Bs Mixing at CDF

1
Bs Mixing at CDF

• Frontiers in Contemporary Physics
• Nashville, May 24 2005
• Gavril Giurgiu for CDF Collaboration
• Carnegie Mellon University

2
Introduction

• - Motivation Constrain the CKM matrix elements

- Within the Standard Model, Bd/s mixing provides
information on Vtd/s
- Although ?md is well measured (0.502 ?
0.007ps-1) determination of Vtd is
affected by 15 error
3
Unitarity Triangle

• - Unitarity of CKM matrix

- Knowledge of both Bd and Bs mixing
frequencies would provide better
constraints on one side of unitarity triangle
x 1.15 ? 0.05 from Lattice QCD
4
Current Bs Status
- Bs mixing not observed yet - Bs oscillates
more than 30 times faster than Bd ? experimental
challenge - At 95 CL lower limit ?ms gt 14.4
ps-1 with sensitivity of 17.8 ps-1
5
B Mixing Phenomenology
- Neutral B system
- Mass eigenstates
- Oscillation frequency of Bq mesons given by
?mq MH-ML
- Lifetime difference ????H??L
- Neglecting ??, mixing probability after time
t is give by
6
CDF Detector
7
CDF Detector Schematic View
Time of Flight for K/p separation placed before
1.4 Tesla Solenoid
Muon Detectors h lt 1.0
Central Tracker (COT) h lt 1.0 dE/dx for PID
Plug Calorimeter 1.3 lt h lt 3.5
Silicon Detector h lt 2.0
Electromagnetic and Hadronic calorimeters
8
Silicon Vertex Trigger (SVT)

• - Silicon Vertex Trigger implemented at Level
  2
• Uses silicon detector information and beamline
  position to
• determine the track impact
  parameter
• Good impact parameter resolution 47 mm
• 33 mm beam size ? 30 mm intrinsic
  SVT resolution
• Trigger on displaced track

9
Mixing Analysis Overview
- Mixing analysis ingredients - Signal
reconstruction - Decay time - B flavor at
decay - B flavor at production inferred
through flavor tagging - lepton
tags - jet charge tags -
Statistical significance of Dms measurement
Decay time resolution
Tagging
Signal Reconstruction
10
SVT Triggers for B Physics
Semileptonic (partially reconstructed) decays
Bs ? lepton Ds X - large number of
events - decay time resolution degraded
due to missing neutrino - triggered by
4 GeV lepton and displaced track with
impact parameter d0gt120 ?m and
d0lt1 mm
Hadronic (fully reconstructed) decays
Bs ? ? Ds - smaller number of events
- good decay time resolution - triggered by
two displaced tracks with impact
parameter d0gt 120 ?m and d0lt1 mm
d0
d0
d0
11
Semileptonic Bs Signals
- Missing neutrino ? cannot see Bs mass peak -
Use Ds mass peak and (lepton, Ds) charge
correlation lD- - right sign combination
l-D- - wrong sign combination - Decay
modes Ds ? ? ? ( 4355 ? 94 ) Ds
? KK ( 1750 ? 83 ) Ds ? 3 ? ( 1573 ? 88
) - Total of 7000 Bs candidates but 20
come from Physics backgrounds B0/ ? Ds
D Bs ? Ds t (D / t / D(s) ? lepton
X) Bs ? Ds D(s)
12
Semileptonic Bs Signals (cont)
Ds ? KK ( 1750 ? 83 )
Ds ? 3 ? ( 1573 ? 88 )
13
Hadronic Bs Signals
- All final state particles reconstructed ?
observe Bs mass peak
- Decay modes Ds ? ? ? ( 526 ? 33 )
Ds ? KK ( 254 ? 21 ) Ds ? 3 ? ( 116 ? 18
) - Total of 900 Bs candidates - Satellite
peak Bs ? Ds ? (Ds ? Ds X)
14
Hadronic Bs Signals (cont)
Ds ? KK ( 254 ? 21 )
Ds ? 3 ? ( 116 ? 18 )
15
Decay Time
- Decay time - In semileptonic modes
missing neutrino is statistically corrected
by
- Hadronic decays do not need correction -
Decay time resolution
Semileptonic
Hadronic
16
Decay Time Bias

• - Because
• (1) In both hadronic and semileptonic
  decays the triggers require
• displaced tracks
• (2) Bs events are selected based on decay
  distance cuts
• the Bs decay time distribution is biased
• - Efficiency as function of decay time obtained
  from Monte Carlo

17
Flavor Tagging
- For Bs mixing analysis CDF used 5 opposite
side flavor taggers - Tag inferred from
opposite side B in event - muon and
electron tag (semileptonic decay of opposite B)
- three jet charge tag types - displaced
vertex - displaced tracks - high pT - Tagging
power given by ?D2 where ? is the tagging
efficiency D 1 2 Pmistag is the tagging
dilution Pmistag mistag probability -
Large dilution (D) means high tagging power
Trigger B meson
18
Flavor Tagging (cont)
- Dependence of dilution on different quantities
enhances tagging power - Dilution of lepton
taggers is calculated as function of -
lepton likelihood (probability that lepton is
real) - (transverse momentum of lepton
w.r.t jet axis) - Dilution of jet charge tagger
is calculated as function of - the jet charge

di - displacement of track i w.r.t the primary
vertex - Total ?D2 ? 1.6 calculated on an
inclusive leptonSVT sample
Jet charge tag
Electron tag
Muon tag
19
Tagger Calibration and Measurement of ?md
- Perform measurement of ?md - Since we
observe B0 oscillations, we can also measure tag
dilutions - Analyze hadronic and semileptonic
decays of B0 and B B0 g D p- B g D0 p B0
g J/y K0 B g J/y K B0/ g D- l X B0/
gD- l X B/0 g D0 l X - Event by event
predicted dilution (D) - Fit the
dilution calibration factor (S) for each of 5 tag
types - Dilution calibration factors are used
for the Bs mixing analysis
Event by event dilution
Dilution Calibration Factor
20
?md Results
- Hadronic Dmd (0.5030.0630.015) ps-1
- Dilution calibration factors S(muon)
0.830.100.03 S(electron)
0.790.140.04 S(vertex)
0.780.190.05 S(track)
0.760.210.03 S(high pT)
1.350.260.02 - Total eD2 1.1
- Semileptonic Dmd (0.4980.0280.015)
ps-1 - Dilution calibration factors
S(muon) 0.930.040.03 S(electron)
0.980.060.03 S(vertex)
0.970.060.04 S(track)
0.900.080.05 S(high pT)
1.080.090.09 - Total eD2 1.4
World average ?md 0.502 ?0.007 ps-1
Muon Tags
21
Lifetime Measurement
- As a cross check of analysis framework measure
Bs lifetime - Lifetime fit projections in both
hadronic and semileptonic modes
- Semileptonic c?(Bs) 443 ? 10 (stat) ? xxx
(syst) ?m - Hadronic c?(Bs) 479 ? 29 (stat) ?
5 (syst) ?m - Good agreement with PDG 2004
c?(Bs) 438 ? 17 ?m
22
Amplitude Scan Method
- Introduce Fourier coefficient A (amplitude) -
Fix ?m at different test values and fit for A
(Moser et.al., NIMA 384 491)
A ? 1 for true value of ?m A ? 0 away from
true value - Test amplitude method on B0
oscillations by scanning for Dmd in
hadronic modes - points A ? 1? -
yellow band A ? 1.645? - dotted line
1.645? - Yellow band bellow 1 ? exclusion at
95 CL
zoom in
23
Bs Analysis
- Performed blind analysis by randomizing the
tag decision tag
tag ? (-1)event number - Evaluate sensitivity
and systematic uncertainties from blind
analysis - Systematic errors evaluated using
pseudo-experiments - include all variables and
distributions determined from data - fit the toy
sample with different Likelihood configurations
- use variations in Amplitude (?A) and
statistical error (??A) to derive the
systematic error
24
Semileptonic Amplitude Scan
- Measurement is statistics dominated - Main
systematic uncertainties from prompt background
and from Physics background
Sensitivity 7.4 ps-1 Limit ?ms gt 7.7 ps-1
25
Hadronic Amplitude Scan
- Measurement is statistics dominated - Main
systematic errors come from tagger calibration
Sensitivity 0.4 ps-1 Limit ?ms
gt 0.0 ps-1
26
Combined CDF result on ?ms
- After combining semileptonic and hadronic
modes Sensitivity 8.4 ps-1
Limit ?ms gt 7.9 ps-1 - With full Bs momentum
reconstruction, hadronic mode will dominate
the measurement at high ?ms
27
Conclusions
?ms limits from CDF Semileptonic 7.4
ps-1 Hadronic 0.0 ps-1 (will become
important at high ?ms with more
statistics) Combined limit 7.9 ps-1,
sensitivity 8.4 ps-1 Results will substantially
improve soon - Same side Kaon tagger -
Improve decay time resolution in hadronic
modes - Add more data Updated analyses expected
soon