Transcript Rare and Semileptonic Decays of B and K Mesons

Radiative and Electroweak Penguin
Decays of B Mesons
Jeffrey D. Richman
University of California, Santa Barbara
BABAR Collaboration
11th International Conference on B Physics at Hadron Machines
Oxford, Sept. 28, 2006
Outline

Overview: a little history, physics goals, and challenges.

B+r+g, B0r0g, B0wg and measurement of |Vtd/Vts|



BK l+l- and BK* l+l-: search for new physics using the
lepton forward-backward asymmetry
Inclusive BXs g: branching fraction measurements and
extraction of heavy-quark expansion parameters from the
Eg spectrum.
Conclusions
My apologies for not covering all results on
radiative/electroweak penguin decays in this talk!
Radiative penguin decays of B mesons
Observation of BK* g
CLEO II (1993): Loops in B decays!
Now it’s a physics program!
PRL 71, 674 (1993): cited >500 times!
B(106 )
B( B  K *g )
 4  105
Rare, but not all that rare!
M ( K *g )
What can we learn from bs, d transitions?

Flavor-changing neutral currents probe SM at 1-loop level.
g
b u, c, t
d
W
s, d
d
(dominated by t quark)

b
g ,Z

s
u, c, t
d
W
d
(+ W+W- box diagram)

New physics can affect the amplitudes at leading order!

As for bc or bu semileptonic decays, the amplitude in
EM/EW penguins is factorizable (only one hadronic current).
What can we learn from bs,d transitions?
b
d


g
Vtb
W
t

*
td
V
d
d
b Vub
d
W



u
d
Presence of only single hadronic current allows us to isolate
non-perturbative QCD parameters in well-defined way. Can
be related to same parameters for other decays.

Exclusive decays: decay form factors fi(q2). bs transition
is similar to bu (heavy to light)

Inclusive decays: parameters of heavy-quark expansion
(mb, mp2,…)
Can extract information on CKM elements if info on hadronic
parameters is available from data, theory, or both.
Observation of bd g and Measurement of |Vtd/Vts|
b
B

u
g
g
u, c, t
b u, c, t
d
W
u
r

B
0
W
d
d

d
r ,
w
0
+ W annihilation diagram (small)
B( B  rg )
Vtd

*
B( B  K g ) Vts
2
m  m 
m  m 
2
B
2
B
*
2 3
r
2 3
K*
T1K (0)
 r
 1.17  0.09
T1 (0)
2
 T (0) 

 1  R 
 T (0) 
r
1
K*
1
1/ 
2
R  0.1  0.1
Ali, Lunghi, Parkhomenko,
PLB 595, 323 (2004)
Ball and Zwicky, JHEP 0604, 046 (2006)
  B   r g   2  B 0  r 0g 
2  B 0  wg 
I-spin (r), quark model (w). Expect small I-spin violation: (1.1+/-3.9)%.
Measurement of bd g Decays (Belle)
Belle, PRL 96, 221601 (2006); 386 M BB.
Nsig  8.5
Signal
continuum background
Nsig  20.7
N sig  5.7
Nsig  36.9
BK*g
Good particle ID is critical
in this measurement to
suppress BK* g feeddown.
Measurement of bd g Decays (BABAR)
BABAR, hep-ex/0607099, 347 M BB
projections of 4-D fit
signal +
bkgnd
bkgnd
B+r+g
Nsig  42.414.1
12.6
B+r+g
B0r0g
B0r0g
Nsig  38.710.6
9.8
signal
Comparison of bd g Branching Fractions
CKM fitter includes CDF
Bs mixing result.
Error on CKM Fitter
prediction includes uncert.
on BVg form-factor
ratio.
I-spin consistency?
(106 )
Mode
BABAR (10-6) (6.3 s signif.) Belle (10-6) (5.1s signif.)
preliminary; hep-ex/0607099
PRL 96, 221601 (2006).
B   r g
1.060.35
0.31  0.09
B0  r 0g
0.770.21
0.19  0.07
B0  wg
 0.84 (90% C.L.)
0.09
1.170.35
0.310.08
0.14
0.580.34
0.310.10
1.01  0.21  0.08
0.10
1.320.34
0.310.09
B  ( r  , r 0 , w) I -avg g
0.12
0.550.43
0.370.11
Extracting |Vtd /Vts| from bd g Decays
Belle, PRL 96, 221601 (2006).
Vtd
0.018
 0.1990.026
0.025 0.015
Vts
courtesy M. Bona (UTfit collab.)
BABAR, hep-ex/0607099
(preliminary)
Vtd
0.017
 0.1710.018
0.021 0.014
Vts
CDF, hep-ex/0606027
(preliminary)
Vtd
0.008
 0.2080.001
0.002 0.006
Vts
Consistent within errors.
Theoretical uncertainties limiting both approaches.
BKl+l- and BK*l+l- in the SM and Beyond
Photon penguin
g
b
d
Z penguin


u, c, t
W
u, c, t
W
d
d

W+W- box
b
Z
b
s
W

s
d


W

s
u, c, t
d
d
• Dependence on kinematic variables in 3-body decays can be used to
study the different amplitudes and their interference effects.
• The mode BKl+l- is allowed as well as BK*l+l- (BKg forbidden
by conservation of angular momentum).
Amplitude for BK*l+lM (B  K
*  

GF EM *
)
VtsVtb C9eff K * s g m PLb B
2p
mix of Z-penguin,
W+W-
box
Kruger and Matias; PRD 71, 094009 (2005)
photon penguin
dom. at v. low q2

mb eff
*

2 2 C7 K sis m q PRb B   g m
q

 C10 K * s g m PLb B

g mg 5

Short-distance physics encoded in Ci’s (Wilson coefficients);
calculated at NNLO in SM:
Ali et al., PRD 61, 074024 (2000)
C7eff 0.3 C9 +4.3 C10 4.7
C9, C10 generate asymm. in lepton angular distribution over most of q2.
Ci’s can be affected by new physics, which enters at same order as SM

Form Factors and Observables
Long distance QCD physics is mainly described in terms of form
factors, which are functions of q2  ( p   p  )2
• 4 semileptonic form factors: A1, A2, V, A0 (similar to BD*l, Brl
• 3 penguin form factors: T1, T2, T3
Form factor uncertainies  35% uncertainty in rate predictions.



K

K
B

K*
d
d
d
cos


0 d cos
1 d cos d cos
1
AFB 
p
0
d
d
d
cos


0 d cos
1 d cos d cos
1
0
( s  q2 )

dAFB
mm
 C10 Re  C9eff VA1 + b B C7eff
ds
s
0.16
2
s0   4.07 0.13
GeV


 mK *
VT2  1 
mB



 mK *   

ˆ
(1

m
)

AT
1


 
1 1
K*
mB   



Precise SM prediction due to ff cancellation.
Predictions for AFB in BK*l+l-: SM and beyond
q q
2


q
2
q2  qmax
2
min
s

K*
K* s
q
C7eff  C7 (SM)
Standard Model
C9eff C10eff  C9 (SM)C10 (SM)
0.16
s0   4.07 0.13
 GeV 2
C7eff  C7 (SM), C9eff C10eff  C9 (SM)C10 (SM)

BKl+l- and BK*l+l- : q2 distributions
B  Kmm
B  K *m  m 
J/yK
SUSY models
Pole from K*g, even in m+m-
y(2S)K
SM nonres
SM nonres
q2
q2
constructive interf.
destructive
BKl+l- and BK*l+l-: the J/y veto




The decays BJ/y K and BJ/y K* are huge backgrounds and
must be carefully removed (also By(2S)K, y(2S)K*).
These backgrounds are restricted in q2, but there is a tail due to
bremsstrahlung in the electron modes.
But BJ/y K and BJ/y K* are valuable control samples; use
them to study efficiency of almost any analysis cut.
Ali, Kramer, Zhu: B( B  K *
 
J/y and y(2S) veto: MC BKe+e-
m
7
;1  q2  7 GeV2 )  (2.920.67
)

10
0.61
m(e+e-) projection: MC BKe+e-
BKl+l- Signal from BABAR
BABAR, PRD 73, 092001 (2006)
9.8
Nsig  45.5-8.9
• summed over all K l+l- modes (K+e+e-, K+m+m- KS e+e-, KS m+m-)
• significance 6.6 s; rarest observed B decay
B( B  K
 
)  (0.34  0.07  0.02) 106
(averaged)
BK*l+l- Signal from BABAR
BABAR, PRD 73, 092001 (2006)
13.7
Nsig  57.1-12.5
B( B  K *
 
6
)  (0.780.19

0.11)

10
0.17
229 M BB
BK(*)l+l- Signals from Belle
Belle, PRL 96, 251801 (2006)
386 M BB
Nsig  96.0  12.0
Nsig  113.6  13.0
(data sample used for study of Wilson coefficients)
BKl+l- and BK*l+l- branching fractions
BABAR (10-6)
Mode
PRD 73, 092001 (2006)
BK
B  K*
 
 
Belle (10-6)
preliminary
0.340.07
0.07  0.02
0.5500.075
0.070  0.027
0.780.19
0.17  0.11
1.650.23
0.22  0.11
BK*l+l-: BABAR results on K* polarization and AFB
1 d
3
3
 FL cos2  K  (1  FL )sin 2  K
 d cos  K 2
4
use in 2 bins
of q2
BABAR, PRD 73, 092001 (2006)
K* polarization
Data
SM
C7  C7 (SM)
Polarization consistent with SM, but doesn’t discriminate
against new physics scenarios with current data sample.
Theory predictions in graphs: Ali et al., PRD 66, 034002 (2002); Ball and Zwicky, PRD 71, 014029 (2005).
BK*l+l-: BABAR results on AFB and L
1 d
3
3
2 *

F
sin


(1  FL )(1  cos2 l* )  AFB cos l*
L
l
*
 d cos l 4
8
use in 2 bins
of q2
Data
C7  C7 (SM)
SM
C9eff C10eff  C9 (SM)C10 (SM)
C7eff  C7 (SM),
C9eff C10eff  C9 (SM)C10 (SM)
excluded at 3.6s!
Any AFB<0 excluded at >2.7s
q2 range (GeV2)
0.1  8.41
 10.42
AFB  B  K
 
AFB
 0.19 (95% C.L.)
  0.15
0.21
0.23
0.720.28
0.26  0.08
 0.08
q
2
FL
0.770.63
0.30  0.07
0.510.22
0.25  0.08
 0.1 GeV 2  (AFB=0 in SM and many BSM)
Belle results on AFB for BK(*)l+lB  K*
SM: A7  0.33, A9  4.07, A10  4.21
 
A7   A7 (SM)
Standard Model
A7  0.28, A9  2.22, A10  3.82
A7  0.28, A9  2.42, A10  3.82
AFB  B  K *
AFB  B  K
  0.50  0.15  0.02
  0.10  0.14  0.01
q
 
 
Belle, PRL 96, 251801 (2006)
2
Belle results on Wilson coefficients for BK(*)l+l• fix |A7| to SM (BXs g)
• fit for A9/A7 and A10/A7
• data consistent with SM SM
• quadrants II, IV allowed
fit
A10>0
SM
12.3
12.8
1400 
A9 A10
 26.4
2
A7
95% C.L.  A9 A10 < 0 excludes quadrants I,III
at 98.2% C.L.
Inclusive B Xs g
• Canonical process for studying bs transition. Theory uncertainties
currently at 10% level (NLO); pushing toward 5% (NNLO).
• Huge theoretical effort to predict branching fractions & photon
energy spectrum.
• Branching fraction measures |C7|; spectrum is insensitive to new
physics but is sensitive to mb and Fermi motion of b-quark (“shape
function”).
Eg  1.6 GeV
T. Hurth, E. Lunghi, W. Porod, Nucl. Phys. B 704, 56 (2005).

B( B  X sg )  3.61 0.24
0.40
mc / mb
ACP ( B  X sg )  (0.42  0.08mc / mb  0.03CKM
0.15
0.08 scale
M. Neubert, Eur. Phys. J. C 40, 165 (2005).

B( B  X sg )  3.47 0.33
0.41

 0.02CKM  0.24param  0.14scale 104
0.32
pert 0.29 param

104
)%
Inclusive B Xs g: some history
CLEO, PRL 74, 2885 (1995); 2.01 fb-1 on Y(4S), 0.96 fb-1 below Y(4S)
Backgrounds: B decays, continuum, e+e-qqg (ISR), e+e-qqp0X
“Event-shape analysis”
“B-reconstruction analysis”
total background
(points w/error bars)
scaled off resonance
Eg
B( B  X sg )  (2.32  0.57 (stat.)  0.35 (sys.)) 104
Challenges of inclusive B Xs g
• Weak experimental signature: single high-energy photon + eventshape cuts. Lots of background from p0’s and h’s! Fully inclusive
analysis is not able to exploit the kinematic constraints (mB, E).
• Difficult to carry analysis down to Eg < 2.0 GeV.
• Want to push toward 5% precision to match the expected precision
of NNLO calculations. (It’s amazing that you can do this analysis
at all!)
• Two methods have evolved from initial CLEO approaches.
Method
Fully inclusive
don’t reconstruct Xs
Advantages
Closest correspondence to
inclusive B(BXs g).
Sum of exclusive Less background due to
BK n(p g
additional kinematic
constraints. Better Eg
resolution.
Disadvantages
Large background; limited
sensitivity at low Eg .
More model dependence due to
finite set of explicitly
reconstructed BXs g decays.
Fully inclusive B Xs g: pushing down the
energy threshold
CLEO, PRL 87, 215807 (2001), 9.1 fb-1
Belle, PRL 87, 061803 (2004), 140 fb-1
Belle, hep-ex/0508005
4
BF  (3.21  0.43  0.270.18
)

10
0.10
0.11
4
BF  (3.55  0.320.30
)

10
0.31 0.07
Measure for Eg>2.0; extrap. to Eg>0.25 GeV Measure for Eg>1.8 GeV; extrap. to full
Fully inclusive, lepton-tagged B Xs g (BABAR)
• Want to suppress large continuum background.
• Strengthen signature for signal by using decay of 2nd B in event.
• Require high energy lepton: pe*>1.25 GeV, pm>1.9 GeV in addition
to event-shape cuts.
• Tag does not compromise inclusiveness of Xs selection.
BB
dom.
qq + ττ
1.9  Eg*  2.7 GeV
BB
(blind)
XSγ
contin.
dom.
lepton tag from 2nd B meson
BABAR Fully Inclusive B Xs g, w/lepton tag
hep-ex/0607071 (preliminary, submitted to PRL)
88.5  106 BB events
Spectrum from
best fit to kinetic scheme.
Spectrum from
best fit to shape function
scheme.
not efficiency corrected
B( B  X sg )  (3.67  0.29  0.34  0.29) 104 Eg  1.9 GeV (measured)
(extrapolated,
B( B  X sg )  (3.94  0.31  0.36  0.21) 104 Eg  1.6 GeV
kinetic scheme)
BABAR B Xs g with Sum of Exclusive Final States
Reconstruct 38 exclusive modes
K (  ,,0)
( 4p )
• |E|<40 MeV
• Fit mES distrib. in bins of
m(Xs)
• Correct for efficiency of
each mode and missing
modes fraction (model
dependence)
summed over all m(Xs):
K (  , ,0)
h
( 2p )
3K (  ,,0)
( 1p )
comb. BB
continuum
peaking bknd
signal
BABAR B Xs g with Sum of Exclusive Final States
BABAR, PRD 72, 052004 (2005)
K*(890)
Energy Range Branching
Fraction (10-4)
0.04
Eg >1.9 GeV 3.27  0.180.55
0.40 0.09
0.04
3.35  0.190.56
0.41 0.09
Eg >1.6 GeV
(extrapolated)
• averages over two shape-function schemes
• errors: stat, sys, variation of shape fcn params
K*(890)
Value (GeV or GeV2)
Eg Moments
2.321  0.0380.017
0.038
Eg
Eg  Eg
2
2
0.0253  0.01010.0041
0.0028
• Eg (min) = 1.897 GeV
Summary of B Xs g Branching Fraction Measurements
B( B  X sg )  (3.55  0.24
0.09
0.10
4
 0.03) 10
HFAG
average
Extraction of heavy-quark expansion
parameters from BXs g
Using heavy-quark expansion (HQE), moments of inclusive B decay
distributions can be expressed in terms of non-perturbative QCD
parameters and quark masses.
• BXs g inclusive Eg spectrum
• BXc l  inclusive El spectrum and M(Xc) hadron mass distrib.
• mb now determined to about 1% and |Vcb| is determined to <2%.
Eg
Eg  Eg
2
mb

2
2
 g (mb , mp2 ,...)
(kinetic energy squared of b-quark)
Fits to moments of inclusive BXc l  and BXs g
distributions
Buchmüller and Flächer, PRD 73, 073008 (2006);
Data from BaBar, Belle, CDF, CLEO, & DELPHI
mp
Vcb
2
(GeV 2 )
b  sg
all moments
(10 3 )
all moments
bc 
bc 
mb
mb
kinetic mass scheme
mb  (4.590  0.025exp  0.030HQE ) GeV mp2  (0.401  0.019exp  0.035HQE ) GeV2
Vcb  (41.96  0.23exp  0.35HQE  0.59sl ) 103 mb used for |Vub| (7.5% error!)
Conclusions
Studies of radiative/electroweak penguins have moved far beyond
BK*g.

Observation of exclusive bd g decays: B(r0, r+, w) g

Use to extract |Vtd/Vts|; consistent with value from Bs mixing.
Precision limited by theoretical uncertainties.

Electroweak penguins decays BK l+ l-, BK* l+ l-, and
BXs l+ l- have been measured. First studies of decay
distributions have been performed and exclude some non-SM
scenarios. Much more data needed to exploit full potential.


Inclusive BXs g measurements provide information on mb
and non-pert. QCD parameters and help improve precision on
|Vcb| and |Vub|. Difficult issues with systematic errors, but goal si
to achieve 5% uncertainty on branching fraction.
Much more to learn about penguins; we will study them for
many years to come at BaBar, Belle, and LHC-b!
Backup slides
BK*l+l-: K* polarization vs. q2
SM
M(l+l-) distributions from BJ/y K+ control
samples: data vs. Monte Carlo
BABAR
points: data
histogram: MC
absolute normalization
Bremsstrahlung tails well described by MC.
Lepton angular distribution in l l rest frame

l
l
sq m 
2
2
l

B
s
q
K
*
use l- if B
use l+ if B

dAFB
 C10 Re  C9eff VA1 
ds
mbmB eff
+
C7
s

 mK *
VT2  1 
mB



 mK *
 (1  mˆ K * )  AT
1 1 1 
mB


 T2 ( s0 )  mK *
s0
mb
C7eff


1 
2
eff
mB
mB Re  C9  s0    A1 ( s0 ) 
mB
0.16
2
s0   4.07 0.13
GeV

 T1 ( s0 )  mK *

1 
mB
 V ( s0 ) 
Ali, Kramer, Zhu, hep-ph/0601034
 

 
  
 
  
BK*l+l- Dalitz plot
Can see AFB behavior and q2 dependence from the Dalitz plot
q2
AFB >0
effect of g pole
1
cos 
1
AFB  0 at q2
4 GeV 2
AFB  0
E
Note: BKl+l- is expected to have very small AFB, even in presence of
new physics; effectively provides a crosscheck.
Extracting AFB and FL in bins of q2
1 d
3
3
2
 FL cos  K  (1  FL )sin 2  K
 d cos  K 2
4
1 d
3
3
2 *
2 *
*

F
sin


(1

F
)(1

cos

)

A
cos

L
l
L
l
FB
l
*
 d cos l 4
8
BABAR, PRD 73, 092001 (2006)
q2  0.1 GeV2
q  0.1 GeV
2
2