Measurement of the Branching fraction B( B ? D* l n )

1
Measurement of the Branching fraction B( B ? D l
n )

• C. Borean, G. Della Ricca
• G. De Nardo, D. Monorchio
• M. Rotondo

Riunione Gruppo I Napoli 19 Dicembre 2002
2
Outline

• Introduction
• Analysis strategy
• Main aspects of the analysis
• Branching ratio results on 2000-2002 data set.

3
B ?Dln and the CKM element Vcb

• The decay B?Dln proceeds by the quark level
  decayb ? c.
• The decay rate therefore depends on Vcb.
• The picture is complicated by the strong
  interaction

4
Measurement of Vcb from B ?Dln
Strong interaction effects can be studied by
means of theHeavy Quark Effective Theory (HQET).
Kinematics function
differential decay rate
CLEO
Fitting d?/dw, VcbF(w) is obtained. Theory (
HQET) provides F(w) allowing the extraction of
Vcb, by means of an extrapolation w ? 1 (zero
recoil).
5
Analysis strategy

• Define selection criteria for the signal.
• Study the backgrounds directly from data as far
  as possible.
• Estimate remaining backgrounds from Monte Carlo
  simulation.
• Determine the selection efficiency. Correct the
  Monte Carlo simulation evaluating the
  corresponding systematic uncertainty.
• Fit signal and residual background from resonant
  and non-resonant semileptonic decays, exploiting
  available physical constraints (we studied both
  electron and muon channel and 4 D0 decay modes)

6
The data set analyzed
2000 20.25 fb-1 on-resonance 2.60 fb-1
off-resonance 2001 35.54 fb-1 on-resonance
3.79 fb-1 off-resonance 2002 24.30 fb-1
on-resonance 3.19 fb-1 off-resonance
2000 MC
2001 MC
2002 MC
7
Event selection

• A Charged D meson and a lepton (muon and
  electron) of opposite charge are reconstructed
• the D meson is reconstructed from the D?D0psoft
  
• D0 reconstructed in 4 modes
• The Neutrino is NOT reconstructed

Constrained vertex fit on D0 decay products, soft
pion and lepton
The signal is measured in 2 ? 4 independent
signal samples .
D0 is reconstructed in 4 decay modes
8
Backgrounds

• The event is not from BB (continuum)
• Reconstructed D is fake (combinatorial)
• lepton is fake (fake lepton)
• B ? D X the other B ? l X (uncorrelated)
• The amount of these backgrounds are be estimated
  directly from experimental data

9
Data samples definition
10
Combinatorial background
The mass difference Dm(MD-MD) discriminate real
D events against combinatorial events.
Maximum likelihood fit, using a double gaussian
for the peaking component and an empirical
distribution for the background. The data set has
been subdivided in several samples sharing the
same peak and combinatorial parameters according
to the different resolutions ( i.e. SVT only
tracks w.r.t. DCH tracks) and background
distributions.
11
Continuum background
The peak yield fitted in off-resonance
experimental data has been scaled according to
the known ratio of data collected on-resoncance
and off-resonance.
12
Fake lepton background

• Signal leptons are tight leptons (maximum purity)
• Fake lepton sample composed by lepton candidates
  which fail the loose lepton identification
  criteria
• The PID selection efficiencies and mis-id
  probabilities as a function of momentum and
  direction are re-weighted according to our sample
  lepton momentum distribution the fraction of
  fake leptons in all samples are extracted

efficiency matrix
true number of events
measured yield
13
Uncorrelated background

• Events with a D and a lepton in the same side
  are enriched in uncorrelated background.

14
Background characterization
15
Fitted Yields
2000
2001
2002
16
Correlated background
e,m
X

• several decay modes contributing
• the overall amount is small
• Estimated from Monte Carlo simulation

17
Extraction of the signal

• After the background subtraction, the signal
  sample is composed by B ? Dl n and B ? D l n
  and non resonant semileptonic decays with a D.
• To disentangle the signal the variable cos qBY is
  used

muons
electrons
signal
D
18
Fit of signal and background components

• Least square fit to the 4 (D mode)? 2 (lepton
  mode) data samples
• all parameters free or imposing equal ratio of D
  / D fractions
• signal and backgrounds shapes taken from Monte
  Carlo

19
Examples of fit results
electron - D0 ? K ? - year 2000
muon - D0 ? K ? - year 2002
20
Monte Carlo correction and systematics
uncertainties

• Correction for PID
• Correction for tracking and ?0 reconstruction
  efficiencies.
• lepton, D0 decay products but most important soft
  ? tracking.

21
Soft pion tracking efficiency

• The soft ? momentum spectrum is really soft 50
  MeV ltpTlt 200 MeV

• From control sample D decays the relative
  tracking efficiency between soft ? and ordinary
  tracks has been measured

22
Other systematics uncertainties

• D background composition and shape modeling in
  Monte Carlo ( 2).
• vertexing algorithm (1)
• stability in the binning (1)
• lepton momentum cut (1)

23
Branching fraction (blind!)
24
Conclusions

• We have analyzed the largest Dl ? sample ever
• Systematics is dominant in the uncertainty.
• Some modes have large uncertainty. Most probably
  the K ? and K 3? mode will be used for the
  measurement leaving the other two for a cross
  check.
• Dominant uncertainties does not depend on us
• f/f0 ratio and D0 PDG branching fractions!
• D modeling is 2.
• It may improve when D measurements will be
  performed.
• soft pion efficiency is historically the main
  concern. We think we have it under control.
• Results are blind. BABAR internal review is
  started. They will give the OK for unblinding.
  Conference target Moriond ( march 2003)

25
Backup slides
26
(No Transcript)