1. Exclusive Semileptonic B Decays and |Vub|: Experimental
4th Flavor and CP Violation Conference
Vancouver, BC
9th – 12th April 2006
Kevin Varvell
University of Sydney

2. Significance of |Vub| Measurement
Precision measurement of SM parameter, size not theoretically predicted
Constrain one side of the
Unitarity Triangle
Test consistency of SM
picture, particularly
between |Vub|/|Vcb| and
sin21 (sin2β)
Current precision
sin21 ~ 4%
|Vcb| ~ 2% (one recent update mentioned here)
|Vub| ~ 7% inclusive (talk this afternoon, E. Barberio)
>10% exclusive (main subject of this talk)
Aim is for ‹ 5% precision for |Vub|excl, compare with |Vub|incl
This side
determined
By |Vub|/|Vcb|
FPCP06 11-Apr-2006
K. Varvell

3. From Exclusive Semileptonic B decay to |Vub|
Measurements made in
Ideally want to measure rate
of tree level process Γ(b  x l )
Complicated by strong interaction
effects – form factors needed
Channel B  π l  is the most
promising for |Vub|
- one dominant form factor
q2 shape and normalization needed
e+e-→Υ(4S) → BB
f+(q2)
FPCP06 11-Apr-2006
K. Varvell

4. Determining the Form Factor f+(q2)
Earlier predictions made with quark models, e.g. ISGW2
Lattice QCD
makes predictions at high q2 (q2 > ~16 GeV2)
unquenched calculations have become available in recent times
e.g. HPQCD
FNAL
Light Cone Sum Rules
makes predictions at low q2 (q2 < ~14 GeV2)
e.g. Ball & Zwicky
Parametrization has traditionally been used to extend LQCD or LCSR to full q2 range
FF normalization main issue when extracting |Vub|
How much theory input needed for |Vub| measurement?
(R. Hill talk)
hep-lat
hep-lat
PRD 71 (2005)
hep-ph/
FPCP06 11-Apr-2006
K. Varvell

5. Approaches to Measuring B(B  π l )
Untagged
initial 4-momentum known
missing 4-momentum = ν
Reconstruct B  π l 
using mB (beam-constrained)
and ΔE = EB-Ebeam
Pros
High efficiency
Cons
ν resolution problematic
Rel. high backgrounds
(rel. low purity)
Semileptonic (SL) Tag
One B reconstructed in
a selection of D(*) l 
modes
Two missing ν in event
Use kinematic constraints
Lower backgrounds
(higher purity)
Rel. low efficiency
Full Recon Tag
One B reconstructed
completely in known
b  c mode.
Many modes used.
Very good ν resolution
Very low backgrounds
Very low efficiency
Rest used to reconstruct 
Signal
Signal
ExampleTag
side
Signal
ExampleTag
side

6. Range of Applicability of Methods
Scenario
Untagged
D(*) l  tag
Full recon tag
140
250
500
1000
fb-1
Boundaries indicative only
Full recon tag will ultimately become method of choice.
Currently
Belle Lint ~ 500 fb-1
BaBar Lint ~ 300 fb-1
FPCP06 11-Apr-2006
K. Varvell

7. CLEO Untagged B  π l  - result mBC ΔE (9.7M BB)
Phys. Rev. D 68 (2003)
(9.7M BB)
Search (and fit for yields) across a
number of exclusive Xu l ν modes
+,0,+,0,ω,η
Isospin constraints imposed in fit
Extract partial BF in 3 q2 bins
for + and +
mBC ΔE
- result
FPCP06 11-Apr-2006
K. Varvell

8. CLEO Untagged B  π l  (cont.)
(9.7M BB)
FF models enter extraction
of BF through
Signal efficiency from MC
b → u crossfeed and b → c
background modelling in MC
Phys. Rev. D 68 (2003)
Different FF
models
- result
B(B0  π- l+ ) = (1.33 ± 0.18 ± 0.11 ± 0.01 ± 0.07) x 10-4
stat
syst
FF sig
FF crossfeed
|Vub| = (3.17 ± ± 0.03) x 10-3
( and  mode
combined)
LCSR + LQCD (unq)
stat
FF shape ρ
syst
theo
FPCP06 11-Apr-2006
K. Varvell

9. BaBar Untagged B0  π- l+ 
Simultaneous extraction of modes +,0,+,0
Phys .Rev. D72 (2005)
79fb-1 (86M BB) ΔE
mBC
- result
Yield 396 -lν, 137 0lν
FPCP06 11-Apr-2006
K. Varvell

10. BaBar Untagged B0  π- l+  (cont)
Isospin constraints imposed
Extract partial BF in
5 q2 bins 
3 q2 bins ρ
“Marginal” agreement with ISGW2
Phys .Rev. D72 (2005)
79fb-1 (86M BB)
- result
B(B0  π- l+ ) = (1.38 ± 0.10 ± 0.16 ± 0.08) x 10-4
stat
syst
FF shape
|Vub| = (3.82 ± 0.14 ± 0.22 ± ) x 10-3
+0.88
-0.52
FNAL LQCD
stat
syst
FF shape
FF norm
FPCP06 11-Apr-2006
K. Varvell

11. Belle SL Tag B  π l+  Mx xB2 hep-ex/0604024 253fb-1 (275M BB)
Simultaneous extraction of modes +,0,+,0
xB2 must lie in range [0,1]
for signal events
Submitted to PLB Mx
xB2
Yield 156 -lν, 69 0lν
FPCP06 11-Apr-2006
K. Varvell

12. Belle SL Tag B  π l+  (cont)p- p0
hep-ex/
Extract partial BF in
3 q2 bins
B(B0  π- l+ ) = (1.38 ± 0.19 ± 0.14 ± 0.03) x 10-4
stat
syst
FF shape
B(B+  π0 l+ ) = (0.77 ± 0.14 ± 0.09 ± 0.00) x 10-4
stat
syst
FF shape
|Vub| = (3.60 ± 0.41 ± ) x 10-3
+0.62
-0.41
+ + 0, q2 ≥ 16 GeV2, FNAL
stat
syst
theo
|Vub| = (4.03 ± 0.46 ± ) x 10-3
+0.59
-0.41
+ + 0, q2 ≥ 16 GeV2, HPQCD
stat
syst
theo
FPCP06 11-Apr-2006
K. Varvell

13. BaBar SL Tag B0  π- l+  hep-ex/0506064 211fb-1 (232M BB)
Single mode - studied
cos2fB  [0,1] for signal events
Yield 61 -lν
FPCP06 11-Apr-2006
K. Varvell

14. BaBar SL Tag B0  π- l+  (cont)
hep-ex/
211fb-1 (232M BB)
Extract partial BF in
3 q2 bins
B(B0  π- l+ ) = (1.03 ± 0.25 ± 0.13) x 10-4
stat
syst
|Vub| = (3.3 ± 0.4 ± ) x 10-3
-0.4
+0.8
+0.8
HPQCD04
stat
syst
FF norm
FPCP06 11-Apr-2006
K. Varvell

15. BaBar SL Tag B+  π0 l+  hep-ex/0506065 81fb-1 (88M BB)
Single mode 0 studied
cosθB0l  [-1, 1] for signal events
Yield 45 0lν
B(B+  π0 l+ ) = (1.80 ± 0.37 ± 0.23) x 10-4
stat
syst
FPCP06 11-Apr-2006
K. Varvell

16. BaBar Full Recon Tag B  π l+ 
hep-ex/
211fb-1 (233M BB)
Tag : Fully reconstructed B  D decays (“Breco” tag)
Select using ΔE, mES (beam constrained B candidate mass)
mmiss2 distribution : peaks at zero for signal (single missing neutrino) p0 p-
mγγ
Yield 36 -lν, 34 0lν
FPCP06 11-Apr-2006
K. Varvell

17. BaBar Full Recon Tag B  π l+ 
hep-ex/
Extract partial
BF in 3 q2 bins p- p0
B(B0  π- l+ ) = (1.14 ± 0.27 ± 0.17) x 10-4
stat
syst
B(B+  π0 l+ ) = (0.86 ± 0.22 ± 0.11) x 10-4
stat
syst
B(B  π l+ ) = (1.28 ± 0.23 ± 0.16) x 10-4
Combined + + 0
stat
syst
|Vub| = (3.7 ± 0.3 ± ) x 10-3
-0.5
+0.8
+ + 0, HPQCD
stat
syst
FF norm
FPCP06 11-Apr-2006
K. Varvell

18. Summary of B(B  π l ) Exclusivep- p0
HFAG Winter 2006
Note: Updated Belle SL tag result
not yet incorporated in plot
Untagged analyses currently the
most precise
Average B(B0  π- l+ ):
(1.34 ± 0.08 ± 0.08) x 10-4
FPCP06 11-Apr-2006
K. Varvell

19. Summary of |Vub| from Exclusive B  π l 
HFAG Winter 2006
Experimental q2 shape input not used (yet)
Using FF calculated in region
of applicability
Using FF extrapolated to
full q2 range
FPCP06 11-Apr-2006
K. Varvell

20. Other B  Xu l  Exclusive Modes
Expt
Mode
BF [10-4]
Reference
CLEO
B0  ρ- l 
2.17± ±0.41
2.69± ±0.50
BaBar
2.57±0.52±0.59
B0  ρ- e 
3.29±0.42±0.47±0.60
2.14±0.21±0.51±0.28
Belle
2.17±0.54±0.31±0.08
B+  η l 
0.84±0.31±0.16±0.09
B+  ρ0 l 
1.33±0.23±0.17±0.05
B+  ω l 
1.3±0.4±0.2±0.3
Tag
+0.47
-0.54
Untagged
PRD 68 (2003)
+0.35
-0.47
Untagged
PRD 61 (2000)
Full
hep-ex/
Untagged
PRL 90 (2003)
Untagged
PRD 72 (2005) SL
hep-ex/
Untagged
PRD 68 (2003) SL
hep-ex/06xxxxx
Untagged
PRL 93 (2004)
HFAG Compilation, except with updated Belle SL result included
FPCP06 11-Apr-2006
K. Varvell

21. BaBar Exclusive |Vcb| Measurement
hep-ex/
Update to
PRD 71 (2005) submitted to PRD
Multi-Dimensional fit to helicity amplitudes for 2 (slope parameter) and R1, R2 (FF ratios), all functions of H+,-,0(w).
w = vB· vD*
79fb-1 (86M BB)
FPCP06 11-Apr-2006
K. Varvell

22. BaBar Exclusive |Vcb| Measurement (cont)
hep-ex/
Improvement in FF uncertainty
Reduces |Vcb| systematic error
Useful for understanding b → c
background in b → u
R1 = ± ± ± 0.027
R2 = ± ± ± 0.013
ρ2 = ± ± ± 0.035 w
cosθl
cosθV χ
|Vcb| = (37.6 ± 0.3 ± ) x 10-3
+1.5
-1.3
theo
stat
syst
FPCP06 11-Apr-2006
K. Varvell

23. Summary of Exclusive |Vcb| Measurements
HFAG Winter 2006
Note:
Does not include latest BaBar result from
hep-ex/
FPCP06 11-Apr-2006
K. Varvell

24. Summary and Future : Exclusive |Vub| Measurements
CLEO, BaBar and Belle
⇒ Xu BF measurements using several channels and several techniques
Xu : +,0,+,0,ω,η
⇒ |Vub| extracted using +,0 modes – Exptl error approaches 5%
- FF normalization error dominates
⇒ Xc BF measurements and |Vcb| continue to be important
⇒ future precision Xu measurements using Full Recon tagging
⇒ FF shape measurements from experiment now possible
Theory progress important (and happening) as well (R. Hill talk)
(P. Mackenzie talk)
|Vub| to better than 5% precision overall from exclusive semileptonic B-factory measurements? ⇒ Feasible.
FPCP06 11-Apr-2006
K. Varvell