Andr

1
Heavy-flavor correlation measurements via
electron azimuthal correlations with open charm
mesons
André Mischke for the STAR Collaboration

• Outline
• Motivation
• Correlation technique
• Data analysis
• Results
• Data-model comparison
• Summary and conclusions

Strangeness in Quark Matter 2429 June 2007,
Levoca (Slovakia)
2
Heavy quark energy loss

• Due to their large mass heavy quarks are
  primarily produced by gluon fusion ?
  production rates can be calculated in
  pQCD ? sensitivity to initial state gluon
  distribution M. Gyulassy and Z. Lin,
  PRC 51, 2177 (1995)
• Heavy quarks lose less energy due to suppression
  of small angle gluon radiation (dead-cone effect)
  Dokshitzer and Kharzeev, PLB 519, 199 (2001)
• Amount of collisional and radiative energy
  losses seems to be similarM.G. Mustafa, PRC72,
  014905 A.K. Dutt-Mazumder et al., PRD71, 094016
  (2005)

parton
hot and dense medium
M. Djordjevic, PRL 94 (2004)
3
Heavy flavour measurement

• Hadronic decay channels
• D0 ? Kp B.R. 3.83
• D? ? Kpp B.R. 9.51
• D ? D0p B.R. 65
• Difficulty large combinatoric background,
  especially in high multiplicity environments
• Event-mixing and/or vertex tracker needed to
  obtain a signal

QM 2005, Nucl. Phys. A774 (2006) 701, publication
in preparation

• Semi-leptonic channels (inclusive modes)
• c ? e X B.R. 9.6
• D0 ? e X B.R. 6.87
• D? ? e? X B.R. 17.2
• b ? e- X B.R. 10.9
• B? ? e? X B.R. 10.2
• Single (non-photonic) electrons sensitive to
  charm and beauty

4
Non-photonic electron spectra in AuAu

• Non-photonic electrons exhibit a similar yield
  suppression at high-pT in central AuAu as light
  hadrons
• Models implying D and B energy loss are
  inconclusive yet
• ? Disentangle D and B contribution to
  non-photonic electron spectrum experimentally
• ? At which pT does B contribution start to
  dominate ?
• Approach Non-photonic electron - D0 meson
  azimuthal correlations

Phys. Rev. Lett. 98, 192301 (2007)
Large suppression not expected due to dead-cone
effect
5
Correlation technique
6
Electron tagged correlations

• Experimental approach
• non-photonic electrons from semi-leptonic D/B
  decays are used to trigger on c-cbar or b-bbar
  pairs
• associate D0 mesons are reconstructed via their
  hadronic decay channel (probe)
• Underlying production mechanism can be
  identified using second c/b particle

trigger side
3.83 54 10
charm production
???0
????
probe side
flavor creation gluon splitting/fragmentation
7
Electron tagged correlationsB production
8
PYTHIA simulations 3ltpTtrglt7 GeV/c

• Different decay channels for D and B
• Charge-sign requirement on (e,K) pairs gives an
  additional constraint on production process
• - Like-sign (e,K) means charge(electron)
  charge(Kaon)
• Separation of D and B contribution to
  non-photonic electrons

9
PYTHIA simulations Electron triggers with
8ltpTtrglt20 GeV/c

• Near-side
• B decays (dominant)
• Away-side
• charm meson pair production (dominant)
• small B contribution

• Away-side
• B decays (dominant)
• small charm contribution

10
Data analysis
11
STAR experiment
Solenoidal Tracker at RHIC Large acceptance
magnetic spectrometer

• Energy measurement
• - Barrel EMC
• ? lt 1
• Pb/scintillator (21 X0)
• dE/E 16/?E
• Shower maximum detector

• PID and tracking
• - TPC
• ? lt 1.5
• ?p/p 2-4
• ?dE/dx/dEdx 8
• - Magnet
• 0.5 Tesla

12
Dataset and triggering

• 2006 pp at ?sNN 200 GeV
• ?Ldt 9 pb-1
• 1.2M events after trigger and vertex cut
• BEMC fully installed
• 97 operational
• Level-0 trigger used to enhance particle yield
  at high-pT
• BEMC tower energy threshold 5.4 GeV

?
Run 04
?
BEMC acceptance
13
Electron selection Procedure

• TPC tracks are extrapolated onto BEMC surface
• Select tracks with well developed shower in SMD
• p measurement in TPC
• E measurement in BEMC
• Quality cuts
• p/Etower ratio
• specific energy loss dE/dx

• Shower Maximum Detector (SMD)
• wire proportional counter with strip read-out
• located after 5 X0
• ????? 0.007 x 0.007

14
Electron selection Quality cuts

• Ratio of momentum and tower-energy should be one
  for electrons
• ? cut 0 lt p/Etower lt 2
• 3.5 lt dE/dx lt 5.0 keV/cm

p/Etower
after p/E and SMD cuts
d
electron candidates
p
K
dE/dx cut
e
?
hadrons (essentially pions)
15
Electron purity and hadron suppression

• Purity 100 for pT lt 7 GeV/c
• Hadron suppression factor 102 - 105
• ? Clean electron sample

16
Photonic background

• Measured electron candidates have a photonic and
  non-photonic contribution
• Photonic contribution from gamma conversions and
  (?0, ?) Dalitz decays
• Procedure
• electron candidates are combined with TPC tracks
  which passed loose dE/dx cuts around the electron
  band
• invariant mass is calculated at dca of these
  pairs
• Electrons having a low invariant mass (minv lt
  150 MeV/c2) are excluded
• Correction for background rejection efficiency
  not implemented yet
• ? Non-photonic electron excess at high-pT

17
Topological reconstruction of open charm mesons

• Non-photonic electron trigger (sub-leading
  particle) present in event
• No measurement of decay vertex
• dE/dx cut (3?) around Kaon band
• Charge sign requirement sign(e) sign(K)
• (K?) invariant mass

18
Results
19
(K?) invariant mass distribution
w/o electron trigger
pp 200 GeV STAR preliminary
dn/dm
combinatorial background is evaluated using
like-sign pairs
20
D0 mesons in pp collisions

• S/B 1/7 ? factor 100 better than in dAu
  w/o trigger
• Signal significance 3.7
• Peak content 200

21
D0 yield versus ??(e,hadron pair)

• Calculate ?? between non-photonic electron
  trigger (pTtrg gt 3 GeV/c) and hadron-pair pT
• Extract D0 yield from invariant mass
  distribution for different ?? bins

22
Non-photonic electron D0 meson azimuthal
correlation

• Near- and away-side correlation peak observed,
  yields are about the same
• First heavy flavour correlation measurement at
  RHIC

23
Data PYTHIA comparison
Procedure (1) Away-side yield for unlike-sign
(e,K) pairs is essentially from B decays ? Scale
PYTHIA distribution to fit measured away-side
yield (2) Compare near-side yield from scaled
PYTHIA??? distribution for like-sign (e,K) pairs
to data ? Difference is expected to come from
gluon splitting (3) Compare away-side yield from
scaled PYTHIA??? distribution for like-sign (e,K)
pairs to data ? Disentangle charm and beauty
contribution
24
Comparison Unlike-sign (e,K) pairs

• Away-side yield
• yielddata 0.012 ? 0.0061
• yieldPYTHIA 0.0042
• ? scaling factor 2.86

25
MC_at_NLO for charm production

• Near-side yield for like-sign (e,K) pairs
• yielddata 0.011?? 0.0046
• yieldPYTHIA (scaled) 0.0096
• Difference is attributed to gluon splitting

• NLO QCD computations plus Herwig
• Remarkable agreement of the away-side peak shape
  between PYTHIA and MC_at_NLO
• Indications for small gluon-splitting
  contribution (3?10-4)
• More statistics needed for final conclusions

private version from S. Frixione (CERN)
26
Comparison Like-sign (e,K) pairs
pp at ?sNN 200 GeV
essentially from B decays only
70 from charm 30 from beauty
27
Summary and conclusions

• Non-photonic electron trigger helps to suppress
  the combinatorial background significantly
• - S/B ratio 1/7 and signal significance 3.7
• First heavy flavour correlation measurement in
  pp at RHIC
• Non photonic electron - D0 meson azimuthal
  correlations allow to disentangle charm and
  beauty contributions to the non-photonic electron
  spectrum
• - near-side essentially from B decays
• - away-side high charm contribution
• Data shows hints for prompt charm meson pair
  production
• Comparison between PYTHIA and MC_at_NLO
• good agreement for LO processes (flavor
  creation)
• small gluon-splitting contribution for 3 lt pT lt
  7 GeV/c

28
Backup
29
D0D- cross section measurement at the Tevatron
D0 or D
??
D-
B. Reisert et al., Beauty 2006, to be published
in Nucl. Phys. B (Proc. Suppl.)

• Within errors near- and away-side yields are the
  same ? gluon splitting as important as flavour
  creation
• Near-side yield PYTHIA underestimates gluon
  splitting
• Note Results are obtained at 10 times higher
  collision energy than at RHIC

30
MC_at_NLO simulations
31
Open charm in dAu
PRL 94, 062301 (2005)
Conventional reconstruction technique
Combination of all positive and negative tracks
after quality and dE/dx cuts
32
Invariant mass spectra
3.1k non-photonic electron trigger, 105
D0 3.3k non-photonic positron trigger,
120?D0
33
minv(K?) for photonic e- trigger
pp at ?sNN 200 GeV

• Di-jet events produce many pions, which can make
  a, e.g., Dalitz decay
• What is the correlation contribution from these
  photonic electrons?

?0 ? ?ee-
?c c
No D0 signal observed
D0