οδηγιες Φυσικη χημεια βιολογια Α ΓΕΛ Α&Β εσπ. ΓΕΛσπ γελ 2012 2013
Mazurov ferrara2014
-
Upload
alexander-mazurov -
Category
Education
-
view
50 -
download
0
description
Transcript of Mazurov ferrara2014
High Level Trigger software performance profiling andχb production study at the LHCb experiment
Alexander MazurovTutor: Concezio Bozzi
University of Ferrara (Italy) & CERN
7 March 2014
1/46
Plan
1 LHCb experiment2 Measuring Υ(NS) (N=1, 2, 3) fraction originating from χb decays andχb(3P ) mass:
MotivationDatasetsDetermination of Υ yieldsDetermination of χb yields in the following decays:
χb(1, 2, 3P )→ Υ (1S)γχb(2, 3P )→ Υ (2S)γχb(3P )→ Υ (3S)γ
Monte-Carlo efficienciesSystematic uncertaintiesResults
3 Summary4 Backup: HLT profiler
2/46
The LHC Accelerator System
The main LHC design parameters:Circumference 27 km
Center-of-mass energy 14 TeVInjection energy 450 GeV
Field at 2 × 450 GeV 0.535 TField at 2 × 7 TeV 8 T
Helium temperature 2KLuminosity 1034 cm−2s−1
Bunch spacing 25 nsLuminosity lifetime 10 hrTime between 2 fills 7 hr
The LHC operated mainly at center-of-mass energies of 7 TeV and 8 TeV in2011 and 2012, respectively. In 2015, after the long shutdown, LHC is plannedto reach energy of 13 TeV.
3/46
The LHCb experiment (1)
Forward spectrometer with planardetectors:
LHCb uniqueness:
Tracking, RICH and calorimeterscover the full detector acceptance(2.0 < η < 5.0); tracking coveragealso in the backward region(−4.0 < η < −1.5)Covers just 4% of the solid anglebut captures 25% of heavy quarkpairs produced at the LHC.Ability to study low-pT processesat large η
Heavy quark pair production at the LHC:
Fraction of bb̄ pairs in the acceptance:√s ATLAS/CMS LHCb
7 TeV 44% 25%14 TeV 41% 24%
4/46
The LHCb experiment (2)
Excellent tracking performance
High quality particle identification: robust hadron ID +photon/lepton/hadron separation
Selective and flexible trigger system
5/46
The LHCb experiment (3)
LHCb 2010–2012 operation:
Integrated luminosities:0.038 fb−1 in 2010
1.107 fb−1 in 2011
2.082 fb−1 in 2012
Visible interactions:
2× 1014 visible pp interactions
6× 1012 visible cc̄ interactions
3× 1011 visible bb̄ interactions∼93% data taking efficiency∼99% of accumulated data are useful for physics analysesLuminosity leveling: constant and moderate interaction rate throughoutthe data taking periodsSmooth data taking in 2011–2012 regardless high luminosity running
6/46
Study of χb production
7/46
Motivation
Bound bb̄ states, which can be produced in different spin configurations, are anideal laboratory for QCD tests. It’s like a hydrogen atom in QCD.
Measured mass
Mass from theory
States with parallel quark spins (S=1):S-wave Υ state.P-wave χb states, composed by 3 spin statesχb(0,1,2).Υ can be readily produced in the radiative decaysof χb.χb(3P ) state recently observed by ATLAS, D0 andLHCb.
This thesis:1 Measurement of Υ(NS) (N=1, 2, 3) fraction
originating from χb decays as function of pT (Υ).2 Measurement of χb(3P ) mass.
8/46
Previous analysis at LHCb
”Measurement of the fraction of Υ(1S) originating from χb(1P ) in ppcollisions at
√s =7 TeV ”, arXiv:1209.0282,
∫L = 32 pb−1
”Observation of the χb(3P ) state at LHCb in pp collisions at√s =7 TeV
”, LHCb-CONF-2012-020,∫L = 0.9 fb−1.
9/46
In this thesis
The results in this study extend the statistical precision of previous LHCbmeasurements and add considerably more decays and higher transversemomentum regions. The measurement of Υ (3S) fraction in radiative χb(3P )decay was performed for the first time.
In each pT (Υ) bin calculate:
σ(pp→χb(mP )X)×Br(χb(mP )→Υ(nS)γ)σ(pp→Υ(nS)X) =
Nχb(mP )→Υ(nS)γ
NΥ(nS)× εΥ(nS)
εχb(mP )→Υ(nS)γ=
Nχb(mP )→Υ(nS)γ
NΥ(nS)× 1
εrecoγ
Calculate for each Υ(nS), n = 1, 2, 3 and χb(mP ),m = 1, 2, 3
Get N from fits: NΥ from m(µ+µ−) and Nχb→Υγ from[m(µ+µ−γ)−m(µ+µ−)] (for better resolution)
Compute efficiency εrecoγ from Monte-Carlo simulation
10/46
Analysis Plan
1 Datasets2 Determination of Υ yields3 Determination of χb yields in the following decays:
χb(1, 2, 3P )→ Υ (1S)γχb(2, 3P )→ Υ (2S)γχb(3P )→ Υ (3S)γ
4 Monte-Carlo efficiencies5 Systematic uncertainties6 Results
11/46
Datasets
Full 2011 dataset at√s =7 TeV.
∫L = 1 fb−1
Full 2012 dataset at√s =8 TeV.
∫L = 2 fb−1
Monte-Carlo simulation of χb inclusive decays, generated 62× 106
events.
12/46
The Υ selection
Description Requirement
Υ rapidity1 2.0 < yΥ < 4.5Track fit quality χ2/ndf < 4
Track pT > 1 GeV/cPrimary vertex probability > 0.5%
Luminous region |zPV | < 0.5m and x2PV + y2
PV < 100mm2
Kullback-Leibler distance > 5000
Muon and hadron hypotheses ∆ logLµ−h > 0Muon probability ProbNN > 0.5
Trigger lines:L0 L0DiMuon
HLT1 Hlt1DiMuonHighMassHLT2 HLT2DiMuonB
1y = 12
ln(E+pLE−pL
), where pL is the component of the momentum along the beam axis13/46
The Υ fit model
9 10 110
5000
10000
15000
20000
25000
30000
Can
dida
tes/
(40
MeV/c
2)
mµ+µ−[
GeV/c2]
√s = 7 TeV
6 < pµ+µ−
T < 40 GeV/c
Υ (1S)
Υ (2S)
Υ (3S)
3 Double Crystal Ball functionsfor signal yields.
Exponential function forcombinatorial background.
µ+µ− transverse momentum intervals, GeV/c6 – 40
√s = 7 TeV
√s = 8 TeV
NΥ (1S) 283,300 ± 600 659,600 ± 900NΥ (2S) 87,500 ± 400 203,300 ± 600NΥ (3S) 50,420 ± 290 115,300 ± 400
Background 296,400 ± 700 721,300 ± 1100
µΥ (1S), MeV/c2 9457.02 ± 0.10 9455.58 ± 0.07σΥ (1S), MeV/c2 42.86 ± 0.10 43.04 ± 0.06
µΥ (2S), MeV/c2 10,019.03 ± 0.21 10,018.05 ± 0.14σΥ (2S), MeV/c2 46.38 ± 0.20 46.45 ± 0.14
µΥ (3S), MeV/c2 10,351.16 ± 0.32 10,349.41 ± 0.16σΥ (3S), MeV/c2 48.63 ± 0.31 48.24 ± 0.11τ -0.3887 ± 0.0023 -0.3819 ± 0.0015
Υ (1S) mass is about 5 MeV/c2
lower than PDG value9460.30± 0.26 MeV/c2
14/46
Υ yields as function of pT
0 10 20 30 40 50
210
310
410
510
610
0 10 20 30 40 50
210
310
410
510
610
0 10 20 30 40 50
210
310
410
510
610
Eve
nts
pT (Υ) [ GeV/c]
Υ (3S)
Eve
nts
pT (Υ) [ GeV/c]
Υ (1S)
Eve
nts
pT (Υ) [ GeV/c]
Υ (2S)
√s =7 TeV√s =8 TeV
√s =7 TeV√s =8 TeV
√s =7 TeV√s =8 TeV
Yields normalized by bin width and luminosity.
The small difference between 7 and 8 TeV data is due to the productioncross-sections, which are expected to be about 10% larger.
15/46
χb selection
In this study photons reconstructed using the calorimeter information.Another approach uses photon conversions in e+e− pairs — this methodhas better invariant mass resolution, but requires more statistics.
Cuts on γ:
Transverse momentum of γ pT (γ) > 600 MeV/cPolar angle of γ in the µ+µ−γ rest frame cos θγ > 0Confidence level of γ cl(γ) > 0.01
Dimuon mass windows:Decay Cut Description
χb(1, 2, 3P )→ Υ(1S) 9310 < µ+µ− < 9600 MeV/c 3σΥ (1S) < µ+µ− < 2.5σΥ (1S) MeV/c
χb(2, 3P )→ Υ(2S) 9870 < µ+µ− < 10090 MeV/c 3σΥ (2S) < µ+µ− < σΥ (2S) MeV/c
χb(3P )→ Υ(3S) 10300 < µ+µ− < 10526 MeV/c σΥ (3S) < µ+µ− < 3σΥ (3S) MeV/c
16/46
χb1,2(1, 2, 3P )→ Υ (1S)γ fit model
10 10.50
500
1000
1500
2000
2500
-4
-2
0
2
4
10 10.50
200
400
600
800
1000
-4
-2
0
2
4
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (1S)
[GeV/c2
]
√s = 8 TeV
14 < pΥ (1S)T
< 40 GeV/c
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (1S)
[GeV/c2
]
√s = 7 TeV
14 < pΥ (1S)T
< 40 GeV/c
χb(1P )
χb(2P )
χb(3P )
χb(1P )
χb(2P )
χb(3P )
One Crystal Ball (CB) for eachχb1,2(1P, 2P, 3P ) state: 6 CB intotal
Exclude the study of χb0 due to itslow radiative branching ratio.
Product of exponential and linearcombination of polynomials forcombinatorial background.
17/46
χb1,2(1, 2, 3P )→ Υ (1S)γ fit model (2)
Free parameters: µχb1(1P ), yields andbackground parameters.Linked parameters for χb1 and χb2 signals:
µχb2(jP ) = µχb1(jP ) + ∆mPDGχb2(jP ),
j=1,2µχb2(3P ) = µχb1(3P ) + ∆mtheory
χb2(3P )
Nχb = λNχb1 + (1− λ)Nχb2σχb2 = σχb1
Other linked parameters:µχb1(2P ) = µχb1(1P ) + ∆mPDG
χb1(2P )
µχb1(3P ) = µχb1(1P ) + ∆mχb1(3P )
(∆mχb1(3P ) measured in this study)Fixed parameters from MC study:
σχb1(1P ) scaled by 1.17,σχb1(2P )
σχb1(1P ),σχb1(3P )
σχb1(1P )
α and n parameters of CB.
Υ(1S) transverse momentum intervals, GeV/c14 – 40
√s = 7 TeV
√s = 8 TeV
Nχb(1P ) 2090 ± 80 5070 ± 130Nχb(2P ) 450 ± 50 1010 ± 80Nχb(3P ) 150 ± 40 220 ± 60
Background 8830 ± 130 23,910 ± 210
µχb1(1P ), MeV/c2 9889.7 ± 1.0 9890.3 ± 0.7σχb1(1P ), MeV/c2 22.0 22.5σχb1(2P )/σχb1(1P ) 1.5 1.5σχb1(3P )/σχb1(1P ) 1.86 1.86
τ -2.6 ± 0.5 -3.27 ± 0.30c0 -0.08 ± 0.12 0.07 ± 0.06c1 1.33 ± 0.04 0.29 ± 0.04
χ2/n.d.f 1.03 1.24
χb1(1P ) mass agrees with the PDG value
9892.78± 0.26 (stat)± 0.31 (syst) MeV/c2 .
18/46
Mass of χb1(1P ) in χb → Υ (1S)γ decay
10 20 30 40
9.882
9.884
9.886
9.888
9.89
9.892
9.894
9.896
√s =7 TeV,
√s =8 TeV
χb1
(1P
)m
ass
GeV/c
2
pΥ (1S)T [ GeV/c]
The major cause of χb1(1P ) mass floating in 10 MeV/c2 range can be theunknown fraction between Nχb1 and Nχb2 yields (λ parameter). We have onlytheoretical prediction for λ value . In this study λ is fixed to 0.5 for all χbdecays.
In the following, this mass was fixed to 9.887 GeV/c2 which is the valuemeasured on the combined 2011 and 2012 datasets. A systematic uncertaintydue to this assumption has been assigned.
19/46
χb yields in χb → Υ (1S)γ decays
10 20 30 400
10
20
30
40
50
60
70
80
90
100
10 20 30 400
500
1000
1500
2000
2500
10 20 30 400
100
200
300
400
500
600
700
800
900
1000
Eve
nts
pΥ (1S)T [ GeV/c]
χb(3P )
Eve
nts
pΥ (1S)T [ GeV/c]
χb(1P )
Eve
nts
pΥ (1S)T [ GeV/c]
χb(2P )
√s =7 TeV√s =8 TeV
√s =7 TeV√s =8 TeV
√s =7 TeV√s =8 TeV
Yields normalized by bin width and luminosity.
The small difference between 7 and 8 TeV data is due to the productioncross-sections, which are expected to be about 10% larger..
20/46
χb1,2(2, 3P )→ Υ (2S)γ fit model
10.2 10.4 10.6 10.8 110
100
200
300
400
500
600
-4
-2
0
2
4
10.2 10.4 10.6 10.8 110
20406080
100120140160180200220
-4
-2
0
2
4
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (2S)
[GeV/c2
]
√s = 8 TeV
18 < pΥ (2S)T
< 40 GeV/c
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (2S)
[GeV/c2
]
√s = 7 TeV
18 < pΥ (2S)T
< 40 GeV/c
χb(2P )
χb(3P )
χb(2P )
χb(3P )
One Crystal Ball (CB) for eachχb1,2(2P, 3P ) state: 4 CB in total
Exclude the study of χb0 due to itslow radiative branching ratio.
Product of exponential and linearcombination of polynomials forcombinatorial background.
21/46
χb1,2(2, 3P )→ Υ (2S)γ fit model (2)
Free parameters: µχb1(2P ), yields andbackground parameters.Linked parameters for χb1 and χb2 signals:
µχb2(2P ) = µχb1(2P ) + ∆mPDGχb2(2P )
µχb2(3P ) = µχb1(3P ) + ∆mtheoryχb2(3P )
Nχb = λNχb1 + (1− λ)Nχb2σχb2 = σχb1
Other linked parameters:µχb1(3P ) = µχb1(1P ) + ∆mχb1(3P )
(∆mχb1(3P ) measured in this study)Fixed parameters from MC study:
σχb1(2P ),σχb1(3P )
σχb1(2P )
α and n parameters of CB.
Υ(2S) transverse momentum intervals, GeV/c18 – 40
√s = 7 TeV
√s = 8 TeV
Nχb(2P ) 237 ± 29 650 ± 50Nχb(3P ) 50 ± 17 78 ± 26
Background 1830 ± 50 4600 ± 80
µχb1(2P ), MeV/c2 10,249.1 ± 2.2 10,249.9 ± 1.3σχb1(2P ), MeV/c2 13.0 13.3σχb1(3P )/σχb1(2P ) 1.65 1.65
τ -7.5 ± 0.8 -7.7 ± 0.5c0 0.431 ± 0.027 0.435 ± 0.016c1 -2.07 ± 0.09 -2.12 ± 0.05c2 0.79 ± 0.36 0.79 ± 0.17
χ2/n.d.f 0.98 1.35
The mass of χb1(2P ) is about 6 MeV/c2 less than
value in PDG (10.25546 GeV/c2).
22/46
Mass of χb1(2P ) in χb → Υ (2S)γ decay
20 25 30 35 4010.235
10.24
10.245
10.25
10.255
10.26
10.265
√s =7 TeV,
√s =8 TeV
χb1
(2P
)m
ass
GeV/c
2
pΥ (2S)T [ GeV/c]
In the following, this mass was fixed to 10.249 GeV/c2 which is the valuemeasured on the combined 2011 and 2012 datasets. A systematic uncertaintydue to this assumption has been assigned.
23/46
χb yields in χb → Υ (2S)γ decays
0 10 20 30 40 500
10
20
30
40
50
60
0 10 20 30 40 500
1
2
3
4
5
6
7
8
9
10
Eve
nts
pΥ (2S)T [ GeV/c]
χb(2P )
Eve
nts
pΥ (2S)T [ GeV/c]
χb(3P )√s =7 TeV√s =8 TeV
√s =7 TeV√s =8 TeV
Yields normalized by bin width and luminosity.
24/46
χb1,2(3P )→ Υ (3S)γ fit model
10.5 10.6 10.70
5
10
15
20
25
30
35
40
45
-4
-2
0
2
4
10.5 10.6 10.70
2
4
6
8
10
12
14
16
18
-4
-2
0
2
4
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (3S)
[GeV/c2
]
√s = 8 GeV
27 < pΥ (3S)T
< 40 GeV/c
Can
dida
tes/
(20
MeV/c2
)
mµ+µ−γ −mµ+µ− +mPDGΥ (3S)
[GeV/c2
]
√s = 7 GeV
27 < pΥ (3S)T
< 40 GeV/c
χb(3P )
χb(3P )
One Crystal Ball (CB) for eachχb1,2(3P ) state: 2 CB in total
Exclude the study of χb0 due to itslow radiative branching ratio.
Product of exponential and linearcombination of polynomials forcombinatorial background.
25/46
χb1,2(3P )→ Υ (3S)γ fit model (2)
Free parameters: µχb1(3P ),σχb1(3P ), yields, backgroundparameters.Linked parameters for χb1 and χb2signals:
µχb2(3P ) =
µχb1(3P ) + ∆mtheoryχb2(3P )
Nχb = λNχb1 + (1− λ)Nχb2σχb2 = σχb1
Fixed parameters from MC study:α and n parameters of CB.
Υ(3S) transverse momentum intervals, GeV/c27 – 40
√s = 7 TeV
√s = 8 TeV
Nχb(3P ) 31 ± 12 72 ± 16
Background 97 ± 14 283 ± 21
µχb1(3P ), MeV/c2 10,517 ± 4 10,504.0 ± 2.5σχb1(3P ), MeV/c2 9 ± 6 8.3 ± 2.7
c0 0.52 ± 0.18 0.52 ± 0.09c1 -0.42 ± 0.19 -0.36 ± 0.10c2 1.3 ± 0.8 -1.23 ± 0.18
χ2/n.d.f 0.38 1.09
In this study the mass of χb1(3P ) wasfixed to 10.508 GeV/c2 which is thevalue measured on the combined 2011and 2012 datasets.
26/46
Mass of χb1(3P ) in χb → Υ (3S)γ decay (1)
Systematic uncertanties due to unknown ratio and mass difference between χb1and χb2 states:
0 0.5 1
10.5
10.502
10.504
10.506
10.508
10.51
10.512
10.514
10.516
Mas
sofχb1(3P
)(G
eV/c2
)
λ
∆mχb1,2(3P ) = 13 MeV/c2
∆mχb1,2(3P ) = 10 MeV/c2
The mass is measured with different ratios (λ) and mass difference (∆mχb1,2(3P )) between χb1(3P )
and χb2(3P ) states. The measurement is performed on the combined 2011 and 2012 datasets.
27/46
Mass of χb1(3P ) in χb → Υ (3S)γ decay (2)
The measured mχb1(3P )=10,508± 2 (stat)± 8 (syst) MeV/c2 is consistentwith the mass measured in another study with converted photons —10,510± 3.0 (stat)+4.4
−3.4 (syst) MeV/c2 (very preliminary results).
ATLAS measured χb1 and χb2 mass barycenter formχb2 −mχb1 = 12 MeV/c2 and λ = 0.5:mχb(3P ) = 10,530± 5 (stat)± 9 (syst) MeV/c2
D0: mχb(3P ) = 10,551± 14 (stat)± 17 (syst) MeV/c2
28/46
MC efficiency (1)
MC true events χb(3P )→ Υ (1S)γ (other decays have the same shape)
0 0.5 1 1.5 20
500
1000
1500
2000
2500
Can
dida
tes
mµ+µ−γ −mµ+µ−[
GeV/c2]
The flat left tails are due to photons which, although being correctly associated to the χb decay,are poorly reconstructed in the calorimeter. In principle, these tails could be modeled in our fit tosignal, but in practice they will not be distinguishable from background. Therefore, the numberof χb events for efficiency calculations is not determined by simple event counting but froma fit where the tails are considered as background. Υ events are obtained by counting mc-trueevents.
29/46
Monte-Carlo photon reconstruction (2)
Example of fits:
9.8 9.9 10 10.10
1000
2000
3000
4000
5000
6000
-4
-2
0
2
4
χb1(1P )→ Υ (1S)γ
6 < pΥ (1S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (1S)
[GeV/c2
]
Can
dida
tes/
(10
MeV/c2
)
10 10.2 10.4 10.60
500
1000
1500
2000
2500
-4
-2
0
2
4
χb1(2P )→ Υ (1S)γ
6 < pΥ (1S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (1S)
[GeV/c2
]C
andi
date
s/(1
0M
eV/c2
)
10.2 10.4 10.6 10.80
200
400
600
800
1000
1200
1400
-4
-2
0
2
4
χb1(3P )→ Υ (1S)γ
6 < pΥ (1S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (1S)
[GeV/c2
]
Can
dida
tes/
(10
MeV/c2
)
10.2 10.25 10.30
10
20
30
40
50
-4
-2
0
2
4
χb1(2P )→ Υ (2S)γ
18 < pΥ (2S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (2S)
[GeV/c2
]
Can
dida
tes/
(10
MeV/c2
)
10.4 10.5 10.60
20
40
60
80
100
120
-4
-2
0
2
4
χb1(3P )→ Υ (2S)γ
18 < pΥ (2S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (2S)
[GeV/c2
]
Can
dida
tes/
(10
MeV/c2
)
10.45 10.5 10.55 10.6 10.650
10
20
30
40
50
60
70
80
-4
-2
0
2
4
χb1(3P )→ Υ (3S)γ
27 < pΥ (3S)T < 40 GeV/c
mµ+µ−γ −mµ+µ− +mPDGΥ (3S)
[GeV/c2
]
Can
dida
tes/
(10
MeV/c2
)
30/46
Data — Monte Carlo comparison
A comparison of the distribution of the relevant observables used in thisanalysis was performed on real and simulated data, in order to assess thereliability of Monte Carlo in computing efficiencies
0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0 0.2 0.4 0.6 0.8 10
0.01
0.02
0.03
0.04
0.05
0.06
0 0.2 0.4 0.6 0.8 1
-0.02
0
0.02
0.04
0.06
0.08
0.1
0 2 4
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 2 4
0
0.02
0.04
0.06
0.08
0 2 4-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 10 20 300
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0 10 20 30
0
0.02
0.04
0.06
0.08
0.1
0 10 20 30
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
15 20 25 30 350
0.01
0.02
0.03
0.04
0.05
0.06
15 20 25 30 35
0
0.02
0.04
0.06
0.08
0.1
15 20 25 30 35
0
0.05
0.1
0.15
0.2
γ confidence level γ confidence level γ confidence level
χ2 of decay tree fitter χ2 of decay tree fitter χ2 of decay tree fitter
pT [χb(1P )][
GeV/c2]
pT [χb(2P )][
GeV/c2]
pT [χb(3P )][
GeV/c2]
pT [Υ (1S)][
GeV/c2]
pT [Υ (1S)][
GeV/c2]
pT [Υ (1S)][
GeV/c2]
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
Arb
itrar
yun
its
χb(1P ) χb(2P ) χb(3P )
χb(1P ) χb(2P ) χb(3P )
χb(1P ) χb(2P ) χb(3P )
χb(1P ) χb(2P ) χb(3P )
The agreement is generally very good.
31/46
Monte-Carlo photon reconstruction efficiency
20 30 400
5
10
15
20
25
30
Effi
cien
cy,%
pΥ (2S)T [ GeV/c]
χb(3P )→ Υ(2S)γ√s =7 TeV√s =8 TeV
20 25 30 35 400
2
4
6
8
10
12
14
16
18
20
22
Effi
cien
cy,%
pΥ (3S)T [ GeV/c]
χb(3P )→ Υ(3S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
Effi
cien
cy,%
pΥ (1S)T [ GeV/c]
χb(3P )→ Υ(1S)γ√s =7 TeV√s =8 TeV
20 30 400
5
10
15
20
25
Effi
cien
cy,%
pΥ (2S)T [ GeV/c]
χb(2P )→ Υ(2S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
Effi
cien
cy,%
pΥ (1S)T [ GeV/c]
χb(1P )→ Υ(1S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
Effi
cien
cy,%
pΥ (1S)T [ GeV/c]
χb(2P )→ Υ(1S)γ√s =7 TeV√s =8 TeV
Photon is more energetic as pT (Υ)increases so it is reconstructed moreefficiently.
32/46
Systematic uncertainties
Since this analysis measures the fraction of Υ(nS) particles originating fromχb decays, most systematic uncertainties cancel in the ratio and only residualeffects need to be taken into account.
Systematic uncertainties on the event yields are mostly due to the fit model of Υand χb invariant masses, while the ones on the efficiency are due to the photonreconstruction and the unknown initial polarization of χb and Υ particles.
33/46
χb yields systematic uncertainties. Unknown σχb2/σχb1ratio (1)
The σ(χb2)/σ(χb1) ratio prediction:
CDFLHCbΧb scaledΧb
Χc
CMS preliminary
5 10 15 20 25 300.0
0.5
1.0
1.5
2.0
2.5
3.0
pT, GeV
σ(χ
2)/σ
(χ1)
Transverse momentum distributions of the dσ [χ2] /dσ[χ1] ratio. Solid and dashed lines stand
for charmonium and bottomonium mesons. The dot-dashed line corresponds to the rescaled
bottomonium ratio: σb2/σb1(Mχc/MχbpT ). The experimental results for charmonium from
LHCb are shown with dots, CDF — with rectangles, and CMS — with triangles.
34/46
χb yields systematic uncertainties. Unknown σχb2/σχb1ratio (2)
χb yields systematic uncertainties (%) related to unknown λ = Nχb1/(Nχb1 +Nχb2) ratio inthe fit model for χb(1, 2, 3P )→ Υ (1S)γ decays.
Systematic uncertanties are measured in each pΥT bin and center-of-mass energy.
Υ(1S) transverse momentum intervals, GeV/c6 – 8 8 – 10
√s = 7 TeV
√s = 8 TeV
√s = 7 TeV
√s = 8 TeV
Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P )
λ = 0.0 13.3 -8.3 — 10.1 -12.9 — 19.3 -8.7 — 15.9 -7.3 —λ = 0.1 9.0 -6.7 — 6.1 -8.9 — 13.9 -7.0 — 11.1 -5.9 —λ = 0.2 5.4 -5.2 — 2.9 -5.4 — 9.1 -5.2 — 6.9 -4.2 —
λ = 0.3 2.5 -3.4 — 0.8 -3.9 — 4.9 -3.2 — 3.6 -2.6 —λ = 0.4 0.7 -1.6 — -0.3 -1.5 — 1.9 -1.5 — 1.2 -1.1 —λ = 0.5 0.0 0.0 — 0.0 0.0 — 0.0 0.0 — 0.0 0.0 —λ = 0.6 0.8 1.5 — 1.3 0.7 — -0.7 1.2 — 0.1 0.8 —λ = 0.7 2.7 1.6 — 4.1 -2.2 — -0.1 1.9 — 1.4 1.0 —
λ = 0.8 6.1 0.8 — 7.1 -8.3 — 1.6 2.2 — 3.7 0.9 —λ = 0.9 9.4 -1.3 — 10.9 -17.2 — 4.3 2.0 — 7.1 0.5 —λ = 1.0 13.9 -5.8 — 14.7 -24.4 — 7.9 1.7 — 11.2 -0.4 —
Υ(1S) transverse momentum intervals, GeV/c10 – 14 14 – 18
√s = 7 TeV
√s = 8 TeV
√s = 7 TeV
√s = 8 TeV
Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P ) Nχb(1P ) Nχb(2P ) Nχb(3P )
λ = 0.0 15.8 -3.1 36.1 13.7 -13.6 36.2 6.8 -0.9 5.3 7.3 -2.0 -9.5λ = 0.1 10.7 -2.6 25.9 5.5 -7.0 -0.6 4.2 -1.0 5.5 4.7 -1.9 -5.7λ = 0.2 6.3 -1.7 16.6 5.2 -6.3 16.8 2.2 -1.0 5.1 2.5 -1.6 -3.1
λ = 0.3 3.0 -0.9 9.0 2.6 -3.4 9.3 0.9 -0.9 3.9 1.1 -1.2 -1.2λ = 0.4 0.9 -0.2 3.6 0.2 -0.9 0.9 0.1 -0.5 2.1 0.2 -0.6 -0.0λ = 0.5 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0λ = 0.6 0.4 -0.2 -2.1 0.6 0.6 -1.5 0.7 0.3 -6.2 0.5 0.9 -0.7λ = 0.7 1.9 -1.0 -2.9 2.5 0.3 1.4 1.6 2.2 -4.8 1.6 2.0 -1.8
λ = 0.8 4.4 -2.0 -3.0 4.1 -0.2 -4.8 3.1 4.4 -3.5 3.2 3.4 -3.4λ = 0.9 7.8 -3.3 -1.8 9.0 -1.6 7.5 5.4 5.7 -9.6 5.5 5.2 -4.7λ = 1.0 12.0 -4.8 0.4 13.3 -3.1 9.4 8.1 8.1 -11.4 8.3 7.3 -5.9
Theory predicts that λ belongs to the range between 0.4 and 0.7.
35/46
Systematic uncertainties due to χb polarization
Efficiencies are evaluated on MC where χb particles are unpolarized. To evaluatesystematic effects due to the unknown polarization of χb, MC events are reweighted asdescribed in HERA-B Collaboration, I. Abt et al., Production of the CharmoniumStates χc1 and χc2 in Proton Nucleus Interactions at s = 41.6-GeV, arXiv:0807.2167and LHCb collaboration, R. Aaij et al.,Measurement of the cross-section ratioσ(χc2)/σ(χc1) for prompt χc production at
√s = 7 TeV, arXiv:1202.1080
For each simulated event in the unpolarised sample, a weight is calculated fromthe distribution of the following angles in the various polarisation hypothesescompared to the unpolarised distribution.
ΘΥ angle between the directions of the µ+ in the Υ rest frame and theΥ in the χb rest frame.
Θχb angle between the directions of the Υ in the χb rest frame and χbin the lab frame.
φ angle between the Υ decay plane in the χb rest frame and the planeformed by χb direction in the lab frame and the direction of the Υin the χb rest frame.
Two hypotheses for χb1 state (w0, w1) and three hypotheses for χb2 (w0, w1,w2).
36/46
Systematic uncertainties due to χb polarization
The ratio of efficiency for unpolarized χb to efficiency for polarized χb.
10 20 30 40
0.99
0.995
1
1.005
1.01
1.015
1.02
1.025
1.03
1.035
10 20 30 40
0.985
0.99
0.995
1
1.005
1.01
pΥ (1S)T [ GeV/c] p
Υ (1S)T [ GeV/c]p
Υ (1S)T [ GeV/c] p
Υ (1S)T [ GeV/c]
ε mχb1/εunpol
ε mχb1/ε unpol
χb1(1P )→ Υ (1S)γ
|mχb1 | = 0
χb1(1P )→ Υ (1S)γ
|mχb1 | = 1
The difference between ratio and 1 is considered as systematic uncertainty.
37/46
Summary of systematic uncertainties
Υ fraction uncertainties common to all χb decays (%)
Υ fit model ±0.7γ reconstruction ±3
Summary of Υ fraction systematic uncertainties (%)(maximum deviations that were found in pΥ
T bins):χb fit model χb polarization
χb(1P )→ Υ (1S)γ +4.3−5.8
+5.1−4.0
χb(2P )→ Υ (1S)γ +4.8−6.2
+5.8−6.8
χb(3P )→ Υ (1S)γ +19.6−16.6
+6.9−6.7
χb(2P )→ Υ (2S)γ +2.3−7.0
+8.7−7.8
χb(3P )→ Υ (2S)γ +19.7−19.9
+4.5−4.2
χb(3P )→ Υ (3S)γ +20.9−27.6
+6.4−7.5
38/46
Υ fractions in χb → Υγ decays
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(1S
)fr
actio
n,%
pΥ (1S)T [ GeV/c]
χb(1P )→ Υ (1S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(1S
)fr
actio
n,%
pΥ (1S)T [ GeV/c]
χb(2P )→ Υ (1S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(1S
)fr
actio
n,%
pΥ (1S)T [ GeV/c]
χb(3P )→ Υ (1S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(2S
)fr
actio
n,%
pΥ (2S)T [ GeV/c]
χb(2P )→ Υ (2S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(2S
)fr
actio
n,%
pΥ (2S)T [ GeV/c]
χb(3P )→ Υ (2S)γ√s =7 TeV√s =8 TeV
10 20 30 400
10
20
30
40
50
60
70
80
90
100
Υ(3S
)fr
actio
n,%
pΥ (3S)T [ GeV/c]
χb(3P )→ Υ (3S)γ√s =7 TeV√s =8 TeV
10 20 30 400
5
10
15
20
25
30
35
40
45
50
Υ(1S
)fr
actio
n,%
pΥ (1S)T [ GeV/c]
χb(1P )→ Υ (1S)γ√s =7 TeV√s =8 TeV√s =7 TeV (2010)
Outer error bars show statistical and systematics
errors, inner error bars — only statistical errors.
39/46
Summary
Measured fractions of Υ (1, 2, 3S) originated from χb decays. About 40%of Υ come from χb, with mild dependence on Υ transverse momentum.
This analysis improves significantly the statistical precision of theprevious work and adds more decays and transverse momentum regions.
The LHCb detector design allows to perform measurements in Υ rapidityand transverse momentum regions, which are complementary to the onesexploited by ATLAS and CMS.
Measured mass of χb(3P ) is 10, 508± 2 (stat)± 8 (stat) MeV/c2,consistent with another determination which uses converted photons.
A software profiling tool was developed.
The results on χb production were regularly presented at the LHCbbottomonium working group, an internal document was prepared by the author,is currently under review and will form the basis of a future LHCb publication.
40/46
High Level Trigger softwareperformance profiling
41/46
Optimizing the LHCb trigger is crucial in many measurement. In particular,decays with muons can be triggered on with unprecedented high efficiency(∼80%).
The dimuon trigger:
is not prescaled.
is used in more than 50% ofLHCb’s papers.
42/46
Trigger software profiling
The trigger needs fastalgorithms!
Most experiments require a trigger in order torecord interesting events at suitable rate.
L0 Hardware Trigger 40MHz → 1MHz.Search for high pt, µ, e, γ, hadroncandidates.High Level Software Trigger Farm
HLT1: Add Impact parameter cutsHLT2: Global event reconstruction.
100 man/years work that has only 20-30ms to process an average event.
29K CPUs or 1700 servers.
2× 1015 bytes stored in 2012.
43/46
CPU Profiler Auditor
CPU profiler tool is vital for trigger optimization.
C++ library.
Deployed into the core softwareframework in LHCb — Gaudi.
Based on Intel R©
VTuneTMAmplifier XE User API
Full and clear reports.
44/46
CHEP2012: Talk and Paper
This work was presented on “Computing in High Energy and Nuclear Physics”conference. The largest of the conferenes in this area. The conference is heldevery 18 months.
A. Mazurov and B. Couturier, “Advancedmodular software performancemonitoring”, Journal of Physics:Conference Series 396 (2012), no.5052054.
45/46
Thank you!
46/46