Jets at RHIC
Sevil Salur UC Davis & Rutgers U.
η φ
Sevil Salur 2
Why we should study jets in heavy ion collisions?
But are we succeeding?
Goal is Unbiased Jet Reconstruc=on:
Reconstruct partonic kinema=cs independent of fragmenta=on details -‐ quenched or unquenched and study the QGP dynamics.
Full Jet Reconstruc=on: Enables study of jet quenching at the partonic level.
• Uniquely large kinema=c reach
• In A+A much reduced geometric biases, full explora=on of quenching.
• Mul=ple channels for consistency checks: Inclusive, di-‐jets, h-‐jets, gamma-‐jets
• Qualita=vely new observables: energy flow, jet substructure, fragmenta=on func=on
Sevil Salur 3 WWND 08 South Padre -TX
dNJet/dE
T (pe
r even
t) Nbin scaled p+p
STAR Prelim
inary
MB-‐Trig O HT-‐Trig
R=0.4 pT cut =1 GeV Seed=4.6 GeV
LOHSC
S. Salur [STAR Collabora=on], Hard Probes 2008
p+p: Phys. Rev. Lec. 97 (2006) 252001
Sevil Salur 4
J. Putschke [STAR Collabora=on], Hard Probes 2008
Hard Probes 2008
Agreement with Nbin scaled p+p
Resolu=on effect corrected assuming Pythia Fragmenta=on.
Trigger Biases due to event recording
Ra=o of A+A to pp in FF~1.
Tower energy threshold bias No modifica=on in FF…
Sevil Salur 5 WWND 09 Big Sky - MT
Sevil Salur 6
E. Bruna [STAR Collabora=on], Quark MaVer 2009
(GeV/c)JetT
p0 10 20 30 40 50 60
-1) (
GeV/
c) d T
dN/(d
pev
ent
1/N
-910
-810
-710
-610
-510
-410
-310
kt R=0.4
anti-kt R=0.4
kt R=0.2
anti-kt R=0.2
STAR Preliminary
=200 GeV/cNNsAu+Au at 10% most central events
M. Ploskon [STAR Collabora=on], Quark MaVer 2009
Quark MaVer 2009
Correct via unfolding:
Exactly the same algorithms and jet defini=ons used as in p+p collisions
Remove the trigger bias:
Apparent suppression of the recoil jet of Au+Au with respect to p+p.
“recoil” “trigger”
dNJet/dE
T (pe
r even
t) Nbin scaled p+p
STAR Prelim
inary
MB-‐Trig O HT-‐Trig
R=0.4 pT cut =1 GeV Seed=4.6 GeV
LOHSC
S. Salur [STAR Collabora=on], Hard Probes 2008
p+p: Phys. Rev. Lec. 97 (2006) 252001
Sevil Salur 7
E. Bruna [STAR Collabora=on], Quark MaVer 2009
J. Putschke [STAR Collabora=on], Hard Probes 2008
(GeV/c)JetT
p0 10 20 30 40 50 60
-1) (
GeV/
c) d T
dN/(d
pev
ent
1/N
-910
-810
-710
-610
-510
-410
-310
kt R=0.4
anti-kt R=0.4
kt R=0.2
anti-kt R=0.2
STAR Preliminary
=200 GeV/cNNsAu+Au at 10% most central events
M. Ploskon [STAR Collabora=on], Quark MaVer 2009
Y. S. Lai [PHENIX Collabora=on] Quark MaVer 2009
PHENIX can reconstruct jets! Large pT range within restricted PHENIX acceptance!
dNJet/dE
T (pe
r even
t) Nbin scaled p+p
STAR Prelim
inary
MB-‐Trig O HT-‐Trig
R=0.4 pT cut =1 GeV Seed=4.6 GeV
LOHSC
S. Salur [STAR Collabora=on], Hard Probes 2008
p+p: Phys. Rev. Lec. 97 (2006) 252001
Sevil Salur 8
E. Bruna [STAR Collabora=on], Quark MaVer 2009
J. Putschke [STAR Collabora=on], Hard Probes 2008
(GeV/c)JetT
p0 10 20 30 40 50 60
-1) (
GeV/
c) d T
dN/(d
pev
ent
1/N
-910
-810
-710
-610
-510
-410
-310
kt R=0.4
anti-kt R=0.4
kt R=0.2
anti-kt R=0.2
STAR Preliminary
=200 GeV/cNNsAu+Au at 10% most central events
M. Ploskon [STAR Collabora=on], Quark MaVer 2009
Y. S. Lai [PHENIX Collabora=on] Quark MaVer 2009
Story is not over yet. Experimental & Theore=cal improvements
IN PROGRESS!
Reconstructed Jets at RHIC: What to do next?
Sevil Salur UC Davis & Rutgers U.
η φ
Experimental Challenges
1) Data taking -‐ Trigger-‐bias
-‐ Biased fragmenta=on func=ons
2) Par=cle and jet level correc=ons
Calorimeter calibra=on
-‐ ~5% uncertainty on calibra=on translates to a larger uncertainty.
Double coun=ng of par=cle energies:
-‐ electrons: Include energy only once!
-‐ hadrons: showering correc=ons such as MIP, constant E-‐frac=on
High momentum tracking
-‐ Do not include cosmics
Unobserved energy
-‐ Energy shio (protons for neutrons)
Jet PT resolu=on
Sevil Salur 10
Cone Algorithm kT Algorithm
hcp://www-‐cdf.fnal.gov/physics/new/qcd/QCD.html
11 Sevil Salur
Jet Reconstruc=on is not a new phenomena…
Guidance from earlier measurements …
Cone Algorithm kT Algorithm
hcp://www-‐cdf.fnal.gov/physics/new/qcd/QCD.html
12 Sevil Salur
Jet Reconstruc=on is not a new phenomena…
A. Bhap et. al. Nucl.Instrum.Meth.A566:375-‐412 (2006)
Frac=onal energy loss vs momentum MC comparisons…
Experimental Challenges
1) Data taking -‐ Trigger-‐bias
-‐ Biased fragmenta=on func=ons
2) Jet Level and Par=cle Correc=ons High momentum tracking
-‐ Including cosmics
Calorimeter calibra=on
-‐ ~5% uncertainty on calibra=on translates to large uncertainty in cross-‐sec=ons.
Double coun=ng of par=cle energies:
-‐ electrons: Include energy only once!
-‐ hadrons: showering correc=ons
such as MIP, constant E-‐frac=on
Unobserved energy:
-‐ Energy shio (protons for neutrons)
Jet PT resolu=on
Sevil Salur 13
Fundamental Assump=on: Two separable components: signal and background.
How might it be violated? Biases in background es=ma=on due to presence of a jet.
a) Ini=al state radia=on (Expected to be small compared to jet energy). b) Final state “out-‐of-‐cone” radia=on.
3) Heavy–Ion Background
Some others like Fake Jets…
Various Jet Reconstruc=on Algorithms
Cone jet An=-‐kT jet
KT jet
Jet
outgoing parton
Fragmentation process
Hard scatter
1. Mid Point Cone: Merging & Splipng 2. SIS CONE
• Insensitive to “soft” radiation • Splitting doesn’t change jets
3. Leading Order High Seed Cone (LOHSC)
4. KT (star=ng point: low pT par=cles) 5. An=-‐KT (star=ng point: high pT par=cles)
Cone Algorithm:
Sequen=al recombina=on: Cluster pairs of objects close in rela=ve pT
14 Sevil Salur
6. Gaussian filtering. Y. Lai, B. Cole arXiv:0806.1499
Different Algorithms respond differently to background!
The FastJet Algorithms
Suite of modern Collinear and infrared safe jet algorithms • sequen=al recombina=on: kT, Cambridge/Aachen, an=-‐kT • cone: SISCone (Seedless Infrared-‐safe Cone)
Mo=vated by high precision jets in high luminosity p+p at LHC (pileup) • but directly applicable to heavy ion collisions
Two important algorithmic advances: 1. Large improvements to processing =me vs. event mul=plicity 2. Rigorous defini=on of jet area for subtrac=on of diffuse event background
M. Cacciari, G. Salam 0707.1378 [hep-‐ph]
M. Cacciari, G. Salam, G. Soyez 0802.1188 [hep-‐ph] FastJet – hcp://www.lpthe.jussieu.fr/~salam/fastjet
pT (Jet Measured) ~ pT(Parton) + ρ × A(Jet) ± σ√A(Jet)
15 Sevil Salur
A, ρ, σ are all measurable quan==es!
A= Jet Area ρ = Diffuse noise, σ=noise fluctuations
Assume: Signal and background can be separated
Sevil Salur 16
Jet Measurements
1) Minimize the kinema=c cuts. 2) Data driven correc=ons :
a. Experimental characteriza=on of background fluctua=ons. b. Detailed unfolding of fluctua=ons & energy resolu=on. Correc=ng for pT smearing
Correc=on depends highly on the background model.
LOHSC
seed=4.6 GeV R=0.4
PyDet PyEmbed PyTrue
ET [GeV]
dNJet/dE
T (a.u.)
STAR Preliminary
S. Salur [STAR Collabora=on], arXiv:0809.1609 [nucl-‐ex]
Sevil Salur 17
(GeV/c)JetT
p0 10 20 30 40 50 60
-1) (
GeV
/c)
d T d
N/(d
pev
ent
1/N
-910
-810
-710
-610
-510
-410
-310
STAR Preliminary
=200 GeV/cNNsAu+Au at 10% most central events
kt R=0.4
anti-kt R=0.4
(GeV/c)JetT
p
0 10 20 30 40 50 60)
-1) (
pb (G
eV/c
)T
dp
/(d2
) d1/
(2
-110
1
10
210
310
410
510
kt R=0.4
anti-kt R=0.4
Published STAR p+p
=200 GeV/cNNsp+p at
STAR Preliminary
Jet Measurements
An=-‐kT and kT jet spectra are consistent.
Unfolding uncertainty corresponds to a factor of 2 in jet cross-‐sec=on.
Jet Energy Scale Uncertainty
1 σ Unfolding Uncertainty
Jet Energy Scale Uncertainty
M. Ploskon [STAR Collabora=on], Quark MaVer 2009
RAA of Jets
(GeV/c)JetT
p10 15 20 25 30 35 40 45 50
Jet
AA
R
-110
1
STAR Preliminary
=200 GeV/cNNsAu+Au and p+p at Au+Au: 10% most central
kt R=0.4
anti-kt R=0.4
R = 0.4
A large frac=on of jets are reconstructed! (Compare pion RπAA = 0.2)
Sevil Salur 18
Jet Energy Scale Uncertainty
We need to constrain the approxima=on of the Jet Energy Scale Uncertainty…
RAA < 1 : unable to recover complete jet energy -‐ jet broadening
RAA = 1 : recover complete jet energy
M. Ploskon [STAR Collabora=on], Quark MaVer 2009
What happens at high pT?
Sevil Salur 19
W. Vogelsang Private Communica=on
p+p
Rela=ve contribu=ons of quark and gluon vary.
What about quenching dependence on parton species?
Rela=ve contribu=on of sub-‐processes to inclusive jet produc=on
What happens at high pT?
Sevil Salur 20
The EMC Effect: Devia=on between structure Func=ons of Au and deuterium.
W. Vogelsang Private Communica=on
p+p
Ini=al state effects at large x ~15% σΑ/σ
d
What about other high effects?
Rela=ve contribu=on of sub-‐processes to inclusive jet produc=on
Rela=ve contribu=ons of quark and gluon vary.
J. Gomez et al., SLAC–PUB–5813 June 7, 2001 D.F. Geesaman et al., Ann. Rev. Nucl. Part. Sci. 45, 337 (1995) B. A. Cole. et al, arXiv:hep-‐ph/0702101
SLAC E139
Nicolas Borghini arXiv: 0902.2951 Ivan Vitev et al JHEP 0811:093 (2008), arXiv:0810.2807
21 Sevil Salur
JEWEL (Jet Evolu=on with Energy Loss): K. Zapp, G. Ingelman, J. Rathsman, J. Stachel, U. A. Wiedemann arXiv:0805.4759 Parton shower with microscopic descrip=on of interac=ons with medium – random process of scacering is used to modify the forma=on =me of the radiated gluon.
Q-‐Pythia: N. Armesto, L. Cunqueiro and C. A. Salgado arXiv:0809.4433[hep-‐ph] MC implementa=on in Pythia of medium-‐induced gluon radia=on through an addi=ve term in the vacuum splipng func=ons.
Analy=c Calcula=ons:
Theore=cal Challenges: Rigorous calcula=ons of full jet variables? Model building of in-‐medium modifica=ons…
Many more…. Mar=ni (S. Jeon), PYQUEN (Lokh=n, Snigriev), PQM (Dainese, Loizides, Paic), HIJING (Gyulassy, Wang)…
YaJEM: T. Renk arXiv:0808.1803
Ivan Vitev DPF 2009
Sevil Salur 22
RAA of Jets at RHIC
Vary Resolu=on Parameter Implica=on of broadening of jets
I. Vitev et al. arXiv:0910.1090v1 [hep-‐ph]
NLO Calcula=on:
Sevil Salur 23
RAA of Jets at RHIC
Vary Resolu=on Parameter Implica=on of broadening of jets
But before can do this: constrain systema=c uncertain=es & jet energy resolu=on
I. Vitev et al. arXiv:0910.1090v1 [hep-‐ph]
NLO Calcula=on:
(GeV/c)JetT
p10 15 20 25 30 35 40 45 50
Jet
AA
R
-110
1
STAR Preliminary
=200 GeV/cNNsAu+Au and p+p at Au+Au: 10% most central
kt R=0.4
anti-kt R=0.4
R = 0.4
(GeV/c)Jet
Tp
15 20 25 30 35 40 45 50
Yie
ld R
ati
o:
R=
0.2
/R=
0.4
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
STAR Preliminary
=200 GeV/cNN
sAu+Au and p+p at
Au+Au: 10% most central p+p kt
p+p anti-kt
Vary Resolu=on Parameter: R=0.2/R=0.4
24
p+p: “Narrowing” of the jet structure with increasing jet energy
Sevil Salur
QM 09
(GeV/c)Jet
Tp
15 20 25 30 35 40 45 50
Yie
ld R
ati
o:
R=
0.2
/R=
0.4
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
STAR Preliminary
=200 GeV/cNN
sAu+Au and p+p at
Au+Au: 10% most central p+p kt
p+p anti-kt
Vary Resolu=on Parameter: R=0.2/R=0.4
25
p+p: “Narrowing” of the jet structure with increasing jet energy
Sevil Salur
NLO Calculation W. Vogelsang – priv. comm. 2009
Broader jet profile further inves=ga=on is required
QM 09
Vary Resolu=on Parameter: R=0.2/R=0.4
26
p+p: “Narrowing” of the jet structure with increasing jet energy
Sevil Salur
QM 09
Au+Au: Broadening of the jet energy profile
Suppression in Au+Au collisions -‐> loss of jet yield for R=0.2
PT Cut
KT KT KT
STAR Prelim
inary
STAR Prelim
inary
STAR Prelim
inary
Au+Au 0-‐10%
MB-‐Trig Nbin Scaled p+p
Au+Au 0-‐10%
MB-‐Trig Nbin Scaled p+p
Au+Au 0-‐10%
MB-‐Trig Nbin Scaled p+p
Vary PT Cut : Bias or Physics?
A good test for quenched Monte Carlo’s and heavy ion background.
Highly sensi=ve to fragmenta=on …
S. Salur [STAR Collabora=on], arXiv:0810.0500 [nucl-‐ex]
27 Sevil Salur
STAR Preliminary
Au+Au Central
η φ
pTcut
R=0.4
K. Zapp, G. Ingelman, J. Rathsman, J. Stachel, U. A. Wiedemann arXiv:0804.3568
Significant uncertain=es due to the sensi=vity to hadronisa=on: Look for new observables unaffected by the hadronisa=on.
Fragmenta=on func=ons: Parton vs Hadron
Clear increase in mul=plicity due to radia=ve energy loss
28 Sevil Salur
See talks of Elena Bruna Ant Timmins Yue Shi Lai
fmed = 3
Ecut = 0 GeV
medium1 jet vacuum2 jets vacuum3 jets vacuum
0.2
0.4
0.6
0.8
1.2
n-jet
frac
tion
0
1
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5 0log10(resolution scale)
Sevil Salur 29
QCD JET Observables
pT cut infrared safe insensi=ve observables! : number of subjets, thurst …
K. Zapp, G. Ingelman, J. Rathsman, J. Stachel, U. A. Wiedemann arXiv:0804.3568
In vacuum (LEP) data well understood in pQCD
Medium Induced Radia=on More Coarser Jet Structure
Cone jet An=-‐kT jet
KT jet
We need to confront with data….
Sevil Salur 30
• Full jet reconstruc=on at RHIC is possible. A new tool to study partonic energy loss…
• First studies indicate significant broadening of the jet energy profile due to quenching
• Experimental & Theore=cal challenges
• Consistency Checks Mul=ple channels: Inclusive, di-‐jets, h-‐jets, γ-‐jets
• New observables: energy flow, jet substructure…
LONG ROAD AHEAD….
Sevil Salur 31
• Full jet reconstruc=on at RHIC is possible. A new tool to study partonic energy loss…
• First studies indicate significant broadening of the jet energy profile due to quenching
• Experimental & Theore=cal challenges
• Consistency Checks Mul=ple channels: Inclusive, di-‐jets, h-‐jets, γ-‐jets
• New observables: energy flow, jet substructure…
LONG ROAD AHEAD….
… MORE TO COME SOON!