Bc lifetime measurement using BcJy e X channel Preblessing cdfnote 7758

1
Bc lifetime measurement using Bc?J/y e X
channel(Preblessing / cdfnote 7758)

• Masato Aoki, Shinhong Kim
• University of Tsukuba
• Ilsung Cho, Intae Yu
• SungKyunKwan University
• Ting Miao
• FNAL

2
Introduction

• We had measured the cross section of Bc in J/ye
  X channel (note7518)
• Electron ID using SoftElectronModule, dE/dx
• Lxygt3sigma to kill prompt background
• Background
• Fake electron estimate fake rate, J/ytrack as
  a control sample
• Residual conversion estimate conversion finding
  efficiency using B0?J/y p0, p0?gg or gee MC. Use
  J/ytagged conversion
• b-bbar use Pythia MC, B?J/yK is used for the
  normalization
• Fake J/y J/y mass sideband subtraction
• We release the lifetime cut and measure the Bc
  lifetime
• New background from prompt events

3
J/ye selection cuts
4
Summary of x-section measurement
Prompt BKG is killed by lifetime cut (Lxygt3sigma)
5
Overview of lifetime measurement procedure

• Same cuts as Bc x-section measurement (note7518)
• Same technique for background fraction estimation
• Background lifetime shapes from fitting
  background samples
• Follow B lifetime measurement(CDF6266) for
  techniques
• Single Gaussian as resolution function
• Systematic error includes study of alternative
  resolution function and Punzi effect
• K-factor estimation similar to that of B?Dln but
  with binning of M(J/ye)

6
Summary after releasing lifetime cut
Excess contains prompt BKG and Bc signal
7
Background fraction

• Background fraction (the denominator includes
  prompt bkg and Bc signal)
• fake e 0.141 /- 0.022
• res. conv 0.086 /- 0.041
• bbbar 0.080 /- 0.022
• fake J/y 0.209 /- 0.012
• Statistical and systematic errors are included
• Constrain the fractions for the final fitting
  using Gaussian

8
K-factor
9
Additional cuts for the lifetime analysis

• Check sLxy distribution
• B?J/yK
• J/yelectron
• Use sLxylt70mm

10
Fitter check using B?J/yK

• Simply check our fitter using B?J/yK
• Result
• ct504.1 ? 9.3mm
• Agree with blessed result from CDF

11
Overview of background shape determination

• Fake electron
• J/ytrack with electron fake rates
• Fake J/y
• Sideband in J/ytrack candidates
• Residual conversion
• J/ytagged conv. electron with conversion finding
  efficiency
• b-bbar
• Pythia MC but with change of GS/FE/FC for
  systematic error
• Prompt
• Assume to be resolution function

12
Fake electron

• PDF for fake electron BKG

e fake rate N normalization factor
can be expected from J/y mass distribution
Use same error scaling factor for both real J/y
and fake J/y here
13
fake J/y parameterization

• PDF for fake J/y

14
Fit results of fake electron fake J/y

• table ? next page

15
Fit results of fake electron fake J/y
l mm
16
Issue on fake J/y shape

• J/ytrack, conversion sample have fake J/y
  component as well as J/yelectron
• Looking at fake J/ytrack, conversion, electron
  events, we found their shapes are similar
• ?see next page
• Use common fake J/y shape
• Use J/ytrack sample for every fake J/y shapes
• Limited stat. for conversion, electron samples

17
J/y sideband event comparison
J/ytrack
J/yelectron
J/yconv.-e

• ?similar shapes

18
Residual conversion

• PDF for residual conversion BKG

Constrain fake J/y and scale factor
19
Fit result of conversion BKG
l mm
Constrained using J/ytrack sample
20
b-bbar background

• PDF for b-bbar BKG
• Background events passing selection cut from each
  production process
• Gluon splitting 70
• Flavor excitation 25
• Flavor creation 5

(scaling factor is not constrained)
?Syst. study GS and FE
21
Fit result of b-bbar BKG
l mm
22
Prompt background

• It is difficult to estimate the size of prompt
  background from either MC or data
• ? Float prompt BKG fraction for the final fitting
  
• We use resolution function as prompt background
  shape (Gaussian)

23
Likelihood definition for the signal fitting

• PDF for signal
• Likelihood

24
Signal fitting
ct(Bc) 142.6 22.2/-19.9 mm
25
signal fitting (contd)
26
Systematic uncertainties

• K-factor
• M(Bc), pT(Bc),lifetime(Bc),decay channel,
• Background shapes
• fake J/y shapes, w/o efficiency weighting,
• Resolution function (follow CDF6266)
• Choice to treat Punzi effect as systematic error
  for now
• Double Gaussians, Gaussiansymmetric exponential
• Silicon alignment ? borrow the result of B
  lifetime analysis using J/yX exclusive mode

27
Systematics from K-factor

• M(Bc) ? 6.291, 6.251 GeV
• ? 142.4, 142.6 mm ? Dct ? 0.2 mm
• t(Bc) ? 0.4, 0.7 ps
• ? 142.3, 142.4 mm ? Dct ? 0.3 mm
• Hb?J/yX spectrum
• ? 141.3 mm ? Dct ? 1.3 mm
• Trigger simulation
• ? 142.8 mm ? Dct ? 0.2 mm
• Inclusive Bc?J/yXen channel (K factor ? next
  page)
• ? 142.1 mm ? Dct ? 0.5 mm

28
K-factor for inclusive Bc decays
29
Systematics from background shapes

• Fake J/y Use J/ye sideband
• 137.5 mm ? Dct -5.1 mm
• Res. conv. Use J/yconv sideband
• 145.1 mm ? Dct 2.5 mm
• b-bbar No error scaling in MC fitting
• ? 140.8 mm ? Dct - 1.8 mm

30
fake rate / finding efficiency weighting
J/yconv.-e
J/ytrack

• Fake e 141.2 mm ? Dct -1.4 mm
  Conv. 141.7 mm ? Dct -0.9 mm

31
b-bbar 100 FE, 100 GS
l mm

• 100 FE ? 152.4 mm Dct 9.8 mm
• 100 GS ? 140.6 mm Dct -2.0 mm

32
Different resolution functions

• Single Gaussian
• Double Gaussians
• Gaussian symmetric exponential
• ? Convolute

33
J/ytrack fit result for GaussianSymmetric Exp.
?ct(Bc)136.5 mm ? Dct -6.1 mm
34
J/ytrack fit result for Double Gaussians
?ct(Bc)136.2 mm ? Dct -6.4 mm
35
ct error distributions for Punzi effect
?ct(Bc)138.0 mm ? Dct -4.6 mm
36
Systematics from ct resolution

• Resolution function
• ? Dct -6.4 mm
• Punzi effect
• (sct of fake J/y, fake e, conv, others)
• ? Dct -4.6 mm
• Silicon alignment effect from note7409
• ? Dct ?1.0 mm

37
Summary of systematic errors
K factor
? 1.5 mm
BKG shapes
10.1 / -6.0 mm
Resolution
1.0 / -7.9 mm
Total10.3/-10.0 mm
38
Summary

• We measured the Bc lifetime using J/yelectron
• ct(Bc)142.6 22.2/-19.9(stat.) ?10.3(syst.) mm
• or
• t(Bc)0.475 0.074/-0.066(stat.) ?0.034(syst.)
  ps
• Details are described in note7758

• Theoretical prediction
• 0.55 ?0.15 ps
• Run1 CDF
• 0.46 0.18/-0.16 ?0.03 ps
• Run2 D0
• 0.448 0.123/-0.096 ?0.121 ps

39
Backup
40
fake J/y with 2 negative exponentials

• Why fake J/y fit quality is so bad?
• ?complicated shape at ctlt0 of fake J/y event
  makes bad fit quality
• ?try to add one more negative exponential
• ? see next page
• ?result of Bc fitting ct(Bc) 142.1 mm
• ?the effect of the negative side is 0.5 mm

41
fake J/y with different parameterization
w/ one negative exponential
w/ two negative exponentials
42
b-bbar FE only fixing s1.25
43
For the lifetime measurement

• Same cuts as Bc x-section measurement (note7518)
• Mass window M(J/ye)4 6GeV/c2
• Background
• fake electron use J/ytrack
• fake J/y use fake J/ytrack
• residual conversion use J/ytagged conv.
• b-bbar Pythia MC
• prompt resolution function (Gaussian)
• Use common fake J/y shape for
• J/ytrack, J/yconv., J/yelectron samples
• Constrain background shapes using Gaussian
• K-factor
• Divide by 4 mass bins (4-4.5, 4.5-5, 5-5.5, 5.5-6
  GeV/c2)

44
(No Transcript)
45
GausGaus Punzi effect

• Resolution function is fixed using B events
• RF parameters from B?J/yK fitting
• s1.271 0.018/-0.017
• fs20.10 0.016/-0.014
• s23.07 0.18/-0.17
• J/ye fit result with new RF Punzi term
• ct(Bc) 131.4 21.5/-19.2 mm
• ?Dct(Bc) -11.2 mm

46
GausSym. Exp Punzi effect

• Resolution function is fixed using B events
• RF parameters from B?J/yK fitting
• s1.284 0.015/-0.015
• fexp0.21 0.03/-0.03
• sexp1.70 0.13/-0.11
• J/ye fit result with new RF Punzi term
• ct(Bc) 134.4 21.8/-19.4 mm
• ?Dct(Bc) -8.2 mm