Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Studies of Semileptonic B Decays at LHCb Marco Fiore
Università di Ferrara -‐ XXVIII Ciclo Tutor: DoD. Concezio Bozzi
26/11/13 Marco Fiore 1
Outline • Semileptonic B0 Decays:
1. Theory & MoPvaPon; 2. B0-‐>D*-‐µ+ν sample; 3. Monte Carlo Studies; 4. Conclusions
• Summary
26/11/13 Marco Fiore 2
Neutral B mixing and CP ViolaPon • Neutral B mesons: mass eigenstates are not flavour
eigenstates ; 〉±〉=〉 qqHL BqBpB ||| ,
3
Prob(B0; B0(t))=
Prob(B0; B0(t))=
!U (t) = e"!t (1+ cos#mt) / 2
!M (t) = e"!t pq
2
(1" cos#mt) / 2
Current WA Δmd= (0.510 ± 0.004) ps-‐1
26/11/13
Flavour specific asymmetries B0 à D*-‐ µ+ ν is a flavour specific decay, we can compute
afs =!(B0 " B0 " D*#µ+!µ )#!(B
0 " B0 " D*+µ#!µ )!(B0 " B0 " D*#µ+!µ )+!(B
0 " B0 " D*+µ#!µ )
4 26/11/13
Current WA afsd= (-‐0.03 ± 0.20) %
afsd (!10"2 )
afss (!10"2 )
B0àD*-‐µ+ν; D*-‐àD0πsl-‐; D0-‐>K+π-‐
26/11/13 5
We have 500k signal events with 2011 data, including also 2012 data (2M events in total) we get a staPsPcal error on Δmd of 0.002 (half the error from the world average) and of 0.18% on afs (from the WA it is 0.20%).
Decay Pme • For oscillaPon and CPV studies it is required to compute the decay Pme:
where k is a correcPon due to the missing neutrino and is parameterized as a funcPon of momentum and invariant mass of detectable parPcles • k-‐factor can be computed from Monte Carlo but there are limitaPons due to:
• StaPsPcs; • ProducPon mechanisms; • Decay dynamics; • Limited knowledge of processes that are indisPguishable from signal (e. g. B-‐>D*µνX);
What is the impact of all this? How can we put remedy to these limitaPons?
26/11/13 Marco Fiore 6
t = mB
pBd = d *mB
pD*µk k =
pD*µpB
MC Studies • The idea is to generate several sets of MC samples (~50M
events each) with different parameterizaPons of poorly known quanPPes, in order to assess their impact on systemaPc uncertainPes;
• Performing these studies on the Full SimulaPon would require excessive compuPng Pme. Therefore, we generate the decay of interest and simulate the detector response parametrically;
• We compare the parameterized Monte Carlo with the Full Monte Carlo SimulaPon. We must check the observables to be in good agreement with the Full SimulaPon.
7 26/11/13 Marco Fiore
Smearing • We apply smearing, detecPon efficiencies and other
correcPons only to the signal parPcles. Given the kinemaPcal range of the signal parPcles, we do not apply any momentum resoluPon;
• B0 decay length; • z(B0) – z(PV); • z(D0) – z(B0); • Cuts on IPPV, p, pT; • µ & K simulated PID; • MagnePc Field Effect; • Trigger
8 26/11/13
Bd
νµ
µ+
D*-‐
πs-‐
_ D0
K+
π-‐
PV
Marco Fiore
ResoluPons from Full SimulaPon
26/11/13 Marco Fiore 9
-0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.20
2000
4000
6000
8000
10000
dReco-dTrue hdEntries 229318Mean 0.002176RMS 0.04804
dReco-dTrue
0 1 2 3 4 5 60
2000
4000
6000
8000
10000
dReco hdREntries 229318Mean 1.237RMS 1.052
dReco
0 1 2 3 4 5 60
2000
4000
6000
8000
10000
dTrue hdTEntries 229318Mean 1.235RMS 1.053
dTrue
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
1000
2000
3000
4000
5000
6000
7000
8000
z(B0)-z(PV) histozPVEntries 229318Mean 0.01068RMS 0.4989Underflow 4921
/ ndf 2χ 388.1 / 141Prob 5.98e-25p0 13.3± 5921 p1 0.0099± 0.3331 p2 0.00117± 0.00439 p3 0.0024± 0.1683 p4 0.00690± 0.04729 p5 0.021± 1.029 p6 0.0070± 0.2161 p7 0.002117± 0.003795 p8 0.0071± 0.3987
z(B0)-z(PV)
-2 -1.5 -1 -0.5 0 0.5 1 1.5 20
1000
2000
3000
4000
5000
z(D0)-z(B0) histozD0Entries 229318Mean -0.008016RMS 0.6525Underflow 6541
/ ndf 2χ 468.8 / 141Prob 8.086e-37p0 60.5± 6046 p1 0.0166± 0.2249 p2 0.002732± 0.004991 p3 0.0077± 0.2494 p4 0.00338± -0.01079 p5 0.0227± 0.5629 p6 0.0127± 0.5073 p7 0.01495± -0.03441 p8 0.2± 1.5
z(D0)-z(B0)
mm
mm
mm
PID efficiency from real data
is expected, since tracks with higher transverse momentum traverse the detector at higher215
polar angles, in the lower occupancy regions. The proton misidentification probability is216
smaller than 0.5% for all pT ranges and momentum above 30GeV/c. It drops quickly with217
momentum for the lowest pT ranges, reaching a plateau at about 30-40 GeV/c. The pion218
and kaon misidentification probabilities have a similar behavior, increasing with decreas-219
ing pT . Above 40 GeV/c, the pion misidentification probability is almost at the level of the220
proton misidentification probability. At low momentum, decays in flight are the dominant221
source of incorrect identification, as can be seen from the difference between the pion/kaon222
and proton curves. While the proton misidentification probability, within the pT intervals223
chosen, lies within 0.1-1.3%, the pion and kaon misidentification probabilities are within224
Momentum [GeV/c]0 20 40 60 80 100
IM!
0.8
0.85
0.9
0.95
1
1.05
<1.7 [GeV/c]T
0.8<p
<3.0 [GeV/c]T
1.7<p
<5.0 [GeV/c]T
3.0<p
>5.0 [GeV/c]T
p
<1.7 [GeV/c]T
0.8<p<3.0 [GeV/c]
T1.7<p
<5.0 [GeV/c]T
3.0<p>5.0 [GeV/c]
Tp
LHCb(a)
Momentum [GeV/c]0 20 40 60 80 100
)µ!
(pIM
"
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
<0.8 [GeV/c]T
p<1.7 [GeV/c]
T0.8<p
<3.0 [GeV/c]T
1.7<p<5.0 [GeV/c]
T3.0<p
LHCb(b)
Momentum [GeV/c]0 20 40 60 80 100
)µ!"(
IM#
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
LHCb(c)
Momentum [GeV/c]0 20 40 60 80 100
)µ!
(KIM
"
0
0.01
0.02
0.03
0.04
0.05
0.06
0.07
LHCb(d)
Figure 5: IsMuon efficiency and misidentification probabilities, as a function of momen-tum, in ranges of transverse momentum: εIM (a), ℘IM(p → µ) (b), ℘IM(π → µ) (c) and℘IM(K → µ) (d).
9
26/11/13 Marco Fiore 10
MagnePc field effect
• The LHCb magnet has a bending power of 4Tm. Since p(GeV) = 0.3B(T)R(m) we correct for the Lorentz force due to the magnet field (along the y axis) by giving a “kick” to px of ±1.2 GeV;
• (x,y)µ,π,K,πsl required to be in the acPve regions of tracking detectors;
• Same for (x,y)µ with respect to muon staPons 11 26/11/13 Marco Fiore
Trigger
• A parametric descripPon of trigger lines is not straighworward. The reason is that informaPon available in the Full SimulaPon are not accessible in parameterized Monte Carlo;
• Therefore, we limit our studies to trigger lines which can be reproduced by simple cuts on signal parPcles;
12 26/11/13 Marco Fiore
k-‐factor matching
26/11/13 Marco Fiore 13
k_binM_3_binP_2_GLb
Entries 44150.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.10
100
200
300
400
500
k_binM_3_binP_2_GLb
Entries 4415
kFactor
k_binM_3_binP_5_GLb
Entries 12200.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.10
20
40
60
80
100
120
140
160
180
k_binM_3_binP_5_GLb
Entries 1220
kFactor
MC11 = yellow GL = red
MC11 = yellow GL = red
Momentum distribuPons
14 0 100 200 300 400 500 600
310×0
2000
4000
6000
8000
10000
B0 momentumpBzero
Entries 244030
Mean 1.188e+05
RMS 7.751e+04
B0 momentum
0 5000 10000 15000 20000 25000 30000 35000 400000
2000
4000
6000
8000
10000
B0 transverse momentumptBzero
Entries 244030
Mean 8098
RMS 4497
B0 transverse momentum
MC11 = yellow MC12 = green GL = red
26/11/13 Marco Fiore
15
-0.4 -0.2 0 0.2 0.4 0.6 0.8 10
2000
4000
6000
8000
10000
12000
14000
16000
theta muonthmu
Entries 213129Mean 0.1129RMS 0.06733
theta muon
0 1 2 3 4 5 6 70
200
400
600
800
1000
1200
1400
1600
1800
phi muonphimu
Entries 213129Mean 3.136RMS 1.807
phi muon
0 1 2 3 4 5 6 7 80
1000
2000
3000
4000
5000
6000
7000
eta muonetamu
Entries 213129Mean 3.069RMS 0.6624
eta muon
0 10 20 30 40 50 60 70 80 90 100310×0
500
1000
1500
2000
2500
3000
3500
4000
4500
muon momentumpMuon
Entries 213129
Mean 2.941e+04
RMS 2.1e+04
muon momentum
0 2000 4000 6000 8000 10000 12000 140000
2000
4000
6000
8000
10000
muon transverse momentumptMuon
Entries 213129Mean 2749RMS 1882
muon transverse momentum
MC11 = yellow MC12 = green GL = red
26/11/13 Marco Fiore
16
-0.4 -0.2 0 0.2 0.4 0.6 0.8 10
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
theta slow pionthslpi
Entries 213129Mean 0.09152RMS 0.0524
theta slow pion
0 1 2 3 4 5 6 70
200
400
600
800
1000
1200
1400
1600
1800
phi slow pionphislpi
Entries 213129Mean 3.138RMS 1.8
phi slow pion
0 1 2 3 4 5 6 7 80
1000
2000
3000
4000
5000
6000
7000
8000
eta slow pionetaslpi
Entries 213129Mean 3.25RMS 0.598
eta slow pion
0 5000 10000 15000 20000 25000 300000
2000
4000
6000
8000
10000
12000
SlowPion momentumpSlowPion
Entries 213129Mean 4529RMS 3297
SlowPion momentum
0 200 400 600 800 1000 1200 1400 1600 1800 20000
2000
4000
6000
8000
10000
12000
SlowPion transverse momentumptSlowPionEntries 213129Mean 327.4RMS 189.5
SlowPion transverse momentum
MC11 = yellow MC12 = green GL = red
Conclusions
• Semileptonic B decays offer the highest staPsPcal power; • The fast simulaPon is very useful to study systemaPc
uncertainPes, however it needs more tuning (work in progress);
• In principle this could also be a very important tool for the whole collaboraPon, provided it is generalized;
26/11/13 17
Summary • In my first year I worked on the development of a fast and simplified Monte Carlo;
• Status of my work was presented regularly at the Ferrara LHCb group, to the LHCb SimulaPon Working Group and to the Semileptonic Asymmetry Working Group;
• I also worked on the data acquisiPon system of a prototype for the LHCb Muon Detector to be installed axer the second LHC long shutdown;
• In October I performed tests on the Front End electronics of the LHCb Muon Detector, to be concluded in December;
26/11/13 Marco Fiore 18
Other AcPviPes • I parPcipated to the Hadron Collider Summer School in Goeyngen, Germany where I gave a seminar talk about “Bayesian Inference”; to the Cabeo School in Ferrara about Physics beyond Standard Model and to a detector school in Legnaro (PD);
• I parPcipated to the InternaPonal Conference on B-‐Physics at Hadron Machines (Beauty) in Bologna;
• I aDended three ParPcle Physics seminars at the Physics Department in Ferrara;
• I aDended the LHCb Analysis&Soxware Week at CERN and will aDend the next LHCb week in December
26/11/13 Marco Fiore 19
26/11/13 Marco Fiore 20
Backup Slides
Muon Detector Front-‐End Board CalibraPon
26/11/13 Marco Fiore 21
LHCb (muon) detector
26/11/13 Marco Fiore 22
A MulP Wire ProporPonal Chamber
• Gas detector (Ar, CO2, CF4); • The signal can be read on: – Wire: anode à negaPve charge collecPon – Pad: cathode à posiPve charge collecPon
• LHCb MWPCs have both readings mode – FEBs for anodic (negaPve) reading [AN] – FEBs for cathodic (posiPve) reading [CP]
• For the short LHC upgrade it’s needed to test the spare FEBs (≈ 800)
26/11/13 23 Marco Fiore
A Front-‐End Board of the Muon Detector
• Each Region(R1,R2,R3,R4) of MD has a different number of FEBs per MWPC;
• Each FEB has 16 channels; • One FEB can be used in anodic or cathodic readings; • One FEB is a PCB with three chips:
2 CARIOCAs: analog part (amplificaPon and discriminaPon 8+8) 1 DIALOG: digital part (logical combinaPon of discriminated signal, internal
counter for measuring the rate)
T O P
B O T T O M
26/11/13 Marco Fiore
24
FEBs TEST: goal
1. Check if they are workable; 2. Measurement of the electronic noise – Dark Rate Threshold scan - EvaluaPon of the RMS
3. FEB Ranking; 4. SubsPtuPon of broken or noisy FEBs
currently on the detector with the tested ones
26/11/13 25 Marco Fiore
FEBs test: setup
• a tester MWPC [M2R2] because it has the most complicated reading approach;
• 6AN + 8CP = 14FEBs connected to the tester; • The tests were done only in the CP posiPons, because we just need the FEB to see a MWPC;
• Readout electronics; • Power supply (HV àM2R2 and LV à FEB)
26/11/13 26 Marco Fiore
1) Tester MWPC
2) 4 FEBs (right side) connected to tester MWPC
3a) Readout electronics 4) PC for data acquisiPon
3b) LV Power supply
27 Marco Fiore
Data AcquisiPon
DIALOG 0 AN 9 Ch Thres Rate(KHz) Mean RMS 01 071 004111.34 71.01 4.06 02 071 002935.45 70.58 4.03 03 065 003117.48 65.31 3.75 04 075 001079.69 74.70 2.88 05 063 000944.92 62.92 2.90 06 066 000825.69 66.17 2.59 07 067 001108.34 66.76 2.59 08 066 000777.94 66.27 2.02 09 069 001859.37 68.77 3.75 10 070 001388.11 69.76 2.88 11 061 001543.64 60.63 3.45 12 068 002721.11 67.87 3.16 13 074 001082.28 74.17 2.95 14 068 000457.75 68.20 2.01 15 068 001995.69 68.13 3.17 16 067 000485.21 66.39 2.04
1) Threshold scan: for each threshold the dark rate is measured; 2) The mean and RMS of the Rate DistribuPon as a funcPon of the threshold are evaluated;
3) As a reference we took the <RMS> of all Channels of all FEBs; 4) Then we evaluated the pull defined as:
!
pulli =RMSi " RMS
#tot26/11/13 28 Marco Fiore
Data Analysis AN (315 FEBs)
• STRATEGY: in each FEB we evaluate the pulli (i=1,…,16) mean value and rank the FEBs in four categories: – A: very good (pull <= -‐0.30) à 30 FEBs = 10% – B: good (-‐0.30 > pull >= 0.15) à 241 FEBs = 76% – C: noisy (0.15 > pull >= 2.00) à 33 FEBs = 10% – D: very noisy (pull >= 2.00) à 11 FEBs = 4%
26/11/13 29 Marco Fiore
Data Analysis CP (301 FEBs)
• STRATEGY: in each FEB we evaluate the pulli (i=1,…,16) mean value and rank the FEBs in four categories: – A: very good (pull <= -‐0.30) à 59 FEBs = 20% – B: good (-‐0.30 > pull >= 0.15) à 141 FEBs = 47% – C: noisy (0.15 > pull >= 2.00) à 97 FEBs = 32% – D: very noisy (pull >= 2.00) à 4 FEBs = 1%
26/11/13 30 Marco Fiore
Conclusion
• 616 FEBs were tested; • We studied the electronic noise and classified them;
• More than 70% of them can be considered good;
• We are aware that CP FEBs are more noisy; • Residual FEBs will be tested in December 2013
26/11/13 31 Marco Fiore