The Discovery and Study of B Mesons
in the
CLEO Experiment
Jeffrey D. Richman
University of California, Santa Barbara
Symposium Celebrating CLEO and CESR, 31 May 2008
Outline

Terra incognita: Mapping the unknown territory of B
physics in the CLEO I era

CLEO II & CLEO II.V: the right stuff


The march of the penguins

The elegant simplicity of semileptonic decays

The triumph of hadronic decays
The success of the CLEO program as a scientific
enterprise
My apologies for not covering all the important measurements
and papers. I had to leave out a lot of them!
Terra Incognita: Mapping the New
Territory of B Physics in the CLEO I Era
[1977: discovery of Y states at FNAL]
courtesy Karl Berkelman
Physics themes of the early CLEO I era


Y(1S), Y(2S), Y(3S), and... discovery of the Y(4S)! (1980)

large decay width

s(Y(4S)) vs. s(continuum)
“Evidence for New Flavor Production at the Y(4S)”
(1981)

Y(4S) as a fountain of BB pairs  can study weak decays!

Inclusive single-lepton final state as signature of weak
decay; issues of continuum background & event shape
variables; off-resonance running

Inclusive properties of B Decay

Critical milestone: observation of fully reconstructed hadronic
decays (1983)
1.09 pb-1 (scan)
s (nb) vs. ECM (GeV)
Scan 10.46—10.64 GeV
 Mass: M(1S)+(1112 +/- 5) MeV

ECM
R2<0.3
M(4S) = (10572 +/- 5 MeV)
PDG: M(4S) = (10579.4 +/ 1.2 MeV)

Width: G = (19.9 +/- 5.5 +/- 5 ) MeV

s(res)/s(non-res) = 1/3

Event shape: more spherical than jetlike; R2 used from the beginning!
2.5 pb-1 (scan)
B( B  Xe )  (13  3  3)%
PDG: B( B  Xe )  (10.78  0.18)%
pe > 1 GeV/c
ECM
76 electron events total!
Predicted spectrum
assuming D*e/De =1
Inclusive properties of B decays (1981-1983)

Decay of b-flavored Hadrons to Single-Muon
and Dimuon Final States (1981)

Decay of B mesons into Charged and Neutral
Kaons (1982)

Charged-Particle Multiplicities in B-Meson
Decay (1982)

Semileptonic Decay of B Mesons (1983)

Ruling out Exotic Models of b Quark Decay
(1983)

Observation of Exclusive Decay Modes of bflavored Mesons (1983)

D0 spectrum from B-Meson Decay (1983)

Observation of Baryons in B-Meson Decay
(1983)
PRL 51, 634 (1983)

M ( )
M ( K  K   )
 sideband
40.7 pb-1
“Until now, the b-flavored mesons themselves had not been found. Here we
report that discovery.”
Signal region
2 evts
5 evts
5 evts
6 evts
D0 sidebands
M(B0)
5274.2+/-1.9+/-2.0 MeV
5279.3+/-0.7 MeV PDG ‘06
M(B-)
wrong-sign combs 5270.8+/-2.3+/-2.0 MeV
5279.1+/-0.5 MeV PDG ‘06
Branching fractions a
little high.
40.8 pb-1 (on res)
M(K-+) vs. p
• B(BD0X) = (0.8+/-0.2 +/- 0.2)
• used B(D0 K-+)=(3.0 +/- 0.6)%
• update to B(D0 K-+)=(3.8+/-0.07)%
• B(BD0X)=(0.63+/-0.2+/-0.2)
• PDG’06  (0.64+/-0.03)
on-res
continuum
M ( K   )
pD0
Beyond the basics


Limit on the bu coupling from
Semileptonic B Decay (1984)
Upper Limit on Flavor-Changing NeutralCurrent Decays of the b Quark (1984)

Two-Body Decays of B Mesons (1984)

Inclusive Decay of B Mesons into Charged
D* (1985)

Observation of the Decay B0 D*+r- (1985)

Decay By X (1985)

Inclusive  production in B-Meson Decay
(1986)

Observation of the Decay BFX (1986)

Inclusive B-Meson Decay to Charm (1987)

Limits on Rare Exclusive Decays of B
Mesons (1987)
Observation of BFX
PRL 56, 2781 (1986)
M ( )

on-res+off-res
on-res, p<2.5 GeV
off-res, p<2.5 GeV
10
117 pb-1
on-resonance
below-resonance
subtracted
Very relevant selection of modes!
Into the era of B0 – B0 mixing
B0-B0

Limits on

Branching Ratios of B Mesons to K+, K-,
and K0/K0 (1987)

Improved Upper limit on Flavor-Changing
Neutral-Current Decays of the b Quark
(1987)




Mixing and
tB0/tB+
(1987)
ARGUS
103 pb-1
79.5 pb-1
Argus
Evidence for Charmed Baryons in B-Meson
Decay ((1987)
G(bul)/G(bcl) from the End Point of
the Lepton Momentum Spectrum in
Semileptonic B Decay (1987)
Exclusive Decays and Masses of the B
Mesons (1987)
B0-B0 mixing at the Y(4S) (1989)
212 pb-1
ARGUS, Phys. Lett. B 192, 245 (1987)
d  0.17±0.05
Time-integrated mixing
rate: 21%
212 pb-1; 240 K BB
Big challenge:
removing contribution
from secondary leptons.
p
N (mix )
r
N ( nomix )
N ( B0 B0 )  N ( B 0 B 0 )

0 0
N (B B )
 (0.19  0.06  0.06)
M
 0.69  0.12  0.12
G
PDG: 0.776  0.008
Observation of B0-B0 Mixing in CLEO using Dileptons
Unlike-sign dileptons
Like-sign dileptons
(For unlike sign, need to subtract contribution from B+B- events.)
BABAR Mixing asymmetry data (raw)
m  0.502 ps-1
D<1 due to mistags
(fixed to PDG'04)
T=2/m
tB=1.53 ps
run out of events
at long lifetimes
0 0
0 0
0 0
NoMix(t )  Mix(t ) N ( B B t )  N ( B B , B B t )
Amix (t ) 

NoMix(t )  Mix(t ) N ( B 0 B 0 t )  N ( B 0 B 0 , B 0 B 0 t )
Searching for bu...and finding it!

Search for Charmless Decays Bpp  and
Bpp   (1989)

Search for bu Transitions in Exclusive
Hadronic B-Meson Decays (1989)

A Search for Exclusive Penguin Decays of
B Mesons (1989)

Study of the decay BD*+ l  (1989)

Observation of B-Meson Semileptonic
Decays to Noncharmed Final States (1990)

Exclusive and Inclusive Decays of B
Mesons into Ds Mesons (1990)

Exclusive and Inclusive Semileptonic
Decays of B Mesons to D Mesons (1991)

Inclusive and Exclusive Decays of B
Mesons to Final States Including Charm
and Charmonium Mesons (1992)
212 pb-1
Vub  0
scaled off res
p
• Major 1st step in the long
struggle to measure |Vub|.
• Inclusive measurement...in
very limited region of phase
space.
• Continuum background
suppression & determination
crucial
• If |Vub|=0, SM would predict
no CP violation.
CLEO II:
The Right Stuff
How long to run below resonance?
20
several more pages of math....
March of the Penguins
BK 
*
B  X s
B  K
B  K (*)
 
B  r
Loops in B decays: probe high mass scales!
1.377 fb-1
cited 560 times

b
d
u , c, t
W

s
d
B( B  K * )  (4.5  1.5  0.9)  105
5
B( B 0  K *0 )  (4.01  0.2)  10HFAG
8 events
3 events
2 events
cited 768 times
2.01 fb-1
B( B  X s )  (2.32  0.57  0.35)  104
Event-shape analysis
Y(4S) data
not all that rare!
B-reconstruction analysis
Y(4S) data
scaled off-resonance
scaled off-resonance
backgroundsubtracted data
E
E
9.1 fb-1
Summary of B(BXs)
Thanks to
Henning Flaecher!
CLEO Phys.Rev.Lett.87,251807(2001)
BR(B→Xsγ) = (3.29± 0.53) 10-4 (9.1 fb-1)
HFAG average:
7% experimental
uncertainty
Belle Semi Phys.Lett.B511:151(2001)
BR(B→Xsγ) = (3.29± 0.53) 10-4 (5.8 fb-1)
BaBar Semi Phys.Rev.D72:052004(2005)
BR(B→Xsγ) = (3.29+0.62-0.50) 10-4 (81.5 fb-1)
BaBar Incl Phys.Rev.Lett.97:171803(2006)
BR(B→Xsγ) = (3.92± 0.56) 10-4 (81.5 fb-1)
BaBar Full Phys.Rev.D77:051103(2008)
BR(B→Xsγ) = (3.91± 1.11) 10-4 (210 fb-1)
BELLE Incl (A. Limosani, Moriond EW08)
BR(B→Xsγ) = (3.37± 0.41) 10-4 (605 fb-1)
HFAG Average 08 (preliminary)
BR(B→Xsγ) = (3.52± 0.25) 10-4
Very good agreement
between experiments
and analysis methods!
huge theoretical effort:
SM predictions:
Misiak et al.
(hep-ph/0609232)
Becher et. al. (hep-ph/0610067)
Andersen et al. (hep-ph/0609250)
BR(B→Xsγ) (10-4)
Radiative penguins: The Next Generation!
BABAR, PRL 98, 151802 (2007)
B /106
B+r+
signal +
bkgnd
bkgnd
42 14.0
12.7 evts
B0r0
38.7 10.6
9.8 evts
B( B  X s ) /10
signal
Yet another generation: electroweak penguins!
Photon penguin

b
d
Z penguin


s
b
d
d
u , c, t
W

W+W- box
b
W
Z
u , c, t
W


W

s
u , c, t
d
d
• BaBar, Belle, CDF have observed BKl+l- and BK*l+l-
• Rarest observed B decay: B( B  K   )  (3.9  0.6)  107
• Kinematic distributions sensitive to new physics (AFB vs. q2)

s
d
B
 
,B  K
 
,B  K
Branching Fraction/10-6
*  
1.37 fb-1
u
Vus
b
B0
Vub
u
s
K

d
d
b
Vts s
u
K
d u

B0
d

30
3.14 fb-1
/105
BK : Direct CP Violation from
Interference between Penguin and Tree Diagrams
BaBar, PRL 93, 131801, 2004
Bkgd symmetric!
n  B 0  K     910
n  B 0  K     696
AK
696  910

 0.133
696  910
AK  0.133  0.030  0.009
N ( BB )  227  106
penguin “pollution”
in B+-
from Steve Olsen’s talk at Aspen Winter Conf., 2008
Direct CPV in B K decays
World Averages:
Acp(K+-) = - 0.097 ± 0.012
5s difference!
Acp(K+0 = +0.047 ± 0.026
Oct 30, 2000
HEPAP Meeting
34
Simple is beautiful: semileptonic B decays

b

Vcb ,Vub
c, u
B
bc:
*
**
D, D , D
q
CKM matrix elements
q
bu:
 , r ,  ,  , ,...
Understanding dynamics:
form factors, HQE params, quark
masses
924 pb-1
B Xc l 
+ continuum
contin.
“strict” contin
suppression
R2 < 0.3
• Quantitative statement
about size of |Vub|
Vub
 0.076  0.008
Vcb
Altarelli model
• Model dependence
studied; part of long, long
struggle.
...and with a factor of 10 more data
9.13 fb-1
Total background, including B decay (histogram)
On-resonance data (points)
Scaled off-res data (shaded region)
Background-subtracted,
efficiency corrected spectrum
Histogram: BXu l  spectrum
predicted from measured BXs 
spectrum
|Vub| Inclusive Measurements: HFAG Averages
Theory framework:
BLNP - B.O. Lange, M. Neubert and G.
Paz, Phys. Rev. D72:073006 (2005)
3
Vub  (3.99  0.140.32
)

10
0.27
Good or Bad?
The full breakdown of the uncertainties on the average |Vub| above is (all errors quoted in percent):
positive errors: +2.0stat +2.3exp +1.3b2c model +1.4b2u model +7.0HQE param +0.5SF func +0.7sub SF +3.6matching +1.3WA = +8.8tot
negative errors: -2.0stat -2.2exp -1.2b2c model -1.4b2u model -5.8HQE param -0.5SF func -0.7sub SF -3.3matching -1.3WA = -7.7tot
Lattice QCD input is essential to fully exploit...
The Heavy Quark Effective Theory Revolution
TopCite 1000+ (cited 1576 times)

Weak Transition Form-Factors Between Heavy Mesons, N. Isgur and M.B. Wise, Phys. Lett.
B237, 527 (1990). Cited 1458 times.

Semileptonic B and D Decays in the Quark Model, N. Isgur, D. Scora, G. Grinstein, and M.
Wise, Phys. Rev. D39, 799, 1989. Cited 1114 times.
3 papers: 4000 citations
Find your favorite
place in the Dalitz
plot.

*
Dc
q
B  D * 

final-state
D* at rest
2
q 2  qmax
V-A: more
points on Rthan L-side


q
c
D
2
q 2  qmin
*
40
2.4 fb-1
Confirmed by BaBar, PRD 74, 692004 (2006)
soft +
gentle fall-off of
form factor
The Triumph of Hadronic Decays
Hadronic B decays have ultimately provided the most
compelling test of the CKM framework through CP-violating
effects. We need interfering amplitudes to do this.
CLEO laid much of the foundation for this work.
CKM fit using angles only
The “Big B Paper”
203 cites
0.89 fb-1
D 
• Huge number of branching fractions
• Color-suppressed decays
• Polarization & factorization studies
• Resonant substructure
*0

D r
*0

D* r 
D 
*

D r
*

D*  D 0 
D *  0
r     0
r  0
B decays involving charmonium
golden mode for sin2b
3.1 fb-1
dE / dx
E
contributing mode for constraining 
B  D0K 
B  D0K 
A path to 
A( B   D 0 K  )  AB rB ei (d  )
A( B   D 0 K  )  AB
u
V 
*
us
b
u
Vcb  A
s
c
K

color suppressed
b Vub  e
 i
2
u
D0
u
Vcs*  1
u
0
c D
su

K
How can we get interference? Need D0 f and D0 f. (For example,
f =KS0+-.) Some observations:
1. Uses charged B decays; method is based on a direct CP asymmetry.
Issues: strong phase d, rB=|A(bu)/A(bc)| =0.1-0.2
2. Uses tree diagrams: no loops/mixing diagrams, no penguin/new
physics issues. Together with |Vub|, gives CKM test with trees only.
9.13 fb-1
contributing mode for constraint on CKM angle a
9.13 fb-1
K

 K *0
K 
 K *0
sin2b with penguins!
B

0
K
0
9.15 fb-1
50
The full glory of the Cabibbo-Kobayashi-Maskawa
framework
We need to see if it all fits: B, Bs, K, penguins, box, trees:
The Success of the CLEO Program as a
Scientific Enterprise

Many experiments have contributed to the huge project of
understanding B physics.

ALEPH, ARGUS, BaBar, Belle, CDF, CLEO, D0,
DELPHI, L3, OPAL, ...

ARGUS contributed enormously, far more than the relative
size of their data sample suggests.

Still, I believe that CLEO, more than any other experiment,
set the standard and created the foundation of this field.

I would like to express my deep appreciation to Wilson
Laboratory and to all the members of the collaboration for
making CLEO such a great project to work on!
Backup Slides
720 BB events
B( B  X m )  (9.4  3.6)%
“The nonobservation of t quarks
has led to the introduction of several
models in which the t quark does not
appear. Some of these models
require flavor-changing neutral
weak currents...”
Dilepton search:
No mm, including J/ym + m1 m e , 2 e+ e- (1 J/ye+ e-)
B(BX l+ l-) <1.3% (90% C.L.)
40.8 pb-1 (on res)
M(K-+) vs. p
• B(BD0X) = (0.8+/-0.2 +/- 0.2)
• used B(D0 K-+)=(3.0 +/- 0.6)%
• update to B(D0 K-+)=(3.8+/-0.07)%
• B(BD0X)=(0.63+/-0.2+/-0.2)
• PDG’06  (0.64+/-0.03)
on-res
continuum
M ( K   )
pD0
yes!
40.6 pb-1
78 pb-1 (doubled)
still < 15
events
in each
mode
Time-integrated mixing probabilities
B0B0 oscillations were measured without explicitly
measuring the time dependence. How was the mixing rate inferred?

 M 


 G 
0



  M 2 
0 PB0 B0 (t )dt  0 PB0 B0 (t )dt 2 1   G  


2
 PB0 B0 (t )dt
Bd system:
Bs system:
M d
  0.51 ps-1  1.53 ps 
Gd
M s
 14.5 ps-1  1.48 ps 
Gs
d
0.2
21.5   s
0.5
0.8

Measurements from BD*+l :
HFAG Averages
 (Dalitz plot): B-  [ D0Ks  - ; D0 Ks  - ]K-,
m2  m 2 ( K S0  )2
Giri, Grossman, Soffer, & Zupan, PRD 68, 054018 (2003),
Bondar (Belle), PRD 70, 072003 (2004)


B
M  (m2 , m2 )  A( B   D 0 K  ) f (m2 , m2 )  rB eid B e  i f (m2 , m2 )
B
M  (m2 , m2 )  A( B   D 0 K  )  f (m2 , m2 )  rB eid B e  i f (m2 , m2 ) 
m2
|M|2 =
D
m2
0
D
2
0
 rB ei (d B  )
m2
m2
Relatively large BFs; all charged tracks; only 2-fold  ambiguity.
Interference depends on Dalitz region: f  K S0 r 0 (CP), f  K *  (DCSD)
 (GLW method): B-DCPK-, DCPfCP
D0 (D0 ) fCP = CP eigenstate from singly-Cabibbo-suppressed decay.
[Gronau & London, PLB 253, 483 (1991), Gronau & Wyler, PLB 265, 172 (1991)].
D
c
u
0
W
Vcd

*
ud
V
d
u
d
u


Vud d


D
c
u
0
W
Vcd*
u
d
u




CP  1  +  , K  K 
CP  1 K S0 0 , K S0 , K S0, K S0 , K S0 
Amp  B  , CPD0  D   AB 1  D rB ei (d B  ) 
Large rate, but
interference is
small: rB << 1
 (ADS method): B-  [ D0K+ -; D0K+ -]KAtwood, Dunietz, & Soni, PRL 78, 3257 (1997),
PRD 63, 036005 (2001)
B  D0 K  ; D0  K  
u
K
s
B
c
b
u
c
D
u
DCSD
0
u
B  D0 K  ; D0  K  
B

Vub
e  i
b
u
A  B  , D  K 
u
c
su
D0
K
D0
c
u
CFD
s
K
u
d
u

d

u
s
u
K
 A A
B

id D
i ( d B  )


r
e

r
e
D  D
B
Interference is large: rB, rD comparable, but overall rate is small!
Extracting |Vtd /Vts| from bd  Decays
Belle, PRL 96, 221601 (2006).
courtesy M. Bona (UTfit collab.)
Vtd
0.018
 0.1990.026
0.025 0.015
Vts
expt thy
BABAR, hep-ex/0607099
(preliminary)
Vtd
0.017
 0.1710.018
0.021 0.014
Vts
thy
expt
CDF, hep-ex/0609040 (preliminary)
Vtd
 0.2060  0.0007 0.0081
0.0060
Vts
expt
thy
Consistent within errors!
(used CDF hep-ex/0606027)
Theoretical uncertainties already or soon limiting both approaches.
Amplitude for BK*l+lM (B  K
*  

GFa EM *
)
VtsVtb C9eff K * s  m PLb B
2
mix of Z-penguin, W+W- box
Kruger and Matias; PRD 71, 094009 (2005)
2
photon penguin
dom. at v. low q2

mb eff
*

m
C
K
si
s
q
P
b
B


7
m
R

q2

 C10 K * s  m PLb B

 m 5

Short-distance physics encoded in Ci’s (Wilson coefficients);
calculated at NNLO in SM:
C7eff
0.3 C9
+4.3
C10
4.7
Ali et al., PRD 61, 074024 (2000)
• Interference terms generate asymmetries in lepton angular
distribution over most of q2 range.
• Ci’s can be affected by new physics; enters at same order as SM amp.

How are CP violating asymmetries produced?
The Standard Model predicts that, if CP violation occurs, it must
occur through specific kinds of quantum interference effects..
source
A1
A1
a
A2
Double-slit experiment: if the final
state does not distinguish between
the paths, then the amplitudes A1
and A2 interfere!
a
A2
A1
A2
fi
fi
Three Kinds of CP Violation
We have seen that CP violation arises as an interference effect.
• Need at least two interfering amplitudes
• Need relative CP-violating phase
• Need relative CP-conserving phase
A single CP-violating amplitude will not produce observable
CP violation!
Classification of CP-violating effects in particle transitions
(based on the sources of amplitudes that are present).
1. CP violation in oscillations (“indirect CP violation”)
2. CP violation in decay (“direct CP violation”)
3. CP violation in the interference between mixing and decay
Two amplitudes with a CP-violating relative phase

Suppose a decay can occur through two processes, with
amplitudes A1 and A2. Let A2 have a CP-violating phase
2.
A  A1  a2e
i 2
A  A1  a2e
 i 2
No CP asymmetry!
(But the decay rate is different
from what it would be without the
phase.)
A  A1  A2
A2
A1  A1
A  A1  A2
2
A2
Two amplitudes with CP-conserving &
CP-violating phases

Next, introduce a CP-conserving phase in addition to the
CP-violating phase.
A  A1  a2e
i (2 d 2 )
A  A1  a2e

i (  2 d 2 )
Now have a CP asymmetry
A  A1  A2
A A
A2
A1  A1
2 
2
d2
A  A1  A2
A2
Amplitude analysis for direct CP violation
A  A1 e
i (1 d1 )
A  ( A1 e
i ( 1 d1 )
2
Asymmetry 
 A2 e
A  A
2
A  A
 A2 e
2
2

i ( 2 d 2 )
i (  2 d 2 )
)e
 i ( P )  ( f )
2sin(1   2 )sin(d1  d 2 )
2
2
A2
A2

 2 cos(1   2 ) cos(d1  d 2 )
A1
A1
Problems with interpreting measurements of direct CP asymmetries:
1. we often don’t know the difference d1-d2 , so we cannot
extract 1-2 from the asymmetry.
2. we often don’t know the relative magnitude of the interfering amps.
Direct CP violation in BK-+
Interference between tree and penguin amplitudes produces a CP
asymmetry in BK- . Both processes are suppressed!
External spectator
Gluonic penguin
*
us
V
b
W
Vub
d

u
s
u
d
K


B
0
b

d
Vts*
Vtb
W
t
s
u
K

u
d 
In our Wolfenstein convention, the CP-violating phase factor comes
 i
from Vub  e .
How the magic works
 m  t 
i ( d D )
cos 

a
e

2


0
B
 m  t 
i ( d D )

a
e

2


CP ( f )  cos 
0
B (t )
B
0
f CP
 m  t 
i ( d D ) 2 iM
i CP ( f )sin 

a
e
e

 2 
Amp( B 0  f CP )
0
B (t )
B
0
B
0
f CP
 m  t 
i ( d D ) 2 iM
i  sin 

a
e
e

 2 
Amp( B 0 (t )  f CP )
In each case, the two interfering amplitudes have the same CP
conserving phase from strong interactions, so it is irrelevant.
A  t   Im( )  sin  m  t 
Time-dependent CP asymmetries from the
interference between mixing and decay amplitudes
By modifying the mixing measurement, we can observe whole new
class of CP-violating phenomena: pick final states that both B0 and
B0 can decay into. (Often a CP eigenstate, but doesn’t have to be.)
B0 ( bd )
fCP
B ( bd )
no net oscillation
no net oscillation
B
0
net oscillation
G( B
0
phys
(t )  fCP )
fCP
0
B
0
B
0
net oscillation
B0
0
G( Bphys
(t )  fCP )
Results on sin2b from charmonium modes
(cc) KS (CP odd) modes
J/ψ KL (CP even) mode
asymmetry is opposite!
sin2b = 0.722  0.040 (stat)  0.023 (sys)
227 M BB events
|| = 0.950 +/- 0.031 (stat) +/- 0.013 (sys)
(raw asymmetry shown above must be corrected for the dilution)
from Steve Olsen talk at Aspen Winter Conference, 2008
1 from the “golden” bccs mode
-
535MBB
PRL 98, 031802 (2007)
Oct 30, 2000
HEPAP Meeting
77