für Kernphysik, Heidelberg Max-Planck-Institut The GSI Anomaly
How large is the LSND anomaly? - University of Chicagoelagin/HARP-CDP_vs_LSND/Elagin...How large is...
Transcript of How large is the LSND anomaly? - University of Chicagoelagin/HARP-CDP_vs_LSND/Elagin...How large is...
-
How large is the "LSND anomaly"?Andrey Elagin
on behalf of the HARP-CDP group
HEP lunch, UChicago, March 12, 2012
-
Outline
● 3.8 σ LSND Final Paper PRD 64, 112007 ● 3.8 σ → 2.9 σ HARP-CDP Paper I arXiv:1110.4265
(accepted to PRD)● 2.9 σ → 2.3 σ HARP-CDP Paper II arXiv:1112.0907
(accepted to PRD)
● LSND corrections: arXiv:1112.2181● HARP-CDP reply: arXiv:1112.3852
-
LSND● Beam dump experiment in Los-Alamos (1993-1998).
● Claimed 3.8 evidence of νµ→ν
e oscillations with ∆m2 ~ 1 eV2. In
stark conflict with three light neutrino flavours (solar and atmospheric oscillations and LEP N = 2.9840±0.0082).
● At least one 'sterile' neutrino is required: 800+ theoretical papers (700+ after 1998)
-
LSND● 1993-1994 data: 16.4 (+9.7 - 8.9) ± 3.3
(alternative analysis by J.E. Hill do not find any excess PRL 75, 2654)
● 1993-1995 data: 51.0 (+20.2 – 19.5)
● Full dataset: 87.9± 22.4± 6.
∆m2 > 0.02 eV2.
● BNL-E776, CCFR, NuTeV and NOMAD
exclude ∆m2 > 10 eV2.
● Bugey and CHOOZ ruled out
∆m2 < 0.2 eV2.
● KARMEN2 ∆m2 < 1 eV2 or ∆m2 ~ 7 eV2.
-
Test by MiniBoone
Neutrino mode: exclude ν
µ→ν
e
+ unexplained excess
Antineutrino mode: does not rule out 'LSND anomaly'
-
LSND neutrino source
-
LSND neutrino source
π+DAR
νµ µ+ ν
µ ν
e e+
π-DIF
νµ µ− ν
µ ν
e e−
p
800 MeV or 1.5 GeV/c
-
LSND neutrino source
π+DAR
νµ µ+ ν
µ ν
e e+
π-DIF
νµ µ− ν
µ ν
e e−
DAR of µ− competes with µ− + (A,Z) → νµ + (Α, Ζ−1)
p
800 MeV or 1.5 GeV/c
-
LSND neutrino source
π+DAR
νµ µ+ ν
µ ν
e e+
π-DIF
νµ µ− ν
µ ν
e e−
DAR of µ− competes with µ− + (A,Z) → νµ + (Α, Ζ−1)
νe flux prediction depends on:
1) Total number of pi-2) Momentum spectra3) Angular spectra4) Geometrical layout and material composition
p
800 MeV or 1.5 GeV/c
-
The HARP experiment
-
The HARP experiment
-
HARP-CDP cross sections
-
HARP-CDP: tuning hadron modelsGeant4 and FLUKA
-
HARP-CDP: tuning hadron modelsGeant4
-
HARP-CDP vs LSND parametrization
-
HARP-CDP simulation
-
Hadroproduction
-
Hadroproduction
-
HARP-CDP simulation
Very complicated taskNeed differential pion production cross-sections
● of p, n, pi+, pi-● on H
2O, Fe, Cu, Al, Mo, Air
● as a function of projectile momentum
-
Pions from different models
-
Pion generations
-
Pion production by 600 MeV neutrons
-
LSND Background I (geniune νe)
● LSND used pions from 1st generation only● LSND ignored pion production by neutrons● LSND used MCNP simulation program only for
geometry description
-
LSND published
DAR νµ
[10-9 / PoT / cm2]
0.8
DAR νe
[10-12 / PoT / cm2]
0.65
LSND Background I (geniune νe)
-
LSND published
LSND „emulation“
DAR νµ
[10-9 / PoT / cm2]
0.8 0.60
DAR νe
[10-12 / PoT / cm2]
0.65 0.59
LSND Background I (geniune νe)
-
LSND published
LSND „emulation“
Geant4 + Exp. data
FLUKA + Exp. data
DAR νµ
[10-9 / PoT / cm2]
0.8 0.60 0.78 0.76
DAR νe
[10-12 / PoT / cm2]
0.65 0.59 0.96 0.88
LSND Background I (geniune νe)
-
LSND Background I (geniune νe)
LSND published HARP-CDP estimate
19.5 ± 3.9 30.6 ± 8.8
LSND detector see this: ν
e + p → e+ + n
„electron“ + delayed 2.2 MeV gamma from neutron capture
-
LSND Background II (fake νe)
νµ + p → µ+ + n
νµ + 12C → µ− + 12N
νµ + 12C → µ+ + 12B
Only π DIF can produce neutrinos above threshold
If Tµ ~< 3 MeV, muons are not identified andthis reaction fakes signal
νe + p → e+ + n
We consider LSND uncertainty too optimistic
-
LSND Background II (fake νe)
We estimate 35% uncertainty on νµ and 60% on ν
µ
(15% by LSND)
-
First reduction of the "anomaly"
-
LSND analysis strategy2100 „electron“ candidates
-
Correlated γ vs uncorrelated γ
-
Beam excess
LSND finds „beam excess“ („beam on“ minus „beam off“) consists of 117.9+/-22.4 correlated events out of 2100 candidates
Uncertainty of 22.4 is consistent with vanishingly small systematics, if any
-
Something is missing...
νe + 12C → e- + 12N
gs followed by 12N
gs → 12C + e+ + ν
e
2.7 events
-
Comparison of R distributions
When fit hypotheses comprise only correlated or accidental gammas 2.3 12N
gs events will be interpreted as correlated
-
"Beam excess" calculation● Generate pseudodata with 120 correlated
events and 1980 accidental events to study systematic uncertainty
● By varying "base distributions" and fitting R distribution we estimate 17.3% of systematic uncertainty
● 2.3 events subtracted from 117.9 LSND yielding115.6 events with total uncertainty of 27.9
-
Final reduction of the "anomaly"
-
Summary● The claim of a 3.8 σ significance of the LSND anomaly
cannot be upheld● LSND didn't take into account pion production by
neutrons● Improved simulation of the LSND beam stop shows that
conventional background increases by a factor of 1.6● Positrons from 12N
gs beta decay were missed in LSND
analysis● We find significance of the "LSND anomaly" not large
than 2.3 σ
-
The HARP-CDP group
-
Back-up
-
Slide 1Slide 2Slide 3Slide 4Slide 5Slide 6Slide 7Slide 8Slide 9Slide 10Slide 11Slide 12Slide 13Slide 14Slide 15Slide 16Slide 17Slide 18Slide 19Slide 20Slide 21Slide 22Slide 23Slide 24Slide 25Slide 26Slide 27Slide 28Slide 29Slide 30Slide 31Slide 32Slide 33Slide 34Slide 35Slide 36Slide 37Slide 38Slide 39Slide 40Slide 41Slide 42Slide 43Slide 44Slide 45Slide 46Slide 47Slide 48Slide 49Slide 50