Matter-Antimatter Asymmetries and CKM Parameters in BABAR

1
Matter-Antimatter Asymmetries and CKM Parameters
in BABAR

• Jeffrey D. Richman
• University of California, Santa Barbara

Representing the BABAR collaboration
Meeting of the Particle Physics Program
Prioritization Panel (P5) Oct. 6, 2005
Version 3.0
2
Outline

• Where are we in B physics?
• A high-precision, benchmark measurement sin2b
  from B?J/y K0
• a a work in progress
• A path to g
• Vcb, Vub, heavy-quark masses, and QCD
  parameters
• Perspective and conclusions

• Zoltan Ligeti (theory) discussion of
  theoretical issues uncertainties new physics
• Luca Silvestrini (theory) new physics
  sensitivity
• Riccardo Faccini BABAR measurements related to
  new physics and rare decays, including sin2b from
  b?s penguin modes

3
Exclusive B decays CLEO (1983)
B0B0 oscillations ARGUS (1987)
Observation of B?K g CLEO II (1993) Loops!
Long B lifetime MAC, Mark II (1983)
B?Xu l n and Vub ARGUS, CLEO (1990)
B?D l n and Vcb ARGUS, CLEO, LEP, Isgur, Wise
(gt1989) HQET!
4
The Current Era in B Physics
Dramatic advances in our knowledge of the
CP-violating phase structure of quark
interactions.

• First achievement clear and unmistakable
  evidence for large (order unity) CP violation in
  the B meson system.
• Amazing stream of surprising results and new
  methods. Many of these would not have appeared in
  an extrapolation from the past.
• Detector technology can search for essentially
  any type of B decay. Trigger on all events
  Tracking/Vertexing CsI PID

Some notable or surprising measurements
s, p wave
5
Probing the CKM quark mixing matrix

• Angles of triangle measure from CP asymmetries
  in B decay
• Sides of triangle measure rates for b?uln, B0B0
  mixing
• Other constraints in r,h plane from CP violation
  in K decay

6
CP asymmetry from interference between mixing and
decay
BABAR, PRD 70, 012007 (2004)
HFAG
1 decay amplitude, q/p1
1 decay amp magnitude strong phase divide out!
7
sin2b as a precision measurement

• The ccs sin2b determination belongs to a special
  class of definitive
• measurements in particle physics.
• We can achieve high statistical precision before
  we are limited by systematic uncertainties.
• It is a data-driven measurement, with very little
  dependence on Monte Carlo or theoretical
  assumptions.
• Theoretical uncertainties lt1, so its
  interpretation is clear (and powerful) Ligeti,
  Silvestrini

8
BABAR sin2b from charmonium (227 M BB)
J/? KL (CP even) mode
asymmetry is opposite!
PRL 94, 161803 (2005), (hep-ex/0408127)
sin2b 0.722 ? 0.040 (stat) ? 0.023 (sys)
l 0.950 /- 0.031 (stat) /- 0.013 (sys)
hypothesis test
(after raw asymmetry shown above is corrected for
the dilution)
9
Foundations of the sin2b measurement
Dt lt 0
Dt gt 0
Background
Mistag rates w(tag)
Dt resolution function
Mixing asymmetry
log scale
MES (GeV)
Si detector alignment, beam spot
Signal 7,730 events (all modes) Control 72,878
events D() p,r,a1,J/yK
10
Dttrec-ttag fits to BFlav control sample
Mixed events
log scale
Mixed events
linear scale
Unmixed events
log scale
Unmixed events
linear scale
11
Mistag (w) measurement from BFlav oscillation
data
Separately determine D for each tag category.
D(1-2w)lt1 due to mistags
T2p/Dm
Overall tagging performance
tB1.6 ps
12
Systematic Errors for sin(2b)
Category s(sin2b) 226 BB s(sin2b) at 1 ab-1 (est.)
Background shape CP content of peaking background 0.012 0.004 to 0.006
Mistag differences between BCP and Bflav samples 0.007 0.003
Composition and content of J/y KL background 0.011 0.005 to 0.009
Dt resolution and detector effects silicon detector alignment and Dt resolution model 0.011 0.004 to 0.008
Beam spot position 0.007 0.004 to 0.007
Fixed Dmd, tB, DG/G, l 0.005 0.002
Tag-side interference DCSD decays 0.003 0.003
MC statistics, bias 0.003 0.001
TOTAL 0.023 0.01 to 0.016
Some systematics scale with 1/sqrt(N) other
partially do.
13
sin2b uncertainties vs. integrated luminosity
Current systematic uncertainty
Range of estimated systematic error 1 ab-1
(109)
At 1 ab-1, we can improve sin2b by nearly a
factor of 2.
14
a A work in progress
Original idea for measuring a Works if B0?pp-
amplitude is dominated by the b?u tree process.
If penguins were negligible, we could extract a
directly from the time-dependent CP asymmetry for
B0?pp- with no additional information.
15
The penguin problem in B0 (B0 ) ?pp-

• In 1998, CLEO performed a search for charmless
  two-body B decays. Did not observe B0?pp- ,
  but found large B0?Kp- rate CLEO, PRL 80, 3456
  (1998).
• We cannot ignore penguin amplitude in B0?pp-.
  (In fact, P-T interference produces direct CP
  violation in B0?Kp- and may also in B0?pp-).

We still measure S and C, but S isnt sin2a!
16
I-spin solution to the penguin problem Gronau
London, PRL, 65, 3381 (1990)
Use I-spin invariance of hadronic matrix elements
to relate B?pp amplitudes. Assume that pions are
identical particles.
?triangle relations
Penguins DI1/2 only, so no contribution to
B?pp0 .
amplitudes cancel
17
Constraining Da with I-spin relations
B?pp0 is pure tree (no gluonic
penguin)?triangles have common side after
rescaling one set by exp(2ig)

• If penguin amp0,
• triangles coincide.
• 4-fold discrete ambiguity (can flip both
  triangles)
• take worst case as penguin error

2Da
Grossman Quinn, PRD 58, 017504 (1998)
18
Measurements of B0?p0p0, B0?pp-, and B?pp0
Mode
B/10-6 (BABAR)
B/10-6 (Belle)
p0p0 amp. isnt small compared to the others.
BABAR PRL 94, 181802 (2005)
BABAR
Red triangles B and B0 decays Purple
triangles B- and B0 decays Difference CP
violating interference between T and P amplitudes.
19
Huge program on B decays to charmless hadronic
final states...
Bigger than pp
BABAR, PRL 94, 131801 (2005)
(10-6)
20
The investigation of B?rr

• BABAR has made intensive effort to study the B?rr
  modes
• Measurement of B?rr0 , B0 ?r0r0 limit PRL 91,
  171802 (2003).
• 1st observation of B?rr- and polarization
  measurement
  PRD 69, 031102 (2004)
• First time-dependent CP asymmetry measurement
  and confirmation of polarization. PRL 93,
  231801 (2005)
• Updated time-dependent CP asymmetry measurement
  with Run 1-4 data. hep-ex/0503049
  ?PRL
• Limit on B0?r0r0 branching fraction PRL 94,
  131801 (2005)

B/10-6 (Belle)
Mode
B/10-6 (BABAR)
BABAR, PRL 94, 131801 (2005)
21
Measurement of CP asymmetry for B?rr-
Is the rr system in a CP eigenstate? If not, get
effective dilution of CP asymmetry.
BABAR, PRL 95, 041805 (2005)
232 M BB
B0 tags
B0 tags
Angular analysis ? almost pure CP1 !
BABAR BELLE (LP2005)
fL
S??
C??
Dt (ps)
Would like to see S, C with 5x data!
22
a combining the BABAR measurements
B ?pp
B ???
PRL, 94, 181802 (2005)
PRL 95, 041805 (2005)
a 100º ? 13º
1s
29º61º excluded _at_ 90 C.L.
79ºlt a lt123º _at_ 90 C.L
hep-ex/ 0408089
B ?pp-p0 Dalitz
1-C.L.
CKM fit excluding a measurements
23
Projections for a measurement in B?rr-
Current a measurement from B?rr
Multiple unresolved solutions within each peak.
1s
90 C.L.
Projected a measurements from B?rr for 1 ab-1
1s
B(B?r0r0) unchanged

• The uncertainty on a depends
• critically on B(B?r0r0).
• Scenarios
• use current central value
• 1s
• - 1s

-1s
24
Critical issue for a measurement B?r0r0
I-spin triangle for B?rr (current measurements)
Projected 2s uncertainties on a
Projected 1s uncertainties on a
25
Goals and issues for the a program

• B?pp
• Resolve issues with S and C Belle observes
  significant direct CP violation in B?pp BABAR
  doesnt.
• BABAR and Belle values of B?p0p0 are higher than
  theoretical expectations (and differ by x2) and
  are not precisely measured.
• B?rp
• Complicated Dalitz-plot measurement currently
  disfavors one of the solution regions allowed
  from B?rr. Will this hold up with more data?
• B?rr
• Need to observe B?r0r0. Value is critical in
  constraining the I-spin triangle and
  determinining penguin-induced uncertainty on a.
• Is I-spin conserved? Does the triangle close?
• Non-resonant background studies indicate is
  small effect but more data would allow more
  detailed investigation.
• Improve measurements of S and Calso investigate
  B?a1p-

26
A path to g
color suppressed
How can we get interference? Need D0 ?f and D0
?f. (Compare with B0?J/y K0.) Some observations

1. Uses charged B decays method is based on a
   direct CP asymmetry. Issues strong phase d,
   rBA(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.

27
g (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).
Large rate, but interference is small rB ltlt 1
28
g (ADS method) B- ? D0?Kp - D0?Kp -K-
Atwood, Dunietz, Soni, PRL 78, 3257 (1997), PRD
63, 036005 (2001)
DCSD
Interference is large rB, rD comparable, but
overall rate is small!
29
g (Dalitz plot) B- ? D0?Ks p p- D0? Ks p
p- K-,
Giri, Grossman, Soffer, Zupan, PRD 68, 054018
(2003), Bondar (Belle), PRD 70, 072003 (2004)
2
M-2
Relatively large BFs all charged tracks only
2-fold g ambiguity.
Interference depends on Dalitz region
(CP), (DCSD)
g ambiguity only 2-fold (g ? gp)
30
Fitting the D0?KSpp- Dalitz plot
BABAR
hep-ex/0504039
Use continuum data D?D0p (91.5 fb-1)
CA K(892)
Nevts 82 K Purity 97
r(770)

• Issue contribution of
• broad, s-wave resonances
• Orig. method 2 BWs
• New K-matrix

Anisovich Saratev Eur. Phys. J A16, 229 (2003)
DCS K(892)
c2/dof?3824/30221.27
31
B/-?D0K/- KS pp- Dalitz plot distributions
B?D0K
B-?D0K-
B?D0K
B-?D0K-
Differences between B and B- signifies direct CP
violation.
Above, D0 is super- position of D0 and D0
Good S/B, but needs more data.
32
g BABAR and Belle results (Dalitz method)
BABAR (statsysmodel) Belle (statsysmodel)
rB (D0K)
rB (D0K)
rB (D0K)
g
direct CP significance
hep-ex/0504039, 0507101
hep-ex/04110439, 0504013
non-K
Importance of rB
The error on g is very sensitive to the value of
rB. Other methods (ADS, GLW) help us to
measure rB.
(degrees)
0.1
0.2
33
rB measurements from ADS channels
Most measurements using interference with DCSD D0
decay indicate rBlt0.2.
34
Projected uncertainty on g for rB 0.1
Projected sys error due to D0 Dalitz plot
We will be able to improve the error on g by at
least a factor of 2.
35
Surprises in semileptonic B decays

• Two complementary experimental and theoretical
  approaches
• Exclusive decays measure (and predict) the rate
  for specific exclusive modes, usually in
  restricted region of phase space.
• Inclusive decays use as much of phase space as
  possible to minimize theoretical input. Extract
  non-perturbative QCD parameters from data.
  Goal Vij (exclusive) Vij (inclusive)!

36
Vcb and the atomic physics of B mesons
Extract Vcb , quark masses, and
non-perturbative QCD parameters from measured
inclusive lepton-energy spectrum and hadron
recoil mass spectrum (masses, QCD params given
below kinetic scheme). Yields Vcb
to about 2. (lattice QCD goal 3 for
B?Dln)
chromomagnetic expec value
Darwin term
spin-orbit
kinetic expectation value
Benson, Bigi, Mannel Uraltsev, hep-ph/0410080
BABAR, PRL 93, 011803 (2004)
Gambino Uraltsev, Eur.Phys.J. C34, 181 (2004)
37
Why measuring Vub is hard
Large b?c background suppression cuts introduce
dependence on theory predictions for kinematic
distributions.
Lepton spectrum endpoint analysis
Fully reconstructed B recoil analysis
p
BABAR
Breco
e-
(hep-ex/0509040)
D
e
n
continuum data (off res)
Xu
l
Brecoil
BABAR
b?c subtraction
b?u
38
Vub inclusive measurements

• Key CKM constraint
• Use mb and QCD parameters extracted from
  inclusive B?Xc l n and B?Xs g spectra.
• Many methods with uncertainties around 10.
• Uncertainty from mb has been reduced to 4.5.
• With more data, the Vub uncertainties could
  be pushed down to 5-6.5.

El endpoint
El vs. q2
mX
mX vs. q2
mb, theory
expt
39
B0?p l- n form-factor predictions
Measuring Vub using B?p l n and lattice QCD
f(q2) is relevant form factor for B?p l n (le,
m)
Fermilab/MILC
HPQCD
At fixed q2, lepton momentum spectrum is
exactly known in this mode, since only one
form factor.
HPQCD hep-lat/0408019
restricted q2 range
Fermilab/MILC hep-lat/0409116
40
Experiment vs. Lattice D?K l n form factor
41
Measuring Vub using B?p l n
Projection to 1 ab-1 (data taken to be on BK fit
curve from present measurement).
BABAR
PRD 72, 051102 (2005)
In the high q2 region alone, we will measure the
branching fraction with an uncertainty of (6-7)
, or (3-3.5) uncertainty on Vub . Lattice
theorists expect to reach 6, so
exclusive/inclusive will be similar.
42
Perspective/Conclusions
Four major measurement programs related to
determining the values of fundamental Standard
Model CKM parameters.
Parameter Goals for 1 ab-1 Methods
sin2b (f1) Measure sin2b to /- 0.025 or bettter. t-dependent CP asymmetry measurements Dt resolution and mistags measured in data
a (f2) Understand B?pp, B?rp, B?rr decays. Measure a to 8-16 degrees (depends on B?r0r0). t-dependent CP asymmetry measurements isospin analysis of related modes B?rr appears to be the most powerful.
g (f3) Understand B?D K decays measure g to 9 degrees or better (depends on rB). Direct CP (t-independent) measurement determine strong phases relative amplitudes from data. D0 Dalitz-plot analysis(D0?CP) (D0?DCSD) also measure sin(2bg).
Vub Understand inclusive B?Xc l n, inclusive B?Xu l n, and B?p l n measure Vub to 6.5 or better. Inclusive determine heavy-quark parameters from kinematic distributions Exclusive pln at high q2 lattice QCD.
43
Perspective/Conclusions

• Many measurements are now multidimensional
  extract not only the
• quantity of interest, but also critical
  information that is difficult
• to get from theory. Examples
• a and g measurements are data-driven isospin
  triangle, rB, etc.
• b-quark mass and other QCD parameters are now
  well determined from Vcb studies this
  information is used as input for the Vub
  measurement.

• CKM measurements go hand-in-hand with other parts
  of the BABAR physics program
• Enormous program of hadronic rare B decay
  studies
• Search for departures from CKM pattern using b?s
  decays
• Studies of electroweak penguin and leptonic
  decays
• Charm physics, including searches for mixing and
  CP violation.

These are great ideas and measurements this is a
great physics program!
44
Backup slides
45
CP Asymmetries formulas and definitions
no net oscillation
no net oscillation
net oscillation
net oscillation
For DG/G ltlt1,
46
Behavior of time-dependent CP asymmetries
Linear scale
Non-exponential decay law for a specific final
state!
Log scale
47
Angles of the unitarity triangle
Consider two complex numbers z1 and z2.
48
The CKM matrix and its mysterious pattern
(Wolfenstein parametrization)

• The SM offers no explanation for this numerical
  pattern.
• But SM framework is highly predictive
• Unitarity triangle (Col 1)(Col 3) 0 etc.
• Only 4 independent parameters A, l, r, h
• One independent CP-violating phase parameter

49
Comments on B physics history (see slide 3)

• Exclusive B decays Reconstruction of b?c modes
  requires charm meson reconstruction. The product
  branching fractions for B?D()X, D()?K(np) modes
  are typically of the order 10-4 to 10-5, so large
  data samples are needed. The 1st exclusive B
  signal from CLEO was made by summing over several
  different modes.
• Long B lifetime showed that Vcb was smaller than
  expected. We began to see the larger pattern of
  the CKM matrix outside the 2x2 Cabibbo sector
  Vcb is proportional to l2 , not l. This
  measurement also demonstrated the critical
  importance of high-precision tracking and
  provided a strong impetus to the development of
  Si vertex detectors.
• BB oscillations this critical discovery was
  made by ARGUS. The oscillation period is about
  12.6 ps (6.3 ps for maximal probability to
  oscillate), which is about 8x larger than the
  mean decay time of 1.6 ps. CP violation in mixing
  is a very small effect in B decays, since the
  off-shell intermediate states such as tt
  completely dominate over on-shell intermediate
  states. CP violation requires interference
  between these two paths. This simplifies the
  BABAR/Belle CP violation measurements, which are
  based on a different effect the interference
  between mixing and decay amplitudes.

50
Comments on B physics history (see slide 3)

• Observation of charmless semileptonic B decays by
  ARGUS and CLEO was a critical discovery. The
  measured value of Vub/Vcb maps out an annular
  region in the r-h plane. The consistency between
  this region and the BB mixing and eK regions
  provided an early test of the CKM framework. In
  the Vub measurement, the lepton spectrum endpoint
  region was used, because backgrounds from b?cln
  decays are suppressed compared with b?uln, where
  the lepton can be more energetic. Later
  measurements use a variety of techniques to
  increase the phase space region used and to
  thereby decrease theoretical uncertainties.
• The observation of B?Kg by CLEO was a major
  discovery, demonstrating the presence of loop
  processes at the rate expected in the SM.
  BABAR/Belle are studying a very large number loop
  processes in both exclusive and inclusive
  measurements. These processes provide a powerful
  probe of physics beyond the SM through virtual
  effects.
• Vcb measurements were given a strong boost by the
  development of Heavy Quark Effective Theory
  (HQET). This and subsequent theoretical advances
  have substantially improved our understanding of
  the dynamics of B decays.

51
A simplified picture of the CKM matrix
Largest phases in the Wolfenstein parametrization
Magnitudes of CKM elements
l3
l
1
1
l2
1
l
1
l3
l2
Note all terms in the inner product between
columns 1 and 3 are of order l3. This produces a
unitarity triangle of roughly equal sides.
52
sin2b measurement signal modes
signal region
yield
signal region
yield
MES (GeV)
MES (GeV)
CP sample NTAG purity ?CP
J/? KS (KS?pp-) 2751 96 -1
J/? KS (KS?p0p0) 653 88 -1
?(2S) KS (KS?pp-) 485 87 -1
?c1 KS (KS?pp-) 194 85 -1
?c KS (KS?pp-) 287 74 -1
Total for ?CP-1 4370 92 -1
J/? K0(K0? KSp0) 572 77 0.51
J/? KL 2788 56 1
Total 7730 78
signal region
yield
BABAR
J/? KL signal
J/? X background
Non-J/? background
?E (MeV)
53
Control samples for sin2b
Use neutral B control sample (BFlav) to
determine tagging dilution and Dt resolution
parameters from simultaneous fit to the data.
MES (GeV)
MES (GeV)
BFlav sample NTAG purity
D- p/?/a1 (6 decay modes) 32974 83.1
D- p/?/a1 (12 decay modes) 35008 89.4
J/? K0(K0? Kp-) (2 modes) 4896 95.8
Total 72878
54
Tagging algorithm performance
e() w() Dw() Q()
Lepton 8.6/-0.1 3.2/-0.4 -0.2/-0.8 7.5/-0.2
KaonI 10.9/-0.1 4.6/-0.5 -0.7/-0.9 9.0/-0.2
KaonII 17.1/-0.1 15.6/-0.5 -0.7/-0.8 8.1/-0.2
K-p 13.7/-0.1 23.7/-0.6 -0.4/-1.0 3.8/-0.2
p 14.5/-0.1 33.9/-0.6 5.1/-1.0 1.7/-0.1
Other 10.0/-0.1 41.1/-0.8 2.4/-1.2 0.3/-0.1
Total 74.9/-0.2 30.5/-0.4
Lepton category not sensitive to mistag
differences due to DCSD decays
55
New Belle result (summer 2005)
Belle-CONF-0569 hep-ex/0507037
J/? KL (CP even) mode
J/? KS (CP odd) mode
386 M BB
56
BABAR and Belle Systematic Errors(for
aficionados)
BABAR, b?ccs modes only
Belle, b?ccs and b?sss modes
57
sin2b results from charmonium summary
from Kazuo Abes talk at LP05
backup slide
58
backup slide
59
BABAR and Belle time-dependent CP asymmetry
results for B0?pp-
Belle observes significant direct CP violation in
B0?pp-.
BABAR-Belle 2.3s
60
Asymmetry for direct CP violation

• Problems with interpreting measurements of direct
  CP asymmetries
• We often dont know the difference d1-d2 , so we
  cannot
• extract f1-f2 from the asymmetry without
  additional information.
• 2. We often dont know the relative magnitude
  of the interfering amplitudes.

61
Direct CP violation in B0?K-p/B0?Kp-
a short digression
The tree-penguin interference that is bothering
us in B0?pp- shows up spectacularly as direct CP
violation in B0?Kp-.
Bkgd symmetric!
62
a direct measurement vs. CKM fit
63
Combining pp, rp, rr measurements of a
B?rp CP(t) asymmetry disfavors rr mirror
solution
a inferred from other CKM measurements
64
Sensitivity to g across the Dalitz plot
Monte Carlo
65
g from B?DK (all methods)
66
B-?DCPK-, DCP?fCP (GLW) Fit Results
67
GLW method D decays to CP eigenstates
3 unknowns g, d, rB 4 observables (3 independent
relations)
In principle can solve for everything! (Just
need a lot of data.)
Dont yet observe significant asymmetries.
68
ADS method
Note both the B and D diagrams are different
now ? two strong phases (but they combine)
Need one more observable measure another D final
state (same dB but different dD. Now have g,
rB, dBdD1, dBdD2
Two observables 4 unknowns g, rB, rD, ddBdD
Basically a good method, but no significant
signals yet!
69
B-?DCPK-, DCP?fCP (GLW) Fit Results
70
B?p l n
q2 distribution from various experiments
Branching fraction measurements,
including restricted q2 region.
71
The CKM Triangle Using Angles Only