Study of the Bc Meson Properties using BcgJ/yen Decay at CDF II

1
Study of the Bc Meson Properties using BcgJ/yen
Decay at CDF II

• Masato Aoki
• University of Tsukuba, Japan

2
The Bc meson

• Ground state of bottom-charm quark bound system
• Both quarks are heavy
• ? Similar to cc and bb heavy quarkonium families
• But the two quarks have different flavor
• ? Only weak decay is possible
• ? Measurable lifetime
• Bc provides unique features of heavy quark bound
  states

3
Bc production and decay

• Differently flavored heavy quarks
• Much smaller production rate
• Unique decay (bgc,cgs,annihilation)
• ? Rich decay modes

Decay BR
bgc BcgJ/y en 1.9
bgc BcgJ/y p 0.13
cgs BcgBs p 16.4
cgs BcgBsr 20.2
ann. Bcgtn 1.6
ann. Bcgcs 4.9
Species Prod. Fraction
B 40
B0 40
Bs 10
b-baryons 10
Bc 0.05
Theoretical calculation 7.4nb Phys. Lett. B605,
311(2005)
hep-ph/0308214(2003)
4
Bc discovery in CDF Run I (9196)

• Bc signal search using BcgJ/yln (le,m) channel
• 20 Bc signal events were observed in 110 pb-1 of
  J/ygmm trigger data

PRL 81, 2432 (1998) and PRD 58, 112004 (1998)
harder pT(Bc) was assumed
5
The CDF II detector _at_Tevatron
HAD Calorimeter
Muon Chamber

• Silicon Detector
• hlt2.0
• svertex30mm
• Central Outer Tracker
• hlt1.0
• spT/pT0.15 pT
• Muon Chamber
• hlt0.6 (1.0)
• EM, HAD Calorimeter
• hlt1.1(EM), lt0.9(HAD)
• sE/Ev (13.52/ET 32) (EM)
• v (502/ET 32) (HAD)

EM Calorimeter
Silicon Detector
Central Outer Tracker
B field 1.4 T
6
Bc reconstruction in this analysis

• Use semileptonic decay BcgJ/yen
• Large BR ? statistical advantage
• Improved J/ygmm trigger
• pT(m)gt1.5 GeV/c (was 2 GeV/c)
• Factor 5 J/y yield (factor 2 BgJ/y yield)
• No narrow mass peak due to neutrino
• Bc signal excess above estimated backgrounds
• Measure
• Production cross section and lifetime
• (Mass to be measured in exclusive channel)

7
J/ygmm trigger data
Additional requirements in this analysis

• Lint 360 pb-1
• pT(mm) gt 3 GeV/c
• Reduce fake
• Reduce prompt
• 2.2M J/y
• Pick M(mm)-M(PDGJ/y) lt 50 MeV/c2

8
Electron reconstruction

• pT(e) gt 2 GeV/c, hlt1.0
• Track-based reconstruction algorithm
• Higher reconstruction efficiency for low pT
  electrons
• Electron ID using both dE/dx and calorimeter
  information

9
Electron ID using dE/dx
Ze/sZ?1.3

• Energy deposit in COT

p
p
e
?2GeV
m
K
90 efficiency
e
m
p
p
K
ZeLog((dE/dx)measured/(dE/dx)predicted)
10
Electron ID using calorimeter

• 10 variables from the calorimeter
• Form a Joint Likelihood Function
• L distribution depends on
• Isolation
• Transverse momentum
• Track charge
• L cut positions as functions of them
• Constant eID efficiency

Choose 70 efficiency
11
Calorimeter 10 variables
E/p
EHad/EEm
E/p (shower max)
Ewire/Estrip (shower max)
c2wire (shower max)
c2strip (shower max)
DZ (shower max)
DX (shower max)
E (preradiator)
DX (preradiator)

• Red electrons from ggee-
• Blue pions from K0sgpp-

12
s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK) Measurement
13
s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK) measurement
strategy

• Reconstruct mass of J/ye pair
• Estimate all of the backgrounds
• Event counting above the backgrounds
• Normalization mode BugJ/yK
• BugJ/yK has similar topology to Bc?J/yen
• Cancel out most of uncertainties

m
m
J/y
J/y
m-
m-
Bu
Bc
n
K
e
14
Backgrounds and control samples

• fake J/y
• J/y mass sideband events
• fake electron
• J/y track
• bb (bgeX, bgJ/yX)
• PYTHIA Monte Carlo, BugJ/yK
• electron from photon conversion
• J/y electron tagged as photon conversion
• prompt J/y
• No control sample
• Only know it has zero lifetime

15
J/ye pair reconstruction
m
J/y
m-
n
Bc
e

• One displaced decay vertex
• Lxy/sLxygt3
• Kill prompt background
• ? Negligible
• pT(J/ye)gt5 GeV/c
• Reduce non-B tracks
• Search window
• Wide mass region due to missing neutrino
• 4ltM(J/ye)lt6 GeV/c2

signal region
16
Background Estimates
17
Fake J/y background

• Fake J/y background can be estimated by J/y mass
  sideband events
• J/yelectron, J/ytrack, J/yconv.-e have fake
  J/y part each other
• To avoid double counting, fake J/y events
  (sideband events) will be subtracted in the
  following background estimations

18
Fake electron background

• Control sample J/y track
• (after dE/dx cut)
• Fake rate
• Fake rates for K/p/P
• Control sample D0gKp, Lgpp
• Combine them with proper fraction
• Nfake N(J/ytrack) x efake
• as a function of pT(track)

19
Fake rate estimates for K/p/p

• Control samples
• D0gKp for K/p,
• Lgpp for proton
• Fit mass distribution to obtain of events
  before and after eID cut

20
Particle composition in J/ytrack sample

• PYTHIA Monte Carlo simulation
• Dominant fake source pion

CDF Preliminary
Data
Kaon
Pion
d0.09
PYTHIA
Kaon
Pion
Proton
Proton
after dE/dx cut
pion fraction from dE/dx fitting ? max difference
0.09
21
Fake rate
p, p-
CDF Preliminary
average fake rate
positive charge negative charge
K, K-
lt 0.8
p, p

• The average fake rate is applied to J/ytrack
  after dE/dx cut

22
Systematic uncertainties

• Isolation dependence
• 14.5
• Trigger bias on fake rate
• 7.2
• Particle fraction difference between MC and data
• 1.9
• Sample statistics
• J/ytrack 2.0
• Fake rate 0.9

23
Estimated fake electron background

• 15.43 ? 2.54 events

24
Photon conversion electrons
g

• Control sample J/ye tagged as ggee
• Find collinear partner track
• These candidates are removed from the J/ye
  candidate list
• Miss-tracking due to very low pT partner track
• ? Not 100 finding efficiency
• ? Residual photon conversion electrons
• Need to understand the finding efficiency

25
Conversion finding efficiency

• Assume g comes from p0 (p0ggg)
• Monte Carlo sample B0gJ/yp0
• etag 50 efficiency
• Residual ggee events
• Systematics study
• Use pT(tracks) in J/ytrack as pT(p0)
• Vary dalitz decay (p0geeg) fraction

26
Systematic uncertainties
CDF Monte Carlo

• pT spectrum
• 43.7
• Lifetime
• 2.0
• Dalitz decay
• 1.0
• Sample statistics
• Finding efficiency 3.8
• J/yconv. e 30.1

27
Estimated residual photon conversion

• 14.54 ? 7.75 events

28
bb background
quark annihilation
gluon fusion
flavor excitation
gluon splitting

• It is possible to make a common vertex with J/y
  from B decay and e from B decay
• bb background

29
bb background estimate

• PYTHIA Monte Carlo simulation
• Normalization with data N(BugJ/yK)
• Dominant contribution from gluon splitting bb

Open angle distribution between J/y and electron
with all kinematical requirements
Bc signal MC
CDF Preliminary
Additional requirement Df lt 90deg.
30
Systematic uncertainties

• Monte Carlo setting (PDF/ISR)
• 31.4
• Isolation dependence of eID efficiency
• 2.9
• Branching ratio of normalization mode BugJ/yK
• 0.9
• Calorimeter fiducial coverage
• 0.9
• Statistics
• MC sample 6.5
• N(BugJ/yK) in MC 1.9
• N(BugJ/yK) in data 1.8
• e(eID by cal) 1.4
• e(eID by dE/dx) 1.0

31
Estimated bb background

• 33.63 ? 11.38 events

32
Summary table
Fake e 15.43?0.31?2.52
Conversion 14.54?4.38?6.39
b-bbar 33.63?2.20?11.17
Total bkg 63.59?4.91?13.59
Data 178.50?14.67
J/y sideband events are subtracted
33
M(J/yelectron) data and excess

• Total background 63.614.4 events
• Excess 115 events
• Significance 5.9s

34
Cross check e-track IP w.r.t. J/y vertex
electron

• Lxy(J/y)gt3s
• Bc should make a peak around IP0

m
IP
m-
PV
Lxy(J/y)
CDF Preliminary
CDF Preliminary
35
s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK) calculation

• Having established the signal !!
• Lets calculate
  s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK)

• acceptance ratio
• efficiency ratio

36
Normalization mode BugJ/yK

• Similar reconstruction criteria as J/ye
• N(Bu)2872?59 was found in the dame data

37
Kinematical acceptance ratio

• RK Akin(Bu)/Akin(Bc) 4.421.02

Systematic uncertainties
MC parameters RK DRK
Central value 4.416?0.082 0
M(Bc)6.291 GeV/c2 4.403?0.082 ?0.013
M(Bc)2.251 GeV/c2 4.394?0.082 ?0.022
t(Bc)0.7 ps 4.076?0.074 -0.34
t(Bc)0.4 ps 5.006?0.096 0.59
pT(Bc) spectrum 3.578?0.062 ?0.838
sLxy 4.576?0.086 ?0.16
J/yenX other decays 4.769?0.090 0.353
Trigger 4.299?0.079 -0.117
e/K tracking 2
38
Reconstruction efficiency ratio

• Most of the efficiencies are expected to be same
  for Bc and B
• u electron ID with calorimeter and dE/dx

39
Kinematical limits
-1 lt y(Bc) lt 1
4GeV

• Choose pT(B) gt 4GeV/c, y(B) lt 1 as our cross
  section definition
• (Run1 pT(B) gt 6GeV/c, y(B) lt 1)

40
Result of s?B ratio measurement

• s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK)
• 0.282 ? 0.038(stat.) ? 0.035(yield) ?
  0.065(acc.)
• (pT(B)gt4GeV/c, h(B)lt1) assuming softer
  pT(Bc)
• Most of the difference to the Run I measurement
  is from the treatment of input pT(Bc) spectrum
• But they are still consistent

41
Lifetime Measurement
42
Introduction

• Three major decay diagrams
• Lifetime should be t(Bc) ? 1/(0.61.20.1) ? 0.5
  ps if all the three diagrams contribute
  to the decay

other contributions G(Pauli interference)
G(penguin)
43
Theoretical calculations

• V. V. Kiselev, hep-ph/0308214 (2003) Review
  paper

t(Bc) ps Author
0.357 0.362 C. H. Chang et al. (Commun. Theor. Phys. 35, 57 (2001))
0.48?0.05 V. V. Kiselev et al. (Nucl. Phys. B 585, 353 (2000))
0.63?0.02 A. Yu. Anisimov et al. (Phys. Atom. Nucl. 62, 1739 (1999))
0.59?0.06 A. Yu. Anisimov et al. (Phys. Lett. B 452, 129 (1999))
0.46 0.47 A. El-Hady et al. (Phys. Rev. D 59, 094001 (1999))
0.38?0.03 L. P. Fulcher (Phys. Rev. D 60, 074006 (1999))
0.4 0.7 M. Beneke et al. (Phys. Rev. D 53, 4991 (1996))
0.55?0.1 V. V. Kiselev (Phys. Lett. B 372, 326 (1996))
lt 1.0 I. I. Bigi (Phys. Lett. B 371, 105 (1996))
0.40 C. H. Chang et al. (Phys. Rev. D 49, 3399 (1994))
1.1 1.2 C. Quigg (FERMILAB-CONF-93/265-T (1993))
0.50 M. Lusignoli et al. (Z. Phys. C 51, 549 (1991))
44
Lifetime measurement strategy

1. Release Lxy(J/ye)gt3s cut
2. Cut on Lxy error (sLxy lt 70 mm)
3. Pick J/ye events in signal region(4-6 GeV/c2)
4. Estimate of background events using same way as
   s?B ratio measurement
5. Un-binned maximum likelihood fit with J/ye data
6. Input pseudo-proper decay length and its error
7. Background PDFs from each control sample
8. Signal PDF with neutrino effect correction
9. Fit J/ye data to extract Bc lifetime

45
Pseudo-proper decay length

• ct proper decay length
• X pseudo-proper decay length
• K correction factor

K-distributions for 4 M(J/ye) bins
46
Background estimates (w/o decay length cut)

• Background events are estimated using same way as
  s?B ratio measurement
• Prompt background and Bc signal are from the
  lifetime fitting directly

M(J/ye) GeV/c2 4-6
fake J/y 164.09.1
fake electron 110.219.0
photon conversion 67.434.8
bb 63.018.4
prompt ??
Bc ??
data 783
systematic uncertainties are included
47
PDF and likelihood function

• Signal PDF
• Background PDF
• Event PDF
• Log likelihood

signal term
backgrounds term
Prompt bkg shape is assumed to be a resolution
function (Gaussian)
48
Background distributions
Fake J/y
Photon conversion
Fake electron
49
Fitter check before J/ye data fitting
BugJ/yK data
CDF Preliminary
ct504.1 9.3(stat.) mm ?good agreement with CDF
Run II result 498.8 8(stat.) 4(syst.) mm
Toy MC
Fitter returns reasonable lifetime result and
error
50
J/ye data fit result
51
Systematic uncertainties
Total systematic uncertainty is order of 7
(10mm)
52
Comparison with theoretical calculations

• CDF Run2 (360pb-1,J/ye)
• ct 142 22/-20 10 mm
• ? t 0.474 0.073/-0.066
  0.033 ps

Operator Product Expansion 0.55 ? 0.15 ps
Bethe-Salpeter Model 0.460.47 ps
Light-Front Constituent Quark Model 0.59 ? 0.06 ps
Light-Front ISGW Model 0.63 ? 0.02 ps
Hard-Soft Factorization 0.55 ? 0.1 ps
QCD Sum Rules 0.48 ? 0.05 ps
53
Experimental results of the Bc meson lifetime
measurement
54
J/ye mass distribution, again

• Normalization 46 GeV/c2 using lifetime fit
  result
• Prompt shape Assume to be J/ytrack with Lxy lt
  3s
• Good agreement !!

55
Summary of the Bc analysis

• Have established the Bc signal in J/yen final
  state with 5.9s significance
• Measurements with good precision (worlds best)
• s(Bc)B(Bc?J/yen) / s(Bu)B(BugJ/yK)
• 0.282 ? 0.038(stat.) ? 0.035(yield) ?
  0.065(acc.)
• Lifetime
• 0.4740.073/-0.066(stat.) ?0.033(syst.) ps
• The results are consistent with CDF Run I
  measurements

56
FIN Thank you

• This lifetime result with minor update was
  accepted for publication in Physical Review
  Letters PRL 97, 012002 (2006)