Charm Physics in BaBar

1
CHARM PHYSICS in
Alexis POMPILI (University I.N.F.N. of
Bari) representing the Collaboration
Les Rencontres de Physique de la Vallee dAoste
La Thuile March 3, 2004
2
Charm Physics _at_ BaBar - Outline
Analyses results reviewed here
D0 mixing with 2-body hadronic decays (wrong sign
CP final states)
states spectroscopy
Other Studies in progress
D0 mixing with semi-leptonic decays
3-body decays of D0, D, Ds ( B. R., Light
Meson Spectroscopy, CP, mixing )
charmed baryons
ISR processes
2
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
3
Charm Physics _at_ PEP-II B-factory
PEP-II (SLAC) asymmetric ee- collider _at_
Y(4S) Integrated luminosity delivered 178fb-1
Effective qq cross sections at the energy of the
Y(4S)
Powerful tool for charm physics
BABAR (90fb-1) 220K Focus
120K E791 36K
3
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
4
The detector
Crucial for charm physics
PID
CERENKOV DETECTOR
e
DRIFT CHAMBER
TRACKING
VERTEXING
SILICON VERTEX
e-
KAON EFFICIENCY
K-Id
CHERENKOV
PION CONTAMINATION
Reconstructed K-K Mass (GeV/c2)
4
La Thuile - 2004
Alexis
Pompili (U. of Bari I.N.F.N.)
5
Selecting charm decays from continuum
Kinematical selection require cut on CMS
momentum of charmed meson
combinatorial background strongly reduced
D() from B-decays rejected
D-TAG with pS
allows flavor-tagging
reduces BKGD through dm m(K-pps) - m(K-p)
D0
Refitting technique with beamspot - constraint
pS
D
improves dm resolution
Beamspot
5
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
6
D-mixing parameters search methods
Mass eigenstates
( flavour eigenstates)
masses M1 , M2 widths G1 , G2
MIXING PARAMETERS
beyond the reach of current experimental
sensitivity
SM
NEW PHYSICS may enhance x
but
FSI SU(3)-breaking can enhance y
Experimental goal to put limits on the
transition D0-D0
1) LIFETIME DIFFERENCE searches
y if CP conserved
2) WRONG-SIGN searches in hadronic decays
x2 , y d unknown strong phase difference
NP may not conserve CP
consider CP when measuring mixing
CP in mixing ( sure sign of NP )
CP in interference mixingdecay
6
7
Lifetime difference searches
Mixing would alter the decay-time distributions
of D0 and D0 that decay into CP eigenstates. They
can be considered to a good approx. as pure
exponential with effective lifetimes
for CSD of D0 (D0) into CP-even final states
(such as KK-,pp-)
for CFD
LIFETIME RATIO
These effective lifetimes can be combined into

where
IF NO MIXING x y 0
IF CP-CONSERVATION IN MIXING rm1
IF ALSO CP-CONSERVATION IN INTERFERENCE MIXING
-DECAY
Systematics effects on lifetime tend to cancel in
the lifetime RATIO
7
La Thuile - 2004
Alexis Pompili (U. of
Bari I.N.F.N.)
8
Mass fits
DATA FIT RESULT
265K
26K
SIGNAL (S)
99.4
97.0
yield
FROM THE FITS
BKGD (B)
S/SB
mass window
145K
13K
68.1
Mass fits determine event-by-event signal
probability
87.9
BKGD constrained in data
sidebandscandidates included as part of proper
time fit
8
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
9
Decay-time distributions fit
within mass signal window
SIGNAL
FROM THE FITS
BKGD
DATA UNBINNED MAXIMUM
LIKELIHOOD FIT RESULT
9
La Thuile - 2004

Alexis Pompili (U. of Bari I.N.F.N.)
10
Summary of lifetime ratio results
Phys.Rev.Lett.91,12(2003)
hep-ex/0308034
BaBar has the most stringent limits on y
BaBar has the first measurement of a CP parameter
with the method of lifetime ratio

10
La Thuile - 2004

Alexis Pompili (U. of Bari I.N.F.N.)
11
Wrong-sign searches in hadronic decays
Study time-evolution of
looking for a signal
dKp unknown relative strong phase (not
measurable in a mixing analysis) rotation of
dKp in the x-y plane
DCSD CFD may have different FSI
Assuming and CP conservation

where
deviation from pure exponential
DCSD
MIXING
INTERFERENCE
enhances mixing signal
Integrating
mixing rate
Additional CP effects are included by measuring
this distribution for D0 and D0 separately
11
CP in DCSD
CP in MIXING
CP in their INTERFERENCE
12
Mass-mass difference Fit
Assign each candidate to one of 4 categories
D0(RS,WS),D0(RS,WS)
Unbinned extended max-likelihood fit to RS and WS
samples simultaneously in 4D variable space

performed in steps
proper time error
1st step fit
m(Kp)
d m
Number of SIGNAL and BKGD candidates from a fit
to m-dm plane (both RS and WS) by modelling BKGD
categories
Signal region
Signal
WS
Wrong slow pion
Combinatorial
Sideband
Double mis-id
With 57fb-1 120,000 RS
440 WS
12
La Thuile - 2004

Alexis Pompili (U. of Bari I.N.F.N.)
13
Decay time Fit
2nd step fit
Simultaneous fit time-distribution to RS and WS
(include mixing parameters). The larger/clean RS
sample fixes D0 lifetime and resolution model
parameters for unmixed decays. BKGD time
distributions determined from m,dm sidebands in
data.
WS
Signal
Wrong slow pion
Double mis-id
Combinatorial
13
La Thuile - 2004

Alexis Pompili (U. of Bari I.N.F.N.)
14
Summary of wrong-sign results
Phys.Rev.Lett.91,17(2003)
Results presented in 4 cases
95 C.L. contours by a frequentistic
approach (based on toy Monte Carlo experiments)
BEST FIT (x2 free) NO CP
BEST FIT (x20)
(stat.syst.)
(stat.)
(Prelim.)
CP allowed
14
La Thuile - 2004
Alexis
Pompili (U. of Bari I.N.F.N.)
15
Two surprising new Charmed Mesons

Phys.Rev.Lett. 90,24(2003)
BaBar discovered a new narrow state in the Ds
p0 invariant mass distribution near 2.32GeV/c2
!
!
Accepted by P.R.L. hep/ex-0310050
BaBar observed a second new narrow state in the
Ds p0 invariant mass distribution near
2.46GeV/c2
sidebands subtraction
15
La Thuile - 2004
Alexis
Pompili (U. of Bari I.N.F.N.)
16
Channel Likelihood Fit
To disentangle these 2 possible DsJ(2458) decay
modes extract signal parameters unbinned
maximum likelihood fit using the channel
likelihood method
contribution from DsJ(2458) decays
combinatorial background
BELLE
CLEO observed them in continuum and Belle both
in continuum and in B decays

CLEO
16
17
SPIN-PARITY
Decay to JP0- mesons only natural
spin-parity allowed 0,1-,2,
JP0 suggested by 1) low mass compared to
Ds1(2535) DsJ(2573)
2) absence of decay to Dsg (not allowed
if JP0) 3)
absence of decay to Dspp- (not allowed if JP0)
Un-natural spin-parity more likely (lack of
decays to DK)
Belle helicity analysis from B-decays favours J1
Decay to Dspp- (by Belle) allowed by JP1
Two narrow states observed, in the inclusive Dsp0
Dsp0 invariant mass distributions, near
2.317GeV/c2 2.458GeV/c2. The widths Glt10MeV
are consistent with experimental resolution. The
most likely assignment for their spin-parity is
0 1.
EXPERIMENTAL SUMMARY
17
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
18
Spectroscopy of states (before after)
Potential models of heavy-quark light-quark
mesons so far reasonable success for
spectroscopy of D, Ds, B, Bs systems
New states do not fit well masses below the
DKDK threshold.
IF interpreted as ordinary states,
they decay mainly by isospin-violating p-emission
thus having widths quite narrow.
?
A possible decay mechanism is through a virtual h
followed by h-p0 mixing Cho-Wise,PRD49.
?
DK
p
DK
p
g
as predicted by models based on HQET chiral
symmetry Bardeen et al., if new states are 0
1
p
NEW STATES
40(!) papers by theorists Exotic (4-quark,
molecule, ) VS Ordinary explanations
(HQETchiral symmetry, )
P-wave
S-wave
j3/2 NARROW by K-emission
Crucial to measure radiative decays di-pion
emission
18
19
SUMMARY
Competitive D0 mixing results obtained with 2
different experimental methods are consistent
with no mixing and no CP.
wrong-sign D0 mixing limits (95
CL,57fb-1CP-allowed)
D0 lifetime ratio (91fb-1)
Wrong-sign D0 mixing limits from semileptonic
decays coming soon
Charmed mesons are a rich laboratory for Heavy
Quark studies.
After the discovery of DsJ(2317) and the
observation of DsJ(2458) further Ds mesons
(spectroscopic) studies are ongoing.
BaBar is a promising place to study charm physics
studies have just began !
19
La Thuile - 2004
Alexis Pompili (U. of Bari
I.N.F.N.)
20
Backup Slides
21
Unbinned maximum likelihood fit
S1
S2
s3
tD
(1-f2)(1-f3)
f2(1-f3)
f3
Psig(mi)
SIGNAL
proper-time RESOLUTION
tBKG
e
BUT with ...
SBKG S1
f0(1-f4)
(1-f0)
SBKG S2
NO-flying BKG
1-Psig(mi)
flying BKG
s3
s4 fixed
BKG
f0f4
for tails
BACKUP-1
La Thuile - 2004
A.Pompili (U. of
Bari I.N.F.N.)
22
Statistical error on lifetime extracted from the
fit
BACKUP-2
La Thuile - 2004
A.Pompili (U. of
Bari I.N.F.N.)
23
Fit allowed x2lt0. Central values for fits of
both separate samples gave x2lt0.
Due to allowing x2lt0 in the fits it is not
clear how to apply a Bayesian ansatz to derive
an error estimate from the 2D likelihood
distributions. Moreover these ones depend
strongly on the most likely fitted values of x2
and y.
95 C.L. limits are determined using toy Monte
Carlo samples at each point on the
contour (frequentistic approach).
(x2,y) points on separate contours are combined
in pairs to determine (x2,y) on 95 CP contours.
Systematic uncertainties included by
calculating equivalent statistical deviation for
each systematic check and expanding the 95 C.L.
contour appropriately.
BACKUP-3
La Thuile - 2004
A.Pompili (U. of
Bari I.N.F.N.)
24
BACKUP-4
La Thuile - 2004
A.Pompili (U. of
Bari I.N.F.N.)
25

• Unbinned maximum likelihood fit
• Assign likelihood to each event for
• DsJ(2458) ? Ds(2112) ?0
• DsJ(2458) ? DsJ(2317) ?
• Combinatorial Ds ?0 ? bkgd
• Ds(2112) random ?0
• DsJ(2317) random ?
• DsJ(2458) ? Ds(2112) ?0
• using wrong ?

The DsJ(2458) signal for a particular decay mode
can be isolated by calculating a weight for
each Dsp0g combination proportional to the
relative likelihood contributed by the decay mode
of interest
BACKUP-5
La Thuile - 2004
A.Pompili (U. of
Bari I.N.F.N.)