EUROPEAN ORGANIZATION FOR NUCLEAR RESEARCH
CERN-OPEN-2012–005
15 February 2012
First Results from Pb+Pb collisions at the LHC
Abstract
At the end of 2010, the CERN Large Hadron Collider started operation with heavy ion beams, col-
liding lead nuclei at a centre-of-mass energy of 2.76 TeV/nucleon and opening a new era in ultra-
relativistic heavy ion physics at energies exceeding previous accelerators by more than an order of
magnitude. This review summarizes the results from the first year of heavy ion physics at LHC ob-
tained by the three experiments participating in the heavy ion program, ALICE, ATLAS, and CMS.
Berndt Mu¨ller
Department of Physics, Duke University, Durham, NC 27708, USA
Ju¨rgen Schukraft
PH Division, CERN, CH-1211 Geneva 23, Switzerland
Bolesław Wysłouch
Department of Physics, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
(To appear in Annual Review of Nuclear and Particle Science)
ar
X
iv
:1
20
2.
32
33
v1
  [
he
p-
ex
]  
15
 Fe
b 2
01
2

First Results from Pb+Pb collisions at the LHC 1
1 Introduction
After almost two decades of design and construction, the LHC accelerator [1] started operation at the
end of 2009 with proton-proton (pp) collisions at 0.9 and 2.36 TeV centre-of-mass energy, reaching its
current top energy of
√
s = 7 TeV in 2010. Heavy ion collisions (A+A) are an integral part of the LHC
physics program. The first four-week run with Pb beams took place in November 2010 at a centre-
of-mass energy of 2.76 TeV per nucleon. About 10 µb−1 of integrated luminosity were delivered to
each of the three LHC experiments which take part in the heavy ion program, the two general purpose
detectors ATLAS [2] and CMS [3] as well as the dedicated heavy ion detector ALICE [4, 5]. This review
summarises the results from the first year of ion physics at the LHC, focusing on global event properties,
high pT and heavy quark physics, and the production of hidden heavy flavour mesons (J/ψ , ϒ).
2 The Physics of Relativistic Heavy Ion Collisions
Relativistic heavy ion collisions make it possible to study the properties of strongly interacting matter at
energy densities far above those of normal nuclear matter. General arguments based on the property of
asymptotic freedom suggest, and quantitative calculations of lattice quantum chromodynamics confirm
[6, 7, 8], that QCD matter undergoes a transition from a hadronic gas to a quark-gluon plasma at a
temperature Tc≈ 160 MeV, corresponding to an energy density of εc≈ 0.5 GeV/fm3. At small net baryon
density, the transition is a smooth cross-over spanning a temperature range of 20−30 MeV, which means
that the precise value of the pseudo-critical temperature Tc depends on the observable used to locate it.
The quark-gluon plasma phase is characterized by a much reduced condensate of light quarks, reflecting
the approximate restoration of chiral symmetry, and by screening of the chromo-electric force between
heavy quarks, implying the absence of quark confinement. At high net baryon density, the confined and
deconfined phases are thought to be separated by a first-order phase transition, beginning with a critical
point whose position is not accurately known.
A large body of experimental and theoretical research conducted over the past decades at AGS, SPS, and
RHIC has led to the following “standard model” of heavy ion collisions at collider energies [9, 10]. The
energy deposited in the mid-rapidity kinematic range is controlled by the density of gluons contained
in the colliding nuclei at moderately small values of the Bjorken scaling variable x [11]. At top RHIC
collision energy the relevant range is 〈x〉 ≈ 10−2, while at the current LHC energy the value is 〈x〉 ≈
10−3. In this range of x the gluon density obeys a highly nonlinear evolution equation that describes
the saturation of the rapidly growing perturbative gluon distribution at low virtuality [12]. The virtuality
scale at which this saturation occurs is called the saturation scale Qs. At present collider energies and for
heavy nuclei Q2s ≈ 1.5−2 GeV2 at RHIC and Q2s ≈ 3−4 GeV2 at LHC [13].
The liberation of the dense gluon sea by interactions during the collision of two nuclei is thought to
result in the formation of a dense, non-thermal QCD plasma with highly occupied gauge field modes,
often called a “glasma” [14]. The glasma thermalizes rapidly through nonlinear interactions including
plasma instabilities [15, 16]. During the longest phase of the heavy ion reaction the QCD matter forms a
nearly thermal quark-gluon plasma, whose evolution can be described accurately by relativistic viscous
hydrodynamics because of its very small kinematic shear viscosity. The temperature of the plasma at the
time of thermalization is approximately 2Tc; it cools by mainly longitudinal expansion until it converts
to a gas of hadron resonances when its temperature falls below Tc. At that temperature the chemical
composition of the produced hadrons gets frozen, but the spectral distribution of the hadrons is still
modified by final-state interactions, which are modelled by a Boltzmann transport equation.
The very small dimensionless ratio of the shear viscosity η to the entropy density s of the QCD matter
produced in heavy ion collisions (4piη/s≤ 2.5 [17]) was one of the main discoveries of the experimental
program at RHIC [10]. Such a small ratio requires that the interactions among the constituent quanta
are extremely strong and the cross sections among them approach the unitarity limit [18]. This picture
2is supported by QCD lattice simulations, which show that the QCD matter in the temperature range
Tc ≤ T ≤ 2Tc is highly nonperturbative as witnessed, e. g., by the large value of the trace anomaly of the
energy-momentum tensor, T µµ = ε−3P [7].
The strongly coupled nature of the quark-gluon plasma created in the heavy ion reaction is confirmed by
its ability to “quench” jets. At RHIC, the two main observables revealing this property are the suppression
of single hadron yields at high transverse momentum pT relative to pp reactions and the additional
suppression of back-to-back emission of high-pT hadrons. The ratios quantifying these suppressions are
known as RAA and IAA, respectively. Compelling arguments related the suppression effects to the ability
of the quark-gluon plasma to degrade the kinetic energy of a hard parton (quark or gluon) traversing it,
either by elastic collisions or collisions followed by gluon bremsstrahlung. The large values of the energy
loss rate of hard partons deduced from the RHIC data [19] are consistent with the low shear viscosity
deduced from the collective flow pattern of low-momentum hadrons [20].
Before the LHC turned on, it was not clear whether this quantitative model of the dynamics of relativistic
heavy ion collisions and the properties of hot QCD matter smoothly extrapolates from RHIC to the one
order of magnitude higher LHC energies. Questions of particular interest were whether the increase in
the initial temperature of the quark-gluon plasma would result in a much more viscous expansion of the
created matter, and whether the gluon saturation effects in the colliding nuclei would be so strong that
they significantly reduce, not only the expected density of the glasma, but also the yield of particles at
moderately large pT . Because of the much higher beam energy and the large acceptance of the LHC
detectors, heavy ion collisions at the LHC were also expected to provide access to new observables
measuring the pattern of jet quenching and extend our knowledge of the interaction of heavy quarks and
heavy quarkonia with hot QCD matter. An extensive compilation of theoretical predictions made before
the commencement of the LHC heavy ion program can be found in [21].
3 Global event properties
Global event properties describe the state and dynamical evolution of the bulk matter created in a heavy
ion collision by measuring characteristics of the vast majority of particles which have momenta below a
few GeV/c, referred to as “soft” particles. They include multiplicity distributions, which can be related
to the initial energy density reached during the collision, yields and momentum spectra of identified
particles, which are determined by the conditions at and shortly after hadronization, and correlations
between particles which measure both size and lifetime of the dense matter state as well as some of its
transport properties via collective flow phenomena.
3.1 Multiplicity distributions
The most basic quantity, and indeed the one measured within days of the first ion collision, is the number
of charged particles produced per unit of (pseudo)rapidity, dNch/dy (dNch/dη), in a central, “head-on”
collision. When the LHC heavy ion program was conceived and detectors had to be designed more
than a decade ago, predictions for dNch/dη were highly uncertain, ranging from below 1000 to above
4500, because experimental results had to be extrapolated from light ions at low energy (Sulphur beams
at
√
s = 20 GeV) by orders of magnitude in both ion mass and beam energy [22, 23]. With results
from RHIC, the uncertainties where substantially reduced, with most predictions concentrating in the
range dNch/dη = 1000− 1700 [21]. The value finally measured at LHC, dNch/dη ≈ 1600 [24], was
on the high side of this range. From the measured multiplicity one can derive a rough estimate of the
energy density with the help of a formula first proposed by Bjorken [25] relating the energy density to
the transverse energy:
ε ≥ dET/dη
τ0 piR2
=
3
2
〈ET/N〉dNch/dητ0 piR2 (1)
First Results from Pb+Pb collisions at the LHC 3  
(G
eV
)
N
N
s
Ce
nt
re
 o
f m
as
s 
en
er
gy
 
1
10
2
10
3
10
)〉
part
N〈 )/(0.5 η /d
ch
N (d
0246810
Sp
 
N
SD
 S
p
pp pp
 N
SD
 R
HI
C
 
N
SD
 F
NA
L
pp pp
 N
SD
 L
HC
A
u
A
u
 
A
G
S
Pb
Pb
 S
PS
A
uA
u 
R
H
IC
 a
ve
ra
ge
Pb
Pb
 
A
LI
CE
 
(sh
ift
ed
)
Pb
Pb
 
AT
LA
S 
(sh
ift
ed
)
Pb
Pb
 C
M
S 
(sh
ift
ed
)
Pb
Pb
 L
HC
 a
ve
ra
ge
0.
15
3
0.
75
s
 
-
 
0.
4
s
0.
78
ln 0
.1
09
0.
80
s
〉
pa
rt
N
 〈
N
um
be
r o
f p
ar
tic
ip
an
ts
 
0
10
0
20
0
30
0
40
0
)〉
part
N〈 )/(0.5 η /d
ch
N (d
0246810
LH
C 
Pb
Pb
 
2.
76
 
Te
V
R
H
IC
 
A
u
A
u
 
20
0 
G
eV
 
x
 
2.
14
pp
 
In
el
 
2.
76
 
Te
V
pp
 In
el
 2
00
 G
eV
 x
 2
.1
4
D
PM
JE
T 
III
 
H
IJ
IN
G
 2
.0
 (s
g=
0.2
3)
A
lb
ac
et
e 
et
 a
l.
Fig. 1: a): Charged particle pseudorapidity density dNch/dη per colliding nucleon pair (0.5Npart) versus centre of
mass energy for pp and A+A collisions. b): dNch/dη per colliding nucleon pair versus the number of participating
nucleons together with model predictions for Pb+Pb at 2.67 TeV
where τ0 denotes the thermalization time, R is the nuclear radius, and ET/N ≈ 1 GeV is the transverse
energy per emitted particle. The value measured at the LHC implies that the initial energy density (at
τ0 = 1 fm/c) is about 15 GeV/fm3 [26], approximately a factor three higher than in Au+Au collisions at
the top energy of RHIC [27, 28, 29, 30, 10]. The corresponding initial temperature increases by at least
30% to T ≈ 300 MeV, even with the conservative assumption that the formation time τ0, when thermal
equilibrium is first established, remains the same as at RHIC.
The high multiplicity at LHC, together with the large experimental acceptance of the detectors, allow for
a very precise determination of the collision geometry (impact parameter and reaction plane orientation)
in each event. Events are classified according to “centrality”, where e.g. the 0-5% centrality bin contains
the 5% of all hadronic interaction events with the largest final-state multiplicity and therefore the smallest
impact parameters. Correspondingly, the 90-100% centrality bin comprises the 10% of events with
the smallest final-state multiplicity and the largest impact parameters. The experimentally measured
centrality can be related analytically or via a probabilistic Glauber Monte-Carlo generator to the number
of participating nucleons Npart which are contained in the nuclear overlap volume for a given centrality
class and therefore take part in the collision [31]. Likewise one can extract the number of nucleon-
nucleon collisions Ncoll between the participants. For Pb+Pb at LHC, the centrality or Npart resolution
varies from about 0.5% for central (Npart ≈ 400) to about 5% for peripheral (Npart < 10) events [32, 33,
34].
The charged particle multiplicity per participant pair [32, 33, 34], dNch/dη /(0.5 Npart), is shown in Fig. 1
together with lower energy data [35, 36, 37] for central A+A collisions (typically 0-5% or 0-6% cen-
trality). Note, however, that even for equal centrality selections, the average number of participants or
collisions (Npart, Ncoll) are different at the various energies because of different nuclear size (Pb or Au)
and the rising pp cross-section. Particle production is no longer compatible with a logarithmic depen-
dence with
√
s, as it was true for the data up to top RHIC energy [37], but follows a power law ≈ s0.15.
Also the pp data are well described by a power law, however with a less steep dependence on energy
(≈ s0.11).
The centrality dependence of particle production is compared in Fig. 1b with the one measured at RHIC,
which is normalised to the LHC result at Npart= 350 by scaling it with a factor 2.14. The results from
the three LHC detectors [32, 33, 34] are in excellent agreement with each other (within 1-2%) and have
been averaged in this figure using the prescription of [35], assuming conservatively that the systematic
errors in Npart are fully correlated between experiments. Comparison to the averaged and scaled 200 GeV
Au+Au data (from [35], updated using more recent STAR [36] and PHOBOS [37] data) shows a remark-
able similarity in the shape of both distributions. For peripheral collisions, however, both distributions
4extrapolate towards respective values measured in pp inelastic collisions (Npart= 2) at 200 GeV and 2.76
TeV and therefore start to separate because of the different energy dependence seen for pp and A+A in
Fig. 1a.
The fact that the shape of the normalised multiplicity distribution varies little with energy had already
been noticed at RHIC [37]. It was still a surprise that it stays almost constant up to TeV energies, because
hard processes, which scale with the number of binary collisions Ncoll, could be expected to contribute
significantly to particle production at LHC and lead to a steeper centrality dependence, as predicted by
the two component (soft + hard) Dual Parton Model DPMJET [38] (dotted line in Fig. 1b). However, a
strong impact parameter dependent shadowing of the nuclear parton distribution function can limit this
rise with centrality and is responsible for the flatter shape seen in the two component model HIJING [39],
which agrees better with the data (full line). Saturation physics based on the “Colour Glass Condensate”
(CGC) description [40], an example [41] is shown by the dashed line in Fig. 1b, naturally predicts such
a strong nuclear modification, as well as the strong rise of particle production as a function of
√
s seen
in Fig. 1a. In saturation models, the energy dependence is predicted to be a power law with roughly
the correct exponent, whereas the centrality dependence changes only weakly (logarithmically) with the
saturation scale and therefore the beam energy.
3.2 Identified particle spectra
Particle production (pi , K, p, Λ,..) is a non-perturbative process and cannot be calculated directly from
first principles (QCD). In phenomenological QCD inspired event generators, particle spectra and ratios
are adjusted to the data in elementary collisions (pp, e+e−) using a large number of parameters. In heavy
ion reactions, however, inclusive particle ratios and spectra at low transverse momentum, which include
the large majority of all produced hadrons even at LHC energies (about 95% of all particles are below 1.5
GeV/c), are consistent with simple descriptions by statistical/thermal [42, 43] and hydrodynamical [44]
models, where particle ratios are determined during hadronisation at or close to the QGP phase boundary
(“chemical freeze-out”, see below), whereas particle spectra reflect the conditions somewhat later in the
collision, during “kinetic freeze-out”.
In pp collisions at high energy, the transverse momentum distribution can be separated into an exponen-
tially decreasing low pT part (< few GeV/c), arising from soft, nonpertubative processes, and a powerlaw
high pT tail from hard QCD scattering and fragmentation. A defining characteristic of heavy ion colli-
sions is the appearance of ordered motion amongst the emitted hadrons in the soft part of the momentum
spectrum [45, 46]. It is called collective flow and implies, in contrast to random thermal motion, a strong
correlation between position and momentum variables (nearby particles have similar velocities in both
magnitude and direction). Flow arises in a strongly interacting medium in the presence of local pressure
gradients. Different flow patterns are observed in heavy ion collisions and classified in terms of their
azimuthal angle ϕ dependence with respect to the reaction plane, i. e. the plane spanned by the beam
direction and the impact parameter. The isotropic (or angle averaged) component is called radial flow.
In the framework of hydrodynamic models, the fluid properties (viscosity, equation-of-state, speed of
sound, ..) together with boundary conditions both in the initial state (collision geometry, pressure gradi-
ents, ..) and in the final state (freeze-out conditions) determine the pattern of collective motions and the
resulting momentum spectra d2N/dpT dϕ .
Fig. 2a shows the transverse momentum distributions of identified particles in central Pb+Pb collisions
at the LHC [47, 48]. The spectral shapes differ significantly from both pp at LHC and Au+Au at RHIC,
most dramatically for protons at low pT . The characteristic mass-dependent blue shift generated by the
radial flow leads to a strong depletion at low pT – the proton spectrum is almost flat between 0.5 and 1.5
GeV/c – and to a harder spectrum at high pT , where the p/pi ratio reaches a value of close to unity around
3 GeV/c.
The data are compared to hydrodynamic calculations, which are normalised to the data individually for
First Results from Pb+Pb collisions at the LHC 5
 
(G
eV
/c)
Tp
0
1
2
3
4
dy (GeV/c)
T
dN/dp
-
2
10
-
1
10
1102 10
3
10
A
LI
CE
 p
re
lim
in
ar
y
Pb
Pb
, 
2.
76
 
Te
V
+
pi
+
K p Ξ Ω
VI
SH
2+
1
VI
SH
NU
H
yd
ro
 n
or
m
al
ize
d 
to
 
<
 3
 G
eV
/c
T
da
ta
 fo
r p
>
 (c
)
β
Fl
ow
 v
el
oc
ity
 <
0.
2
0.
3
0.
4
0.
5
0.
6
0.
7
 (GeV)
fo
Temperature T
0
0.
050.
1
0.
15
ST
A
R
, 
K
, p
pi
A
u
-
A
u
, 
0.
2 
Te
V,
 
 
A
LI
CE
 p
re
lim
in
ar
y , 
K
, p
 
pi
Pb
-
Pb
, 
2.
76
 
Te
V,
 
pe
rip
he
ra
l
ce
n
tra
l
Fig. 2: a): Transverse momentum spectra of identified particles for central Pb+Pb collisions. The lines are
scaled results from a boost invariant hydrodynamic model with (VISHNU, dashed lines) and without (VISH2+1,
full lines) rescattering in the hadronic phase. b): Kinetic freeze-out temperatures Tf o and average radial flow
velocities 〈β 〉 extracted from identified particle spectra (pi , K, p) at LHC (Pb+Pb) and RHIC (Au+Au) for different
centralities.
each particle species in order to better compare the shapes (absolute particle yields and ratios, which
are an external input to the hydro models, are discussed below). The dashed line (VISH2+1) is a pure
hydro calculation [49, 50, 51] for Ξ and Ω (0-20% centrality); the full line (VISHNU) for pi , K, p (0-
5% centrality) includes additional final-state rescattering, calculated with the URQMD transport code,
which increases the radial flow velocity and improves the agreement with the data [52]. The hydro model
describes the data fairly well up to intermediate momenta. Around 2 GeV/c, first pions and later protons
start to deviate from the exponential hydro slope, indicating a progressive decoupling of high momentum
particles from the thermalized and flowing bulk matter.
The radial flow velocity can be estimated by fitting the spectra with a hydrodynamically inspired function
called a “blastwave” fit [53]. The resulting kinetic freeze-out temperatures Tf o and average radial flow
velocities 〈β 〉 are shown in Fig. 2b for different centrality selections [47]. Quantitative results from such
fits have to be taken with caution, as they in general depend on the particle types and momentum ranges
included in the fit. The two fit parameters (Tf o,〈β 〉) are strongly correlated as indicated by the confidence
contours in Fig. 3a (each confidence contour corresponds to a different centrality class). However, a
relative comparison to a similar fit to RHIC data [30] shows for the most central collisions that the
average flow velocity increases significantly at LHC, reaching about 0.65c – the leading edge of the
fireball expands collectively at essentially the speed of light – and that the kinetic freeze-out temperature
drops below the one at RHIC.
Fig. 3a shows the ratio of Λ/K0S as a function of pT for central and peripheral nucleus-nucleus collisions
at LHC and RHIC as well as pp at
√
s = 7 TeV [54]. Like the p/pi ratio mentioned above, the Λ/K0S ratio
is strongly enhanced compared to pp, reaching a maximum of≈ 1.5 around 3 GeV/c. Both the maximum
value as well as its position in pT change only modestly compared to RHIC, where these unusual baryon
to meson ratios had been observed previously and were referred to as “baryon anomaly” [55, 56, 57].
However, unlike at RHIC, the enhancement for central collisions of roughly a factor of three relative to
the pp value persists out to at least 6 GeV/c. Qualitatively such an enhancement is expected from radial
flow (the heavier baryons are pushed further out in pT than the lighter mesons), as shown by the full line,
which represents the Λ/K0S extracted from the hydro model VISH2+1 already shown in Fig. 2a. Quan-
titatively, however, the measured shape is very different indicating a progressive decoupling of hadrons
from pure hydrodynamic flow above 2 GeV/c (see Fig. 2a). Global observables like momentum spec-
tra, particle ratios and anisotropic flow seem to be governed by collective and thermal processes below
pT < 2−3 GeV/c, whereas hard processes and fragmentation dominate above pT > 6−10 GeV/c. In the
6 
(G
eV
/c)
Tp
0
2
4
6
8
0
S
 / K Λ
0
0.
51
1.
5
Λ/
Λ
 
×
ST
A
R
: 
A
u
-
A
u
 
20
0 
G
eV
 
w
ith
 1
0%
 fe
ed
-d
ow
n 
co
rre
ct
io
n
A
LI
CE
 p
re
lim
in
ar
y:
 
=
 
2.
76
 
Te
V
N
N
s
Pb
-P
b 
st
at
. e
rro
rs
 o
nl
y
sy
st
. e
rro
r ~
10
 %
 
A
LI
CE
, 
|y|
<
0.
75
Pb
-P
b 
0-
5%
Pb
-P
b 
60
-8
0%
 
=
 
7 
Te
V
s
pp
 
ST
A
R
A
u-
A
u 
0-
5%
A
u-
A
u 
60
-8
0%
H
yd
ro
 V
IS
H2
+1
R
ec
om
bi
na
tio
n
x
 0
.8
5
Particle Ratio
-
4
10
-
3
10
-
2
10
-
1
10
=
2.
76
 
Te
V,
 
0-
20
%
 
ce
n
tr
al
ity
N
N
s
Pb
-P
b 
 
=
7 
Te
V
s
pp
  
 
=
 1
sγ
=
 
1 
M
eV
, 
bµ
Th
er
m
al
 
M
o
de
l T
=
16
4 
M
eV
, 
A
LI
CE
 p
re
lim
in
ar
y
+
pi/
+ K
-
pi/
- K
+
pi
p/
-
pi/p
+
pi/
- Ξ
-
pi/
+ Ξ
+
pi/
-
Ω
-
pi/
+
Ω
Fig. 3: a): The Λ/K0S ratio as a function of transverse momentum for two centrality selections in nuclear collisions
(Pb+Pb at LHC, Au+Au at RHIC) as well as pp collisions at 7 TeV. The full and dashed lines show the correspond-
ing ratios from a hydro and a recombination model (scaled by 0.85), respectively. b) Particle ratios measured in
central Pb+Pb (squares) and pp (triangles) collisions at the LHC. The full lines are the predictions from a thermal
model.
intermediate momentum region, various mechanisms have been proposed to describe the interplay be-
tween soft and hard processes, including coalescence (recombination) of constituent quarks [58], strong
colour fields with enhanced baryon transport [59], and colour transparency [60]. The dashed line shows
Λ/K0S calculated in a recombination model [58] where radial flow and jet-quenching have been adjusted
to the measured data at LHC; it describes the shape of the distribution above ≈ 3 GeV/c very well but is
slightly off in normalisation (the model in the figure has been scaled by 0.85). A quantitative understand-
ing at intermediate momenta remains a challenge and will require additional experimental constraints, in
particular high precision data on identified particles out to at least 10 GeV/c.
3.3 Identified particle yields
Bulk particle production can be very successfully described in the framework of the thermal (statistical)
hadronisation model [42, 43]. It assumes that particles are created in thermal (phase space) equilibrium
governed by a scale parameter T, interpreted as a temperature. Production of a particle with mass m is
suppressed by a Boltzmann factor e−m/T . Conservation laws introduce additional constraints, like the
baryochemical potential µB which accounts for baryon number conservation. An additional parameter
γs is introduced to describe the observation that in some collision systems particles containing strange
quarks are suppressed compared to the grand canonical thermal expectation. The temperature parameter
is found in all high energy collisions (pp, e+e−, A+A) to be about 160 - 170 MeV, while γs increases
from 0.5−0.6 in pp to 0.9−1 in A+A [61, 62]. The disappearance of strangeness suppression in nuclear
collisions [63, 64], usually called strangeness enhancement, was one of the first signals predicted for the
QGP [65], and the fact that the bulk of all particles are produced in heavy ion reactions with thermal ratios
to very good approximation (typically < 10-20%) is considered to be an essential and well established
fact.
Fig. 3b shows a number of particle ratios from central Pb+Pb collisions, together with the (particle anti-
particle averaged) values measured in pp at 7 TeV [47, 66], as well as the prediction from a thermal model
using the canonical conditions expected for nuclear collisions at the LHC [67]. Within experimental
errors, particles and anti-particles are produced in Pb+Pb at midrapidity in equal numbers, consistent with
the very small value of the baryochemical potential µB of the thermal model. Strangeness is enhanced in
Pb+Pb compared to pp, by a factor which increases with strangeness content from about 1.3 in K/pi to up
to > 3 in Ω/pi (note that the pp data in Fig. 3b correspond to a higher energy than the Pb+Pb data). All
strange particle ratios in Pb+Pb are very well described by the model, implying that they are produced in
accordance with fully thermal ratios (i.e. γS = 1). However, in a complete surprise, protons were found
First Results from Pb+Pb collisions at the LHC 7
〉 η/d
ch
dN〈
0 500 1000 1500 2000
)3
 
(fm
lo
ng
R
si
de
R
o
u
t
 
R
3/
2 )
pi(2
0
1000
2000
3000
4000
5000
6000
E895 2.7, 3.3, 3.8, 4.3 GeV
NA49 8.7, 12.5, 17.3 GeV
CERES 17.3 GeV
STAR 62.4, 200 GeV
PHOBOS 62.4, 200 GeV
ALICE 2760 GeV
1/3〉 η/d
ch
dN〈
0 2 4 6 8 10 12 14
 
(fm
/c)
fτ
0
2
4
6
8
10
12 E895 2.7, 3.3, 3.8, 4.3 GeV
NA49 8.7, 12.5, 17.3 GeV
CERES 17.3 GeV
STAR 62.4, 200 GeV
PHOBOS 62.4, 200 GeV
ALICE 2760 GeV
Fig. 4: a): Local freeze-out volume as measured by identical pion interferometry at LHC compared to central
gold and lead collisions at lower energies. b): The system lifetime (decoupling time) τ f compared to results from
lower energies.
to be strikingly off – too low by a factor of about 1.5 – and well outside the usual precision of the thermal
model. The p/pi ratio, which would have been expected to increase from its value measured in pp, stays
essentially unchanged or even decreases slightly.
Before concluding that something is wrong or missing in current implementations of the thermal model,
the data, which are still preliminary, and various corrections (e.g. feed-down from weak decays) have to
be thoroughly checked. Also the proton ratios at RHIC should be revisited, as there are indications of a
smaller, but still significant, tension between model fits and data [68]. If the anomaly persists and cannot
be described with extensions to the standard thermal model [69, 70], it may be due to inelastic processes
which can change the particle composition after hadronisation. This mechanism is already thought to
be responsible for the abundance of resonances with a large width and therefore a large hadronic cross
section (like ρ and K∗), but hard to describe quantitatively because of partially unknown cross sections
and the difficulty to account for multi-particle initial states in hadronic transport codes.
4 Particle correlations
4.1 Identical particle (HBT) correlations
The freeze-out volume (the size of the matter at the time when strong interactions cease) and the total
lifetime of the created system (the time between collision and freeze-out) can be measured by identi-
cal particle interferometry (also called Hanbury-Brown–Twiss or HBT correlations) [71]. For identical
bosons (fermions), quantum statistics leads to an enhancement (depletion) for particles emitted close-by
in phase space. This modifies the two-particle correlation function, measured in energy and momen-
tum variables, and can be related via a Fourier transformation to the space and time distribution of the
emitting source, i.e. the space-time hyper surface of last rescattering.
Results from HBT correlation measurements are shown in Fig. 4 for central collisions from very low
energies up to LHC as a function of the charged particle density dNch/dη [72]. The total freeze-out vol-
ume is given as the product of a geometrical factor and the radii measured in three orthogonal directions
(called Rlong,Rside, and Rout), whereas the lifetime was estimated from the pair-momentum dependence
of Rlong. The locally comoving freeze-out volume is directly proportional to the particle multiplicity
(Fig. 4a) and therefore increases by a factor two compared to top RHIC energy to about 5000 fm3. The
system lifetime is proportional to the cube root of the particle density (Fig. 4b) and increases by about
30% to 10 fm/c. Incidentally, the freeze-out volume roughly extrapolates at low beam energy to the
volume of a Pb nucleus (≈ 800 fm3) and the lifetime vanishes at zero particle density. The evolution
from RHIC to LHC of the individual radius parameters (Rlong,Rside,Rout) as well as their pair momen-
8tum dependence is in satisfactory agreement with the predictions of hydrodynamical models [21, 72].
The stronger transverse flow generated during the hydrodynamic phase at LHC energy reduces the im-
portance of pre-equilibrium flow, in comparison with RHIC, and thus makes the agreement with the
hydrodynamical predictions less dependent on extraneous assumptions [73].
4.2 Anisotropic Flow
4.2.1 General considerations
The nuclear overlap zone in collisions with non-zero impact parameter is not azimuthally symmetric but
has an almond shape whose deformation changes with centrality. Consequently, the pressure gradients
between the centre of the overlap zone and its periphery in an average collision vary with azimuth, being
strongest in the direction of the reaction plane angle ΨRP, which coincides with the direction of the
minor axis of the almond. The developing collective flow is proportional to the pressure gradient and
therefore strongest towards the reaction plane, leading to an anisotropic distribution dN/dϕ of particles.
Anisotropic particle distributions were first suggested in [74] as a signal of collective flow in ultra-
relativistic heavy ion collisions and the flow pattern is usually quantified via a Fourier expansion [75]:
E
d3N
d3 p
=
1
2pi
d2N
pT dpT dy
(
1+2
∞
∑
n=1
vn cos [n(ϕ−Ψn)]
)
(2)
The Fourier (or flow) coefficients vn depend on pT and are given by vn(pT ) = 〈cos [n(ϕ−Ψn)]〉, where
the brackets denote an average over particles in a given pT bin and over events in a given centrality class.
In the above equations, n is the order of the harmonic, ϕ is the azimuthal angle of the particle, and Ψn
is the angle of the spatial plane of symmetry of harmonic n, the plane which maximises the expectation
value of vn in each event. Note that in general the symmetry angles Ψn can be independent from each
other and point towards different directions in each individual event.
Several experimental methods exist to measure the symmetry plane angles – using the phi-asymmetry
generated by the flow itself to find the directions Ψn event-by-event – and the coefficients vn (which
are usually averaged over events), via two- or many-particle correlation measures [76, 45, 46]. The v1
coefficient is called directed flow. It is most prominent near beam rapidity but is not discussed further
here (see [77]). Most attention was given in the past to v2, the elliptic flow, which is very strong in
non-central collisions and directly linked to the almond shape overlap zone (Ψ2 ≈ ΨRP). Higher-order
harmonics have usually been neglected because they were expected to be small for symmetry reasons
(this assumption turned out to be surprisingly poor because of fluctuations, see below).
The elliptic flow magnitude increases continuously with
√
s from SPS to RHIC [45, 46]. At top RHIC
energy, v2 reaches a value compatible with the one predicted by hydrodynamics for a “perfect fluid”, i.e.
a fluid without internal friction and vanishing shear viscosity [27, 28, 29, 30, 10]. The shear viscosity,
usually quoted in units of η/s (shear viscosity η over entropy density s), for a good relativistic quantum
fluid is of order h¯/kB. Using gauge gravity duality (conformal field theory in Anti de Sitter space, or
AdS/CFT) as a proxy for strongly coupled QCD, it has been conjectured that a lower bound exists on
η/s, i.e. 4piη/s≥ 1 (h¯ = kB = 1) [78]. This value is reached when the ’t Hooft coupling tends to infinity,
and the mean free path approaches the quantum limit, the Compton wavelength. In such a “perfect” fluid,
pressure gradients are not dissipated away because momentum transport perpendicular to the gradient is
restricted by the ultra-short mean free path.
Bounds on the η/s ratio can be derived from the data in two different ways. Either one compares the mo-
mentum dependent elliptic flow parameter v2(pT ) with results from viscous hydrodynamics calculations,
or one extracts a value of η/s by fitting the centrality (i. e. transverse size) dependence of the average pT
integrated elliptic flow. Both approaches yield a η/s ratio at RHIC that is very close to the lower bound,
First Results from Pb+Pb collisions at the LHC 9
 
(G
eV
/c)
Tp
0
5
10
15
20
n
Flow coefficient v
0
0.
050.
1
0.
150.
2
0.
25
Pb
-
Pb
 
2.
76
 
Te
V 
Ce
n
tr
al
ity
 
30
-
40
%
 
AL
IC
E 
pr
e
lim
in
a
ry
{E
P}
2
v
 
CM
S
{2P
}
2
v
 
AT
LA
S
{E
P}
2
v
>
1.
0) 
pre
l.
η|∆
 
AL
IC
E 
(|
{S
P}
3
v
 
CM
S
{2P
}
3
v
 
AT
LA
S 
pr
e
lim
in
a
ry
{E
P}
3
v H
yd
ro
 (G
lau
be
r)
/s
 =
 0
.0
)
η
 
(
2
v
/s
 =
 0
.0
8)
η
 
(
2
v
/s
 =
 0
.0
)
η
 
(
3
v
/s
 =
 0
.0
8)
η
 
(
3
v
Ce
nt
ra
lit
y 
pe
rc
en
til
e
0
20
40
60
80
n
Flow coefficient v
0
0.
050.
1
A
LI
CE
 
pr
el
im
in
a
ry
 
Pb
-
Pb
 
2.
76
 
Te
V
 
>
 1
}
η∆
{2,
 
2v
 
>
 1
}
η∆
{2,
 
3v
 
H
yd
ro
 V
IS
H2
+1
/s
=0
.0
8
η
 
G
la
ub
er
 
3
,
v
2v
/s
=0
.1
6
η
 
CG
C 
3
,
v
2v
Fig. 5: a): Elliptic flow coefficients v2 and v3 as a function of transverse momentum compared to a hydro
model [81] using two different values for the viscosity η/s b): Integrated elliptic flow coefficients as a function of
collision centrality compared to a hydro calculation [82] using two different models for the initial state geometry.
at most 3-5 times 1/4pi [79, 80]. The most advanced model–data comparison using the second method
gives an even tighter bound: 4piη/s≤ 2.5 [17].
Assuming no or only small changes in the transport properties of the quark-gluon plasma between RHIC
and LHC, hydrodynamic models predicted that the (particle type averaged) elliptic flow coefficient v2,
measured as a function of pT , should change very little, whereas the momentum integrated flow values
should either stay approximately constant or rise by at most 30% [21]. This prediction was quickly
confirmed [83]: the pT differential flow for charged particles is essentially unchanged at LHC whereas
the pT integrated flow increases by some 30%, a direct consequence of the stronger radial flow discussed
above. The matter created at LHC therefore still behaves like the (almost) perfect liquid discovered at
RHIC.
In order to make progress in the precision determination of η/s, some major obstacles have to be over-
come. An example on the experimental side are non-flow correlations introduced e.g. by jets and reso-
nance decays. On the theory side, uncertainties arise from model assumptions made about the initial state
pressure gradients and their event-by-event fluctuations. It was in fact only recently realized [84, 85] that
the statistical nature of individual nucleon-nucleon collisions can lead to highly irregular shapes of the
reaction zone and the corresponding initial energy/pressure distributions. These shapes fluctuate from
one event to the next, even at a fixed impact parameter. The irregular pressure gradients show no symme-
try with respect to the reaction plane and therefore induce higher harmonic flow patterns. Fluctuations
also tilt the symmetry angle of v2 away from the geometrical reaction plane (Ψ2 6= ΨRP), leading to
event-by-event fluctuations of the elliptic flow direction and magnitude.
4.2.2 Charged particle anisotropic flow
The elliptic (v2) and triangular (v3) flow coefficients for mid-central Pb-Pb collisions are shown in Fig. 5a
as functions of pT . The agreement between the LHC experiments [86, 87, 88, 89, 90, 91] is again
remarkable. Making use of the large rapidity acceptance of the LHC detectors, the coefficients vn were
measured by methods that minimize the influence of non-flow correlations, which are dominantly short-
range in rapidity (2P = trigger di-hardon correlations, EP = Event Plane, SP = two-particle cumulant with
eta-gap; see [45]). The elliptic flow coefficient rises approximately linearly with pT to a maximum of
v2 ≈ 0.23 around 3 GeV/c. The corresponding asymmetry in dN/dϕ is very large indeed: almost three
times as many 3 GeV/c particles are emitted in-plane compared to out-of-plane (see eq 2). The coefficient
then decreases, at first rapidly then more gradually, but stays finite out to the highest pT measured. Also
the triangular flow coefficient is very significant out to about 10 GeV/c and similar in shape, reaching
about half the value of v2 at the maximum.
10
 
[G
eV
/c]
Tp
0
1
2
3
4
2
Flow coefficient v
0
0.
1
0.
2
0.
3
A
LI
CE
 p
re
lim
in
ar
y
 
=
 
2.
76
 
Te
V
N
N
s
Pb
-P
b 
Ce
nt
ra
lit
y 
20
-4
0%
pi K p Ξ ΩAL
IC
E 
pr
el
im
in
ar
y
 
=
 
2.
76
 
Te
V
N
N
s
Pb
-P
b 
Ce
nt
ra
lit
y 
20
-4
0%
pi K p Ξ Ω
H
yd
ro /
s=
0.
20
η
CG
C,
 
VI
SH
NU
   
VI
SH
2+
1
pi K p
Ξ Ω
 
[G
eV
]
q
)/n 0
-
m
T
(m
0
0.
5
1
1.
5
q
/n
2
v
0
0.
050.
1
A
LI
CE
 p
re
lim
in
ar
y
 
=
 
2.
76
 
Te
V
N
N
s
Pb
-P
b 
Ce
nt
ra
lit
y 
20
-4
0%
pi K p Ξ Ω
Fig. 6: a): Elliptic flow v2 as a function of pT for identified particles compared with a hydrodynamical model [94,
95]. b): Elliptic flow for mesons and baryons, scaled by the respective number of valence quarks nq, versus scaled
transverse kinetic energy.
At high pT , where spectral shapes and relative particle yields indicate that hadrons are no longer in
local thermal equilibrium, the nonzero v2 is thought to arise from differential parton energy loss rather
than collective flow. The energy loss should depend on the path length inside the matter and therefore
correlate naturally with the orientation of the elongated reaction zone.
At lower pT , the data are compared to a hydrodynamic calculation [81] using two different values of the
shear viscosity(4piη/s = 0 and 1). The effect of a nonvanishing viscosity is clearly visible: The initial
pressure gradients are dissipated and the resulting collective flow is reduced. The influence of viscosity is
seen to be much stronger on v3, as expected. The triangularity corresponds to shorter wavelength density
variations and therefore larger local pressure gradients, which are more sensitive to viscosity. While both
values of the viscosity are compatible with the v2 results, only 4piη/s = 1 describes v3. A larger value of
the viscosity would fall significantly below the v3 data, but could in principle be accommodated by using
a different model for the initial state which generates a larger pressure gradient.
This is illustrated in Fig. 5b, which shows the pT integrated flow as a function of centrality [86]. The el-
liptic flow rises strongly towards peripheral collisions, following the increasing elongation of the reaction
region. Because of event-by-event fluctuations v2 stays non-zero even for the most central collisions. The
triangular flow shows little centrality dependence, as it is driven solely by shape fluctuations. The hydro-
dynamic model [82] shown in Fig. 5b employs two different, frequently used sets of initial conditions.
The energy density distribution in the transverse plane is taken either from a Glauber calculation (solid
line) combined with 4piη/s = 1, or from a saturation model (dashed line) combined with 4piη/s = 2. The
latter predicts larger pressure gradients and consequently needs a larger viscosity to describe the v2 data.
However, the saturation model significantly underpredicts v3. Glauber initial conditions can describe
both v2 and v3 better, in particular, if flow fluctuations are taken into account (see [82] for details).
The measurement of higher harmonics thus seems to be able to overcome the degeneracy between shear
viscosity and initial conditions and give a tighter limit of 4piη/s ≤ 2 [82]. Recent measurements of
higher harmonic flow at RHIC have led to a similar conclusion [92, 93]. However, none of the currently
used initial state models can perfectly describe all experimental flow observations. With the on-going
effort in both experiment and theory, the new high precision flow data from LHC promise significant
further progress toward the goal of a precision measurement of the viscosity and other properties of the
quark-gluon plasma.
The interference between v2 and higher harmonic flow coefficients [86, 96, 91, 97, 98], which are signif-
icant up to n = 6, also gives a most natural explanation for some unusual long range structures observed
in two-particle correlation at RHIC [84], traditionally called the “soft near-side ridge” and the “away-
First Results from Pb+Pb collisions at the LHC 11
side cone”. The interpretation of these structures was controversial [99] until higher harmonic flow was
firmly established at LHC and simultaneously at RHIC.
4.2.3 Identified particle flow
As for radial flow, a most stringent test of the collective flow interpretation of azimuthal anisotropies is
the characteristic dependence on particle mass. The LHC data for v2 for various particles (pi , K, p, Ξ,Ω)
are shown in Fig. 6a [100, 101]. To understand the origin of the characteristic mass splitting, seen also in
v3 [100], one needs to keep in mind that collective radial flow tends to equalise the velocities of particles,
not their momenta, and therefore shifts heavy particles out to higher pT than light ones [102]. The effect
of a given azimuthal flow asymmetry thus manifests itself at higher momenta for particles with a larger
mass. The hydrodynamical model [95, 94], which incorporates this effect, describes the data very well
for all particle species up to intermediate pT . It also predicted the observation that the mass splitting is
larger at LHC than at RHIC as a consequence of the increased radial flow.
The mesons (pi , K) deviate from the predicted flow above 1.5 GeV/c, whereas the baryons follow the
hydrodynamical curves, within the still large experimental errors, out to about 3 GeV/c. The different
behaviour of mesons and baryons, also seen in RHIC data [27, 28, 29, 30, 10], has been interpreted
as a sign of quark recombination or coalescence [103, 104, 105, 106], already mentioned above as a
possible reason for the large baryon-to-meson ratios. In this picture, hadrons at intermediate pT of a few
GeV/c are formed by coalescence of constituent quarks from deconfined quark matter, with the three-
quark baryons receiving a proportionally larger share of the collective quark flow than the two-quark
mesons. Indeed, RHIC data indicate a universal scaling of flow [107] for all particles of mass m when
plotting v2/nq versus (mT −m)/nq), where nq is the number of constituent quarks (nq = 2 or 3) and
(mT −m) is the transverse kinetic energy (mT =
√
p2T +m2).
The same scaled distribution is shown in Fig. 6b [100, 101] for LHC data. No scaling is apparent
for pT < 1.5 GeV/c, where all particles follow more closely the mass dependent pattern predicted by
hydrodynamics (see Fig. 6a). It therefore seems that the empirically observed low pT “quark scaling”
at RHIC is a coincidence, only approximately valid for a particular value of the radial flow and the
corresponding mass splitting. At higher pT , quark scaling between mesons and baryons may hold at
LHC within the current experimental errors, but is definitely less convincing so far. Reduced errors and
a measurement of flow for the φ , a meson with a mass comparable to the proton, are required before a
strong case for the coalescence picture can be made with LHC flow results.
5 Hard processes at high pT and high mass
The increased energy of heavy ion collisions at LHC relative to RHIC leads to much larger cross sections
for hard processes, i.e. those involving high momentum or high mass scales. Energetic quarks or gluons
can be observed as jets or single particles with pT reaching 100 GeV/c and beyond. Similarly, high
pT photons, charmonium and bottonium states (i.e. the J/ψ and ϒ families), and even the weak vector
bosons W and Z are copiously produced. The details of production and propagation of these high pT
probes can be used to explore the mechanisms of parton energy loss and deconfinement in the medium.
Some of the salient questions to be explored in this way are: How do colour charge and mass of a parton
influence the energy loss? The perturbative QCD formalism of radiative energy loss [108] predicts that
the energy loss of a parton should be proportional to the Casimir eigenvalue of its colour charge, i. e. that
gluons should lose energy at roughly twice the rate of quarks, and that the rate of energy loss of heavy
quarks should be reduced by the so-called dead-cone effect [109]. What is the relative contribution of
radiative and elastic mechanisms of energy loss? How does the energy loss depend on the thickness
of the medium? Strongly coupled gauge theories generally predict a stronger dependence on the path
length in the medium than perturbative QCD [110]. Where does the lost energy go? Does it remain
12
inside the jet cone and manifest itself as a modification of the fragmentation function, or does it get
rapidly thermalized in the medium, making the jet look like a jet in vacuum, but with reduced energy?
There are several experimental techniques available to explore these questions at the LHC. One is to
measure single particle cross sections and compare them to the equivalent cross sections in pp colli-
sions. The nuclear modification factor RAA of single particles, which has been extensively studied at
RHIC [27, 28, 29, 30], allows for direct comparison with the RHIC results and their extension to higher
momenta. The other approach is to reconstruct jets directly and compare jet yields in Pb+Pb with jet
production in pp, as well as to measure jet–jet and jet–charged particle energy and angular correlations.
The large kinematic range and the excellent calorimetry of the LHC detectors facilitate jet reconstruction,
especially for jet energies in excess of 50 GeV.
5.1 Single particle spectra
The single particle production rates at RHIC have shown a large suppression of hadrons in nuclear
collisions relative to pp, whereas particles that do not interact strongly, e. g. photons, are not modified.
The LHC can significantly extend the accessible pT range and allow the measurement of additional
particles, such as the Z and W . The suppression effects of a given particle are typically expressed in
terms of the nuclear modification ratio:
RAA(pT ) =
d2NAA/d pT dη
〈TAA 〉d2σNN/d pT dη , (3)
where NAA and σNN represent the particle yield in nucleus-nucleus collisions and the cross section in
nucleon-nucleon collisions, respectively. The nuclear overlap function 〈TAA〉 is the ratio of the number of
binary nucleon-nucleon collisions, 〈Ncoll〉, calculated from the Glauber model, and the inelastic nucleon-
nucleon cross section (σNNinel = (64± 5) mb at
√
sNN = 2.76 TeV). In the absence of nuclear effects the
factor RAA is unity by construction. As observed at RHIC in 200 GeV Au+Au collisions [29, 30], the
yield of 5−10 GeV/c charged particles is suppressed in the most central events by more than a factor of
five. Instead of RAA one can also approximate the centrality dependence by measuring RCP, the ratio of
central over peripheral events.
While the charged particle RAA is the measurement with best statistical and systematic precision [111,
112, 113, 114], it is also interesting to measure the nuclear suppression factor for individual particle
species to distinguish the exact mechanisms of energy loss. Measurements exist for identified pi , K0S,
Λ [112, 115], isolated photons [116], Z,W [117, 118, 119, 120], D-mesons [115], jets [114], J/ψ [121,
119, 122] and ϒ [122]. The suppression factors of prompt and non-prompt J/ψ mesons, the latter being
produced from B-mesons and identified by their displaced decay vertex, were measured separately. In
addition, ATLAS has presented results for the relative suppression of charged particles in central versus
peripheral events, RCP [123].
A summary of RAA measurements for different particle species is shown in Fig. 7 for the most central
events. The inclusive charged particle RAA follows, up to about 10−15 GeV/c, the characteristic shape
discovered at RHIC (left panel, full circles). The pronounced maximum at a few GeV/c, which is some-
times attributed to initial or final state interactions in nuclei (“Cronin effect”), is at very high energies
more likely to be yet another manifestation of collective flow. It is qualitatively described by the dashed
line, which shows the RAA obtained by dividing the inclusive charged particle distribution calculated by
viscous hydrodynamics [50] for central Pb+Pb by the experimentally measured pp spectrum. This inter-
pretation is also supported by the fact that the apparent “suppression” factor is slightly larger for kaons
and significantly larger for the Λ, as expected from flow. The peak region is followed by a steep decline
and a minimum, around 5−7 GeV/c, where the suppression reaches a factor of about seven, very similar
to but slightly larger than the one measured at RHIC.
Heavy quarks, as shown by the RAA of prompt D mesons (open squares) and non-prompt J/ψ (from
First Results from Pb+Pb collisions at the LHC 13
 
(G
eV
/c)
Tp
0
5
10
15
20
AA
 R
0
0.
2
0.
4
0.
6
0.
81
1.
2
1.
4
1.
6
1.
82
 
=
 2
.
76
 
Te
V
N
N
s
Pb
Pb
 
Ch
ar
ge
d 
Pa
rti
cl
e 
0 
- 5
%
 (C
MS
)
Ch
ar
ge
d 
Pa
rti
cl
e 
0 
- 5
%
 (A
LIC
E)
 
fro
m
 B
 0
 - 
10
%
 (C
MS
)
ψ
J/
 
 
 
0 
- 2
0%
 (A
LIC
E)
0 D
 
 
 
0 
- 5
%
 (A
LIC
E)
0
K  
 
 
0 
- 5
%
 (A
LIC
E)
Λ
Th
eo
ry
Ch
.
 
ha
dr
o
n
 (Y
a
JE
M
-
D
)
Ch
.
 
ha
dr
o
n
 
(H
T-
M
)
Ch
. h
ad
ro
n 
(JE
W
EL
)
+
K+
p 
(V
is
hn
u
)
π Ch
.
 
ha
dr
o
n
 (H
T-
W
)
Ch
.
 
ha
dr
o
n
 
(G
LV
)
 
(G
eV
/c)
Tp
0
20
40
60
80
10
0
AA
 R
0
0.
2
0.
4
0.
6
0.
81
1.
2
1.
4
1.
6
1.
82
 
=
 
2.
76
 T
eV
N
N
s
Pb
Pb
 
Ch
ar
ge
d 
Pa
rti
cl
e 
0 
- 5
%
 (C
MS
)
Ch
ar
ge
d 
Pa
rti
cl
e 
0 
- 5
%
 (A
LIC
E)
Is
ol
at
ed
 P
ho
to
n 
0 
- 1
0%
 (C
MS
)
 
 
 
 
 
 
0 
- 1
0%
 (C
MS
)
0 Z
Th
eo
ry
Ch
. h
ad
ro
n
 (Y
a
JE
M
-
D
)
Ch
. h
ad
ro
n
 (H
T-
M
)
Ch
. h
ad
ro
n 
(JE
W
EL
)
Ch
. h
ad
ro
n
 (H
T-W
)
Ch
.
 
ha
dr
o
n
 
(G
LV
)
Fig. 7: Nuclear modification factor RAA as a function of pT for a variety of particle species together with theoret-
ical predictions. Experimental error bars correspond to the total error (statistical and systematic errors added in
quadrature). a) Low momentum region pT < 20 GeV; b) Entire momentum range measured at LHC. The curves
show the results of various QCD-based models of parton energy loss [124, 125, 126, 127, 128]. For details, see
text.
the decay of bottom quarks, closed diamond) in Fig. 7, are almost as strongly suppressed as inclusive
charged particles. A similar conclusion can be drawn from the measurement of leptons from heavy
flavour decays [115]. This seems contrary to the expectation that gluons, which are the dominant source
of inclusive charged particles at LHC, should suffer twice as much energy loss as light quarks and that, in
addition, the energy loss of heavy quarks should be even less than that of light quarks because of the mass
dependence of radiation (“dead-cone” effect [109]). The strong suppression found for hadrons containing
c- and b-quarks confirms observations made at RHIC and may indicate that the energy loss rate depends
less strongly on the parton mass than expected for radiative energy loss. Reasons for this behaviour
could be nonperturbatively large elastic energy loss in the strongly coupled quark-gluon plasma or heavy
meson formation within the medium [124]. More data and a quantitative comparison with models will be
required to see how the small, with current statistics not very significant, difference between light hadron
and heavy quark suppression can be accommodated by theory.
Above pT ≈ 8 GeV/c, the suppression becomes universal for all particle species (with the possible ex-
ception of the non-prompt J/ψoriginating from B-meson decays shown in the left panel). With increasing
pT , RAA rises gradually towards a value of 0.5 (see right panel), a feature which was not readily apparent
in the RHIC data. Isolated photons and the Z boson are not suppressed, within the currently still large
statistical errors. This finding is consistent with the hypothesis that the suppression observed for hadrons
is due to final-state interactions with the hot medium.
The observed rise of RAA with pT allows a better discrimination between competing models of energy
loss than the rather flat high pT dependence observed at RHIC. The rise can be understood as a decrease
of the parton fractional energy loss with increasing pT , reflecting the weak energy dependence of pQCD
radiative energy loss on parton energy. At RHIC this trend is compensated by the softening of the
underlying parton spectrum, whereas at LHC the spectrum stays hard up to the highest measured pT
which remains much farther away from the kinematic threshold than at RHIC.
The observed trend is semi-quantitatively described by several models implementing the perturbative
QCD (pQCD) formalism for energy loss [124, 125, 126, 127, 128]. The rate of induced gluon radiation
in pQCD is governed by the rate of transverse momentum broadening, encoded in the jet quenching
14
parameter qˆ, of the trajectory of a hard parton in the quark-gluon plasma [108]. The value of qˆ is
proportional to the density of gluons in the medium and grows approximately as the entropy density of
the plasma, or the charged particle multiplicity in the final state of the heavy ion collision. It should thus
be slightly more than twice as large in Pb+Pb collisions at LHC than at the top RHIC energy.
The band of curves labelled “GLV” [124] in Figure 7 shows results of a pQCD calculation using light-
cone wave functions for the produced hadrons and assuming static scattering centres corresponding to a
gluon density dNg/dy ≈ 2,800 and a QCD coupling constant αs ≈ 0.3. The curve labeled “YaJEM-D”
[125] is a Monte-Carlo implementation of pQCD jet quenching that accounts for the limited virtuality
evolution of the jet in a finite medium. The curves labelled “HT-W” [126] are based on the higher-
twist formalism and correspond to values of qˆ0τ0 = 1.0− 1.4 GeV2 at the time of thermalization τ0.
The curve labelled “HT-M” [127] is a different implementation of the higher-twist formalism with the
parameter choice qˆ0 = 1.3 GeV2/fm at T0 = 344 MeV. The “JEWEL” model [128] uses a Monte-Carlo
implementation of an in-medium parton shower including LPM radiation suppression.
The various calculations apply models of the nuclear reaction that are of widely varying detail. The
GLV model uses a schematic description of the integrated medium density based on the boost-invariant
Bjorken model of hydrodynamics. The YaJEM-D model combines ideal two-dimensional hydrodynam-
ics with a saturation model for the initial energy density, but has not adjusted the parameters to the
measured global properties of Pb+Pb collisions at LHC. The HT-W and HT-M calculations use three-
dimensional ideal and two-dimensional viscous hydrodynamics, respectively, with initial conditions cho-
sen to reproduce the charged particle multiplicity measured at LHC. The JEWEL model uses a variant
of the schematic Bjorken hydrodynamical model.
It is an important theoretical challenge to identify the origin of the differences between the five calcu-
lations, in particular, the much more rapid rise of RAA in the YaJEM-D and HT-M models. Systematic
studies of the differences between various implementations of radiative parton energy loss in pQCD
at RHIC energies have revealed large sensitivities to poorly controlled aspects of lowest-order, collinear
gluon radiation. These differences can be further enhanced by the use of different models for the medium
evolution. Nevertheless, the quantitative success of several models in correctly reproducing the rise and
saturation of RAA with pT suggests that the energy loss of the leading parton in a jet shower may be
described by perturbative QCD radiation in a strongly coupled medium.
5.2 Jets
Studying the modification of fully reconstructed jets is a particularly useful tool for probing the properties
of the hot quark-gluon plasma. Jets are formed by fragmentation from high pT partons as they propagate
through the produced matter. Measuring the energy of fully reconstructed jets allows one to distinguish
between energy redistribution among the leading parton and the remainder of the jet and energy dissipa-
tion out of the jet into the thermal medium. One of the promising channels are dijets, in particular their
transverse energy balance and azimuthal angle correlation. The energy dissipation into the medium can
be studied by measuring the asymmetry in pT of dijets in heavy ion collisions as a function of centrality
and by comparing them to data from pp collisions. Similarly, the measurement of the relative azimuthal
angle of the two jet axes gives information about the degree of scattering of partons as they traverse the
medium.
The measurement of the dijet asymmetry AJ = (pT 1− pT 2)/(pT 1 + pT 2), where “1” and “2” refer to the
leading and subleading jet, respectively, was performed by both ATLAS [129] and CMS [130]. Events
containing at least two jets, with the leading (sub-leading) jet having pT of at least 120 (50) GeV/c for
CMS and at least 100 (25) GeV/c for ATLAS, were selected for further study.
The most striking observation by both experiments is the large centrality-dependent increase of the im-
balance in the energy of the two jets, as measured in the calorimeters (See Fig. 8). While their energies
First Results from Pb+Pb collisions at the LHC 15
J
A
0
0.
2
0.
4
0.
6
0.
8
1
J
) dN/dA
evt
(1/N
01234
40
-1
00
%
J
A
0
0.
2
0.
4
0.
6
0.
8
1
J
) dN/dA
evt
(1/N
01234
20
-4
0%
J
A
0
0.
2
0.
4
0.
6
0.
8
1
J
) dN/dA
evt
(1/N
01234
10
-2
0%
J
A
0
0.
2
0.
4
0.
6
0.
8
1
J
) dN/dA
evt
(1/N
01234
0-
10
%
A
TL
A
S
Pb
+P
b
=
2.
76
 T
eV
N
N
s
-
1
bµ
=
1.
7 
in
t
L
φ∆
2
2.
5
3
φ∆ ) dN/d
evt
(1/N
-
2
10
-
1
10
110
φ∆
2
2.
5
3
φ∆ ) dN/d
evt
(1/N
-
2
10
-
1
10
110
φ∆
2
2.
5
3
φ∆ ) dN/d
evt
(1/N
-
2
10
-
1
10
110
φ∆
2
2.
5
3
φ∆ ) dN/d
evt
(1/N
-
2
10
-
1
10
110
Pb
+P
b 
Da
ta
p+
p 
Da
ta
H
IJ
IN
G
+P
YT
HI
A
Fig. 8: Calorimetric jet imbalance in dijet events (top) and azimuthal angle between the leading and subleading
jets (bottom) as a function of collision centrality for pp and Pb+Pb collisions.
were very different, the two jets were observed to be very close to back-to-back in the azimuthal plane,
implying little or no angular scattering of the partons during their traversal of the medium [130] as shown
in Fig. 8 [131].
The distribution of particle momenta inside jets normalized to the jet energy is the same, within experi-
mental uncertainties, to that of jets produced in pp collisions as shown in Fig. 9 [132]. This suggests that
most of the additional energy radiated by the leading parton inside the medium gets absorbed by the mat-
ter, and the fragments observed within the jet cone are produced outside of the medium. Several model
calculations [133, 134, 135, 136], which combine elastic and inelastic parton energy loss with deflection
of radiated gluons by the medium, have been able to reproduce the increased energy asymmetry of dijets
in Pb+Pb.
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
ξ dN/d
-
4
10
-
3
10
-
2
10
-
1
10
1102 10
3
10
CM
S 
Pr
el
im
in
ar
y
-
1
bµ
L 
dt
 
=
 
7.
2 
∫
 
=
 
2.
76
 
Te
V
N
N
s
Pb
Pb
 
0-
30
%
<
0.
13
J
0<
A
CM
S 
Pr
el
im
in
ar
y
-
1
bµ
L 
dt
 
=
 
7.
2 
∫
 
=
 
2.
76
 
Te
V
N
N
s
Pb
Pb
 
0-
30
%
<
0.
13
J
0<
A
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
PbPb/pp
0
0.
51
1.
52
2.
53
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
ξ dN/d
-
4
10
-
3
10
-
2
10
-
1
10
1102 10
3
10
<
0.
24
J
0.
13
<A(R=
0.3
) P
Flo
w 
Je
ts 
wi
th 
PU
T
a
n
ti-
k
>
40
 G
eV
/c
Je
t 2
T
>
10
0 
G
eV
/c
, p
Je
t 1
Tp
pi
 
>
 2
/3
φ∆
<
0.
24
J
0.
13
<A(R=
0.3
) P
Flo
w 
Je
ts 
wi
th 
PU
T
a
n
ti-
k
>
40
 G
eV
/c
Je
t 2
T
>
10
0 
G
eV
/c
, p
Je
t 1
Tp
pi
 
>
 2
/3
φ∆
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
PbPb/pp
0
0.
51
1.
52
2.
53
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
ξ dN/d
-
4
10
-
3
10
-
2
10
-
1
10
1102 10
3
10
<
0.
35
J
0.
24
<A
R
 <
 0
.3
)
∆
Tr
a
ck
s 
in
 
co
n
e
 
(
>
4 
G
eV
/c
Tp
<
0.
35
J
0.
24
<A
R
 <
 0
.3
)
∆
Tr
a
ck
s 
in
 
co
n
e
 
(
>
4 
G
eV
/c
Tp
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
PbPb/pp
0
0.
51
1.
52
2.
53
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
ξ dN/d
-
4
10
-
3
10
-
2
10
-
1
10
1102 10
3
10
Le
ad
in
g 
jet
pp
 re
fe
re
nc
e
Su
bl
ea
di
ng
 je
t
pp
 re
fe
re
nc
e <
0.
70
J
0.
35
<A
Le
ad
in
g 
jet
pp
 re
fe
re
nc
e
Su
bl
ea
di
ng
 je
t
pp
 re
fe
re
nc
e <
0.
70
J
0.
35
<A
)
Tr
a
ck
T
/p
Je
t
T
 
=
 
ln
(p
ξ
0
0.
5
1
1.
5
2
2.
5
3
3.
5
4
4.
5
PbPb/pp
0
0.
51
1.
52
2.
53
Le
ad
in
g 
Je
t
Su
bl
ea
di
ng
 J
et
Le
ad
in
g 
Je
t
Su
bl
ea
di
ng
 J
et
Fig. 9: Ratio of Pb+Pb to pp jet fragmentation function as a function of jet asymmetry AJ for central Pb+Pb events.
16
The strong effect of jet quenching is confirmed by the ATLAS measurement of the jet RCP [131]. The
comparison of jet production cross section between the central and peripheral events indicates that for
the most central events the energy loss appears to reduce the cross section by about a factor of two (see
Fig. 10). This shows that the energy lost by the leading parton is not simply redistributed within the jet
cone but lost from the reconstructed jet.
 
 
[G
eV
]
TE
10
0
15
0
20
0
25
0
30
0
CP
R
0.
51
1.
5
Ce
nt
ra
lit
y 
 0
-1
0%
AT
LA
S
Pr
el
im
in
ar
y
 
 
[G
eV
]
TE
10
0
15
0
20
0
25
0
30
0
0.
51
1.
5
Ce
nt
ra
lit
y 
50
-6
0%
 
=
 
2.
76
 
Te
V
N
N
s
Pb
+P
b 
 
-
1
bµ
 
=
 
7 
in
t
L R
 =
 0
.4
Fig. 10: Jet central-to-peripheral event suppression factor RCP for different centralities.
To find the “missing energy”, the calorimetric measurement in CMS was complemented by a detailed
study of low pT charged particles in the tracker and by using missing pT techniques. The apparent
missing energy in highly unbalanced dijet events was found among the low pT particles, predominantly
with 0.5 < pT < 2 GeV/c, emitted outside of the sub-leading jet cone [130].
5.3 Quarkonium suppression
Heavy quarkonia are important probes of the quark gluon plasma since they are produced early in the
collision and their survival is affected by the surrounding medium [137]. The bound states of charm and
bottom quarks are predicted to be suppressed in heavy ion collisions in comparison with pp, primarily
as a consequence of deconfinment (“melting”) in the QGP [138]. The magnitude of the suppression for
different quarkonium states should depend on their binding energy, with strongly bound states such as
the ϒ showing less or no modification [139].
However, J/ψ production, the classical deconfinement signal, has confounded expectations and interpre-
tations ever since the first nuclear suppression was measured with Oxygen beams at the SPS [140]; now
attributed to cold nuclear matter effects rather than deconfinement [141]. The ”anomalous” suppression
seen later with heavier beams [142] turned out, surprisingly, to be rather similar in magnitude at SPS and
RHIC. This could indicate suppression of only the high mass charmonium states ψ’ and χc, which popu-
late about 40% of the observed J/ψ , and which should dissociate very close to or even below the critical
transition temperature. Alternatively, it has been suggested that the increasing (with energy) J/ψ suppres-
sion is more or less balanced by enhanced production via recombination of two independently produced
charm quarks [143, 144]. A resolution to this puzzle may come from measuring J/ψ production at the
LHC where coalescence effects should be stronger because of the abundant charm production, and by
comparing the suppression patterns of the J/ψ and ϒ families.
A compilation of first LHC results on quarkonia production [121, 122] for both J/ψ (a) and ϒ (b) is
First Results from Pb+Pb collisions at the LHC 17
pa
rt
N
0
50
10
0
15
0
20
0
25
0
30
0
35
0
40
0
AA
R
0
0.
2
0.
4
0.
6
0.
81
1.
2
1.
4
1.
6
1.
82
 
=
 
20
0 
G
eV
N
N
s
Au
Au
  
PH
EN
IX
: |y
|<0
.35
PH
EN
IX
: 1
.2
<|y
|<2
.2
 
>
 5
 G
eV
/c
, |y
|<1
.0
T
ST
AR
: 
p =
 
2.
76
 
Te
V
N
N
s
Pb
Pb
  
 
<
 3
0 
G
eV
/c
, |y
| <
 2.
4
T
CM
S:
 6
.5
 <
  p
AL
IC
E:
 2
.5
 <
 y
 <
 4
.0
pa
rt
N
0
50
10
0
15
0
20
0
25
0
30
0
35
0
40
0
AA
R
0
0.
2
0.
4
0.
6
0.
81
1.
2
1.
4
1.
6
1.
82
 
=
 
2.
76
 
Te
V
N
N
s
Pb
Pb
  
(1S
)
ϒ
CM
S:
 
|y|
<2
.4
 
=
 
20
0 
G
eV
N
N
s
Au
Au
  
(1S
+2
S+
3S
)
ϒ
ST
AR
: 
|y|
<0
.5
Fig. 11: Nuclear modification factor RAA as a function of centrality for J/ψ (right) and ϒ (left).
shown in Fig. 11 as a function of centrality (Npart), together with data from RHIC [145, 146, 147].
While errors are still large, and the overall amount of suppression at LHC remains qualitatively similar
to RHIC, the detailed pattern is quite different and intriguing. The pT integrated RAA measured for
the J/ψ at forward rapidity (closed circles) of about 0.5 depends very little on centrality and is almost
a factor of two larger than the one measured at RHIC in central collisions, also at forward rapidity
(open circles); the difference is smaller but still significant when comparing with RHIC midrapidity
data (open squares). On the contrary, the high pT data at LHC (full squares), which is compatible with
an independent RCP measurement [119], show a stronger suppression than the high pT RHIC results
(open stars). While such a pattern would be unexpected in a pure suppression scenario, it is qualitatively
consistent with the recombination model, which predicts substantial regeneration effects only at low
transverse momentum. The ϒ suppression (right panel) is very similar at RHIC and LHC. As only about
50% of the observed ϒ(1S) are directly produced, and the ϒ(2S/3S) states seem to be more suppressed
than the ground state [148], the measured RAA is compatible at both RHIC and LHC with suppression
of the high mass bottonium states only.
However, there are still a number of unknown aspects that must be clarified before firm conclusions can
be drawn. Foremost are cold nuclear matter effects, in particular shadowing or saturation of the nuclear
parton distribution functions, which for gluons are very poorly constrained at the small x values relevant
at LHC. Likewise, very little is known about final state absorption of quarkonia in nuclei, which may
(or may not) be important at LHC. For this reason a p+Pb run at LHC, which will address these nuclear
effects, is mandatory and anticipated for 2012.
6 Conclusions
The LHC has entered the field of ultra-relativistic heavy ion physics with an impressive first year per-
formance, having benefited from many years of preparation, a strong and very complementary set of
state-of-the-art detectors, and, crucially, decades of experience and progress made at previous machines.
Global event properties were found to evolve smoothly from RHIC to LHC, to be qualitatively similar
but quantitatively different, and to reflect the expected increases in energy density, volume, and lifetime
of the hot and dense matter created in the collisions. The data have generally confirmed that also at LHC
energy the bulk features imply the formation of a very strongly interacting, nearly perfect quark-gluon
plasma liquid. The prominent presence of various collective flow patterns has opened a path forward to
18
precision measurements of its material properties, including shear viscosity, equation-of-state, and sound
velocity.
Concerning bulk properties, and with the possible exception of thermal particle production, where the
unexpectedly low proton yield needs confirmation and clarification, the standard reaction model estab-
lished at RHIC has proven itself to be predictive and detailed enough to permit quantitative extrapolation
to much higher collision energies.
For observables where the energy reach of LHC is unique, i. e. low-x parton physics and hard processes,
already the first low luminosity run has revealed a wealth of significant, and at times surprising results.
The energy loss of high momentum partons traveling through the matter was measured with several dif-
ferent particle species as well as with fully reconstructed jets. While its magnitude is roughly in the
expected range, the more differential analysis of parton type and momentum dependence, dijet corre-
lations, and energy flow around the jet axis has revealed features which were not anticipated and may
spur substantial refinements of our understanding and modelling of parton interactions with the hot QCD
medium.
The results on quarkonium production, where precision measurements of the ϒ resonances are possible at
LHC, in addition to those of the charmonium states, have yet to provide decisive information on the sup-
pression mechanism. However, intriguing hints have been found in the charmonium and bottomonium
suppression pattern, suggesting significant differences to the observations at RHIC.
The near future of heavy ion running at LHC will bring a significant increase in integrated luminosity –
already the second heavy ion run in November 2011 has increased the statistics for hard probes by more
than an order of magnitude – as well as new signals like γ–jet and Z–jet correlations. The p+Pb run
planned for 2012 should not only deliver the much needed comparison data for the heavy ion program
but also far extend the kinematic reach for the study of initial-state effects in the nuclear parton wave
function, such as saturation and other low-x physics. And another factor of two increase in energy is in
sight when LHC will reach the design value after 2014.
The combination of the on-going heavy ion program at RHIC, which includes high-luminosity mea-
surements with upgraded detectors, the new higher energy data from LHC, and the demonstrated solid
foundation in theory promises rapid progress in a quantitative understanding of the properties of the
quark-gluon plasma.
Acknowledgments
The authors thank the ALICE, ATLAS and CMS collaborations for providing the preliminary data used
in this review; U. Heinz, C. Shen and R. J. Fries for the model calculations used in Figs. 2a, 3a 6a
(UH, CS) and Fig. 3a (RJF); M. Floris, M. Poghosyan and R. Shahoyan for help in preparing data and
figures in chapters 3 and 4; T. Dahms, Y. Kim, P. Steinberg and A. Yoon for preparing data and figures
in chapter 5; A. Majumder, T. Renk, I. Vitev and X.-N. Wang for providing data files with the results
of their calculations for Fig. 7; and M. Floris, P. Steinberg and S. Voloshin for careful reading of the
manuscript and helpful discussions. This work was supported in part by grants from the US DOE Office
of Science (DE-FG02-05ER41367 and DE-FG02-94ER40818).
References
[1] Evans L, Bryant P, JINST 3:S08001 (2008).
[2] ATLAS Collaboration, Aad G, et al., JINST 3:S08003 (2008).
[3] CMS Collaboration, Adolphi R, et al., JINST 3:S08004 (2008).
[4] ALICE Collaboration, Aamodt K, et al., JINST 3:S08002 (2008).
First Results from Pb+Pb collisions at the LHC 19
[5] Fabjan C, Schukraft J, (2011), 1101.1257.
[6] Aoki Y, Fodor Z, Katz S, Szabo K, Phys.Lett. B643:46 (2006), hep-lat/0609068.
[7] Borsanyi S, Endrodi G, Fodor Z, Jakovac A, Katz SD, et al., JHEP 1011:077 (2010), 1007.2580.
[8] Bazavov A, Bhattacharya T, Cheng M, DeTar C, Ding H, et al., (2011), 1111.1710.
[9] F. Antinori, A. Billmeier and J. Zaranek, AIP Conf. Proc. 631:294 (2003).
[10] Mu¨ller B, Nagle JL, Ann.Rev.Nucl.Part.Sci. 56:93 (2006), nucl-th/0602029.
[11] Banerjee D, Nayak JK, Venugopalan R, Lect.Notes Phys. 785:105 (2010), 0810.3553.
[12] Kovchegov YV, Nucl.Phys. A854:3 (2011), 1007.5021.
[13] Albacete JL, Dumitru A, Nara Y, J.Phys.Conf.Ser. 316:012011 (2011), 1106.0978.
[14] Lappi T, McLerran L, Nucl.Phys. A772:200 (2006), hep-ph/0602189.
[15] Romatschke P, Venugopalan R, Phys.Rev. D74:045011 (2006), hep-ph/0605045.
[16] Kurkela A, Moore GD, JHEP 1111:120 (2011), 1108.4684.
[17] Song H, Bass SA, Heinz U, Hirano T, Shen C, Phys.Rev.Lett. 106:192301 (2011), 1011.2783.
[18] Gyulassy M, McLerran L, Nucl.Phys. A750:30 (2005), nucl-th/0405013, RIKEN BNL Research
Center Scientific Articles, Vol.9.
[19] Bass SA, Gale C, Majumder A, Nonaka C, Qin GY, et al., Phys.Rev. C79:024901 (2009),
0808.0908.
[20] Majumder A, Mu¨ller B, Wang XN, Phys.Rev.Lett. 99:192301 (2007), hep-ph/0703082.
[21] Armesto N, et al., J.Phys.G 35:054001 (2008), 0711.0974.
[22] Satz H, Heavy ion physics at very high-energies, in ECFA Large Hadron Collider Workshop,
Aachen, Germany, 4-9 Oct 1990, volume 1, p. 188, 1990.
[23] Heinz U, Strangeness and antimatter production at the LHC, in ECFA Large Hadron Collider
Workshop, Aachen, Germany, 4-9 Oct 1990, volume 2, p. 1079, 1990.
[24] ALICE Collaboration, Aamodt K, et al., Phys.Rev.Lett. 105:252301 (2010), 1011.3916.
[25] Bjorken J, Phys.Rev. D27:140 (1983).
[26] CMS Collaboration, Krajczar K, J.Phys.G 38:124041 (2011), CMS-PAS-HIN-11-003.
[27] PHOBOS Collaboration, Back BB, et al., Nucl. Phys. A757:28 (2005), nucl-ex/0410022.
[28] BRAHMS Collaboration, Arsene I, et al., Nucl. Phys. A757:1 (2005), nucl-ex/0410020.
[29] PHENIX Collaboration, Adcox K, et al., Nucl. Phys. A757:184 (2005), nucl-ex/0410003.
[30] STAR Collaboration, Adams J, et al., Nucl. Phys. A757:102 (2005), nucl-ex/0501009.
[31] Miller ML, Reygers K, Sanders SJ, Steinberg P, Ann.Rev.Nucl.Part.Sci. 57:205 (2007), nucl-
ex/0701025.
[32] ALICE Collaboration, Aamodt K, et al., Phys.Rev.Lett. 106:032301 (2011), 1012.1657.
[33] ATLAS Collaboration, Aad G, et al., (2011), 1108.6027.
[34] CMS Collaboration, Chatrchyan S, et al., JHEP 1108:141 (2011), 1107.4800.
[35] PHENIX Collaboration, Adler S, et al., Phys.Rev. C71:034908 (2005), nucl-ex/0409015.
[36] STAR Collaboration, Abelev B, et al., Phys.Rev. C79:034909 (2009), 0808.2041.
[37] PHOBOS Collaboration, Alver B, et al., Phys.Rev. C83:024913 (2011), 1011.1940.
[38] Bopp FW, Engel R, Ranft J, Roesler S, (2007), 0706.3875.
[39] Deng WT, Wang XN, Xu R, Phys.Lett. B701:133 (2011), 1011.5907.
[40] Gelis F, Iancu E, Jalilian-Marian J, Venugopalan R, Ann.Rev.Nucl.Part.Sci. 60:463 (2010),
1002.0333.
[41] Albacete JL, Dumitru A, (2010), 1011.5161.
[42] Braun-Munzinger P, Redlich K, Stachel J, (2003), nucl-th/0304013, To appear in Quark Gluon
20
Plasma 3, eds. R.C. Hwa and Xin-Nian Wang, World Scientific Publishing.
[43] Becattini F, (2009), 0901.3643.
[44] Huovinen P, Ruuskanen P, Ann.Rev.Nucl.Part.Sci. 56:163 (2006), nucl-th/0605008.
[45] Voloshin SA, Poskanzer AM, Snellings R, Landolt-Boernstein: Relativistic Heavy Ion
Physicsvolume 1/23 (Springer-Verlag, 2010), pp. 5–54, 0809.2949.
[46] Snellings R, New J.Phys. 13:055008 (2011), 1102.3010.
[47] ALICE Collaboration, Floris M, J.Phys.G 38:124025 (2011), 1108.3257.
[48] ALICE Collaboration, Nicassio M, et al., Acta Phys. Polon. SQM2011 Proceedings:(in print)
(2012).
[49] Song H, Heinz UW, Phys. Lett. B 658:279 (2008), arXiv:0709.0742 [nucl-th].
[50] Song H, Heinz UW, Phys. Rev. C 77:064901 (2008), arXiv:0712.3715 [nucl-th].
[51] Song H, Heinz UW, Phys. Rev. C 78:024902 (2008), 0805.1756.
[52] Song H, Bass SA, Heinz U, Phys.Rev. C83:024912 (2011), 1012.0555.
[53] Schnedermann E, Sollfrank J, Heinz UW, Phys.Rev. C48:2462 (1993), nucl-th/9307020.
[54] ALICE Collaboration, Belikov I, et al., J.Phys.G 38:124078 (2011).
[55] PHENIX Collaboration, Adler S, et al., Phys.Rev.Lett. 91:172301 (2003), nucl-ex/0305036.
[56] STAR Collaboration, Abelev BI, et al., Phys. Rev. Lett. 97:152301 (2006), 0606003.
[57] Lamont M, Eur.Phys.J. C49:35 (2007).
[58] Fries RJ, Greco V, Sorensen P, Ann.Rev.Nucl.Part.Sci. 58:177 (2008), 0807.4939.
[59] Topor Pop V, Gyulassy M, Barrette J, Gale C, Phys.Rev. C84:044909 (2011), 1105.3940.
[60] Brodsky SJ, Sickles A, Phys.Lett. B668:111 (2008), 0804.4608.
[61] Andronic A, Braun-Munzinger P, Stachel J, Acta Phys.Polon. B40:1005 (2009), 0901.2909.
[62] Becattini F, Castorina P, Milov A, Satz H, Eur.Phys.J. C66:377 (2010), 0911.3026.
[63] Blume C, Markert C, Prog.Part.Nucl.Phys. 66:834 (2011), 1105.2798.
[64] Margetis S, Safarik K, Villalobos Baillie O, Ann.Rev.Nucl.Part.Sci. 50:299 (2000).
[65] Rafelski J, Mu¨ller B, Phys. Rev. Lett. 48:1066 (1982).
[66] ALICE Collaboration, Kalweit A, et al., Acta Phys. Polon. SQM2011 Proceedings:(in print)
(2012).
[67] Andronic A, Braun-Munzinger P, Stachel J, Phys.Lett. B673:142 (2009), 0812.1186.
[68] STAR Collaboration, Aggarwal M, et al., Phys.Rev. C83:024901 (2011), 1010.0142.
[69] Letessier J, Rafelski J, Eur.Phys.J. A35:221 (2008), nucl-th/0504028.
[70] Rafelski J, Letessier J, Phys.Rev. C83:054909 (2011), 1012.1649.
[71] Lisa MA, Pratt S, Soltz R, Wiedemann U, Ann.Rev.Nucl.Part.Sci. 55:357 (2005), nucl-
ex/0505014.
[72] ALICE Collaboration, Aamodt K, et al., Phys.Lett. B696:328 (2011), 1012.4035.
[73] Bozek P, J.Phys.G 38:124043 (2011), 1106.5953.
[74] Ollitrault JY, Phys.Rev. D46:229 (1992).
[75] Voloshin S, Zhang Y, Z.Phys. C70:665 (1996), hep-ph/9407282.
[76] Poskanzer AM, Voloshin S, Phys.Rev. C58:1671 (1998), nucl-ex/9805001.
[77] ALICE Collaboration, Selyuzhenkov I, et al., J.Phys.G 38:124167 (2011), 1106.5425.
[78] Kovtun P, Son D, Starinets A, Phys.Rev.Lett. 94:111601 (2005), hep-th/0405231.
[79] Teaney DA, Quark-Gluon Plasma 4 (World Scientific, Singapore, 2010), chap. Viscous Hydro-
dynamics and the Quark Gluon Plasma, 0905.2433.
[80] Nagle J, Steinberg P, Zajc W, Phys.Rev. C81:024901 (2010), 0908.3684.
First Results from Pb+Pb collisions at the LHC 21
[81] Schenke B, Jeon S, Gale C, Phys.Lett. B702:59 (2011), 1102.0575.
[82] Qiu Z, Shen C, Heinz UW, Phys.Lett. B707:151 (2012), 1110.3033.
[83] ALICE Collaboration, Aamodt K, et al., Phys.Rev.Lett. 105:252302 (2010), 1011.3914.
[84] Sorensen P, J.Phys.G 37:094011 (2010), 1002.4878.
[85] Alver B, Roland G, Phys.Rev. C81:054905 (2010), 1003.0194.
[86] ALICE Collaboration, Aamodt K, et al., Phys.Rev.Lett. 107:032301 (2011), 1105.3865.
[87] ALICE Collaboration, Dobrin A, et al., J. Phys. G 38:124170 (2011).
[88] ATLAS Collaboration, Aad G, et al., Phys.Lett. B707:330 (2012), 1108.6018.
[89] ATLAS Collaboration, Trzupek A, et al., J.Phys.G 38:124163 (2011).
[90] CERN Report No. ATLAS-CONF-2011-074, 2011 (unpublished).
[91] CMS Collaboration, Chatrchyan S, et al., (2012), 1201.3158.
[92] PHENIX Collaboration, Adare A, et al., (2011), 1105.3928.
[93] STAR Collaboration, Sorensen P, et al., J.Phys.G 38:124029 (2011), 1110.0737.
[94] Shen C, Heinz U, Huovinen P, Song H, Phys.Rev. C84:044903 (2011), 1105.3226.
[95] Heinz UW, Shen C, Song H, (2011), 1108.5323.
[96] ATLAS Collaboration, Jia J, et al., J.Phys.G 38:124012 (2011), 1107.1468.
[97] ALICE Collaboration, Adare A, et al., J.Phys.G 38:124091 (2011), 1107.0285.
[98] CMS collaboration, Li W, et al., J.Phys.G 38:124027 (2011), CMS-PAS-HIN-11-005.
[99] Nagle J, Nucl.Phys. A830:147C (2009), 0907.2707.
[100] ALICE Collaboration, Krzewicki M, et al., J.Phys.G 38:124047 (2011), 1107.0080.
[101] ALICE Collaboration, Yin Z, et al., (2012), 1202.0365.
[102] Huovinen P, Kolb P, Heinz UW, Ruuskanen P, Voloshin S, Phys.Lett. B503:58 (2001), hep-
ph/0101136.
[103] Voloshin S, Nucl.Phys. A715:379 (2003), nucl-ex/0210014.
[104] Fries R, Mu¨ller B, Nonaka C, Bass S, Phys.Rev.Lett. 90:202303 (2003), nucl-th/0301087.
[105] Fries R, Mu¨ller B, Nonaka C, Bass S, Phys.Rev. C68:044902 (2003), nucl-th/0306027.
[106] Greco V, Ko C, Levai P, Phys.Rev. C68:034904 (2003), nucl-th/0305024.
[107] PHENIX Collaboration, Adare A, et al., Phys.Rev.Lett. 98:162301 (2007), nucl-ex/0608033.
[108] Majumder A, Van Leeuwen M, Prog. Part. Nucl. Phys. 66:41 (2011), 1002.2206.
[109] Dokshitzer YL, Khoze VA, Troian SI, J. Phys. G17:1602 (1991).
[110] Mu¨ller B, Nucl. Phys. A 855:74 (2011), 1010.4258.
[111] ALICE Collaboration, Aamodt K, et al., Phys.Lett. B696:30 (2011), 1012.1004.
[112] ALICE Collaboration, Appelsha¨user H, et al., J.Phys.G G38:124014 (2011), 1110.0638.
[113] CMS Collaboration, Chatrchyan S, et al., (2012), 1202.2554.
[114] ATLAS Collaboration, Milov A, et al., J.Phys.G G38:124113 (2011), 1107.0460.
[115] ALICE Collaboration, Dainese A, et al., J.Phys.G G38:124032 (2011), 1106.4042.
[116] CMS Collaboration, Chatrchyan S, et al., (2012), 1201.3093.
[117] ATLAS Collaboration, Steinberg P, et al., J.Phys.G G38:124004 (2011), 1107.2182.
[118] CMS Collaboration, Chatrchyan S, et al., Phys.Rev.Lett. 106:212301 (2011), 1102.5435.
[119] ATLAS Collaboration, Aad G, et al., Phys.Lett. B697:294 (2011), 1012.5419.
[120] CERN Report No. ATLAS-CONF-2011-078, 2011 (unpublished).
[121] ALICE Collaboration, Abelev B, et al. (2012), 1202.1383.
[122] CMS Collaboration, Chatrchyan S, et al., (2012), 1201.5069.
[123] CERN Report No. ATLAS-CONF-2011-079, 2011 (unpublished).
22
[124] Sharma R, Vitev I, Zhang BW, Phys.Rev. C80:054902 (2009), 0904.0032.
[125] Renk T, Holopainen H, Paatelainen R, Eskola KJ, Phys.Rev. C84:014906 (2011), 1103.5308.
[126] Chen XF, Hirano T, Wang E, Wang XN, Zhang H, Phys.Rev. C84:034902 (2011), 1102.5614.
[127] Majumder A, Shen C, (2011), 1103.0809.
[128] Zapp KC, Krauss F, Wiedemann UA, (2011), 1111.6838.
[129] ATLAS Collaboration, Aad G, et al., Phys.Rev.Lett. 105:252303 (2010), 1011.6182.
[130] CMS Collaboration, Chatrchyan S, et al., Phys.Rev. C84:024906 (2011), 1102.1957.
[131] ATLAS Collaboration, Cole BA, et al., J.Phys.G 38:124021 (2011), ATLAS-CONF-2011-075.
[132] CMS Collaboration, Yilmaz Y, et al., J.Phys.G 38:124157 (2011), CMS-PAS-HIN-11-004.
[133] Qin GY, Mu¨ller B, Phys.Rev.Lett. 106:162302 (2011), 1012.5280.
[134] Lokhtin I, Belyaev A, Snigirev A, Eur.Phys.J. C71:1650 (2011), 1103.1853.
[135] Young C, Schenke B, Jeon S, Gale C, (2011), 1103.5769.
[136] Vitev I, J.Phys.G 38:124087 (2011), 1107.5741.
[137] Rapp R, van Hees H, (2009), 0903.1096.
[138] Matsui T, Satz H, Phys.Lett. B178:416 (1986).
[139] Digal S, Petreczky P, Satz H, Phys.Rev. D64:094015 (2001), hep-ph/0106017.
[140] NA38 Collaboration, Baglin C, et al., Phys.Lett. B220:471 (1989).
[141] Gerschel C, Hu¨fner J, Ann.Rev.Nucl.Part.Sci. 49:255 (1999), hep-ph/9802245.
[142] NA50 Collaboration, Abreu M, et al., Phys.Lett. B410:337 (1997).
[143] Rapp R, Blaschke D, Crochet P, Prog. Part. Nucl. Phys. 65:209 (2010), 0807.2470.
[144] Braun-Munzinger P, Stachel J, (2009), 0901.2500.
[145] PHENIX Collaboration, Adare A, et al., Phys.Rev.Lett. 98:232301 (2007), nucl-ex/0611020.
[146] PHENIX Collaboration, Adare A, et al., Phys.Rev. C84:054912 (2011), 1103.6269.
[147] Reed R, J.Phys.G G38:124185 (2011), 1109.3891.
[148] CMS Collaboration, Chatrchyan S, et al., Phys.Rev.Lett. 107:052302 (2011), 1105.4894.