# No Slide Title - at STONY BROOK

W
Shared by:
Categories
Tags
-
Stats
views:
2
posted:
2/1/2012
language:
pages:
69
Document Sample

```							    The first two years of RHIC:
predictions vs. reality

Summary of the workshop:
Who wins the wine, and why?

And, by the way,

What did we learn from the exercise?
Particle yields and spectra

 global quantities
What do the data say?
G. Roland

dNch/dh = 640

Rises somewhat faster than Npart
Rapidity distribution

PHOBOS

dNp/dy ~ 220-230 per charge
dNK+/dy ~ 40
dNp/dy ~ 28
Net baryon density at mid-y
small, but not 0  mB small
Transverse energy
PHENIX preliminary
ET/particle
~ 0.9 GeV

Similar cent.
PHENIX preliminary         dependence
as <pt>

But <pt> goes
up with s by
20% while
ET is constant
 particle mix
is changing
Anti-particle/particle ratios
I. Bearden

Ratios similar to those in p+p!

p+p collisions

BRAHMS 200 GeV

At mid-rapidity:
Net-protons: dN/dy  7         ISR
proton yield: dN/dy  29       extrapolation
 ¾ from pair-production
What model can reproduce the net baryons?

Net baryon central
plateau (y=0 to ~ y=2)
Cannot (yet) differentiate
AMPT vs. HIJING/BJ
AMPT - CheMing Ko
Degree of difficulty = 3.5

 Ingredients:
   ET = 750 GeV at y=0 (50% off  *)
data say: 3.3 GeV x (300/2) = 495 GeV
   80 baryons at y= 3.9 (data say 34 at y=3.5)
at y=0: 14 p, 10 pbar; pbar/p = 0.6 (data say 29, 22, 0.74)
(ratio is within 25% of data  ***)
   dNch/dh ~ 800               (dNch/dh within 25%  ***)
430 p, 60 K per unit y at mid-y (data say 640 ,230, 40)
   Central plateau |y|<1.5 for mesons
(pion data says 1.5  *****)

Total score: 3.5 + 10.5 + 10.5 + 17.5= 42
What did we learn?

 To get proper particle yields must tweak model so it no
longer agrees with pp collisions
Changed fragmentation function to match lower s
data, rationale: fragmentation in dense matter

 Must add a partonic phase with large scattering cross
sections to reproduce v2 and HBT

rescattering channels
Then get f  K+K- correct in s = 130 GeV/A data
LEXUS – Joe Kapusta
Degree of difficulty = 2

 Ingredients: parameterized p+p collision results, Glauber,
NN hard collision probability parameter l = 0.6
   Minimalist approach, which works at SPS
   Net proton density = 13 (data say 7  *)
   dNch/dy = 1200, but should have been 950 using p+p at
proper s
Correcting by 15% for yh, get 1020 or 800
(800 is within 25% of data, but –1 for p+p oops  **)
   Particle spectra are too steep, but missing power law tail
   proton <pT> ~ 0.925 GeV/c (data say 0.94  *****)

Total score: 2 + 4 + 10 = 16
(so their next model will be a Bentley…?)
What did we learn?

 Create more hadrons in LEXUS than in wounded
nucleon model, since wounded nucleons are not sterile
in LEXUS. Overprediction  some destructive
interference among stopped nucleons at mid-y?

 Total multiplicity is fixed by energy conservation
 Baryon density fixed by Dy in each collision

 Minimalist picture works ~ OK for the simplest
observables, but not for more complex ones
 Caution in interpreting scaling with Ncoll or Npart !
Particle Spectra @ 200 GeV

BRAHMS: 10% central
PHOBOS: 10%
PHENIX: 5%
STAR: 5%

QM2002 summary slide (Ullrich)
Feed-down matters !!!
<pT> vs. Npart          J. Velkovska
<pT> [GeV/c]

<pT> [GeV/c]
•Systematic error on
200 GeV data
p (10 %), K (15 %),
p (14 %)

open symbol :
130 GeV data

• Increase of <pT> as a function of Npart and tends to saturate
p < K < proton (pbar)
• Consistent with hydrodynamic expansion picture.

<pT> prediction with Tth
and <b> obtained from
STAR
blastwave fit (green line)

<pT> prediction for
preliminary                 Tch = 170 MeV
F. Wang                     and <b>=0
pp no rescattering,
no flow
no thermal equilibrium

<pT> of X and W from
exponential fits in mT
Do they flow ? Or is
<pT> lower due to
different fit function?
Does it flow? Fits to Omega mT spectra
M. van Leeuwen (NA49)                                C. Suire (STAR)

STAR preliminary

RHIC
SPS/NA49

bT is not well constrained !

• At SPS W and X are now found to be consistent with common freeze-out
• Maybe W and X are consistent with a blastwave fit at RHIC
• Need to constrain further  more data & much more for v2 of W
UrQMD - Bleicher
Degree of difficulty = 2

 Ingredients: excitation and fragmentation of color strings,
rescattering
   dET/dh = 600, dNch/dh = 750, ET/Nch = 0.85 GeV
Data say 495, 640, 0.9
Get ET to 20%, Nch to 17%  *** and ***
   y=0: 12 net protons, 400 p-, 45 K+
Data: 7, 230, 40  *, *, and ****
   <pT> = 375, 500, 780 for p, K, p
Data: 400, 650, 940  ****
Score = 32
   v2 ~ 1% (way too low as the strings don’t collide)
Dense set of non-interacting strings… a problem…
We learned that

 Need QGP-type equation of state to get the v2 and
UrQMD has insufficient initial pressure as the strings
don’t scatter.

 Mass shifts of resonances very sensitive to breakup
dynamics. Resonances are not dissolved  implies fast
freeze-out
Statistical model summary - Magestro
Degree of difficulty = 1
 Johanna: chemical equilibrium with T=170 MeV, mB =
10 MeV
 Johann: sudden freezeout with incomplete chemical
equilibrium
Predictions (200 GeV)
Exptl. (130 GeV)   Exptl. (200 GeV)
0.66                0.75
0.074               0.076       T=177MeV
0.90                0.95
mB =29 MeV
0.15                0.15
0.75
STAR     PHENIX
0.58     0.89      Scores:
0.66     0.95      Johanna – within ~15%
0.021                       ****
0.19
Johann - within ~ 40%
0.0015                      **
Lessons from statistical analyses

 See chemical equilibrium populations at RHIC as at SPS
mB is lower, but not as low as predicted
No anomalous strangeness enhancement

 Simple thermal emission produces proton spectra flatter
than pion spectra, so they must cross someplace!
Of course the big question is where and why there??
Elliptic flow
Centrality dependence of v2
Note possible dependence on low pt cut
200 GeV: 0.2< pt < 2.0
130 GeV: 0.075< pt < 2.0
200 GeV: 0.150< pt < 2.0
4-part cumulants
STAR

v2=0.05

STAR                     Preliminary
Preliminary

200 GeV: Preliminary

- Consistent results
- At 200 GeV better pronounced decrease
of v2 for the most peripheral collisions.

QM2002 summary slide (Voloshin)
A puzzle at high pT   Nu Xu

 Still flowing at pT = 8 GeV/c? Unlikely!!
v2 of mesons & baryons
Au+Au at sNN=200GeV

1) High quality M.B. data!!!

2) Consistent between
PHENIX and STAR

pT < 2 GeV/c
v2(light) > v2(heavy)
v2
pT > 2.5 GeV/c
v2(light) < v2(heavy)

Model: P.Huovinen, et al., Phys. Lett.
B503, 58 (2001)
Hydrodynamics – Ulrich Heinz, Peter Kolb
Predictions of major importance!

 Ingredients: thermal with some initial conditions, QGP EOS
early with transition to resonance gas, geometry + Glauber,
hydrodynamics
 Predictions:
Thermalization by 0.6 fm/c at RHIC
v2 as function of pion multiplicity density (to fix initial cond.)
v2 has a dip (~5%) due to phase transition softening EOS
RHIC is near this point (data says v2 ~ 6%)
v2 vs. pT increases to 2 GeV/c
v2(mesons) > v2 (baryons)
spectra (once initial condition is fixed)
 Lessons: v2 requires early rescattering! Hadronization follows
thermalization by 5-7 fm/c. But, final state decoupling needs
work (get HBT wrong)
Hydrodynamics –Teaney & Shuryak

 Ingredients: hydrodynamics + RQMD for hadronic
state and freeze-out
 Predictions:
RHIC should be near softest point in EOS
s dependence of v2 correctly predicted for b=6 fm
fixed initial conditions, then got spectra correct
Predict particle yields without rescaling
Initial entropy too high, HBT radii too large!
 Lessons: hydro good to pT ~ 1.5 GeV/c
Viscosity corrections may be important; cause v2 to
bend over at 1 GeV/c pT (compared to ideal gas).
Also helps reduce HBT radii. Maybe small viscosity
early, but increases in hadron gas phase?
Parton transport theory – Denes Molnar
Degree of difficulty = 5

 Next step beyond hydro – calculate parton transport, fixing
s (i.e. transport opacity c)
 Predictions & insights:
ET loss due to pdV work so (ET)cent < (ET)peripheral
ET results require small s (3 mb)
can’t easily fix up with inelastic collisions
need parton subdivision to avoid numerical “viscosity”
Can reproduce v2 if dNgluon/dy very large or sel= 45 mb
But large opacity underpredicts HBT spectra!

pQCD fixes dNgluon/dy at large pT
pQCD fixes partons at large Q2
 Picture doesn’t want to hang together!
Next, jets and high pT

summary from
Thomas Peitzmann,
QM2002

200 GeV results
from all experiments

Shape changes from peripheral  central

Preliminary sNN = 200
GeV

Preliminary sNN = 200 GeV                 C. Roland,
PHOBOS
Parallel Saturday
p/p at high pT

Higher than in p+p
collisions or fragmentation
of gluon jets in e+e-
collisions

Vitev & Gyulassy nucl-th/0104066

Can explain by combination of
hydro expansion at low pT with
jet quenching at high pT
Jet Quenching – Gyulassy, Wang, Vitev, Levai
Degree of difficulty = 5

 HIJING: Beam jets @ pt<2 GeV (LUND), pQCD mini
jets @ pt>2 GeV (PYTHIA), geometry (Glauber), 1D
expansion, conservation laws; tuned to pp data 10-103 GeV
 + nuclear shadowing and parton energy loss “knobs”

GLV “Thin” Plasma Limit

L/ lg Opacity Expansion
No Quench
dEg/dx=0.5 GeV/fm
BDMS “Thick” Plasma Limit
dEg/dx=2.0
Nbinary
? 2003 ?

PHENIX 130

BRAHMS
hch
STAR 130
PRL88(02)

Npart/2

15% too many particles, baryons over-quenched, but predicted the suppression
BUT: dE/dx =2 GeV/fm or 0.5 GeV/fm or not linear with x?
Vitev: they can get v2 right

• There is a quantitative difference
Calculations/fits with flat  v2 = const  .
or continuously growing  v2  ln pT / m  .

Check against high-pT data (200 AGeV)

b~7 fm                                        b=7 fm

Same for 0-50%            C. Adler et al. [STAR Collab.],
arXiv: nucl-ex/0206006

• The decrease with pT is now
supported by data
• For minimum bias this rate is
slightly slower
K. Filimonov [STAR Collab.],
See: N.Borghini, P.Dinh, J-Y.Ollitrault,
arXiv: nucl-ex/0210027                             Phys.Rev. C 64 (2001)
Other penetrating probes

 Open Charm
 J/Y
 Dileptons

Need (a lot) more statistics in the data
But getting a first sniff of physics already
J/Y

Energy/Momentum

Data consistent with:
Hadronic comover breakup (Ramona Vogt) w/o QGP
Limiting suppression via surface emission (C.Y. Wong)
Dissociation + thermal regeneration (R. Rapp)
Open charm - Lin

predicted curves
with or w/o
energy loss

no x4 suppression
from peripheral to
central,
as predicted for
dE/dx=-0.5GeV/fm

But -
Is 40-70% peripheral
enough? error bars
still big!
Some old things and some new things

 HBT
 High pT baryons
 Dijets vs. monojets
Well, there was a prediction but for 10x the pT
 Parton saturation
HBT – lots of questions
Panitkin, Pratt

• How to increase R
without increasing
Rout/Rside?
 EOS, initial T and r
profiles (Csőrgó),
emissivity?

• Why entropy looks low?
Low entropy implies
equilibrated QGP ruled
out
Baryons at high pT      Jia, Sorenson

protons

p0, h

Yields scale with Ncoll near
pT = 2 – 3 GeV/c
Then start to fall
Meaning of Ncoll scaling?
Accident? Complex hard/soft interplay?
Medium modified jet fragmentation function?
Away-side Jet Suppression
D. Hardtke

 trigger-jet
not much
modification
(the trigger
particles
from jets!)

 Away side:
strong jet
suppression
Strong jet suppression  surface emission of jets?
Color glass back-to-back jets simply not created…
Parton saturation
Dima Kharzeev, Jamal Jalilian-Marian

 Hadron multiplicities imply a coherent initial state
Initial NN interactions are NOT independent!
High parton density  weak coupling  CGC
 Saturation at y=0, and even more so at forward y
affects QCD evolution, even at Q2 > Qs2
causes multiplicity to scale with Npart, even at high pT
hard parton scattering suppressed by CGC 
monojets
 does saturation set in already at s ~ 20GeV? I doubt this!
 Should measure in forward y in p+A, where Qs is larger
and CGC is magnified.
This should clarify initial vs. final state effect in AA!
conclusions

 Have early pressure buildup – high dNg/dy & they scatter!
 success of hydro, need for string melting, large s…
 High pT, high mass data look like pQCD + something
Jet quenching works; surface emission??
Baryon flow is a nuclear effect!
Color glass is intriguing, but where does the collectivity
come from?
 Event generators (still) a valuable tool to investigate
sensitivity of observables to physics ingredients
 Integrated quantities are simple (conservation laws!)
 Caution in interpreting scaling with Npart or Ncoll

 e+e- scaling with Npart is arbitrary, agreement irrelevant

Experiments: homework to allow quantitative comparisons
(multiple 15% factors = sloppy interpretations!)
And the winners are…

 Best predictions of general features by event generator
AMPT (Ko, Lin, Zhang)

 Novel approach, theoretically intriguing (+ agrees with data)
Baryon junctions (Kharzeev, Vance, Gyulassy, Wang)

 Important prediction with potential great insights to QGP
Hydrodynamics (Heinz & Kolb, Teaney & Shuryak, Bass &
Dumitru, Ollitrault for “inventing” v2 analysis)

 Much promise for understanding properties of QGP
Jet energy loss (Gyulassy,Wang, Vitev, Levai)
yield in AuAu vs. p-p collisions
D. d’Enterria

Yieldcentral/  N binary  central
Yield ratio s=200/130 GeV
Consistent at at high pT with                Yieldpp
pQCD predictions (STAR)

PHENIX Preliminary

70-80% Peripheral
Ncoll =12.3 ±4.0
kT dependence of R
Centrality is in top 30%

•Broad <kT> range : 0.2 - 1.2 GeV/c
•All R parameters decrease as a function of kT
 consistent with collective expansion picture.
• Stronger kT dependent in Rlong have been observed.
kT : average momentum of pair
Comparison of kaon to pion
In the most 30% central
Comparison with hydrodynamic model
Centrality is in top 30%

Recent hydrodynamic calculation
by U.Heinz and P. F. Kolb
(hep-ph/0204061)

Hydro w/o FS
• Standard initialization and freeze out
which reproduce single particle spectra.

Hydro at ecrit
• Assuming freeze out directly at the

kT dependence of Rlong indicates the
early freeze-out?
kT dependence of Rout/Rside
A. Enikizono
QM2002

C.M. Kuo, QM2002 poster (PHOBOS) 200 GeV:
1.16  0.09  0.25( syst .) @0.25 GeV/c
HBT PUZZLE
Small Rout implies small Dt

P.Kolb

Small Rbeam implies
Large Rside implies large R      small breakup t,~10 fm/c
Jet Evidence in Azimuthal Correlations at RHIC

 near-side correlation of           also seen in g (p0) triggered
charged tracks (STAR)                events (PHENIX)
trigger particle pT = 4-6            trigger particle pT > 2.5
GeV/c                                GeV/c
Df distribution for pT = 2-4
Df distribution for summary slide (Peitzmann)
QM2002 pT > 2
GeV/c                                GeV/c
 signature of jets                

M. Chiu, PHENIX Parallel Saturday
Identifying Jets - Angular Correlations

 Remove soft background
by subtraction of mixed event distribution

 Fit remainder:                                  raw differential yields
Jet correlation in Df;             PHENIX Preliminary 2-4 GeV
shape taken from
PYTHIA
to correct flow effects
Verify PYTHIA using p+p collisions

Df(neutral E>2.5 GeV + 1-2 GeV/c charged partner)

Make cuts in Dh to enhance
near or far-side correlations
Blue = PYTHIA
|Dh|<.35                         |Dh|>.35
In Au+Au collisions
Df(neutral E>2.5 GeV + charged partner)

1-2 GeV partner
Correlation after mixed event background
subtraction
Clear jet signal in Au + Au
Different away side effect than in p+p

|Dh|<.35                            |Dh|>.35
1/Ntrig dN/dDf

1/Ntrig dN/dDf
jets or flow correlations? fit pythia + 2v2vjcos(2f)

partner = .3-.6 GeV   .6-1.0 GeV/c   1-2 GeV/c   2-4 GeV/c
1/Ntrig dN/dDf

Df
v2vj

Jet strength
See non-zero jet strength as partner pT increases!
How do protons scale with Ncoll/Npart?

Scale with Ncoll (unlike p)?!
High pT baryons scale with Ncoll!
J. Velkovska

Low pT near Npart scaling

But baryons with pT > 2 GeV/c
behave very differently!
From jets? Unsuppressed??
Charm cross section at RHIC
Centrality dependence of charm
Homework assignment (PHENIX & STAR)

Yieldcentral/  N binary  central
Yieldperipheral /  N binary  peripheral

Charged larger than p0
But difference not same
as for RAA
PHENIX and STAR RAA
not the same
Different reference in
each case!
Systematic difference
between experiments
Charged hadron correlations - small Df

pT
jT   Correlation width  jT/pT
Correlation width

•Fit charged correlations with v2 + Gaussian (fixed pT)
•Jet signal visible via s
Width of near-side Gaussian decreases with pT
No significant centrality dependence on near-side
How do high pT yields scale?
 vs. binary collisions:
continuous decrease as
function of centrality
factor ~ 3.5 from
peripheral to central
 vs. participants:
first increase, then
decrease as function of
centrality
for Npart > 100 have 3s
change (scaling or no?)
surface emission?
re-interactions?
accident?
18% scaling uncertainty from corrections
dN/dy

dN/dy / (0.5 Npart)        PHENIX Preliminary                  PHENIX Preliminary

Au+Au at sqrt(sNN) =200GeV           Au+Au at sqrt(sNN) =200GeV

p+                                   p+

K+                                    K-

p
open symbol :                                            pbar
130 GeV data
Positive                             Negative

Npart                                 Npart
• Similar centrality dependence 130 GeV and 200 GeV
Opaque, expanding source would mean:

R - R = b ( Dt ) + ( X - Y ) - 2 b s
2
o
2
s
2

2          2        2                     2
xt
Opaque              Expanding

Y (side)

X (out) 

Rs( half -shell )   5
Rischke RIKEN workshop (2002):                        ( sphere)
=   = 1.29
Rs                 3
Such strong xt correlations probably
require a lack of boost-invariance...        Ro half -shell )
(
5
( sphere)
=    = 0.65
Ro                12
Energy Dependence

Assumptions:
PHENIX preliminary
in Lab             in C.M.
dX dX             dX       dX
                   1.2
dy dh             dy       dh

Energy density (Bjorken):
1 dEt         R = 1.18 fm  A1/ 3
= 2
pR t dy       t = 1 fm / c
PHENIX preliminary
2% most central at sNN=200 GeV:
  5.5 GeV/fm3

From AGS, SPS to RHIC:
Transverse energy and charged particle
multiplicity densities per participant
consistent with logarithmic behaviour
p,K,P spectra from Star
 High quality data over
9 centrality selections
 Shape described by
blast wave fit
K-/K+ and p/p from AGS to RHIC
I. Bearden (BRAHMS)

Becattini caluclation using
statistical model:
T=170, gs=1 (weak dependency)
vary mB/T  K+/K- andp/p

K- /K+=(p/p)1/4 is
a empirical fit to the data points

K-/K+ driven by ms
~ exp(2ms/T)

p/p driven by mB
~ exp(-2mB/T)

ms = ms (mB) since <S> = 0
QM2002 summary slide (Ullrich)
BUT: Holds for y  0 (BRAHMS y=3)
The K*0 story

STAR Preliminary
   K*0/Ksuppressed in AA versus pp
 f/K*0 appears enhanced versus pp

pp  uncorrected for trigger bias
and vertex finding efficiency

STAR QM Talks: E. Yamamoto and P. Fachini
v2 at high pT

min bias 200 GeV Au+ Au
Centrality dependence of p/pi

•Ratios reach ~1 for
central collisions                           +
•Peripheral collisions
lower, but still above gluon
jet ratios at high pT

•Maybe not so surprising
1)“peripheral” means 60-
91.4% of stotal                              -
2) p/pi = 0.3 at ISR
Note pbar/p behavior

Centrality dependence only
for pT > 3 GeV/c

Peripheral collisions have
quite a few protons at mid-y

```
Related docs
Other docs by ewghwehws
Patent US2100036
Child__039;s hobbyhorse