B0s h h decays at CDF

1
B0(s) ? hh- decays at CDF
HEP2005 International Europhysics Conference on
High Energy Physics July 21st 27th, 2005
Lisboa, Portugal Diego Tonelli tonel_at_fnal.gov
Istituto Nazionale di Fisica Nucleare, Pisa for
the CDF Collaboration
2
Motivation
Joint study of B0 and B0s 2-body decays into
charged kaons and pions (KK, ?? and K?) a
privileged insight into flavor physics and a
useful tool in searching for New Physics.

CDF has simultaneous access to both B0
/B0s ? hh'- decays ? flavor physics program
complementary to Y(4S). R. Fleischer
PLB459306-320, 1999 - constrain hadronic
unknowns with SU(3) symmetry. Use approximated s
? d quark symmetry (i.e. measure jointly B0 and
B0s) to extract phase ?.
WEAK amplitude theoretically clean
HADRONIC/ EW theoretically uncertain
3
Motivation (contd)
Fleischers method needs time-dependent
asymmetries in b-flavor tagged samples, size of
SU(3) breaking, sin(2?) and ?ms
CDF ultimate long term goal.
Currently accessible BR can constrain theory too
compare CDF measurements with allowed regions
in spaces of B0? ??- and B0s? KK- observables
(Y(4S) and theory) a probe for both ? and NP
Fleischer and Matias PRD66 054009,2002 -
London and Matias PRD70031502, 2004.
many other interesting measurement, e.g. ??s/?s
in B0s? KK- see talk by M. Donega - Friday
17.08 Room 5.
4
Experimental challenge 1 write signal events
to tape

• S/N at production 10-9
• high track multiplicity per event
• generic final states (2 tracks, K
• and/or ?) in huge QCD BCKG.
• Crucial role of trigger
• Reconstruct silicon tracks online. Trigger on
  displaced secondary vertices with 30 ?m impact
  parameter (IP) resolution.

• large ?(B) ? large B decay length and IP of
  tracks
• pp ? bb X ? B from primary vertex, small IP of
  B.

5
Analysis overview
SIGNAL RECONSTRUCTION unbiased optimization of
the selection requirements (use MC and data)
dE/dx CALIBRATIONS Accurate calibrations of dE/dx
with D decays
FIT OF COMPOSITION Unbinned ML fit kinematics
and dE/dx to distinguish each B0(s) ? hh'-
mode
SIGNAL SIMULATIONRealistic MC sample
BR RATIOS and ASYMMETRIES Correct
raw fit results for trigger / selection
efficiencies (use MC and data)
CONTROL SAMPLES B0s and B0 exclusive decays
in J/?X and D?
6
Experimental challenge 2 signal extraction
Unbiased optimization of selection cuts maximize
S/?(S B) with signal from MC and background
from data sidebands.
isolation fraction of pT carried by the B
candidate after fragmentation. Rejects 15 of
signal and 400 of background.
7
Experimental challenge 3 peak composition

• Excellent CDF mass resolution
• 5 silicon layers drift chamber measure pT in
  1.4 T solenoidal field with 132 cm lever arm
• ? resolution (0.7 ? 0.1 pT)
• Still, the four (expected) major modes overlap
  into an unresolved mass peak.
• Extract composition statistically with an
  unbinned 5-dimensional ML fit. Combine
  information from
• kinematics (mass and p)
  
• particle ID (dE/dx).

B0 modes
B0s modes
simulated signals
8
Peak composition handle 1 kinematics
??-mass vs signed momentum imbalance (1-
pmin/pmax)qmin discriminates among modes (and
flavors in K? modes).
9
Peak composition handle 2 specific
ionization in the drift chamber

• Tracks have hits in drift chamber
• 96 layers, ? 1.0
• 44 cm lt r lt 132 cm, 30k channels
• s(hit) 180 µm
• dE/dx encoded in hit pulse-width.

TIME OF FLIGHT
Accurate, time dependent dE/dx calibration using
95 pure K / ? samples from 70k decays D ?
D0 ? ? K-? ? c.c. Strong D decay tags
the D0 flavor.
10
Peak composition handle 2 dE/dx (contd)
Angle, gas pressure, time and hit-multiplicity
dependences corrected.
1.4? K/? separation at p gt 2 GeV/c (? 60 of
perfect separation) Residual gain fluctuations
cause correlated dE/dx shifts measured and
included in the fit of composition.
11
Raw fit results
What we measure
900 evts/180 pb-1 in initial CDF data, taken
with still non optimized detector/trigger.
Now much better 2700 / 360 pb-1
12
Efficiency/acceptance corrections
Correct for relative acceptance, trigger and
selection effic. 5 -10
Kinematics and nuclear interaction efficiencies
(Monte Carlo)
Trigger-bias correct for dE/dx-dependent trigger
efficiency (D data)
Isolation efficiency, measured from exclusive
decays (B0 and B0s data)
13
Dominant systematic uncertainties
(stat. 7.8) syst. 1.2
(stat. 29) syst. 13
(stat. 17) syst. 15
(stat. 24) syst. 14

• dE/dx track-to-track correlations (partially
  reduces with statistics)
• B-meson masses input to the fit (reduces with
  statistics)
• Relative isolation efficiency between B0s and B0
  (reduces with statistics)

• Effect of final state radiation
• Trigger bias on efficiency
• Charge-dependence of dE/dx

• Background shape
• Charge asymmetries in background
• Others.

14
Final results B0s sector
B0s ? KK- decay established. BR ratio may favor
large SU(3) breaking as predicted from sum rules
(Khodjamirian et al. PRD68114007, 2003).
Allows first comparisons with Y(4S) and theory
expectations, test of NP.
No evidence for B0s ? K-?, set a limit a factor
40 better than PDG04.
Great improvement on annihilation mode B0s ? ?
-?. A factor gt100 below PDG04 (time-evolutions
of B0s ? ?-? and B0s ? K-K assumed the same).
15
Final results B0 sector
ACP compatible with B-factories, systematic
uncertainty comparable as well, Babar statistic
uncertainty just 30 better with same sample
size. With
currently available data (3x statistics), we
expect lt 4.5 statistical uncertainty to be
compared with current world best 2.2 (Belle).
Limit on pure annihilation/exchange mode B0 ?
KK-. A factor 2 above B-factories, expect much
better performance on current sample.
Consistent with B-factories. Valuable cross-check
for other measurements.
16
Concluding remarks

• CDF has unique joint access to B0 / B0s ? hh'-
  modes rich physics program complementary to
  B-factories. Results shown ready for PRL
  submission in first B0(s) ? hh'- paper from an
  hadronic collider
• B0s? KK- decay established and BR measured
  (first B0s ? PP observed)
• x100 improvement with respect to PDG04 on the
  upper limit on
• BR(B0s ? ??-) and x40 improvement on the
  limit on BR(B0s ? K-?)
• measurement of ACP (B0 ? K?-) with small
  systematics.
• Just the beginning now 900 pb-1 on tape, of
  which 360 ready for analysis with 2x increase in
  yield/pb-1 (optimized efficiencies), improved
  mass resolution (better tracking alignment and
  reconstruction). By end of 2005 we expect
• Observe B0s ? K-? and reduce to 10
  statistical error on BR(B0s ? KK- )
• ACP (B0 ? K?-) much closer to Y(4S) with 4
  statistical uncertainty
• World best limits on BR(B0s ? ??-) and on
  BR(B0 ? KK-).

17
ADDITIONAL MATERIAL
18
The Tevatron pp collider
Superconducting proton-synchrotron 36 p ? 36 p
bunches
collision every 396 ns at vs
1.96 TeV Luminosity. record peak is
1.3 ? 1032 cm-2 s-1
10 pb-1 /
week recorded on tape interactions /
bunch-crossing.. lt N gtpoisson 6 (at 2 ?
1032 cm-2s-1) Luminous region size.. 30 cm
(beam axis) ? 30 ?m (transverse)

need long Si-vertex small wrt c?(B)
450 ?m
2005, regularly exceeds 1032 cm-2s-1
19
Delivered Luminosity
900 pb-1 on tape
1 fb-1 milestone!
data for physics
Feb 2002
April 2001
Jul 2002
first data for analyses
detector commiss.
Stable data taking efficiency 85 The results
shown here from an analysis that uses 180 pb-1
20
Heavy Flavor physics at the Tevatron
The Good bb production x-section O(105) larger
than ee- at ?(4S) /Z0. Incoherent strong
production of all b-hadrons B?, B0, Bs, Bc, ?b,
?b . The Bad Total inelastic x-section 103 ? ?
(bb). BRs for interesting
processes O(10-6). and
The Ugly Messy environments with large
combinatorics.
Need highly selective trigger
21
CDF Detector Upgrades
22
Triggering bs (and cs)
conventional
new approach
Di-lepton B ? charmonium Rare B ? ?? Two muons
with pTgt 1.5 GeV ?lt 1 pTgt 2.5-4.5 GeV
?lt2
electron or ? and displaced track Semileptonic
decays Electron (?) with pTgt 4 (1.5) GeV
?lt 1 and one track with pT gt 2.0 GeV IP gt
120 ?m
Two displaced tracks n-body hadronic B Two tracks
with pT gt 2.0 GeV ?pT gt 5.5 GeV IP gt 120
(100) ?m
Displaced track trigger at Level 2 a revolution
in hadronic environment ! Makes it accessible
rare hadronic decays with high S/B.
23
Systematics - detailed
24
Raw physics results
25
Raw physics results
26
Corrected results
27
Correlation matrix
Legenda
28
p.d.f. projection momentum imbalance
29
p.d.f. projection p(1) p(2)
30
p.d.f. projection dE/dx
31
p.d.f. projection dE/dx
32
Momentum p.d.f.

Binned ML fit
Monte Carlo
?p
?
?p
?
signal
background
33
dE/dx model

34
Silicon Vertex Trigger

35
Silicon Vertex Trigger (contd)

(35 ? 33) mm SVT ? beam ? s 48mm
36
B0s ?KK- lifetime

B0 lifetime from PDG04. B0s ?KK- lifetime
depends on - ??s/?s width difference between
long and short
eigenstates -the relative composition in short
and long-lived components Assume Standard Model
?s ?d and ??s / ?s -0.12 ? 0.06 B0s?KK-
is expected almost 100 short-eigenstate
Fleischer
and MatiasPRD66-054009,2002) Therefore the
lifetime of B0s ?KK- is determined as
37
Physics Motivations (backup)
The combination of Bd and Bs decays provides a
promising way to extract CP-related physical
parameters avoiding the penguin pollution. (R.
Fleischer PLB459 (1999) 306) Assume U-spin
symmetry (d ? s), the ACP are function of the
CKM angles ? and ? and of the amplitude ratio P/T
( dei?) ? 4 equation with 4 unknowns (?, ?, d,
?). A
combined fit of the 4 CP asymmetries measures ?,
? and P/T ratio. Above strategy need
time-dependent analysis with tagged samples long
term goal