Transcript MINOS Results From The 1 Year in the NuMI Beam

MINOS Results
From The 1st Year in the NuMI
Beam
Patricia Vahle
University College, London
on behalf of the
MINOS Collaboration
Overview
●
●
●
●
2
Introduction to NuMI and MINOS
–
Physics Goals
–
NuMI beam and MINOS detectors
Experiment operation
–
Data collection
–
Event reconstruction and selection
–
Near and far detector distributions
Oscillation Analysis
–
Prediction of the Far Detector spectrum (no oscillations)
–
Oscillation fits, parameter extraction
Future
P. Vahle, NuFact 2006
The MINOS Experiment
●
●
●
●
●
3
Main Injector Neutrino
Oscillation Search
Far detector
Muon neutrino beam
produced by 120 GeV/c Main
Injector at Fermilab
Two functionally identical
detectors, separated by
735km
Neutrino beam
Near Detector (Fermilab)
measures beam before
oscillations
Far Detector (Soudan, MN)
measures distortions w.r.t.
the Near Detector
Near detector
P. Vahle, NuFact 2006
MINOS Collaboration
32 institutions, 175 scientists
Funding: DOE, NSF, PPARC
Argonne • Athens • Benedictine • Brookhaven • Caltech • Cambridge • Campinas • Fermilab
College de France • Harvard • IIT • Indiana • ITEP-Moscow • Lebedev • Livermore
Minnesota-Twin Cities • Minnesota-Duluth • Oxford • Pittsburgh • Protvino • Rutherford
Sao Paulo • South Carolina • Stanford • Sussex • Texas A&M
Texas-Austin • Tufts • UCL • Western Washington • William & Mary • Wisconsin
4
P. Vahle, NuFact 2006
MINOS Physics Goals
• Test the νμ → ντ oscillation
hypothesis
– Measure precisely |Δm232| and
sin22θ23
 e

 

 
  U e1
 
  U  1
 U
  1
U2
U 2
U e 3   1 
 
U  3  2 
 
U 3 
 3 
ν3
• Search for sub-dominant
ν2
νμ → νe oscillations
ν1
• Search for or constrain
Useful Approximations:
exotic phenomena
– Sterile ν, ν decay
U e2
Δm232 = m32 –
m2 2
νμ Disappearance (2 flavors):
• Compare ν, ν oscillations P(νμ→νx) = 1 - sin22θ23sin2(1.27Δm232L/E)
– Test of CPT violation
• Atmospheric neutrino
oscillations
– Phys. Rev. D73, 072002 (2006)
5
νe Appearance:
P(νμ→νe) ≈ sin2θ23 sin22θ13sin2(1.27Δm231L/E)
Where L, E are experimentally optimized and
θ23, θ13, Δm232 are to be determined
P. Vahle, NuFact 2006
Oscillation Measurement
●
Long baseline νμ disappearance experiment
●
Predict unoscillated CC spectrum at Far Detector
●
Compare with measured spectrum to extract oscillation parameters
P(    )  1  sin 2 2 sin 2 (1.267m2 L / E)
νμ spectrum
spectrum ratio
Unoscillated
Oscillated
Monte Carlo
6
Characteristic
Shape
Monte Carlo
( Input parameters: sin22θ = 1.0, Δm2 = 3.35x10-3 eV2 )
P. Vahle, NuFact 2006
Oscillation Measurement
●
Long baseline νμ disappearance experiment
●
Predict unoscillated CC spectrum at Far Detector
●
Compare with measured spectrum to extract oscillation parameters
P(    )  1  sin 2 2 sin 2 (1.267m2 L / E)
νμ spectrum
spectrum ratio
Unoscillated
Oscillated
Monte Carlo
7
sin2(2θ)
Monte Carlo
( Input parameters: sin22θ = 1.0, Δm2 = 3.35x10-3 eV2 )
P. Vahle, NuFact 2006
Oscillation Measurement
●
Long baseline νμ disappearance experiment
●
Predict unoscillated CC spectrum at Far Detector
●
Compare with measured spectrum to extract oscillation parameters
P(    )  1  sin 2 2 sin 2 (1.267m2 L / E)
νμ spectrum
spectrum ratio
Unoscillated
Oscillated
Monte Carlo
8
Δm2
Monte Carlo
( Input parameters: sin22θ = 1.0, Δm2 = 3.35x10-3 eV2 )
P. Vahle, NuFact 2006
The NuMI facility
• Neutrinos at the Main Injector
•120 GeV/c protons from the Main Injector
•Performance
•~10 μs spill
•2.2 second cycle time (2.0 sec)
•2.3x1013 Protons/Spill (3.0 x 1013 peak)
•0.17 MW (0.29 MW peak)
•Variable target position → variable beam energy!
9
P. Vahle, NuFact 2006
The NuMI neutrino beam
νμ= 92.9%
νμ= 5.8%
νe+νe = 1.3%
Expected number of events (no osc.)
in Far Detector
●
●
10
Majority of data in the LE-10
configuration
Took ~1.5e18 pot in ME and HE for
commissioning and systematics
studies
Beam
Target z
position (cm)
FD Events per
1e20 pot
LE-10
-10
390
ME
-100
970
HE
-250
1340
Changing Horn Currents gives
even more beam options
P. Vahle, NuFact 2006
Detector Technology
Scintillator
2.54 cm Fe
Steel
Extruded
PS scint.
4.1 x 1 cm
5.9 cm
U V planes
+/- 450
WLS fiber
Clear
Fiber cables
•Two functionally equivalent detectors
•Tracking sampling Calorimeters
•2.54 cm thick Steel Absorber
•1 cm thick active plastic scintillator
•Segmentation
•5.94 cm longitudinal
•4.1 cm transverse
•Alternating planes rotated +/- 90 deg
•WLS collects/routes light to PMTs
•Magnetized detectors <B>=1.3 T
M64 PMT
M16 PMT
Scintillator
Multi-anode PMT
11
P. Vahle, NuFact 2006
The Detectors
The Near Detector
The Far Detector
PMT+FEE Racks
Veto Shield
Scint. Modules
Beam
Me
Coil Hole
Beam
PMT Boxes
•Measures beam before oscillations
•Predicts Far Spectrum
•1 kton
•1 km from target
•103 m underground
•3.8 x 4.8 x 15 m3
•High rates—fast electronics
12
• Measures beam after oscillations
• 5.4 kton
• 735 km from target
• 705 m underground
• 8 x 8 x 30 m3
• Low rate environment
• Taking data since 2001 (completed in
2003)
P. Vahle, NuFact 2006
Near Detector Spill
One near detector spill
• High rate in Near
detector results in
multiple neutrino
interactions per MI spill
• Events are separated by
topology and timing
(19ns resolution)
7.1 m
beam direction
13
P. Vahle, NuFact 2006
Near Detector rates
•High event rates in the Near detector
• ~8 events / spill
• ~35 million events
(~2.5 million in fiducial volume for 1.27 x 1020 pot)
• Image detector with Neutrinos!
Reconstructed ν event vertices
!
14
Reconstruction stability w.r.t
intensity
P. Vahle, NuFact 2006
νμ-CC event selection
1.
2.
Event contains at least one good, reconstructed track
Reconstructed vertex within the fiducial volume:
–
–
NEAR: 1m < z < 5m, R< 1m from beam center (4.5% of total).
FAR: z>50cm, z>2m from rear face, R< 3.7m from center of detector
(72.9% of total).
NEAR DETECTOR
FAR DETECTOR
μCalorimeter
3.
4.
5.
15
Spectrometer
Fitted tracks with negative charge (selects )
Additional cuts in FD to remove events polluted by light injection,
steep cosmic tracks
Cut on likelihood-based Particle ID parameter which is used to
separate CC and NC events.
P. Vahle, NuFact 2006
CC/NC classification
NC event
νμ CC event
3.5m
• Low level shape variables used
•Event length
•Track PH per plane
•Fraction of event PH in track
• Purity 98%
• Ave. Efficiency 74% (FD) 67% (ND)
(for fid. vol. events)
• NC contamination limited to lowest
energy bins (below 1.5 GeV)
16
1.8m
Event Classification Parameter
P. Vahle, NuFact 2006
Energy Spectra in ND
LE 10
17
ME
HE
P. Vahle, NuFact 2006
Energy Spectra in ND
LE 10
ME
HE
Discrepancy between Data/MC changes energy with
different beam tunes suggests production of Hadrons off
the target is to blame.
18
P. Vahle, NuFact 2006
Energy Spectra in ND
LE 10
19
ME
HE
•Flux prediction based on
•Atherton—400 GeV/c p-Be
•Barton—100 GeV/c p-C
•SPY—450 GeV/c p-Be
•Data extrapolated to MINOS beam energy, target composition, target thickness
P. Vahle, NuFact 2006
Energy Spectra in ND
LE 10
20
ME
HE
• Parameterize predicted flux as function of pT and pz
• Perform fit to reweight neutrino hadron parents
• Horn focusing, beam alignment, x-section, NC BG, energy scale included as
nuisance parameters
• Remaining data/MC discrepancies at ~5-10% level across 6 beam configurations
P. Vahle, NuFact 2006
Prediction of FD spectrum
Far Spectrum without oscillations is similar, but
not identical to the Near spectrum!
p Target
π+
FD
Horn 1 Neck
Edge of Decay Pipe
Horn 2 Neck
Decay Pipe
Eν ~ 0.43Eπ / (1+γπ2θν2)
• Difference in Decay angle to ND
vs. FD changes Energy spectrum
• MINOS must take the differences
into account when predicting the
FD unoscillated spectrum from
the measured ND spectrum
21
Far/Near Ratio
ND
π Lifetime
NuMI Beam MC
Energy (GeV)
P. Vahle, NuFact 2006
Predicting the FD Spectrum
●
●
22
Several procedures for predicting the FD
spectrum:
•
“Far/Near” & “Beam Matrix”—directly extrapolate ND data
spectrum using best knowledge of beam kinematics, use MC
to correct for efficiency, purity, resolution.
•
“ND-Fit” & “2D-Grid”—Describe ND distributions by fitting
physics quantities, predict FD spectrum from best fit (e.g., by
reweighting MC)
All yield compatible results, at 1.27 x 1020 POT
exposure, for all sources of systematic error we
have studied
P. Vahle, NuFact 2006
Beam Matrix: Near→Far
X
=
• Simplest method—use ratio of F/N from MC, multiply by Near
Data
• Next level of sophistication—2-D Beam Matrix generated using
beam MC, relates neutrino energy in Near to energy in Far
• Both encapsulate the knowledge of pion 2-body decay
kinematics & beam geometry.
23
P. Vahle, NuFact 2006
FD Prediction From
All Methods
1.27×1020 POT
MINOS
All methods agree
to within ~ 5% binby-bin
Acceptable
considering
current exposure
24
Size of
statistical
error
P. Vahle, NuFact 2006
Selecting FD beam events
●
Time stamping of the neutrino events is provided by
two GPS units (located at Near and Far detector sites).
–
●
FD Spill Trigger reads out 100μs of activity around beam spills
Far detector neutrino events have very distinctive
topology and are easily separated from cosmic muons
(0.5 Hz)
●
●
●
●
25
In 2.6 million “fake” triggers, 0
events survived the selection
cuts
No accepted events outside of
expected spill duration
Upper limit on cosmic events,
0.5 events
Upper limit on rock
interactions, 0.4 events
P. Vahle, NuFact 2006
Observed Number of Events
Data Sample
26
FD
Expected
Data/MC
Data
(Matrix Method;
Unoscillated)
(Matrix Method)
νμ (<30 GeV)
215
336.0±14.4
0.64±0.05
νμ (<10 GeV)
122
238.7±10.7
0.51±0.05
νμ (<5 GeV)
76
168.4±8.8
0.45±0.06
●
An energy dependent deficit
●
Below 10 GeV a 49% deficit is observed
●
Significance is 6.2σ (stat+syst)
P. Vahle, NuFact 2006
Far Detector Distributions
●
●
●
Predicted no oscillations (solid)
Best fit (dashed)
Clear deficit of CC like events
Track Vertex r2 (m2)
MINOS
1.27×1020 POT
Event Classification Parameter
MINOS
1.27×1020 POT
y = Eshw/(Eshw+Pμ)
MINOS
1.27×1020
POT
27
P. Vahle, NuFact 2006
Systematic Errors
Shift in Δm2
Shift in
(10-3 eV2)
sin22θ
Near/Far normalization ±4%
0.050
0.005
Absolute hadronic energy scale ± 11%
0.060
0.048
NC contamination ± 50%
0.090
0.050
All other systematic uncertainties
0.044
0.011
Total systematic (summed in quadrature)
0.13
0.07
Statistical error (data)
0.36
0.12
Preliminary Uncertainty
• Systematic shifts in the fitted parameters are computed using MC
“fake data” samples
• Magnitude of systematic error is ~40% of statistical error for Δm2
• Several systematic uncertainties are data driven, expected to
improve with more data and study
• Three largest included in fit as nuisance parameters
28
P. Vahle, NuFact 2006
Fit to Oscillation
Hypothesis
m
2
32
 2.74
0.44
 0.26
(stat  syst)  10 eV
sin 2 2 23  1.00 0.13 (stat  syst)
Normalizat ion  0.98
29
3
Measurement errors are 1σ, 1 DOF
Fit constrained to sin2(2θ)≤1
2
 
2
nsys
nbins
 2e  o   2o ln o e    s
i 1
i
i
i
i
i
j1
2
j
 s2
j
P. Vahle, NuFact 2006
Allowed Region
• Fit includes penalty
terms for three main
systematic
uncertainties
• Fit is constrained to
physical region:
sin2(2θ23)≤1
2
3
2
m 32
 2.74 00..44
26  10 eV
sin 2 2 23  1.00 0.13
30
P. Vahle, NuFact 2006
Allowed Region
• Fit includes penalty
terms for three main
systematic
uncertainties
• Fit is constrained to
physical region:
sin2(2θ23)≤1
2
3
2
m 32
 2.74 00..44
26  10 eV
sin 2 2 23  1.00 0.13
31
P. Vahle, NuFact 2006
Projected Sensitivity
32
P. Vahle, NuFact 2006
Summary
●
MINOS has completed an analysis of the first year of NuMI beam data
–
Exposure used in analysis: 1.27 x1020 POT
–
Exclude no oscillations at 6.2σ (based only on event rate)
–
Results are consistent with the oscillation hypothesis with parameters:
2
3
2
m 32
 2.74 00..44

10
eV
26
sin 2 2 23  1.00 0.13
–
Constraining the fit to sin2(2θ23) = 1 yields:
2
m 32
 2.74  0.28 10 3 eV 2
●
Systematic uncertainties under control
–
●
●
33
Significant improvements expected with data driven studies & more statistics
Paper submitted to PRL
– Available online at: hep-ex 0607088
Second year of running is underway!
P. Vahle, NuFact 2006
Calibration
•LED based light injection system tracks
channel gain over time
•Cosmic rays intercalibrate strips
•Stopping muons intercalibrate detectors
•Dedicated calibration module in CERN test
beam for absolute shower energy scale
PMTs
1m
Beam
Optical Cables
Calibration Detector
34
P. Vahle, NuFact 2006
Event topologies
νμ CC event
NC event
νe CC event
UZ
VZ
3.5m
• long μ track+ hadronic
activity at vertex
1.8m
• short event, often
diffuse
Eν = Eshower+Pμ
2.3m
•short, with typical
EM shower profile
Muon Energy Resolution
6% range, 10% curvature
Shower Energy Resolution: 55%/√E
35
P. Vahle, NuFact 2006
Near detector distributions
X Vertex
—Data
• Detector acceptance well
modeled in MC
• Beam points down 3°—points to
Soudan
—MC
Track Angle
w.r.t
vertical
Z Vertex
36
P. Vahle, NuFact 2006
Hadron Production Tuning
Model
GFLUKA
Sanf.-Wang
CKP
Malensek
MARS – v.14
MARS – v.15
Fluka 2001
Fluka 2005
Fluka2005 Tuned
pT (GeV/c)
0.37
0.42
0.44
0.50
0.38
0.39
0.43
0.364
0.355
• Weights ~20% in region of pT
vs pz that produces MINOS
neutrinos
• Hadron production tuning
changes mean pT less than
model spread
Region of LE10 Beam
37
P. Vahle, NuFact 2006