2004/8/17ICHEP20041 Measurement of |V ub | at Belle Toru Iijima Nagoya University for Belle Collaboration August 17, 2004 32 nd International Conference.

2004/8/17 ICHEP2004 1

Measurement of |Vub| at Belle

Toru Iijima

Nagoya University

for

Belle Collaboration

August 17, 2004

32nd International Conference on High Energy Physics

2004/8/17 ICHEP2004 2

Introduction|Vub| measurement

Crucial to test the standard model of the CP violation.

Both inclusive/exclusive B→Xu l decays are useful.

Belle explores B tagging technique to extract B→Xu l in good quality.

Good signal-to-noise

Sufficient statistics

low middle high

poor

mod.

goodS/N

Eff.

trad. recon.

Full recon.

advanced recon.

Semilep. decay

Belle (mX,q2)PRL92,071802

Lum.

with 140fb-1 (152M BB)Results are preliminary

In this talk, we report measurements of

Inclusive B→Xu l

with fully reconstructed tag.

Exclusive B→Xu l with semileptonic decay tag.

2004/8/17 ICHEP2004 3

Inclusive B→Xu l w/ Fully Reconstructed Tag

Tagging side: Full reconstruction of Btag using hadronic decay modes.

Signal side: Lepton Detection

Neutrino reconstruction using Pmiss, Emiss.

Reconstruction of (mX, q2)

Υ(4S)e- e+

Btag

Bsig

K

mX: Hadronic inv. mass

v

Tagging side Btag→D(*) etc.

q2: Leptonic inv. mass 2Signal side Bsig→Xu l

2004/8/17 ICHEP2004 4

Full Reconstruction of Btag

Reconstructed modes (180 in total)

7 modes

6 modes

0 (*) (*)1/ / / SB D a D

(*)0 (*)1/ / / SB D a D

0 0/D D

0 0 0/D D sD

sD 0D

D

sD 2 modes

B0 tag.1.58 ×105 eff.=0.21% purity = 47%

B0 tag.1.58 ×105 eff.=0.21% purity = 47%

B+ tag.2.47 ×105 eff.=0.33% purity = 50%

B+ tag.2.47 ×105 eff.=0.33% purity = 50%

Inclusive B→Xu l

Mbc (beam constrained mass) distribution 140fb-1 data

Signal region: Mbc > 5.26GeV/c2, -0.2 GeV < E < 0.05 GeV

2004/8/17 ICHEP2004 5

Reconstruction of Signal Side

Semileptonic lepton selectionLepton momentum Pl > 1.0 GeV/c

Lab. angle 26 ~ 140deg.

+ J/ veto, conversion veto etc.

Normalization to B→X l

Inclusive B→Xu l

mX < 1.7 GeV/c2, q2 > 8 GeV2

Phase space cut to minimize theory error

※ same as our previous (mX,q2) measurement with simulated annealing [PRL 92, 071802 (2004)]

Reduction of bc background• Number of leptons Nl = 1• Total net charge ∑Qi=0• Missing mass -1.0<MM 2<0.5 Ge

V2

• Missing mom. Direction cosMM < 0.95• Number of kaons NK+ = NKS = 0

0 2 4 6 8 141210-4 -2MM2 (GeV2/c2)

Expected MM2 dist. b→u b→c

2004/8/17 ICHEP2004 6

Signal Yield ExtractionFit Mbc distribution in each mX region with Gaussian+’Argus’ to subtract the combinatorial background in the Btag sample.

Fit the obtained mX distribution with expected b → u and b → c distribution.

Nb→ u = 174±26

Nsl = (5.07±0.4)×104

Inclusive B→Xu l

The number of semileptonic leptons is extracted by similar fitting on Mbc before the b → u selection cut.

mX<1.7GeV/c2

q2>8GeV2

Mbc distribution for each mX region mX distribution with Btag bkg. Subtracted.

Expected dist. of b→u b→c

Btag background

0 0.5 1.0 1.5 2.0mX (GeV/c2)

2004/8/17 ICHEP2004 7

Event Distribution in mX and q2

Event distribution in mX and q2 with the normalization fixed to the fitting with the coarse bin.

Inclusive B→Xu l

140fb-1

q2 distribution (mX>1.7GeV/c2)

Expected dist. of b→u b→c

-5 0 5 10 15 20q2(GeV2)

mX distribution (q2>8GeV2)

Expected dist. of b→c b→u

0 0.5 1.0 1.5 2.0 2.5 3.0mX (GeV/c2)

Subtracted with the Expected b→c bkg.

2004/8/17 ICHEP2004 8

Fraction of semileptonic leptons with Pl > 1.0GeV/c

Partial Branching FractionRelative partial branching fraction

( )

( )

1b us

sll

b u

slfrec

b ufrec

b

l ls l

u u

e

FB B X N

B B X N

b→ u selection efficiency: 0.274

Full recon. efficiency ratio:

Corrections: Unfolding the mX /q2 resol. 0.984±0.014

3( ) 0.99 0.15 0.18 0.04 0.07 10uB B X

( ) 0.1073 0.0028Br B X (PDG2004)

stat. syst. b→u b → c

Partial branching fraction

140fb-1 preliminary

mX < 1.7 GeV/c2, q2 > 8 GeV2

Inclusive B→Xu l

0.75±0.048

2004/8/17 ICHEP2004 9

|Vub| Determination

Extrapolation to the total br.

New determination of b-quark shape function (SF)

parameters from Belle’s B → Xs photon spectrum.

( ) ( ) /u u uB B X B B X f

0.294 0.044uf

+ Higher order correction (FN → BLL)+ Contribution from sub-leading SF + W.A.

Inclusive B→Xu l

A.Limosani, T.NozakiHep-ph/0407052

3( ) 3.37 0.50 0.60 0.14 0.24 0.50 10uB B X

from fu error1/ 2( ) 1.61

| | 0.004240.002

uub

B

B B X psV

+latest HQ parametersPDG formula

3| | 5.54 0.42 0.50 0.12 0.19 0.42 0.27 10ubV stat. syst. b→u. b → c. fu Br→|Vub

|140fb-1 preliminary

2004/8/17 ICHEP2004 10

Exclusive B→Xu l w/ Semileptonic Decay Tag

New method for clean extraction of

B→ l l decays.

Υ(4S)e- e+

Btag

Bsig

v

Tagging side Btag→D(*) l

Signal side Bsig→ l

v

D(*)

“Double semileptonic”decay.

2004/8/17 ICHEP2004 11

Analysis Method

* /tagB D D 0 0/D D

decay modesdecay modes

Tag side reconstruction

o have intersection,must be 20 1Bx

Kinematics of double semileptonic decay Back-to-back correlation of the two B constrains their direction to the intersection of the 2 cones.

Exclusive B→Xu l

sig uB X

0( ) 1, ( ) 1N N 0 or

Signal side reconstruction0.8 /P GeV c

2004/8/17 ICHEP2004 12

The ratio Nobs/Nexpected = 0.89 ±0.08 is used to correct the MC efficiency for l and l detection.

Calibration with Bsig→D*l Decays

Validity of the method for double semileptonic decay detection has been tested with *

sigB D 0D

K

M(K+--) on the signal side x2 distribution

Nobs=147±12

The method works !

Dots: DataHisto: MC

Exclusive B→Xu l

( Nexpected =165)

M(K) GeV/c2 x2

2004/8/17 ICHEP2004 13

Signal ExtractionWe extract l / l signals simultaneously by fitting 2D

(mX, x2) distribution.

Fitting components: l , l , other Xu l , BB background.

PDF’s are based on MC.

Constraint for extracted Br: Br( l +Br( l +Br(other Xu l )=Br(Xu l )

Exclusive B→Xu l

Fitting results for all q2 data.l (82±13)

l (65±20)other Xul

BB

x2 dist. for mass region (I)

x2 dist. for mass region (II)

I II

mX GeV/c2 x2 x2

ldecays are cleanly extracted !

140fb-1

2004/8/17 ICHEP2004 14

Extraction of q2 Distributionq2 distribution is extracted by fitting the (mX,x2) distribution for three q2 intervals.

Exclusive B→Xu l

△ UKQCD○Ball’01□ ISGW IIl+

△ UKQCD○Ball’01□ ISGW II

◇ Melikhov’98

l+

At the present accuracy, the obtained q2 dist. does not exclude any tested models

32.54 0.78 0.85 0.30 10totalB 31.76 0.28 0.20 0.03 10totalB

Extracted q2 dist.

FF-dep.

140fb-1

Preliminary

q2<8GeV2 q2:8-16GeV2 q2>16GeV2mX dist. for three q2 intervals

mX (GeV/c2)

2004/8/17 ICHEP2004 15

|Vub| from B0→ l+

|Vub| determined from B>16GeV2 (B0→ l+ ) with lattice QCD.

0.62 30.48(3.90 0.71 0.23 ) 10

0.320.121.92 0.47thy

Exclusive B→Xu l

( )ub

phy B

B BV

w/ quenched LQCD [FNAL/JLQCD/APE/UKQCD]

Average

4th error fromFF-dep. in Br is small for l data

w/ unquenched LQCD [FNAL/HPQCD]

Preliminary results reported at Lattice’04.

FNAL’04

HPQCD

1.96 0.51 0.39thy

1.31 0.33thy

0.85 30.51(3.87 0.70 0.22 ) 10

0.74 30.50(4.73 0.85 0.27 ) 10

140fb-1, preliminary thy

2004/8/17 ICHEP2004 16

Summary

We have obtained preliminary |Vub| results from

Inclusive B→Xu l w/ fully reconstructed tag.

Results are compatible and consistent

with the previous exp’s.

Exclusive B→Xu l (B0 → l )

w/ semileptonic decay tag.Results are compatible and consistent

with the previous exp’s and inclusive results.

New method for clean signal extraction

has been demonstrated.

These measurements are promising

approach at B-factories in the coming years !

Belle |Vub|

2004/8/17 ICHEP2004 17

Backup Slides

2004/8/17 ICHEP2004 18

Systematic Uncertainties

bu MC statistics = 4%model dep. = 4.2%

Detector sim. for bc & bu MC = 6.5%Correlated errors, added/subtracted in linear for the two MC’s

Subtracting:Lept. sel. = 1% clusters = 1%

Adding:Kaon ID = 6%tracking = 2%

Systematic error of Refreco and Re

lept ... 6.4%

bc MC statistics = 15%model dep. = 5%

Variate inclusive parameters within errors

Dl (D =1.19±0.19)D*l (A1=1.51±0.13)D**l

... 5.6%

... 3.9%

... 1.6%

Inclusive B→Xu l

2004/8/17 ICHEP2004 19

Systematic UncertaintyMajor contribution

D*lcalibration

Statistics of detected D*l8.3%

Error of Br(B0D*l) 4.3%

BB background shape

tested B(l) (B( l )) in MC by

removing charged track by 1%

-4.2(+23.5) %

removing by -1.1(+12.8) %

Replacing K with by 2%

-0.5(+16.6) %

Exclusive B→Xu l

Significant change in B( l ), due to the broad width of meseon.

2004/8/17 ICHEP2004 20

Four quenched calculations from FNAL, JLQCD, APE and UKQCD (used in CLEO2003).

Average

Preliminary unquenched results from FNAL and HPQCD

(reported at Lattice’04).

FNAL (M.Okamoto)

HPQCD(J.Shigemitsu)

B Form-Factor from LQCD

0.320.121.92 0.47thy Include 15%

quenching error

1.96 0.51 0.39thy

1.31 0.33thy

( given for q2>16GeV2 )thy

Exclusive B→Xu l

2004/8/17 ICHEP2004 21

FNAL’04Statistical ~10%

Lattice-continuum matching ~7%

Chiral extrapolation ~6%

Discritization ~5%

HPQCDStatistical ~5%

Higher oder operator matching ~9%

Chiral extrapolation ~5%

Relativistic+discritization ~5%

Error of FF from Unquenched LQCD

1.96 0.51( .) 0.39( .)phy stat syst

1.31 0.33[0.13( ) 0.30( )]phy stat syst

2

Number given by M.Okamoto (FNAL) and J.Shigemitsu (HPQCD) in private communications.

2

Exclusive B→Xu l