Neutrino Oscillations and and Lorentz Violation Results from MiniBooNE Outline: z - LSND - signal for ν oscillations y x - sidereal analysis and LV - Tandem Model - MiniBooNE - experiment, analysis, ν results - LV results R. Tayloe, Indiana University CPT 2007 R. Tayloe, Indiana University CPT '07 1
ν e events vs energy The LSND Result The LSND experiment observed an excess of ν e event s in beam of ν µ 87.9 ± 22.4 ± 6.0 (4 σ ) consistent with ν µ →ν e oscillations. However, this result, with large ∆ m 2 ,does not fit in a 3 generation neutrino model osc parameter likelihood regions (given results from other oscillation experiments) since ∆ m 12 2 + ∆ m 13 2 + ∆ m 23 2 = 0 If LSND is correct ⇒ new physics. - additional (sterile) neutrinos - a different model for oscillations R. Tayloe, Indiana University CPT '07 2
Review: Sidereal variation in the LSND signal - In AK, MM, PRD 70 , 076002, a short-baseline approximation for neutrino oscillations (allowing for sidereal variation) was developed. - In PRD 72 , 076004 we (with LSND collaboration) reported the results of a search for sidereal variation in the LSND signal... all are f(a L , c L and ν beam direction in sun-centered frame) R. Tayloe, Indiana University CPT '07 3
Sidereal variation in the LSND signal - LSND sidereal variation, results: consistent with no sidereal variation... 5-param 3-param 1-param (flat) bkgd R. Tayloe, Indiana University CPT '07 4
Sidereal variation in the LSND signal - LSND sidereal variation, results: extraction of SME parameter combinations. Log-likelihood (1s) contours from 3-parameter fit - allowed regions include sidereal variations (non-zero A s , A c ) - extracted parameter square-sum: - (noted by AK,MM before this analysis) - regardless of sidereal variation, if the SME is used to explain LSND then, a L or E x c L ~10 -19 GeV (~ expected Planck-scale effects) R. Tayloe, Indiana University CPT '07 5
A “global model” of ν oscillations (with the SME) - The biggest challenge in constructing a global model of ν oscillations within the SME is the E-dependence. SK-atmospheric and KAMLAND report an L/E dependence... How to model with with E 0 and E 1 terms? - AK, MM noted that the mixed energy dependence in the coeffs can lead to a LV “see-saw” mechanism that occurs in certain energy ranges (“pseudomass”) - the “bicycle-model” h bicycle ab 0 c E a a a 0 0 a 0 R. Tayloe, Indiana University CPT '07 6
global oscillation probabilities The “tandem model” - T. Katori, V. A. Kosteleck`y, R. Tayloe, Phys.Rev.D74:105009,2006. - start with bicycle model - add additional m 2 term which generates a 2 nd seesaw... - 3 parameters, rotationally invariant - explain solar, atmospheric, KamLAND, LSND - only 3 parameters (remember, standard 3 ν has 4-6) - no MSW needed for solar - prediction for MiniBooNE (among others) R. Tayloe, Indiana University CPT '07 7
atmos. ν /anti- ν oscillations oscillation probabilities solar neutrino oscillations short-baseline ν /anti- ν oscillations long-baseline anti- ν oscillations R. Tayloe, Indiana University CPT '07 8
MiniBooNE experimental strategy - Test the LSND observation via ν µ → ν e appearance. - Keep L/E same, change beam, energy, and systematic errors P (ν µ →ν e ) = sin 2 2θ sin 2 (1.27∆ m 2 L /Ε) neutrino energy (E): baseline (L): MiniBooNE: ~500 MeV MiniBooNE: ~500 m LSND: ~30 MeV LSND: ~30 m target and horn decay region absorber dirt detector ν µ → ν e ??? K + π + Booster primary beam secondary beam tertiary beam (protons) (mesons) (neutrinos) R. Tayloe, Indiana University CPT '07 9
MiniBooNE Collaboration R. Tayloe, Indiana University CPT '07 10
MiniBooNE beam: total ν flux - mean energy ~800MeV MB ν flux - ν e / ν µ = 0.5% π → µ ν µ K → µ ν µ µ → e ν µ ν e K → π e ν e R. Tayloe, Indiana University CPT '07 11
ν Events in MiniBooNE Background - Recall: search for ν e in a ν µ beam - signature of a ν e reaction (signal): electron - need to distinguish from backgrounds Signal (due to ν µ reactions) that consist of a muon or π 0 - ν interaction products create (directed, prompt) Cerenkov light and (isotropic, delayed) scintillation light Background - pattern and timing of the detected light allows for event identification (and position, direction, energy meas.) R. Tayloe, Indiana University CPT '07 12
ν interactions in detector: -predicted ν events and fractions predicted # ν events in data set (no efficiency corrections) from event generator* CC quasielastic 340k - extensively tuned using MiniBooNE data NC elastic 150k CC π + 180k CC π 0 30k NC π 0 48k NC π +/- 27k CC/NC DIS, multi- π 35k all channels 810k ν osc. events ~1k µ, e ν "CC": charged- W current X N ν ν "NC": neutral- Z *NUANCE ( D. Casper, NPS, 112 (2002) 161) current X N R. Tayloe, Indiana University CPT '07 13
oscillation analysis: strategy signal reaction: e n e − p - need accurate, efficient particle identification algorithm to separate (signal) electron-like events from ubiquitous ν e e − (background) muon, pion events W p - To avoid experimenter bias, this was done with “blind” n procedure, signal data set kept in “box” until algorithms set. background: n − p Two algorithms were used: - “track-based” (TB) ν µ µ − Uses direct reconstruction of particle types and likelihood ratios for particle-ID W - “boosted decision trees” (BDT) p n Set of low-level variables combined with background: BDT algorithm -> PID “score” 0 , 0 p ,n p ,n ν µ ν µ - In the end, the TB analysis had slightly better sensitivity, so is used for primary results. π 0 Z BDT analysis is a powerful “double-check” p,n ∆ p,n R. Tayloe, Indiana University CPT '07 14
oscillation analysis: backgrounds intrinsic- ν e backgrounds (from ν e produced at ν source) - µ → ν e : (indirectly) measured in ν µ CCQE events via π -decay chain - π → ν e : “ “ “ “ “ “ “ - K → ν e : measured in high-energy ν µ ,ν e CCQE (from Kaons), extrapolate to low-E “mis-ID” backgrounds (mainly from ν µ ) - CC Inclusive: includes CCQE, measured, simulated - NC π 0 : measured, simulated TB analysis predicted backgrounds - NC ∆ →N γ : constrained in data, simulated - NC coherent, radiative γ : calculated, negligible - Dirt: ν interactions outside tank, simulated, measured - beam-unrelated events, measured, very small correlated errors on all backgrounds are considered R. Tayloe, Indiana University CPT '07 15
oscillation analysis: box-opening With... - algorithms finalized, - cuts determined, - backgrounds predicted, - the neutrino oscillation box was opened on March 26, 2007 R. Tayloe, Indiana University CPT '07 16
oscillation analysis: results track-based analysis: - E ν > 475MeV cut for oscillation analysis region - no sign of an excess in the analysis region - visible excess at low E null - 2 best =0.94 ● 2 No evidence for ν µ → ν e appearance in the analysis region R. Tayloe, Indiana University CPT '07 17
oscillation analysis: results track-based analysis: Counting Experiment: 475<E ν <1250 MeV data: 380 events expectation: 358 ± 19 (stat) ± 35 (sys) significance: 0.55 σ No evidence for ν µ → ν e appearance in the analysis region R. Tayloe, Indiana University CPT '07 18
oscillation analysis: results Limit curves: solid: TB, primary result dashed: BDT - MiniBooNE and LSND incompatible at a 98% CL for all ∆ m 2 under a 2 ν mixing hypothesis R. Tayloe, Indiana University CPT '07 19
oscillation results: low-energy region Track-based analysis E ν distributions: For: 300<E ν <475 MeV 96 ± 17 ± 20 events Excess: 3.7 σ background subtracted data: The energy-dependence of excess is not consistent with ν µ →ν e appearance assuming standard energy dependence Best Fit (sin 2 2 θ , ∆ m 2 ) = (1.0, 0.03 eV 2 ) P (ν µ →ν e ) = sin 2 2θ sin 2 (1.27∆ m 2 L /Ε) R. Tayloe, Indiana University CPT '07 20
Recommend
More recommend
