The Thermal Photon PuzzleExperimental Issues

Introduction

Thermal PhotonsHigh photon yieldCentrality dependenceAngular anisotropy

Theoretical ModelsThermal Photon PuzzleB-fields and flowing glasma

Summary and Outlook

Axel Drees, April 20th, RIKEN-BNL Research Center

Black Body Radiation Real or virtual photonsSpectrum and yield sensitive to temperature

Avg. inv. slope T, Yield T3

Space-time evolution of matter collective motion Doppler

shift anisotropy

Microscopic view of thermal radiationQGP: hadron gas:

photons low mass lepton pairs

Thermal Radiation from Hot & Dense Matter

g

p

r

p

q

qg

g

Axel Drees2

g

gg*

e+e-

Need realistic model simulation for rates and space-time evolution for quantitative comparison with data

High yield high T early emissionLarge Doppler shift late emission

Experimental Issue: Isolate Thermal Radiation

Axel Drees3

1 10 107 log t (fm/c)

, * from A+A

Direct

Hadron Decays“Prompt” hard scattering

Pre-equilibrium

Quark-Gluon Plasma

Hadron Gas

ThermalNon-thermal Need to subtract decay and prompt

contributions

g

g

PC1

DC

magnetic field &tracking detectors

e-

e+

Axel Drees

Thermal Radiation at RHIC Energies: PHENIX

Photons, neutral pion p0 g g

Calorimeter

e+e- identification

4

E/p and RICH

Photons

HBD

g

Searches for thermal photons ongoing since late 1980’s at SPSWA80 & successors, HELIOS, CERES … Established mostly upper limits

Breack through at RHIC: Measuring direct photons via virtual photons

Method originally proposed by UA1 for prompt photons

Using virtual photons:any process that radiates g will also radiate * gfor m<<pT *g is “almost real”extrapolate * g e+e- yield to m = 0 direct g yield m > mp removes 90% of hadron decay backgroundS/B improves by factor 10: 10% direct g 100% direct g*

Using * g to Measure Direct Photons

Axel Drees

Fit e+e- Mass Distribution to Extract the Direct Yield:

Example: one pT bin for Au+Au collisions

and

normalized to da

( )

ta

(

f

)

or 30

dir eec

e

e

e

ef

m

m

V

f m

Me

Direct * yield fitted in range 120 to 300 MeVInsensitive to 0 yield

dydp

dML

MdydpdM

d

TT

ee2222

)(1

3

1/m

6

access above cocktail

fraction or direct photons:

dir dir

incl incl

r

Direct Photons from Virtual Photons

Axel Drees7

Prompt photon consistent with pQCD in p+pSignificant additional yield in Au+Au

PHENIX Phys.Rev.C87 (2013) 054907 PHENIX Phys.Rev.Lett 104 (2010) 132301

2006 data

Axel Drees

pQCD

g* (e+e-) m=0 g

T ~ 220 MeV

First Measurement of Thermal Radiation

Direct photons from real photons:Measure inclusive photonsSubtract p0 and h decay photons at S/B < 1:10 for pT<3 GeV

Direct photons from virtual photons:Measure e+e- pairs at mp < m << pT

Subtract h decays at S/B ~ 1:1 Extrapolate to mass 0

8

First thermal photon measurement: Tini > 220 MeV > TC

large photon yield!

Direct Photons from Photon ConversionsDouble ratio tagging method

Clean photon sample with photon conversion Explicit cancelation of systematic uncertainties

Axel Drees9

Almost 20% direct photons! Approx. independent of pT

Extended range to 400 MeV

measured raw yields

PHENIX arXiv:1405.3940

MChadr

Data

tag

incl

hadr

incl

N

N

Y

Yf

N

NR

0

0

simulated basedOn hadron data

measured raw yieldsconditional

tagging efficiency fN

N

fapN

apN

Y

Ytag

incl

eeeeconvtag

eeeeconvincl

hadr

incl

00

“pQCD-fit”

Teff ~ 220 MeV

g* (m0)

g

g e+e-

Direct Photons Au+Au Collisions

10 Axel Drees

PhD thesis Bannier & Petti,

PHENIX arXiv:1405.3940

Centrality Dependence of Thermal Component

No significant change in shapeInverse slope ~ const.Fit range 0.6-2.0 GeV:

Rapidly increasing yield with centrality like Npart

Data = 1.48±0.08±0.04 ~ 3/2

Faster than volume ~ Npart

Faster than prompt componentNcoll ~ Npart

4/3

Slower than naïve expectationYield ~ Npart

2

Axel Drees11

Collective Behavior: Elliptic Flow

initial state of non-central Au+Au collisionspatial asymmetry asymmetric pressure gradientsmomentum anisotropy in final stateexpressed as elliptic flow “v2”

Axel Drees

x

z

Non-central Collisions

in-plane

out-of-plane

y

Au nucleus

Au nucleus

3 3

R30T T

2 cosnn

d N d NE v nd p p d dp dy

Thermal radiation emitted from moving matter Doppler Shift

Anisotropy

blue shift

red shift

Thermal Photon Show Large Anisotropy

Axel Drees13

R (pT) ~ constant > 1

At 2 GeV v2inc ~ v2

dec

Large v2~ 0.2 at 2 GeV/cInsensitive to sys. uncertainty

IF there are direct photons

v2dir ~ v2

inc

PHENIX Phys.Rev.Lett 109 (2012) 122302

Thermal Photon Anisotropy Update

Axel Drees14

Two new independent analysisCalorimeter measurementphoton conversions →e+e

for v3 discussion see T.Sakaguchi’s talk

Consistent with published resultsLarge v2~ 0.2 at 2 GeV/c

Indication for const. v2 at low pT

Theory Comparison: Thermal Photon Puzzle (I)

Reasonable agreement with:Tini = 300 to 600 MeV

Shape similar, butyield is underestimated

Transport model: Linnyk, Cassing, Bratkovskaya, PRC89 (2014) 0034908

Fireball model: van Hees, Gale, Rapp, PRC84 (2011) 054906

Hydrodynamic model: Shen, Heinz, Paquet, Gale, PRC89 (2014) 044910

Theoretical Models Underestimate Yield

About factor of 2 at high pt – with large errors

Factor 5-10 at lower pt (central collisions)

Axel Drees16

Theory Comparison: Thermal Photon Puzzle (II)

Axel Drees17

Models fail to describe simultaneously photon yield, T and v2!

Large B-field Enhances Thermal Radiation

Large magnetic fieldEnhanced thermal emissionAnisotropy with respect to reaction plan

Axel Drees18

x

z

in-plane

MagneticB-field

nucleus +

nucleus+

Basar, Kharzeev, Skokov PRL 109 (2012) 202303

B.Müller, S.Y.Wu, D.LYang PRD 89 (2014) 026013

Radiation from Flowing Glasma

Geometrical scaling behavior of direct photon yield

Describes centrality dependence of Au+Au dataDescribes pp, dAu, AuAu at RHICDescribes scaling of spectra RHIC-LHC

Ideal fluid (hydrodynamic expansion) prior to thermal QPD may result in observed anisotropic “flow”

Axel Drees19

C.Klein-Bösing, McLerran, Phys. Lett.B734 (2014) 282

Flow vs B Field Effect

Look at different collision systems

Axel Drees20

Au+Aucentral

U+Ucentral

Cu+Ausemi-central

B = 0v2 = 0

B = 0v2 ≠ 0

B ≠ 0v2 = 0

U+U and Cu+Au data sitting on tape waiting to be analysis

Summary and Outlook

Well established measurement of “thermal” photon in Au+Au at 200 GeV from PHENIX

Large yield above expected contribution from pQCDCentrality dependence of yield ~ Npart

3/2

Large anisotropy v2 with respect to reaction plan

Thermal photon puzzleModels based on standard rates and time evolution fail to describe simultaneously photon yield, T and v2New additional sources early in collision

Enhanced emission due to large B fieldsEmission from flowing glasma state

Additional experimental measurements from RHICVary collision geometry U+U and Cu+Au62.4 (and 39 GeV) Au+AuNew large Au+Au data samples to measure vn

Axel Drees21