Download BABAR Belle Belle

BdK0K0 and Other Hadronic bd Decays
Jed Biesiada
Princeton University
(for the BaBar Collaboration)
4th International Workshop on the CKM Unitarity Triangle
Nagoya, Japan
December 12, 2006
Virtual Loops: Low-Energy Probe of High Energy Physics




Virtual-loop, or “penguin”, diagrams are
sensitive to new physics contributions
Many models beyond the Standard Model
contribute additional CP-violating phases
Penguin modes are suppressed in the
standard model relative to the treedominated decays
Small branching fractions
Many interesting bs penguin decays
have been investigated at the B Factories

fK,
h’K0,
K0K0K0,
K00,
…
Standard Model
Supersymmetry
2
BdK0K0 : b → d Penguin



B0K0K0 is dominated by the bd
transition
Analogous to bs penguins in fKS
Even more suppressed in the SM:
2
(b  d ) Vtd
~
~ 4%
2
(b  s ) Vts


New physics might affect bd
differently than bs
Same penguin as in B0+- and
B+K0K+

Useful for extracting the UT angle 
in the former and the annihilation
contribution in the latter
d
K
BK0K0 : Expectations from the Standard Model

Assuming top-quark dominance:


BF ~ 10-6
Decay weak phase exactly cancels
mixing phase
d
s
s
q A  Vtb*Vtd  VtbVtd* 
 *   1

 
* 
p A  VtbVtd  VtbVtd 

Expect zero indirect CP violation



K0
R. Fleischer and S. Recksiegel,
S(K0K0) = 0
Eur.Phys.J.C38:251-259,2004
→ any non-zero value would indicate
2
sin
2
α

2
r
cos
δ
sin
(
2
β

γ)-r
sin 2 β
New Physics
S KK (SM) 
2
Expect a small contribution from uand c-penguins
CKK (SM ) 

K0
S(K0K0) < 0.10, sensitive to
combination of all three UT angles
Predictions (using QCD FA):
1  r  2r cos δ cos γ
1
Pct
 2r sin  sin 
r
 2  h 2 Put
1  r 2  2r cos  cos 
0.02  S KK (SM)  0.13
 0.17  C KK (SM)  0.15
Evidence for B→K0K0 (ICHEP 2004)
N ( K S0 K S0 )  23.0 76..77
1.9
 2.0
(4.5 )
40
( B 0  K 0 K 0 )  (1.19 00..35
 0.13) 10 6
PRL 95: 221801, 2005 (227 million BB pairs)
N ( K S0 K S0 )  15.6  5.810..16 (3.5 )
( B 0  K 0 K 0 )  (0.8  0.3  0.1) 10 6
PRL 95: 231802, 2005 (275 million BB pairs)
Next Step: Observe the mode and measure CP asymmetries with ~350-450 million BB pairs
5
Measuring SKK
Challenge: Need to vertex a decay with no primary tracks from the e+einteraction point
Solution: Exploit precise knowledge of the interaction point and fit the
entire (4S) decay chain using beam-spot constraints.
Method developed by BaBar, described in PRL 93: 131805, 2004 and
PRD 71: 111102, 2005
+
1. Constrain Brec decay
y
to beam-spot in x-y
z
Ks
B0
e  4s 
B
e+
Beam
0
2. Use a neural-net
algorithm to determine
the Btag flavor
Ks
Btag- Standard Method
3. Determine the proper-time
difference t between Brec and
Btag decay vertices
6
Use KS’s that decay in the Silicon Vertex Tracker
Classes of KS decays:
35%
25%
40%
Have two KS‘s but need only one to vertex the signal B
Class I – both pions have
hits in inner layers
Class II – not Class I, both
pions have hits in SVT
Class III and IV: not used for
time-dependent measurement
Class I B0’s: 58%
Class II B0’s: 26%
Class I
Class II
7
Experimental Method: find the signal…
Kinematics
e

B
*2
mES  Ebeam
 p*B2
BK0K0KSKS
B
Candidate KS mass spectrum
KS+-
e
pB* ~ 300 MeV / c
*
E B*  Ebeam
KS mass selection
*
E  EB*  Ebeam
Main Background is
“Continuum”
e+e- qq
Background Suppression
Background
Use event-shape variables to separate
from signal
Need rejection of >106, but a fake
signal event must have two highmomentum KS’s, pLAB  3 GeV
 low-background mode
Signal
Analysis Overview
Common Features


BK0K0 reconstructed in K0K0KSKS and KS+Loose selection cuts



KS mass and lifetime significance ( or vertex displacement)
mES( or mbc) – E fit region
Extended unbinned maximum-likelihood fit
Differences
Dataset
Vertexing
BaBar
Belle
347 million BB pairs (316 fb-1)
449 million BB pairs (409 fb-1)
beam-constrained
none
Cut on likelihood ratio
Event-shape continuum
suppression
Efficiency (including
secondary BFs)
Fit variables
Physics quantities in the
fit


Loose cut on sphericity angle
Optimized Fisher discriminant used in the fit
(Legendre moments)

Modified Fox-Wolfram moment (Fisher)
B polar angle

Optimization dependent on tagging
quality of the other B

8.5  0.3%
5.9%
mES , E, Fisher, t
mbc , E
signal and background yields, time-dependent
CP-violating parameters S and C
signal and background yields, partialrate flavor asymmetry, ACP
Unbinned Maximum Likelihood Fit and Discriminating Variables for BaBar
ni = candidate category, signal or background
Pi = probability density for category i
N = number of events
Kinematic
Event-shape
mES
Fisher discriminant
E
F  a0  pi*  a2  pi* cos 2  i*
i
*
i
i
    pB* , pi* 
Solid histogram
=
Dashed histogram =


Signal
Background
t PDFs
B0tag
B0tag
MC
Fit result closely
tracks the
generated CPviolating structure
Data
Resolution function
characterizes the data well
11
Results: Observation of bdg Penguins
PRL 97: 171805, 2006
BABAR:
BABAR
N ( K S0 K S0 )  32  8  3 (7.3 )
347 M BB
( B 0  K 0 K 0 )  (1.08  0.28  0.11) 10 6
Dominant systematic uncertainties:
 Fitter bias
 Uncertainty in PDF shapes in the fit
BABAR
M.Pivk and F.R.Le Diberder,
“sPlot: A Statistical Tool to Unfold Data Distributions,”
Nucl. Instrum. Meth. A 555, 356 (2005)
Belle:
Belle
Belle
hep-ex/0608049
N ( K S0 K S0 )  23  6  3 (6.0 )
449 M BB
( B 0  K 0 K 0 )  (0.86  0.23  0.09)  106
Dominant systematics:
Cuts on variables not shown to enhance signal
 KS reconstruction
 Event-shape likelihood selection
Results: CP Asymmetry
Likelihood Projection Plots PRL 97: 171805, 2006
n Contours
physical region
BABAR
1 = 68% 2-D C.L.
2 = 95% 2-D C.L.
3 = 99.7% 2-D C.L.
fit result
S ( K S0 K S0 )  1.2800..80
73
0.11
 0.16
C ( K S K S )  0.40  0.41  0.06
0
0
Belle:
ACP ( K 0 K 0 )  0.57 00..72
6513 0.13
B+→K0K+
Penguin + Annihilation Amplitudes



Same penguin amplitude as in B0K0K0
Annihilation contribution may affect branching fraction
Need comparison with B0K0K0 to estimate the size of this effect
K+
K0
15
Analysis Overview



B+K0h+ reconstructed in KS+-, h=K, 
No vertexing
BaBar PID



Belle KID:



Cut on a K- likelihood ratio to form separate
candidate samples for each species
Constrain crossfeed and signal yields in the two
samples in a simultaneous fit
Pion mass is assumed for the track


Use the Detector of Internally Reflected Cherenkov
light to separate pion and kaon bachelor tracks in the
fit
DIRC model is the same as in the BK+-/+analysis
Additional PID from E, where the KSK+ peak is
displaced -45 MeV relative to the KS+ peak
Efficiency:
BaBar: 12.9  0.4 % KS+, 12.6  0.4 % KSK+
Belle: 12.2 % KS+
6.7 % KSK+
BaBar Particle ID: DIRC
cos C 
1
n
D*  D0  , D0  K  
Results
BABAR:
Belle:
347 M BB
449 M BB
N K 0   1072  46 32
37
N K 0   1252 41
39
N K 0 K   71  19  4 (5.3 )
N K 0 K   36.6 89..37 (5.3 )
( B   K 0  )  (23.9  1.1  1.0)  10 6
( B   K 0  )  (22.9 00..78  1.3) 10 6
( B   K 0 K  )  (1.61  0.44  0.09)  10 6
33
( B   K 0 K  )  (1.22 00..28
AK 0   0.029  0.039  0.010
AK 0   0.03  0.03  0.01
AK 0 K   0.10  0.26  0.03
23
AK 0 K   0.1300..24
 0.02
S
S
S
S
S
S
 0.13
 0.16
S
S
Dominant systematic uncertainties:
Dominant systematics:
 Uncertainty in PDF shapes in the fit
 Cross-feed
 KS reconstruction
 Cross-feed constraint
17
) 10 6
Projection Results
BABAR
KS+
KSK+
KS+
KSK+
KSKS
Belle
KSKS
Belle
PRL 97: 171805, 2006
hep-ex/0608049
18
B0→K+K-
W Exchange Amplitude



B(B0→K+K-) ~ (0.78) x 10-8 in the
standard model
Beneke and Neubert, Nucl. Phys. B
675, 333 (2003)
Rescattering or new physics could
enhance the branching fraction
Yield extracted from the B0K+/+- fit


Use DIRC and E to separate from
the other two components
K+K- peak in E lies on the low tail
of the large K+- peak
Difficult to measure
KK+
+K+K+K- 40
20
Results
Belle:
BABAR:




227 million BB pairs
3±13±7 events
B(B0→K+K-) < 0.40  10-6
(90% confidence level)
Submitted to PRD




449 million BB pairs
2.5 53..07 events
B(B0→K+K-) < 0.25  10-6
(90% confidence level)
hep-ex/0608049
21
Summary

Observations of B0K0K0 and B+ K0K+, dominated by the b → dg
penguin amplitude



Non-observation of B0 K+K- is so far consistent with the standard model


First observations of BdKK
Branching fractions at SM predicted values
The only Bhh’ (h=K,) charmless mode left to be observed
First time-dependent CP measurement in a b → d penguin


Large positive values of S are disfavored
Method is feasible at BaBar



Opens a new area of enquiry for the B factories
More data will provide stronger constraints in this new sector
K*K modes not yet seen but will be available at  1 ab-1
22
State of the Art
Backup Slides
Im(23)RR
Flavor-Changing New Physics in Theory and Experiment
M. Ciuchini and L. Silvestrini,
Phys. Rev. Lett. 97 (2006) 021803
bs
Im(13)LL
Re(23)RR
M. Ciuchini et al.,
hep-ph/0512141
bd
The experimental program is
just beginning:
B  
B  KK …
Re(13)LL
Need to explore the bd sector
and add another constraint
BK0K0 : Example of a Potential New Physics Scenario
d
d
d
A. Giri and R. Mohanta, JHEP 11, 084 (2004)
Predictions for
various assumptions
on the weak phase
NP and strong
phase NP between
NP amplitudes:
 NP  4 , 3 , 2 , 
0   NP  2
Note: this scenario does not include all experimental constraints
New-Physics Predictions for TDCP



A. Giri and R. Mohanta, JHEP 11, 084 (2004)
Predictions for various assumptions
on the weak phase NP and strong
phase NP between NP amplitudes
Depending on the NP phase, SKK
(NP) could be large
Current BF measurement consistent
with NP scenario


 NP  4 , 3 , 2 , 
0   NP  2
New physics in b → d penguins is
highly constrained assuming threegeneration unitarity
But there’s still room for NP

Measure TDCP in b →d penguins for
the first time and add another
constraint
Vertexing Results
Signal MC
28
KSKS PDFs
double Gaussian
Argus
double Gaussian
bifurcated Gaussian
1-degree polynomial
double Gaussian
Background Parameters Floated
29
KSh+ PDFs
Cruijff
double Gaussian
bifurcated Gaussian
Argus
1-degree polynomial
double Gaussian
30
KK Vector Decays

Useful for constraining sin2 in vector bs decays (e.g. BfK0)



Not yet seen
Require Dalitz-plot window selection



Grossman, et al., Phys. Rev. D 68, 015004 (2003).
Loss of efficiency
Full Dalitz plot analysis with more data
Single vector modes expected at ~ 0.5  10-6 level (bigger if
NP/rescattering effects are sizable)
Current upper limits (HFAG):





K*0 K0 < 1.9  10-6
K*0 K+ < 5.3  10-6
K*0 K*0 < 22  10-6
K*+ K*0 < 71  10-6
K*+ K*- < 141  10-6
within reach of the B factories
31
Future Outlook with 1 ab-1

15% statistical error on B0K0K0 and B+ K0K+ BF’s from each
experiment





Meaningful comparison with theoretical predictions
Constraint on annihilation contribution in B+K0K+
Time-dependent measurement from Belle?
S ~ 0.40, C ~ 0.25 from each experiment
Waiting for the first K*K signals with 1ab-1
32