 0 σ excess at low-E [3]; we found no specific problems with the background es- timates, it was found that raising the minimum E QE of ν • this excess is incompatible with 2 ν oscillations; the fit region to 475 MeV greatly reduced a number of backgrounds with little impact on the fit’s sensitivity to • therefore, data with E QE < 475 MeV have not oscillations. We thus performed our oscillation fits in the ν energy range 475 < E QE < 3000 MeV and opened the ν been used to check LSND. full data set. ⇒ Omission of low-energy bins in based on the hypothesis of two-flavor oscillations ! • Is it possible to do something about these data in more sophisticated models? [3] A.A. Aguilar-Arevalo et al. [MiniBooNE collab], Phys. Rev. Lett. 98 (2007) 231801 [arXiv:0704.1500]. Michele Maltoni <michele.maltoni@csic.es> M - “W ν ?”, 28/06/2012
Recommend
More recommend
