Elena G. Ferreiro Universidade de Santiago de Compostela, Spain

EMMI workshop “Quarkonia in Deconfined Matter”, Acitrezza Sicily, 27 Sept ‐ 1 Oct 2011

Elena G. Ferreiro Universidade de Santiago de Compostela, Spain

Work done in collaboration with F. Fleuret, J‐P. Lansberg , N. Matagne and A. Rakotozafindrabe EPJC61 (2009), PLB680 (2009), PRC81 (2010), NPA855 (2011)

Some definitions… Charmonium: heavy quark bound states made of charm J/Ψ meson: bound state of a charm quark and its antiquark QGP: deconfined matter made of quarks and gluons, supposed to exist in the first instants after Big Bang The goal: search of a QGP in heavy‐ions collisions (high T and density)

Looking for QGP signals: Matsui & Satz, PLB178 (1986) 416

E. G. Ferreiro USC

unambiguous” signature of QGP Onset of quarkonia melting above a certain temperature / energy density threshold

CNM effects on quarkonium @ RHIC and LHC

EMMI 28 Sept 2011

Introduction I: the intringuing story of J/ψ production Potential between q‐anti‐q pair grows linearly at large distances

4 αs V(r) = − + kr 3 r

V(r)

Screening of long range confining potential at high enough temperature or density.

V(r)

r

What happens when the range of the binding force becomes smaller than the radius of the state? different states “melting” at different temperatures due to different binding energies. Matsui and Satz: J/ψ destruction in a QGP by Debye screening

J/Ψ suppression = QGP signature

E. G. Ferreiro USC


EMMI 28 Sept 2011

Introduction I: the intringuing story of J/ψ production • J/Ψ suppression at SPS Suppression beyond nuclear absorption observed in central Pb+Pb at √s ~ 17 GeV CERN communicate: SPS results presented a compelling evidence for the existence of a new state of matter in which quarks, instead of being bound up into more complex particles such as protons and neutrons are liberated to roam freely.

• J/Ψ suppression at RHIC J/Ψ are suppressed, but not as much as expected if we have complete color screening Puzzle at RHIC: Same amount of suppression at RHIC and SPS √s≈200 GeV √s≈20GeV

At RHIC, stronger suppression at forward y => Recombination? E. G. Ferreiro USC


BNL 6‐18 June 2011

Introduction II: too many effects on J/ψ production … nuclear absorption

CGC

gluon

low x

cronin effect J Ïˆ

sequential suppresion

percolation

gluon shadowing

hadronic comovers cc

cc

J/ψ c

D

pomeron shadowing

co movers

D

c-bar c

cc

recombination partonic comovers QGP parton saturation E. G. Ferreiro USC

c

c Color Screening


EMMI 28 Sept 2011

Introduction III: COLD or HOT effects? •cold effects:

wo thermalisation NO QGP

gluon shadowing nuclear structure functions in nuclei ≠ superposition of constituents nucleons

nuclear absorption multiple scattering of a preresonance c-cbar pair within the nucleons of the nucleus

•hot effects: QGP E. G. Ferreiro USC

partonic comovers

percolation parton saturation

hadronic comovers

non-lineal effects favoured by the high density of partons become important and lead to eventual saturation of the parton densities

NI@SPS, IMP@RHIC

IMP@SPS, RHIC?

CGC

non thermal colour connection

dissociation of the c-cbar pair with the dense medium produced in the collision partonic or hadronic

suppression by a dense medium, not thermalized

Others: Cronin effect energy loss

w thermalisation QGP

sequential suppression CNM effects on quarkonium @ RHIC and LHC

recombination EMMI 28 Sept 2011

Introduction: motivation • A lot of work trying to understand A+A data (since J/ψ ≡ QGP signal)

Quarkonium as a hint of deconfinement • If we focalise on p+A data (where no QGP is possible) only cold nuclear matter (CNM) effects are in play here: shadowing and nuclear absorption EMC and energy loss

Quarkonium as a hint of coherence • In fact, the question is even more fundamental: p+p data we do not know the specific production kinematics at a partonic level: (2→2,3,4) vs (2→1)

Quarkonium as a hint of QCD E. G. Ferreiro USC


EMMI 28 Sept 2011 1

Introduction : contents Our goal: To investigate the CNM effects and the impact of the specific partonic production kinematics 3 ingredients: •J/ψ partonic production mechanism •Shadowing •Nuclear absorption • Results on J/ψ production @ RHIC and LHC

•

To extend our study to ϒ CNM effects : •fractional energy loss •gluon EMC effect • Results on ϒ production @ RHIC E. G. Ferreiro USC


EMMI 28 Sept 2011 2

Quarkonium as a tool of COLD and HOT effects •cold effects:

wo thermalisation NO QGP

gluon shadowing gribov shadowing nuclear structure functions v in nuclei ≠ superposition of constituents nucleons

nuclear absorption multiple scattering of a preresonance c-cbar pair within the nucleons of the nucleus

•hot effects: QGP E. G. Ferreiro USC

partonic comovers

percolation parton saturation

hadronic comovers

non-lineal effects favoured by the high density of partons become important and lead to eventual saturation of the parton densities

NI@SPS, IMP@RHIC

IMP@SPS, RHIC?

CGC

non thermal colour connection

dissociation of the c-cbar pair with the dense medium produced in the collision partonic or hadronic

suppression by a dense medium, not thermalized

Others: Cronin effect EMC effect, energy loss v

w thermalisation QGP

sequential suppression CNM effects on quarkonium @ RHIC and LHC

recombination EMMI 28 Sept 2011

J/ψ production mechanisms • Color Singlet Model: – – –

perturbative creation of the ccbar pair in color singlet state with subsequent binding to J/ψ with same quantum numbers hard gluon emission underpredicts J/ψ production cross section, predicts no polarization

2→2 g+g → J/ψ+g

• Color Evaporation Model: – – –

phenomenological approach perturbative creation of the ccbar pair in the color octet state with subsequent non‐perturbative hadronization to color singlet via unsuppressed soft gluon emission 2→1 predicts no polarization

• NRQCD Color Octet Model: – – –

g+g → J/ψ

uses NRQCD formalism to describe the non‐perturbative hadronization of the ccbar color octet to the color singlet state via soft gluon emission factorizes the charmonium production into a short distance hard part and a long distance matrix element which is claimed to be universal Predicts large transverse polarization at high pT (not seen by data) E. G. Ferreiro USC

Abigail Bickley, August 9, 2007 CNM effects on quarkonium @ RHIC and LHC

10 EMMI 28 Sept 2011

Shadowing: an initial cold nuclear matter effect • Nuclear shadowing is an initial‐state effect on the partons distributions • Gluon distribution functions are modified by the nuclear environment • PDFs in nuclei different from the superposition of PDFs of their nucleons Shadowing effects increases with energy (1/x) and decrease with Q2 (mT)

antishadowing

The shadowing corrections shadowing EMC depend on the partonic process producing the J/Ψ since it affects kinematics (x,Q2) E. G. Ferreiro USC


EMMI 28 Sept 2011 3

Nuclear absorption: a final cold nuclear matter effect Particle spectrum altered by interactions with the nuclear matter they traverse => J/Ψ suppression due to final state interactions with spectator nucleons • Usual parameterisation: (Glauber model)

Sabs = exp(‐ρ σabs L )

nuclear matter density break‐up cross section path length Energy dependence • At low energy: the heavy system undergoes successive interactions with nucleons in its path and has to survive all of them => Strong nuclear absorption • At high energy: the coherence length is large and the projectile interacts with the nucleus as a whole => Smaller nuclear absorption In terms of formation time:

C. Lourenço et al.

Rapidity dependence of nuclear absorption? E. G. Ferreiro USC

σabs @ mid y Investigating two production mechanisms (including pT for the J/ψ):

2→1

g+g → J/ψ

• intrinsic scheme: the pT of the J/ψ comes from initial partons

Not relevant for, say, pT>3 GeV Only applies if COM(LO, αs2) is the relevant production mechanism at low pT

g+g → J/ψ+g, gg,ggg,… 2→2, 3, 4 •extrinsic scheme: the pT of the J/ψ is balanced by the outgoing parton(s) COM, CSM (NLO, NNLO)

E. G. Ferreiro USC

for a given y, larger x in extrinsic scheme => modification of shadowing effects


EMMI 28 Sept 2011 5

Intrinsic J/ψ production kinematics • Intrinsic scheme: 2 → 1 process CEM @ LO • y, pT can be determined using PHENIX p+p data Phys. Rev. Lett. 98, 232002 (2007)

• Easy to handle : yJ/ψ and pTJ/ψ directly give x1,2

Q2=(2mc)2+(pT)2=mT • Straightforward evaluation of the gluon PDF shadowed in the nucleus at x2 (and x1 in AA) E. G. Ferreiro USC


EMMI 28 Sept 2011

Extrinsic J/ψ production kinematics • We deal with a 2 → 2 partonic process with collinear initial gluons • The quadri‐momentum conservation results in a complex expression of x2 as a function of (x1, y, pT) • Information from the data alone – the y and pT spectra– is not sufficient to determine x1 and x2: the presence of a final‐state gluon authorizes much more freedom to choose (x1, x2) for a given set (y, PT) • Models are mandatory to compute the proper weighting of each kinematically allowed (x1, x2) •We use s—channel cut mechanism Extension of CSM Haberzettl and Lansberg, Phys.Rev.Lett.100,032006 (2008) Also 2 ‐> 2: CSM @ LO, CEM @NLO E. G. Ferreiro USC


Good results at low pT EMMI 28 Sept 2011

On the kinematics of J/ψ production: equations

shadowing

partonic cross section

nuclear absorption

fit to data E. G. Ferreiro USC

kinematic variables CNM effects on quarkonium @ RHIC and LHC

your preferred model EMMI 28 Sept 2011

Extrinsic vs intrinsic kinematics I Intrinsic scheme

Extrinsic scheme

both implemented in a Monte Carlo code: JIN E.G. Ferreiro, F. Fleuret, and A. Rakotozafindrabe ,Eur. Phys. J. C61, 859 (2009) E.G. Ferreiro, F. Fleuret, J‐P. Lansberg and A. Rakotozafindrabe, Phys.Lett.B680, 50 (2009) E. G. Ferreiro USC


EMMI 28 Sept 2011

INTRINSIC (2→1) vs EXTRINSIC (2 → 2) kinematics 2 → 1

2 → 2 For a given set (y, pT): extrinsic scheme: more freedom for x for a given y => larger x in extrinsic scheme

We expect different shadowing effects in both cases

E. G. Ferreiro USC


EMMI 28 Sept 2011 7

2→2 g+g → J/ψ+g

2→1 g+g → J/ψ

Results d+Au @ RHIC: J/ψ rapidity dependence of RdAu

• shadowing depends on the partonic process: 2→1 or 2→2 arXiv:0912.4498 • antishadowing peak shifted toward larger y in the extrinsic case • in order to reproduce data @ RHIC: nuclear absorption σabs extrinsic > σabs intrinsic the kinematics matter for the extraction of σabs E. G. Ferreiro USC


EMMI 28 Sept 2011 6

2→2 g+g → J/ψ+g

2→1 g+g → J/ψ

Results d+Au @ RHIC: J/ψ rapidity dependence of RdAu

for a given y x larger in extrinsic

• shadowing depends on the partonic process: 2→1 or 2→2 arXiv:0912.4498 • antishadowing peak shifted toward larger y in the extrinsic case • in order to reproduce data: nuclear absorption σabs extrinsic > σabs intrinsic the kinematics matter for the extraction of σabs E. G. Ferreiro USC


EMMI 28 Sept 2011 7

Fit of σabs with EKS, EPS and nDS(g) from RdAu

EKS98: compatible with intrinsic & extrinsic

EPS08: extrinsic scheme is favorized

nDSg: neither extrinsic nor intrinsic… E. G. Ferreiro USC


EMMI 28 Sept 2011

Results d+Au @ RHIC: J/ψ rapidity dependence of RCP Extrinsic scheme: σabs= 0, 2, 4, 6 mb in 3 shadowing models

Data dependence on y: • Suppression for the most forward points in the three centrality ranges • In the negative rapidity region, dominated by large x, no (or compensated) nuclear effects abs(y)? Data at back and mid‐y can be described with a σabs of 2–4 mb, while the most forward points seem to decrease more than our evaluation E. G. Ferreiro USC CNM effects on quarkonium @ RHIC and LHC EMMI 28 Sept 2011

σ

Fit of σabs with EKS, EPS and nDS(g) from RdAu and RCP σabs and χ2from RdAu intrinsic extrinsic

EKS

σabs int progressive increase of RdAu vs Ncoll • in the forward region: shadowing => progressive decrease of RdAu vs Ncoll E. G. Ferreiro USC


EMMI 28 Sept 2011

Results d+Au @ RHIC: J/ψ transverse momentum dependence Extrinsic scheme: σabs= 0, 2, 4, 6 mb in 3 shadowing models

EKS98

EPS08

nDSg

Growth of RdAu not related to Cronin effect: it comes from the increase of x for increasing PT • in the mid and forward‐y region: x goes through the antishadowing region => enhancement in RdAu • In the backward region: x sits in an antishadowing region=> decrease in RdAu E. G. Ferreiro USC


EMMI 28 Sept 2011 9

Results Au+Au @ RHIC: J/ψ centrality and y dependence mid‐y & forward‐y Intrinsic scheme: 2→1

2→1 & 2→2 process

Extrinsic scheme: 2→2

Extrinsic scheme : RAA @ forward y stronger suppression at mid y E. G. Ferreiro USC


EMMI 28 Sept 2011 11

Work in progress: J/ψ @ LHC rapidity dependence (2 → 2)

RHIC

Extrinsic EKS98 sabs=0, 1.5, 2.8 NPA855 (2011)

This behaviour is attenuated when going to higher pT

Opposite CNM RAA behaviour vs rapidity @ RHIC and LHC: • At RHIC=> stronger suppression at forward y • At LHC => stronger suppression at mid y E. G. Ferreiro USC



Work in progress: J/ψ @ LHC centrality dependence

Extrinsic EKS98 sabs=0, 1.5, 2.8 NPA855 (2011) E. G. Ferreiro USC



Work in progress: J/ψ @ LHC centrality dependence (2 → 2) “CEM NLO” before kT smearing

“Traditional” 2 → 2

CMS preliminary

E. G. Ferreiro USC


CMS preliminary



E. G. Ferreiro USC



EMMI 28 Sept 2011



with kt smearing

without kt smearing kt smearing procedure is applied after the (x1,x2) integration

E. G. Ferreiro USC

underlying partonic model



Note on the underlying partonic model 2 different 2‐> 2 models can give different results z Example : with the existing code for CEM @ NLO, the kt smearing procedure is applied after the (x1,x2) integration z

Before the smearing (left) the distribution overhsoots the data z More weight on low pT's=> the distribution used is closer to a 2 ‐> 1 process z The CEM @ NLO is a mix between a pure collinear 2‐>2 and a pure 2‐>1 with intrinsic kt z

E. G. Ferreiro USC


BNL 6‐18 June 2011

Note on the shadowing and its uncertainties at LHC energies As we have seen, different 2‐>2 partonic models can give different results z We have used 2 'toy' models : z

z

z

We use nDSg and EKS98 as possible gluon shadowings (non‐exhaustive) Finally we vary μF from 0.5 x mT to 2 x mT (as done in pp for g(x,μF)

E. G. Ferreiro USC


EMMI 28 Sept 2011


E. G. Ferreiro USC



EMMI 28 Sept 2011

CEM NLO inspired 2‐> 2 peacked at low pT (to be smeared out) For pT>0: Stronger shadow suppression at mid rapidity

For pT>6.5: Slightly stronger shadowing suppression at mid rapidity

nDSg shadowing > EKS shadowing

E. G. Ferreiro USC


EMMI 28 Sept 2011

CEM NLO inspired 2‐> 2 peacked at low pT (to be smeared out) For pT>0: Stronger shadow suppression at mid rapidity



E. G. Ferreiro USC


EMMI 28 Sept 2011

“Traditional” 2 ‐> 2 For pT>0: Stronger shadow suppression at mid rapidity



E. G. Ferreiro USC


EMMI 28 Sept 2011

“Traditional” 2 ‐> 2 For pT>0: Stronger shadow suppression at mid rapidity



E. G. Ferreiro USC


EMMI 28 Sept 2011

Work in progress: J/ψ @ LHC pT dependence (2 → 2)

Shadowing decreases with increasing pT Stronger variation for EKS than nDSg EKS: 25‐40% nDSg: 15‐30%

pT matters!!! E. G. Ferreiro USC



CNM effects: Comparing A+A results @ RHIC and LHC RHIC

Opposite CNM behaviour vs y Data @ RHIC: stronger supp. at frwrd y Data @ LHC: stronger suppression at mid y (pT>6)

LHC

Data & CNM: similar tendency @ RHIC & LHC

RHIC

Same CNM behaviour vs pT Data @ RHIC: stronger supp. at low pT Data @LHC: similar suppression with pT ?

LHC

Data & CNM: similar tendency @RHIC tendency @ LHC?

E. G. Ferreiro USC



Comparing A+A experimentalresults @ RHIC and LHC

E. G. Ferreiro USC



Comparing A+A experimental results @ RHIC and LHC Mid y and large pT RAA RHIC > RAA LHC

Forward y and low pT RAA LHC > RAA RHIC

Recombination at LHC? (stronger at mid y and pT=0)

ALICE data at mid y needed! If recombination, RAA at mid y > RAA at forward y (pT>0)

Place for recombination effects @ LHC

E. G. Ferreiro USC



On the kinematics of ϒ production Results at 1.8 TeV: • CSM describes well dσ/dpT at NNLO • LO CSM is sufficient to describe low pT data

2 → 2 process Results at 200 GeV: LO upper line: mb = 4.5 GeV, μR = MT , μF = 2MT LO lower line: mb = 5.0 GeV, μR = 2MT , μF = MT We take the parameters of the upper curve in the following.

E. G. Ferreiro USC


EMMI 28 Sept 2011

Results for d+Au: ϒ rapidity dependence Intrinsic vs extrinsic scheme

• Different shadowing effects in the 2 approaches • Antishadowing peak shifted toward larger y in the extrinsic case

E. G. Ferreiro USC


EMMI 28 Sept 2011

Other CNM effects: ϒ rapidity dependence in dAu @ RHIC

• Gluon EMC effect • Fractional energy loss

E. G. Ferreiro USC


EMMI 28 Sept 2011

Other CNM effects: ϒ rapidity dependence in dAu @ RHIC Extrinsic scheme: σabs=0 mb, σabs= 0.5mb , σabs= 1 mb in 3 shadowing models

• backward: ok within uncertainties • central: reasonable job • forward : clearly too high (for any σabs) Physical interpretation • backward: EMC effect • central: antishadowing • forward : shadowing≈1 energy loss is needed E. G. Ferreiro USC



Work in progress: EMC effect

EMC antishadowing

Let us try to increase the suppression of g(x) in the EMC region, keeping momentum conservation : ʃxg(x) dx = Cte

Works better for backward region

E. G. Ferreiro USC



Work in progress: Energy loss effect • Basic idea: An energetic parton traveling in a large nuclear medium undergoes multiple elastic scatterings, which induce gluon radiation => radiative energy loss (BDMPS)

• Intuitively: due to parton energy loss, a hard QCD process probes the incoming PDFs at higher x, where they are suppressed, leading to nuclear suppression

• The problem: This energy loss is subject to the LPM bound => Δ E is limited and does not scale with E (Brodsky‐Hoyer)

• At RHIC and LHC (contrary to SPS), typical partons (for x1 ~ 10‐2) have energies of the order of hundreds of GeV in the nucleus rest frame => radiative energy loss has a negligible effect on the parton x1

E. G. Ferreiro USC



Work in progress: Energy loss effect • Still, in order to explain large xF data at RHIC, it would be useful to have => a fractional energy loss: Δ E α E (Old idea by Gavin Milana, thought to be ruled out by LPM bound) • Recently (Arleo, Peigner, Sami arxiv:10006.0818) it has been probed that the notion of radiated energy associated to a hard process is more general than the notion of parton energy loss. The medium‐induced gluon radiation associated to large‐xF quarkonium hadroproduction: arises from large gluon formation times tf >> L scales as the incoming parton energy E cannot be identified with the usual energy loss qualitatively similar to Bethe‐Heitler energy loss the Brodsky‐Hoyer bound does not apply for large formation times Thus, the Gavin‐Milana assumption of an “energy loss” scaling as E turns out to be qualitatively valid for quarkonium production provided this “energy loss” is correctly interpreted as the radiated energy associated to the hard process, and not as the energy loss of independent incoming and outgoing color charges. • Note that space effect through Sudakov suppression can also induce a fractional energy loss but for x1 > 0.5 (Kopeliovich) ) E. G. Ferreiro USC



Work in progress: Energy loss effect When the longitudinal momentum pL >> mT

Due to tf of the order of nuclear size, this energy loss is not applicable in the backward rapidity regions. Note that, independently of the gluon PDF parameterization, this energy loss will induce a minimum suppression of 75% ‐ 80% up to a maximum one of 40% in the forward region

E. G. Ferreiro USC



Work in progress: ϒ centrality dependence Extrinsic scheme: σabs=0 mb, σabs= 0.5mb , σabs= 1 mb in 3 shadowing models EKS

EPS

nDSg

• in the mid region: antishadowing=>progressive increase of RdAu vs Ncoll • in the forward region: shadowing => progressive decrease of RdAu vs Ncoll

E. G. Ferreiro USC


EMMI 28 Sept 2011

Work in progress: ϒ transverse momentum dependence Extrinsic scheme: σabs=0 mb, σabs= 0.5mb , σabs= 1 mb in 3 shadowing models EKS

EPS

nDSg

Growth of RdAu not related to Cronin effect: it comes from the increase of x for increasing PT • in the forward region: x goes through the antishadowing r => enhancement in RdAu • In the backward region: x sits in an antishadowing and EMC => decrease in RdAu E. G. Ferreiro USC


EMMI 28 Sept 2011

CEM NLO inspired 2‐> 2 peacked at low pT (to be smeared out)

E. G. Ferreiro USC


EMMI 28 Sept 2011

E. G. Ferreiro USC


EMMI 28 Sept 2011

Conclusions • We have studied the influence of specific partonic kinematics within 2 schemes: intrinsic (2→ 1) and extrinsic (2→2) pT for different shadowings: EKS98, EPS08, nDSg including nuclear absorption and different partonic models • for J/ψ A+A collisions @ RHIC: RAA forward y RAA mid y as CNM in 2→2 but… RAA forward y @ LHC > RAA forward y @ RHIC Place for recombination effects, to be checked with ALICE data (pT>0) at mid y • for ϒ in d+Au collisions @ RHIC: EMC effect in the backward region fractional energy loss in the central & forward region http://phenix‐france.in2p3.fr/software/jin/index.html E. G. Ferreiro USC


BNL 6‐18 June 2011 21