TM
The Penguin &
The Elephant:
Half a Billion b Quarks
at BaBar
Jeffrey Berryhill
University of California, Santa Barbara
For the BaBar Collaboration
Rutgers High Energy Physics Seminar
March 21, 2005
Outline
•
•
•
•
Quark Flavor and CP Violation
B Factories
Gluon Penguin B Decays
Photon Penguin B Decays
• The Electroweak Penguin B Decay b→ sll
Brand new results for B →Kll, K*ll
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
2
TM
Quarks,
Flavor Violation,
and CP Violation
Quarks and the problem of mass
Standard Model “explanation” of quark mass:
Six quark species with unpredicted masses
Spanning almost six orders of magnitude
Up type
(q=+2/3)
Up
u
Charm c
Top
t
Mass
(GeV/c2)
10-3
1
175
Down type
(q=-1/3)
Down
d
Strange s
Bottom b
Mass
(GeV/c2)
5 10-3
10-1
5
The origin of different fermion generations, masses, flavor
violation, and CP violation are all arbitrary parameters of
electroweak symmetry breaking.
A comparative physics of the quark flavors directly probes this
little-known sector.
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
4
Quarks and Flavor Violation
Photon, gluon or Z boson: quark flavor conserving interactions
W boson: changes any down type flavor to any up type flavor
d
u
qI = s
qJ = c
b
Vij
t
 Vud
V   Vcd

V
 td
Vus Vub 
Vcs Vcb 

Vts Vtb 
The (Cabibbo-Kobayashi-Maskawa) CKM matrix: complex amplitude of
each possible transition
Conservation of probability → CKM matrix is unitary
3X3 unitary matrix has (effectively) four degrees of freedom:
3 angles + 1 complex phase
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
5
CKM Unitarity
Inner product of first and third columns of CKM matrix is zero:
Rescale, rotate and reparameterize to describe a
Unitarity Triangle in the complex plane

 , 

VudVub
VcdVcb  2 
0, 0
Rutgers 21 Mar 05
 3 
VtdVtb Measure sides (decay rates)
VcdVcb and angles (CP violation)
to test unitarity
 1 
Jeffrey Berryhill (UCSB)
1, 0 
6
Three Paths to Flavor Violation
1.
Tree diagram decay: down → up (b→c, b→u)
Not generally sensitive to new physics
2. Box diagram: neutral meson mixing (K0,Bd0,Bs0)
down-type → anti-down type (b→b)
via double W exchange
3. Penguin diagram:
(b→s, b→d)
down-type → down-type
via emission & reabsorption of W;
top-quark couplings Vtd, Vts dominate
g,,Z
SM penguins are suppressed; new physics can compete directly!
+ 4th gen. quarks, technicolor,
LED, etc., etc., with possible
enhanced flavor couplings
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
7
Three Paths to CP Violation
CP violation → an observable O of particles (f1,f2,…) such that
O(f1,f2,…) ≠ O(CP(f1,f2,…)) = O(f1, f2, …)
1. CP violation in meson mixing
f
B
0
B
0
2
f
≠
B
0
B
2
0
Mixing rate of meson to final state f not the same
as
Mixing rate of anti-meson to same final anti-state
In the Standard Model, very small ~10-3
CP violation in K0 decays first observed through
this path forty years ago!
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
8
Three Paths to CP Violation
2. CP violation in meson decay → “Direct CP violation”
f
B
2
0
f
≠
B
Decay rate of meson to final
state f not the same as
Decay rate of anti-meson to
same final anti-state
0
BABAR
Recently observed in B0
→K+ p- at the 10% level!
Rutgers 21 Mar 05
2
Jeffrey Berryhill (UCSB)
Candidate mass
Blue
Red
mB
B 0  K p 
B 0  K p 
9
Three Paths to CP Violation
3.Time dependent asymmetry of meson/anti-meson decay rate to a
common final state
fCP
B
0
B
2
fCP
≠
B
0
2
0
B
0
If B0 and B0 decay to the same final state fCP, there is interference between
amplitude of direct decay
and
amplitude of mixing
followed by decay
In the presence of CP violating phases in these amplitudes, can induce large
time-dependent asymmetry with
frequency equal to mixing frequency: AfCP  C fCP cos(mt )  S fCP sin(mt )
C fCP  0 implies Direct CP Violation
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
10
B Factories
TM
b Quarks and CKM Unitarity
B decay rates, CP asymmetries measure the entire triangle!
Triangle sides:
B+→ 0 l- n decay rate measures b→u transition (|Vub|)
B0 mixing rate measures |Vtd|
Angles:
B0 → + - time dependent CP asymmetry measures sin 2
B0→ J/y KS0 time dependent CP asymmetry measures sin 2
B+→ D(*)K+ decay rates measure 

 , 
VudVub
VcdVcb  2 
Rutgers 21 Mar 05
0, 0
 3 
Need large sample
of B0, B+ mesons
produced under
controlled conditions

td tb

cd cb
V V
V V
 1 
Jeffrey Berryhill (UCSB)
1, 0 
12
Asymmetric B Factories
Resonance on top of
qq continuum,
S:B = 1:4
e+
3.1 GeV
B0
e9 GeV
Y(4S)
Y(4S)
B0
Y(4S) meson: bb bound state with mass 10.58 GeV/c2
Just above 2 × B mass → decays exclusively to B0 B0 (50%) and B+ B- (50%)
B factory: intense e+ and e- colliding beams with ECM tuned to the Y(4S) mass
Use e beams with asymmetric energy → time dilation due to relativistic speeds
keeps B’s alive long enough to measure decay time (decay length ~0.25mm)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
13
Time-Dependent CP Violation: Experimental technique
µ+
µ-
J/ψ
Fully
reconstruct
decay to CP
eigenstate f
π+
π-
KS
B0
e-
e+
B0
Asymmetric energies
produce boosted
Υ(4S), decaying into
coherent BB pair
Δz=(βγc)Δt
Determine time
between decays
from vertices
s(t) ≈ 1 ps
-
K
l-
Determine flavor and vertex
position of other B decay
Compute CP violating asymmetry A(t) = N(f ; t) – N(f ; t)
N(f ; t) + N(f ; t)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
14
PEP-II at the Stanford Linear Accelerator Center
Linac
PEP-II Storage Rings
SF Bay View
BaBar detector
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
15
PEP-II performance
PEP-II top luminosity:
9.2 x 1033cm-2s-1
(more than 3x design goal 3.0 x 1033)
1 day record : 681 pb-1
About 1 Amp of current per beam,
injected continuously
Run1-4 data: 1999-2004
On peak
211 fb-1
s(e+e- →Y(4S)) = 1.1 nb →
229M Y(4S) events produced
458 million b quarks produced!
Also ~108 each of u, d, s, c, and t
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
16
USA
[38/300]
INFN,
INFN,
INFN,
INFN,
The BABAR
Collaboration
California Institute of Technology
UC, Irvine
UC, Los Angeles
UC, Riverside
UC, San Diego
UC, Santa Barbara
UC, Santa Cruz
U of Cincinnati
U of Colorado
U of South Carolina
Colorado State
Stanford U
Florida A&M
U of Tennessee
Harvard
U of Texas at Austin
U of Iowa
U of Texas at Dallas
Iowa State U
Vanderbilt
LBNL
U of Wisconsin
LLNL
Yale
U of Louisville
U of Maryland
[4/20]
U of Massachusetts, Amherst Canada
U of British Columbia
MIT
McGill U
U of Mississippi
U de Montréal
Mount Holyoke College
U of Victoria
SUNY, Albany
U of Notre Dame
Ohio State U
China
[1/5]
U of Oregon
Inst. of High Energy Physics, Beijing
U of Pennsylvania
Prairie View A&M U
France
[5/51]
Princeton U
LAPP, Annecy
SLAC
LAL Orsay
11 Countries
80 Institutions
593 Physicists
The Netherlands [1/5]
NIKHEF, Amsterdam
Norway
[5/31]
Ruhr U Bochum
U Dortmund
Technische U Dresden
U Heidelberg
U Rostock
Italy
[1/3]
U of Bergen
LPNHE des Universités Paris
VI et VII
Ecole Polytechnique,
Laboratoire Leprince-Ringuet
CEA, DAPNIA, CE-Saclay
Germany
Perugia & Univ
Roma & Univ "La Sapienza"
Torino & Univ
Trieste & Univ
[12/101]
INFN, Bari
INFN, Ferrara
Lab. Nazionali di Frascati dell' INFN
INFN, Genova & Univ
INFN, Milano & Univ
INFN, Napoli & Univ
INFN, Padova & Univ
INFN, Pisa & Univ &
ScuolaNormaleSuperiore
Jeffrey Berryhill (UCSB)
Russia
[1/11]
Spain
[2/2]
Budker Institute, Novosibirsk
IFAE-Barcelona
IFIC-Valencia
United Kingdom
[10/66]
U of Birmingham
U of Bristol
Brunel U
U of Edinburgh
U of Liverpool
Imperial College
Queen Mary , U of London
U of London, Royal Holloway
U of Manchester
Rutherford Appleton Laboratory
May 3, 2004
The BaBar detector
Electromagnetic Calorimeter
6580 CsI crystals
e+ ID, p0 and  reco
Instrumented Flux Return
19 layers of RPCs
m and KL ID
e+ [3.1 GeV]
Cherenkov Detector
(DIRC)
144 quartz bars
K,p separation
Drift Chamber
40 layers
Tracking + dE/dx
Silicon Vertex
Tracker
e- [9 GeV]
Rutgers 21 Mar 05
5 layers of double
sided silicon strips
Jeffrey Berryhill (UCSB)
18
Penguin Portrait
Muon
from
other
B decay
B+ → Ks p 
Candidate
High energy
photon
Ks → pp
and p+
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
19
Our Friendly Competitors
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
20
TM
Gluon Penguins
and CP Violation
CKM Unitarity Triangle: Experimental Constraints
Constraints from
Remarkable
validation of
the CKM
mechanism
for both flavor
violation and
CP violation!
Decay rates and
Mixing Rates
CP violation in
kaon decay
CP violation in
b→ ccs
Rutgers 21 Mar 05
Do penguin
decays agree
with this
picture??
Jeffrey Berryhill (UCSB)
22
B0 →  KS0 and sin 2
Decay dominated by a single “gluonic penguin”
Feynman diagram: b → sss
b
Decay rate 100X smaller than J/y KS
→ small signal, large background
B0
W
u,c , t
g
d
s
s
 , ,( KK )C
s 0
K
d S
114 ± 12 B0 →  KS events out of ½ billion
b quarks produced!
Time-dependent CP violating asymmetry A
can be measured in the same way as J/y KS
Same combination of CKM complex phases as
J/y KS → same relation between A and sin 2
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
23
B0 →  K0 and sin 2
B0KS
SK 0  0.29  0.31
S
B0KL
SK 0  1.05  0.51
L
0
Btag
K  1
0
S
0
Btag
0
Btag
K  1
0
L
0
Btag
sin 2 = S( K0) = +0.50 ± 0.25 ±0.07
vs.
S(y K0) = +0.72 ± 0.04 ±0.02
Penguin decay consistent with tree decays, about 1 s low
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
24
Trees(green) vs. Penguins(yellow):
World Average
Averaging over
many penguin
decays:
World discrepancy
with tree decays
= -3.7
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
s
25
Trees(green) vs. Penguins(yellow): World
Average
Averaging over
many penguin
decays:
World discrepancy
with tree decays
= -3.7
s
Red boxes:
estimate from
theory of errors
due to neglecting
other decay
amplitudes
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
26
New Physics Scenarios
•New physics at the electroweak scale generically introduces many new large
flavor-violating or CP-violating couplings to quarks
Ex: Right handed (b →s) squark mixing in gluino penguins could introduce a
new phase in b → sss penguins
without affecting B mixing nor b → ccs nor b → s
Effects in gluon penguins should generically produce
effects in electroweak penguins
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
27
Photon Penguins
TM
Photon Penguin Decays
Radiative penguin decays: b → s and b → d FCNC transitions
SM leading order = one EW loop
Vts, Vtd dependent
precise SM prediction:
Measured two different ways:
Inclusive photon spectrum
Rutgers 21 Mar 05
BF(b→s)TH = 3.57 ± 0.30 x 10-4 (SM NLO)
Sum of B → Xs  exclusive states
Jeffrey Berryhill (UCSB)
29
b  s  Branching Fractions
Excellent agreement with SM
Systematics, theory limited:
both could improve
to 5% precision
“Standard Candle” of flavor physics
Constrains numerous models
Strongest bound on H+
b→ s
Colliders
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
30
b  s  Asymmetries
•CP asymmetries consistent with SM (0.4%) at the ~5% level
•K* isospin asymmetry 0- consistent with C7 < 0
•Statistics limited up to ~1 ab-1
sgn 0- = -sgn C7
Can we exclude C7 > 0 ?
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
31
Search for b → d  Penguins: B → , w 
Simplest and most “common” b → d
exclusive decays are  and w:
B+ →  ,  → p p0
B0 → 0 , 0 → p p
B0 → w , w → p p p0
BF ≈ BF K* x |Vtd/Vts|2 ≈ 1.6 x 10-6
Huge backgrounds suppressed by
neural net
multi-dimensional likelihood fit
excellent K-pi separation
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
BF = (0.6 ± 0.3 ± 0.1) x10-6
significance = 2.1s
BF < 1.2 x10-6 90% CL
32
b→d CKM matrix constraint
Ali et al. hep-ph/0405075
SU(3) breaking of
form factors z2= 0.85 ± 0.10
weak annhilation
correction R = 0.1 ± 0.1
(z2,R) = (0.75,0.00)
theory error
(z2,R) = (0.85,0.10)
no theory error
Penguin data now
provide strong CKM
constraint
Direct competitor with
Bs mixing
Reduction of theory errors
needed!
 95% C.L. BaBar allowed region (inside the blue arc)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
33
TM
The Ultimate Electroweak
Penguin Decay:
b → s l+l-
The Physics of b  s l l
Three separate FCNC diagrams contribute→ multiple ways for new physics to interfere
Three body decay→ non-trivial kinematic and angular distributions sensitive to magnitude
and phase of different amplitudes
Rare decay→ 10-6 branching fractions approach limits of B-factory sensitivity
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
35
The Physics of b  s l l
b→ s ll rate computed from operator product expansion with s and L/mb corrections
general Hamiltonian
of b→s transitions
Wilson coefficients Ci encode short distance physics (order s)
C7 photon penguin
from b → s 
C9 (mostly) Z penguin
C10 (mostly) W box
unique to b→ s ll
BF(b→smm) = 4.2 ± 0.7 10-6
15% uncertainty in inclusive rate
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
36
The Physics of B  K l l, K*l l
Dominant exclusive decays are B → Kll, K*ll
K ll, K*ll rates computed from (perturbative) b→ s ll amplitude convolved with
(non-perturbative) B→ K, K* form factors for each Oi
et al.
3 B → K form factors
BF(B→Kmm) = 0.35 ± 0.12 10-6
7 B → K* form factors
et al.
BF(B→K*mm) = 1.2 ± 0.4 10-6
30% form factor uncertainty
(Light-cone QCD sum rules)
Exclusive branching fractions not a precision test of the SM →
Asymmetries and distributions are higher precision tests
(form factor uncertainty cancels)
EX: Direct CP asymmetry ACP ≈ 10-4 in SM, could be order 1 beyond SM
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
37
Dilepton Mass Distribution
Dilepton mass distribution
probes Wilson coeffcients
J/Y K(*)
Y(2S)K(*)
Pole at q2 ≈ 0 for K*ee
(nearly on-shell B→K*
Huge long-distance
contribution from
B →charmonium decays
Rutgers 21 Mar 05
Dilepton q2
Jeffrey Berryhill (UCSB)
38
SUSY Higgs physics at a B Factory
b→ smm = Bs → mm turned sideways
Sensitive to “neutral Higgs penguin”
for SUSY with large tan 
Ratio R(K) = BF(B→ Kmm)/BF(B→Kee)
isolates Yukawa enhancement in muon mode
In SM, equal to unity with very high precision
RK limit of 1.2 =
Bs→mm limit of < 10-6
Also contributes to R(K*)
(=1 above the photon pole)
Complementary to Tevatron Bs→mm limit
Rutgers 21 Mar 05
Hiller & Kruger hep-ph/0310219
Jeffrey Berryhill (UCSB)
39
Forward-Backward Asymmetry
Angular asymmetry of lepton (anti-lepton)
angle with B (anti-B) momentum in dilepton
rest frame
Standard Model
Theoretically clean probe of relative size
and phase of C7/C9/C10
dAFB/dq2
Varies with dilepton q2
Can be dramatically modified by
new physics
Zero of AFB provides simple, precise (15%)
relation between C7 and C9 (for LHCb)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
Ci modified by SUSY
Dilepton q2
40
B  K l l, K*l l Measurement
Find ~50 needles in a haystack of 500 million B’s + 2 billion light quarks
Full B decay reconstruction to charged tracks:
(brem photons added for ee modes)
B→ Kll
B+ →K+ e+ eB0→Ks e+ e-
B→K*ll
B+→K*+ e+ eB0→K*0 e+ e-
B+ →K+ m+ m-
B+→K*+ m+ m-
B0→K*0 m+ m-
K*+→Ksp+
K*0→K+p-
B0 →Ks m+ mKs→p p
Strict particle ID requirements
BLIND ANALYSIS
Veto “peaking” backgrounds of B decays similar to signal
Construct multivariate discriminants to suppress “combinatorial” backgrounds
Signal yield extraction via multi-dimensional unbinned maximum likelihood fit
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
41
“Killer App” of the Y(4S): precision kinematic constraints
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
42
Particle Identification
0.3 GeV
0.7 GeV
m, K fake rates < 2%
e fake rate < 0.1 %
Huge control samples for efficiency and misid studies
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
43
Charmonium Background
Huge (100 times signal) J/y K(*) and y(2S) K(*) background eliminated with
2D veto on dilepton mass and E
Mismeasured mass correlated with E
Bigger veto for electron modes (Bremsstrahlung tail)
Energy loss and tracking response to leptons well-calibrated in MC
(checked with huge radiative Bhabha and mm samples)
Reduced to < 1 event expected per mode
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
Veto sample
is huge control
sample for validating
signal efficiency!
44
Continuum Background Suppression
linear combination of variables for
which multi-dimensional gaussian dist.
of signal and background are maximally
separated
Signal shape from MC; background shape from off-resonance data
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
45
Continuum Suppression
Kaon-lepton mass in Fisher
reduces large background from
semileptonic D decays
Background cuts off at D mass
Large charmonium samples validate signal shape; sidebands validate background shape
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
46
Semileptonic B Decay Background
Large fraction (10%) of B’s decay to leptons→ large “combinatorial” B background
Separation of B background from signal using likelihood function (product of shapes):
Signal vs.
Background
missing energy
Missing energy
B production angle
Vertex probability
Cut values scanned
For maximal S2/(S+B)
Likelihood shapes from MC
Sideband
data
Charmonium
data
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
47
Hadronic B Decay Background
B decays to hadrons misidentified as leptons will have same kinematics as signal
Electron modes: misid ~0.1%, negligible
Muon modes: misid ~1-2% for pions and kaons
Misid rate calibrated from large B→D*p, D* →D p, D→ K p samples
K mm, K*mm events with Km mass consistent with D decay vetoed
Non-resonant B→K(*)pp background estimated from data:
Convolve misid rates with B →K(*)mp events in data
Extract background from a binned fit to mES distribution
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
48
Maximum Likelihood Fit
Unbinned maximum likelihood fit of (m ES, E, m(Kp)):
maximizes LH function
K*0ee MC
Components: signal, peaking backgrounds,
combinatorial background
Pi = Pi(mES)*Pi(E)*Pi(m(Kp)) (negligible correlation)
Signal shape parameters fixed from charmonium data
Signal shapes (narrow peaks)
Signal yield, background yield and background shape
parameters are floating in fit
K+em data
mES: “ARGUS function” models phase space
cutoff at MB
E: linear or quadratic
M(Kp: quadratic (+ small 5% K* fraction)
Background shapes (not peaked)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
49
Branching Fraction Systematics
Measure single cut efficiencies for particle ID, Fisher, likelihood
directly from data using vetoed charmonium events
Ex: B background likelihood cut, data vs. MC
Charmonium
data
Systematic uncertainty on cut efficiency reduced to a few percent
Dominant systematic on efficiency (4-7%) is model dependence
(form factor models change q2 distribution)
Fit systematics: change in signal yield from choices of
Signal shape parameters (mES, E, m(Kp) mean and width)
Background shape scheme (introduce correlations, higher polynomial terms)
Peaking background mean rates (total varied within uncertainty)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
50
Fit Validation: Charmonium
KsJ/y(ee)
data
Feed-down from
J/y K*+
Test fit procedure on charmonium
J/y K(*) branching fractions agree with PDG
ACP consistent with 0 (bounds detector bias)
K*0J/y(ee)
data
Signal and background well-modeled by PDFs
Also:
Y(2S) K(*) branching fractions
K(*)em
Lower-dimensional and/or smaller range fits
Sideband yields agree with MC
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
51
Kll Fits
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
52
Kll Fits
K+ee
K+mm
Ksee
Ksmm
Eff. Syst.
Fit Syst.
Kll modes consistent
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
53
Kll Combined BF
Simultaneous fit to four individual decay modes
Partial rates constrained to same value
K*ll feed-down
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
Kll signal
54
Kll Sidebands
Fit describes background shape well
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
55
K*ll
slide
Fits
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
56
K*ll Fits
K*0ee
Rutgers 21 Mar 05
K*+ee
Jeffrey Berryhill (UCSB)
57
K*ll Fits
K*0mm
Rutgers 21 Mar 05
K*+mm
Jeffrey Berryhill (UCSB)
58
K*ll Combined Branching Fraction
Simultaneous fit with
constraint
G(B→K*mm)/G(B→K*ee)=0.752
(“K*ll BF” ≡ K*0mm BF)
Alternatively, with
pole region removed
and equal partial rates:
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
59
K*ll Sidebands
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
60
BF Summary
BaBar and Belle branching fractions agree with SM predictions
Experimental uncertainty already better than theory
Difference between BaBar and Belle becoming significant?
Smallest B branching fractions ever measured!!
difference = 2.6s
difference = 1.9s
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
“Sum-of-exclusive”
b→sll BF also agrees
with SM
61
Ratios of BFs and ACP
Ratio of Kmm/Kee BFs consistent
with unity:
ACP measurements consistent with 0
Ratio of K*mm/K*ee BFs consistent
with SM (=0.752):
Ratio of K*mm/K*ee BFs (excluding
pole) consistent with unity:
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
Dominant systematic from
unknown asymmetry of peaking
background
62
Towards AFB
First, need partial
BF measurement
vs. q2
Signal candidates span
entire range
Separate fits for optimal
q2 bins
Fit method for cosq*:
4D LH fit
Check with charmonium
and Kll
J/yK*+
Need asymmetric efficiency
and background PDFs
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
63
Towards AFB
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
64
Summary
•Strong penguins beginning to crack the standard model?
•Decay rate measurements of b→ s penguins are well into the precision era.
•CP asymmetries of b→ s penguins are statistics limited and will continue to test
the SM
b→ d penguins are only now beginning to reveal themselves in B-factory data.
They could also uncover new physics, or measure the poorly known |V td|.
BF(B → ,w ) < 1.2 X 10-6
|Vtd/Vts| < 0.19 (90% CL)
“Ultimate” b →sll penguin will ultimately disentangle all electroweak
penguin effects, SM or not.
B→Kll ratios and K*ll angular asymmetries are high precision tests of the
electroweak scale and beyond
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
65
b
B0
W
u,c , t
g
d
s
s
s 0
K
d S
Penguin decays are strongly challenging
the Standard Model!
TM
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
66
Search for B → , w (191 fb-1)
Simplest and most “common” b → d exclusive decays are  and w:
B+ →  ,  → p p0
B0 → 0 , 0 → p p
B0 → w , w → p p p0
BF ≈ BF K* x |Vtd/Vts|2 ≈ 1.6 x 10-6
BF ≈ ½ BF 
Comparable to rarest
BF ≈ ½ BF 
B decays measured!
Previous BaBar limit: BF  < 1.9 x 10-6 90% CL (78 fb-1)
Preliminary Belle evidence (3.5s): BF  = 1.8 ± 0.6 ± 0.1 x 10-6 (140 fb-1)
Ratio of exclusive b → d and b → s decay rates measures |Vtd/Vts| via
Ali et al. hep-ph/0405075
R weak annhilation correction
z2 SU(3) symmetry breaking of exclusive form factors
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
67
Continuum Background Suppression
e+e- → qq continuum events are
the dominant background
e = 76%
cut
Combine properties discriminating
B →  + X from continuum
into a single-output neural net:
•Event shape variables: spherical B vs.
jet-like continuum
•B flavor tagging variables: kaons and
leptons in the rest of the event
•B candidate vertex separation from
rest of the event
Neural net trained on simulated signal
and continuum samples
Neural net distribution for both signal and
background described well by simulation
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
68
B Background Suppression
Large (40 x signal) K*  background
reduced by strict p+ PID requirements
(track dE/dx, DIRC qc , DIRC N )
K+ misid 1-2% for most of acceptance
E provides additional separation
of K* from 
cut
B decays to p0, , wp0, w
suppressed by cut on helicity angle
Combine helicity angle, B production angle,
Dalitz angle (w only) into B Fisher discriminant
for additional background separation in signal fit
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
69
Signal Fit: K* Control Sample
Validate fit procedure on
data with K* sample
(0 candidates with
no PID requirement)
Unbinned extended ML
fit of three background
components
(qq, Xs , 0p0/
and K* component
K* + background
preliminary
E shifted
background
to 4D data (mES,E,NN,Fisher)
qq, Xs , K* yield are free
parameters
Large K* signal yield
consistent with BaBar
BF measurement
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
70
b  d  Branching Fractions: Summary
Comparison of measurements with predictions for individual modes and combined BF
central value
90% C.L. upper limit
2.0s difference in
BaBar, Belle BF
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
71
CKM matrix constraint
BaBar BF ratio upper limit < 0.029 → |Vtd/Vts| < 0.19 (90% CL)
Ali et al. hep-ph/0405075
(z2,R) = (0.85,0.10)
no theory error
Penguins are starting to
provide meaningful CKM
constraint
 95% C.L. BaBar allowed region (inside the blue arc)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
72
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
73
Quarks and their Strong Interactions
Quarks cannot be detected in isolation, only as bound states
A quark/anti-quark pair forms a bound state (mesons)
Flavor
u,d
s
c
b
t
spin 0 p+ (ud)
mesons p0 (uu-dd)
K+ (us)
KS0 (ds+sd)
D+ (cd)
D0 (cu)
B0 (db)
B+ (ub)
none
spin 1 + (ud)
mesons 0 (uu-dd)
w (uu+dd)
K*+ (us)
K*0 (ds)
 (ss)
D*+ (cd) U (bb)
D*0 (cu)
J/y (cc)
none
b quarks are the heaviest flavor with measureable bound states→
B mesons are a natural starting point for studying the other flavors
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
74
Trees(green) vs. Penguins(yellow): BaBar Data
Averaging over
many penguin
decays:
BaBar discrepancy
with tree decays
= -2.7
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
s
75
Quarks and Flavor Violation: Mixing
Pairs of down (or pairs of up) type quarks can spontaneously
swap flavor for anti-flavor via two flavor-violating exchanges
“Meson Mixing” aka “Flavor oscillation”
Sensitive to Vtd, top quark!
B
0
B
0
Prob(B0 → B0) ≈ exp( -G t)/2 * ( 1 – cos(m t) )
Similar to neutrino oscillation, except decay term added
Mixing time ~few ps
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
76
Quarks and CP Violation
For a particle(s) f with momentum p and helicity l
C : Charge conjugation operator C f(p, l) = f(p, l)
P : Parity reversal operator
P f(p, l) = f(-p, -l)
CP f(p, l) = f(-p, -l)
CP eigenstate: particle = anti-particle (Ex: qq mesons)
CP conservation → left-handed particles have the same
physics as right-handed anti-particles
Obviously violated for our (baryon-asymmetric) local universe!
In the Standard Model CP violation originates from
complex phase in CKM matrix → in general, Vij ≠ Vij*
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
77
B0 → J/y KS0 and sin 2
Decay dominated by a single “tree-level”
Feynman diagram: b → ccs
J/y identified cleanly by decay to a lepton pair;
KS identified cleanly by decay to pion pair.
Both particles are CP eigenstates → both B0 and B0 decay to them
Time-dependent CP violation has amplitude sin 2 and frequency m
Works for several other b →ccs decays as well; results can be combined
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
78
sin 2 fit results
Raw asymmetry A(t) ≈ ( 1 – 2w ) sin 2 sin mt
(cc) KS (CP odd) modes
J/ψ KL (CP even) mode
Signal yield, background yield, sin 2, flavor tagging, t resolution function all
from simultaneous maximum likelihood fit to signal+control samples
sin2β = 0.722  0.040 (stat)  0.023 (sys)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
79
CKM Unitarity Triangle: Experimental Constraints
Constraints from
Decay rates and
Mixing Rates
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
80
Consistency checks
Cleanest
charmonium
sample
Χ2=11.7/6 d.o.f.
Prob (χ2)=7%
Rutgers 21 Mar 05
J/ψKS(π+π-)
Purest flavor
tagging mode
Lepton tags
ηF=-1 modes
Χ2=1.9/5 d.o.f.
Prob (χ2)=86%
Jeffrey Berryhill (UCSB)
81
CKM Unitarity Triangle: Experimental Constraints
Constraints from
Decay rates and
Mixing Rates
CP violation in
Kaon decay
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
82
Future and Follow-up Measurements
•Both B factories hope to collect 4-5 X more data over the next 4-5 years
•Significance of the penguin problem could double and unambiguously
falsify the Standard Model!
•Improved measurements of rates and asymmetries in other penguin decays
(b →s , b→ d , b→ s l l, B →  K*, ……)
•Fermilab Tevatron can measure Bs , Lb decays
•LHCb, BTeV: scheduled to produce billions of B’s in pp collisions
•Super B Factory: 50X version of B factories
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
83
Direct CP Violation: BaBar Data
Direct CP
violation
consistent with
0 for all modes
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
84
Direct CP Violation: World Average
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
85
CKM Constraints
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
86
Direct CP Asymmetry: b →s and B  K* 
< 1% in the SM, could receive ~10% contributions from new EW physics
Either inclusive or exclusive decays could reveal new physics
B or K charge tags the flavor of the b quark with ~1-2% asymmetry systematic
Sum of 12 exclusive,
self-tagging
B→Xs  final states
b→s
b→s
Xs = K/Ks + 1-3 pions
E > 2.14 GeV
b →s ACP = (N – N)/(N + N) = 0.025 ± 0.050 ± 0.015
PRL 93 (2004) 021804, hep-ex/0403035
Asymmetries also measured precisely in exclusive K* decays:
B →K*
0- =
G(K*0 )
ACP = -0.013 ± 0.036 ± 0.010
–
G(K*-
submitted to PRL, hep-ex/0407003
) = 0.050 ± 0.045 ± 0.028 ± 0.024
preliminary
G(K*0 ) + G(K*- )
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
87
Time-Dependent CP Asymmetry in B →K* (113 fb-1)
As in B0→J/y KS, interference between mixed and
non-mixed decay to same final state required for CPV.
In the SM, mixed decay to K* requires wrong photon
helicity, thus CPV is suppressed:
In SM: C = -ACP ≈ -1%
S ≈ 2(ms/mb)sin 2 ≈ 4%
Measuring t of K*(→KSp0)  events requires novel beam-constrained vertexing techinque:
Vertex signal B with intersection of K S trajectory and beam-line
Usable resolution for KS decaying inside the silicon tracker
Validated with B0→J/y KS events
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
88
Time-Dependent CP Asymmetry in B →K* (113 fb-1)
Likelihood fit of three components
(qq, BB, K*)
to 5D data
(mES,E,Fisher,mK*,t)
preliminary
K* signal = 105 ± 14 events
S = +0.25 ± 0.63 ± 0.14
C = -0.57 ± 0.32 ± 0.09
submitted to PRL, hep-ex/0405082
Consistent with SM
For C fixed to 0, S = 0.25 ± 0.65 ± 0.14
First ever measurement of time-dependent CP asymmetries in radiative penguins!
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
89
Radiative Penguin Decays and New Physics
Radiative penguin decays: b → s and b → d FCNC transitions
SM leading order = one EW loop
Vts, Vtd dependent
FCNCs probe a high virtual energy scale comparable to high-energy colliders
Radiative FCNCs have precise SM predictions:
BF(b→s)TH = 3.57 ± 0.30 x 10-4 (SM NLO)
BF(b→s)EXP = 3.54 ± 0.30 x 10-4 (HFAG)
Multiple new BF(b→s)
measurements coming
soon from BaBar
Decay rate agreement highly constrains new physics at the electroweak scale!
Further tests presented here:
•Exclusive b→s decay rates
•b→s CP asymmetries
•b→d penguins
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
90
Measuring Exclusive Radiative Penguin Decays
Small rates (BF < 10-4) on top of large backgrounds
Common strategy :
select photon + hadrons with strict particle ID
cut when possible on meson masses
reduce continuum background with multivariate methods
extract signal with multi-dimensional maximum likelihood fit
Best kinematic constraints
from B candidate momentum and
energy, compared with Ebeam:
Ex: signal fit for B → K*0, K*0 → K+p583±30 events
* denotes CM frame
For signal, mES ≈ mB, smES) ≈ 3 MeV
E* ≈ 0, s(E*) ≈ 50 MeV
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
91
B  K*  Branching Fractions: Summary
•BaBar preliminary measurements on 82 fb-1
•Becoming systematics limited
•Exp. vs. theory: data more accurate
than form factor predictions!
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
92
Direct CP Asymmetry: b →s and B  K* 
< 1% in the SM, could receive ~10% contributions from new EW physics
Either inclusive or exclusive decays could reveal new physics
B or K charge tags the flavor of the b quark with ~1-2% asymmetry systematic
Sum of 12 exclusive,
self-tagging
B→Xs  final states
b→s
b→s
Xs = K/Ks + 1-3 pions
E > 2.14 GeV
b →s ACP = (N – N)/(N + N) = 0.025 ± 0.050 ± 0.015
PRL 93 (2004) 021804, hep-ex/0403035
Asymmetries also measured precisely in exclusive K* decays:
B →K*
0- =
G(K*0 )
ACP = -0.013 ± 0.036 ± 0.010
–
G(K*-
submitted to PRL, hep-ex/0407003
) = 0.050 ± 0.045 ± 0.028 ± 0.024
preliminary
G(K*0 ) + G(K*- )
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
93
Time-Dependent CP Asymmetry in B →K* (113 fb-1)
As in B0→J/y KS, interference between mixed and
non-mixed decay to same final state required for CPV.
In the SM, mixed decay to K* requires wrong photon
helicity, thus CPV is suppressed:
In SM: C = -ACP ≈ -1%
S ≈ 2(ms/mb)sin 2 ≈ 4%
Measuring t of K*(→KSp0)  events requires novel beam-constrained vertexing techinque:
Vertex signal B with intersection of K S trajectory and beam-line
Usable resolution for KS decaying inside the silicon tracker
Validated with B0→J/y KS events
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
94
Time-Dependent CP Asymmetry in B →K* (113 fb-1)
Likelihood fit of three components
(qq, BB, K*)
to 5D data
(mES,E,Fisher,mK*,t)
preliminary
K* signal = 105 ± 14 events
S = +0.25 ± 0.63 ± 0.14
C = -0.57 ± 0.32 ± 0.09
submitted to PRL, hep-ex/0405082
Consistent with SM
For C fixed to 0, S = 0.25 ± 0.65 ± 0.14
First ever measurement of time-dependent CP asymmetries in radiative penguins!
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
95
Signal Fit: B+ → + 
+ + background
N(signal) = 26 ± 15 ± 2
significance = 1.9 s
e = 13.2 ± 1.4 %
preliminary
background
BF = (0.9 ± 0.6 ± 0.1) x10-6
BF < 1.8 x10-6 90% CL
Fit of four backgrounds
(qq, Xs , +p0/), K*)
and + signal
qq, Xs , + yield are free
parameters
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
96
Signal Fit: B0 → 0 
small peaking background
N(signal) = 0.3
+7.2 +1.7
preliminary
- 5.4 - 1.6
significance = 0.0 s
e = 15.8 ± 1.9 %
BF = (0.0 ± 0.2 ± 0.1) x10-6
BF < 0.4 x10-6 90% CL
Fit of four backgrounds
(qq, Xs , 0p0/), K*)
and 0 signal
qq, Xs , 0 yield are free
parameters
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
97
Signal Fit: B0 → w 
+5.7 +1.3
N(signal) = 8.3 - 4.5 – 1.9
preliminary
significance = 1.5 s
e = 8.6 ± 0.9 %
BF = (0.5 ± 0.3 ± 0.1) x10-6
BF < 1.0 x10-6 90% CL
Fit of two backgrounds
(qq, wp0/))
and w signal
qq, w yield are free
parameters
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
98
Summary
•Decay rate measurements of b→ s penguins are well into the precision era.
•CP asymmetries of b→ s penguins are statistics limited and will continue to test
the SM
ACP in B → Xs  PRL 93 (2004) 021804, hep-ex/0403035
BF, ACP, and Isospin asymmetry in B→ K* submitted to PRL, hep-ex/0407003
time-dependent CPV in B0 → K*0  submitted to PRL, hep-ex/0405082 First measurement!
b→ d penguins are only now beginning to reveal themselves in B-factory data.
They could also uncover new physics, or measure the poorly known |V td|.
BaBar finds no evidence for B →  in 211 million BB events
submitted to PRL, hep-ex/0408034
BF(B → ,w ) < 1.2 X 10-6
Rutgers 21 Mar 05
|Vtd/Vts| < 0.19 (90% CL)
Jeffrey Berryhill (UCSB)
99
B → K* (82 fb-1): Signal fits
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
100
B → K* (82 fb-1): K* lineshape
K* lineshape for
events with:
mES > 5.27 GeV/c2
-0.2 GeV < E < 0.1 GeV
Background subtracted
with shape from mES
sideband
Fit to relativistic
Breit-Wigner with
PDG GK*), mK*
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
101
Time-Dependent CP Asymmetry in B →K*
Z resolution vs. KS decay radius
Projection of fit with likelihood cut
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
102
B → : Neural net performance
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
103
B → : K Control Sample
0, no PID
, no PID
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
104
preliminary
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
105
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
106
CKM matrix constraint
Ali et al. hep-ph/0405075
SU(3) breaking of
form factors z2= 0.85 ± 0.10
weak annhilation
correction R = 0.1 ± 0.1
(z2,R) = (0.75,0.00)
theory error
(z2,R) = (0.85,0.10)
no theory error
Penguins are starting to
provide meaningful CKM
constraint
Reduction of theory errors
necessary to be competitive
with Bd,Bs mixing
 95% C.L. BaBar allowed region (inside the blue arc)
Rutgers 21 Mar 05
Jeffrey Berryhill (UCSB)
107