Gustavo Conesa Balbastre

1
High pT direct photon measurement and
correlations with hadrons and jets in ALICE

• Gustavo Conesa Balbastre

2
Direct photons as probes for QGP

• Direct EM probes convey unperturbed information
  and their production probe the medium
• Tag medium-modified jets
• Prompt photons from 2-gt2 hard process (Eg gt 10
  GeV)
• Medium modified production
• Fragmentation photons (Eg lt Ejet)
• Medium produced photon
• Bremsstrahlung and jet conversion (Eg lt Ejet)
• Thermal photons (Eg lt 10 GeV)
• Challenge
• Disentangle the different sources.
• Neutral mesons decay.
• But decay photons provide a first choice probe of
  medium effects
• Identify real photons (EM calorimetry, trigger)
  and ee- from virtual and converted photons
  (tracking and PID, trigger)

3
Direct photon sources
prompt
LO
Isolated photons
4
Estimates with a thermal model
Turbide, Gale, Jeon, and Moore PRC (2004)

• Jet bremsstrahlung/fragmentation correlated with
  hadrons
• Jet-plasma thermal, uncorrelated

Gale QM2008
5
How can we distinguish different direct photon
sources?

• Prompt RAA 1, v20
• Fragmentation RAAlt1, v2gt0
• Thermal, Bremsstrahlung, Jet Conversion RAAgt1,
  v2lt0

6
RHIC News PHENIX RAA
Gale QM2008

• Run 2 No ? suppression (PRL 94, 232301
  (2005)).
• Run 4 (QM06) High pT ? suppression
• Isospin (PDF) effect
• Fragmentation photon suppresion?
• Something else?

7
RHIC News PHENIX v2
Gale QM2008
v2 small! Consistent with zero (within errors)
8
Why photon-tagged jets?

• Medium effects redistribute (?qL) the parton
  energy, Ejet, inside the hadron jet
  (multiplicity, jT).

A
B

• Redistribution can be best measured with the
  Fragmentation Function... If we know Ejet.
• HI environment hinders precise reconstruction of
  Ejet.

If we measure Eg Ejet
9
RHIC news PHENIXg-h correlations
Nguyen QM2008
Use Near Side peak to determine direct ?
associated with h, I.e. fragmentation
photons. Statistical subtraction of decay photons.
10
RHIC news PHENIXh-g correlations I
Nguyen QM2008

• Idea By triggering on a hadron and looking for
  near-side direct photon partners one can measure
  the fragmentation photon yield directly
• Measure hadron - inclusive g and hadron - decay g
  correlations
• Decay corrs are made by tagging p0 and h by
  invariant mass
• Must know tagging efficiency and false tagging
  rate precisely ? dominant source of systematic
  uncertainty

2.5 lt pT,g lt 3.5
Near Side
11
RHIC news PHENIXh-g correlations II
Nguyen QM2008

• First measurement of its kind at RHIC!
• Will measure jet shape distributions, e.g. xE,
  pout
• Constrain photon FF?
• Nfrag/Ninc 0.1 at intermediate pT
• Measure in AuAu

11
12
RHIC news STARg -h correlations I
Hamed QM2008
The ?-rich sample has lower near-side yield than
?0.
12
13
RHIC news STARg -h correlations II
Hamed QM2008
IAA of direct??
IAA of p0

• Within the current uncertainty in the scaling the
  IAA of direct ? and ?0
• are similar.

14
How can we measure direct photons in ALICE?
EMCal
D? 110º ? lt 0.7
E gt 10 GeV ? DE/E lt 3
sx 3,50 mm
TPC
D? 360º ? lt 0.9
Dp/p 2, ? 1.1º
PHOS
ITS
D? 100º ? lt 0.12
E gt 10 GeV ? DE/E lt 1.5, sx 0.5,2.5 mm
15
How many direct photons?
10k/year
Large sample of direct LO g-jet for pT lt 30
GeV/c in PHOS and pT lt 50 GeV/c in EMCal
16
?-?0 discrimination
Three regions of analysis
increasing pT

• well separated clusters
• invariant mass analysis
• lt 10 GeV/c in EMCal
• lt 30 GeV/c in PHOS

17
PHOS identified spectrum pp and PbPb annual
statistics
ALICE-INT-2005-014 G. Conesa et al., NIM A 580
(2007) 1446
Yaxian M. poster QM2008
IC R 0.3, S(pT)2 GeV/c
IC R 0.2, pTgt2 GeV/c
Statistics limits to 100 GeV
18
Direct g identification in EMCalEvent generation

• pp PYTHIA collisions
• 3 simulation cases
• pp _at_ v14 TeV
• pp _at_ v5.5 TeV, merged with HIJING, no quenching
  in PYTHIA.
• pp _at_ v5.5 TeV, merged with HIJING, quenching
  qhat 50.
• gjet in final state (MSEL10) ? ? jet
• Prompt g is the signal under study.
• 2 jets in final state (MSEL1) ? jet jet
• These events constitute the background decay g,
  fragmentation g and hadrons.
• Jet-Jet Event generation with jets containing at
  least one p0 with pTgt5 GeV/c in the acceptance of
  EMCAL. production of fragmentation photons
  suppressed.
• Also did a pp simulation without trigger in the
  same bins, similar number of events.
• 10 k events in different pT hard bins.
• Pb-Pb collisions _at_ vs 5.5A TeV pp simulations
  2) and 3) merged with HIJING central events (b lt
  3fm) _at_ vs 5.5A TeV.
• Simple PID with shower shape used to discriminate
  photons from other particles if l02 lt 0.25
  cluster is a photon (Cynthia H., PWG4 meeting,
  12/07)
• Full ALICE simulation with AliRoot 4.07 Release.

19
Generated spectra g-jet pp _at_ v14 TeV
PYTHIA
50 of g convert in the material before EMCal
(5-10 in ITS-TPC). Cluster reconstruction
efficiency is almost 95 but with shower shape
selection the efficiency descends to 60-80 due
to the non photon shape of some converted
clusters. PID rejects 15 of clusters generated
by real photons and from 50 to 25 of converted
photons.
20
Generated spectra g-jet ppPbPb _at_ v5.5 TeV
Cluster reconstruction efficiency is almost
90. With shower shape selection the efficiency
descends to 45-75 PID rejects from 45 to 10 of
clusters generated by photons and from 60 to 30
of converted photons.
21
jet-jet clusters rejection with photon PID
PbPb _at_ v5.5 TeV
pp _at_ v14 TeV
Decay g not overlapped, p0 invariant mass not done
Overlapped decay g rejection no more feasible
The jet cluster rejection goes from 0.3 (large
value due to decay g correctly identified as g,
no invariant mass analysis is done) to around
0.02 in pp collisions. There is an increase at
30-40/c GeV because PID for larger pT cannot
separate effectively overlapped g from p0 decay.
In Pb-Pb rejection worsens, from 0.5 to 0.05.
22
Prompt g / jet clusters
pp _at_ v14 TeV
PbPb _at_ v5.5 TeV
With PID prompt g to jet clusters ratio increases
significatively but it is not enough.
23
Isolation Cut Signal/Background
Ratio isolated clusters in g-jet / isolated
clusters in jet-jet Clusters selected with PID
l20lt 0.25
pp _at_ v14 TeV
PbPb _at_ v5.5 TeV
PbPb _at_ v5.5 TeV, qhat 50
Prompt photons signal larger than jet-jet
clusters background for pT larger than around 15
GeV/c for pp and quenched PbPb events
24
Isolated Spectra in EMCal
IC Parameters R0.4, pTth 0.5 (pp), 2 (PbPb)
GeV/c
pp _at_ 14 TeV
PbPb _at_ 5.5 TeV, qhat0
PbPb _at_ 5.5 TeV, qhat50
No PID
No PID
No PID
With PID
With PID
With PID
25
Direct photon in EMCal Summary conclusions

• jet-jet and g-jet events, pp and Pb-Pb
  collisions, quenched and not quenched, have been
  generated and fully reconstructed.
• A simple PID and Isolation Cut have been applied
  and the measurements seem feasible for prompt g
  energies larger than 10-20 GeV/c in pp and Pb-Pb
  quenched collisions.
• Need a large production of jet-jet events to
  improve background estimation after isolation and
  use a more sophisticated PID.

26
Other approach photon conversionsStudy
performed by Ana Marin (GSI)

• Identify photons converting in the beampipe, ITS
  and TPC
• Clean photon identification
• Provide directional information
• Non vertex background (important source of
  systematic errors in measurement of direct
  photons) can be rejected.
• Independent measurement of the same quantities,
  with different systematics compared to
  PHOS/EMCAL. Increase level of confidence in the
  results
• Counting annual statistics for pTg gt 20 GeV/c
  (very very rough stimations)

Loss of efficiency at high pT under
investigation Needs to be improved !
27
Azimuthal correlation Direct g converted
charged particles Study performed by Ana Marin
(GSI)
g detected in Central Barrel
Isolation Cut R0.2, pTgt0.7GeV
28
Prompt Fragmentation g in PYTHIA

• Have fragmentation and prompt photon similar
  properties? Looks the correlation with jets the
  same for both kind of photons?
• I have studied the prompt and fragmentation
  photon production generating several millions of
  jet-jet and g-jet events in same pT bins as in
  previous study.
• pp collisions vs14 TeV.
• Pure PYTHIA generation.
• I have looked to the isolation cut and
  correlation and fragmentation function with the
  back to back jet at this generation level.
• Only difference with previous simulations now is
  the Parton Distribution Function (PDF). Before
  CTEQ4L now CTEQ5L.

29
Prompt Fragmentation g pT distribution.
PYTHIA pp collisions vs14 TeV

• Pythia produces non hadronic decay photons
• Prompt photons Compton and annihilation
  processes
• Jet-Jet events
• Initial state photons (ISR), radiated by hard
  parton before scattering. Few
• Final state photons (FSR), scattered partons
  radiate / fragment into bremsstrahlung /
  fragmentation photons
• Spectrum of all photons in ? lt 1 in all plots.
• Isolation done on the pure PYTHIA particles,
  R0.4, pTth 1 GeV/c. Particles with status code
  1 except neutrinos enter in the cone.

? lt 1
30
Direct Fragmentation gIsolation efficiency.
PYTHIA pp collisions vs14 TeV
Direct / Fragmentation ?lt1
Isolation efficiency

• Prompt and ISR are isolated, FSR at most 50 are
  not isolated.
• Fragmentation photon (FSR) yield is larger than
  prompt photon for pT lt 50 GeV/c but with
  isolation at pT lt 20 GeV/c
• ISR yield is small, I will not consider it in
  next slides. No away side correlation.

31
Fraction of parton energy carried by the
fragmentation photon
PYTHIA pp collisions vs14 TeV
pTg gt 5 GeV/c
Obviously, fragmentation photon isolation is more
efficient when it has an small part of the
parton (near side jet) energy. Is hadron
correlation with fragmentation photons and prompt
photons the same? Could we reduce further the
amount of low z fragmentation photons correlating
with the away jet and putting a threshold on the
energy? I will check.
32
Photon and back jet type
PYTHIA pp collisions vs14 TeV

• Probability that a photon is back to a gluon or a
  quark
• Prompt photons are mainly back to quark jets
  (Compton).
• Fragmentation photons are mainly back to gluon
  jets

33
g-jet energy/phi smearing
PYTHIA pp collisions vs14 TeV
Jet energy reconstructed with the PYTHIA jet
finder. We have to be careful when we say that g
is back and has the energy of the jet. There is
an smearing to take into account.
34
Correlation Function Df
PYTHIA pp collisions vs14 TeV
All charged hadrons with pT gt 2 GeV/c
Plots for quark jets (similar for gluon jets) As
expected, without isolation fragmentation g have
a correlation in the near and away sides, direct
g only in the away side. Isolated fragmentation g
only correlate in the away side To understand
why away side descends a bit in isolated FSR
photons.
35
Fragmentation Function zT g
PYTHIA pp collisions vs14 TeV
All charged hadrons with pT gt 0.1 GeV/c inside
cone of size R1 around jet axis
Gluon Jet
Quark Jet
Difference in FF for fragmentation and prompt due
to zlt1 for fragmentation photons?
36
Prompt and Fragmentation g in PYTHIA Summary
conclusions

• PYTHIA predicts that prompt and fragmentation
  photons are produced in pp collisions at 14 TeV,
  being their ratio 0.6 at 10 GeV and increasing
  linearly to 1.4 at 90 GeV.
• Isolation cut rejects few of the fragmentation
  photons, at best around 50.
• Remaining fragmentation photons carry a
  significant amount of the original parton/jet
  energy.
• Are isolated fragmentation photons similar to
  prompt photons?
• Prompt photons and isolated fragmentation photons
  have a correlation with particles in the away
  side, not in the near side.
• But correlation/fragmentation function of both
  isolated fragmentation and prompt photons seems
  not to follow the same trend due to the fact that
  isolated fragmentation photons do not carry all
  the parton energy, and something else?

37
Back-up
38
?-?0 discrimination Shower Shape Analysis
Low particle environment
EMCal identified g l02 lt 0.25 identified p0 l02
gt 0.25
PHOS (PPR Vol2) Bayesian
g identified as g
g identified as p0
PID efficiency
pp
p0 as p0
p0 as g
see Cynthias H. talk about EMCal PID during PWG4
meeting in December for more details
pT (GeV/c)
39
PHOS identified spectrum PbPb annual statistics
Reminder
ALICE-INT-2005-014 - G. Conesa et al, NIM A 580
(2007) 1446
Particles identified as g
Corrected spectrum, systematic errors
Factor 5 suppression
Signal
Background
IC R 0.2, pTgt2 GeV/c
Statistics limits to 100 GeV
40
Direct g identification in EMCalEvent
generation and full reconstruction II
New

• pp PYTHIA collisions
• kT default value (1 GeV/c).
• 3 simulation cases
• pp _at_ v14 TeV
• pp _at_ v5.5 TeV, merged with HIJING, no quenching
  in PYTHIA.
• pp _at_ v5.5 TeV, merged with HIJING, quenching
  qhat 50.
• Pb-Pb collisions _at_ vs 5.5A TeV pp simulations
  2) and 3) merged with HIJING central events (b lt
  3fm) _at_ vs 5.5A TeV.
• Simple PID with shower shape used to discriminate
  photons from other particles if l02 lt 0.25
  cluster is a photon (see talk from Cynthia in
  December PWG4)
• Full ALICE simulation with AliRoot 4.07 Release.

41
Direct g identification in EMCalEvent
generation and full reconstruction I
New

• pp PYTHIA collisions
• gjet in final state (MSEL10) ? ? jet
• Prompt g is the signal under study.
• Generated pT hard bins (GeV) 5-10, 10-20,
  20-30, 30-40, 40-50, 50-60, 60-70,
  70-80, 80-90, 90-100 and gt100 with 10k
  events per bin.
• 2 jets in final state (MSEL1) ? jet jet
• These events constitute the background decay g,
  fragmentation g and hadrons.
• Generated pT hard bins (GeV) 12-16, 16-20,
  20-24, 24-29, 29-35, 35-41, 41-50,
  50-60, 60-72, 72-86, 86-104, 104-124,
  124-149, 149-179, 179-215, 215-258 and
  gt258 with 10k events per bin.
• Jet-Jet Event generation with jets containing at
  least one p0 with pTgt5 GeV/c in the acceptance of
  EMCAL.
• Also did a pp simulation without trigger in the
  same bins, similar number of events.

42
Prompt g / p0At generation level
PYTHIA
Theory
Ratios for pp collisions at vs14 and 5.5 TeV
agree with theoretical predictions. With qhat50
ratio a factor 2 larger than prediction.
43
jet-jet clusters effect of the p0 trigger 1 p0
with pT gt 5 GeV/c in EMCal per event
Number of clusters found in p0 triggered pp
jet-jet events divided by clusters found in non
triggered events. From now on pp p0 triggered pp
jet-jet events will be multiplied by the function
fitted in the figure to apply correction for the
triggering bias on the hadron contribution. The
fragmentation photon yield is much more
suppressed (not considered in the cluster points
and next plots)
f(pTlt30) 0.861.55 e -0.072 pT
Reconstructed clusters
Pythia generated particles (more events than in
rec. clusters)
44
Isolation cut method
Reminder
ALICE-INT-2005-014 - G. Conesa et al, NIM A 580
(2007) 1446

• Prompt g are likely to be produced isolated.
• Two parameters define g isolation
• Cone size

• pT threshold candidate isolated if
• no particle in cone with pT gt pT thres
• pT sum in cone, SpT lt SpTthres
• p T min of all particles (charged and neutral)
• in cone is at least 0.5 GeV/c
• Consider in cone particles
• Charged in TPC acceptance ?lt0.7, 0ºlt?lt360º
• Neutral in
• EMCal acceptance ?lt0.7, 80ºlt?lt190º
• PHOS acceptance ?lt0.12, 220ºlt?lt320º

45
Isolation Cut Efficiency
Ratio isolated clusters / total clusters No PID
pp _at_ v14 TeV
PbPb _at_ v5.5 TeV
PbPb _at_ v5.5 TeV, qhat 50
g-jet
jet-jet
Isolation rejection for jet clusters can be
better than 99 in pp and quenched PbPb events
and pT gt 20 GeV/c
46
Isolation Cut Efficiency pp _at_ v14 TeV
Ratio isolated clusters / total clusters No PID
pTth 0.5 GeV/c
R0.3
g-jet
g-jet
jet-jet
jet-jet
47
Isolation Cut Efficiency ppPbPb _at_ v5.5 TeV, no
quenching
Ratio isolated clusters / total clusters No PID
pTth 2 GeV/c
R0.3
g-jet
g-jet
jet-jet
jet-jet
48
Isolation Cut Efficiency ppPbPb _at_ v5.5 TeV,
qhat 50
Ratio isolated clusters / total clusters No PID
pTth 2 GeV/c
R0.3
g-jet
g-jet
jet-jet
jet-jet
49
Prompt photon at parton level
Partonic photons and back-to-back quarks / gluons
do not have the same momentum distribution, there
is an energy smearing, and they are not exactly
back-to-back.
There is also a difference in pT (not in azimuth)
between the partonic photon (status 21) and the
final state photon (status 1). Where does it come
from?, kT? This difference is usually small but
not always.
50
Direct Fragmentation g pT distribution and ratio
pT spectra (full f and ?, and ?lt0.7)
Direct/Fragmentation ratio (full f and ?, and
?lt0.7)
51
Direct Fragmentation g ? distribution and ratio
pT spectra (full f, pT gt 5 GeV/c)
Direct/Fragmentation ratio (full f, pT gt 5 GeV/c)
Particle production seems independent of
pseudo-rapidity in the range -5 to 5. I have
studied the isolation cut in the EMCal ?
acceptance, results should be equivalent for the
PHOS ? acceptance
52
Correlation fuction for ISR
53
Correlation Function zT
PYTHIA pp collisions vs14 TeV
All charged hadrons with pT gt 2 GeV/c and at ?g
lt 1, hadrons in all eta. 0.5p lt f g f hadron
lt 1.5p
Gluon Jet
Quark Jet
The correlation function of isolated
fragmentation photons seems not to be the same
as of prompt photons, effect of the zlt1 for
fragmentation photons?
54
CF versus FF
PYTHIA pp collisions vs14 TeV
Quark Jet
Gluon Jet
Prompt
Dont understand prompt-gluon ratios. Other plots
show similar trend, needs check.
Fragmentation