1012 Degrees in the Shade:
Physics of the quark-gluon plasma
Craig Ogilvie
1. Motivation: non-perturbative quantum chromodynamics (QCD)
2. Results from RHIC
a. Properties of quark-gluon plasma to study QCD
b. Outstanding questions, next steps
Non-perturbative QCD
Strength of QCD interaction determined by the
coupling (αs) between quarks (q) and gluons (g)
g
q
Perturbation series in αs
As αs nears 1, perturbative tools fail,
what happens
pp
to the p
physics
y
when the
interaction is very strong?
What new phenomena occur?
PDG, W.-M. Yao et al., Journal of
Physics G 33, 1 (2006)
Oct 3, 2007
Craig Ogilvie
2
1
Two systems to study non-perturbative QCD
proton
QGP
Hive of activity,
fluctuations of q-⎯q pairs, g
John Lajoie
confined byy
vacuum condensate
correlated <q⎯ q>
Oct 3, 2007
Strongly interacting system
With large average number q and g
Easier to use to study QCD?
⇒Statistical tools
⇒ reduced role of vacuum
3
Craig Ogilvie
Views of heavy-ion collision: Au+Au
Thermal freezeout
Hadronization
Equilibration
Oct 3, 2007
Craig Ogilvie
4
2
Assertion
„
In these complicated events, we have (a posteriori ) control
over the event geometry:
‰ Degree of overlap
Central
‰
Peripheral
Orientation with respect
p
to overlap
p
Oct 3, 2007
5
Craig Ogilvie
Transverse Dynamics
p
pT
θ
Oct 3, 2007
Craig Ogilvie
6
3
Physics of the QGP
Is plasma locally
thermalized?
Oct 3, 2007
Craig Ogilvie
7
Chemical Equilibrium
„
Abundance of hadrons determined by their mass and a
single common temperature (+ a chemical potential)
Yield ~ e-(m/kT)
Spectacular
agreement over
3 orders
=> Consistent with
system
t
att
chemical equilibrium
STAR Preliminary
Oct 3, 2007
Craig Ogilvie
8
4
Exponential Spectra:
„
Random motion + collective expansion (β=v/c ~ 0.6)
‰ Heavier particles larger increase in pt due to expansion
mt =
„
pt2 + m 2
Expanding, thermal system
Oct 3, 2007
9
Craig Ogilvie
Expansion is not isotropic
⇔
coordinate-space-anisotropy
momentum-space-anisotropy
y
x
dn/dφ ~ 1 + 2 v2(pT) cos (2 φ) + ...
„
„
If thermalized =>
> pressure gradient is largest in x-direction
x direction
‰ Expansion larger in that direction
Anisotropy characterized by elliptic flow, v2
Oct 3, 2007
Craig Ogilvie
10
5
Expansion is hydrodynamic
dn
dφ
2v2
Initial asymmetry propagates to final state
⇒Low momentum particles reproduced by hydro
( 99% of the particles)
⇒System thermalizes early (Δt < 0.6 fm/c)
Oct 3, 2007
11
Craig Ogilvie
What is flowing?
Universal scaling
=> system flows
while it is
deconfined quarks
„
„
Quark scaling: divide by nq valence quarks in particle
Convert to energy axis KET = pT2 + m02 (system does work)
Oct 3, 2007
Craig Ogilvie
12
6
Physics of the QGP
Does the QGP
behave like a gas
of quarks/gluons, or?
Plasma appears to be
locally thermalized
O
Ongoing
i question:
ti
temperature
t
t
via emission of photons
Oct 3, 2007
13
Craig Ogilvie
Current :) predicted QCD phase diagram
Marzia Rosati, LHC
T (MeV)
Hot QGP
Attractive interactions in plasma
N
t b ti l strong
t
Correlated q⎯ q Non-perturbatively
αs(q) large, since
sQGP
momentum transfer ~ T
200
packed
hadrons
μ (MeV)
nuclei
Oct 3, 2007
Craig Ogilvie
14
7
Probe of strongly coupled fluid: shear viscosity
„
Shear gradient of velocity distribution du1/dx2
u1
2
1
„
u1
Exchange of
constituents
Viscosity η reduces shear gradients, e.g. by transport of
constituents from one fluid-cell to another
Long mean-free path
Eff. Transport of p
Strong coupling between
fluid elements increases η
Minimum
viscosity
OctcQGP,
3, 2007
coupling
Nucl-th/0601029
Craig Ogilvie
15
Z Donko, Phys Plasma 7, 45 (2000
Shear viscosity reduces elliptic flow
Quantum
lower bound?
P Romatschke,
P.
Romatschke nucl-th:0706
nucl-th:0706.1522
1522
Perturbative QCD
QC η/s
/ >0.5
0 => sQGP
QG
Calc η/s in non-perturbative QCD…
RHIC viscosity close to conjectured lower quantum bound
perfect fluid?
Oct 3, 2007
Craig Ogilvie
16
8
Viscosity in other systems
„
Perfect fluid has η=0 (or lower bound η/s > 1/4π)
Kovtun et al, PRL 94
111601 (2005)
Physics Today
Today, 58(5)
RHIC, 4π*η/s~ 1
at 1012 K
Oct 3, 2007
17
Craig Ogilvie
Physics of the QGP
Plasma appears to be
locally thermalized
QGP behaves like
a strongly coupled liquid, η key
Ongoing question: Charm-quark binding
with di-quark
q
=> flow, spectra,
p
Λc
How dense is the QGP?
Oct 3, 2007
Craig Ogilvie
18
9
Hard Scattered Parton {Quark or Gluon}
hard-scattered parton:
high pt-transfer,
calc. with perturbative QCD
cone of hadrons, pt leading hadron
hadrons have less energy,
broader angular spread ?
parton loses energy
within plasma
high pt
p
hard-scattered
parton during Au+Au
p
Oct 3, 2007
Craig Ogilvie
19
Baseline p+p π0 spectra compared to pQCD
dσ = f a / A ( xa , μ 2 ) ⊗ f b / A ( xb , μ 2 )
•parton distribution functions,
for p
partons a and b
•measured in DIS, universality
⊗ σˆ (a + b → c + d )
•perturbative cross-section (NLO)
•requires hard scale
•factorization between pdf and cross section
⊗ Dπ / c ( zπ , μ 2 )
•fragmentation function
•measured in e+ePhys. Rev. Lett. 91, 241803 (2003)
Oct 3, 2007
Craig Ogilvie
20
10
High-pt parton loses energy within QGP
„
High-pt parton scatters in plasma, radiates gluons
Radiated gluon has formation time ~ 1/Egluon
‰ During formation time
time, scatter from many partons in plasma
L
Scattering centers = color charges
q
q
g
‰
Multiple-scatters treated coherently
„ Reduces total gluon emission rate
„ Reduces energy-loss (though still large :)
Oct 3, 2007
Craig Ogilvie
Au+Au π0 spectra
21
RAA = ratio (Au+Au)/(p+p)
Enhanced
„
Suppresse
ed
Oct 3, 2007
Craig Ogilvie
22
11
How Opaque is QGP?
L
q
g
q
μ 2 momentum transferred
qˆ =
=
λ
mean free path
Scattering power of
the QCD medium:
Range probable values
Oct 3, 2007
Craig Ogilvie
23
Strong energy-loss not fully understood
RHIC data
QGP
R. Baier
Pion gas
Cold nuclear matter
Larger than expected from QGP that interacts perturbatively
Calculate q in non-perturbative QCD, compare to expt
Oct 3, 2007
Craig Ogilvie
24
12
Fate of Gluons
parton loses energy
within plasma
Do radiated gluons
produce broader jjet?
p
„
Jets broader in Cu+Cu than p+p
‰ Fragmentation of induced gluon radiation?
Oct 3, 2007
Nathan Grau, Hua Pei
25
Craig Ogilvie
Physics of the QGP
QGP behaves like
a strongly coupled liquid
Plasma appears to be
locally thermalized
QGP is opaque colored system,
Scattering power, q, is key
Ongoing question: is the modeling of gluon radiation right?
Energy-loss of charm-quarks
Oct 3, 2007
Craig Ogilvie
26
13
VTX $8M upgrade, C. Ogilvie project manager
Designed to provide early time probe
Charm, beauty: open questions sQGP
silicon pixel+strip detectors
Tracks extrapolate back to collision
Displaced vertices
=> charm (D), beauty (B)
Requires ~ 50 μm precision
e
X
D
D→e+X
X
Au
B →e+X
X
Oct 3, 2007
Craig Ogilvie
Au
e
27
VTX
VTX
Oct 3, 2007
Craig Ogilvie
28
14
Wafers Testing at BNL
Rachid Nouicer (BNL), Kieran Boyle (SBU), Hua Pei (ISU)
and Junkichi Asai (RBRC)
BNL
SBU
Oct 3, 2007
RBRC
ISU
29
Craig Ogilvie
Strip Front-End Electronics
On the ladder
Signal from hit strip => digitized => collected by read-out-card (ROC)
Led by Alan Dion (ISU post-doc)
1) Testing, debugging ROCs
2) Optimizing ROC+sensor
Control Room
Off-Detector
FEM
GTM
Clock & Control parsing
ARCNet
Serial Control parsing
DCM
Data Formatting
ROC
RCC
14 6 14
12 SVX4s
Oct 3, 2007
Craig Ogilvie
Data Signals
18x2
13x2
Detector Acceptance
Bus
Pilot
Module
Control Signals
ROC
RCC
14 6 14
12 SVX4s
ROC
RCC
14 6 4
12 SVX4s
ROC
RCC
14 6 14
12 SVX4s
ROC
RCC
14 6 14
12 SVX4s
30
15
eRHIC
To establish the existence, then study
the strongest possible gluon field
Process continues until saturation
⇒Maximum strength color field
Kirill Tuchin
Gluon splitting
=> lower momenta
Oct 3, 2007
Craig Ogilvie
31
Summary
Non-perturbative QCD
As αs nears 1, perturbative tools fail,
what happens to the physics when the
interaction is very strong?
Strongly coupled QGP
⇒ Colored, opaque, and low-viscosity
⇒ Use these quantities (q, η) to test QCD
Open questions <= VTX upgrade
1. Energy-loss + gluon radiation
2. Charm coupled to di-quarks in fluid?
Oct 3, 2007
Craig Ogilvie
32
16
Thanks!
Oct 3, 2007
Craig Ogilvie
33
Craig Ogilvie
34
Backup
Oct 3, 2007
17
Viscosity lower bound
F
= η∇ y v x
A
For dilute systems, η ~ mean free path, transport of momentum across fluid cells
η ~ npλ
Limit on η , use Heisenberg
pλ ≥ =
η ≥ n= or
η
n
≥=
For relativistic systems, number of particles not so well defined,
Entropy density s ~ kBn
η
s
Oct 3, 2007
≥
=
kB
35
Craig Ogilvie
Expansion
spectra (π)
HBT
⟨β⟩
Oct 3, 2007
Craig Ogilvie
36
18
PHENIX Experiment at RHIC
„
„
In operation since 2000
Designed for
penetrating
p
gp
probes to
characterize QGP
Focus on mid-rapidity arms
Charged tracking +
photons,, π0…+
Calorimetryy => p
J/ψ,…
Oct 3, 2007
Craig Ogilvie
√s=200 GeV, p+p => x
NLO QCD agrees well with data
Oct 3, 2007
Craig Ogilvie
37
D. d’Enterria
nucl-ex/0611012
38
19
Partons lose energy as they travel through QGP
π0 spectra at √s=200 GeV
nucl-ex/0611007
PHENIX
Eloss
„
„
p+p cross-section scaled by parton-flux in Au+Au
Fewer high-pt π0 in Au+Au
‰ Energy-lost by parton => info on density of QGP
Oct 3, 2007
39
Craig Ogilvie
Ratio of (Au+Au)/(scaled p+p spectra)
Drop possibly due to
isospin difference
p+p and A+A
„
„
Mesons suppressed μ 5 → energy-lost in QGP
γ scale with parton flux
Oct 3, 2007
Craig Ogilvie
40
20
Particles correlated with high-pt trigger
Δφ = φ1-φ2
STAR
p+p, PHENIX
PRD, 74 072202 (06).
A+A
„
Correlation survives high-multiplicity environment of A+A
Oct 3, 2007
41
Craig Ogilvie
Response of medium to passage of high-pt parton
Au+Au 0-10%
preliminary
„
„
3 < pt,trigger < 4 GeV
pt,assoc. > 2 GeV
Near-side, generation of ridge => strength large Δη (STAR talk)
Far-side: does super-sonic parton generate a mach-cone ?
hep-ph/0410067;
H.Stocker…
Jorge Casalderry-Solana
Oct 3, 2007
Craig Ogilvie
42
21
Far-side Production of Particles
PHENIX preliminary
nucl-ex/0611019
1<pt,ass<2.5<pt,trig<4 GeV/c
centrality
Observation
Ob
ti off particles
ti l produced
d
d
~1 radian away from back-to-back!
Fit with 2 Gaussians, each D radians away from π
D scales with system size
=> emission consistent with medium’s response to jet
Oct 3, 2007
Craig Ogilvie
43
Suppression of heavy-quark spectra in A+A
PHENIX
nucl-ex/0611018
PRL in press
Suppression of e, pT>3.0 GeV/c
Slightly smaller than light quarks
Challenge for models to reproduce both light, heavy-q Eloss
Expt. need to increase statistics, reduce systematics
silicon upgrade=> displaced vertices
Oct 3, 2007
Craig Ogilvie
44
22
Spectra:
„
Random motion + collective expansion (β=v/c)
‰ Heavier particles are boosted more by expansion
„
Expanding, thermal system
Oct 3, 2007
Craig Ogilvie
45
23