probing the low-x structure of the nucleus with the phenix detector

1

Probing the Low-x Structure of the Nucleus with the PHENIX Detector

Mickey Chiu

INT, Seattle, 20 October 2011

2

PRL 107.172301

PRL 107.142301

PLB 679 (2009) 321-329

1

2

3

1. Di-hadron correlations in d+Au2. J/ Production in d+Au3. UPC (diffractive) J/ in Au+Au

Low-x nucleon/nuclear structure is a very difficult business! We’ll want to test it with as many probes as we can.

3

Forward di-Hadron Production in d+Au

44

PHENIX Muon Piston Calorimeter

Small cylindrical holes in Muon Magnet Pistons, Radius 22.5 cm and Depth 43.1 cm

SOUTH

PbWO4

North

)( 43 yyTd ee

s

px )( 43 yyT

Au ees

px

•Fwd-Fwd, x~(0.001,0.005)•Mid-Fwd, x~(0.008,0.040)•Mid-Bwd, x~(0.050,0.100)

d(forward) Au(backward)

5

MPC Performance

North MPC

Decay photon impact positions for low and high energy 0s. The decay photons from high energy 0s merge into a single cluster

Sometimes use (EM) clusters, but always corrected to 0 energy

Clusters 80% 0 (PYTHIA)

“Trigger”Near

Far

Jet1

Jet2

6

•Large suppression in RdA•That increases with centrality•And increases with larger rapidity

•Consistent with previous measurements•However, x covered by single inclusive measurement is over wide range

•Includes shadowing, anti-shadowing, (EMC effect)

RdAu in 2 forward rapidity Bins

Guzey, Strikman, Vogelsang, PLB603, 173

Guzey, Strikman, Vogelsang, PL B603, 173

7

CNM effects: dynamical shadowing, Energy Loss, Cronin

RdA Past, di-Hadron Future

Kharzeev, NPA 748, 727 (2005)

•Di-Hadron Correlations allow one to select out the di-jet from the underlying event•Constrains x range (probe one region at a time)•Probe predicted angular decorrelation of di-jets (width broadening)

Kharzeev, Levin, McLerran Nucl. Phys. A748 (2005) 627

Color Glass Condensate

(Qiu, Vitev PLB632:507,2006)

8

di-Hadron Signal

Peripheral d+Au Correlation Function

d

dN

NN

NCY

assoc

trigassoctrig

pair 1

CORRELATED

Npair

“Di-Hadron Nuclear Modification factor”

trigdA

trigdAdA RIJ

• Possible indicators of nuclear effects

• JdA < 1, RdA < 1

• Angular decorrelation of widths

pppairpp

dApairdA

colldA N

J

/

/1

ppsglpp

dAsgldA

colldA N

R

/

/1

“Sgl-Hadron Nuclear Modification factor”

“Conditional Yield”

Number of di-jet particle pairs per trigger particle after corrections for efficiencies, combinatoric background, and subtracting off pedestal

Caveats:

1. Low pT (but back-to-back peak is selected)

2. Pedestal Determination (Assumed up to twice the width as a systematic).

3. Di-Hadrons instead of di-jets (but ok if fragmentation unmodified)

9

0 (trigger,central)/0 (associate,forward)

<pTa>=0.55 GeV/c

p+p

Cor

rela

tion

Fu

nct

ion

d+Au 0-20%

d+Au 60-88%

<pTa>=0.77 GeV/c <pT

a>=1.00 GeV/c3.0 < pTt < 5.0 GeV/c

for all plots

pTt, 0

pTa, 0 PHENIX Preliminary

10

Correlation Widths, d+Au and p+p

• No significant broadening between p+p and d+Au within large experimental uncertainties

Trigger p0: |h| < 0.35, 2.0 < pT < 3.0 GeV Trigger p0: |h| < 0.35, 3.0 < pT < 5.0 GeV

dAu 0-20%

ppdAu 40-88%

10

•Widths are consistent between p+p and d+Au (all centralities) within large statistical and systematic errors•No broadening seen (within errors)

11

JdA vs Ncoll, pTmid, pT

fwd

MPC 0 pT

•Suppression of di-hadron correlation (relative to p+p binary scaling hypothesis) with

Increasing Centrality

Decreasing pTmid

Decreasing pTfwd

pTt, 0

pTa, 0

12

Fwd-Fwd: p+p vs d+Au Peripheral

Peripheral d+Au collisions are similar to p+p collisions

Beam view of d+Au peripheral collisionpT

t, 0

pTa, 0

13

Fwd-Fwd: p+p vs d+Au Central

“Monojet” in central d+Au collisions

Beam view of d+Au peripheral collisionpT

t, 0

pTa, 0

14

JdA for Fwd-Fwd

fragAuTTfrag

Au zxs

epepx

21

21

MPC 0 pT

(assuming LO)•Better way to plot:

•Suppression of JdA gets larger in fwd-fwd correlations•Trend with pT, centrality also consistent with mid-fwd correlations

15

xAufrag Dependence

s

epepx TTfrag

Au

21

21

Note: points for mid-fwd JdA are offset for visual clarityStatistical and systematic errors are added in quadrature

•Plotting vs suggests that the effect is due to something happening in the nucleus as one probes to lower x•Does it prove CGC?

•Shadowing? Initial state energy loss? Multi-Parton Interactions (MPI)?

60-88%(Peripheral)

0-20%(Central)

16

d

cdab

pbpp

ap

d

cdab

Aub

Auda

d

pppairsppcoll

dAupairsdAu

dA

zzDxfxf

zzDxfxf

NJ

c

c

,)()(

,)()(

/

/

Extending the LO picture

b=0-100%Q2 = 4 GeV2

xAu

EPS09 NLO gluonsEskola , Paukkunen, Salgado, JHP04 (2009)065

),(

),(),(

2

22

QxAxG

QxxGQxR

p

AuAuG

Low x, mostly gluons JdA ~ RGAuHigh x, mostly quarks

Weak effects expected

17

Where is the Saturation Scale if we are actually seeing the CGC?

Iancu and Venugopalan, hep-ph/0303204

Extended scaling?

H. Kowalski and D. Teaney. Phys. Rev.D, 68:114005, 2003

•We evaluated in PYTHIA the ~ coverage for Q2 and x for the fwd-fwd and cnt-fwd correlations

•No nuclear modifications evaluated yet•Not clear that we are in the saturation region – possibly in extended region?

•Can we explore Qs from the data?•Nuclear Scaling: Look at impact parameter dependence by varying centrality

Fwd-Cnt?

Fwd-Fwd?

18

d+Au MC Glauber

d Au

bnucleon

bnucleon

•From Glauber Monte Carlo we can determine the number of nucleons in the path of each nucleon in the deuteron

Centrality0-20%

20-40%40-60%60-88%

19

JdA Centrality Dependence

•Fit using EPS09 parametric function:•Evaluate JdA at xfrag = 6x10-4, 6x10-3, 1.5x10-2

])[()( 210axxA

g eexaaaxR

20

Can we determine Qs?

)(~

~

xG

RAu

Aug

L~

0

3/122

2 ~),(1

x

xAQxxG

RQ sS

•If we are measuring gluons w/ JdA, then we can perhaps extract length and x dep of Qs, as well as possibly extracting the value of Qs at RHIC????•Eg, are we seeing an approx linear dependence on length????

xfrag ~ 1.5x10-2

xfrag ~ 6x10-3

xfrag ~ 6x10-4

21

J/ Production in d+Au

22

Traditional shadowing from fits to DIS or from coherence models

high xlow x

D

Dccmoversco-

Absorption (or dissociation) of into two D mesons by nucleus or co-movers

cc

Energy loss of incident gluon shifts effective xF and produces nuclear suppression which increases with xF

R(A/p)

R=1 xF

Gluon saturation from non-linear gluon interactions for the high density at small x - Amplified in a nucleus.

pA

What are the CNM effects that are so strong in Quarkonia production?

shadowing

anti-shadowing

RG in A

u

23


Reasonable agreement with EPS09 nPDF + br=4 mb for central collisions but not peripheral

EPS09 with linear thickness dependence fails to describe centrality dependence of forward rapidity region.

J/ψ in d+Au – learning about CNM thickness dependence

PHENIXarXiv:1010.1246v1

24

Quarkonia Suppression in A+A Collisions – key observations and questions

Overall suppression of J/ψ is very similar between:• SPS (17.2 GeV) • RHIC (200,62,39 GeV)• and LHC (2.76 TeV)

62 GeV 39 GeV

SPS

PHENIX forward

PHENIX y=0

CMS:0 <|y|< 2.4pT > 6.5

(more on LHC in a minute)Npart

Npart

NpartNpart

RA

A

RA

A

2525

CMS pT > 6.5 GeV/c

• LHC suppressed more than RHIC at y~0(but CMS is pT> 6.5 GeV/c)• LHC suppressed less than RHIC at forward y(here ALICE is pT> 0)Features expected from regeneration, which is concentrated at small pT

High-pT suppressed more than low pT

(but ALICE y~3; ATLAS y~0)

However suppression roughly flat with rapidity for pT>6.5

So may also be consistent with regeneration at small pT

Missing LHC data – y~0, pT> 0 RAA?(where regeneration may be rather large)

Quarkonia Suppression in A+A Collisions – comparing RHIC & LHC

Mid Rapidity

Forward Rapidity

all pT

Npart

y R

cp

CMS

ALICE

ATLAS

caution caution

ALICE, all pT

26

Vary the strength of suppression (a) & see what relationship between RdAu and RCP is given strictly by Glauber geometry for different dependences on density-weighted thickness

0

1( ) ( , )T Tr dz z r

Woods-

Saxon

• Break-up has exponential dependence• EPS09 & initial-state dE/dx have unknown dependences


J/ψ in d+Au – learning about CNM thickness dependence

The forward rapidity points suggests a quadratic or higher geometrical dependence

)()( :lExponentia TraT erM

)(1)( :Linear TT rarM 2)(1)( :Quadratic TT rarM PHENIX

arXiv:1010.1246v1

27

Rapidity or x Coverage

28

Does di-hadron data match J/Psi?

•Comparison not so bad, considering many other uncertainties (production model, energy loss, breakup cross-section). Also J/ is generally at higher Q2

•Real way to do this is to try to extract G(x) from di-hadron data, and then predict J/

29

Ultraperipheral J/

30

Hadronic Interaction:Au-Au --> X ~7 barns

-:AuAu --> AuAu + e+e- ~33 kbarnsAuAu --> AuAu + 2(e+e-) ~680 barnsAuAu --> AuAu + 3(e+e-) ~50 barns

-N: L(-N )=1029 cm-2s-1 2<E<300GeVAuAu --> Au+Au* 92 barns X+neutronsAuAu --> Au*+Au* 3.670.26 barns X+neutrons Y+neutrons

“Hadronic” Collider Processes•You’re probably familiar with the “Hadronic Interactions”•But there are a lot more processes going on at a hadron collider

Hadronic Interaction

Ultra-Peripheral Interaction

31

Au J/ Au* measurement in PHENIX

e+

e

• UPC dedicated trigger– Rapidity gap 3<||<4

MB interaction veto (BBC veto)– Large probability to exchange

additional photons by GDR 1 or 2 ZDC trigger

– EmCal trigger (E>0.8GeV)– DiMuon Trigger

Au J/ ( l+l-) Au*– DC & PC tracking detectors– RICH & EmCal electron identification

devices– Muon Tracker

n

l+

l-

||<0.35

32

J/ cross section vs theoretical calculations• Compatible with coherent

predictions, • With more statistics,

sensitive to the shadowing parameterizations,

[ 1) P.R.L.89 012301 (2002)…][ 2) P.L.B626 (2005) 72 ]

[ 3) arXiv0706.2810 [hep-ph] ][ 4) arXiv:0706.1532 [hep-ph] ]

[ 1) ][ 2) ]

[ 3) ][ 4) ]

d/dy |y=0 = 76 31 (stat) 15 (syst) b

[Filho et al, PRC78 044904 (2008)]

coherent

incoherent

coherent

33

Impact Parameter DependenceHorowitz, INT-PUB-11-005 (arxiv:1102.5058) UPC J/ψ pT

2

distribution(Theoretical)

Coherent(γAu): low pt peakIncoherent(γn): wider pt distribution(Incoherent + neutron tagged :Yellow shadow )

Strikman et al, PLB 626 p. 72-79

EIC Workshop, INT, Seattle 2010

•Fourier transform of t distribution can distinguish the density of the nucleus vs b•However, incoherent contribution is a potentially large source of background

34

PT dependence of UPC J/ψ+Xn(N)+Xn(S)

•UPC J/ pT (~t1/2) also confirms existence of incoherent contribution•Strategy: measure at forward rapidities to get incoherent, subtract from total to get remainder

•Major challenges: momentum resolution of 3 MeV! (technical driver for EIC detector)•Statistics (EIC is good, RHIC/LHC is poor)

35

+, +p, +A “Applications”•Higgs as well as many SUSY ptcls should be produced at the LHC in + and +p (+A)

•High energy photon interactions at the LHC, de Jeneret et al, arXiv:0908.2020•FP220, FP420

•Observation of exclusive charmonium production and gamma+gamma to mu+mu- in p+pbar collisions at sqrt{s} = 1.96 TeV, CDF, PRL102:242001,2009

x10-3 10-2 10-1 10-3 10-2 10-1 10-3 10-2 10-1

EPS09: A New Generation of NLO and LO Nuclear PDF’s, Eskola,Paukannen,Salgado JHEP 0904:065 2009

•Direct measurement of G(x) at from photoproduction (~g2(x))222/ 105.1/ xWmx AJ

•Possibly study dynamics of J/ propagation through nuclear matter•Feature or Bug?

•Test of QED in strong-coupling regime?: =ZEM~0.6

Nucleus-Nucleus Interactions

36




x10-3 10-2 10-1 10-3 10-2 10-1 10-3 10-2 10-1



PHENIX UPC J/Psi




37




x10-3 10-2 10-1 10-3 10-2 10-1 10-3 10-2 10-1



PHENIX UPC J/Psi




38

Summary•Three Tests of Saturation in PHENIX (or probes of g(x))•FWD-FWD di-hadron yields in d+Au relative to p+p (JdA)

•Suppression depends strongly on centrality•And gets stronger as both particles go toward more forward rapidities•Nuclear Shadowing? We see extreme Shadowing in most central.•Gluon Saturation/Color Glass Condensate?

•If so, we can extract a wealth of information on Qs from our measurements

•Initial State Energy Loss? MPI?•Angular Broadening of Away Side Jet?

•Mid-Fwd, no increase seen within errors•Mid-MidFwd, also no increase •Fwd-Fwd, currently inconclusive

•J/ Production in d+Au not well understood•Forward rapidities not well described – something extra going on?•Highly important to understand CNM effects for HI interpretation

•Ultraperipheral J/ is a third, very different probe of gluon distribution•Can get a ~10% measurement of G(x) at x~10-2

•Statistics and detector challenged for G(x,b) impact parameter dep measurement

Are these enough to see the elephant in the room?

39

Backup Slides

40

IdA vs JdA: Can we decouple effects?

d

dN

NN

d

dN

N

N

N

N

NN

d

dN

N

d

dN

NI pp

trigassocpp

evtcoll

dAtrigassoc

dAevt

dAevt

dAtrig

ppevt

pptrigcoll

ppassoc

pptrig

dAassoc

dAtrig

dA)|(

)|(

1

1

1

1

dAtrigdA

trigdA J

RI

1

•IdA is the per trigger comparison of d+Au jet associated counts relative to p+p•JdA is the rate of the associated pairs from a jet (per minbias event)•Can we use this to tell if the jets are modified, or do they disappear?•From the CNT-MPC corrrelations, we get IdA ~ 0.5, and RdA ~ 1.1

•JdA ~ 0.5•The rate of correlated pairs is about half of p+p

•Does this imply that the missing jets have disappeared, and not that they are modified, since IdA ~ JdA?•But not true for STAR FMS triggered-central barrel, where IdA ~ 1 and JdA ~ 0.5

41

JdA, RdA vs Ncoll

MPC 0 pT

Qiu-Vitev Shadowing + Energy Loss

(private communication)

42

Muon-Central IdA & Widths, 2003 d+Au

Phys.Rev.Lett.96:222301,2006

dAu

43

Nuclear Modification in d+Au at Forward(Backward) Rapidity

Phys. Rev. Lett. 94, 082302 (2005)

Punch through hadrons & Hadron decay muons •Forward η suppression •No backward η suppression

•Gluon Saturation?•Cronin, Shadowing, E-loss?

•Look at 2 particle correlations …

43

44

• Rough comparison with HERA e-p data, if coherent incoherent ratio is 50% - 50%

• HERA (H1 & ZEUS) input

• Result: coh = 1.01 0.07

incoh = 0.92 0.08

~ 1, good agreement with HERA data hard probes scaling

UPC Comparison with HERA data

[ZEUS, Eur.Phys.J. C24 (2002) 345]

[H1, Eur.Phys.J. C46 (2006) 585]

probing the low-x structure of the nucleus with the phenix detector

Documents

p p vs d au centralmonojet

fwdfwd correlationstrend

midfwd jda

dihadron correlations

midfwd correlations

p pno significant broadening

p0 trigger

hadron futurekharzeev