B Physics and CP Violation
Jeffrey D. Richman
UC Santa Barbara
CTEQ Summer School
Madison, June 7-8, 2002
Outline (Lecture 1)

Overview of B decays

Why B physics is interesting; overview of decay
diagrams; introductory discussion of CP violation.

Accelerators and b-quark production

The BaBar Detector

Identifying B decays

B-meson lifetimes and mixing

CP Violation (CPv) and the CKM matrix

the CKM hierarchy and the prediction of large CP
asymmetries in B decays
Outline (Lecture 2)


CP Asymmetries:

sin(2b): the golden measurement

the struggle for the other angles
Rare decays


Semileptonic decays, decay dynamics, and the
magnitudes of CKM elements.


Penguins are everywhere!
Heavy-quark symmetry and Vcb
Prospects and future directions
A reference: J. Richman, Les Houches lectures, 1997.
http://hep.ucsb.edu/papers/driver_houches12.ps
(or send e-mail asking for a copy: richman@charm.physics.ucsb.edu)
Remarks/disclaimers

I will be unashamedly pedagogical, and I will not aim for
the level of impartiality that is customary in a review talk
or article.

I will be unashamedly selective: many important topics have
been left out.

There will be a strong bias towards recent results from e+ecolliders at the Y(4S). This is probably not too misleading
for now, since BaBar, Belle, and CLEO have to some extent
defined the state of the art, especially in CPv and rare
decays. However, soon-to-come measurements from the
Fermilab Tevatron (CDF, D0) will be of major importance.

My own background in b physics: BaBar, CLEO

I strongly encourage you to ask questions!
Goals of B (and Bs) Physics
1.
Can CP violation be understood quantitatively within
the Standard Model, or is new physics needed?
Perform a comprehensive set of measurements to
check for the presence non-SM CP-violating phases.
2.
Make precise measurements of the Standard Model
CKM parameters: |Vcb |, | Vub |, |Vtd |, |Vts |, a, b, g,...
3.
Map out and understand rare B decays, especially
processes with loops that can be very sensitive to
particles outside the Standard Model.
4.
Understand the dynamics of B decays: underlying
weak interaction process with overlay of complex
strong interaction effects. Progress: HQET, lattice
QCD, many measurements to test predictions.
Overview of B Decays

b is the heaviest quark that forms bound states with other
quarks (t-quark decays too rapidly).

m(b)<m(t) => the b-quark is forced to decay outside of its own
generation

Dominant decays are CKM suppressed:
(b  cW )  Vcb  (0.04) 2
 Relatively long B lifetime:   1.6 ps
B
2
(c B  480  m)
Silicon tracking systems have been essential tools.

Largest single branching fraction:
B( B   D*0l  )  (6.50  0.20  0.43)% (CLEO, hep-ex/0203032)

Many interesting rare decay processes are experimentally
accessible (b->uW, gluonic penguins, electroweak
penguins).
Leptonic and Semileptonic Decays
q 2  mW2  ml2
Ds
f Ds  Vcs

Leptonic B+ decay not yet
observed! (Amp  Vub )

Largest expected mode is:
Fj (q 2 ) Vcb

Used to measure
magnitudes of CKM
elements: Vcb and Vub

Amplitude can be
rigorously parametrized
in terms of form factors.
B(B      )  7 105

Ignoring photon radiation:
( M  l ) 
2

m 
G
 VqQ  f M2  ml2  1 

8
M


2
F
2
l
2
2
Hadronic Decays: Tree Diagrams


Theoretical predictions
very difficult.

“Color suppressed”

Naïve factorization model
probably breaks down.
(New data on B->D00
and B->D*00.)
Naïve factorization model
works reasonably well in
predicting pattern of
decays.
A( B 0  D   )
 a1  f  FBD (q 2  m2 )
A( B 0  D0 0 )

The color allowed
and color suppressed
amplitudes interfere
constructively in
charged B decays.
(Opp. effect for D+.)
 a2  f D  FB (q 2  mD2 )
Processes with loops: sensitivity to new particles
Z,

Both gluonic and electoweak penguins have been observed!

The SM mixing rate is dominated by tt (off-shell) intermediate states.
Processes used for sin2b measurement
0
0
Direct decay of B (or B ) to f CP
b
B
0
d
c
W+
c
s
d
J / ,  (2S),  c
0
S
0
L
K ,K ,K
*
A color suppressed decay! However, in this case, the rate is
enhanced by the relatively large decay constant of the J/:
f ( J / )  400 MeV
Decay modes for sin2b measurement
Mode
J/ KS
J/ KS
J/ KL
 2S) KS
 2S) KS
c KS
J/ K*
Subsequent Decays
Branching Ratio (10-6)
CP
J/  e+e- or 
KS
J/  e+e- or 
KS00
J/  e+e- or 
 2S)  e+e- or 
KS
 2S)  J/ 
J/  e+e- or 
KS
c  J/ g
J/  e+e- or 
KS
J/  e+e- or 
K*S0 KS
34
-1
15
-1
50
5
+1
-1
9
-1
6
-1
33
mixed
The C, P, and T Transformations

C, P, and T are discrete transformations: there is no
continuously varying parameter, and these operations
cannot be constructed from successive infinitesimal
transformations.
C : a  a (a is the antiparticle of a)
P : r  r (spatial inversion)
T : t  t (motion- or time-reversal; antilinear op.)

In all well-behaved quantum field theories, CPT is
conserved. A particle and its antiparticle must have
equal mass and mean lifetime.
M (a)  M (a )
 (a)   (a )
( a )  ( a )  / 
P and C violation in Weak Interactions
is Maximal (V-A)
e



e
e


e


P
e
  

Allowed
Not Allowed
Allowed
e
C

A First Look at CP violation

The discovery of CP violation in 1964 was based on the
demonstration that the mass eigenstate KL is not an
eigenstate of CP, so [ H , CP ]  0 .
c  K S0 )  2.7 cm
c  K L0 )  15.5 m
B( K L0     )  (2.0  0.4) 103
CP (   , L  0)  1

Remove Ks from
beam using lifetime
difference.
CPv small in kaon system!
The lifetime separation between BH and BL is tiny, so we
must use a different method, in which we compare the rates
for CP-conjugate processes.
( B  f )  ( B  f )  CP violation
  decay rate
f  a particular final state (often pick f  f CP )
The Legacy of Kaon Physics
“...the effect is telling us that at some tiny level there is a
fundamental asymmetry between matter and antimatter,
and it is telling us that at some tiny level interactions will
show an asymmetry under the reversal of time. We know
that improvements in detector technology and quality of
accelerators will permit even more sensitive experiments in
coming decades. We are hopeful then, that at some epoch,
perhaps distant, this cryptic message from nature will be
deciphered.” ...J.W. Cronin, Nobel Prize lecture*.
J.W. Cronin and V.L. Fitch, Nobel Prize 1980.
*J.W. Cronin, Rev. Mod. Phys. 53, 373 (1981).
J.H. Christenson, J.W. Cronin, V.L. Fitch, and R. Turlay,
Phys. Rev. Lett. 13, 138 (1964).
CP violation and alien civilizations

We can use our knowledge of CP violation to
determine whether alien civilizations are made of
matter or antimatter, without having to touch them.
( K   e e )  ( K   e e )
3
 3.3 10
 
 
( K   e e )  ( K   e  e )
0
L
0
L
 
Long-lived neutral kaon
0
L
0
L
 
We have these inside of us
CP Violation and Cosmology

A. Sakharov noted (1967) that CP violation has an
important connection to cosmology.

3 conditions for an asymmetry between N(baryons)
and N(anti-baryons) in the universe (assuming equal
numbers initially due to thermal equilibrium).

baryon-number-violating process

both C and CP violation (helicities not relevant to particle
populations)

departure from thermal equilibrium
Bnet   ( X  Yi )  ( X  Yi )   Bi
i
How can CP asymmetries arise? (I)

When we talk about CP violation, we need to talk about the
phases of QM amplitudes.

This is usually very confusing.


some phases are physical; others are not.

many treatments invoke specific phase conventions, which
acquire a magical aura.
Need to consider two types of phases

CP-conserving phases: don’t change sign under CP.
(Sometimes called strong phases since they can arise from
strong, final-state interactions.)

CP-violating phases: these do change sign under CP.
How can CP asymmetries arise? (II)

Suppose a decay can occur through two different
processes, with amplitudes A1 and A2.

First, consider the case in which there is a (relative)
CP-violating phase between A1 and A2 only.
A  A1  a2e
i 2
A  A1  a2e
 i 2
No CP asymmetry!
(Decay rate is different
from what is would be
without the phase.)
A  A1  A2
A2
A1  A1
A  A1  A2
2
A2
How can CP asymmetries arise? (III)

Next, introduce a CP-conserving phase in addition to
the CP-violating phase.
A  A1  a2e
i (2  2 )
A  A1  a2e
i (  2  2 )
A  A1  A2

Now have a CP asymmetry
A A
A2
A1  A1
2 
2
2
A  A1  A2
A2
Measuring a CP-violating phase

To extract the CP-violating phase from an observed
CP asymmetry, we need to know the value of the CPconserving phase.
Asymmetry 

2
2
2
2
A  A
A  A

2 A1 A2 sin(1   2 )sin(1  2 )
A1  A2  A1 A2 cos(1   2 ) cos(1  2 )
2
2
In direct CP-violating processes we usually do not
know the relative CP-conserving phase because it is
produced by strong-interaction dynamics that we do
not understand.
B production at the Y(4S)
 
0
0
e e  g  bb  (4S )  B (bd ) B (bd )


 B (bu) B (bu )
No accompanying pions!
The B-meson energy is known from the beam energy.
Rate of events vs. total energy in e+e- CM frame:
TM
BB threshold
 ( (4S ))  1.1 nb
33
2
 1.110 cm
f 
 1.058  0.084  0.136
f 00
(CLEO, CLNS 02/1775)
The New e+e- B factories

The machines have unequal (“asymmetric”) energy e+ and ebeams, so two separate storage rings are required.
PEP-II: E(e-)=8.992 GeV


E(e+)=3.120 GeV bg=0.55
The machines must bring the beams from the separate rings
into collision.

KEK-B: +-11 mrad crossing angle

PEP-II: magnetic separation
With two separate rings, the machines can store huge numbers
of beam bunches without parasitic collisions.

KEK-B: 1224 bunches/beam; I(e+)=716 mA; I(e-)=895 mA

PEP-II: 831 bunches/beam; I(e+)=418 mA; I(e-)=688 mA

CESR (single ring): 36 bunches/beam; I(e+)=I(e-)=365 mA
PEP-II e+e- Ring and BaBar Detector
LER (e+, 3.1 GeV)
Linac
HER (e-, 9.0 GeV)
BaBar
PEP-II ring: C=2.2 km
BaBar
May 26, 1999: 1st events recorded by BaBar
The Y(4S) Boost

The purpose of asymmetric beam energies is to boost the B0B0
system relative to the lab frame.
 B 1.6 ps


By measuring z, we can follow time-dependent effects in B decays.
The distance scale is much smaller than in the kaon decay experiments that
first discovered CP violation!
From CESR (1 ring, E symmetric) to
PEP-II (2 rings, E asymmetric)
Pretzel orbits in CESR
(36 bunches, 20 mm excursions)
Top view of PEP-II interaction
region showing beam trajectories.
(10X expansion of vertical scale)
The race between BaBar/PEP-II and Belle/KEK-B
Belle
L(max)  4.6 1033 cm-2s-1
Best day: 303 pb
-1
L(max)  7.25 1033 cm-2s -1
Exceeds design
luminosity!
Best day: 395 pb-1
e+e- vs. pp and pp

Production cross sections
Y(4S):

pp at Tevatron:  (bb ) 100  b
pp at LHC:
 (bb ) 500  b



 ( BB ) 1.1 nb

b fraction (ratio of b cross section to total hadronic cross section)

Y(4S):
0.25

pp at Tevatron:
0.002

pp at LHC:
0.0063
Comments

Triggering: so far, most B branching fractions have been measured at
e+e- machines, because CDF, D0 triggers were very selective in Run 1.
Also, PID & g detection are better at Y(4S) experiments so far.)

Hadron colliders produce Bs and b-baryons. (LEP also.)

New displaced-vertex triggers at hadron-collider experiments should
make a dramatic improvement.
The BABAR Detector
1.5 T solenoid
DIRC
(particle ID)
CsI (Tl) Electromagnetic Calorimeter
e+ (3.1GeV)
Drift Chamber
Instrumented Flux Return
e- (9 GeV)
Silicon Vertex Tracker




SVT: 97% efficiency, 15m z resol. (inner layers, perpendicular tracks)
Tracking : pT)/pT = 0.13% PT  0.45%
DIRC : K- separation >3.4 for P<3.5GeV/c
EMC: E/E = 1.33% E-1/4  2.1%
BaBar Detector
center line
DIRC:
quartz bars
standoff box
PM tubes
Superconducting
magnet (1.5 T)
Drift chamber
e
+
e
CsI crystals
Silicon Vertex
Tracker
Muon detector
& B-flux return
EM Calorimeter:
6580 CsI(Tl)
crystals (5% g
BaBar Event Display
(view normal to beams)
energy res.)
Cerenkov ring imaging
detectors: 144 quartz
bars (measure velocity)
Tracking volume:
B=1.5 T
Silicon Vertex Tracker
5 layers: 15-30 m res.
Rdrift chamber=80.9 cm
(40 measurement points, each with
100-200 m res. on charged tracks)
Innermost Detector Subsystem: Silicon Vertex Tracker
Installed SVT Modules
Be beam pipe: R=2.79 cm
(B mesons move 0.25 mm along beam direction.)
BaBar Silicon Vertex Tracker

5 layers of double-sided silicon-strip detectors (340)
50m
300m
80 e-/hole
pairs/m
Particle Identification (DIRC)
(Detector of Internally Reflected Cherenkov Light)
Quartz bar
Particle
c
Cherenkov light
• Measure angle of Cherenkov cone
Active
Detector
Surface
1
cos  c 
nb
p  mbg
– Transmitted by internal reflection
– Detected by PMTs
n  1.473
No. light bounces (typical)=300
Particle Identification with the DIRC.

DIRC c resolution and K- separation measured in data  D*+ D0+ 
(K-+)+ decays
>9
(c)  2.2 mrad
K/ Separation
2.5
Particle Identification
E/p from
E.M.Calorimeter
Electrons – p* > 0.5 GeV
Shower Shape
0.8 < p < 1.2 GeV/c
E/p > 0.5
1 < p < 2 GeV/c
•shower shapes in EMC

•E/p match
e
e

• Muons – p* > 1 GeV
• Penetration in iron of IFR
• Kaons
• dE/dx in SVT, DCH
• C in DRC
dE/dx from Dch
c from Cerenkov Detector
0.8 < p < 1.2 GeV/c
0.5 < p < 0.55 GeV/c

e

e
Identifying B Decays in BaBar
• Select “candidate daughter particles” using particle ID, etc.
• Compute the total 4-momentum: (E, p)=(E1+E2+E3, p1+ p2 +p3)
• Compute invariant mass: m2=E2-|p|2
Gives 10x improvement
in mass resolution.
mes  3 MeV
 E  15 MeV
All Ks CP modes
Nsig  750
Purity 95%
E
mes
sin2b Signal and Control Samples
J/ Ks
Bflav
J/ Ks
(Ks  )
(Ks  +-)
mixing
sample
CP=-1
J/ KL
CP=+1
J/ Ks
(Ks Ks00)
J/
(2s) Ks
J/ K*0
*0
J/
(K*0 KKs0)
c1 Ks
(Ks  00)
(K*0  Ks0)
The Lorentz Boost

The asymmetric beam energies of PEP-II allow us to
measure quantities that depend on decay time.
e-
e+
9.0 GeV
3.1 GeV
Tag B
z ~ 170 m
CP B
z ~ 70 m
J/
U(4s)
bg = 0.56
K0
z
t  z/gbc
gbcB  250 m
1 ps   170 μm
Measurement of Decay Time Distributions
τ( B  )
 1.082  0.026 (stat)  0.011 (sys)
0
τ( B )
B0 decay time
distribution
(linear scale)
background
0
0
B B Oscillations (Mixing)

B0 and anti-B0 mesons spontaneously oscillate into one
another! (Mixing also occurs with neutral kaons.)

Neutral B mesons can be regarded as a coupled, twostate system.
B

0
1
 
 0
B
0
 0
 
 1
To find the mass eigenstates we must find the linear
combinations of these states that diagonalize the
effective Hamiltonian.
Interpretation of the Effective Hamiltonian

The effective Hamiltonian for the two-state system is not
Hermitian since the mesons decay.
Quark masses, strong,
and EM interactions
 H11
H 
 H 21
H12   M
 *

H 22   M 12
B 0  f  B 0 transitions
f  off-shell f  on-shell
M 12  i  
  *

M  2   12

0

i 1 0
 ( M  ) 


2 0 1  * i *
 M 12  12
2

12 
 
Decays
i

M 12  12 
2


0

CP Violation in Mixing

Compare mixing for particle and antiparticle
off-shell
off-shell
on-shell
on-shell
CP-conserving phase
CP B 0  e 2iCP B 0
CP B 0  e 2iCP B 0
CP, H   0
 a =1
 0
CP   2i
CP
e

where
a
e2iCP 

0 
M  

M 12  12
*
12
i
2
i
2
*
12
B 0 H B0
B0 H B 0
CP violation in mixing, continued

To produce a CP asymmetry in mixing, M12 and 12 must not
be collinear and both must be nonzero:
No CP violation in mixing
CP violation in mixing
CP is a convention-dependent phase
*
Im( M1212
)  M12 12 sin( M12  12 )  0  CP violation in mixing
Time evolution of states that are initially flavor
eigenstates
B 0 (t ) 
B (t ) 
0
1
a
f  (t ) B 0
 a  f  (t ) B 0
f  (t ) B 0

f  (t ) B 0
1 iM t t / 2  iM t t / 2
f  (t )   e
e
e
e
)
2
1 iM t t / 2  iM t t / 2
f  (t )   e
e
e
e
)
2
a
*
M 12*  2i 12
M 12  2i 12
General case;
allows CP
violation.
CP Violation in B Mixing is Small

When CP violation in mixing is absent (or very small), we have
M  2 M12

  2 12
In the neutral B-meson system, the states that both B0 and B0 can decay
into have small branching fractions, since
bc
and
b c
normally lead to different final states. Can have ccdd (Cabibbo
suppressed) and uudd (b->u is CKM suppressed). So the SM predicts
12
 O(mb2 / mt2 )
M 12
1
Expect CPv in B 0 B 0 mixing to be O(10 -4 ). not yet observed
Time evolution of states that are initially flavor
eigenstates
B (t ) 
0
B 0 (t ) 

 t
2  iMt
M  t 0

e e  cos
B
2


 t
i
M  t 0
2  iMt 
e e  sin
B
2
a
M  t 0 
 ia  sin
B 
2

M  t 0 

cos
B 
2

In these formulas, we have assumed that
/<<1 and have set
  1   2
1
M  (M   M  )
2
The Oscillation Frequency (m)

In the neutral B-meson system, the mixing amplitude is
completely dominated by off-shell intermediate states (m)
[contrast with the neutral kaon system].

Calculation of the mixing frequency
m  VtbVtd

) strong+weak interactions  0.5 ps
 2
-1
Time-dependent mixing probabilities and asymmetry
dN nomix 1

 et  1  cos(m  t ) 
dt
4τ B
dN mix
1
t

 e  1  cos(m  t ) 
dt
4τ B
( a  1)
NoMix(t) - Mix(t)
 Asym(t)=
 cos(m  t )
NoMix(t)  Mix(t)
Tagging
CP asymmetry is between B0  fcp and B0  fcp
Must tag flavor at t=0 (when flavor of two Bs is opposite).
Use decay products of other (tag) B.
Leptons :
Cleanest tag. Correct 91%
e-
W-
W+

b
W-
Second best.
c
W+

b
c
Kaons :
b
e+
s
u
d
Correct 82%
K
b
c
W+
c
-
W
s
u
d
K+
Effect of Mistagging and t Resolution
No mistagging and perfect t
w=Prob. for wrong tag
D=1-2w=0.5
Nomix
Mix
t
t
D=1-2w=0.5
t res: 99% at 1 ps; 1% at 8 ps
t
t
Measure mixing on control sample:
• constrain model of t resolution
• measure dilution D = (1-2w)
t  trec  ttag
T=2/m
~D
m = (0.516  0.016  0.010) ps-1
CP violation in the Standard Model

In the SM, the couplings of quarks to the W are
universal up to factors that are elements of a unitary,
3x3 rotation matrix Vij of the quark fields. This
matrix originates in the Higgs sector (mass
generation of quarks).
W-
W-
e-
g e
b
b
W+
gVub
*
u
gVub
u
The Standard Model “Unitarity Triangle”
Cabibbo-Kobayashi-Maskawa (CKM) matrix
 d    Vud
  
s

V
   cd
 b   V
   td
[Col 1][Col 3]*=0
Vus Vub  d 
 
Vcs Vcb  s 

Vts Vtb 
b
 
Weak interaction
eigenstates
Quark mass
eigenstates
1 of 6 equal-area triangles:
orientation is just an unphysical
phase
V has only 4 real parameters,
including 1 CP-violating phase.
CPv
If just 2 quark generations:
no CP phase allowed!
The Structure of the CKM Matrix
The CKM matrix exhibits a simple, hierarchical
structure (which we do not understand) with 4 real
parameters.
λ 0.22
 Vud

V   Vcd
V
 td
1 2


3
1



A

(


i

)


2
Vus Vub  


1
  O(4 )
Vcs Vcb   

1  2
A2
2



3
2
Vts Vtb   A (1    i)  A
1

0.04




5
*
O ( )  O ( )  O ( )  0
(col 1)  (col 2)
O ( 3 )  O( 3 )  O( 3 )  0
(col 1)  (col 3) *
O ( 4 )  O ( 2 )  O( 2 )  0
(col 2)  (col 3)*
(All unitarity triangles have same area, corresponding to the sizes of
interference terms between 1st order weak amps. But we care about CP
asymmetries, so the angles of the triangles also matter.)
End of Lecture 1
Outline (Lecture 2)


CP Asymmetries:

sin(2b): the golden measurement

the struggle for the other angles
Rare decays


Semileptonic decays, decay dynamics, and the
magnitudes of CKM elements.


Penguins are everywhere!
Heavy-quark symmetry and Vcb
Prospects and future directions
A reference: J. Richman, Les Houches lectures, 1997.
http://hep.ucsb.edu/papers/driver_houches12.ps
(or send e-mail asking for a copy: richman@charm.physics.ucsb.edu)
Decay rates for B0(t) and B0 (t) to fCP
f CP H B 0 (t ) 
e
f CP H B 0 (t ) 
0
f CP H B (t )
0
f CP H B (t )
2
2

 t
2  iMt
e
e
e
t
e
t

 t
2  iMt
e


 a 
f CP
0

f
H
B

M

t
M  t 
CP
0
H B  cos
 ia 
sin

0

2
2 
f CP H B


f CP
0

f
H
B

M

t
M  t 
CP
0
H B  sin
- ia 
cos

0

2
2 
f CP H B


f CP H B
1
a
0
2
f CP H B 0
f CP H B 0
f CP H B
0
2
2

) 
)

) 
)
1
2
2
1

  1    1   cos  M  t )  Im(  )sin M  t 
2
2

1
2
2
1

  1    1   cos  M  t )  Im(  )sin M  t 
2
2


f CP H B 0
M  

M 12  12 f CP H B 0
*
12
i
2
i
2
*
12
Calculating the CP Asymmetry
0
AfCP (t ) 
f CP H B (t )
0
f CP H B (t )
2
 f CP H B (t )
2
2
 f CP H B (t )
2
0
0

  B 0 (t )  f CP )    B 0 (t )  f CP )
  B 0 (t )  f CP )    B 0 (t )  f CP )
AfCP (t )  C  cos( m  t )  S  sin( m  t )
C
1 
2
1 
2
S
2  Im( )
1 
2
If there is just one direct decay amplitude, we will see that
 1
AfCP (t )  Im(  )  sin( m  t )
If CP violation is due to interference between mixing
and one direct decay amp: pure sin(m t) time dependence.
Calculating 

Piece from mixing (a)
2
GF2 M W2 B mB BB f B2 *
i ( 2CP )
M 12 
V
V
S
(
x
)
e

td td )
0
t
2
12
a

mt2
xt  2
mW
M 12* Vtb*Vtd i 2CP
i 2(CP  M )


e

e
M 12 VtbVtd*
Piece from decay
f CP H B 0  a ei ( D )
f CP H B 0  CP ( f CP )e 2iCP a ei ( D )
f CP H B
0
f CP H B
0
 CP ( f CP )e
2 i (CP  D )
if just one direct decay
amplitude to fCP
  CP ( f CP )e
2 i ( M  D )
Hadronic physics divides out!
Calculating  for specific final states
  CP ( f CP )e 2i (
M
 D )
Vtb*Vtd Vud* Vub
=

*
VtbVtd VudVub*
Im( )=sin(2a )
Vtb*Vtd Vcs*Vcb Vcd* Vcs
 =  -1) 


*
*
VtbVtd VcsVcb VcdVcs*
Im( )=sin(2b )
B 0    
(b  uud )
B 0  J / K S0
(b  ccs)  ( K  K )
0
B  J / K
0
0
L
0
S
*
tb td
*
tb td
*
cs cb
*
cs cb
*
cd cs
*
cd cs
VV VV V V
 =  +1) 


VV VV V V
(b  ccs )  ( K 0  K L0 )
Im( )=-sin(2b )
Why it is magic
CP violating phase
CP conserving phase!
AJ / K 0
S ,L
t )  J / K
0
S ,L
 sin  2 b )  sin  m  t )
Graphical
Analysis
asdf
Analogy: “Double-Slit” Experiments
with Matter and Antimatter
A1
source
A1
A2
A2
In the double-slit experiment, there are two paths to
the same point on the screen.
In the B experiment, we must choose final states that
both a B0 and a B0 can decay into.
We perform the B experiment twice (starting from B0
and from B0). We then compare the results.
CP violation due to interference between mixing and
decay: non-exponential decay law
Ingredients of the CP Asymmetry
Measurement
  B (t )  f )    B (t )  f )
0
ACP (t ) 
0
  B (t )  f )    B (t )  f )
0
0
Determine initial state:
Measure t dependence
“tag” using other B.
Reconstruct the
final state system.
The Lorentz Boost

The asymmetric beam energies of PEP-II allow us to
measure quantities that depend on decay time.
e-
e+
9.0 GeV
3.1 GeV
Tag B
z ~ 170 m
CP B
z ~ 70 m
J/
U(4s)
bg = 0.56
K0
z
t  z/gbc
gbcB  250 m
1 ps   170 μm
Tagging
We must classify each neutral B according to whether it “started”
as a B0 or a B0. The start time is defined as the decay time of
the accompanying B meson (“tag B”). We use flavor-specific final
states of the tag B.
Leptons :
Cleanest tag. Correct 91%, Efficiency 11%
e-
W-
W+

b
b
c
Kaons :
W-
b
e+

c
Second cleanest.
c
W+
s
u
Correct 82%, Efficiency 35%
W+
c
s
b
K
u
W
d
d
K+
The Correlated State

At the Y(4S), the two neutral B mesons evolve as a
correlated quantum state until one of them decays.

1
(t1 , t2 ) C 1 
B0 (t1 ); p B 0 (t2 );  p  B 0 (t1 );  p B0 (t2 ); p
2

As a consequence, the asymmetry of time-integrated
rates is identically zero!

At the Y(4S), we must measure the CP asymmetry as
a function of time. The experiment would not work
with the silicon vertex detector.
)
Experimental aspects of the sin2b measurement:
F(t) F(t)
Acp(t)
sin2b
Everything perfect 
D sin2b
Add tag mistakes 
Dilution: D=1-2w
Add imperfect
t resolution 
t(ps)
t(ps)
Must understand tagging/mistagging and t resolution !!
Blind Analysis
• The whole analysis is performed blind.
• All studies are performed in such a way as
to hide information on the value of the
final answer.
• Avoids any subconscious experimenter bias
 e.g. agreement with the Standard Model!
When we are ready, we
have an unblinding party…..
Fit results
sin2b
(cc) Ks CP = -1
0.76  0.10  0.04
J/ KL CP = +1
0.73  0.19  0.07
All modes
0.75  0.09  0.04
(stat)
(syst)
56 fb-1: 62 M BB pairs.
CP asymmetry in CP -1 and +1 modes
J/  Ks
CP = -1
J/ KL
CP = +1
Note: likelihood curves are normalized to the total number of tagged
events, not B0 and anti-B0 separately.
Crosscheck: fit Bflav events as a CP sample
Expect no CP asymmetry
ACP = -0.004  0.027
sin2b fit results
Systematic errors
sin2b
(cc) Ks CP = -1
0.76  0.10  0.04
J/ KL CP = +1
0.73  0.19  0.07
All modes
0.75  0.09  0.04
(stat)
Belle :
(syst)
CP = -1 background
0.019
t resolution and detector effects
0.015
md and B (PDG 2000)
Monte Carlo statistics
0.014
0.014
J/ KL background
0.013
Signal mistag fractions
0.007
Total systematic error
0.04
Fit without ||=1 constraint (CP=-1 only)
|| = 0.92  0.06 (stat)  0.03 (syst)
Im/|| = 0.76  0.10
sin2b =0.82  0.12  0.05
42 fb-1
Cross checks
sin2b by decay mode
sin2b in sub-samples
Individual modes and sub-samples are all consistent.

CKM interpretation
Our sin2b measurement is
consistent with current
Standard Model constraints
from measurements of
other parameters.
 = (1-2/2)
 = (1-2/2)
Method as in Höcker et al,
Eur.Phys.J.C21:225-259,2001
(also other recent global CKM
matrix analyses)

Michael Peskin’s viewpoint
Conclusions so far...

We have observed CP violation in the neutral-B
meson system.

The asymmetry is large, unlike the O(10-3) effects
observed in the neutral-K system.

The asymmetry displays consistent behavior across
all observed channels, including CP odd and CP even
final states.

The time dependence of the asymmetry agrees with
the expectation based on interfering amplitudes
involving mixing and direct decay.
Conclusions so far...

With the present data sample, the region allowed by
the measurement is consistent with the Standard
Model CKM framework constrained by

CP-violation measurements in K decay

non-CP-violating observables in B decay
Hadronic Rare B Decays: Towards sin(2a)

B-> would measure sin(2a)…

…except there is a second direct decay amplitude!
Hadronic Rare B Decays: B->, B->K+
mES
E
B->
(5.4  0.7  0.4)  106
mES
E
B->K+
(17.8  1.1  0.8)  106
Mixing and CP Asymmetry Measurement
in B->
Mixing
Belle Mixing and Asymmetry
Measurement in B->
b  26
Belle :
0.16
S  1.210.38
0.270.13
BaBar : S  0.01  0.37  0.07
P / T  0.28
0.25
C  0.94 0.31
 0.09
C  0.02  0.29  0.07
B  K(*)l+l- in the SM and Beyond

Flavor changing neutral current (b to s): proceeds via
“penguin’’ or box diagrams in the SM.

New physics at the EW scale (SUSY, technicolor, 4th generation
quarks, etc.) can compete with small SM rate.

Complementary to studying b to s g due to presence of W and Z
diagrams.
Branching Fraction Predictions in the
Standard Model
New Ali et al. predictions lower by 30-40%
long-distance contribution from
 resonances excluded
Decay rate vs. q2 in the SM and SUSY
B  K  
B  K *   
J/
SUSY models
Pole from K*g, even in +-
(2S)K
SM nonres
SM nonres
q2
q2
constructive interf.
destructive
Generator-level q2 Distributions from FormFactor Models
 
 
B

K


B  Ke e
Ali et al. 2000
(solid line)
*  
BK e e
BK  
*


Colangelo 1999
(dashed line)
Melikhov 1997
(dotted line)
Shapes are
very similar!
J/ and Large Sideband Control Sample Study: B
Likelihood Variable
J/ Sample:
signal-like
Large SB
Sample:
backgroundlike
K *0 e  e 
keep
K *0    
K *0 e  e 
off resonance
K *0    
log LB
-10
4
log LB
Kl+l- Fit Regions,
Unblinded Run 1+2 data (56.4 fb-1)
E
mES
Fit Results (preliminary)
B(BK*ee)/B(BK*)=1.21
from Ali, et al, is used in combined K*ll
fit.
Belle results
(29.1 fb-1)
 0.6
4.12.7
2.1 0.8 evts
1.0
6.33.7
3.0 1.1 evts
 0.8
9.53.8
3.11.0 evts
 0.9
2.12.9
2.11.0 evts
 0.9
13.64.5
3.81.1 evts
Bkgd shape fixed from MC
Results

We obtain the following preliminary results:
 0.10
6
B( B  K    )  (0.84 00..30
)

10
24 0.18
B( B  K *
 
0.84
)  (1.890.72
 0.31) 106
< 3.5 10 6 90% C.L.

The statistical significance for B  K l+l- is computed
to be > 4 including systematic uncertainties.
Belle :
B( B  K
 
6
)  (0.750.25

0.09)

10
0.21
BaBar and Belle results are both higher than typical
theoretical predictions, but the uncertainties are still very large.
Measuring Magnitudes of CKM
Elements with Semileptonic B Decays
2
B( M Qq  X qq l  )  g theory  VqQ  M

Expt.
Need input
from theory!
Expt.
Kinematic Configurations in Semileptonic
Decay

b->cl processes are
dominant and are much
easier to understand than
b->ul decays.

reliable theoretical
predictions for b->cl at
zero recoil (Heavy Quark
Symmetry/HQET).

zero recoil: b->c without
disturbing the light
degrees of freedom

expansion in LQCD/mQ
zero recoil
Semileptonic decays: Dalitz plot

Effect of V-A coupling on lepton angular distribution
and energy spectrum.
* 
BDl 
zero
recoil
Contributions of different helicities to the rate
BDl 
* 
Zero recoil
Max recoil
New CLEO measurement of |Vcb |
B D
0
l
* 
B D l

*0 
CLEO Measurement of |Vcb | : w distribution and
extrapolation to zero recoil
Systematic Errors on CLEO |Vcb | Measurement
Recent |Vcb | measurements


Uncorrected for common inputs
Corrected for common inputs
(Compilation by Artuso and Barberio, hep-ph/0205163, May 2002.)
Recent |Vcb | measurements
Form Factor at Zero Recoil and |Vcb|

The experimental extrapolation to zero recoil velocity of the
daughter hadron provides the quantity
F ( w  1) Vcb  0.0383  0.005  0.009 (world average)

Zero recoil form factor (“consensus value”)
F (1)  QED A (1   1/ m2  ...)
Luke's theorem: no 1/ m corrections
QED  1.007
 A  0.960  0.007
F (1)  0.91  0.04

World average |Vcb|
Vcb  0.0421  0.0010 (expt.)  0.0019 (theory)
Bumps in the road: Crystal Ball
observation of the z(8.3) (1984)
Photon energy
spectrum.
(1S )  gz (8.3)
First observation of exclusive B decay

CLEO I data (1983)
Some free advice

Almost every measurement is very hard, even if it is of a
quantity that no one cares about. So, try to find an important
measurement that will have real scientific impact.

Never determine your event-selection criteria using the same
event sample that you will use to measure your signal.

Don’t use more cuts than you need. A simple analysis is easier
to understand, check, duplicate, and present.

Look at all the distributions you can think of for your signal
and compare them with what you expect.

Look at the distributions of events that you exclude. Do you
understand the properties of your background?
More free advice

When possible, use data rather than Monte Carlo events to
measure efficiencies and background levels.

Do not use Monte Carlo samples blindly. Find out where the
information came from that went into the MC. The MC may
do well in someone else’s analysis, but in may never have been
checked for the modes or region of phase space relevant to
your analysis.

Be careful not to underestimate the systematic errors
associated with ignorance of

signal efficiency

background shapes, composition, and normalization
Yet more advice

Don’t be afraid to…

ask any question

pursue a crazy idea

jump into something you don’t already understand

question what people say is established fact

look into the details and assumptions
Conclusions

We have two remarkable new facilities for B physics:

KEK-B/Belle

PEP-II/BaBar

The performance of these accelerators is a major
achievement for the laboratories.

The clear observation of CP asymmetries in the B
meson system is a milestone for particle physics.

The measurement of sin(2b) is very well accomodated
by the SM. It suggests that the dominant source of CP
violation in B decays is due to the CKM phase. In
spite of this, we have a long way to go before we have
fully tested the SM/CKM framework.
Conclusions (continued)

Hadron-collider experiments will soon start to play a
major role: the observation and precise measurement
of Bs mixing is one of the next major goals.

We are just beginning to scratch the surface of rare B
decays. They have interesting sensitivity to new
physics.

The next few years will be very exciting.
Backup slides
PEP-II

Very high current, multibunch operation

2 rings helps avoid beam instabilities and parasitic
beam crossings (crossings not at the IP)

I(e+)=1.3 A (LER), I(e-)=0.7 A (HER)

Bunch spacing: 6.3-10.5 ns

Beam spot:

x=120 m y=5.6 m z=9 mm

Number bunches/beam: 553-829 (to 1658)

High-quality vacuum to keep beam-related
backgrounds tolerable for experiments
PEP-II/BaBar Construction

1993: Start of PEP-II construction

1994: Start of BaBar construction

Summer 1998: 1st e+e- collisions in PEP-II

Spring 1999: BaBar moves on beamline

May 26, 1999: 1st events recorded by BaBar

Oct 29, 2000: PEP-II achieves design luminosity

Intense competition with KEK-B/Belle in Japan
PEP-II/BaBar

The Standard Model predicts O(1) CP asymmetries in B
decays! However, these asymmetries occur in processes that
are relatively rare, so a large data sample is required.

To perform these measurements, a two-ring e+e- storage ring
with unequal beam energies was built by SLAC/LBNL/LLNL
with unprecedented luminosity. We now have >60 MU (4S)
events.
The BaBar Collaboration
(9 countries)
BaBar DIRC quartz bar
Overall length (4 bars): 4.9 m
No. light bounces (typical)=300
Surface roughness (r.m.s.)= 0.5 nm
 (typical) = 400 nm
3.5 cm
BaBar DIRC Principle
1
cos  C 
b n
n  1.473
Num. r.l.=0.19 X0
(C) = 3 mrad
Number of Cherenkov photons=20-60
Experimental aspects of CP measurement
F(t) F(t)
True t, Perfect tagging:
Acp(t)
sin2b
True t, Imperfect tagging:
D sin2b
D = (1-2w) where w is mistag fraction.
Must measure flavor tag Dilution.
Measured t, Imperfect tagging:
Must measure t resolution properties.
t(ps)
t(ps)
B0 mixing measurement: D and R(t,t’)
True t, Perfect tagging:
Fmix(t)
Fnomix(t)
Amix(t)
True t, Imperfect tagging:
D
Amplitude of mixing asymmetry is
the dilution factor D.
Measured t, Imperfect tagging:
Mixing sample has 10x statistics of
CP sample. Shape of t determines
resolution function R(t,t’)
t(ps)
t(ps)
B->K*g