Semileptonic B Decays at BABAR

Vladimir Golubev
Budker Institute of Nuclear Physics,
Novosibirsk, Russia
(for the BaBar Collaboration)

Beauty 2005
Assisi (Perugia), June 20–24, 2005
1
 Semileptonic B Decays – Why are they
Interesting
2
+
G 2
G 2 5
(b  c )  |
F
Vcb | 2
mb
2
( mb  mc ) 3
(b  u )  |F
Vub | mb
192 2 192 2

 Semileptonic decays

theoretically simple at parton level W 
 rate depends on CKM elements |Vcb| and |
Vub|, the quark masses mb and mc b
 the leptonic current factors out cleanly,
Vcb , Vub
thus one can probe strong interactions in B c, u
mesons
 sensitive to QCD corrections, OPE
 A precise determination of |Vcb| and |Vub| with
reliable errors
important for

understanding B decay rates

testing of the unitarity of the CKM matrix
and predictions of CP violation in B mesons

V.Golubev - BABAR Beauty2005, Assisi 2

 Semileptonic B decay studies at BABAR

Inclusive decays Exclusive decay modes

bcl B  Xc l  B  D* l B  D l 

|Vcb|, total decay rate, Branching fractions, decay
HQE parameters form factors, |Vcb|

b  u l  B  Xu l  B   l B  l 

|Vub|, total decay Branching fractions, decay
rate, lepton form factors, |Vub|
momentum and
hadron mass spectra
2-nd B tagging: untagged, lepton tag, hadronic tag

V.Golubev - BABAR Beauty2005, Assisi 3

 Inclusive B  Xc l  decays: HQE Expansions
 Heavy Quark Expansions, tool to correct for QCD effects

Expansion in terms of 1/mb and s(mb)
 Separate short- and long-distance effects at ~ 1 GeV

Perturbative corrections calculable from mb mc s(mb)

Non-perturbative parameters cannot be calculated
 We choose calculation by Gambino & Uraltsev hep-ph/0401063 &
0403166
3
 Kinetic mass scheme2 toO (1/ mb )

Eℓ moments O ( s )

mX moments O ( s ) kinetic O (1/ mb2 )
 8 parameters to be fitted chromomagnetic

Vcb mb mc B ( B  X c  ) 2 G2  D3  LS

3

spin-orbit
O (1/ mb3 )
Darwin

 Measure 8 moments, each as a function of minimum lepton energy El

V.Golubev - BABAR Beauty2005, Assisi 4
 Measurement of Electron Energy Moments

BABAR Albert et al. Phys.Rev. D 69, 111104 (2004)

 Inclusive e spectrum in 3x10
± 6

electron-tagged (1.4<p*<2.3GeV/c) 47 + 9 /fb

BB events
 Corrected for detector effects
 Corrected for non-prompt electrons M 0 
d 
B
– lepton charge correlation and MC 156,000 e±
 Corrected for B0 – B0 mixing

 Moments for Ecut = 0.6 … 1.5 GeV

 Corrected for the final state radiation

Translated to B rest frame
Partial BF
 Subtracted B  Xuℓdecays

Lepton
Energy Moments

V.Golubev - BABAR Beauty2005, Assisi 5

 Measurement of Hadron Mass Moments
Aubert et al. Phys.Rev. D 69, 111103

Events/0.5 GeV2
(2004)
BABAR
 BB events tagged by a fully reconstructed
hadronic B decays 7114
 e± or ± with El > Ecut (0.9…1.6 GeV/c) Events
D, D*
 Lepton charge – B flavor correlation
 Improve mx measurement by kinematic fit to high mass
whole event, resolution ~350 MeV charm
 To eliminate dependence of moments on states
uncertain BF and unknown masses of high mass
(Mx ) 350 MeV / c 2
charm mesons we calibrate mx measurement mX2 (GeV2)
 Calibrate mx based on MC simulation Mx2[GeV2/c4]

mxmeas (Gev)
 Linear relation between mxmeas and mXtrue
 Validate calibration with excl. B D(*) l 
 Moments corrected for detector effects and bkg

Hadron mass
moments

mXtrue (GeV)
V.Golubev - BABAR Beauty2005, Assisi 6
 HQE Fits to Hadron Mass and Lepton Energy
Moments
Aubert et al. Phys.Rev.Lett. 93, 011803 (2004)

● = used, ○ = unused
BABAR MX moments in the nominal fit

M1x M2x M3x M4x

M0l M1l M2l M3l

2/ndf = 20/15 Eℓ moments Red line: HQE fit

Yellow band: theory errors

V.Golubev - BABAR Beauty2005, Assisi 7

 HQE Fit Results (kinetic mass scheme, scale =1GeV)

Vcb (41.4 0.4exp 0.4HQE 0.6 th ) 10 3

Bc (10.61 0.16exp 0.06HQE )% BABAR
mb (4.61 0.05exp 0.04HQE 0.02s ) GeV
mc (1.18 0.07 exp 0.06HQE 0.02s ) GeV
2 (0.45 0.04exp 0.04 HQE 0.01s ) GeV 2
G2 (0.27 0.06exp 0.03HQE 0.02s ) GeV 2
 D3 (0.20 0.02exp 0.02HQE 0.00s ) GeV 3
3
 LS ( 0.09 0.04exp 0.07 HQE 0.01s ) GeV 3

 Separate fits to hadron and lepton moments give consistent

results
 G2 and LS3 are consistent with B-B* mass splitting and HQ sum
rules
 Considerable improvement in precision for |Vcb| (±2%) and
V.Golubev - BABAR Beauty2005, Assisi 8
 |Vub| from total B Xu l  decay rate

 Total b  u ℓ  decay rate from OPE:

 |Vub| from inclusive BF(B  Xu l ) (theory error 5%):

 Experimentally challenging due to charmed semileptonic

background: 2
(b  u ) Vub 1
 
(b  c ) Vcb 2 50

V.Golubev - BABAR Beauty2005, Assisi 9

 |Vub| from partial B Xu l  decay rates

 Charmed semileptonic background can be controlled

in limited regions of phase space

 Non-perturbative Shape Functions are used to extract |Vub|

from partial branching fractions. The SF parameters are
extracted from b  sg photon spectrum

 New approach to |Vub| extraction:

• Differential decay rate parametrization from BLNP
Bosch, Lange, Neubert, and Paz, Nucl. Phys. B 699, 335
(2004);
Lange, Neubert, and Paz, hep-ph/0504071
• New source of information on SF parameters – moments
of the b  c ℓ  lepton spectrum
Neubert, Phys.Lett. B 612, 13 (2005)

V.Golubev - BABAR Beauty2005, Assisi 10

 |Vub| from partial B Xu l  decay rates

 Different types of partial decay rate

measurements:
- electron endpointhep-ex/0408075

- Ee and q2 regions withhep-ex/0506036

neutrino reconstruction
- Mx and q2 regions hep-ex/0408068

q2(Gev/c)2 vs. Ee(GeV) q2(Gev/c)2 vs. Mx(GeV)

q2=(pl+p)2

Kinematic regions
of B Xcl decays

V.Golubev - BABAR Beauty2005, Assisi 11

 |Vub| from B Xu l  endpoint spectrum

On-peak data
Off-peak data

(On-peak)-(Off-peak) data

Histogram – B Xcl MC

(On-peak)-(Off-peak) data

Histogram – B Xul MC

Differential
Xuebranching fraction
as a function of electron
momentum, corrected for
detector effects and final-state
Signal extraction: subtraction of radiation
BB and continuum background (On-peak)−(Off-peak) data
Histogram – B  Xul MC

V.Golubev - BABAR Beauty2005, Assisi 12

 Endpoint spectrum analysis: |Vub|
extraction

 Determination of Br for pe>pmin

 Extrapolation to total charmless semileptonic decay Br using
shape function parameters from B  Xs spectrum and
calculations of DeFazio-Neubert (*) and Kagan-Neubert(**)
 Extraction of |Vub| using the OPE relation.
 BABAR preliminary results, based on BELLE shape function
parameters and DFN approach (hep-ex/0408075):
Br(pe>2.0GeV/c)=(0.5310.032stat0.049syst)×10-3
Br=(2.16 0.24exp 0.27SF)×10-3
|Vub|=(4.40 0.24exp 0.28SF 0.21OPE+mb)×10−3
 Analysis based on BLNP calculation is being completed

* F.De Fazio and M.Neubert, JHEP 9906, 017 (1999)

( )

(
**) A.L.Kagan and M.Neubert, Eur.Phys.J. C 7, 5 (1999)

V.Golubev - BABAR Beauty2005, Assisi 13

 q2 − Ee analysis with neutrino
reconstruction

 Event selection:
 High energy electron Ee>2GeV
 Missing momentum used for neutrino
parameters estimation
 Cuts on missing momentum magnitude
and direction, and event shape
 Suppression of bcebyshmax=m2D:

2
q
shmax mB2  q 2  2mB ( Ee  ),
4 Ee
(slightly modified for B -meson
motion in the (4S) rest frame)

Acceptance ~ 14%, signal/bkg.~0.6

V.Golubev - BABAR Beauty2005, Assisi 14

 BABAR
q2 − Ee analysis results preliminary

Test of neutrino energy resolution

using pure control sample of
BD(*)e events
Signal extraction for Ee>2.0 GeV and sh<3.5
GeV2,
BB background normalization for sh>4.25
GeVBr(
2
, Ee>2.0GeV,
s <3.5GeV2)=(3.54±0.33±0.3410-4
| Vub | h B /(   B ) ,  - normalized partial rate (*),  B 1.604 0.023ps

| Vub |(3.95 0.18 0.19 00..58

42 0.25) 10
3 (*)
B.O.Lange, M.Neubert
and G.Paz,
hep-ph/0504071
V.Golubev - BABAR Beauty2005, Assisi 15
 Mx−q2 analysis

 Select events with a fully reconstructed B  D(*)X decay (Breco)

 Study charmless semileptonic decay of B recoiling against
Breco
- require lepton with p*lep > 1 GeV

 2D fit to measure partial BF

hep-ex/0408068
in [Mx<1.7 GeV, q2>8 GeV2]
Mx < 1.7 GeV 1.7 < Mx < 2.2 GeV 2.2 < Mx < 2.8 GeV

downfeed Signal 82 fb-1

bcl + other

V.Golubev - BABAR Beauty2005, Assisi 16

 BABAR
Mx−q analysis: the results
2
preliminary
(CKM 2005)

 Mx spectrum unfolding and fitting:

| Vub |(5.22 0.30stat 0.31exp.syst 0.21fu 0.25OPE ) 10 3 (DFN* )

 Mx−q2 2-dimentional fit:

Br ( M X  1.7GeV, q 2  8GeV 2 ) (8.96 1.43stat 1.44syst ) 10 4

| Vub |(5.18 0.41stat 0.40exp.syst 0.22fu 0.25OPE ) 10  3 (DFN* )

| Vub |(4.98 0.40stat 0.39 exp.syst 0.47 fu OPE ) 10  3 (BLL** )
(*)
F.De Fazio and M.Neubert, JHEP 9906, 017 (1999)
(**)
Bauer, Ligeti and Luke, hep-ph/0111387

V.Golubev - BABAR Beauty2005, Assisi 17

 BABAR
Inclusive B Xu l  decay summary preliminary

Technique |Vub|×103

El>2GeV
4.40 0.13stat 0.25exp.syst 0.28fu 0.21OPE
(ICHEP2004)

El vs. q2
3.95 0.18stat 0.19exp.syst  ( 00..58
42 ) SF 0.25 theo
(NEW)

Mx vs. q2
(CKM 2005)
5.18 0.41stat 0.40exp.syst 0.22fu 0.25OPE
Average
(using the
4.22 0.57 CL 0.45
procedure of
HFAG)

V.Golubev - BABAR Beauty2005, Assisi 18

 |Vub| from exclusive B Xu l  decays

Strong nonperturbative effects included in form factors

d( B   ) GF2 3 2 2 2

2
 3
| k  | | Vub | | f  ( q )|
dq 24
q2 = (plep + p)2
B ( B   )
| Vub |
x B
ISGW2
GF2
48 3  
3 2 2 2 Ball’04
x | k | | f  ( q ) | dq

FNAL’04
HPQCD’04
Calculations
 Light-Cone Sum Rules: Ball,Zwicky (hep-ph/0406232)
 Unquenched Lattice QCD: HPQCD (hep-lat/0408019)
and FNAL (hep-lat/0409116)
 ISGW2: quark model

V.Golubev - BABAR Beauty2005, Assisi 19

 Exclusive decays, experimental
techniques
 Different experimental approaches to measure BF:

PURITY EFFICIENCY
Neutrino reconstruction
Untagged analysis
B  ℓ , B  ℓ 

Recoil of B  D(*) ℓ 
Semileptonic Tags
B  ℓ 

Recoil of B  D(*)X
Hadronic Tags 9 channels (,,,,’,a0)

V.Golubev - BABAR Beauty2005, Assisi 20

 Exclusive B  l  decays, tagged analysis

l D(*)
v

tag B
signal B

 v
l
 Use BD(*)l decays of Discriminating Variables
opposite B to “tag” event  B0  -l+angle between the B momentum and the
 Use events with BD(*)l plane defined by the D(*)l and l momenta: cos2B
decays on both sides to
check tagging efficiency
 B+  0l+angle between B and 0l system: cos B-l
 Measure B(B0-l+) in
bins of q2
 BF for B+0l+

V.Golubev - BABAR Beauty2005, Assisi 21

 BABAR
B  l  tagged analysis: the results preliminary
(CKM 2005)

B0-l+
on-peak data
signal MC
232 million B lv MC
BB pairs b ¨ ulv MC
B0B0 MC
B+B− MC
off-peak data
B+0l+

Branching Fraction Results

B ( B 0     ) (1.03 0.25 stat 0.13syst ) 10  4

in 3 bins of q2
88 million
BB pairs

B ( B    0  ) (1.80 0.37 stat 0.23syst ) 10  4

V.Golubev - BABAR Beauty2005, Assisi 22

 Exclusive B  l  decays, untagged
analysis

• Reconstruct neutrino from full event + ensure good

reconstruction quality

• Signal yield extracted from common fit to E and mES in 5

q2 bins for +l and 3 bins for landleach
 q2 resolution
mES  s / 4 | p B* |2 σ1 0.163 GeV 2
* σ 2 0.495 GeV 2
E E  B s /2

q 2 ( pl  p ) 2

neutrino momentum scaled so

E=0  improved q2 resolution

V.Golubev - BABAR Beauty2005, Assisi 23

 BABAR
Untagged B  l signal extraction
preliminary

Bl+
Data
Signal MC
Comb. Sig.
Crossfeed
bcl
qq

Bl
Branching Fraction Results
B( B 0     ) 
(1.38 0.10 stat 0.18 syst 0.08 FF ) 10  4

B ( B 0     ) 
( 2.14 0.21stat 0.53syst 0.28 FF ) 10  4

83 million
BB pairs
V.Golubev - BABAR Beauty2005, Assisi 24
 BABAR
Form factor fits and |Vub| extraction preliminary
(CKM 2005)

Untagged l l  The l untagged

sample begins to
distinguish among FF
calculations: ISGW II
is disfavored:
Prob(2) = 3%
 Other FF calculations
tested are
compatible with data

Extraction of |Vub|

V.Golubev - BABAR Beauty2005, Assisi 25

 Summary

 HQE fits to moments hadronic mass and lepton energy moments:

Based on 80 fb-1, ~30% of the current BABAR data, we have made significant
advances in the analysis of semileptonic B decays.

| Vcb |Br(B X0c.4l exp

(41.4 )= 0(10.61
.4 HQE 0±0.16 10±0.06
.6theory )exp 3
HQE)%

High precision measurement of mb and mc, and exp. determination of HQE

parameters.

 Inclusive B Xu l  decays, three different approaches:

The most precise determination of |Vub|
Measurement of partial semileptonic branching fractions for different kinematic
regions

 Exclusive B () l  decay modes in tagged and untagged samples:

Determination of |Vub|
Begin to test decay form factor calculations

The analyses will be repeated for larger data samples and more precise
shape function parameters derived from from B Xs photon spectrum

V.Golubev - BABAR Beauty2005, Assisi 26