PHENIX Single NonPhotonic Electron Spectra and v2

1
PHENIX Single Non-Photonic Electron Spectra and v2

• Nathan Grau
• Journal Club
• April 12, 2006

2
Outline

• What do single electrons tell us?
• Light quarks, heavy quarks, direct production
• Why is that interesting?
• Heavy quarks have a perturbative scale mQ
• Light vs. heavy quark differences
• How do we measure them?
• Need to remove large backgrounds
• What do we conclude?

3
Sources of electrons

• Physics sources of electrons
• Light quarks/hadrons
• f ?ee-, w ? ee-
• K?pen, etc.
• Dalitz decay p0 ?gee-, etc.
• Heavy quarks/hadrons
• J/y ? ee-, Y ? ee-
• D ?Ken, etc.
• Direct production

• Other sources of electrons
• Internal conversion of photons in material
• Note almost everything here is true about muons
  as well.

4
Two definitions

• Inclusive electrons are all of these sources
• Non-photonic electrons are those not from light
  hadron decay and from internal conversions and
  virtual direct photon production
• Primarily from heavy flavor decays and Drell-Yan
• Drell-Yan is small component down by a factor of
  100 because of aEM
• New sources of electrons in AA?
• Enhancement of low mass dileptions?
• Thermal radiation?

5
Why not just measure heavy quarks directly?

• Typically charm and bottom are measured from
  their quarkonia spectra
• PHENIX does this at least for J/y
• Open charm and bottom are also typically measured
  from displaced vertices
• ct 100 mm for D and 200 mm for B
• PHENIX cant do this yet
• Measure open charm in the hadronic decay channel
• D?Kp, D?ppp
• After three years still dont see it (but STAR
  does)
• Measuring electrons maximizes usage of statistics
• Catch more of the branching ratio

6
Interest in Heavy Flavors

• In HIC we would like a probe that is
• Strongly interacting with the medium
• Heavy quarks have color charge
• Survive the hadronization process of the plasma
• See the next couple of slides
• Heavy flavors compared to jets
• Can be calculated perturbatively aS(mQ) ltlt LQCD
• Auto-generated in the interaction in similar
  processes.

7
(No Transcript)
8
Initial Expectations for Heavy Quark Energy Loss

• Heavy quarks from hard scattering traverse the
  medium and lose energy
• Survives QGP hadronization.
• Dead cone effect
• Can someone please explain the dead cone effect
  to me. I really couldnt find a clear explanation
  in the literature.

9
Heavy-to-Light Comparison

• Ratio of heavy quark RAA to light quark RAA.
• 20 higher RAA predicted for heavy quarks at 5
  GeV.

10
Anisotropy of Heavy Quarks (I)

• Flow results from 2 sources
• Pressure gradients in the overlap region of the
  nuclei
• Low pT, hydrodynamics
• Path length dependent energy loss
• High pT
• Question Do heavy quarks couple as strongly to
  the medium as light quarks?
• We should measure it!

11
Anisotropy of Heavy Quarks (II)

• Another question Less energy loss for heavy
  quarks, but does that necessarily reduce the
  anisotropy?

if
(Good to lt10 from Dokshitzer and Kharzeev)
!
We should measure it!
12
Electrons in PHENIX

• Identification by
• Charged track in DC/PC
• Momentum, charge, position
• Associated hit in RICH
• Electrons only fire up to 3.5 GeV
• Muons and pions then fire
• Muons are rare
• Associated EM cluster in
  calorimeter

13
Final Spectra

• Inclusive Electrons
• Need to determine the photonic contribution

10-20
60-80
0-10
14
Cocktail Method

• Parameterize the measured p0 spectrum as a
  function of centrality
• Assume that all other light mesons mT scale,
  confirmed by h spectrum
• Conversion photon spectrum determined from PISA
  simulation
• Direct photons parameterized from NLO fit
• Kaon spectrum parameterized from data
• Run EXODUS which randomly picks from the given
  distribution and decays if necessary

15
Non-Photonic Spectrum (I)

• Comparison of the minimum bias cocktail and
  converter spectra
• Note that the cocktail is much more precise
• Excellent agreement

16
Non-Photonic Spectrum (II)

• Published spectrum
• The line indicates a fit to the pp spectra
• Note no centrality above 60?
• Suppression observed at high-pT in all centrality

17
RAA

• A dramatic suppression is seen at high pT.
• Comparable to suppression of p0
• Is this misleading, shouldnt we shift the
  electron spectrum to the left in order to compare
  heavy and light quark suppression?

18
What about gt60 Centrality?

• We have spectra that compares well to the
  converter method
• But RAA looks terrible! Was PHENIX just sneaky?
• The paper claims More peripheral collisions have
  insufficient electron statistics to reach pT 5
  GeV/c.
• The p0 spectra do not reach to the same pT in all
  centrality bins

19
What can we say about heavy quark Eloss?

• Comparison of data to theory
• 1a-1c BDMPS (next weeks talk) calculation of
  charm only for
• a no medium, only Cronin
• b
• c
• 2a-2b GLV calculation with charm and bottom,
  bottom pulls up the RAA because of dead cone.
• a
• b
• Very extreme range of densities and opacities!

20
Gluon Contribution to Spectrum?

• A hard gluon from a hard process could split
  (fragment?) to Q-Qbar and create two hard mesons
• If the formation time for such a splitting is
  longer than say the lifetime of the plasma, the
  gluon would lose the energy and this would be
  reflected in the resulting charm hadrons.
• Because the gluon is fast, gamma is large and
  there will be a time dilation in its decay
• No calculation of this I have found
• pp spectrum errors leave room for this
  production
• Is it implemented in pythia?

21
Summary on Spectra

• This is an open topic at the moment
• No calculation can reproduce the observed spectra
  based on both charm and bottom contributions
• On the face it seems that the charm and bottom
  loose as much energy as light quarks and gluons
• What about the coupling to the medium
• i.e. do heavy quarks flow?

22
Extracting Inclusive Electron v2

• Measure the azimuthal angle wrt Y for both
  candidates and background
• Subtract background from total to get signal and
  fit

23
Inclusive Electron v2
24
Obtaining Non-photonic electron v2
25
Obtaining Photonic v2

• Just use a cocktail similar to the singles
  spectra
• EXODUS modified to produce a random RP and f
  distribution of the generated particles.
• Study electron v2 given input v2 and spectra

p/- and p0 as input
26
Cocktail Sources

• Cocktail sources (in order of importance)
• p0 Dalitz(previous slide) and conversion (run
  through PISA)
• Not suprisingly similar v2.
• h Dalitz decay, assume v2 kaon v2, spectrum mT
  scales
• K decay, use measured v2 and spectra of K and
  STARs Ks0
• Nothing else without further assuming about
  heavier particle v2 (r, w, f, J/y, etc.)

27
Cocktail Results
e v2 from p0 Dalitz
e v2 from K
e v2 from h Dalitz

• The resulting v2 for the different components
• Relative contribution to the total is also known
  from the cocktail

28
Non-photonic Electron v2 Results

• The paper claims a 90 confidence level that
  non-photonic electron v2 !0
• Why does that seem too low?
• All points except on are gt0 at 1.5s?

29
But Im Missing the Point

• Non-zero non-photonic electron v2!
• And it is consistent with charm flow!
• Is recombination believable?

30
The Summary

• PHENIX has measured single non-photonic electron
  spectra and v2 and found that
• High-pT electrons are suppressed wrt binary
  scaled pp collisions to the level of p0
• There is a non-zero v2.
• In RUN-4 these results have been extended to
• Better the stats
• Centrality binning
• Other things that are necessary
• Extending the pT reach of the electron spectra
• Only reason stopping them at 5 GeV/c was pion
  turnon in RICH
• Need to do this in pp as well
• Measure charmed hadrons and measure there v2
• J/y v2 ongoing analysis (but Tatia will let us
  know if we can distriminate between partonic flow
  recombination, etc. with the J/y)

31
Backup Slides
32
Electron ID details

• Exactly the same cuts for both analyses
• High quality tracks
• Excellent p resolution, S/B?
• 2s matching to EMCal
• Cluster association, multiple scattering
• n0gt3, n3gt1 (number of pmts with good timing
  fired)
• ?
• -2s lt E/p lt 3s

Overall S/B for 0.5-5 GeV/c is very good 10/1
33
Electron ID Background

• Background is determined by the swap variables
• z ? -z of hits reassociate RICH and EMCal hits
• Good for determining random association
• Why is the background not the same shape as the
  tails?
• Effect on the single particle spectrum and for
  the flow analysis
• Just subtract off the background spectrum and
  dn/df shape from the measured spectrum and dn/df

34
Acceptance and Efficiency

• Acceptance
• Amount of dead area within the fiducial region
• Study by PISA with detector response tuned to
  data
• Efficiency
• In active area probability for finding the
  electrons given the cuts in the analysis
• Study by embedding single particles into real
  events

1/(AccEff)
pT
35
Measuring the RP

• wi are weights, could be n for number of
  particles in the ith bin, pT for pT flow
  correlations