High-pT Photon and Hadron Probes of Dense Matter

Created in Heavy Ion Collisions @ RHIC

Prof. Brian A. Cole.Columbia University

Outline 1. Introduction 2. Single high-pT photons, hadrons. 3. Azimuthal anisotropy. 4. Di-jet distortion. 5. Inclusive -jet correlations 6. “High”-pT photons (reprise) 7. Summary, conclusions

Thanks to:

D. Winter, J. Jin, S. Bathe, H. Buesching, T. Isobe, C. Klein-Boesing, N. Grau, P. Constantine, C. Gale, S. Turbide, M. Gyulassy, I. Vitev, A. Dainese, W. Horowitz, T. Dietel, M. van Leeuwen …

I must particularly acknowledge S. Mioduszewski and J. Jia with whom I work closely and whose insights have contributed significantly to this talk.

This Talk: Physics Overview•Single high-pT photon & hadron production

– How sensitive to medium properties?

– What does constant (?) RAA tell us?

•Azimuthal anisotropy– Origin at intermediate-high pT?

– Energy loss contributions?– Consistent geometric picture?

•Di-jet distortion at intermediate pT

– Real or experimental artifact?– Dip or not?– Dependence on angle wrt reaction plane?

•Inclusive -h correlations– First steps down the path suggested by Xin-Nian …

•“High”-pT photons: medium effects

Hard Scattering in p-p Collisions

•Factorization: separation of into– Short-distance physics: – Long-distance physics: ’s, D’s

p-p di-jet Event

STARSTAR

From Collins, Soper, Sterman Phys. Lett. B438:184-192, 1998

dt

d

z

QzDQxQxdxdx

dp

dE c

abcaBbaAaba

ˆ),,(

),,(),,(2

/2/

2/3

30

a/A

b/B

A

B

ab

D(z)

p-p Baseline @ RHIC (PHENIX)

KKP

Kretzer

data vs pQCD

Phys. Rev. Lett. 91, 241803 (2003)

p-p0+xp-pprompt + x

Hard Processes: Rate Calibration

•Plot total/ decay

•Use double ratio to reduce systematic errors.

•See clear “high” pT direct photon signal.

•Excess above 1 is direct photon signal.

•Curves = pQCD calculation.

PHENIX Direct : Phys. Rev. Lett. 94:232301, 2005

Hard Processes: Rate Calibration (2)

•In A+B collisions pQCD (factorization) implies:–

•RAA measures degree of violation of factorization

–

•High-pT prompt yields consistent with TAB scaling p-p Au-Au.

•In Au+Au, hard scattterings occur at the expected rate!

binary

NNhardNN

inelAB

NNhard

AB

NNhard

ABhard N

dp

dnbT

dp

dnbT

dp

d

dp

dn2222

)()(

binary

NNhard

ABhard

AB

NNhard

ABhard

AA Ndp

dn

dp

dnbT

dp

d

dp

dnR

2222)(

A simple picture – but there’s more going on than meets the eye.

Come back to this later …

PHENIX: Au-Au High-pT 0

Suppression

constancy for pT > 4 GeV/c for all centralities?

We are now measuring out to truly high pT

Flat RAA – Cautionary Note

•We are presumably seeing the EMC effect.– And its impact parameter dependence!

•As we were warned, flat RAA is mixture of effects.

0 Suppression: dE/dx Comparisons

•Measured RAA shows little/no variation with pT

•Consistent w/ energy loss calculations– Suppressed hard production over “whole” pT range?

GLV dNg/dy = 2700

STAR 200 GeV Cu-Cu RAA

•Crucial new data on jet-quenching systematics.•Note: Wang & PQM both agree well with data !(?)

Poster by M. van Leeuwen

PHENIX: Cu-Cu 0 RAAR

AA

Single Hadron Suppression Status•We have excellent agreement between models and data for both Au+Au AND Cu+Cu!– This is good, right? Wrong!

•There is a serious outstanding disagreement re: the opacity of the medium.

•More important, it’s unclear how well we can determine the properties of the medium.– Wiedemann, Salgado, Armesto + Dainese, Loizides,

Paic: single high-pT hadrons originate near surface.

– Vitev: GLV RAA still sensitive for dNg/dy > 1100.

•GLV: fluctuations in # emitted gluons important– Weakens “blackness” of the matter.

•How can such fundamentally different physics produce equally good description of data???

Understanding Single Hadron Suppression

•Both GLV and PQM claim Npart2/3 … but for

completely different reasons.•PQM: surface emission, GLV: L + thickness

Plot from PHENIX white paper testing GLV predicted Npart

2/3 dependence of effective mean energy loss.

PHENIX: High-pT 0 v2 (Reaction Plane)

•Clear observation of decreasing v2 @ high pT

From parallel session talk by D. Winter

V2(pT): Energy Loss Calculations (2)

•Many more (partial list)– Gyulassy, Vitev, Wang: Phys. Rev. Lett. 86:2537 (2001)– Drees, Feng, Jia: Phys. Rev. C71:034909 (2005)– Muller: Phys. Rev. C67:061901

(2003)– X.N. Wang: Phys. Lett. B595:165 (2004)

•All attempts to describe v2(pT) using E-loss w/ realistic geometry have failed @ intermediate pT

***

•Remember that the shown v2 is for pions

•We now observe energy loss dominated azimuthal anisotropy at high pT.– But, our statistics do not yet carry us deep into E-loss

dominated region …*** What about effects of jets crossing (or not) flow pattern?

V2(pT): Energy Loss Calculationsv

2PQM: Dainese, Loizides, Paic, Eur. Phys. J. C38:461 2005

V2(pT): Energy Loss Calculations (2)

Turbide et al, Phys. Rev. C72:014906, 2005 See plenary talk by C. Gale Monday morning.

•AMY - BDMS-like energy loss calculation – w/ HTL screening masses– Additional diagrams

beyond bremsstrahlung

•Calculation an extension of work in:

PHENIX Preliminary

0-20%

PHENIX Preliminary

20-40%

PHENIX Preliminary

40-60%

•Rather than separating v2, RAA(pT) plot RAA(, pT)

RAA(, pT)

Understanding RAA(Npart, )

•Can we find a consistent description of centrality and dependence of RAA?

• Define with

•Perform survival-weighted average over all production points, emission directions.

•Survival weighting:

ldlLeff )( 0

0)(

21 neffsurv Lkp• Energy-loss motivated Leff consistent

description of centrality and dependence of 0 RAA above 7 GeV/c.

At low pT the same prescription badly fails.

Leff

Something other than radiative energy loss is affecting azimuthal anisotropy at lower pT.

Leff

Lessons from RAA(, pT)•Constancy of RAA below 7 GeV/c not “intrinsic”. – Some additional physics

varying w/ pT.

•That physics must be require spatial /flow anisotropy.– Not there/much weaker

in central collisions.

•“bump” below 3 GeV/c in all centrality bins ?!– Recombination?

•Extra yield in plane ??

Collisional Energy Loss a la Molnar

•Denes: many soft scatterings can push partons to high pT

– pushed partons have larger v2 than quenched partons

– And a stronger pT dependence to v2.

•MPC under-predicts v2 magnitude

– But gets a v2 that varies strongly with pT!

– Too low a value for RAA ????•Is possible to combine w/ “traditional” energy loss approaches as parton pT gets high?

Data 30-40%Calculation 8 fm

Data 20-30%Calculation 8 fm

MPC: Implications•Suppose the medium does accelerate low-pT hadrons to high pT.

•Pushed partons would naturally reduce di-jet correlation strength!

•In a (“trigger”) pT – independent way:– As hadron (jet) energy , push fraction – So Denes should predict the IAA that would result.

•Pushed partons could explain the pT dependence of RAA in the reaction plane.– But not in central collisions!!!

•Is the spatial anisotropy an important ingredient?– i.e. “Pushed” partons escape the easy way ???

•Imagine telling your high-energy colleagues that our matter can self-generate > 6 GeV/c jets …

PHENIX: 200 GeV Au+Au h-h Correlations (2)

•With increasing centrality see more distortion of the di-jet signal.– Development of a

“shoulder”– And a dip.

•Yes, the shape is sensitive to systematic errors on v2.

•Yes, the signal is small.•But the “shoulder” is absolutely, 100%, unequivocally robust

•Not di-jet broadening– Try to prove me wrong ….

Di-jet Distortion: STAR

•Does STAR see the dip? Certainly not ruled out …

pT trigger 2.5-4 GeV/c, pT associated 1-2.5 GeV/c, || < 0.35

0-5% 10-20%

5-10% 20-30%

Parallel talk by J. Ulery, Monday.

STAR Preliminary

PHENIX: Reaction Plane Angle Dependence

•Study (di)jet correlations vs angle of trigger hadron relative to reaction plane– J. Bielcikova et al,

Phys. Rev. C69:021901, 2004

trig = trig - –6 bins from 0 to /2.

•Flow systematics change completely vs trig

•Can study dependence of distortion on geometry.

Shoulder and dip seen in all trig bins.

trig

?

From Poster by J. Jia

PHENIX: Reaction Plane Angle Dependence(2)

•For PHENIX reaction plane resolution & chosen bin sizes, trig bin 4 has smallest flow effects.

•Even without subtracting flow contribution, a dip is seen for central collisions.

Look in bin #4

PHENIX Preliminary

Photon - Jet(Hadron) Measurements(?)

• Study of di-jet correlations affected by energy loss of both jets.

•Photon-jet “cleaner” because on parton escapes– No geometrical bias.

Photon-hadron more practical for now.

Old idea (Wang et al, Phys. Rev. Lett. 77:231-234, 1996)

Use photon-jet pairs to study medium-induced energy loss under better controlled conditions

STAR (inclusive) -h Correlations

•STAR sees the effect of reduced jet strength due to contribution of direct photons in inclusive -h.

•Consistent w/ direct/total photon rates + 0 suppression.

•1st attempt at inclusive -h di-jet per-trigger yields.– Compared to various assumptions re: di-jet / direct -jet ratios.

From parallel session talk by T. Dietel

PHENIX Inclusive -h Correlations

•“Trigger” photons w/ pT > 5 GeV/c

•Compared to 0-h correlations w/ 0 pT > 5 GeV/c.– Reduced “jet” yields in more central collisions– Consistent w/ larger (relative) prompt contributions

•Reminder: ~30% of prompt ’s from fragmentation.

From poster by J. Jin

Jet Conversion Photons

•There is a new source of “hard” photons in QGP– High pT quarks/gluons convert into photons in

medium

•This extra contribution must be present – @ large enough t, incident jet sees unscreened partons

•What about at low-t ?– In principle, pole in the t channel produces “large”

•But medium screens @ low-t & regulates pole.– Jet-conversion rate sensitive to screening mass.– And potentially also to quark/gluon thermal masses.

Jet Quenching: Photon Bremstrahlung

•For light quarks (and gluons??), in-medium energy loss dominated by radiation.– Interference between vacuum & induced radiation.

– For large parton pT (> ~10 GeV/c) coherence crucial.

•Unfortunately, we can’t measure the gluons.•But we could measure photon bremstrahlung!

Direct measurement of medium properties.

From S. Turbide et al

•Inclusion of extra QGP contributions improves agreement with data ??

•Too soon for conclusion – but real motivation!

Summary / Conclusions

•Incredible range of measurements of single hadron suppression & photon non-suppression in A-A collisions @ RHIC.– Nonetheless, we have a crucial outstanding issue:

Just how opaque is the medium??–Hopefully charm E-loss, di-jets, -jet will help …

Summary / Conclusions (2)

• We now have valuable information on azimuthal anisotropy at intermediate & high pT.

See expected radiative energy loss v2 for pT > 7.

At such high pT, centrality and dependence of RAA fit into consistent E-loss pattern.

Something extra is needed for pT < 7 GeV/c.

–But only in collisions w/ spatial anisotropy.

RA

A

pT (GeV/c)

Summary / Conclusions (3)

•What is it? Mach cone, Cherenkov cone, jet deflection, …. We don’t know yet.

•The “shoulder” is absolutely robust.

•PHENIX reaction plane study control of v2 systematic errors.

•“It” doesn’t/only weakly depends on trig

STAR 5-10%

PHENIX 0-5%

Summary / Conclusions (4)

•We have taken first steps on the path …– These measurements will further revolutionize our

understanding of parton energy loss …– Will provide insight on anisotropy, di-jet distortion, …

•Comment: Fragmentation ’s are not just background. They may (will) include in-medium radiation from jet.

The Big Picture

•Are we probing properties of the medium with high-pT hadrons / photons ?

•Yes, in more ways than I ever thought possible.

trigg

assoc

D

Compare with charged hadrons

pT Dependence of Di-jet Distortion

It appears that with increasing pT, the jet fills the dip.

RAA(,pT) vs. Npart

In-plane

Out-of-plane

Grey Bands: Inclusive RAA w/ Error

PHENIX: 200 GeV Au+Au h-h Correlations

“Flow” contribution w/ v2 systematic error

bands

Reaction Plane Biases?

•Can hard scattering bias the reaction plane measurement ?

•Evaluate using Pythia:– Calculate between pions in – And charged particles in

– For different pion pT bins (2 GeV/c, 4 GeV/c, 10 GeV/c).

35.043

dn/d

d

3<

<

4

Au-Au dn/d*2v2

10-20% 133

20-30% 122

30-40% 91.6

40-50% 5.81.3

Much larger than hard scattering correlation.

Hard Photon Production in pQCD•@ LO in pQCD, photon production is simple.•Two contributions:

– “partonic” photons: direct from hard scattering– “Fragmentation” photons – from fragmentation of

jet(s)

•But, @ NLO things are much more complicated– Distinction between partonic & fragmentation

contributions becomes ambiguous.

– In principle, “isolation” cuts possible – but matching those cuts with pQCD is difficult (virtual radiation).

A different NLO pQCD CalculationINCNLO (v1.4): J.Ph. Guillet, M. Werlen et al•NLO pQCD calculation

•With calculated frag. function.– Compares well with

HERA, CDF, D0 data

•Fragmentation ~ 25% at high pT

•Large increase in fragmentation yield below 5 GeV.– Mostly from NLL

contributions to FF– Right where we want

to measure “thermal”

What about Frag. Contribution?

•p-p analyzed with and w/o an isolation cut. – By eye: not unreasonable.– data less fragmentation, but too soon to conclude.

PHENIX Comparison to INCNLO

•No K factors, no fudge factors, absolute comp.•Completely independent calculation.

– Good control over pQCD prompt photon calculation @ RHIC.

Points: PHENIXCurve: INCNLO

Jet Studies: Hadron-Hadron Correlations

dN

/d

0º 180º

(di)jet Situation is Complicated

•Recent preliminary result from PHENIX showing strong distortion of opposite-side jet.– STAR sees similar but strongly pT dependent.

•Problem: – gluon radiation strongly couples to the medium

PHENIX Preliminary4 < pT

trig < 6 GeV/c 2.5 < pTtrig < 3 GeV/c

0.3<pTassoc<0.8 GeV/c

0.8<pTassoc<1.3 GeV/c

1.3<pTassoc<1.8 GeV/c

STAR Preliminary

Put it all together …

•Extremely rich mixture of physics contributing to the photon spectrum in ~ 4-10 GeV/c range.

•How to unravel all of the different pieces?

Shamelessly stolen from Simon’s talk.

How to Measure Frag. & Brem. ? (2)

•But, we can measure prompt photons produced in associate with hadrons of given pT.

– e.g. for hadron pT > 3 GeV, with kT<1 GeV

– confusion from IS radiation, jet intermingling, inter-jet radiation strongly reduced•Need statistical

subtraction of decay photons. 0 “tagging” will help

significantly. – In p-p can obtain ~

60% 0 rejection.

•Pythia study: – E > 4 GeV– pT hadron > 3 GeV/c.

dN

/d

(arb

. N

orm

.) -hadron pairs

Bremsstrahlung in Heavy Ion Collisions

•Bremsstrahlung contribution only!

•Potential increase in bremsstrahlung yield in medium.

•More important:– Energy & kT spectrum

will directly reflect medium properties.

•In my opinion: “Holy Grail” of energy-loss physics– Can “see” the radiation itself.– Photon bremsstrahlung calculation is much less model

dependent than the gluon radiation calculations.– Will not be easy to measure but it’s worth trying …

No quenching

w/ quenching

Zakharov (hep-ph/0405101)

How to Measure Frag. & Brem. ?

•Remember: separation of prompt spectrum into direct, fragmentation contributions is “arbitrary”.

•This may be particularly an issue @ RHIC:– photon/jet pT scales are < x10 pT scale of IS

radiation.– And jet cones are broad.

Jet 2

Jet 1 Acoplanarity exaggerated – but inter-jet radiation, hard radiation from one jet not shown.

Bremsstrahlung Measurements•Real opportunity for qualitatively new insight on the physics of in-medium parton scattering.

•Let’s be clear – this measurements won’t be easy.– In worst case, need to dig out bremsstrahlung out from

under x10 larger decay signal. Jet quenching no longer helps when you require the

photon to be in a jet !!– But, if the bremsstrahlung is enhanced, angular

distribution is broadened, then life is better.

•Observation by Axel: “trigger bias effect”– Will be an issue.– But potentially controllable by using opposite-side

“jet”/high-pT hadron requirement.

– Guidance from complete in-medium interaction calculation like AMY would be a big help.

How to Measure Jet Conversion ?•Brute force:

– Measure p-p accurately enough to provide a baseline with ~ 10% accuracy.

– Measure the Au-Au, Cu-Cu yield vs pT well enough to see >30% effects.

•Might work if the jet-conversion yield is as large as has been predicted.

•But, Cronin ???•Bremsstrahlung? (measure it – but enough of total yield ??)

One Hope: Reaction Plane Dependence

•Both Bremsstrahlung and jet-conversion photons could contribute to prompt photon v2.

– Observation of prompt photon v2 one or both mechanisms are present (high priority!)

– If we observe prompt photon v2, then we need to unravel the contributions.

•Jet conversion will produce more photons out-of-plane than in-plane

Negative v2 for these “jet quenching” photons

Radiative Effects on (di)Jets

• Large radiative component to di-jet acoplanarity– Also see Vitev, Qiu : Phys.Lett.B570:161-170,2003.

• Radiative effects are so large that we may have to re-think p-p and d-Au analysis– Cannot subtract off “constant background”

Analysis of STAR di-hadron distribution by Boer & Vogelsang,

Phys. Rev. D69 094025, 2004

PHENIX d-Au/p-p, - h, Correlations

•“Trigger” pion pT > 5 GeV/c

•Four different associated hadron pT bins

•Clearly see role of constant jT, contribution from kT

0.4-1 GeV/c1-2 GeV/c

2-3 GeV/c 3-5 GeV/c

PHENIX preliminary

p-pd-Au

Jet Properties in d-Au

•Compare pout dist’s in p-p and d-Au.

•Evidence for effects of re-scattering, modified radiation, … ?– Not so far!– But this is just the

beginning!

•Such measurements w/ one jet @ > 2 would be very interesting!!– But not possible yet

Alternative Method for Studying (di)jets

•By measuring pout pair-by-pair, more directly see the shape of the jT/kT dist’s.

•See non-Gaussian tails – expected due to hard radiation.

pp

Radiative tails

PHENIX, From J. Jia, DNP’04 Talk

PHENIX Preliminary

JetPout

Pout

PHENIX d-Au Production

•PHENIX sees small Cronin effect– Approx. consistent within errors with STAR Ks result – Enhancement seen in charged (baryons) all the more striking!

PHENIX 200 GeV Au-Au Production

•Observe high-pT suppression in Au-Au collisions.– Consistent with 0

0 / ratio consistent w/ vacuum fragmentation.