Recent Results on D0D0 Mixing from BaBar

1
Recent Results on D0-D0 Mixing from BaBar

• William Lockman
• for the BaBar Collaboration
• Lepton-Photon 2007, Daegu, S. Korea

2
Topics from BaBar Experiment

• Introduction
• D0-D0 Mixing in Lifetime Ratio of D0?KK?, ???
  vs D0?K??
• Search for CP Violation in D0?KK? and D0?????
• to be submitted to PRL
• D0-D0 Mixing in the Decay D0?K?????
• Conclusion

3
Charm meson mixing

• Why would observation of charm mixing be
  interesting?
• It would complete the picture of quark mixing
  already seen in the K, B, and Bs systems.
• K 1956
• Bd 1987
• Bs 2006
• It would provide new information about processes
  with down-type quarks in the mixing loop diagram.
• It would be a significant step toward observation
  of CP violation in the charm sector.
• It could indicate new physics.

4
Current Evidence for D0-D0 mixing
D0?K?-
PRL 98,211802
Combined
3.9s signal
5.7s signal
BELLE
y ()
2.4s signal
arXiv0704.1000
D0?K?K?, ???-
D0?Ks???-
x ()
5
Flavor States Mixing

• Flavor eigenstates can mix through weak
  interaction
• Mass eigenstates
• Flavor state time evolution
• Mixing if either or
  nonzero

6
Standard Model Predictions

• Short-distance contributions from mixing box
  diagrams in the Standard Model are expected to be
  small
• b quark is CKM-suppressed
• s and d quarks are GIM suppressed
• mainly contributes to the mass difference
• x O(10-5) or less
• Long-distance contributions dominate but hard to
  estimate precisely
• expect y 0.01
• x 0.1 - 1y

7
BABAR Charm Factory 1.3 million Charm events per
fb-1
Integrated luminosity 384 fb-1 used for mixing
results presented here 500M cc events
BaBar is a large acceptance general purpose
detector providing excellent tracking, vertexing,
particle ID and neutrals detection
8
BaBar Generic Mixing Analysis

• Identify the D0 flavor at production
• using the decays
• select events around the expected
• The charge of the soft pion determines the
  flavor of the D0
• Identify the D0 flavor at decay
• using the charge of the Kaon
• Vertexing with beam spot constraint
• determines decay time,
  
• and decay time error,

D0 decay vertex
Beam spot ?x ¼ 100 ?m, ?y ¼ 6 ?m
right-sign (RS) wrong-sign (WS)
D0 production vertex
9
D0-D0 Mixing in Lifetime Ratio of D0?KK?, ???
vs D0?K??

• D0?K??? CP-mixed D0(t)? K?K?, ????
  CP?even
• Determine the quantities
• If CP is conserved in mixing and decay, but
  violated in the interference
• between them, these quantities are related to the
  mixing parameters

CPV in interference of mixing and decay
10
Previous lifetime ratio results
BELLE, PRL 98, 211803 (2007) 540 fb-1
11
BaBar (yCP, ?Y) analysis overview

• Select D??D0????D0?K?????K?K??????? decays from
  384 fb-1
• Event selection chosen to minimize backgrounds
• systematics affecting signal mostly cancel
• background systematics dont cancel between
  modes
• Unbinned likelihood fit to (t,?t) to obtain ?hh
• signal resolution determined from fitting data
• Backgrounds taken from MC and sidebands
• Determine ycp and ?Y from
  lifetimes

12
Decay time fits to determine (yCP, ?Y)

?409.30.7 fs
?401.32.5 fs
?404.52.5 fs
?407.63.7 fs
?407.33.8 fs
K? and KK lifetimes differ!
13
BaBar (yCP, ?Y) results

• Tagged results from 384 fb-1
• Result in good agreement with BELLE measurement

14
BaBar (yCP, ?Y) systematics

• Systematic uncertainties ()
• Variations
• Signal PDF shape, polar angle dependent
  resolution offset, signal interval
• Charm backgrounds yields and charm lifetime
• Combinatorial backgrounds yields, shape and
  sideband region
• Selection ?t criterion, treatment of multiple
  candidates
• Detector Alignment and energy loss

15
Search for CPV in D0? K?K?, ?????
2 weak amplitudes with phase difference
strong phase difference

• Two amplitudes with different strong weak
  phases needed to observe CPV (in SM from tree and
  penguins)

Standard model predictions for direct CPV
asymmetries in these modes O(0.001 - 0.01)
F. Bucella et al., Phys. Rev. D51, 3478 (1995) S.
Bianco et al., Riv. Nuovo Cim. 26N7, 1(2003)
16
Search for CPV in D0? K?K?, ?????

• Measure the time integrated CP asymmetries
• Experimental procedure
• fit m,?m distributions to determine raw signal
  weights
• Determine relative D0/D0 soft pion tagging
  efficiency using D0?K?? data
• greatly reduces systematic uncertainties
• correct for forward-backward asymmetries in
  e?e??cc production
• extract aCP

17
Search for CPV in D0? K?K?, ?????
KK
??
No evidence for CPV in either mode
18
Mixing in D0?K?????

• Two types of WS Decays
• Doubly Cabbibo-supressed (DCS)
• Mixing followed by Cabibbo-Favored (CF) decay
• Two ways to reach same final state ? interference!

mix

• Time dependent WS rate
• where
• and

?K??? strong phase difference between CF and
DCS decay amplitudes
19
RS and WS (mK??, ?m) fits

• Determine signal and background yields in
  subsequent Dalitz analyses.

signal and sideband regions
m
?m
m
?m
20
D0?K????? RS Dalitz fit
Time-integrated analysis to determine CF
amplitudes,
21
D0(t)?K????? WS Dalitz fit results
Through t-dependence, distinguish DCS amplitudes
from the CF amplitudes arising from mixing.
22
Mixing parameter contours and results
no-mix x best fit
Results are consistent with no mixing at 0.8,
including systematics
23
BaBar D0-D0 Mixing Summary

• Presented more evidence for D0-D0 mixing from
  BaBar experiment
• D0? K????to D0? K?K?, ?????lifetimes
• D0?K????? time-dependent Dalitz analysis
• In D0? K?K?, ?????decays,
• no evidence for CP violation
• no evidence for CP violation in mixing

No mixing excluded at ??
24
Backup Slides
25
K??backup
26
Time evolution of WS D0?K????decays

• Two types of WS Decays
• Doubly Cabbibo-supressed (DCS)
• Mixing followed by Cabibbo-Favored (CF) decay
• Two ways to reach same final state ? interference!

mix
Discriminate between DCS and Mixing decays by
their proper time evolution
(assuming CP-conservation and x1, y1)
DCS decay
Mixing
Interference between DCS and mixing
?K? strong phase difference between CF and DCS
decay amplitudes
27
D0?K? Fit Procedure

• Unbinned maximum likelihood fit performed in
  stages
• Fit m(K?) and ?m distribution
• Separate signal from background in subsequent
  decay time fits
• Fit RS decay time distribution
• Determine D0 lifetime and decay time resolution
  function R(t)
• Fit WS decay time distribution
• Use D0 lifetime and decay time resolution
  function from RS fit
• Fit WS signal to
• Compare fits with and without mixing to determine
  significance
• Fit D0 and D0 samples separately to search for CP
  violation
• In this analysis, all parameters are determined
  by fitting data, not MC

28
RS and WS mK? ,?m Distributions
Selected RS data
Selected WS data

• Separate signal from background by fitting over
  the full range shown in the plots
• 1.81 GeV/c2 lt mK? lt 1.92 GeV/c2 and 0.14 GeV/c2 lt
  ?m lt 0.16 GeV/c2
• For displaying decay time fits, integrate over a
  signal box
• 1.843 GeV/c2 lt mK? lt 1.883 GeV/c2 and 0.1445
  GeV/c2 lt ?m lt 0.1465 GeV/c2

29
RS Proper Time Fit
D0 lifetime and resolution functionfitted in RS
sample
Consistent with PDG
Systematics dominated by signal resolution
function
30
?m - m(Kp) Fit Results
31
Wrong-sign mK? , ?m fit

• The mK? , ?m fit determines the WS branching
  ratio RWS

4,030 90 WS signal events
BABAR (384 fb-1) RWS (0.353 0.008 0.004)
(PRL 98,211802 (2007)) BELLE (400 fb-1) RWS
(0.377 0.008 0.005) (PRL 96, 151801 (2006))
32
WS Fit with Mixing

• Fit results allowing mixing

RD (3.030.160.10)x10-3 x2
(-0.220.300.21)x10-3 y (9.74.43.1)x10-3
WS mixing fit projection in signal region 1.843
GeV/c2 lt m lt 1.883 GeV/c2 0.1445 GeV/c2 lt ?m lt
0.1465 GeV/c2
33
Mixing contours
Best fit Best fit, x2 0 No
mixing (0,0)

• Fit D0 and D0 samples together assuming no CP
  violation
• y, x2 contours computed bychange in log
  likelihood
• Best fit point in non-physical region
• ?? contour extends into physical region
• correlation -0.95
• Accounting for systematicerrors, no-mixing point
  is atthe 3.9? contour

1 CL 3.17 x 10-1 (1s) 4.55 x 10-2 (2s) 2.70 x
10-3 (3s) 6.33 x 10-5 (4s) 5.73 x 10-7 (5s)
RD (3.03 ? 0.16 ? 0.10) x 10-3 x2 (-0.22 ?
0.30 ? 0.21) x 10-3 y (9.7 ? 4.4 ? 3.1) x 10-3
34
Allowing for CP Violation
Fit D0 () and D0 (-) samples separately
CP violation if any () parameter differs from
corresponding (-)
x2 (-0.240.430.30)x10-3 y
(9.86.44.5)x10-3
x-2 (-0.200.410.29)x10-3 y-
(9.66.14.3)x10-3
RD(0.3030.0160.010) AD(-2.15.21.5)
No evidence for CP violation
35
BaBaR/BELLE D0!K? comparison
Results consistent within 2?
400 fb-1
PRL 96,151801
stat. only
no-mixing excluded at 2s
BELLE 2? statistical
36
Systematics, Validations

• Systematics variations in
• Functional forms of PDFs
• Fit parameters
• Event selection
• Computed using full difference with original
  value
• Results are expressed in units of the statistical
  error

Validations and cross-checks Alternate fit (RWS
in time bins) Fit RS data for mixing x2
(-0.010.01)x10-3 y (0.260.24)x10-3 Fit
generic MC for mixing x2 (-0.020.18)x10-3 y
(2.23.0)x10-3 Fit toy MCs generated with
various values of mixing Reproduces generated
values Validation of proper frequentist coverage
in contour construction Uses 100,000 MC toy
simulations
37
Lifetime ratio backup
38
D0(t)? K?K?, ?????

• Using the D??D0????D0?K?????K?K??????? decays
  from 384 fb-determine the quantitiesand
• CP violating quantities
• Lifetimes with CP violation
• If CP is conserved in mixing

where
CPV in mixing
CPV in interference of mixing and decay
39
Event Categories
40
BELLE Ratio Measurement
BELLE
PRL 98, 211803
3.2s signal
41
Mass Projections

• Mass Projections (??????????m ?????????GeV/c?)
• Signal Purities (1.8495 lt m lt 1.8795 GeV/c2)

42
Lifetime difference Cross Checks

• Performed several cross checks to ensure unbiased
  fit results
• Fits to generic and signal MC
• Fits with independent resolution functions
• Subdivided fit results into different running
  periods, D0 lab angles (cos, phi, psiangle
  between D0 decay plane and bending plane)
• use high statistics Kpi untagged data sample
• Conclusions
• No hidden differences between the modes observed
  which could bias the mixing parameters, except in
  the polar angle variation where a small
  difference in mixing parameters was observed.
  This is accounted for in the Signal systematic.

43
Direct CPV backup
44
Direct CPV Results aFB
KK
??
There is a significant FB asymmetry
45
Direct CPV cross Validations
46
D0? K?K?, ????? CPV in Decay

• Soft pion tagging efficiency determined using CF
  decay
• Yields

• no-tag D0?K???? sample determines the efficiency
  D0?K??? relative to D0?K???
• tagged K????sample determines the slow pion
  efficiency D0?K??? relative to D0?K???
• Slow pion efficiency correction is then applied
  to D0??????and D0?K?K??

47
Production asymmetries and CPV

• Forward-backward asymmetries in cc production
• Interference in e-e? ? cc as mediated by either a
  virtual photon or a virtual Z0.
• Higher-order QED box- and Bremsstrahlung-diagram
  interference effects
• Both effects are antisymmetric in cos?, the polar
  angle of the D0 CMS momentum
• Direct CPV is symmetric in this variable
• Construct symmetric (aCP) and antisymmetric (aFB)
  combinations of the yield asymmetries versus cos?

48
Direct CPV Systematic Variations
49
D0(t)?K????? WS Dalitz fit

• The WS signal contains both DCS and CF
  amplitudes.
• The CF amplitudes are determined in the RS fit
  and fixed in the WS fit
• The total time dependent WS PDF iswhere
  yields are determined from the (m,?m) fit
• The Dalitz and time distributions for mis-tag
  events are taken from the RS Dalitz model and RS
  time distributions
• The term is determined
  by a (m,?m,t) interpolation to the signal box
  from the sideband regions

50
HFAG Rmix world average
51
D0(t)?K????? Systematics/Checks

• Systematics
• Checks
• extensive Toy MC studies comparing generated and
  fitted mixing parameters. No bias seen with high
  statistics toy samples