Disordered systems

Cargse

Leticia F. Cugliandolo
[email protected]
Laboratoire de Physique Thorique et Hautes Energies de Jussieu

May 30, 2011

1
Contents
1 Introduction 6
1.1 Falling out of equilibrium . . . . . . . . . . . . . . . . . . . . . 6
1.2 Structural and quenched disorder: glassy physics . . . . . . . . 8
1.3 Interdisciplinary aspects . . . . . . . . . . . . . . . . . . . . . 14
1.3.1 Optimization problems . . . . . . . . . . . . . . . . . . 14
1.3.2 Biological applications . . . . . . . . . . . . . . . . . . 15
1.3.3 Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2 Phase transitions 19
2.1 Order and disorder . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 Order parameters . . . . . . . . . . . . . . . . . . . . . 20
2.2.2 Thermodynamic limit . . . . . . . . . . . . . . . . . . . 21
2.2.3 Pinning field . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2.4 Broken ergodicity . . . . . . . . . . . . . . . . . . . . . 21
2.2.5 Spontaneous broken symmetry . . . . . . . . . . . . . . 22
2.3 Mean-field theory . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3.1 The naive mean-field approximation . . . . . . . . . . . 23
2.3.2 The fully-connected Ising ferromagnet . . . . . . . . . 26

3 Disordered systems 32
3.1 Quenched and annealed disorder . . . . . . . . . . . . . . . . . 32
3.2 Bond disorder: the case of spin-glasses . . . . . . . . . . . . . 33
3.2.1 Lack of homogeneity . . . . . . . . . . . . . . . . . . . 33
3.2.2 Frustration . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2.3 Gauge invariance . . . . . . . . . . . . . . . . . . . . . 34
3.2.4 Self-averageness . . . . . . . . . . . . . . . . . . . . . . 35
3.3 Models with quenched disorder . . . . . . . . . . . . . . . . . 37
3.3.1 Spin-glass models . . . . . . . . . . . . . . . . . . . . . 37
3.3.2 Random ferromagnets . . . . . . . . . . . . . . . . . . 39
3.3.3 Random manifolds . . . . . . . . . . . . . . . . . . . . 40
3.4 The spin-glass transition . . . . . . . . . . . . . . . . . . . . . 41
3.4.1 The simplest order parameter . . . . . . . . . . . . . . 42
3.4.2 Pure states and more subtle order parameters . . . . . 43
3.4.3 Pinning fields . . . . . . . . . . . . . . . . . . . . . . . 46

2
 3.4.4 Divergent susceptibility . . . . . . . . . . . . . . . . . . 46
3.4.5 Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . 48
3.4.6 Scaling . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.5 The TAP approach . . . . . . . . . . . . . . . . . . . . . . . . 48
3.5.1 The complexity or configurational entropy . . . . . . . 51
3.5.2 Weighted averages . . . . . . . . . . . . . . . . . . . . 53
3.5.3 Metastable states in two families of models . . . . . . . 55
3.5.4 Low temperatures . . . . . . . . . . . . . . . . . . . . . 55
3.6 The replica method . . . . . . . . . . . . . . . . . . . . . . . . 62
3.6.1 Interpretation of replica results . . . . . . . . . . . . . 70
3.6.2 The pinning field . . . . . . . . . . . . . . . . . . . . . 76
3.7 Finite dimensional systems . . . . . . . . . . . . . . . . . . . . 77
3.7.1 The Griffiths phase . . . . . . . . . . . . . . . . . . . . 78
3.7.2 Droplets and domain-wall stiffness . . . . . . . . . . . . 78
3.7.3 The droplet theory . . . . . . . . . . . . . . . . . . . . 83
3.8 The random manifold . . . . . . . . . . . . . . . . . . . . . . . 84

4 The Langevin equation 87

4.1 Derivation of the Langevin equation . . . . . . . . . . . . . . . 88
4.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
4.2.1 Irreversibility and dissipation. . . . . . . . . . . . . . . 92
4.2.2 Discretization of stochastic differential equations . . . . 93
4.2.3 Markov character . . . . . . . . . . . . . . . . . . . . . 93
4.2.4 Generation of memory . . . . . . . . . . . . . . . . . . 93
4.2.5 Smoluchowski (overdamped) limit . . . . . . . . . . . . 94
4.3 The basic processes . . . . . . . . . . . . . . . . . . . . . . . . 94
4.3.1 A constant force . . . . . . . . . . . . . . . . . . . . . 95
4.3.2 Relaxation in a quadratic potential . . . . . . . . . . . 102
4.3.3 Thermally activated processes . . . . . . . . . . . . . . 105

5 Dynamics through a phase transition 109

5.1 Time-dependent Ginzburg-Landau description . . . . . . . . . 111
5.2 Relaxation and equilibration time . . . . . . . . . . . . . . . . 116
5.2.1 Quench from T Tc to T > Tc . . . . . . . . . . . . . 116
5.2.2 Quench from T Tc to T Tc . . . . . . . . . . . . . 117
5.2.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.3 Short-time dynamics . . . . . . . . . . . . . . . . . . . . . . . 119
5.4 Growing length and dynamic scaling . . . . . . . . . . . . . . 119

3
 5.5 Critical coarsening . . . . . . . . . . . . . . . . . . . . . . . . 121
5.6 Sub-critical coarsening . . . . . . . . . . . . . . . . . . . . . . 122
5.6.1 Dynamic scaling hypothesis . . . . . . . . . . . . . . . 122
5.6.2 R(t) in clean one dimensional cases with non-conserved
order parameter . . . . . . . . . . . . . . . . . . . . . . 126
5.6.3 R(t) in non-conserved curvature driven dynamics (d > 2)126
5.6.4 R(t) in conserved order parameter and the role of bulk
diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.6.5 Crossover between critical and sub-critical coarsening:
(T ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.6.6 The 2d xy model . . . . . . . . . . . . . . . . . . . . . 131
5.6.7 Role of disorder: thermal activation . . . . . . . . . . . 133
5.6.8 Temperature-dependent effective exponents . . . . . . . 135
5.7 Scaling functions for subcritical coarsening . . . . . . . . . . . 136
5.7.1 Breakdown of dynamic scaling . . . . . . . . . . . . . . 136
5.8 Annealing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.9 An instructive case: the large N approximation . . . . . . . . 137
5.10 Nucleation and growth . . . . . . . . . . . . . . . . . . . . . . 141

6 Classical dynamic generating functional and symmetries 143

6.1 Classical statics: the reduced partition function . . . . . . . . 143
6.2 Classical dynamics: generating functional . . . . . . . . . . . . 144
6.3 Generic correlation and response. . . . . . . . . . . . . . . . . 146
6.4 Time-reversal . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.5 An equilibrium symmetry . . . . . . . . . . . . . . . . . . . . 148
6.5.1 Invariance of the measure . . . . . . . . . . . . . . . . 149
6.5.2 Invariance of the integration domain . . . . . . . . . . 149
6.5.3 Invariance of the action functional . . . . . . . . . . . . 150
6.5.4 Invariance of the Jacobian (Grassmann variables) . . . 150
6.6 Consequences of the transformation . . . . . . . . . . . . . . . 151
6.6.1 The fluctuation-dissipation theorem . . . . . . . . . . . 151
6.6.2 Fluctuation theorems . . . . . . . . . . . . . . . . . . . 152
6.7 Equations on correlations and linear responses . . . . . . . . . 155
6.8 Classical statics: the reduced partition function . . . . . . . . 157

7 Quantum formalism 158

7.1 Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158
7.1.1 Feynman path integral . . . . . . . . . . . . . . . . . . 159

4
 7.1.2 The Matsubara imaginary-time formalism . . . . . . . 159
7.1.3 The equilibrium reduced density matrix . . . . . . . . 159
7.2 Quantum dynamics . . . . . . . . . . . . . . . . . . . . . . . . 161
7.2.1 Schwinger-Keldysh path integral . . . . . . . . . . . . . 162
7.2.2 Green functions . . . . . . . . . . . . . . . . . . . . . . 163
7.2.3 Generic correlations . . . . . . . . . . . . . . . . . . . . 165
7.2.4 Linear response and Kubo relation . . . . . . . . . . . 166
7.2.5 Quantum FDT . . . . . . . . . . . . . . . . . . . . . . 166
7.2.6 The influence functional . . . . . . . . . . . . . . . . . 170
7.2.7 Initial conditions . . . . . . . . . . . . . . . . . . . . . 171
7.2.8 Transformation to MSR-like fields . . . . . . . . . . . 174
7.2.9 Classical limit . . . . . . . . . . . . . . . . . . . . . . . 175
7.2.10 A particle or manifold coupled to a fermionic reservoir 175
7.2.11 Other baths . . . . . . . . . . . . . . . . . . . . . . . . 179

8 Quantum glassiness 180

8.1 Quantum driven coarsening . . . . . . . . . . . . . . . . . . . 180

A Conventions 183
A.1 Fourier transform . . . . . . . . . . . . . . . . . . . . . . . . . 183
A.2 Commutation relations . . . . . . . . . . . . . . . . . . . . . . 183
A.3 Time ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . 183

B The instanton calculation 184

5
1 Introduction

1.1 Falling out of equilibrium

This introduction is not really adapted to this set of lectures but it gives
an idea of the out of equilibrium glassy problems. See the slides for a more
adequate introduction to the lectures.
In standard condensed matter or statistical physics focus is set on equi-
librium systems. Microcanonical, canonical or grand canonical ensembles
are used depending on the conditions one is interested in. The relaxation
of a tiny perturbation away from equilibrium is also sometimes described in
textbooks and undergraduate courses.
More recently, attention has turned to the study of the evolution of sim-
ilar macroscopic systems in far from equilibrium conditions. These can
be achieved by changing the properties of the environment (e.g. the tem-
perature) in a canonical setting or by changing a parameter in the systems
Hamiltonian in a microcanonical one. The procedure of rapidly (ideally in-
stantaneously) changing a parameter is called a quench. Right after both
types of quenches the initial configuration is not one of equilibrium at the
new conditions and the systems subsequently evolve in an out of equilibrium
fashion. The relaxation towards the new equilibrium (if possible) could be
fast (and not interesting for our purposes) or it could be very slow (and thus
the object of our study). There are plenty of examples of the latter. Dissi-
pative ones include systems quenched through a phase transition and later
undergoing domain growth, and problems with competing interactions that
behave as glasses. Energy conserving ones are of great interest at presence
due to the rapid growth of activity in cold-atom systems.
Out of equilibrium situations can also be established by driving a sys-
tem, that otherwise would reach equilibrium in observable time-scales, with
an external pertubation. In the context of macroscopic systems an interest-
ing example is the one of sheared complex liquids. Yet another interesting
case is the one of powders that stay in static metastable states unless ex-
ternally perturbed by tapping, vibration or shear that drives them out of
equilibrium and makes them slowly evolve towards more compact config-

6
urations. Such situations are usually called non-equilibrium steady states
(NESS). But also small systems can be driven out of equilibrium with ex-
ternal perturbations. Transport in nano-structures is the quantum (small)
counterpart phenomenon of these cases, also of special interest at present.
In the study of macroscopic out of equilibrium systems a number of ques-
tions one would like to give an answer to naturally arise. Among these are:
1. Is the (instantaneous) structure out of equilibrium similar to the one
in equilibrium?
2. What microscopic/mesoscopic relaxation mechanism takes place
after the quench?
3. Does the system quickly settle into a stationary state? In more techni-
cal terms, is there a finite relaxation time to reach a steady state?
4. Can one describe the states of the system sometime after the quench
with some kind of effective equilibrium-like measure?
5. Are there thermodynamic concepts, such as temperature, entropy,
free-energy, playing a role in the non-equilibrium relaxation?

In the last 20 years or so a rather complete theory of the dynamics of

classical macroscopic systems evolving slowly out of equilibrium,
in a small entropy production limit (asymptotic regime after a quench,
small drives), that encompasses the situations described above has been de-
veloped. This is a mean-field theory type in the sense that it applies
strictly to models with long-range interactions or in the infinite dimensional
limit. It is, however, proposed that many aspects of it also apply to systems
with short-range interactions although with some caveats. A number of finite
dimensional problems have been solved demonstrating this fact.
In several cases of practical interest, quantum effects play an impor-
tant rle. For instance, glassy phases at very low temperatures have been
identified in a large variety of materials (spin-glass like systems, interacting
electrons with disorder, materials undergoing super-conductor transitions,
metallic glasses, etc.). Clearly, the driven case is also very important in
systems with quantum fluctuations. Take for instance a molecule or an in-
teracting electronic system driven by an external current applied via the
coupling to leads at different chemical potential. It is then necessary to
settle whether the approach developed and the results obtained for the clas-
sical dynamics in a limit of small entropy production carries through when
quantum fluctuations are included.

7
 In these notes we start by exposing some examples of the phenomenol-
ogy of out of equilibrium dynamics we are interested in. We focus on clas-
sical problems and their precise setting. Next we go into the formalism
used to deal with these problems. The basic techniques used to study clas-
sical glassy models with or without disorder are relatively well documented
in the literature (the replica trick, scaling arguments and droplet theories,
the dynamic functional method used to derive macroscopic equations from
the microscopic Langevin dynamics, functional renormalization, Montecarlo
and molecular dynamic numerical methods). On the contrary, the techniques
needed to deal with the statics and dynamics of quantum macroscopic sys-
tems are much less known in general. I shall briefly discuss the role played
by the environment in a quantum system and introduce and compare the
equilibrium and dynamic approaches.
Concretely, we recall some features of the Langevin formalism and its
generating function. We dwell initially with some emblematic aspects of
classical macroscopic systems slowly evolving out of equilibrium. Concerning
models, we focus on two, that are intimately related: the O(N) model in
the large N limit that is used to describe coarsening phenomena, and
the random manifold, that finds applications to many physical problems
like charge density waves, high-Tc superconductors, etc. Both problems are of
field-theoretical type and can be treated both classically and quantum
mechanically. These two models are ideal for the purpose of introducing
and discussing formalism and some basic ideas we would wish to convey in
these lectures. Before entering the technical part we explain the two-fold
meaning of the word disorder by introducing the glass problem and some
of the numerous questions it raises.

1.2 Structural and quenched disorder: glassy physics

While the understanding of equilibrium phases, the existence of phase
transitions as well as the characterization of critical phenomena are well un-
derstood in clean systems, as soon as competing interactions or geometric
frustration are included one faces the possibility of destroying this simple pic-
ture by giving way to novel phenomena like glassy behaviour.
Competing interactions can be dynamic, also called annealed or quenched.
A simple example illustrates the former: the Lennard-Jones potential (see

8
Fig. 1-left) that gives an effective interaction between soft1 particles in a liq-
uid has a repulsive and an attractive part, depending on the distance between
the particles, a set of dynamic variables. In this example, the interactions
depend on the positions of the particles and evolve with them. Quenched
competing interactions are fixed in the observational time-scale and they
transmit contradictory messages. Typical examples are systems with fer-
romagnetic and/or antiferromagnetic exchanges that are not organised in a
simple way with respect to the geometry and connectivity of the lattice such
as spin-glasses [2] (see Fig. 1-right).

0.75
LJ potential

0.25
V

-0.25

-0.75
0 0.5 1 1.5 2 2.5 3
r

Figure 1: Left: The Lennard-Jones potential. Right: the Edwards-Anderson

3d spin-glass.

When competing interactions are present the low-temperature configura-

tions may look disordered but still have macroscopic properties of a kind of
crystalline state. Again, cooling down a liquid to obtain a glass is helpful
to exemplify what we mean here: the liquid cannot support stress and flows
while the glass has solid-like properties as crystals, it can support stress and
does not easily flow in reasonable time-scales (this is why glasses can be made
of glass!) However, when looked at a microscopic scale, one does not iden-
tify any important structural difference between the liquid and the glass: no
simple long-range structural order has been identified for glasses. Moreover,
there is no clear evidence for a phase transition between the liquid and the
glass. At present one can only talk about a dynamic crossover. The glassy
regime is however usually called a glassy phase and it is sometimes said to
be a disordered phase due to the lack of a clear structural order this does
not mean that there is no order whatsoever (see Fig. 3 for an example of a
1
Soft means that the particles can overlap at the price of an energy cost. In the case
this is forbidden one works with hard particles.

9
system with a liquid, a crystal and glassy phase). Lennard-Jones binary mix-
tures are prototypical examples of systems that undergo a glass transition (or
crossover) when cooled across the glass temperature Tg or when compressed
across a density ng [1].
In the paragraphs above we characterized the low temperature regime
of certain particle models and claimed that their structure is disordered (at
list at first sight). Another sense in which the word disorder is used is to
characterize the interactions. Quenched interactions are initially drawn
from a probability distribution. Annealed interactions may have a slow
time-dependence. Both lead to disorder. These can be realized by coupling
strengths as in the magnetic example in Fig. 1, but also by magnetic fields,
pinning centers, potential energies, etc. Disordered interactions usually lead
to low-temperature behaviour that is similar to the one observed in systems
with dynamic competing interactions.
There are many types of glasses and they occur over an astounding range
of scales from macroscopic to microscopic. Macroscopic examples include
granular media like sand and powders. Unless fluidized by shaking or dur-
ing flow these quickly settle into jammed, amorphous configurations. Jam-
ming can also be caused by applying stress, in response to which the material
may effectively convert from a fluid to a solid, refusing further flow. Tem-
perature (and of course quantum fluctuations as well) is totally irrelevant for
these systems since the grains are typically big, say, of 1mm radius. Col-
loidal suspensions contain smaller (typically micrometre-sized) particles
suspended in a liquid and form the basis of many paints and coatings. Again,
at high density such materials tend to become glassy unless crystallization
is specifically encouraged (and can even form arrested gels at low densities if
attractive forces are also present). On smaller scales still, there are atomic
and molecular glasses: window glass is formed by quick cooling of a silica
melt, and of obvious everyday importance. The plastics in drink bottles and
the like are also glasses produced by cooling, the constituent particles being
long polymer molecules. Critical temperatures are of the order of 80C for,
say, PVC and these systems are glassy at room temperature. Finally, on the
nanoscale, glasses are also formed by vortex lines in type-II superconductors.
Atomic glasses with very low critical temperature, of the order of 10mK,
have also been studied in great detail.
In these lectures we shall only deal with a canonical setting, the micro-
canonical one being more relevant to quantum systems that we shall not
discuss here. Disordered systems (in both senses) are usually in contact

10
with external reservoirs at fixed temperature; their description is done in
the canonical (or grand-canonical in particle systems with the possibility of
particle exchange with the environment) ensemble.
Many questions arise for the static properties of systems with compet-
ing interactions. Some of them, that we shall discuss in the rest of the course
are:
1. Are there equilibrium phase transitions between low-temperature and
high temperature phases?
2. Is there any kind of order at low temperatures?
3. At the phase transition, if there is one, does all the machinery developed
for clean systems apply?
4. Are these phases, and critical phenomena or dynamic crossovers, the
same or very different when disorder is quenched or annealed?
5. What is the mechanism leading to glassiness?

Figure 2: A crystal in a 2d colloidal suspension of hard spheres

In practice a further complication appears. Usually, disordered phases are

prepared with a relatively rapid quench from the high temperature phase.
When approaching a characteristic temperature the systems cannot follow
the pace of evolution dictated by the environment and fall out of equi-
librium [3]. Indeed, without entering into the rich details of the glass phe-
nomenology we just mention here that their key feature is that below some
characteristic temperature Tg , or above a critical density g , the relaxation
time goes beyond the experimentally accessible time-scales and the system is
next bound to evolve out of equilibrium. Although the mechanism leading to
such a slow relaxation is unknown and might be different in different cases
the out of equilibrium relaxation presents very similar properties. The left

11
 Figure 3: A liquid or a glass in a 2d colloidal suspension of hard spheres.

panel in Fig. 4 shows one aspect of glassy dynamics, aging, as shown by

the two-time relaxation of the self-correlation of a colloidal suspension, that
is remarkably similar to the decay of the magnetic correlation in the Ising
model shown in the right panel and in Fig. 26.
A purely static description, based on the use of the canonical (or grand-
canonical) partition function is then not sufficient. One is forced to include
the time evolution of the individual agents (spins, particles, molecules) and
from it derive the macroscopic time-dependent properties of the full system.
The microscopic time-evolution is given by a stochastic process. The macro-
scopic evolution is usually very slow and, in probability terms, it is not a small
perturbation around the Gibbs-Boltzmann distribution function but rather
something quite different. This gives rise to new interesting phenomena.
The questions that arise in the non-equilibrium context are

1. How to characterize the non-equilibrium dynamics of glassy systems

phenomenologically.
2. Which are the minimal models that reproduce the phenomenology.
3. Which is the relation between the behavior of these and other non-
equilibrium systems, in particular, those kept away from equilibrium
by external forces, currents, etc.
4. Which features are generic to all systems with slow dynamics.
5. Whether one could extend the equilibrium statistical mechanics ideas;
e.g. can one use temperature, entropy and other thermodynamic con-
cepts out of equilibrium?
6. Related to the previous item, whether one can construct a non-equilibrium
measure that would substitute the Gibbs-Boltzmann one in certain
cases.

12
 a) tw
Various shear histories
1

b)
0.14

C(t,tw)
0.12 tw=2
4
2

0.10
|g1(tw,t)|

8
0.08
0.1 16
0.06
32
64
0.04 128
0.02
256
512
0.01 0.1 1 10 100 1000 1 10 100 1000
t (sec) t-tw

Figure 4: Left: two-time evolution of the self-correlation in a colloidal suspen-

sion initialized by applying a shearing rate (data from Viasnoff and Lequeux)
The longer the waiting time the slower the decay. Right: two-time evolution
in the bi-dimensional Ising model quenched below its phase transition at Tc .
A two-scale relaxation with a clear plateau at a special value of the correla-
tion is seen in the double logarithmic scale. We shall discuss this feature at
length in the lectures.

An out of equilibrium situation can be externally maintained by applying

forces and thus injecting energy into the system and driving it. There are
several ways to do this and we explain below two quite typical ones that
serve also as theoretical traditional examples.
Rheological measurements are common in soft condensed matter;
they consist in driving the systems out of equilibrium by applying an ex-
ternal force that does not derive from a potential (e.g. shear, shaking, etc.).
The dynamics of the system under the effect of such a strong perturbation
is then monitored.
The effect of shear on domain growth is one of great technological and
theoretical importance. The growth of domains is anisotropic and there
might be different growing lengths in different directions. Moreover, it is not
clear whether shear might interrupt growth altogether giving rise to a non-
equilibrium stationary state or whether coarsening might continue for ever.
Shear is also commonly used to study the mechanical properties of diverse
glasses.
Another setting is to couple the system to different external reservoirs
all in equilibrium but at different temperature or chemical potential thus
inducing a heat or a particle current through the system. This set-up is

13
relevant to quantum situations in which one can couple a system to, say, a
number of leads at different chemical potential. The heat transport problem
in classical physics also belongs to this class.

1.3 Interdisciplinary aspects

The theory of disordered systems has become quite interdisciplinary in
the sense that problems in computer science, biology or even sociology and
finance have disorder aspects and can be mimicked with similar models and
solved with similar methods to the ones we shall discuss here.

1.3.1 Optimization problems

The most convenient area of application is, most probably, the one of com-
binatorial optimisation in computer science [4]. These problems can usually
be stated in a form that corresponds to minimizing a cost (energy) function
over a large set of variables. Typically these cost functions have a very large
number of local minima an exponential function of the number of variables
separated by barriers that scale with N and finding the truly absolute min-
imum is hardly non-trivial. Many interesting optimisation problems have the
great advantage of being defined on random graphs and are then mean-field
in nature. The mean-field machinery that we shall discuss at length is then
applicable to these problems with minor (or not so minor) modifications due
to the finite connectivity of the networks. Let us illustrate this with one ex-
ample. In graph partitioning the problem is how to partition, in the optimal
way, a graph with N vertices and K links between them in two groups of
equal size N/2 and the minimal the number of edges between them. This
problem is encountered, for example, in computer design where one wishes
to partition the circuits of a computer between two chips.
Complexity theory in computer science, and the classification of optimi-
sation problems in classes of complexity P for problems solved with algo-
rithms that use a number of operations that grows as a polynomial of the
number of variables, e.g. as N 2 or even N 100 , NP for problems for which no
polynomial algorithm is known and one needs a number of operations that
grow exponentially with N, etc. applies to the worst instance of a problem.
Worst instance, in the graph-partitioning example, means the worst possible

14
realization of the connections between the nodes. Knowing which one this is
is already a very hard problem!
But one can try to study optimisation problems on average, meaning that
the question is to characterize the typical and not the worst realization of
a problem. The use of techniques developed in the field of disordered physical
systems, notably spin-glasses, have proven extremely useful to tackle typical
single randomly generated instances of hard optimization problems.
Note that in statistical mechanics information about averaged macro-
scopic quantities is most often sufficiently satisfactory to consider a problem
solved. In the optimisation context one seeks for exact microscopic configu-
rations that correspond to the exact ground state and averaged information
is not enough. Nevertheless, knowledge about the averaged behaviour can
give us qualitative information about the problem that might be helpful to
design powerful algorithms to attack single instances.

1.3.2 Biological applications

In the biological context disordered models have been used to describe
neural networks, i.e. an ensemble of many neurons (typically N 109 in
the human brain) with a very elevated connectivity. Indeed, each neuron
is connected to 104 other neurons and receiving and sending messages
via their axons. Moreover, there is no clear-cut notion of distance in the
sense that axons can be very long and connections between neurons that are
far away have been detected. Hebb proposed that the memory lies in the
connections and the peculiarity of neural networks is that the connectivity
must then change in time to incorporate the process of learning.
The simplest neural network models [5] represent neurons with Boolean
variables or spins, that either fire or are quiescent. The interactions link
pairs of neurons and they are assumed to be symmetric (which is definitely
not true). Memory of an object, action, etc. is associated to a certain
pattern of neuronal activity. It is then represented by an N-component
vector in which each component corresponds to the activity of each neuron.
Finally, sums over products of these patterns constitute the interactions.
As in optimization problems, one can study the particular case associated
to a number of chosen specific patterns to be stored and later recalled by
the network, or one can try to answer questions on average, as how many
typical patterns can a network of N neurons store. The models then become
fully-connected or dilute models of spins with quenched disorder.

15
 Another field of application of disordered system techniques is the de-
scription of hetero-polymers and, most importantly, protein folding. The
question is how to describe the folding of a linear primary structure (just the
sequence of different amino-acids along the main backbone chain) into an
(almost) unique compact native structure whose shape is intimately related
to the biological function of the protein. In modeling these very complex sys-
tems one proposes that the non-random, selected through evolution, macro-
molecules may be mimicked by random polymers. This assumption is based
on the fact that amino-acids along the chain are indeed very different. One
then uses monomer-monomer and/or monomer-solvent interactions that are
drawn from some probability distribution and are fixed in time (quenched
disorder). Still, a long bridge between the theoretical physicists and the
biologists approaches remain to be crossed. Some of the important missing
links are: proteins are mesoscopic objects with of the order of 100 monomers
thus far from the thermodynamic limit; interest is in the particular, and not
averaged, case in biology, in other words, one would really like to know what
is the secondary structure of a particular primary sequence; etc.

1.3.3 Dynamics
In all these interdisciplinary problems dynamical questions are very im-
portant. In the combinatorial optimisation context, the complexity of a
problem is given by the time needed to solve it. In this context, the micro-
scopic dynamics is not dictated by physical rules but it can be chosen at will
to render the resolution of the problem as fast as possible. However, glassy
aspects, as the proliferation of metastable states separated by barriers that
grow very fast with the number of variables can hinder the resolutions of
these problems in polynomial time for any algorithm.
In the neural network context the microscopic dynamics cannot be chosen
at will but, in general, will not be as simple as the single spin flip ones used
in more conventional physical problems. Still, if the disordered modelling
is correct, glassy aspects can render recall very slow due to the presence of
metastable states for certain values of the parameters.
In the protein folding problem it is clear that the time needed to reach
the secondary structure from an initially stretched configuration depends
strongly on the existence of metastable states that could trap the (hetero)

16
Figure 5: Left: random graph with finite connectivity; right: the fully-
connected case.

polymer. Glassy aspects have been conjectured to appear in this context too.

1.4 Summary
The main steps in the development and application of Statistical Me-
chanics ideas to macroscopic cooperative systems have been (1) the devel-
opment of the basic ideas (Boltzmann-Gibbs); (2) the recognition of collec-
tive phenomena and the identification and mean-field description of phase
transitions (Curie-Weiss); (3) the correct description of critical phenomena
with scaling theories and the renormalization group (Kadanoff, Widom, M.
Fisher, Wilson) and more recently the development of conformal field theo-
ries for two-dimensional systems; (4) the study of stochastic processes and
time-dependent properties (Langevin, Fokker-Planck, Glauber, etc.).
To describe disordered systems the same route has been followed. There
is no doubt that, in principle, Equilibrium Statistical Mechanics yields the
static properties of these systems. Although this might be a little irrelevant
from the practical point of view since, most disordered systems are out of
equilibrium in laboratory time-scales, it is certainly a necessary step on which
one can try to build a truly dynamic theory. The mean-field study step
(2) of the equilibrium properties of disordered systems, in particular those
with quenched disorder, has revealed an incredibly rich theoretical structure.
We still do not know whether it carries through to finite dimensional cases.
Even though, it is definitely interesting per se and it finds a very promising
field of application in combinatorial optimisation problems that are defined
on random networks with mean-field character. Scaling arguments have been
applied to describe finite dimensional disordered systems but they remain as

17
their parent ones for clean systems quite phenomenological and difficult to
put to sufficiently restrictive numerical or experimental test. The extension
of renormalisation group methods to systems with quenched disorder is also
under development and still needs quite a lot of work step (3). As for the
out of equilibrium dynamics of these systems, again, it has been solved at the
mean-field level but little is known in finite dimensions apart from numerical
simulations or the solution to toy models. As in its static counterpart, the
results from the study of dynamic mean-field models have been very rich
and they have suggested a number of new phenomena later searched for in
numerical simulations and experiments of finite dimensional systems. In this
sense, these solutions have been a very important source of inspiration.

18
2 Phase transitions
2.1 Order and disorder
Take a piece of material in contact with an external reservoir. The material
will be characterized by certain observables, energy, magnetization, etc.. To
characterize macroscopic systems it is convenient to consider densities of
energy, magnetization, etc, by diving the macroscopic value by the number
of particles (or the volume) of the system. The external environment will
be characterized by some parameters, like the temperature, magnetic field,
pressure, etc. In principle, one is able to tune the latter and the former will
be a function of them.
Sharp changes in the behavior of macroscopic systems at critical point
(lines) in parameter space have been observed experimentally. These corre-
spond to phase transitions, a non-trivial collective phenomenon appearing in
the thermodynamic limit. In this Section we shall review the main features
of, and analytic approaches used to study, phase transitions.
When one cools down a magnetic sample it undergoes a sharp change in
structure, as shown by a sharp change in its macroscopic properties, at a
well-defined value of the temperature which is called the critical temperature
or the Curie temperature. Assuming that this annealing process is done in
equilibrium, that is to say, that at each temperature step the system manages
to equilibrate with its environment after a relatively short transient an
assumption that is far from being true in glassy systems but that can be safely
assumed in this context the two states above and below Tc are equilibrium
states that can be studied with the standard Statistical Mechanics tools.
More precisely, at Tc the equilibrium magnetization density changes from
0 above Tc to a finite value below Tc . The high temperature state is a disor-
dered paramagnet while the low temperature state is an ordered ferromagnet.
One identifies the magnetization density as the order parameter of the
phase transition. It is a macroscopic observable that vanishes above the
transition and takes a continuously varying value below Tc . The transition
is said to be continuous since the order parameter grows continuously from
zero at Tc .
If one looks in more detail into the behavior of the variation of the mag-

19
netization density close Tc one would realize that the magnetic susceptibility,
 !
mh  
= fh 
(2.1)
h h=0 h h
 
 
h=0

i.e. the linear variation of the magnetization density with respect to its
conjugate magnetic field h diverges when approaching the transition from
both sides. As the second identity shows, the susceptibility is just a second
derivative of the free-energy density. Thus, a divergence of the susceptibility
indicates a non-analyticity of the free-energy density. This can occur only
in the infinite volume or thermodynamic limit, N . Otherwise the free-
energy density is just the logarithm of the partition function, a finite number
of terms that are exponentials of analytic functions of the parameters, and
thus an analytic function of the external parameters itself.
What is observed near such a critical temperature are called critical phe-
nomena. Since the pioneering work of Curie, Langevin and others, the two
phases, paramagnetic and ferromagnetic are well-understood. Qualitative ar-
guments as well as the mean-field approach capture the two phases and their
main characteristics. However, what happens close to the critical point has
remained difficult to describe quantitatively until the development of scaling
and the renormalization group.

2.2 Discussion
Let us discuss some important concepts, pinning fields, broken ergodicity
and broken symmetry, with the help of a concrete example, the Ising model.
The dicussion is however much more general and introduces the concepts
mentioned above.

2.2.1 Order parameters

An order parameter is generically defined as a quantity the average of an
observable that vanishes in one phase and is different from zero in another
one (or other ones). One must notice though that the order parameter is not
unique (any power of an order parameter is itself an order parameter) and
that there can exist transition without an order parameter as the Kosterlitz-
Thouless one in the 2d xy model. In the rest of this course we focus on
problem that do have an order parameter defined as the thermal average of
some observable.

20
2.2.2 Thermodynamic limit
The abrupt change in the order parameter at a particular value of the external
parameters (T, h) is associated to the divergence of some derivative of the
free-energy with respect to one of these parameters. The free-energy is a sum
of positive terms. In a system with a finite number of degrees of freedom (as,
for instance, in an Ising spin model where the sum has 2N terms with N the
number of spins) such a sum is an analytic function of the parameters. Thus,
no derivative can diverge. One can then have a phase transition only in the
thermodynamic limit in which the number of degrees of freedom diverges.

2.2.3 Pinning field

In the absence of a magnetic field for pair interactions the energy is an
even function of the spins, E(~s) = E(~s) and, consequently, the equilibrium
magnetization density computed as an average over all spin configurations
with their canonical weight, eH , vanishes at all temperatures.
At high temperatures, m = 0 characterizes completely the equilibrium
properties of the system since there is a unique paramagnetic state with van-
ishing magnetization density. At low temperatures instead if we perform an
experiment we do observe a net magnetization density. In practice, what hap-
pens is that when the experimenter takes the system through the transition
one cannot avoid the application of tiny external fields the experimental
set-up, the Earth... and there is always a small pinning field that actually
selects one of the two possible equilibrium states, with positive of negative
magnetization density, allowed by symmetry. In the course of time, the ex-
perimentalist should see the full magnetization density reverse, however, this
is not seen in practice since astronomical time-scales would be needed. We
shall see this phenomenon at work when solving mean-field models exactly
below.

2.2.4 Broken ergodicity

Introducing dynamics into the problem2 , ergodicity breaking can be stated
as the fact that the temporal average over a long (but finite) time window is
different from the statical one, with the sum running over all configurations
2
Note that Ising model does not have a natural dynamics associated to it. We shall see
in Section ?? how a dynamic rule is attributed to the evolution of the spins.

21
with their associated Gibbs-Boltzmann weight:

A 6= h A i . (2.2)

In practice the temporal average is done in a long but finite interval 0

< . During this time, the system is positively or negatively magnetized
depending on whether it is in one or the other degenerate equilibrium states.
Thus, the temporal average of the orientation of the spins, for instance,
yields a non-vanishing result A = m 6= 0. If, instead, one computes the
statistical average summing over all configurations of the spins, the result
is zero, as one can see using just symmetry arguments. The reason for the
discrepancy is that with the time average we are actually summing over
half of the available configurations of the system. If time is not as large
as a function of N, the trajectory does not have enough time to visit all
configurations in phase space. One can reconcile the two results by, in the
statistical average, summing only over the configurations with positive (or
negative) magnetization density. We shall see this at work in a concrete
calculation below.

2.2.5 Spontaneous broken symmetry

In the absence of an external field the Hamiltonian is symmetric with respect
to the simultaneous reversal of all spins, si si for all i. The phase tran-
sition corresponds to a spontaneous symmetry breaking between the states
of positive and negative magnetization. One can determine the one that is
chosen when going through Tc either by applying a small pinning field that is
taken to zero only after the thermodynamic limit (h 0 N ), or by im-
posing adequate boundary conditions like, for instance, all spins pointing up
on the borders of the sample. Once a system sets into one of the equilibrium
states this is completely stable in the N limit.
Ergodicity breaking necessarily accompanies spontaneous symmetry break-
ing but the reverse is not true. Indeed, spontaneous symmetry breaking
generates disjoint ergodic regions in phase space, related by the broken sym-
metry, but one cannot prove that these are the only ergodic components in
total generality. Mean-field spin-glass models provide a counterexample of
this implication with a number of equilibrium states not related by symmetry.

22
2.3 Mean-field theory
In spite of their apparent simplicity, the statics of ferromagnetic Ising models
has been solved analytically only in one and two dimensions. The mean-field
approximation allows one to solve the Ising model in any spatial dimension-
ality. Even if the qualitative results obtained are correct, the quantitative
comparison to experimental and numerical data shows that the approxima-
tion fails below an upper critical dimension du in the sense that it does not
capture correctly the behaviour of the systems close to the critical point. It
is however very instructive to see the mean-field approximation at work.

2.3.1 The naive mean-field approximation

The naive mean-field approximation consists in assuming that the probabil-
ity density of the systems spin configuration is factorizable in independent
factors
N
Y 1 + mi 1 mi
P ({si }) = Pi (si ) with Pi (si ) = si ,1 + si ,1 (2.3)
i=1 2 2

and mi = h si i, where the thermal average has to be interpreted in the

restricted sense described in the previous sections, i.e. taken over one ergodic
component, in such a way that mi 6= 0. Note that one introduces an order-
parameter dependence in the probabilities. Using this assumption one can
compute the total free-energy

F = U TS (2.4)

where the energy average is taken with the factorized probability distribution
(2.3) and the entropy S is given by (kB = 1)
X
S= P ({si }) ln P ({si }) . (2.5)
{si }

One can use this approximation to treat finite dimensional models. 3

Applied to the d-dimensional pure ferromagnetic Ising model with nearest-
neighbor interactions on a cubic lattice Jij = J/2 for nearest-neighbors and
3
P Note that this approximation Pamounts to replace the exact equation mi = htanh (h+
J s
j ij j )i by m i = tanh (h + j Jij mj ).

23
zero otherwise. One finds the internal energy
JX X JX X
U = hsi sj i h hsi i = mi mj h mi , (2.6)
2 hiji i 2 hiji i

and the entropy

N
X Y N
Y N
X X Y
S = Pk (sk ) ln Pl (sl ) = Pl (sl ) ln Pl (sl ) Pk (sk )
si =1 k=1 l=1 l=1 sn =1 k(6=l)
N
X X
= Pl (sl ) ln Pl (sl )
l=1 sl =1
X  1 + mi 1 + mi 1 mi 1 mi

= ln + ln . (2.7)
i 2 2 2 2

For a uniformly applied magnetic field, all local magnetization equal the
total density one, mi = m, and one has the order-parameter dependent
free-energy density:
1+m 1+m 1m 1m
 
2
f (m) = dJm hm + T ln + ln . (2.8)
2 2 2 2
The extrema, df (m)/dm = 0, are given by

m = tanh (2dJm + h) , (2.9)

with the factor 2d coming from the connectivity of the cubic lattice. The
stable states are those that also satisfy d2 f /dm2 > 0. This equation of state
predicts a second order phase transition at Tc = 2dJ when h = 0.
Taking the derivative of m with respect to h and the limit h 0 one
easily finds that

m/h|h0 = 2 (2.10)
cosh (2dJm) 2dJ
that, when T Tc approaches
T /Tc 1
, (2.11)
|T Tc | |T Tc |

and diverges as |T Tc |.

24
 This result is qualitatively correct in the sense that Tc increases with
increasing d but the actual value is incorrect in all finite dimensions. In
particular, this treatment predicts a finite Tc in d = 1 which is clearly wrong.
The critical behavior is also incorrect in all finite d, with exponents that
do not depend on dimensionality and take the mean-field values. Still, the
nature of the qualitative paramagnetic-ferromagnetic transition in d > 1 is
correctly captured. We postpone the study of the solutions to eq. (2.9) to the
next Subsection where we shall treat a similar, and in some respects, more
general case.
Having an expression for the free-energy density as a function of the order
parameter, that is determined by eq. (2.9), one can compute all observables
and, in particular, their critical behavior. The Taylor expansion of the free-
energy in powers of m, close to the critical point where m 0, yields the
familiar crossover from a function with a single minima at m to the double
well form:
1 T
f (m) (T 2dJ)m2 + m4 hm . (2.12)
2 12
Indeed, below T = 2dJ = Tc the sign of the quadratic term becomes negative
and the function develops two minima away from m = 0.
Another way of deriving the mean-field approximation is to first single
out a spin si and then write

sj = m + sj (2.13)

for the surrounding spins, with sj m. One then replaces sj by this

expression in the quadratic term in the energy, where m is the global mag-
netization density. Keeping only the leading term, one finds a model with
N non-interacting Ising spins coupled to a field, hef f , that depends on m,
H(m) = hef f i si = (Jzm + h) i si , where h is a uniform external field
P P

and z is the connectivity of the lattice, z = 2d for the cubic case. One has to
determine m self-consistently by requiring hsi i = m. This way of presenting
the approximation makes the mean field character of it more transparent.
This approach is what is usually called Weiss mean-field theory.
One can see that the more spins interact with the chosen one the closer
the spin sees an average field, i.e. the mean-field. The number of interacting
spins increases with the range of interaction and the dimension in a problem
with nearest neighbour interactions on a lattice.

25
2.3.2 The fully-connected Ising ferromagnet
Let us now solve exactly the fully-connected Ising ferromagnet with interac-
tions between all p uplets of spins in an external field:
X X
H= Ji1 ...ip si1 . . . sip hi si , (2.14)
i1 6=...6=ip i

si = 1, i = 1, . . . , N. For the sake of generality we use a generic interaction

strength Ji1 ...ip . The ferromagnetic model corresponds to

J
Ji1 ...ip = (2.15)
p!N p1
with 0 < J = O(1), i.e. finite, and p is a fixed integer parameter, p = 2
or p = 3 or ..., that defines the model. The normalization with N p1 of
the first term ensures an extensive energy in the ferromagnetic state at low
temperatures, and thus a sensible thermodynamic limit. The factor p! is
chosen for later convenience. This model is a source of inspiration for more
elaborate ones with dilution and/or disorder.
Naive mean-field approximation
Using the factorization of the joint probability density that defines the
mean-field approximation, one finds
X X
F ({mi }) = Ji1 ...ip mi1 . . . mip hi mi
i1 6=...6=ip i
N 
1 + mi 1 + mi 1 mi 1 mi
X 
+T ln + ln . (2.16)
i=1 2 2 2 2

Note that a Taylor expansion of the entropic contribution around mi = 0

leads to a polynomial expression that is the starting point in the Landau
theory of second order phase transitions.
The local magnetizations, mi , are then determined by requiring that they
minimize the free-energy density, f ({mj })/mi = 0 and a positive definite
Hessian, 2 f ({mj })/mi mj (i.e. with all eigenvalues being positive at the
extremal value). This yields

X
mi = tanh p Jii2 ...ip mi2 . . . mip + hi (2.17)
i2 6=...6=ip

26
If Ji1 ...ip = J/(p!N p1) for all p uplets and the applied field is uniform, hi = h,
one can take mi = m for all i and these expressions become (2.19) and (2.22)
below, respectively. The mean-field approximation is exact for the fully-
connected pure Ising ferromagnet, as we shall show below. [Note that the
fully-connected limit of the model with pair interactions (p = 2) is correctly
attained by taking J J/N and 2d N in (2.9) leading to Tc = J.]
Exact solution
Let us solve the ferromagnetic model exactly. The sum over spin config-
urations in the partition function can be traded for a sum over the variable,
x = N 1 N i=1 si , that takes values x = 1, 1 + 2/N, 1 + 4/N, . . . , 1
P

4/N, 1 2/N, 1. Neglecting subdominant terms in N, one then writes

eN f (x)
X
Z= (2.18)
x

with the x-parameter dependent free-energy density

J 1+x 1+x 1x 1x
 
f (x) = xp hx + T ln + ln . (2.19)
p! 2 2 2 2
The first two terms are the energetic contribution while the third one is of
entropic origin since N!/(N(1 + x)/2)!(N(1 x)/2)! spin configurations have
the same magnetization density. The average of the parameter x is simply
the averaged magnetization density:
N
1 X
hxi = h si i = m (2.20)
N i=1

In the large N limit, the partition function and all averages of x can
be evaluated in the saddle-point approximation

eN f (xsp )
X
Z , (2.21)

where xsp are the absolute minima of f (x) given by the solutions to f (x)/x|xsp =
0, !
J p1
xsp = tanh x + h , (2.22)
(p 1)! sp
together with the conditions d2 f (x)/dx2 |xsp > 0. Note that the contributing
saddle-points should be degenerate, i.e. have the same f (xsp ) for all ,

27
 -0.8 0.25 -0.25

p=3
0
f (m)

-1.2 -0.3
-0.25
p=2

-1.6 -0.5 -0.35

-1.5 -0.5 0.5 1.5 -1.5 -0.5 0.5 1.5 -1.5 -0.5 0.5 1.5
m m m

Figure 6: The free-energy density f (m) of the p = 2 (left), p = 3 (center)

and p = 4 (right) models at three values of the temperature T < Tc (light
dashed line), T = Tc (dark dashed line) and T > Tc (solid line) and with no
applied field. (The curves have been translated vertically.)

otherwise their contribution is exponentially suppressed. The sum over

then just provides a numerical factor of two in the case h = 0. Now, since
xsp = f (x)/h|xsp = h x i = m , (2.23)
as we shall show in Eq. (2.24), the solutions to the saddle-point equations
determine the order parameter. We shall next describe the phases and phase
transition qualitatively and we shall later justify this description analytically.
Model in a finite field
In a finite magnetic field, eq. (2.22) has a unique positive negative
solution for positive negative h at all temperatures. The model is
ferromagnetic at all temperatures and there is no phase transition in this
parameter.
2nd order transition for p = 2
In the absence of a magnetic field this model has a paramagnetic-ferromagnetic
phase transition at a finite Tc . The order of the phase transition depends on
the value of p. This can be seen from the temperature dependence of the free-
energy density (2.19). Figure 6 displays f (x) in the absence of a magnetic
field at three values of T for the p = 2 (left), p = 3 (center) and p = 4 (right)
models (we call the independent variable m since the stationary points of f (x)

28
are located at the magnetization density of the equilibrium and metastable
states, as we shall show below). At high temperature the unique minimum
is m = 0 in all cases. For p = 2, when one reaches Tc , the m = 0 mini-
mum splits in two that slowly separate and move towards higher values of
|m| when T decreases until reaching |m| = 1 at T = 0 (see Fig. 6-left). The
transition occurs at Tc = J as can be easily seen from a graphical solution
to eq. (2.22), see Fig. 7-left. Close but below Tc , the magnetization increases
1
as m (Tc T ) 2 . The linear magnetic susceptibility has the usual Curie
behavior at very high temperature, , and it diverges as |T Tc |1
on both sides of the critical point. The order parameter is continuous at Tc
and the transition is of second-order thermodynamically.
1st order transition for p > 2
For p > 2 the situation changes. For even values of p, at T two minima
(and two maxima) at |m| = 6 0 appear. These coexist as metastable states
with the stable minimum at m = 0 until a temperature Tc at which the
three free-energy densities coincide, see Fig. 6-right. Below Tc the m = 0
minimum continues to exist but the |m| = 6 0 ones are favored since they
have a lower free-energy density. For odd values of p the free-energy density
is not symmetric with respect to m = 0. A single minimum at m > 0
appears at T and at Tc it reaches the free-energy density of the paramagnetic
one, f (m ) = f (0), see Fig. 6-center. Below Tc the equilibrium state is the
ferromagnetic minimum. For all p > 2 the order parameter is discontinuous
at Tc , it jumps from zero at Tc+ to a finite value at Tc . The linear magnetic
susceptibility also jumps at Tc . While it equals on the paramagnetic side, it
takes a finite value given by eqn. (2.25) evaluated at m on the ferromagnetic
one. In consequence, the transition is of first-order.
Pinning field, broken ergodicity and spontaneous broken symmetry
The saddle-point equation (2.22) for p = 2 [or the mean-field equation
(2.9)] admits two equivalent solutions in no field. What do they corre-
spond to? They are the magnetization density of the equilibrium ferro-
magnetic states with positive and negative value. At T < Tc if one com-
putes m = N 1 N
P N f (x)
i=1 h si i = xe x summing over the two minima
P

of the free-energy density one finds m = 0 as expected by symmetry. In-

stead, if one computes the averaged magnetization density with the partition
sum restricted to the configurations with positive (or negative) x one finds

29
 1.5 1.2 1.2
p=2 p=3 p=4
0.5 0.8 0.8
eq(m)

-0.5 0.4 0.4

-1.5 0 0
-1.5 -0.5 0.5 1.5 0 0.5 1 1.5 0 0.5 1 1.5
m m m

Figure 7: Graphical solution to the equation fixing the order parameter x

for p = 2 (left), p = 3 (center) and p = 4 (right) ferromagnetic models at
three values of the temperature T < T , T = T and T > T and with
no applied field. Note that the rhs of this equation is antisymmetric with
respect to m m for odd values of p while it is symmetric under the same
transformation for even values of p. We show the positive quadrant only
to enlarge the figure. T is the temperature at which a second minimum
appears in the cases p = 3 and p = 4.

30
m = |msp | (or m = |msp |).
In practice, the restricted sum is performed by applying a small magnetic
field, computing the statistical properties in the N limit, and then
setting the field to zero. In other words,
N
! 
1 X 1 ln Z  f (xsp ) 
m h si i = = = |xsp | (. 2.24)

N i=1 N h h h0
 

h0

By taking the N limit in a field one selects the positive (or negatively)
magnetized states.
For all odd values of p the phase transition is not associated to symmetry
breaking, since there is only one non-degenerate minimum of the free-energy
density that corresponds to the equilibrium state at low temperature. The
application of a pinning field is then superfluous.
For any even value of p and at all temperatures the free-energy density
in the absence of the field is symmetric with respect to m m , see
the left and right panels in Fig. 6. The phase transition corresponds to a
spontaneous symmetry breaking between the states of positive and negative
magnetization. One can determine the one that is chosen when going through
Tc either by applying a small pinning field that is taken to zero only after the
thermodynamic limit, or by imposing adequate boundary conditions. Once a
system sets into one of the equilibrium states this is completely stable in the
N limit. In pure static terms this means that one can separate the sum
over all spin configurations into independent sums over different sectors of
phase space that correspond to each equilibrium state. In dynamic terms it
means that temporal and statistical averages (taken over all configurations)
in an infinite system do not coincide.
The magnetic linear susceptibility for generic p is a simple generalization
of the expression in (2.10) and it is given by
 
m  xsp 
= = p1 . (2.25)
h h0 2 J
h h0 cosh ( (p1)! xsp ) J
xp2
 
(p2)! sp

For p = 2, at T > Tc , xsp = 0 the susceptibility is given by (T J)1

predicting the second order phase transition with a divergent susceptibility
at Tc = J. Approaching Tc from below the two magnetized states have the
same divergent susceptibility, (Tc T )1 .
For p > 2, at T > Tc , xsp = 0 and the susceptibility takes the Curie
form = . The Curie law, = , jumps to a different value at the critical
temperature due to the fact that xsp jumps.

31
3 Disordered systems
No material is perfectly homogeneous: impurities of different kinds are
distributed randomly throughout the samples.
A natural effect of disorder should be to lower the critical temperature.
Much attention has been payed to the effect of weak disorder on phase tran-
sitions, that is to say, situations in which the nature of the ordered and dis-
ordered phases is not modified by the impurities but the critical phenomenon
is. On the one hand, the critical exponents of second order phase transitions
might be modified by disorder, on the other hand, disorder may smooth out
the discontinuities of first order phase transitions rendering them of second
order. Strong disorder instead changes the nature of the low-temperature
phase and before discussing the critical phenomenon one needs to under-
stand how to characterize the new ordered glassy phase.
In this Section we shall discuss several types of quenched disorder and
models that account for it. We shall also overview some of the theoretical
methods used to deal with the static properties of models with quenched
disorder, namely, scaling arguments and the droplet theory, mean-field equa-
tions, and the replica method.

3.1 Quenched and annealed disorder

First, one has to distinguish between quenched and annealed disorder.
Imagine that one mixes some random impurities in a melt and then very
slowly cools it down in such a way that the impurities and the host remain
in thermal equilibrium. If one wants to study the statistical properties of
the full system one has to compute the full partition function, summing
over all configurations of original components and impurities. This is called
annealed disorder. In the opposite case in which upon cooling the host and
impurities do not equilibrate but the impurities remain blocked in random
fixed positions, one talks about quenched disorder. Basically, the relaxation
time associated with the diffusion of the impurities in the sample is so long
that these remain trapped:

o 1012 1015 sec tobs 104 sec tdif f , (3.1)

32
where o is the microscopic time associated to the typical scale needed to
reverse a spin.
The former case is easier to treat analytically but is less physically rele-
vant. The latter is the one that leads to new phenomena and ideas that we
shall discuss next.

3.2 Bond disorder: the case of spin-glasses

Spin-glasses are alloys in which magnetic impurities substitute the original
atoms in positions randomly selected during the chemical preparation of the
sample. The interactions between the impurities are of RKKY type:

cos(2kF rij )
Vrkky = J si sj (3.2)
rij3

with rij = |~ri ~rj | the distance between them and si a spin variable that
represents their magnetic moment. Clearly, the location of the impurities
varies from sample to sample. The time-scale for diffusion of the magnetic
impurities is much longer than the time-scale for spin flips. Thus, for all
practical purposes the positions ~ri can be associated to quenched random
variables distributed according to a uniform probability distribution that in
turn implies a probability distribution of the exchanges. This is quenched
disorder.

3.2.1 Lack of homogeneity

It is clear that the presence of quench disorder, in the form of random
interactions, fields, dilution, etc. breaks spatial homogeneity and renders sin-
gle samples heterogenous. Homogeneity is recovered though, if one performs
an average over all possible realizations of disorder, each weighted with its
own probability.

3.2.2 Frustration
Depending on the value of the distance rij the numerator in eq. (3.2)
can be positive or negative implying that both ferromagnetic and antifer-
romagnetic interactions exist. This leads to frustration, which means that
some two-body interactions cannot be satisfied by any spin configuration.
An example with four sites and four links is shown in Fig. 8-left, where

33
 + +
+ + + +
+

Figure 8: A frustrated (left) and an unfrustrated (center) square plaquette.

A frustrated triangular plaquette (right).

we took three positive exchanges and one negative one all, for simplicity,
with the same absolute value, J. Four configurations minimize the energy,
Ef = 2J, but none of them satisfies the lower link. One can easily check
that any closed loop such that the product of the interactions takes a neg-
ative sign is frustrated. Frustration naturally leads to a higher energy and
a larger degeneracy of the number of ground states. This is again easy to
grasp by comparing the number of ground states of the frustrated plaque-
tte in Fig. 8-left to its unfrustrated counterpart shown on the central panel.
Indeed, the energy and degeneracy of the ground state of the unfrustrated
plaquette are Eu = 4J and nu = 2, respectively.
Frustration may also be due to pure geometrical constraints. The canon-
ical example is an anti-ferromagnet on a triangular lattice in which each
plaquette is frustrated, see Fig. 8-right. This is generically called geometric
frustration.
In short, frustration arises when the geometry of the lattice and/or the
nature of the interactions make impossible to simultaneously minimize the
energy of all pair couplings between the spins. Any loop of connected spins
is said to be frustrated if the product of the signs of connecting bonds is
negative. In general, energy and entropy of the ground states increase due
to frustration.

3.2.3 Gauge invariance

The gauge transformation

si = i si , Jij = i Jij j , with i = 1 (3.3)

leaves the energy and the partition function of an Ising spin model with
two-body interactions invariant:

EJ [{s}] = EJ [{s }] ZJ = ZJ . (3.4)

34
This invariance means that all thermodynamic quantities are independent of
the particular choice of the quenched disordered interactions.
Whenever it exists a set of i s such that frustration can be eliminated
from all loops in the model, the effects of disorder are less strong than in
trully frustrated cases, see the example of the Mattis model in Sect. 2.3.2.

3.2.4 Self-averageness
If each sample is characterized by its own realization of the exchanges,
should one expect a totally different behavior from sample to sample? Fortu-
nately, many generic static and dynamic properties of spin-glasses (and other
systems with quenched disorder) do not depend on the specific realization
of the random couplings and are self-averaging. This means that the typical
value is equal to the average over the disorder:
Atyp
J = [ AJ ] (3.5)
in the thermodynamic limit. More precisely, in self-averaging quantities
sample-to-sample fluctuations with respect to the mean value are expected to
be O(N a ) with a > 0. Roughly, observables that involve summing over the
entire volume of the system are expected to be self-averaging. In particular,
the free-energy density of models with short-ranged interactions is expected
to be self-averaging in this limit.
An example: the disordered Ising chain
The meaning of this property can be grasped from the solution of the ran-
dom bond Ising chain defined by the energy function E = i Ji si si+1 with
P

spin variables si = , for i = 1, . . . , N and random bonds Ji independently

taken from a probability distribution P (Ji ). For simplicity, we consider peri-
odic boundary conditions. The disorder-dependent partition function reads
P
e Ji si si+1
X
ZJ = i (3.6)
{si =1}

and this can be readily computed introducing the change of variables i

si si+1 . One finds.
Y X
ZJ = 2 cosh(Ji ) FJ = ln cosh(Ji ) + N ln 2 . (3.7)
i i

The partition function is a product of i.i.d. random variables and it is itself

a random variable with a log-normal distribution. The free-energy density

35
instead is a sum of i.i.d. random variables and, using the central limit the-
orem, in the large N limit becomes a Gaussian random variable narrowly
peaked at its maximum. The typical value, given by the maximum of the
Gaussian distribution, coincides with the average, limN fJtyp [ fJ ] = 0.
General argument
A simple argument justifies the self-averageness of the free-energy den-
sity in generic finite dimensional systems with short-range interactions. Let
us divide a, say, cubic system of volume V = Ld in n subsystems, say also
cubes, of volume v = d with V = nv. If the interactions are short-ranged,
the total free-energy is the sum of two terms, a contribution from the bulk of
the subsystems and a contribution from the interfaces between the subsys-
tems: FJ = ln ZJ = ln conf eEJ (conf ) = ln conf eEJ (bulk)EJ (surf )
P P

ln bulk eEJ (bulk) + ln surf eEJ (surf ) = FJbulk FJsurf (we neglected
P P

the contributions from the interaction between surface and bulk). If the in-
teraction extends over a short distance and the linear size of the boxes
is , the surface energy is negligible with respect to the bulk one and
FJ ln bulk eEJ (bulk) . In the thermodynamic limit, the disorder de-
P

pendent free-energy is then a sum of n = (L/)d random numbers, each

one being the disorder dependent free-energy of the bulk of each subsystem:
FJ nk=1 ln bulkk eEJ (bulkk ) . In the limit of a very large number of
P P

subsystems (L or n 1) the central limit theorem implies that the to-

tal free-energy is Gaussian distributed with the maximum reached at a value
FJtyp that coincides with the average over all realizations of the randomness
[ FJ ]. Morever, the dispersion
about the typical value vanishes in the large
n limit, FJ /[ FJ ] n/n = n1/2 0 in the large n limit. Similarly,
fJ /[fJ ] O(n1/2 ) where fJ = FJ /N is the intensive free-energy. In a suf-
ficiently large system the typical FJ is then very close to the averaged [ FJ ]
and one can compute the latter to understand the static properties of typical
systems.
Lack of self-averageness in the correlation functions
Once one has [FJ ], one derives all disordered average thermal averages
by taking derivatives of the disordered averaged free-energy with respect to
sources introduced in the partition function. For example,

[ FJ ] 
[ h si i ] = , (3.8)
hi hi =0


36
 
[ FJ ] 
[ h si sj i h si ih sj i ] = T , (3.9)
hi hj hi =0


with E E i hi si . Connected correlation functions, though, are not

P

self-averaging quantities. This can be seen, again, studying the random bond
Ising chain:

h si sj iJ h si iJ h sj iJ = ZJ1 ... ZJ = tanh(Ji ) . . . tanh(Jj )(3.10)
,
Ji Jj
where we used h si i = 0 (valid for a distribution of random bonds with zero
mean) and the dots indicate all sites on the chain between the ending points i
and j. The last expression is a product of random variables and it is not equal
to its average (3.9) not even in the large separation limit |~ri ~rj | .

3.3 Models with quenched disorder

3.3.1 Spin-glass models

In the early 70s Edwards and Anderson proposed a rather simple model
that should capture the main features of spin-glasses. The interactions (3.2)
decay with a cubic power of the distance and hence they are relatively short-
ranged. This suggests to put the spins on a regular cubic lattice model
and to trade the randomness in the positions into random nearest neighbor
exchanges taken from a Gaussian probability distribution:
J2
ij
2 12
X
Eea = Jij si sj with P (Jij ) = (2 ) e 2 2 . (3.11)
hiji

The precise form of the probability distribution of the exchanges is suppose

not to be important, though some authors claim that there might be non-
universality with respect to it.
A natural extension of the EA model in which all spins interact has been
proposed by Sherrington and Kirkpatrick
X X
E= Jij si sj hi si (3.12)
i6=j i

and it is called the SK model. The interaction strengths Jij are taken from a
Gaussian pdf and they scale with N in such a way that the thermodynamic

37
is non-trivial:
J2
ij

2 21 2 2 2
P (Jij ) = (2N ) e N N = 2 N . (3.13)
The first two-moments of the exchange distribution are [Jij ] = 0 and [Jij2 ] =
J 2 /(2N). This is a case for which a mean-field theory is expected to be exact.
A further extension of the EA model is called the p spin model
X X
E= Ji1 ...ip si1 . . . sip hi si (3.14)
i1 <...<ip i

with p 3. The sum can also be written as i1 <i2 <...<ip = 1/p! i1 6=i2 6=ip .
P P

The exchanges are now taken from a Gaussian probability distribution

J2
ij

2 21 2 2 2
P (Jij ) = (2N ) e N N = J 2 p!/(2N p1 ) . (3.15)
J p! 2
with [Ji1 ...ip ] = 0 and [Ji21 ...ip ] = 2N p1 . Indeed, an extensive free-energy is
(p1)/2
achieved by scaling Ji1 ...ip with N . This scaling can be justified as
follows. The local field hi = 1/(p 1)! ii2 6=ip Jii2 ...ip mi2 . . . mip should be
P

of order one. At low temperatures the mi s take plus and minus signs. In
particular, we estimate the order of magnitude of this term by working at
T = 0 and taking mi = 1 with probability 12 . In order to keep the discussion
simple, let us take p = 2. In this case, if the strengths Jij , are of order one,
hi is a sum of N i.i.d. random variables, with zero mean and unit variance4 ,
and hi has zero mean and variance equal to N. Therefore, one can argue
that
ih is of order N. To make it finite we then chose Jij to be of order
1/ N or, in other words, we impose [ Jij2 ] = J 2 /(2N). The generalization
to p 2 is straightforward.
Cases that find an application in computer science are defined on random
graphs with fixed or fluctuating finite connectivity. In the latter case one
places the spins on the vertices of a graph with links between couples or
groups of p spins chosen with a probability c. These are called dilute spin-
glasses.
Exercise. Study the statics of the fully connected p-spin ferromagnet in
which all coupling exchanges are equal to J. Distinguish the cases p = 2
from p > 2. What are the order of the phase transitions?
4
0 and h Fi2 i
P
The calculation goes as follow: h Fi i = j Jij h mj i = =
P P 2
jk Jij Jik h mj mk i = j Jij

38
3.3.2 Random ferromagnets
Let us now discuss some, a priori simpler cases. An example is the Mattis
random magnet in which the interaction strengths are given by

Ji1 ...ip = i1 . . . ip with j = with p = 1/2 . (3.16)

In this case a simple gauge transformation, i i si , allows one to trans-

form the disordered model in a ferromagnet, showing that there was no true
frustration in the system.
Random bond ferromagnets (RBFMs) are systems in which the strengths
of the interactions are not all identical but their sign is always positive. One
can imagine such a exchange as the sum of two terms:

Jij = J + Jij , with Jij small and random . (3.17)

There is no frustration in these systems either.

Models with site or link dilution are also interesting:

Esite dil = J si sj i j , Elink dil = J si sj ij , . (3.18)

P P
hiji hiji

with P (i = 0, 1) = p, 1 p in the first case and P (ij = 0, 1) = p, 1 p in

the second.
Link randomness is not the only type of disorder encountered experimen-
tally. Random fields, that couple linearly to the magnetic moments, are also
quite common; the classical model is the ferromagnetic random field Ising
model (RFIM):

1 h2
i
P (hi ) = (2 2 ) 2 e 22 . (3.19)
X X
Erf im = J si sj si hi with
hiji i

The dilute antiferromagnet in a uniform magnetic field is believed to behave

similarly to the ferromagnetic random field Ising model. Experimental real-
izations of the former are common and measurements have been performed
in samples like Rb2 Co0.7 Mg0.3 F4 .
Note that the up-down Ising symmetry is preserved in models in which
the impurities (the Jij s) couple to the local energy (and there is no applied
external field) while it is not in models in which they couple to the local
order parameter (as the RFIM).

39
 The random fields give rise to many metastable states that modify the
equilibrium and non-equilibrium behaviour of the RFIM. In one dimension
the RFIM does not order at all, in d = 2 there is strong evidence that the
model is disordered even at zero temperature, in d = 3 it there is a finite
temperature transition towards a ferromagnetic state. Whether there is a
glassy phase near zero temperture and close to the critical point is still and
open problem.
The RFIM at zero temperature has been proposed to yield a generic
description of material cracking through a series of avalaches. In this problem
one cracking domain triggers others, of which size, depends on the quenched
disorder in the samples. In a random magnetic system this phenomenon
corresponds to the variation of the magnetization in discrete steps as the
external field is adiabatically increased (the time scale for an avalanche to
take place is much shorter than the time-scale to modify the field) and it is
accessed using Barkhausen noise experiments. Disorder is responsible for the
jerky motion of the domain walls. The distribution of sizes and duration of
the avalanches is found to decay with a power law tail cut-off at a given size.
The value of the cut-off size depends on the strength of the random field and
it moves to infinity at the critical point.

3.3.3 Random manifolds

Once again, disorder is not only present in magnetic systems. An example

that has received much attention is the so-called random manifold. This is
a d dimensional directed elastic manifold moving in an embedding N + d
dimensional space under the effect of a quenched random potential. The
simplest case with d = 0 corresponds to a particle moving in an embedding
space with N dimensions. If, for instance N = 1, the particle moves on a
line, if N = 2 it moves on a plane and so on and so forth. If d = 1 one has a
line that can represent a domain wall, a polymer, a vortex line, etc. The fact
that the line is directed means it has a preferred direction, in particular, it
does not have overhangs. If the line moves in a plane, the embedding space
has (N = 1) + (d = 1) dimensions. One usually describes the system with an
N-dimensional coordinate, , ~ that locates in the transverse space each point
on the manifold, represented by the internalqd-dimensional coordinate ~x,
The elastic energy is Eelas = dd x 1 + ((~x))2 with the de-
R

formation cost of a unit surface. Assuming the deformaiton is small one

40
can linearize this expression and get, upto an additive constant, Eelas =
R d
2
d x ((~x))2 .
Disorder is introduced in the form of a random potential energy at each
point in the N + d dimensional space. The manifold feels, then a potential
~ x), ~x) characterized by its pdf. If the random potential is the result
V ((~
of a large number of impurities, the central limit theorem implies that its
probability density is Gaussian. Just by shifting the energy scale one can set
its average to zero, [ V ] = 0. As for its correlations, one typically assumes,
for simplicity, that they exist in the transverse direction only:
~ x), ~x)V (
[ V ((~ ~ (~x ), ~x ) ] = d (~x ~x )V(,
~ ~) . (3.20)
If one further assumes that there is a statistical isotropy and translational
~
invariance of the correlations, V(, ~ ) = W/2 V(| ~ ~ |/) with a
correlation length and (W d2 )1/2 the strength of the disorder. The disorder
can now be of two types: short-ranged if V falls to zero at infinity sufficiently
rapidly and long-range if it either grows with distance or has a slow decay
to zero. An example involving both cases is given by the power law V(z) =
( + z) where is a short distance cut-off and controls the range of the
correlations with > 1 being short-ranged and < 1 being long-ranged.
The random manifold model is then

Z  
H= d x d 2 ~
((~x)) + V ((~x), ~x) . (3.21)
2
This model also describes directed domain walls in random systems. One
can derive it in the long length-scales limit by taking the continuum limit of
the pure Ising part (that leads to the elastic term) and the random part (that
leads to the second disordered potential). In the pure Ising model the second
term is a constant that can be set to zero while the first one implies that
the ground state is a perfectly flat wall, as expected. In cases with quenched
disorder, the long-ranged and short-ranged random potentials mimic cases in
which the interfaces are attracted by pinning centers (random field type) or
the phases are attracted by disorder (random bond type), respectively. For
instance, random bond disorder is typically described by a Gaussian pdf with
zero mean and delta-correlated [V ((~~ x), ~x), V (
~ (~x ), ~x )] = W d2 d (~x
~
~x )( ~ ).

3.4 The spin-glass transition

41
 Let us now discuss a problem in which disorder is so strong as to modify
the nature of the low temperature phase. If this is so, one needs to define a
new order parameter, capable of identifying order in this phase.
3.4.1 The simplest order parameter
The spin-glass equilibrium phase is one in which spins freeze in randomly-
looking configurations. In finite dimensions these configurations are spatially
irregular. A snapshot looks statistical identical to a high temperature para-
magnetic configuration in which spins point in both directions. However,
while at high temperatures the spins flip rapidly and another snapshot taken
immediately after would look completely different from the previous one, at
low temperatures two snapshots taken at close times are highly correlated.
In a spin-glass state the local magnetization is expected to take a non-
zero value, mi = h si i =
6 0, where the average is interpreted in the restricted
sense introduced in the discussion of ferromagnets, that we shall call here
within a pure state (the notion of a pure state will be made more precise
below). Instead, the total magnetization density, m = N 1 N i=1 mi , vanishes
P

since one expects to have as many averaged local magnetization pointing up

(mi > 0) as spins pointing down (mi < 0) with each possible value of |mi |.
Thus, the total magnetization, m, of a spin-glass vanishes at all temperatures
and it is not a good order parameter.
The spin-glass transition is characterized by a finite peak in the linear
magnetic susceptibility and a diverging non-linear magnetic susceptibility.
Let us discuss the former first and show how it yields evidence for the freezing
of the local magnetic moments. For a generic magnetic model such that the
magnetic field couples linearly to the Ising spin, E E i hi si , the linear
P

susceptibility is related, via the static fluctuation-dissipation theorem to the

correlations of the fluctuations of the magnetization:

h si ih 
ij = h (si h si i)(sj h sj i) i . (3.22)
hj h=0


The averages in the rhs are taken without perturbing field. This relation is
proven by using the definition of h si ih and simply computing the derivative
with respect to hj . In particular,
 
ii = h (si h si i)2 i = 1 m2i 0 , (3.23)

with mi = h si i. The total susceptibility measured experimentally is

N 1 ij ij . On the experimental side we do not expect to see O(1) sample-
P

42
to-sample fluctuations in this global quantity. On the analytical side one can
use a similar argument to the one presented in Sect. 2.3.2 to argue that
should be self-averaging (it is a sum over the entire volume of site-dependent
terms). Thus, the experimentally observed susceptibility of sufficiently large
samples should be given by
 
= [ ] = N 1 [ ij ] N 1 [ ii ] = N 1 1 [ m2i ] , (3.24)
X X X

ij i i

since we can expect that cross-terms will be subleading in the large N limit
under the disorder average (note that ij can take negative values). The
fall of at low temperatures with respect to its value at Tc , i.e. the cusp
observed experimentally, signals the freezing of the local magnetizations, mi ,
in the non-zero values that are more favourable thermodynamically. Note
that this argument is based on the assumption that the measurement is done
in equilibrium.
Thus, the natural and simpler global order parameter that characterizes
the spin-glass transition is
q N 1 [ m2i ]
X
(3.25)
i

as proposed in the seminal 1975 Edwards-Anderson paper. q vanishes in

the high temperature phase since all mi are zero but it does not in the low
temperature phase since the square power takes care of the different signs.
Averaging over disorder eliminates the site dependence. Thus, q is also given
by
q = [ m2i ] . (3.26)
These definitions, natural as they seem at a first glance, hide a subtle dis-
tinction that we discuss below.

3.4.2 Pure states and more subtle order parameters

Let us keep disorder fixed and imagine that there remain more than two
pure or equilibrium states in the selected sample. A factor of two takes into
account the spin reversal symmetry. in the following we shall consider half
the phase space, getting rid of this trivial symmetry. Consider the disorder-
dependent quantity
qJ = N 1 m2i
X
(3.27)
i

43
where the mi depend on the realization of the exchanges. Then, two possi-
bilities for the statistical average in mi = hsi i have to be distinguished:
Restricted averages
If we interpret the statistical average in the same restricted sense as the
one discussed in the paramagnetic - ferromagnetic transition of the usual
Ising model, i.e. under a pinning field that selects one chosen pure state,
in (3.27) we define a disorder and pure state dependent Edwards-Anderson
parameter,
N
qJ ea = N 1 (mi )2 ,
X
(3.28)
i

where we label the selected pure state. Although qJ ea could depend on

it turns out that in all known cases it does not and the label is superfluos.
In addition, qJ ea could fluctuate from sample to sample since the indi-
vidual mi do. It turns out that in the thermodynamic limit qJ ea does not
fluctuate. With this in mind we shall later use

qea = qea = qJ ea (3.29)

for the intra-state average. This is the interpretation of the order parameter
proposed by Edwards-Anderson who did not take into account the possibility
that is discussed next.
In the clean Ising model, had we taken into account all the phase space,
(1) (2)
= 1, 2 and mi = hsi i with mi = mi > 0. If we kept only half of it
= 1 and mi = m > 0, say. The dependence on J does not exist in this case
from the very definition of the model.
Full statistical averages
If, instead, the statistical average in mi runs over all possible equilibrium
states (on half the phase space, that is to say, eliminating spin-reversal) the
quantity (3.27) has non-trivial contributions from overlaps between different
states. Imagine each state has a probability weigth wJ (in the ferromagnetic
phase of the Ising model one has only two pure states with w1 = w2 = 1/2
and mi = 1/2hsi i1 + 1/2hsi i2 = 0) then
N
!2
1
wJ mi
X X
qJ = N . (3.30)
i=1

44
 In the ferromagnetic transition = 1, q = qea = m2 , and qea and q are
identical order parameters.
In the disorder case, qJ ea takes the same value on all equilibrium states
independently of there being only two (as in the usual ferromagnetic phase)
or more (as we shall see appearing in fully-connected spin-glass models).
Therefore it does not allow us to distinguish between the two-state and the
many-state scenarii. Instead, qJ does.
Having defined a disorder-dependent order parameter, qJ , and its disorder
average, q, that explains the decay of the susceptibility below Tc , we still have
to study whether this order parameter characterises the low temperature
phase completely. It will turn out that the the knowledge of q is not enough,
at least in fully-connected and dilute spin-glass models. Indeed, one needs
to consider the probability distribution of the fluctuating qJ quantity, P (qJ ).
The more pertinent definition of an order parameter as being given by such
a probability distribution allows one to distinguish between the simple, two-
state, and the many-state scenarii.
In practice, a way to compute the probability distribution of the order
parameter is by using an overlap or correlation between two spin config-
urations, say {si } and {i }, defined as
J
= N 1
X
qs h si i i (3.31)
i

J
where h . . . i is an unrestricted thermal average. qs takes values between
1 and 1. It equals one if {si } and {i } differ in a number of spins that is
smaller than O(N), it equals 1 when the two configurations are totally an-
ticorrelated with the same proviso concerning a number of spins that is not
O(N) and it equals zero when {si } and {i } are completely uncorrelated.
Note that the self-overlap of a configuration with itself is identically one for
Ising spins. Other values are also possible. A related definition is the one of
the Hamming distance:
N
dJs = N 1 h (si i )2 i = 2(1 qs
J
X
). (3.32)
i=1

The overlap can be computed by running a Monte Carlo simulation, equi-

librating a sample and recording many equilibrium configurations. With
them one computes the overlap and should find a histogram with two peaks
at qea and qea (the values of the overlap when the two configurations fall

45
in the same pure state or in the sign reversed ones) and, in cases with many
J
different pure states, other peaks at other values of qs . This is observed in
J
the SK model as exemplified in Fig. ??. Note that qs is related to the q
definition above.
3.4.3 Pinning fields
In the discussion of the ferromagnetic phase transition one establishes that
one of the two equilibrium states, related by spin reversal symmetry, is chosen
by a small pinning field that is taken to zero after the thermodynamic limit,
limh0 limN .
In a problem with quenched disorder it is no longer feasible to choose
and apply a magnetic field that is correlated to the statistical averaged local
magnetization in a single pure state since this configuration is not known!
Moreover, the remanent magnetic field that might be left in any experience
will certainly not be correlated with any special pure state of the system at
hand.
Which is then the statistical average relevant to describe experiments?
We shall come back to this point below.

3.4.4 Divergent susceptibility

In a pure magnetic system with a second-order phase transition the suscep-
tibility of the order parameter to a field that couples linearly to it diverges
when approaching the transition from both sides. In a paramagnet, one
induces a local magnetization with a local field
N
X
mi = h si i = ij hj (3.33)
j=1

with ij the linear susceptibilities, the magnetic energy given by E = E0

i si hi , and the field set to zero at the end of the calculation. Using this
P

expression, the order parameter in the high temperature phase becomes

N N X N X N
1 X 1 X
q = qea = [ m2i ] = [ ij ik hj hk ] (3.34)
N i=1 N i=1 j=1 k=1
If the applied fields are random and taken from a probability distribution
such that hj hk = 2 jk one can replace hj hk by 2 jk and obtain
N N X N
1 X 1 X
q= [ m2i ] = [ 2ij ] 2 SG 2 . (3.35)
N i=1 N i=1 j=1

46
 2 acts as a field conjugated to the order parameter. (One can also argue
that a uniform field looks random to a spin-glass sample and therefore the
same result holds. It is more natural though to use a trully random field since
a uniform one induces a net magnetization in the sample.) The spin-glass
susceptibility is then defined as

1 X 2 2 X 2 X
SG [ ij ] = [ (h si sj i h si ih sj i)2 ] = [ h si sj i 2 ]
N ij N ij N ij
(3.36)
in the high T phase and one finds that it diverges as T Tc+ as expected in a
second-order phase transition. (Note that there is no cancelation of crossed
terms because of the square.) Indeed, the divergence of SG is related to
the divergence of the non-linear magnetic susceptibility that is measurable
experimentally and numerically. An expansion of the total mangnetization
in powers of a uniform field h acting as E E h i si is
P

(3) 3
Mh = h h + ... , (3.37)
6
and the first non-linear susceptibility is then given by
  !4 +
3 Mh  4
3N
*
1 ln Zh 

(3)
X
= = si (3.38)
h3 h=0 h4 h=0 3
 
i c

with the subindex c indicating that the quartic correlation function is con-
nected. Above Tc , mi = 0 at zero field,

3
(3) = 3 (h si sj sk sl i 3h si sj ih sk sl i) = 3 4N 6 h si sj i2 ,
X X

ijkl N ij
(3.39)
and one can identify SG when i = k and j = l plus many other terms that
we assume are finite. Then,
2
(3) = (SG 2 ) . (3.40)
3
This quantity can be accessed experimentally. A careful experimental mea-
surement of (3) , (5) and (7) done by L. Lvy demonstrated that all these
susceptibilities diverge at Tc .

47
3.4.5 Calorimetry
No cusp in the specific heat of spin-glasses is seen experimentally. Since one
expects a second order phase transition this means that the divergence of
this quantity must be very weak.

3.4.6 Scaling
Having identified an order parameter, the linear and the non-linear suscepti-
bility one can now check whether there is a static phase transition and, if it
is of second order, whether the usual scaling laws apply. Many experiments
have been devoted to this task. It is by now quite accepted that Ising spin-
glasses in 3d have a conventional second order phase transition. Still, the
exponents are difficult to obtain and there is no real consensus about their
values. There are two main reasons for this: one is that as Tc is approached
the dynamics becomes so slow that equilibrium measurements cannot really
be done. Critical data are thus restricted to T > Tc . The other reason is
that the actual value of Tc is difficult to determine and the value used has an
important influence on the critical exponents. Possibly, the most used tech-
nique to determine the exponents is via the scaling relation for the non-linear
susceptibility:
h2
!

nl = t f + (3.41)
t
with t = |T Tc |/Tc and one finds, approximately, the values given in Table 1.

d
1 1 2 -1 12 0
3 0.5 4 9

Table 1: Critical exponents in the Ising spin-glass transitions.

3.5 The TAP approach

Disordered models have quenched random interactions. Due to the fluctuat-
ing values of the exchanges, one expects that the equilibrium configurations
be such that in each equilibrium state the spins freeze in different directions.

48
The local averaged magnetizations need not be identical, on the contrary one
expects h si i = mi and, if many states exist, each of them can be identified
by the vector (m1 , . . . , mN ).
One may try to use the naive mean-field equations (2.17) to character-
ize the low temperature properties of these models at fixed quenched dis-
order and determine then the different (m1 , . . . , mN ) values. It has been
shown by Thouless-Anderson-Palmer (TAP) [?, ?] that these equations are
not completely correct even in the fully-connected disordered case: a term
which is called the Onsager reaction term is missing. This term repre-
sents the reaction of the spin i: the magnetization of the spin i produces
a field hj = Jji mi = Jij mi on spin j; this field induces a magnetization
mj = jj hj = jj Jij mi on the spin j and this in turn produces a mean-
field hi = Jij mj = Jij jj Jij mi = jj Jij2 mi on the site i. The equilib-
rium fluctuatio-dissipation relation between susceptibilities and connected
correlations implies jj = h (sj h sj i)2 i = (1 m2j ) and one then
has hi = (1 m2j )Jij2 mi . The idea of Onsager or cavity method is
that one has to study the ordering of the spin i in the absence of its own
effect on the rest of the system. Thus, the field hi has to be subtracted
from the mean-field created by the other spins in the sample, i.e. hcorr i =
P 2 2
j Jij mj + hi mi j Jij (1 mj ) where hi is the external field.
P

The generalization of this argument to p spin interactions is not so straigh-

forward. An alternative derivation has been given by Biroli []. The TAP
equations for p-spin fully connected models read

1
Jii2 ...ip mi2 . . . mip + mi Jii2 2 ...ip (1 m2i2 ) . . . (1 m2ip ) + hi .
X
mi = tanh
i2 6=...6=ip (p 1)!
(3.42)
the first contribution to the internal field is proportional to Ji12 ...ip N (p1)/2
and once the p 1 sums performed it is of order one. The reaction term in-
stead is proportional to Jii2 2 ...ip and, again, a simple power counting shows
that it is O(1). Thus, In disordered systems the reaction term is of the
same order of the usual mean-field; a correct mean-field description has to
include it. In the ferromagnetic case this term can be neglected since it
is subleading in N. Using the fact that there is a sum over a very large
number of elements, Ji21 ...ip can be replaced by its site-independent variance
[Ji21 ...ip ] = p!J 2 /(2N p1) in the last term in (3.42). Introducing the Edwards-
Anderson parameter qea = N 1 i=1 m2i (note that we study the system in
P

49
one pure state) the TAP equations follow:

1 J 2 p
mi (1 qea )p1 .
X
mi = tanh Jii2 ...ip mi2 . . . mip + hi
(p 1)! i2 6=...6=ip 2
(3.43)
The argument leading to the Onsager reaction term can be generalized to
include the combined effect of the magnetization of spin i on a sequence
of spins in the sample, i.e. the effect on i on j and then on k that comes
back to i. These higher order terms are indeed negligible only if the series
of all higher order effects does not diverge. The ensuing condition is 1 >
2 (1 2qea + N 1 i m4i ).
P

The importance of the reaction term becomes clear from the analysis
of the linearized equations, expected to describe the second order critical
behaviour for the SK model (p = 2) in the absence of an applied field. The
TAP equations become mi ( j Jij mj J 2 mi + hi ). A change of basis to
P

the one in which the Jij matrix is diagonal leads to m (J 2 )m +h .

The staggered susceptibility then reads

m   1
= 1 2J + (J)2 . (3.44)
h h=0


The staggered susceptibility for the largest eigenvalue of an interaction matrix

in the Gaussian ensemble, Jmax = 2J, diverges at c J = 1. Note that without
the reaction term the divergence appears at the inexact value T = 2Tc (see
Sect. 3.6 for the replica solution of the SK model).
The TAP equations are the extremization conditions on the TAP free-
energy density:
1 X X
f ({mi }) = Ji1 ...ip mi1 . . . mip Ji21 ...ip (1 m2i1 ) . . . (1 m2ip )
p! i1 6=...6=ip 4p i1 6=...6=ip
N 
1 + mi 1 + mi 1 mi 1 mi
X X 
hi mi + T ln + ln . (3.45)
i i=1 2 2 2 2

The free-energy density as a function of the local magnetizations mi defines

what is usually called the free-energy landscape. Note that this function
depends on N 1 variables, mi , and these are not necessarily identical
in the disordered case in which the interactions between different groups of

50
spins are different. The stability properties of each extreme {ml } are given
by the eigenvalues of the Hessian matrix

f ({mk }) 
Hij . (3.46)
mi mj {m }

l

The number of positive, negative and vanishing eigenvalues determine then

the number of directions in which the extreme is a minimum, a maximum or
marginal. The sets {ml } for which f ({ml }) is the absolute minima yield a
first definition of equilibrium or pure states.
The TAP equations apply to {mi } and not to the configurations {si }.
The values of the {mi } are determined as minima of the TAP free-energy
density, f ({mi }), and they not need to be the same as those of the energy,
H({si }), a confusion sometimes encountered in the glassy literature. The
coincidence of the two can only occur at T 0.

3.5.1 The complexity or configurational entropy

There are a number of interesting questions about the extreme of the TAP
free-energy landscape, or even its simpler version in which the Onsager term
is neglected, that help us understanding the static behaviour of disordered
systems:
1. For a given temperature, T , how many solutions to the mean-field
equations exist? The number of solutions can be calculated using
 
YZ 1 YZ 1  eqi 
NJ = dmi (mi mi ) = dmi (eqi ) det  . (3.47)

i 1 i 1  mj 
{mi } are the solutions to the TAP equations that we write as {eqi = 0}.
The last factor is the normalization of the delta function after the
change of variables, it ensures that we count one each time the integra-
tion variables touch a solution to the TAP equations independently of
its stability properties.
We define the complexity or the configurational entropy as the logarithm
of the number of solutions at temperature T divided by N:
J (T ) N 1 ln NJ (T ) . (3.48)
The normalization with N suggests that the number of solutions is ac-
tually an exponential of N. We shall come back to this very important
point below.

51
 2. Does NJ (T ) dependend on T and does it change abruptly at particular
values of T that may or may not coincide with static and dynamic
phase transitions?
3. One can define a free-energy level dependent complexity

J (f, T ) N 1 ln NJ (f, T ) (3.49)

where NJ (f, T ) is the number solutions in the interval [f, f + df ] at

temperature T .
4. From these solutions, one can identify the minima as well as all saddles
of different type, i.e. with different indeces K. These are different kinds
to metastable states. Geometry constrains the number of metastable
states to satisfy Morse theorem that states N l=1 (1)
l
= 1, where l
P

is the number of negative eigenvalues of the Hessian evaluated at the

solution l, for any continuous and well-behaved function diverging at
infinity in all directions.
One can then count the number of solutions to the TAP equations of
each index, NJ (K, T ), and define the corresponding complexity

J (K, T ) N 1 ln NJ (K, T ) , (3.50)

or even work at fixed free-energy density

J (K, f, T ) N 1 ln NJ (K, f, T ) . (3.51)

Even more interestingly, one can analyse how are the free-energy den-
sities of different saddles are organized. For instance, one can check
whether all maxima are much higher in free-energy density than min-
ima, etc.
5. What is the barrier, f = f1 f0 , between ground states and first
excited states? How does this barrier scale with the actual free-energy
difference, f between these states? The answer to this question is
necessary to estimate the nucleation radius for the reversal of a droplet
under an applied field, for instance.

The definitions of complexity given above are disorder-dependent. One

might then expect that the complexity will show sample-to-sample fluctua-
tions and be characterized by a probability distribution. The quenched com-
plexity, quenched, is then the most likely value of J , it is defined through

52
max P (J ) = P (quenched). In practice, this is very difficult to compute.
Most analytic results concern the annealed complexity

ann N 1 ln [ NJ ] = N 1 ln[ eN J ] . (3.52)

One can show that the annealed complexity is smaller or equal than the
quenched one.

3.5.2 Weighted averages

Having identified many solutions to the TAP equations in the low-T phase
one needs to determine now how to compute statistical averages. A natural
proposal is to give a probability weight to each solution, w , and to use it to
average the value the observable of interest:

wJ O
X
hOi = (3.53)

where labels the TAP solutions, O = O({mi }) is the value that the ob-
servable O takes in the TAP solution , and wJ are their statistical weights,
satisfying the normalization condition wJ = 1. Two examples can illus-
P

trate the meaning of this average. In a spin-glass problem, if O = si , then

O = mi . In an Ising model in its ferromagnetic phase, if O = si , then
O = mi = m and w = 1/2. Within the TAP approach one proposes
J
eF
wJ = P F J (3.54)
e

with FJ the total free-energy of the -solution to the TAP equations. The
discrete sum can be transformed into an integral over free-energy densities,
introducing the degeneracy of solutions quantified by the free-energy density
dependent complexity:
1 1
Z Z
N f
hOi = df e NJ (f, T )O(f ) = df eN (f J (f,T )) O(f ) . (3.55)
Z Z
The normalization is the partition function
Z Z
Z= df eN f NJ (f, T ) = df eN (f J (f,T )) . (3.56)

We assumed that the labelling by can be traded by a labelling by f that

implies that at the same free-energy density level f the observable O takes

53
the same value. In the N limit these integrals can be evaluated by
saddle-point, provided the parenthesis is positive. The disorder-dependent
complexity is generally approximated with the annealed value introduced in
eq. (3.52).
The equilibrium free-energy
The total equilibrium free-energy density, using the saddle-point method
to evaluate the partition function Z in eq. (3.56), reads

feq = N 1 ln Z = min [f T J (f, T )] min J (f, T ) (3.57)

f f

It is clear that J (f, T ) is the Landau free-energy density of the problem

with f playing the rle of the energy and J of the entropy. If we use
f = (E T S)/N = e T s with E the actual energy and S the microscopic
entropy one has
J (f, T ) = e T (s + J (f, T )) . (3.58)
Thus, J is an extra contribution to the total entropy that is due to the
exponentially large number of metastable states. Note that we do not dist-
ninguish here their stability.
Note that J is subtracted from the total free-energy. Thus, it is possible
that in some cases states with a higher free-energy density but very numerous
have a lower total free-energy density than lower lying states that are less
numerous. Collectively, higher states dominate the equilibrium measure in
these cases.
The order parameter
The Edwards-Anderson parameter is understood as a property of a single
state. Within the TAP formalism on then has

qJ ea = N 1 (mi )2 .
X
(3.59)
i

An average over pure states yields wJ (mi )2 .

P

Instead, the statistical equilibrium magnetization, mi = hsi i = wJ mi ,

P

squared is
!2
2
m2i wJ mi wJ wJ mi mi .
X X
qJ h si i = = = (3.60)

54
If there are multiple phases, the latter sum has crossed contributions from
terms with 6= . These sums, as in a usual paramagnetic-ferromagnetic
transition have to be taken over half space-space, otherwise global up-down
reversal would imply the cancellation of all cross-terms.

3.5.3 Metastable states in two families of models

High temperatures
For all models, at high temperatures f (mi ) is characterized by a single sta-
ble absolute minimum in which all local magnetizations vanish, as expected;
this is the paramagnetic state. The mi = 0 for all i minimum continues
to exist at all temperatures. However, even if it is still the global absolute
minimum of the TAP free-energy density, fT AP , at low temperatures it be-
comes unstable thermodynamically, and it is substituted as the equilibrium
state, by other non-trivial configurations with mi 6= 0 that are the absolute
minima of . Note the difference with the ferromagnetic problem for which
the paramagnetic solution is no longer a minimum below Tc .

3.5.4 Low temperatures

At low temperature many equilibrium states appear (and not just two as
in an Ising ferromagnetic model) and they are not related by symmetry (as
spin reversal in the Ising ferromagnet or a rotational symmetry in the Heisen-
berg ferromagnet). These are characterized by non-zero values of the local
magnetizations mi that are different in different states.
At low-temperatures both the naive mean-field equations and the TAP
equations have an exponential in N number of solutions and still an exponen-
tial in N number of them correspond to absolute minima of the mi -dependent
free-energy density. This means that J (T ) and even J (0, f0 , T ) are quan-
tities O(1). These minima can be identified as different states that could be
accessed by applying the corresponding site-dependent pinning fields.
The derivation and understanding of the structure of the TAP free-energy
landscape is quite subtle and goes beyond the scope of these Lectures. Still,
we shall briefly present their structure for the SK and p-spin models to give
a flavor of their complexity.
The SK model

55
 The first calculation of the complexity in the SK model appeared in 1980.
After 25 years of research the structure of the free-energy landscape in this
system is still a matter of discussion. At present, the picture that emerges
is the following. The temperature-dependent annealed complexity is a de-
creasing function of temperature that vanishes only at Tc but takes a very
small value already at 0.6Tc . Surprisingly enough, at finite but large N
the TAP solutions come in pairs of minima and saddles of type one, that is
to say, extrema with only one unstable direction. These states are connected
by a mode that is softer the larger the number of spins: they coalesce and
become marginally stable in the limit N . Numerical simulations show
that starting from the saddle-point and following the left direction along
the soft mode one falls into the minimum; instead, following the right di-
rection along the same mode one falls into the paramagnetic solution. The
free-energy difference between the minimum and saddle decreases for increas-
ing N and one finds, numerically, an averaged f N 1.4 . The extensive
complexity of minima and type-one saddles is identical in the large N limit,
(0, T ) = (1, T ) + O(N 1 ) [Aspelmeier, Bray, Moore (06)] in such a way
that the Morse theorem is respected. The free-energy dependent annealed
complexity is a smooth function of f with support on a finite interval [f0 , f1 ]
and maximum at fmax . The Bray and Moore annealed calculation (with su-
persymmetry breaking) yields fmax = 0.654, max = 0.052, (fmax ) = 8.9.
The probability of finding a solution with free-energy density f can be ex-
pressed as
s
N (f, T ) eN (f,T ) N (fmax ) N | (fmax )|(f fmax )2
p(f, T ) = = e 2 ,
N (T ) N (T ) 2
(3.61)
N (f,T )
where we evaluated the total number of solutions, N (T ) = df e
R
, by
steepest descent The complexity vanishes linearly close to f0 : (f, T )
(f f0 ) with < .
Only the lowest lying TAP solutions contribute to the statistical weight.
The complexity does not contribute to in the large N limit:
= f ann (f, T ) f (f f0 )

>0 iff > (3.62)
f
and min fmin = f0 . See Fig. ??. The total free-energy density in the
exponential is just the free-energy density of each low-lying solution.

56
 Free energy (3)
(5)

(2)

(4)

(1)

Ts T* Tmax
(Ec) T T
tap

Figure 9: The TAP free-energy as a function of T in the spherical p-spin

model. (1) : free energy of the paramagnetic solution for T > T , Ftot for
T < T ; (2) : free energy of the lowest TAP states, with zero temperature
energy Emin ; (3) : free energy of the highest TAP states, corresponding to
Ec ; (4) : an intermediate value of E0 leads to an intermediate value of f
at any temperature; (5) : feq (T ); the difference between curves (5) and (1)
gives the complexity T Sc (feq (T ), T ).

The (spherical) p-spin model

The number and structure of saddle-points is particularly interesting in
the p 3 cases and it is indeed the reason why these models, with a random
first order transition, have been proposed to mimic the structural glass arrest.
The p 3 model has been studied in great detail in the spherical case, that
is to say, when spins are not Ising variables but satisfy the global constraint,
PN 2
i=1 si = N.
Although in general the minima of the mean-field free energy do not
coincide with the minima of the Hamiltonian, they do in the spherical p-spin
model. Their positions in the phase space does not depend on temperature,
while their self-overlap does. At T = 0 a state (stable or metastable) is just a

57
minimum (absolute or local) of the energy. For increasing T energy minima
get dressed up by thermal fluctuations, and become states. Thus, the states
can be labeled by their zero-temperature energy E 0 .
The complexity is given by
pz 2 p 1 2
" #
1 2 2p
(E) = ln + z 2 2+ , (3.63)
2 2 2 p z p
where z is  q 
z = E 0 E 0 2 Ec2 /(p 1) . (3.64)
The complexity vanishes at

E 0 = Emin = f (p) , (3.65)

the ground state of the system, and it is real for zero-temperature energies
E < Eth with s
2(p 1)
Eth = . (3.66)
p
Emin is the zero-T energy one finds with the replica calculation using a 1-step
RSB Ansatz as we shall see below. The finite-T energy of a state is
p
1 h i
E = q2 E0 (p 1)qp pqp1 + 1 . (3.67)
2T
This means that:
1. There can be only a finite number of states with E < E0 .
2. It can be shown that below Eth minima dominate on average.
3. Above Eth one can show that there are states but these are unstable.
Each zero-temperature state is characterized byqunit N-vector si and it
gives rise to a finite-T state characterized by mi = q(E, T )si with q(E, T )
given by q
2 2
p2 2 2
(E + E 2 Eth )
q (1 q) = T . (3.68)
(p 1)2
(q(E, T = 0) = 1 and at finite T the solution with q closest to 1 has to be
chosen.) The self-overlap at the threshold energy, E Eth , is then
p2 2
qth (1 qth )2 = T 2 . (3.69)
p(p 1)

58
 Another way for the q equation to stop having solution, is by increasing
the temperature, T > Tmax (E 0 ), at fixed bare energy E 0 . This means that,
even though minima of the energy do not depend on the temperature, states,
i.e. minima of the free energy do. When the temperature becomes too large,
the paramagnetic states becomes the only pure ergodic states, even though
the energy landscape is broken up in many basins of the energy minima. This
is just one particularly evident demonstration of the fundamental different
between pure states and energy minima. Tmax (E 0 ) is obtained as the limiting
temperature for which eq. (3.68) admits a solution. It is given by
! ! p2
0 2 p1 p2 2
Tmax (E ) = q . (3.70)
p E 0 E 0 2 Eth
2 p

Tmax is a decreasing function of E 0 . The last states to disappear then are the
ones with minimum energu Emin , ceasing to exist at TT AP = Tmax (Emin ).
Below a temperature Td , an exponential (in N) number of metastable
states contribute to the thermodynamics in such a non-trivial way that their
combined contribution to the observables makes them those of a paramagnet.
Even if each of these states is non-trivial (the mi s are different from zero)
the statistical average over all of them yields results that are identical to
those of a paramagnet, that is to say, the free-energy density is 1/(4T ) as
in the mi = 0 paramagnetic solution. One finds
v
u p(p 2)p2
u
Td = t . (3.71)
2(p 1)p1

At a lower temperature Ts (Ts < Td ) there is an entropy crisis, less than

an exponential number of metastable states survive, and there is a static
phase transition to a glassy state.
In the p-spin models there is a range of temperatures in which high lying
states dominate this sum since they are sufficiently numerous so as to have
a complexity that renders the cimbined term f J (f, T ) smaller (in ac-
tual calculations the disorder dependent complexity is approximated by its
annealed value). In short:

1. Above Td the (unique) paramagnetic solution dominates, q = 0 and

= f = 1/(4T ).

59
 2. In the interval T [Ts , Td ] an exponentially large number of states
(with q 6= 0 given by the solution to pq p2 (1 q) = 2T 2 ) dominate
the partition sum. = 1/(4T ) appearing as the continuation of the
paramagnetic solution.
3. At T < Ts the lowest TAP states with E 0 = Emin control the partition
sum. Their total free-energy is different from 1/(4T ).

This picture is confirmed with other studies that include the use of pinning
fields adapted to the disordered situation, the effective portential for two
coupled real replicas, and the dynamic approach.
Low temperatures, entropy crisis
The interval of definition of (E, T ) is the same as (E), that is E
[Emin : Eth ]. Assuming that at a given temperature T the energy Eeq (T )
minimizing lies in this interval, what happens if we lower the temperature?
Remember that the complexity is an increasing function of E, as of course
is f (E, T ). When T decreases we favor states with lower free energy and
lower complexity, and therefore Eeq decreases. As a result, it must exist a
temperature Ts , such that, Eeq (Ts ) = Emin and thus, (Eeq (T )) = (Emin ) =
0. Below Ts the bare energy Eeq cannot decrease any further: there are no
other states below the ground states Emin . Thus, Eeq (T ) = Emin for each
temperature T Ts . As a result, if we plot the complexity of equilibrium
states (Eeq (T )) as a function of the temperature, we find a discontinuity of
the first derivative at Ts , where the complexity vanishes. A thermodynamic
transition takes place at Ts : below this temperature equilibrium is no longer
dominated by metastable states, but by the lowest lying states, which have
zero complexity and lowest free energy density.
We shall show that Ts is the transition temperature found with a replica
calculation. The temperature where equilibrium is given for the first time by
the lowest energy states, is equal to the static transition temperature. Above
Ts the partition function is dominated by an exponentially large number of
states, each with high free energy and thus low statistical weight, such that
they are not captured by the overlap distribution P (q). At Ts the number
of these states becomes sub-exponential and their weight nonzero, such that
the P (q) develops a secondary peak at qs 6= 0.
The threshold
The stability analysis of the states on the threshold shows that these are

60
marginal, in the sense that they have many flat directions. This feature is
very important for the relaxational dynamics of these systems.
Organization
The ordering of the TAP states in terms of their free-energy value and
their stability has been very deeply analyzed in a number of papers. For a
summary see the review articles by Cavagna and Zamponi.
Finite dimensions
In finite-dimensional systems, only equilibrium states can break the er-
godicity, i.e. states with the lowest free energy density. In other words, the
system cannot remain trapped for an infinite time in a metastable state, be-
cause in finite dimension free energy barriers surrounding metastable states
are always finite.
The extra free energy of a droplet of size r of equilibrium phase in a back-
ground metastable phase has a positive interface contribution which grows
as r d1 , and a negative volume contribution which grows as r d ,
f = r d1 f r d (3.72)
where here is the surface tension and f is the bulk free energy difference
between the two phases. This function has always a maximum, whose finite
height gives the free energy barrier to nucleation of the equilibrium phase
(note that at coexistence f = 0 and the barrier is infinite). Therefore, if
initially in a metastable states the system will, sooner or later, collapse in the
stable state with lower free energy density. For this reason, in finite dimension
we cannot decompose the Gibbs measure in metastable components. When
this is done, it is always understood that the decomposition is only valid
for finite times, i.e times much smaller than the time needed for the stable
equilibrium state to take over. On the other hand, in mean-field systems
(infinite dimension), barriers between metastable states may be infinite in
the thermodynamic limit, and it is therefore possible to call pure states also
metastable states, and to assign them a Gibbs weight wJ . We will analyze
a mean-field spin-glass model, so that we will be allowed to perform the
decomposition above even for metastable states.
Comments
There is a close relationship between the topological properties of the
model and its dynamical behavior. In particular, the slowing down of the

61
dynamics above but close to Td is connected to the presence of saddles, whose
instability decreases with decreasing energy. In fact, we have seen that the
threshold energy level Eth separating saddles from minima, can be associ-
ated to the temperature Tth = Td , marking the passage from ergodicity to
ergodicity breaking. In this context the dynamical transition can be seen as
a topological transition. The plateau of the dynamical correlation function,
which has an interpretation in terms of cage effect in liquids, may be reinter-
preted as a pseudo-thermalization inside a saddle with a very small number
of unstable modes.

3.6 The replica method

A picture that is consistent with the one arising from the naive mean-field
approximation but contradicts the initial assumption of the droplet model
arises from the exact solution of fully-connected spin-glass models. These
results are obtained using a method which is called the replica trick and that
we shall briefly present below.
In Sect. 2.3.2 we argued that the typical properties of a disordered system
can be computed from the disorder averaged free-energy
Z
[ FJ ] dJP (J)FJ . (3.73)

One then needs to average the logarithm of the partition funtion. In the
annealed approximation one exchanges the ln with the average over disorder
and, basically, considers the interactions equilibrated at the same tempera-
ture T as the spins:
[ ln ZJ ] ln[ ZJ ] . (3.74)
This approximation turns out to be correct at high temperatures but incor-
rect at low ones.
The replica method allows one to compute [ FJ ] for fully-connected mod-
els. It is based on the smart use of the identity
ZJn 1
ln ZJ = lim . (3.75)
n0 n
The idea is to compute the right-hand-side for finite and integer n = 1, 2, . . .
and then perform the analytic continuation to n 0. Care should be taken
in this step: for some models the analytic continuation may be not unique.

62
It turns out that this is indeed the case for the emblematic Sherrington-
Kirkpatrick model, as discussed by Palmer and van Hemmen in 1979 though
it has also been recently shown that the free-energy f (T ) obtained by Parisi
with the replica trick is exact!
The disorder averaged free-energy is given by
1
Z Z 
[ FJ ] = dJP (J) ln ZJ = lim dJP (J)ZJn 1 , (3.76)
n0 n

where we have exchanged the limit n 0 with the integration over the
exchanges. For integer n the replicated partition function, ZJn , reads
1 n
ZJn = e[EJ ({si })+...+EJ ({si }] .
X
(3.77)
{sa
i}

Here {sai } {s1i =1} . . . {sni =1} . ZJn corresponds to n identical copies of
P P P

the original system, that is to say, all of them with the same realization of
the disorder. Each copy is characterized by an ensemble of N spins, {sai }.
We label the copies with a replica index a = 1, . . . , n. For p-spin disordered
spin models ZJn takes the form
Pn h P P i
Ji1 ...ip sa a
i ...sip + hi sa
ZJn
X i
a=1 i1 6=...6=ip i
= e 1
. (3.78)
{sa
i}

The average over disorder amounts to computing a Gaussian integral for each
set of spin indices i1 , . . . ip . One finds
2J 2
P P P P
( sa ...sa 2
ip ) + hi sa a
[ ZJn ] = eF ({si }) . (3.79)
X X
e 2N p1 i1 6=...6=ip a i1 a i i

{sa
i} {sa
i}

The function F ({sai }) is not random. It depends on the spin variables only
but it includes terms that couple different replica indices:
!p
N 2 J 2 X 1 X a b
F ({sai}) hi sai .
XX
si si + n (3.80)
2 a6=b N i a i

In writing the last expression we have dropped terms that are subleading
in N (in complete analogy with what we have done for the pure p spin
ferromagnet). The constant term Nn 2 J 2 /2 originates in the terms with
a = b, for which (sai )2 = 1.

63
 To summarize, we started with an interacting spin model. Next, we en-
larged the number of variables from N spins to N n replicated spins by
introducing n non-interacting copies of the system. By integrating out the
disorder we decoupled the sites but we payed the price of coupling the repli-
cas. Hitherto the replica indices act as a formal tool introduced to compute
the average over the bond distribution. Nothing distinguishes one replica
from another and, in consequence, the free-energy F ({sai }) is invariant un-
der permutations of the replica indices.
The next step to follow is to identify the order parameters and transform
the free-energy into an order-parameter dependent expression to be rendered
extremal at their equilibrium values. In a spin-glass problem we already
know that the order parameter is not the global magnetization as in a pure
magnetic system but the parameter q or more generally the overlap between
states. Within the replica calculation an overlap between replicas

qab N 1 sai sbi

X
(3.81)
i

naturally appeared in eq. (3.80). The idea is to write the free-energy density
as a function of the order parameter qab and look for their extreme in complete
analogy with what has been done for the fully-connected ferromagnet. This
is, of course, a tricky business, since the order parameter is here a matrix
with number of elements n going to zero! A recipe for identifying the form of
the order parameter (or the correct saddle-point solution) has been proposed
by G. Parisi in the late 70s and early 80s. This solution has been recently
proven to be exact for mean-field models by two mathematical physics, F.
Guerra and M. Talagrand. Whether the very rich physical structure that
derives from this rather formal solution survives in finite dimensional systems
remains a subject of debate.
Introducing the Gaussian integral
2
Z
= e 2 ( N J )
N 1
P
sa sb N sa sb
P
q2
dqab eJqab i i i 2 ab i i i (3.82)

for each pair of replica indices a 6= b, one decouples the site indeces, i, and
the averaged replicated partition function can be rewritten as
Z Y
[ZJn ] = dqab eF (qab ) (3.83)
a6=b

64
and

N 2 J 2 X p
F (qab ) = qab + n N ln (qab ) , (3.84)
2 a6=b

eH(qab ,sa ) , H(qab , sa ) = J

X X X
(qab ) = qab sa sb h sa(3.85)
sa ab a

where for simplicity we set hi = h. The factor N in front of ln comes from

the decoupling of the site indeces. Note that the transformation (3.82) serves
to uncouple the sites and to obtain then the very useful factor N in front of
the exponential. The partition function Z(qab ) is the one of a fully-connected
Ising model with interaction matrix qab .
Saddle-point evaluation
Having extracted a factor N in the exponential suggests to evaluate the
integral over qab with the saddle-point method. This, of course, involves the
a priori dangerous exchange of limits N and n 0. The replica theory
relies on this assumption. One then writes
1 sp
lim [ fJ ] lim f (qab ) (3.86)
N n0 n

and searches for the solutions to the n(n 1)/2 extremization equations

f (qab ) 
=0. (3.87)
qcd qsp

ef

In usual saddle-point evaluations the saddle-point one should use is (are)

the one(s) that correspond to absolute minima of the free-energy density. In
the replica calculation the number of variables is n(n 1)/2 that becomes
negative! when n < 1 and makes the saddle-point evaluation tricky. In order
to avoid unphysical complex results one needs to focus on the saddle-points
with positive (or at least semi-positive) definite Hessian

f (qab ) 
H , (3.88)
qcd qef qsp

ab

and these sometimes corresponds to maxima (instead of minima) of the free-

energy density.

65
 The saddle-point equations are also self-consistency equations
sp
qab = hsa sb iH(qab ,{sa }) = [ hsa sb i ] (3.89)

where the second member means that the average is performed with the
single site Hamiltonian H(qab , sa ) and the third member is just one of the
averages we would like to compute.
The partition function in eq. (3.85) cannot be computed for generic qab
since there is no large n limit to exploit on the contrary, n 0. Thus, one
usually looks for solutions to eqs. (3.87) within a certain family of matrices
qab . We discuss below the relevant parametrizations.
Replica symmetry (RS)
In principle, nothing distinguishes one replica from another one. This is
the reason why Sherrington and Kirkpatrick looked for solutions that preserve
replica symmetry:
qab = q , for all a 6= b . (3.90)
Inserting this Ansatz in (3.84) and (3.85) and taking n 0 one finds
s
Z dz 2 pq p1
q= ez /2 tanh2 z + h . (3.91)
2 2

This equation resembles strongly the one for the magnetization density of
the p-spin ferromagnet, eq. (2.22).
Let us first discuss the case p = 2, i.e. the SK model. In the absence of
a magnetic field, one finds a second order phase transition at Ts = J from
a paramagnetic (q = 0) to a spin-glass phase with q 6= 0. In the presence
of a field there is no phase transition. SK soon realized though that there
is something wrong with this solution: the entropy at zero temperature is
negative, S(0) = 1/(2), and this is impossible for a model with discrete
spins, for which S is strictly positive. de Almeida and Thouless later showed
that the reason for this failure is that the replica symmetric saddle-point is
not stable, since the Hessian (3.88) is not positive definite and has nega-
tive eigenvalues. The eigenvalue responsible for the instability of the replica
symmetric solution is called the replicon.
Comparison with the TAP equations shows that the RS Ansatz corre-
sponds to the assumption that the local fields hi = ii1 ...iip Ji1 ...ip mi1 . . . mip +
P

66
 q
2
q
q1
q1 2
o
q q1 q
2 q
0 q
2
0
q1
q1 q
2
o
q1 q
2

o o
o o
o o

q q q
2
q1
0 0
q1 q
2

q1 q
2

Figure 10: Left: a one-step replica symmetry breaking (1RSB) Ansatz. Right:
a two-step replica symmetry breaking Ansatz. The elements on the main
diagonal vanish identically. In the 1RSB case the diagonal blocks have size
m m. In the 2RSB the proceudre is repeated and one has blocks of size
m1 m1 with smaller diagonal blocks of size m2 m2 .

h are independent and have a Gaussian distribution with average h and vari-
ance 2 = J 2 q p1. Numerical simulations clearly show that this assumption
is invalid.
Interestingly enough, the numerical values for several physical quantities
obtained with the replica symmetric solution do not disagree much with
numerical results. For instance, the ground state zero-temperature energy
density is E 0 = 0.798 while with numerical simulations one finds E 0
0.76.
For the p > 2 model one finds that the replica symmetric solution is
stable at all temperatures. However, the problem of the negative entropy
remains and should be solved by another solution. The transition must then
have aspects of a first-order one, with another solution appearing at low
temperatures and becoming the most convenient one at the transition.
One step replica symmetry breaking
The next challenge is to device a replica symmetry breaking Ansatz, in
the form of a matrix qab that is not invariant under permutations of rows or
columns. There is no first principles way of doing this, instead, the structure
of the Ansatz is the result of trial and error. Indeed, a kind of minimal way

67
to break the replica symmetry is to propose a structure in blocks as the one
shown in Fig. 10-left. The diagonal elements are set to zero as in the RS
case. Square blocks of linear size m close to the main diagonal are filled with
a paramater q1 . The elements in the rest of the matrix take a different value
q0 and one takes 0 q0 q1 . The matrix qab depends on three parameters
q0 , q1 , m and one has to find the values such that the free-energy density is
maximized! The conditions for a extreme are
f (q0 , q1 , m) f (q0 , q1 , m) f (q0 , q1 , m)
= = =0. (3.92)
q0 q1 m
In the SK model (p = 2) the 1RSB Ansatz yields a second order phase
transition (q0 = q1 = 0 and m = 1 at criticality) at a critical temperature
Ts = J, that remains unchanged with respect to the one predicted by the
RS Ansatz. The 1RSB solution is still unstable below Ts and in all the
low temperature phase. One notices, however, that the zero temperature
entropy, even if still negative and incorrect, takes a value that is closer to
zero, S(T = 0) 0.01, the ground state energy is closer to the value
obtained numerically, and the replicon eigenvalue even if still negative has
an absolute value that is closer to zero. All this suggest that the 1RSB Ansatz
is closer to the exact solution.
Instead, in all cases with p 3 the 1RSB Ansatz is stable below the static
critical temperature Ts and all the way up to a new characteristic temperature
0 < Tf < Ts . Moreover, one can prove that in this range of temperatures
the model is solved exactly by this Ansatz. The critical behaviour is quite
peculiar: while the order parameters q0 and q1 jump at the transition from
a vanishing value in the paramagnetic phase to a non-zero value right below
Ts , all thermodynamic quantities are continuous since m = 1 at Ts and all q0
and q1 dependent terms appear multiplied by 1 m. This is a mixed type
of transition that has been baptized random first-order. Note that disorder
weakens the critical behaviour in the p 3-spin models. In the limit p
the solutions become m = T /Tc , q0 = 0 and q = 1.
k-step replica symmetry breaking
The natural way to generalize the 1RSB Ansatz is to propose a k-step one.
In each step the off-diagonal blocks are left unchanged while the diagonal ones
of size mk are broken as in the first step thus generating smaller square blocks
of size mk+1 , close to the diagonal. At a generic k-step RSB scheme one has
0 q0 q1 . . . qk1 qk 1 , (3.93)

68
 n = m0 m1 . . . mk mk+1 , (3.94)

parameters. In the n 0 limit the ordering of the parameters m is reversed

0 = m0 m1 . . . mk mk+1 . (3.95)

In the SK model one finds that any finite k-step RSB Ansatz remains
unstable. However, increasing the number of breaking levels the results con-
tinue to improve with, in particular, the zero temperature entropy getting
closer to zero. In the p 3 case instead one finds that the 2RSB Ansatz
has, as unique solution to the saddle-point equations, one that boils down to
the 1RSB case. This suggests that the 1RSB Ansatz is stable as can also be
checked with the analysis of the Hessian eigenvalues: the replicon is stricly
positive for all p 3.
Full replica symmetry breaking
In order to construct the full RSB solution the breaking procedure is
iterated an infinite number of times. The full RSB Ansatz thus obtained
generalizes the block structure to an infinite sequence by introducing a func-
tion

q(x) = qi , mi+1 < x < mi (3.96)

with 0 x 1. Introducing q(x) sums over replicas are traded by integrals

over x; for instance
1X l
Z 1
qab = dx q l (x) . (3.97)
n a6=b 0

The free-energy density becomes a functional of the function q(x). The ex-
tremization condition is then a hard functional equation. A Landau expan-
sion expected to be valid close to the assumed second order phase transition
simplifies the task of solving it. For the SK model one finds
x
(
2
, 0 x x1 = 2q(1) ,
q(x) = (3.98)
qea qmax = q(1) , x1 = 2q(1) x 1 ,

at first order in |T Tc |, with q(1) = |T Tc |/Tc and x1 = 2q(1). The

stability analysis yields a vanishing replicon eigenvalue signalling that the
full RSB solution is marginally stable.
One can also recover the particular case of the 1RSB using a q(x) with
two plateaux, at q0 and q1 and the breaking point at x = m.

69
 1 1
RS 1RSB
qea qea

q
0 0
0 1 0 x1 1
x x
1
Full RSB
qea
q

0
0 x1 1
x

Figure 11: The function q(x) for a replica symmetric (left), one step replica
symmetry breaking (center) and full replica symmetry breaking Anstze.

Marginality condition
In the discussion above we chose the extreme that maximize the free-
energy density since we were interested in studying equilibrium properties.
We could, instead, use a different prescription, though a priori not justified,
and select other solutions. For example, we can impose that the solution
is marginally stable by requiring that the replicon eigenvalue vanishes. In
the p = 2 this leads to identical results to the ones obtained with the usual
prescription since the full-RSB Ansatz is in any case marginally stable. In the
p-spin models with p 3 instead it turns out that the averaged properties
obtained in this way correspond to the asymptotic values derived with the
stochastic dynamics starting from random initial conditions. This is quite a
remarkable result and we shall discuss it in more detail in Sect. ??.

3.6.1 Interpretation of replica results

Let us now discuss the implications of the solution to fully-connected dis-
ordered models obtained with the, for the moment, rather abstract replica
formalism.
The interpretation uses heavily the identification of pure states. Their

70
definition is a tricky matter that we shall not discuss in detail here. We
shall just assume it can be done and use the analogy with the ferromagnetic
system and its two pure states and the TAP results at fixed disorder. As
we already know, which are the pure states, its properties, number, etc. can
depend on the quenched disorder realization and fluctuate from sample to
sample. We shall keep this in mind in the rest of our discussion.
Let us then distinguish the averages computed within a pure state and
over all configuration space. In a ferromagnet with no applied magnetic
field this is simple to grasp: at high temperatures there is just one state, the
paramagnet, while at low temperatures there are two, the states with positive
and negative magnetization. If one computes the averaged magnetization
restricted to the state of positive (negative) magnetization one finds meq >
0 (meq < 0); instead, summing over all configurations meq = 0 even at
low temperatures. Now, if one considers systems with more than just two
pure states, and one labels them with Greeks indices, averages within such
states are denoted hOi while averages taken with the full Gibbs measure
are expressed as
wJ h O i .
X
hOi = (3.99)

wJ is the probability of the state given by

J
eF J
wJ eF
X
= , with ZJ = (3.100)
ZJ

and thus satisfying the normalization condition wJ = 1. FJ can be

P

interpreted as the total free-energy of the state . These probabilities, as well

as the state dependent averages, will show sample-to-sample fluctuations.
One can then define an overlap between states:

qJ N 1 hsi i hsi i = N 1 mi mi
X X
(3.101)
i i

and assume rename the self-overlap the Edwards-Anderson parameter

qJ N 1
X
hsi i hsi i qJ ea . (3.102)
i

The statistics of possible overlaps is then characterized by a probability func-

tion
wJ wJ (q q ) ,
X
PJ (q) (3.103)

71
where we included a subindex J to stress the fact that this is a strongly
sample-dependent quantity. Again, a ferromagnetic model serves to illustrate
the meaning of PJ (q). First, there is no disorder in this case so the J label
is irrelevant. Second, the high-T equilibrium phase is paramagnetic, with
q = 0. P (q) is then a delta function with weight 1 (see the left panel in
Fig. 12). In the low-T phase there are only two pure states with identical
statistical properties and qea = m2 . Thus, P (q) is just the sum of two delta
functions with weight 1/2 (central panel in Fig. 12).
Next, one can consider averages over quenched disorder and study
Z
wJ wJ (q q ) .
X
[ PJ (q) ] dJ P (J) (3.104)

How can one access PJ (q) or [ PJ (q) ]? It is natural to reckon that

!
PJ (q) = Z 2 eEJ () eEJ (s) N 1
X X
i si q (3.105)
s i

that is to say, PJ (q) is the probability of finding an overlap q between two real
replicas of the system with identical disordered interactions in equilibrium at
temperature T . This identitiy gives a way to compute PJ (q) and its average
in a numerical simulation: one just has to simulate two independent systems
with identical disorder in equilibrium and calculate the overlap.
But there is also, as suggested by the notation, a way to relate the pure
state structure to the replica matrix qab . Let us consider the simple case

ZJn1 X 1 1
[ mi ] = ZJ1 si eEJ ({si }) = eEJ ({si })
X
s

{si }
ZJn {s1 } i
i

1 X 1 Pn EJ ({sai })
= n se a=1 (3.106)
ZJ {sa } i
i

where we singled out the replica index of the spin to average. This relation
is valid for all n, in particular for n 0. In this limit the denominator
approaches one and the average over disorder can be simply evaluated
ef f ({sa })
s1i eE
X
[ mi ] = i (3.107)
{sa
i}

72
 Pn
EJ ({sa
and introducing back the normalization factor Z n = 1 = e i })
P
a=1
Pn {sa
i}
EJ ({sa ef f ({sa })
=[ e i }) ] = eE we have
P
a=1 i
{sa
i}

[ mi ] = h sai iEef f (3.108)

with a any replica index. The average is taken over the Gibbs measure of a
system with effective Hamiltonian Eef f . In a replica symmetric problem in
which all replicas are identical this result should be independent of the label
a. Instead, in a problem with replica symmetry breaking the averages on the
right-hand-side need not be identical for all a. This could occur in a normal
vectorial theory with dimension n in which not all components take the same
expected value. It is reasonable to assume that the full thermodynamic
average is achieved by the sum over all these cases,
n
1X
[ mi ] = lim h sai i . (3.109)
n0 n
a=1

Let us now take a less trivial observable and study the spin-glass order
parameter q

1
q [ h si i 2 ] = Z J si eEJ ({si }) ZJ1 i eEJ ({i })
X X

{si } {i }

Z n2 X
= n si i eEJ ({si })EJ ({i })
Z {si },{i }

1 X 1 2 Pn EJ ({sai })
= n ss e a=1 (3.110)
ZJ {sa } i i
i

In the n 0 limit the denominator is equal to one and one can then perform
the average over disorder. Introducing back the normalization one then has
q = h sai sbi iEef f ({sai }) (3.111)
for any arbitrary pair of replicas a 6= b (since h sai sai i = 1 for Ising spins).
The average is done with an effective theory of n interacting replicas char-
acterized by Eef f ({sai }). Again, if there is replica symmetry breaking the
actual thermal average is the sum over all possible pairs of replicas:
1
q ab .
X
q = lim (3.112)
n0 n(n 1)
a6=b

73
A similar argument allows one to write
1
q (k) = [ h si1 . . . sik i2 ] = lim k
X
qab . (3.113)
n0 n(n 1)
a6=b

One can also generalize this argument to obtain

1
(q q ab )
X
P (q) = [ PJ (q) ] = lim (3.114)
n0 n(n 1)
a6=b

Thus, the replica matrix qab can be ascribed to the overlap between pure
states.
Note that a small applied field, though uncorrelated with a particular pure
state, is necessary to have non-zero local magnetizations and then non-zero
q values.
The function P (q) then extends the concept of order parameter to a
function. In zero field the symmetry with respect to simultaneous reversal of
all spins translates into the fact that PJ (q) must be symmetric with respect
to q = 0. [ PJ (q) ] can be used to distinguish between the droplet picture
prediction for finite dimensional spin-glasses two pure states that simply
corresponds to
1 1
[ PJ (q) ] = (q qea ) + (q + qea ) (3.115)
2 2
(see the central panel in Fig. 12) and a more complicated situation in which
[ PJ (q) ] has the two delta functions at qea plus non-zero values on a finite
support (right panel in Fig. 12) as found in mean-field spin-glass models.
The linear susceptibility
Taking into account the multiplicity of pure states, the magnetic suscep-
tibility, eq. (3.24), and using (3.99) becomes
1 X Z
[ h si i2 ] = 1 [ wJ wJ ] q = dq (1 q) P (q) .
X
T = T [ J ] = 1
N i
(3.116)
There are then several possible results for the susceptibiliy depending on the
level of replica symmetry breaking in the system:
1. In a replica symmetric problem or, equivalently, in the droplet model,

= (1 qea ) . (3.117)

74
 PM FM

-1 0 1 -1 2 0 2 1
-m m
q q

p-spin SK

-1 -qEA 0 qEA 1 -1 -qEA 0 qEA 1

q q

Figure 12: [ PJ (q) ] in a paramagnet, a ferromagnet, a replica symmetric

system, a system with one-step replica-symmetry breaking a system with
full RSB.

This is also the susceptibility within a pure state of a system with a

higher level of RSB.
2. At the one step RSB level, this becomes

= [1 (1 m)qea ] . (3.118)

3. For systems with full RSB one needs to know the complete P (q) to
compute , as in (3.116).

Note that in systems with RSB (one step or full) the susceptibility is larger
than (1 qea ).
A system with qea = 1 in the full low-temperature phase (as the REM
model or p limit of the p spin model, see below) has just one configu-
ration in each state. Systems with qea < 1 below Tc have states formed by
a number of different configurations that is exponentially large in N. (Note
that qea < 1 means that the two configurations differ in a number of spins
that is proportional to N.) The logarithm of this number is usually called
the intra-state entropy.
Even if the number of pure states can be very large (exponential in N)

75
only a fraction of them can have a non-negligible weight. This is the case if
one finds, for example, w2 < +
P

Symmetry and ergodicity breaking

In all p 2 spin models there is a phase transition at a finite Ts at which
the rather abstract replica symmetry is broken. This symmetry breaking is
accompanied by ergodicity breaking as in the usual case. Many pure states
appear at low temperatures, each one has its reversed si si counterpart,
but not all of them are related by real-space symmetry properties.
The one-step RSB scenario
In this case the transition has first-order and second-order aspects. The
order parameters q0 and q1 jump at the critical point as in a first-order
transition but the thermodynamic quantities are continuous.
The full RSB scenario
Right below Tc an exponential in N number of equilibrium states appear.
The transition is continuous, the order parameter approaches zero right below
Tc . Lowering further the temperature each ergodic component breaks in
many other ones. In this sense, the full spin-glass phase, T < Tc , is critical
and not only the single point Tc .

3.6.2 The pinning field

We can nevertheless choose a possible direction, given by another field (x),
and compute the freeenergy of our system when it is weakly pinned by this
external quenched field
1 Z
g
R 2
F [, g, ] = log d(x) eH[] 2 dx((x)(x)) (3.119)

where g > 0 denotes the strength of the coupling. This free-energy (3.119)
will be small when the external perturbing field (x) lies in a direction corre-
sponding to the bottom of a well of the unperturbed free-energy. Therefore,
we should be able to obtain useful information about the free-energy land-
scape by scanning the entire space of the configurations (x) to locate all the
states in which the system can freeze after spontaneous ergodicity breaking
(g 0). According to this intuitive idea, we now consider the field (x) as a
thermalized variable with the Hamiltonian F [, g, ]. The free-energy of

76
the field at inverse temperature m where m is a positive free parameter
therefore reads
1
Z
F (m, ) = lim+ log d(x) emF [,g,] (3.120)
g0 m
When the ratio m between the two temperatures is an integer, one can easily
integrate (x) in eq.(3.120) after having introduced m copies (x) ( =
1...m) of the original field to obtain the relation
m
1 Z Y

P
H[ ]+ 12
P R
g dx (x) (x)
F (m, ) = lim+ log d (x) e ,
g0 m =1
(3.121)
1
where g = g( m ). Let us define two more quantities related to the field
: its internal energy W (m, ) = (mFm
)
and its entropy S(m, ) = m2 F m

.
Since the case m = 1 will be of particular interest, we shall use hereafter
Fhs () W (m = 1, ) and Shs () S(m = 1, ) where hs stands for
F
hidden states. We stress that S(m, ) and 2 which are respectively
the entropies of the fields and are two distinct quantities with different
physical meanings.
When the pinning field (x) is thermalized at the same temperature as
(x), that is when m = 1, one sees from eq.(3.121) that F () = F (m =
1, ). The basic idea of this letter is to decompose F into its energetic and
entropic contributions to obtain
 
Shs () = Fhs () F () (3.122)

To get some insights on the significance of the above relation, we shall now
turn to the particular case of disordered mean-field systems. We shall see how
it rigorously gives back some analytical results derived within the mean-field
TAP and dynamical approaches. We shall then discuss the physical meaning
of identity (3.122) for the general case of glassy systems.

3.7 Finite dimensional systems

We start now the discussion on the statics of spin-glass models by describing
briefly scaling arguments and the droplet theory. Similar arguments can be
used to study other models with strong disorder, as a manifold in a random
potential.

77
3.7.1 The Griffiths phase
The effects of quenched disorder show up already in the paramagnetic phase
of finite dimensional systems. Below the critical point of the pure case (no
disorder) finite regions of the system can order due to fluctuations in the
couplings. Take the case of random ferromagnetic interactions. Fluctuations
in bonds can be such that in a given region they take higher values than on
average. In practice, at the working temperature T that is higher than the
transition temperature of the full system, Tcdis , a particular region can behave
as if it had have an effective Tcloc that is actually higher than Tc , see Fig. ??.
Similarly, fluctuations can make a region more paramagnetic than the average
if the Jij s take smaller values [ Jij ]. (Note that Tc is typically proportional
to J, the strength of the ferromagnetic couplings. In the disordered case we
normalize the Jij s in such a way that [ Jij ] = Jpure . We can then compare
the disordered and the pure problems.)
These properties manifest in non-analyticities of the free-energy that ap-
pear in a full interval of temperatures above (and below) the critical temper-
ature of the disordered model, as shown by Griffiths. For instance, deviations
from Curie-Weiss ( = 1/T ) behaviour appear below the Nel temperature
of dilute antiferromagnets in a uniform field. These are sometimes described
with a Lorentzian distribution of local temperatures with the corresponding
Curie-Weiss law at each T . It is clear that Griffiths effects will also affect the
relaxation of disordered systems above freezing. We shall not discuss these
features in detail here.

3.7.2 Droplets and domain-wall stiffness

Let us now just discuss one simple argument that is at the basis of what is
needed to derive the results of the droplet theory without entering into the
complications of the calculations.
It is clear the structure of droplets, meaning patches in which the spins
point in the direction of the opposite state, plays an important role in the
thermodynamic behaviour of systems undergoing a phase transition. At criti-
cality one observes ordered domains of the two equilibrium states at all length
scales with fractal properties. Right above Tc finite patches of the system
are indeed ordered but these do not include a finite fraction of the spins in
the sample and the magnetization density vanishes. However, these patches
are enough to generate non-trivial thermodynamic properties very close to

78
Tc and the richness of critical phenomena. M. Fisher and others developed
a droplet phenomenological theory for critical phenomena in clean systems.
Later D. S. Fisher and D. Huse extended these arguments to describe the
effects of quenched disorder in spin-glasses and other random systems; this
is the so-called droplet model.
Critical droplet in a ferromagnet
Let us study the stability properties of an equilibrium ferromagnetic phase
under an applied external field that tends to destabilize it. If we set T = 0
the free-energy is just the energy. In the ferromagnetic case the free-energy
cost of a spherical droplet of radius R of the equilibrium phase parallel to
the applied field embedded in the dominant one (see Fig. 13-left) is

F (R) = 2d Rd hmeq + d1 Rd1 0 (3.123)

where 0 is the interfacial free-energy density (the energy cost of the do-
main wall) and d is the volume of a d-dimensional unit sphere. We assume
here that the droplet has a regular surface and volume such that they are
proportional to Rd1 and Rd , respectively. The excess free-energy reaches a
maximum !d1
d dd1 d1
Fc = d
0d (3.124)
d d 2dhmeq
at the critical radius
(d 1)d1 0
Rc = , (3.125)
2dd hmeq
see Fig. 13 (h > 0 and m > 0 here, the signs have already been taken into
account). The free-energy difference vanishes at

d1 0
F (R0 ) = 0 R0 = . (3.126)
2d hmeq

Several features are to be stressed:

1. The barrier vanishes in d = 1; indeed, the free-energy is a linear func-

tion of R in this case.
2. Both Rc and R0 have the same dependence on hmeq : they monotoni-
cally decrease with increasing hmeq vanishing for hmeq and di-
verging for hmeq 0.

79
 droplet free-energy

f
fc

R 0
h

m
0 Rc
R

Figure 13: Left: the droplet. Right: the free-energy density f (R) of a spher-
ical droplet with radius R.

3. In dynamic terms that we shall discuss later, the passage above the
barrier is done via thermal activation; as soon as the system has reached
the height of the barrier it rolls on the right side of potential F and
the favorable phase nucleates.
4. As long as the critical size Rc is not reached the droplet is not favorable
and the system remains positively magnetized.

In this example the field drives the system from one state to the other.
In studies of phase transitions at zero external field, temperature generates
fluctuations of different size and the question is whether these are favourable
or not. The study of droplet fluctuations is useful to establish whether an
ordered phase can exist at low (but finite) temperatures. One then studies
the free-energy cost for creating large droplets with thermal fluctuations that
may destabilize the ordered phase, in the way we have done with the simple
Ising chain. Indeed, a fundamental difference between an ordered and a
disordered phase is their stiffness (or rigidity). In an ordered phase the free-
energy cost for changing one part of the system with respect to the other
part far away is of the order kB T and usually diverges as a power law of the
system size. In a disordered phase the information about the reversed part
propagates only a finite distance (of the order of the correlation length, see
below) and the stiffness vanishes.
The calculation of the stiffness is usually done as follows. Antiparallel
configurations (or more generally the two ground states) are imposed at the

80
opposite boundaries of the sample. A domain wall is then generated some-
where in the bulk. Its free-energy cost, i.e. the difference between the free-
energies of the modified configuration and the equilibrium one, is measured
and one tests when creating a wall is favourable.
The Imry-Ma argument for the random field Ising model
Take a ferromagnetic Ising model in a random field, defined in eq. (3.19).
In zero applied field and low enough temperature, if d > 1 there is phase
transition between a paramagnetic and a ferromagnetic phase. Under the
effect of a random field with very strong typical strength, the spins align
with the local external fields and the system is paramagnetic. It is, however,
non-trivial to determine the effect of a relatively weak random field on the
ferromagnetic phase at sufficiently low temperature. The long-range ferro-
magnetic order could be preserved or else the field could be enough to break
up the system into large but finite domains of the two ferromagnetic phases.
A qualitative argument to estimate whether the ferromagnetic phase sur-
vives or not in presence of the external random field due to Imry and Ma.
Let us fix T = 0 and switch on a random field. If a domain D of the opposite
order (say down) is created within the bulk of the ordered state (say up) the
system pays an energy due to the unsatisfied links lying on the boundary
that is
Eborder 2JLd1 (3.127)
where L is the linear length of the border and d 1 is the dimension of
the border of a domain embedded in d a dimensional volume, assuming it
is compact. By creating a domain boundary the system can also gain a
magnetic energy in the interior of the domain due to the external field:

Erf hLd/2 (3.128)

since there are N = Ld spins inside the domain

of lineard/2
length L and, using
the central limit theorem, h jD si h N = hL . The comparison
P

between these two energy scales yields

! 2

d/2 h d2
JLd1
0 hL0 L0 (3.129)
J

In the large L limit E diverges to + with increasing L in d > 2. The

marginal case d = 2 is more subtle and we do not discuss it in detail here.

81
One can also search for an extreme in E(L) finding
!2
4J(d 1)
Lc . (3.130)
hd
Several comments are in order:
1. In d = 1 the energy difference is a monotonically decreasing function
of L thus suggesting that the creation of droplets is very favorable and
there is no barrier to cross to do it.
2. In d > 2 the energy difference first decreases from E(L = 0) = 0
to reach a nagative minimum at Lc , and then increases back to pass
through zero at L0 and diverge at infinity. This indicates that the
creation of domains at zero temperature is not favorable in d > 2.
Just domains of finite length, upto L0 can be created. Note that L0
increases with h/J in d > 2 and thus a higher field tends to generate
larger droplets and thus disorder the sample.
With this argument one cannot show the existence of a phase transition at
hc nor the nature of it. The argument is such that it suggests that order can
be supported by the system at zero temperature and small fields.
An elastic line in a random potential
Let us take an interface model of the type defined in eq. (3.21) with
N = 1. If one assumes that the interfaces makes an excursion of longitudinal
length L and transverse length the leastic energy cost is
c
Z
Eelast = dd x ((~x))2 Eelast cLd (L1 )2 = cLd2 2
2
(3.131)
d d+1
If the excursion is sufficiently large, the interface meets L / impurities
(that is to say the volume of the displacement over the typical volume between
impurities given by the correlation length of disorder to the power given by
the number of dimensions).
q Each impurity applies a pinning force of the
d
order of dV /d W/ and then the energy gain due to the random
potential is q
Erandom W/d . (3.132)
The balance between the cost of elastic energy and the gain in random energy
leads to
(L/)(4d)/3 (3.133)

82
where = (c2 4 /W )1/(4d) is the Larkin length and = (4 d)/3 is the
Flory exponent for the roughness of the surface. One then concludes that for
d > 4 disorder is irrelevant and the interface is flat ( 0 when L ).
Since the linearization of the elastic energy [see the discussion leading to
eq. (3.21)] holds only if /L 1, the result (3.133) may hold only for d > 1
where < 1.
The 3d Edwards-Anderson model in a uniform magnetic field
A very similar reasoning is used to argue that there cannot be spin-glass
order in an Edwards-Anderson model in an external field. The only difference
is that the domain wall energy is here assumed to be proportional to Ly with
an a priori unknown d-dependent exponent y that is related to the geometry
of the domains.
Comments These arguments are easy to implement when one knows the
equilibrium states. They cannot be used in models in which the energy is
not a slowly varying function of the domain wall position.

3.7.3 The droplet theory

The droplet theory is a phenomenological model that assumes that the low
temperature phase of a spin-glass model has only two equilibrium states
related by an overall spin flip. It is then rather similar to a ferromagnet,
only that the nature of the order in the two equilibrium states is not easy
to see, it is not just most spins pointing up or most spins pointing down
with some thermal fluctuations within. At a glance, one sees a disordered
paramagnetic like configuration and a more elaborate order parameter has
to be measured to observe the order. The spin-glass phase is then called a
disguised ferromagnet and a usual spontaneous symmetry breaking (between
the two equilibrium states related spin reversal symmetry) leading to usual
ergodicity breaking is supposed to take place at Tc .
Once this assumption has been done, renormalization group arguments
are used to describe the scaling behavior of several thermodynamic quanti-
ties. The results found are then quantitatively different from the ones for a
ferromagnet but no novelties appear.
We shall not develop these arguments here.

83
3.8 The random manifold
The problem of an oriented manifold embedded in a space with random
impurities is relevant for a large class of physical systems. The dimension
of the manifold is called d while the dimension of the embedding space is
called N; there are then N transverse dimensions to the manifold. Both d
and N are important to determine the behaviour of the manifold. Different
choices of d and N are associated with different physical problems. If N = 1
the model describes the interface between two coexisting phases. For d = 1
the model is that of a directed polymer which also describes the interaction
of a flux line in a type II superconductor. This problem shares many of the
interesting features of other disordered system such as spin-glasses.
Attention has been first paid to the study of the equilibrium properties
of the manifold. Mainly two approaches have been applied. On the one
hand, Mzard and Parisi [24] proposed a Gaussian Variational method com-
plemented by the replica analysis with an ensuing replica symmetry breaking
solution. This study allowed to obtain very interesting results such as non-
trivial - Flory-like - exponents for the displacement of the manifold, sample
to sample susceptibility fluctuations, etc. On the other hand, the functional
renormalization group and the relation between the two in the large N limit
was discussed by Le Doussal and collaborators [11]. The extension of the
large N treatment and the Gaussian Variational Approach (GVA) to the out
of equilibrium relaxation has been developed and we discuss it below [14].
The dynamic functional renormalization group to treat the out of equilibrium
dynamics is not ready yet.
The model of a manifold of internal dimension d embedded in a random
medium is described, in terms of an N component displacement field, ~ =
(1 , 2 , . . . , N ), by the Hamiltonian
1 ~~ ~ x), ~x) +
Z  
H= d
d x ((~x))2 + V ((~ ~ 2 (~x) . (3.134)
2 2
is a mass, which effectively constraints the manifold to fluctuate in a re-
stricted volume of the embedding space. The elastic terms has to be read
as
d X N
~ x) =
~ (~
X
(~x) (~x) . (3.135)
i=1 =1 xi xi
~ (,
The Hamiltonian is invariant under rotations of the vector . ~ ~x) = (~
~ x)
is a point in the N + d dimensional space. V is a Gaussian random potential

84
with zero mean and correlations

~ ~ 2
V (, ~ , ~x ) = N d (~x ~x ) | | .
~ ~x)V ( (3.136)
N

The factors N are introduced to have a good N limit. The overline

represents the average over the distribution of the random potential V . We
label with greek indices the N-components of the field , ~ , = 1, . . . , N.
The internal coordinate ~x has d components ~x = (x1 , . . . , xd ) and we label
them with latin indices, xi , i = 1, . . . , d. Note the similarity with the O(N)
coarsening problem. We study this problem in the infinite volume limit
< < and < xi < .

Equilibrium properties

The equilibrium properties of the manifold are often described in terms

of the displacement of the manifold D st (~x, ~x ) that is characterized by the
roughness exponent :

~ x ) 2 i |~x ~x |2 .
~ x) (~
 
D st (~x, ~x ) h (~ (3.137)

The angular brackets denote a thermal average. The random potential and
the thermal fluctuations make the manifold roughen and this impies that the
~ diverge at large distances |~x ~x | 1.
fluctuations of the field
In the absence of disorder the manifold is flat if the internal dimension is
larger than two: for d > 2, = 0. Instead, if the internal dimension is smaller
that two the manifold roughens: for d < 2, = (2 d)/2. The situation
changes in the presence of disorder where depends on d and N non-trivially.
For instance, for d > 4 the manifold remains flat; for 2 < d < 4 there is only
a disorder phase with a non-trivial exponent , finally, for d < 2, and large
enough N there is a high-temperature phase and a randomness dominated
low-temperature phase with increasing with N and approaching 1/2 for
N as suggested by numerical simulations for finite N and calculations
for N . Whether there is a finite upper critical Nc such that for N > Nc
= 1/2 is not clear.
There are several models which can be studied for N corresponding
to different choices of the random potential correlation (z). A common

85
choice is the power law model

( + z)1
(z) = . (3.138)
2(1 )

There are two physically distinct cases. If (1d/2) < 1 the correlations grow
with the distance z and the potential is called long-range. If (1 d/2) > 1
the correlations decay with the distance and the potential is called short-
range. The statics of these models, studied with the Gaussian Variational
method complemented by the replica method [24], is solved with a one step
replica symmetric ansatz in the short-range case and with a full replica sym-
metry breaking scheme in the long-range one.

86
4 The Langevin equation
Examples of experimental and theoretical interest in condensed matter
and biophysics in which quantum fluctuation can be totally neglected are
manifold. In this context one usually concentrates on systems in contact
with an environment: one selects some relevant degrees of freedom and treats
the rest as a bath. It is a canonical view. Among these instances are col-
loidal suspensions which are particles suspended in a liquid, typically salted
water, a soft condensed matter example; spins in ferromagnets coupled to
lattice phonons, a hard condensed matter case; and proteins in the cell a
biophysics instance. These problems are modeled as stochastic processes
with Langevin equations, the Kramers-Fokker-Planck formalism or master
equations depending on the continuous or discrete character of the relevant
variables and analytic convenience.
The Langevin equation is a stochastic differential equation that describes
phenomenologically a large variety of problems. It models the time evolution
of a set of slow variables coupled to a much larger set of fast variables that
are usually (but not necessarily) assumed to be in thermal equilibrium at a
given temperature. We first introduce it in the context of Brownian motion
and we derive it in more generality in Sect. 2.
The Langevin equation5 for a particle moving in one dimension in contact
with a white-noise bath reads

mv + 0 v = F + , v = x , (4.1)

with x and v the particles position and velocity. is a Gaussian white noise
with zero mean and correlation h(t)(t)i = 20 kB T (t t ) that mimics
thermal agitation. 0 v is a friction force that opposes the motion of the
particle. The force F designates all external deterministic forces and depends,
in most common cases, on the position of the particle x only. In cases in which
the force derives from a potential, F = dV /dx. The generalization to higher
dimensions is straightforward. Note that 0 is the parameter that controls
the strength of the coupling to the bath (it appears in the friction term as
5
P. Langevin, Sur la thorie du mouvement brownien, Comptes-Rendus de lAcadmie
des Sciences 146 (1908), 530-532.

87
well as in the noise term). In the case 0 = 0 one recovers Newton equation
of motion. The relation between the friction term and thermal correlation is
non-trivial. Langevin fixed it by requiring hv 2(t)i hv 2 ieq . We shall give a
different argument for it in the next section.

4.1 Derivation of the Langevin equation

Let us take a system in contact with an environment. The interacting
system+environment ensemble is closed while the system is open. The
nature of the environment, e.g. whether it can be modeled by a classical or a
quantum formalism, depends on the problem under study. We focus here on
the classical problem. A derivation of a generalized Langevin equation with
memory is very simple starting from Newton dynamics of the full system [40,
32]. We shall then study the coupled system
env .
Htot = Hsyst + Henv + Hint + Hcounter = Hsyst + H (4.2)

For simplicity we use a single particle moving in d = 1: Hsyst is the Hamil-

tonian of the isolated particle,

p2
Hsyst = + V (x) , (4.3)
2M
with p and x its momentum and position. Henv is the Hamiltonian of a
thermal bath that, for simplicity, we take to be an ensemble of N independent
Harmonic oscillators with masses ma and frequencies a , a = 1, . . . , N
N
X a2 ma a2 2
Henv = + q (4.4)
a=1 2ma 2 a

with a and qa their momenta and positions. This is indeed a very usual
choice since it may represent phonons. Hint is the coupling between sys-
tem and environment. We shall restrict the following discussion to a linear
interaction in the oscillator coordinates, qa , and in the particle coordinate,
N
X
Hint = x ca qa , (4.5)
a=1

with ca the coupling constants. The counter-term Hcounter is added to avoid

the generation of a negative harmonic potential on the particle due to the

88
coupling to the oscillators that may render the dynamics unstable. We
choose it to be
N
1X c2a 2
Hcounter = x . (4.6)
2 a ma a2
The generalization to more complex systems and/or to more complicated
baths and higher dimensions is straightforward. The calculations can also
be easily generalized to an interaction of the oscillator coordinate with a
more complicated dependence on the systems coordinate, V(x), that may
be dictated by the symmetries of the system, see Ex. 1.
Hamiltons equations for the particle are
N N
p(t) c2a
= V [x(t)]
X X
x(t)
= , p(t) ca qa (t) 2
x(t)(4.7)
m a=1 a=1 ma a

(the counter-term yields the last term) while the dynamic equations for each
member of the environment read
a (t)
qa (t) = , a (t) = ma a2 qa (t) ca x(t) , (4.8)
ma
showing that they are all massive harmonic oscillators forced by the chosen
particle. These equations are readily solved by
a (0) ca t
Z
qa (t) = qa (0) cos(a t)+ sin(a t) dt sin[a (tt )]x(t ) (4.9)
ma a ma a 0

with qa (0) and a (0) the initial coordinate and position at time t = 0 when
the particle is set in contact with the bath. It is convenient to integrate by
parts the last term. The replacement of the resulting expression in the last
term in the rhs of eq. (4.7) yields
Rt
= V [x(t)] + (t)
p(t) 0 dt (t t )x(t
) , (4.10)

with the symmetric and stationary kernel given by

PN c2a
(t t ) = a=1 ma a2 cos[a (t t )] , (4.11)

(t t ) = (t t), and the time-dependent force given by

h   i
PN a (0) ca x(0)
(t) = a=1 ca ma a
sin(a t) + qa (0) + ma a2
cos(a t) . (4.12)

89
This is the equation of motion of the reduced system.
The third term on the rhs of eq. (4.10) represents a rather complicated
friction force. Its value at time t depends explicitly on the history of
the particle at times 0 t t and makes the equation non-Markovian.
One can rewrite it as an integral running up to a total time T > max(t, t )
introducing the retarded friction:

(t t ) = (t t )(t t ) . (4.13)

Until this point the dynamics of the system remain deterministic and
are completely determined by its initial conditions as well as those of the
reservoir variables. The statistical element comes into play when one
realizes that it is impossible to know the initial configuration of the large
number of oscillators with great precision and one proposes that the initial
coordinates and momenta of the oscillators have a canonical distribution at
an inverse temperature . Then, one chooses {a (0), qa (0)} to be initially
distributed according to a canonical phase space distribution:

P ({a (0), qa (0)}, x(0)) = 1/Zenv [x(0)] e Henv [{a (0),qa (0)},x(0)] (4.14)

env = Henv + Hint + Hcounter , that can be rewritten as

with H
!2
N
env m 2
a a qa (0) +
ca 2 (0)
+ a .
X
H = x(0) (4.15)
a=1 2 ma a2 2ma

The randomness in the initial conditions gives rise to a random force acting
on the reduced system. Indeed, is now a Gaussian random variable,
that is to say a noise, with

h(t)i = 0 , h(t)(t )i = kB T (t t ) . (4.16)

One can easily check that higher-order correlations vanish for an odd number
of factors and factorize as products of two time correlations for an even
number of factors. In consequence has Gaussian statistics. Defining the
R t
inverse of over the interval [0, t], 0 dt (t t )1 (t t ) = (t t ), one
has the Gaussian pdf:
1
Rt Rt
1 2kB T dt dt (t)1 (tt )(t )
P [] = Z e 0 0 . (4.17)

90
Z is the normalization. A random force with non-vanishing correlations on
a finite support is usually called a coloured noise. Equation (4.10) is now
a genuine Langevin equation. A multiplicative retarded noise arises from a
model in which one couples the coordinates of the oscillators to a generic
function of the coordinates of the system, see Ex. 1 and eq. (4.24).
The use of an equilibrium measure for the oscillators implies the re-
lation between the friction kernel and the noise-noise correlation, which are
proportional, with a constant of proportionality of value kB T . This is a gen-
eralized form of the fluctuation-dissipation relation, and it applies to the
environment.
Different choices of the environment are possible by selecting different
ensembles of harmonic oscillators. The simplest one, that leads to an ap-
proximate Markovian equation, is to consider that the oscillators are coupled
to the particle via coupling constants ca = ca / N with ca of order one. One
defines
c2a
S() N1 N a=1 ma a ( a ) (4.18)
P

a function of , of order one with respect to N, and rewrites the kernel as

R S()
(t t ) = 0 d
cos[(t t )] . (4.19)

A common choice is
|| 1
   
S() ||

= 20

fc
. (4.20)

The function fc (x) is a high-frequency cut-off of typical width and is usually

chosen to be an exponential. The frequency is a reference frequency
that allows one to have a coupling strength 0 with the dimensions of vis-
cosity. If = 1, the friction is said to be Ohmic, S()/ is constant when
|| as for a white noise. When > 1 ( < 1) the bath is superOhmic
(subOhmic). The exponent is taken to be > 0 to avoid divergencies at
low frequency. For the exponential cut-off the integral over yields
cos[ arctan(t)]
+1
(t) = 20 a() (4.21)
[1 + (t)2 ]/2
with a(x) the Gamma-function, that in the Ohmic case = 1 reads

(t) = 20 , (4.22)
[1 + (t)2 ]

91
and in the limit becomes a delta-function, (t) 20 (t). At long
times, for any > 0 and different from 1, one has

+1 cos(/2)a() 1 t1 ,
lim (t) = 20 (4.23)
t

a power law decay.

Time-dependent, f (t), and constant non-potential forces, f np , as the ones
applied to granular matter and in rheological measurements, respectively,
are simply included in the right-hand-side (rhs) as part of the deterministic
force. When the force derives from a potential, F (x, t) = dV /dx.
In so far we have discussed systems with position and momentum degrees
of freedom. Other variables might be of interest to describe the dynamics
of different kind of systems. In particular, a continuous Langevin equation
for classical spins can also be used if one replaces the hard Ising constraint,
si = 1, by a soft one implemented with a potential term of the form V (si ) =
u(s2i 1)2 with u a coupling strength (that one eventually takes to infinity
to recover a hard constraint). The soft spins are continuous unbounded
variables, si (, ), but the potential energy favors the configurations
with si close to 1. Even simpler models are constructed with spherical
spins, that are also continuous unbounded variables globally constrained to
satisfy N 2
i=1 si = N. The extension to fields is straightforward and we shall
P

discuss one when dealing with the O(N) model.

Exercise Prove that or a non-linear coupling Hint = V[x] a ca qa there is a
P

choice of counter-term for which the Langevin equation reads

Z t

= V [x(t)] + (t)V [x(t)] V [x(t)]
p(t) dt (t t )V [x(t )]x(t
) (4.24)
0

with the same as in eq. (4.11) and (t) given by eq. (4.12) with x(0)
V[x(0)]. The noise appears now multiplying a function of the particles
coordinate.
Another derivation of the Langevin equation uses collision theory and
admits a generalization to relativistic cases [41].

4.2 Properties

4.2.1 Irreversibility and dissipation.

92
 The friction force 0 v in eq. (4.1) or its retarded extension in the
non-Markovian case explicitly breaks time-reversal (t t) invariance, a
property that has to be respected by any set of microscopic dynamic equa-
tions. Newton equations describing the whole system, the particle and all
the molecules of the fluid, are time reversal invariant. However, time-reversal
can be broken in the reduced equation in which the fluid is treated in an
effective statistical form and the fact that it is in equilibrium is assumed from
the start.
Even in the case in which all forces derive from a potential, F = dV /dx,
the energy of the particle, mv 2 /2 + V , is not conserved and, in general, flows
to the bath leading to dissipation. At very long times, however, the particle
may reach a stationary regime in which the particle gives and receives energy
from the bath at equal rate, on average.
Exercise Prove the time-irreversibility of the Langeving equation and the
fact that the symmetry is restored if 0 = 0. Show that dhHi/dt 6= 0 when
0 6= 0.

4.2.2 Discretization of stochastic differential equations

The way in which a stochastic differential equation with white noise is to
be discretized is a subtle matter that we shall not discuss in these lectures,
unless where it will be absolutely necessary. There are basically two schemes,
called the It and Stratonovich calculus, that are well documented in the
literature.
In short, we shall use a prescription in which the pair velocity-position of
the particle at time t + , with an infinitesimal time-step, depends on the
pair velocity-position at time t and the value of the noise at time t.

4.2.3 Markov character

In the case of a white noise (delta correlated) the full set of equations
defines a Markov process, that is a stochastic process that depends on its
history only through its very last step.

4.2.4 Generation of memory

The Langevin equation (4.1) is actually a set of two first order differential
equations. Notice, however, that the pair of first-order differential equations

93
could also be described by a single second-order differential equation:

m
x + 0 x = F + . (4.25)

Having replaced the velocity by its definition in terms of the position x(t)
depends now on x(t ) and x(t 2). This is a very general feature: by
integrating away some degrees of freedom (the velocity in this case) one gen-
erates memory in the evolution. Generalizations of the Langevin equation,
such as the one that we have just presented with colored noise, and the ones
that will be generated to describe the slow evolution of super-cooled liquids
and glasses in terms of correlations and linear responses, do have memory.

4.2.5 Smoluchowski (overdamped) limit

In many situations in which friction is very large, the characteristic time
for the relaxation of the velocity degrees of freedom to their Maxwellian
distribution, tvr , is very short (see the examples in Sect. 2.3). In consequence,
observation times are very soon longer than this time-scale, the inertia term
mv can be dropped, and the Langevin equation becomes

0 x = F + (4.26)

(for simplicity we wrote the white-noise case). Indeed, this overdamped

limit is acceptable whenever the observation times are much longer than the
characteristic time for the velocity relaxation. Inversely, the cases in which
the friction coefficient 0 is small are called underdamped.
In the overdamped limit with white-noise the friction coefficient 0 can
be absorbed in a rescaling of time. One defines the new time

t = 0 (4.27)

the new position, x( ) = x(0 ), and the new noise ( ) = (0 ). In

the new variables the Langevin equation reads x ( ) = F (
x, ) + ( ) with

h( )( )i = 2kB T ( ).

4.3 The basic processes

We shall discuss the motion of the particle in some 1d representative
potentials: under a constant force, in a harmonic potential, in the flat limit

94
 V

V
x x x

Figure 14: Three representative one-dimensional potentials.

of these two (Fig. 14) and the escape from a metastable state and the motion
in a double well potential (Fig. 17).

4.3.1 A constant force

Let us first consider the case of a constant force, F . The first thing to
notice is that the Maxwell-Boltzmann measure
 
v2
2m
+V (x)
Pgb (v, x) e (4.28)
is not normalizable if the size of the line is infinite, due to the exp[V (x)] =
exp(F x) term. Let us then study the evolution of the particles velocity and
position to show how these variables behave and the fact that they do very
differently.
The problem to solve is a set of two coupled stochastic first order differ-
ential equations on (v(t)x(t)), one needs two initial conditions v0 and x0 .
The velocity
The time-dependent velocity follows from the integration of eq. (4.1) over
time
1 t 0 (tt )
0
Z
v(t) = v0 e m t + dt e m [ F + (t ) ] , v0 v(t = 0) .
m 0
The velocity is a Gaussian variable that inherits its average and correla-
tions from the ones of . Using the fact that the noise has zero average
0 F  0 
hv(t)i = v0 e m t + 1 e m t .
0
In the short time limit, t tvr = m/0 , this expression approaches the
Newtonian result (0 = 0) in which the velocity grows linearly in time v(t) =

95
v0 + F/m t. In the opposite long time limit, t tvr = m/0 , for all initial
conditions v0 the averaged velocity decays exponentially to the constant value
F/0 . The saturation when the bath is active (0 6= 0) is due to the friction
term. The two limiting values match at tvr 1. The relaxation time
separating the two regimes is
m
tvr = 0
. (4.29)

The velocity mean-square displacement is

kB T  0 
v2 (t) h(v(t) hv(t)i)2i = 1 e2 m t (4.30)
m
independently of F . This is an example of the regression theorem accord-
ing to which the fluctuations decay in time following the same law as the
average value. The short and long time limits yield
20
(
kB T t t tvr
v2 (t) 2
h(v(t) hv(t)i) i m (4.31)
m 1 t tvr

and the two expressions match at t tvr /2. The asymptotic limit is the result
expected from equipartition of the kinetic energy, h(v(t)hv(t)i)2i h(v(t)
hvistat )2 istat that implies for the kinetic energy hKistat = kB T /2 (only if the
velocity is measured with respect to its average). In the heuristic derivation of
the Langevin equation for F = 0 the amplitude of the noise-noise correlation,
say A, is not fixed. The simplest way to determine this parameter is to
require that equipartition for the kinetic energy holds A/(0m) = T /m and
hence A = 0 T . This relation is known under the name of fluctuation
dissipation theorem (fdt) of the second kind in Kubos nomenclature.
It is important to note that this fdt characterizes the surrounding fluid and
not the particle, since it relates the noise-noise correlation to the friction
coefficient. In the case of the Brownian particle this relation ensures that
after a transient of the order of tvr , the bath maintains the mean kinetic
energy of the particle constant and equal to its equilibrium value.
The velocity two-time connected correlation reads
kB T h 0 |tt | 0
i
h[v(t) hv(t)i][v(t) hv(t )i]i = e m e m (t+t ) .
m
This is sometimes called the Dirichlet correlator. This and all other
higher-order velocity correlation functions approach a stationary limit when

96
 F/0 kB T /m

hvi

v2
v0 (a) (b)
0
tvr tvr
t t

Figure 15: Results for the constant force problem. (a) Mean velocity as
a function of time. (b) Velocity mean-square displacement as a function
of time. In both cases the linear behavior at short times, t tvr and the
saturation values are shown.

the shortest time involved is longer than tvr . At t = t on recovers the mean-
square displacement computed in eq. (4.30). When both times are short
compared to tvr the two-time correlator behaves as 2kB T 0 /m2 max(t, t ).
When at least one of the two times is much longer than tvr the second term
vanishes and one is left with an exponential decay as a function of time delay:
kB T 0 |tt |
c
Cvv (t, t ) h[v(t) hv(t)i][v(t ) hv(t )ii e m t, t tvr .
m
(4.32)
The two-time connected correlation falls off to, say, 1/e in a decay time
tvd = m/0 . In this simple case tvr = tvd but this does not happen in more
complex cases.
More generally one can show that for times t1 t2 . . . tn tvr :

hv(t1 + ) . . . v(tn + )i = hv(t1 ) . . . v(tn )i (TTI) (4.33)

for all delays . Time-translation invariance (TTI) or stationarity is

one generic property of equilibrium dynamics. Another way of stating
(4.33) is
hv(t1 ) . . . v(tn )i = f (t1 t2 , . . . , tn1 tn ) . (4.34)
Another interesting object is the linear response of the averaged velocity
to a small perturbation applied to the system in the form of V V hx,

97
i.e. a change in the slope of the potential in this particular case. One finds

hv(t)ih  1 0 (tt )
Rvx (t, t )
= e m (t t ) (4.35)
h(t ) h=0 m


1
h[v(t) hv(t)i][v(t ) hv(t )i]i (t t ) (4.36)
kB T
the last identity being valid in the limit t or t tvr . This is an fdt relation
between a linear response, Rvx (t, t ), and a connected correlation, Cvv c
(t, t ),
that holds for one of the particle variables, its velocity, when this one reaches
the stationary state.

kB T Rvx (t, t ) = Cvv

c
(t, t ) (t t ) (FDT) . (4.37)

In conclusion, the velocity is a Gaussian variable that after a character-

istic time tvr verifies equilibrium-like properties: its average converges to
a constant (determined by F ), its multi-time correlation functions become
stationary and a fluctuation-dissipation theorem links its linear response to
the connected correlation at two times. More general FDTs are discussed in
the exercise proposed below.
The position
Rt
The particles position, x(t) = x0 + 0 dt v(t ) is still a Gaussian random
variable:
!
F F 0
x(t) = x0 + v0 tvr + (t tvr ) + tvr v0 e m t
0 0
1
Z t Z t 0 )
+ dt dt e m (t t (t ) . (4.38)
m 0 0

Its noise-average behaves as the Newtonian result, ballistic motion, hx(t)i =

x0 + v0 t + F/(2m) t2 at short times t tvr and it crossover to

F
hx(t)i x0 + v0 tvr + 0
(t tvr ) (4.39)

for t tvr . Note the reduction with respect to ballistic motion (x F t2 )

due to the friction drag and the fact that this one-time observable does not
saturate to a constant.

98
 The position mean-square displacement approaches

kB T
x2 (t) h(x(t) hx(t)i)2 i 2Dx t with Dx 0
(Diffusion) (4.40)

in the usual t tvr limit, that is to say normal diffusion with the diffusion
constant Dx . This expression can be computed using x(t) hx(t)i as ob-
tained from the v(t)hv(t)i above (and it is quite a messy calculation) or one
can simply go to the Smoluchowski limit, taking advantage of the knowledge
of what we have just discussed on the behaviour of velocities, and obtain
diffusion in two lines. In contrast to the velocity mean-square displacement
this quantity does not saturate at any finite value. Similarly, the particle
displacement between two different times t and t is

xx (t, t ) h[x(t) x(t )]2 i 2Dx |t t | . (4.41)

It is interesting to note that the force dictates the mean position but it does
not modify the fluctuations about it (similarly to what it did to the velocity).
xx is stationary for time lags longer than tvr .
The two-time position-position connected correlation reads
c
Cxx (t, t ) = h(x(t) hx(t)i)(x(t ) hx(t )i)i = (4.42)

Exercise: compute it.

2T 01 01
Rxx
Cxv

0 (a) (b)
0
tvr tvr
t t t

Figure 16: Results for the constant force problem. (a) The correlation be-
tween the position and the velocity of the particle measured at different times.
(b) The linear response of the position to a kick applied linearly to itself at
a previous time. In both cases the linear behavior at short times, t tvr and
the saturation values are shown.

99
 Another way to measure the diffusion coefficient directly from the velocity
that is commonly used in the literature is
R
Dx = lim limt 0 dt hv( + t )v(t )i . (4.43)

One can check that it gives the same result.

The linear response of the particles position to a kick linearly applied to
itself at a previous time, in the form V V hx at t < t, is

hx(t)ih  1 0
Rxx (t, t )
= [1 e m (tt ) ] (t t ) , (4.44)
h(t ) h=0 0



with the limits

(
m1 (t t ) (t t ) t t tvr ,
Rxx (t, t ) (4.45)
01 (t t ) t t tvr .

A simple calculation proves that in the short time-differences limit this is the
results for Newton dynamics (Exercise: Show it.)
The correlation between the position and the velocity reads
" ! #
2kB T m m 0 t
h(x(t) hx(t)i)(v(t ) hv(t )i)i = 1+ e m
m 0 0
2kB T
(4.46)
0

and it is only a function of t . One notices that in the asymptotic limit in

which both sides of the equation saturate

2kB T Rxx (t, t ) = Cxv

c
(t, t ) for t t tvr and t tvr , (4.47)

with a factor of 2 different from the relation in eq. (4.37).

In conclusion, the position is also a Gaussian variable but it is explicitly
out of equilibrium. Its average and variance grow linearly in time, the latter
as in normal diffusion, and the fluctuation-dissipation relation has an addi-
tional factor of 1/2 (or 2, depending on on which side of the equality one
writes it) with respect to the form expected in equilibrium.
The energy

100
 The averaged potential energy diverges in the long-time limit since the
potential is unbounded in the x limit: hH(t)i = kB T /2 F hx(t)i
kB T /2 + F/0 t for t tvr .
Two kinds of variables
This example shows that even in this very simple problem the velocity and
position variables have distinct behaviour: the former is in a sense trivial,
after the transient tvr and for longer times, all one-time functions of v
F/0 saturate to their equilibrium values and the correlations are stationary.
Instead, the latter remains non-trivial and evolving out of equilibrium. One
can loosely ascribe the different behaviour to the fact that the velocity feels
a confining potential K = mv 2 /2 while the position feels an unbounded
potential V = F x in the case in which a force is applied, or a flat potential
V = 0 if F is switched off. In none of these cases the potential is able to take
the particle to equilibrium with the bath. The particle slides on the slope
and its excursions forward and backward from the mean get larger and larger
as time increases.
Quite generally, the classical problems we are interested in are such that
the friction coefficient 0 is large and the inertia term can be neglected, in
other words, all times are much longer than the characteristic time tvr . We
shall do it in the rest of the lectures.
Ergodicity
The ergodic hypothesis states that, in equilibrium, one can exchange en-
semble averages by time averages and obtain the same results. Out of equi-
librium this hypothesis is not expected to hold and one can already see how
dangerous it is to take time-averages in these cases by focusing on the simple
velocity variable. Ensemble and time averages coincide if the time-averaging
is done over a time-window that lies after tvr but it does not if the integration
time-interval goes below tvr .
Tests of equilibration have to be done very carefully in experiments and
simulations. One can be simply mislead by, for instance, looking just at the
velocities statistics.
A measure for the time dependent fluctuating position and velocity can be
written down, taking advantage of the fact that both variables are Gaussian:
1
 Z Z 
P (v, x) exp dt dt y t(t)A(t, t )y(t) (4.48)
2

101
with the 2 2 matrix A being the inverse of the matrix of correlations,
A1 ij (t, t ) = hyi (t)yj (t )i with i, j = 1, 2, y t(t) = (v(t) x(t)) and v(t) =
v(t)hv(t)i (similarly for x). The correlations are given above so the dynamic
pdf can be easily constructed.
Exercise. Confront

hv m (t)xn (t)xk (t )i and hv m (t)xn (t)kxk1 (t )v(t )i ; (4.49)

conclude.
Effect of a colored bath: anomalous diffusion
The anomalous diffusion of a particle governed by the generalized
Langevin equation, eq. (4.10), with colored noise characterized by power-
law correlations, eq. (4.11), a problem also known as fractional Brownian
motion, was studied in detail by N. Pottier [?]. The particles velocity equi-
librates with the environment although it does at a much slower rate than
in the Ohmic case: its average and mean-square displacement decay as a
power law - instead of exponentially - to their asymptotic values (still sat-
isfying the regression theorem). The particles mean square displacement
is determined by the exponent of the noise-noise correlation, hx2 (t)i t ,
i.e. the dynamics is subdiffusive for < 1, diffusive for = 1 and
superdiffusive for > 1. A time-dependent diffusion coefficient verifies
Dx (t) 1/2 dhx2 (t)i/dt t1 : it is finite and given by eq. (4.41) for nor-
mal diffusion, it diverges for superdiffusion and it vanishes for subdiffusion.
The ratio between the linear response and the time-derivative of the corre-
lation ratio reads T Rxx (t, t )/t Cxx (t, t ) = Dx (t t )/[Dx (t t ) + Dx (t )].
It approaches 1/2 for normal diffusion and the two-time dependent function
1/[1 + (t /(t t ))1 ] in other cases.

4.3.2 Relaxation in a quadratic potential

Another relevant example is the relaxation of a particle in a harmonic
potential, with its minimum at x 6= 0:
k
V (x) = (x x )2 , (4.50)
2
in contact with a white noise. The potential confines the particle and one
can then expect the coordinate to reach an equilibrium distribution.

102
 This problem can be solved exactly keeping inertia for all values of 0
but the calculation is slightly tedious. The behaviour of the particle velocity
has already been clarified in the constant force case. We now focus on the
overdamped limit,
0 x = k(x x ) + , (4.51)
with k the spring constant of the harmonic well, that can be readily solved,
Z t
k t k (tt )
x(t) = x0 e 0 + 01 dt e 0 [(t ) + kx ] , x0 = x(0) . (4.52)
0

This problem become formally identical to the velocity dependence in the

previous example.
Convergence of one-time quantities
The averaged position is

k t
hx(t) x i = (x0 x )e 0 0 (Convergence) txr 0 /k
(4.53)
Of course, one-time quantities should approach a constant asymptotically if
the system equilibrates with its environment.
Two-time quantities
The two-time connected correlation (where one extracts, basically, the
asymptotic position x ) reads
 
1 k (t+t ) 2 k min(t,t )
hx(t)x(t )i = kB T k e 0 e 0 1 . (4.54)

Again, the Dirichlet correlator (x(t) = x(t) hx(t)i). For at least one of
the two times going well beyond the position relaxation time txr = 0 /k the
memory of the initial condition is lost and the connected correlation becomes
stationary:
k |tt |
Cc (t, t ) = hx(t)x(t )i kB T k 1 e 0 min(t, t ) txr . (4.55)

For time-differences that are longer than txd = 0 /k the correlation decays to
1/e and one finds txd = txr . Interestingly enough, the relaxation and decay
times diverge when k 0 and the potential becomes flat.

103
 Note that when the time-difference t t diverges the average of the
product factorizes, in particular, for the correlation one gets

hx(t)x(t )i hx(t)ihx(t )i x hx(t )i (4.56)

for any t , even finite. We shall see this factorization property at work later
in more complicated cases.
Fluctuation-dissipation theorem (FDT)
One can also compute the linear response to an infinitesimal perturbation
that couples linearly to the position changing the energy of the system as
H H f x at a given time t :

hx(t)if 
R(t, t ) = . (4.57)
f (t ) f =0


The explicit calculation yields

1
(tt )
R(t, t ) = 01 ek0 (t t )
1 Cc (t,t )
R(t, t ) = kB T t
(t t ) (FDT) (4.58)

The last equality holds for times that are longer than txr . It expresses the
fluctuation-dissipation theorem (fdt), a model-independent relation be-
tween the two-time linear response and correlation function. Similar - though
more complicated - relations for higher-order responses and correlations also
exist in equilibrium. There are many ways to prove the fdt for stochas-
tic processes. We shall discuss one of them in Sect. 4.1 that is especially
interesting since it applies easily to problems with correlated noise.
It is instructive to examine the relation between the linear response and
the correlation function in the limit of a flat potential (k 0). The linear
response is just 01 (tt ). The Dirichlet correlator approaches the diffusive
limit:
hx(t)x(t )i = 201 kB T min(t, t ) for k0 (4.59)
and its derivative reads t hx(t)x(t )i = 201 kB T (t t ). Thus,
1
R(t, t ) = t hx(t)x(t )i (t t )
2kB T
1
R(t, t ) = 2kB T
t Cc (t, t ) (t t ) (FDR for diffusion) (4.60)

104
A factor 1/2 is now present in the relation between R and Cc . It is an-
other signature of the fact that the coordinate is not in equilibrium with the
environment in the absence of a confining potential.
Exercise Evaluate the two members of the FDT, eq. (4.58), in the case of
the tilted potential V (x) = F x.
Reciprocity or Onsager relations
Let us compare the two correlations hx3 (t)x(t )i and hx3 (t )x(t)i within
the harmonic example. One finds hx3 (t)x(t )i = 3hx2 (t)ihx(t)x(t )i and
hx3 (t )x(t)i = 3hx2 (t )ihx(t )x(t)i. Given that hx2 (t)i = hx2 (t )i hx2 ieq
and the fact that the two-time self-correlation is symmetric,

hx3 (t)x(t )i = hx3 (t )x(t)i . (4.61)

With a similar argument one shows that for any functions A and B of x:

hA(t)B(t )i = hA(t )B(t)i

CAB (t, t ) = CAB (t , t) (Reciprocity) (4.62)

This equation is known as Onsager relation and applies to A and B that

are even under time-reversal (e.g. they depend on the coordinates but not
on the velocities or they have an even number of verlocities).
All these results remain unaltered if one adds a linear potential F x and
works with connected correlation functions.

4.3.3 Thermally activated processes

The phenomenological Arrhenius law 6 yields the typical time needed to
escape from a potential well as an exponential of the ratio between the height
of the barrier and the thermal energy scale kB T , (with prefactors that can be
calculated explicitly, see below). This exponential is of crucial importance
for understanding slow (glassy) phenomena, since a mere barrier of 30kB T
is enough to transform a microscopic time of 1012 s into a macroscopic time
scale. See Fig. 17-right for a numerical study of the Coulomb glass that
demonstrates the existence of an Arrhenius time-scale in this problem. In
the glassy literature such systems are called strong glass formers as opposed
to weak ones in which the characteristic time-scale depends on temperature

105
 V

V
x x

Figure 17: Left: sketch of a double-well potential. Center: sketch of a

potential with a local minimum. Right: correlation function decay in a
classical model of the 3d Coulomb glass at nine temperatures ranging from
T = 0.1 to T = 0.05 in steps of 0.05 and all above Tg . In the inset the
scaling plot C(t) f (t/tA ) with a characteristic time-scale, tA , that follows
the Arrhenius activated law, tA 0.45/T . Figure taken from [25].

in a different way.
In 1940 Kramers estimated the escape rate from a potential well as the
one shown in Fig. 17-center due to thermal fluctuations that give sufficient
energy to the particle to allow it to surpass the barrier 7 . After this seminal
paper this problem has been studied in great detail [41] given that it is of
paramount importance in many areas of physics and chemistry. An example
is the problem of the dissociation of a molecule where x represents an effective
one-dimensional reaction coordinate and the potential energy barrier is,
actually, a free-energy barrier.
Kramers assumed that the reaction coordinate is coupled to an equili-
brated environment with no memory and used the probability formalism in
which the particle motion is described in terms of the time-dependent proba-
bility density P (x, v, t) (that for such a stochastic process follows the Kramers
partial differential equation).
If the thermal energy is at least of the order of the barrier height, kB T
V , the reaction coordinate, x, moves freely from the vicinity of one well to
the vicinity of the other.
6
S. A. Arrhenius, On the reaction velocity of the inversion of cane sugar by acids,
Zeitschrift fr Physikalische Chemie 4, 226 (1889).
7
H. A. Kramers, Brownian motion in a field of force and the diffusion model of chemical
reactions Physica 7, 284 (1940).

106
 The treatment we discuss applies to the opposite weak noise limit in which
the thermal energy is much smaller than the barrier height, kB T V , the
random force acts as a small perturbation, and the particle current over the
top of the barrier is very small. Most of the time x relaxes towards the
minimum of the potential well where it is located. Eventually, the random
force drives it over the barrier and it escapes to infinity if the potential has
the form in Fig. 17-center, or it remains in the neighbourhood of the second
well, see Fig. 17-left.
The treatment is simplified if a constant current can be imposed by inject-
ing particles within the metastable well and removing them somewhere to the
right of it. In these conditions Kramers proposed a very crude approximation
whereby P takes the stationary canonical form
v2
Pst (x, v) = N e 2
V (x)
. (4.63)

If there is a sink to the right of the maximum, the normalization constant

N is fixed by further assuming that Pst (x, v) 0 for x x > xmax . The
resulting integral over the coordinate can be computed with a saddle-point
approximation justified in the large limit. After expanding the potential
about the minimum and keeping the quadratic fluctuations one finds
2
N 1 = q eV (xmin ) .

V (xmin )

The escape rate, r, over the top of the barrier can now be readily computed
by calculating the outward flow across the top of the barrier:
q
1
Z V (xmin )
r dv vP (xmax , v) = e(V (xmax )V (xmin )) . (4.64)
tA 0 2
Note that we here assumed that no particle comes back from the right of the
barrier. This assumption is justified if the potential quickly decreases on the
right side of the barrier.
The crudeness of the approximation (4.63) can be grasped by noting that
the equilibrium form is justified only near the bottom of the well. Kramers
estimated an improved Pst (x, v) that leads to
1/2
2
 q

4
+ V (xmax ) 2 V (xmin )
r= q e(V (xmax )V (xmin )) . (4.65)
V (x
max )
2

107
This expression approaches (4.64) when V (xmax ), i.e. close to the
underdamped limit, and
q
V (xmax )V (xmin )
r= e(V (xmax )V (xmin )) (4.66)
2

when V (xmax ), i.e. in the overdamped limit (see Sect. 4.2.5 for the
definition of these limits).
The inverse of (4.65), tA , is called the Arrhenius time needed for ther-
mal activation over a barrier V V (xmax ) V (xmin ). The prefactor
that characterises the well and barrier in the harmonic approximation is the
attempt frequency with which the particles tend to jump over the barrier.
In short,
tA e|V | (Arrhenius time) (4.67)
The one-dimensional reaction coordinate can be more or less easily identi-
fied in problems such as the dissociation of a molecule. In contrast, such a sin-
gle variable is much harder to visualize in an interacting problem with many
degrees of freedom. The Kramers problem in higher dimensions is highly
non-trivial and, in the infinite-dimensional phase-space, is completely out
of reach.

108
5 Dynamics through a phase transition
Take a piece of material in contact with an external reservoir. The ma-
terial will be characterized by certain observables, energy density, magneti-
zation density, etc. The external environment will be characterized by some
parameters, like the temperature, magnetic field, pressure, etc. In principle,
one is able to tune the latter and study the variation of the former. Note
that we are using a canonical setting in the sense that the system under
study is not isolated but open.
Sharp changes in the behavior of macroscopic systems at critical points
(or lines) in parameter space have been observed experimentally. These cor-
respond to equilibrium phase transitions, a non-trivial collective phe-
nomenon appearing in the thermodynamic limit. We shall assume that the
main features of, and analytic approaches used to study, phase transitions
are known.
Imagine now that one changes an external parameter instantaneously
or with a finite rate going from one phase to another in the (equilibrium)
phase diagram. The kind of internal system interactions are not changed.
In the statistical physics language the first kind of procedure is called a
quench and the second one an annealing and these terms belong to the
metalurgy terminology. We shall investigate how the system evolves by trying
to accomodate to the new conditions and equilibrate with its environment.
We shall first focus on the dynamics at the critical point or going through
phase transitions between well-known phases (in the sense that one knows the
order parameter, the structure, and all thermodynamic properties on both
sides of the transition). Later we shall comment on cases in which one does
not know all characteristics of one of the phases and sometimes one does not
even know whether there is a phase transition.
The evolution of the free-energy landscape (as a function of an order
parameter) with the control parameter driving a phase transition is a guide-
line to grasp the dynamics following a quench or annealing from, typically, a
disordered phase to the phase transition or into the ordered phase. See Fig. 18
for a sketch. We shall discuss quenches to the phase transition and below
it. In the former case, the system can get to a critical point (Fig. 18-left)
in which the free-energy is metastable in the sense that its second deriva-

109
 upper critical upper critical
critical critical
lower critical lower critical
f

Figure 18: Left: second-order phase transition. Right: first order phase
transition.

tive vanishes (second order phase transition cases) or to a first-order phase

transition (Fig. 18-right) in which various minima are degenerate. In the
latter case the initial state becomes unstable, that is to say a maximum,
and the phase transition is of second-order (see Fig. 18-left) or metastable,
that is to say a local minimum, and the phase transition is of first order (see
Fig. 18-right) in the final externally imposed conditions.8 In the former case
the ordering process occurs throughout the material, and not just at
nucleation sites. Two typical examples are spinodal decomposition, i.e.
the method whereby a mixture of two or more materials can separate into
distinct regions with different material concentrations, or magnetic domain
growth in ferromagnetic materials. Instead, in the latter case, the stable
phase conquers the system through the nucleation of a critical localized
bubble via thermal activation and its further growth.
Having described the dependence of the free-energy landscape on the ex-
ternal parameters we now need to choose the microscopic dynamics of the
order parameter. Typically, one distinguishes two classes: one in which the
order parameter is locally conserved and another one in which it is not. Con-
served order parameter dynamics are found for example in phase separation
in magnetic alloys or inmiscible liquids. Ferromagnetic domain growth is an
exemple of the non-conserved case.

8
Strictly speaking metastable states with infinite life-time exist only in the mean-field
limit.

110
5.1 Time-dependent Ginzburg-Landau description
The kinetics of systems undergoing critical dynamics or an ordering pro-
cess is an important problem for material science but also for our generic
understanding of pattern formation in non-equilibrium systems. The late
stage dynamics is believed to be governed by a few properties of the systems
whereas material details should be irrelevant. Among these relevant prop-
erties one may expect to find the number of degenerate ground states, the
nature of the conservation laws and the hardness or softness of the domain
walls that is intimately related to the dimension of the order parameter.
Thus, classes akin to the universality ones of critical phenomena have been
identified. These systems constitute a first example of a problem with slow
dynamics. Whether all systems with slow dynamics, in particular structural
and spin glasses, undergo some kind of simple though slow domain growth is
an open question.

200 200 200

data data data

150 150 150

100 100 100

50 50 50

0 0 0
0 50 100 150 200 0 50 100 150 200 0 50 100 150 200

Figure 19: Monte Carlo simulations of a 2d Ising model. Three snapshots at

t = 1, 3 105 , 3 106 MCs after a quench to Tc .

Take a magnetic system, such as the ubiquitous Ising model with ferro-
magnetic uniform interactions, and quench it to its Curie point or into the
low temperature phase starting from a random initial condition. Classically,
the spins do not have an intrinsic dynamics; it is defined via a stochastic
rule of Glauber, Metropolis or similar type with or without locally conserved
magnetization. For the purpose of the following discussion it is sufficient to
focus on non-conserved local microscopic dynamics. Three snapshots taken
after times 1, 3 105 and 3 106 MCs in a critical and two sub-critical
quenches are shown in Figs. 19, 20, and 21.
Time-dependent macroscopic observables are then expressed in terms of
the values of the spins at each time-step. For instance, the magnetization

111
 200 200 200
data data data

150 150 150

100 100 100

50 50 50

0 0 0
0 50 100 150 200 0 50 100 150 200 0 50 100 150 200

Figure 20: Monte Carlo simulations of a 2d Ising model. Three snapshots at

t = 1, 3 105 , 3 106 MCs after a quench to 0.5 Tc . Thermal fluctuations
within the domains are visible.

density and its two-time self correlation function are defined as

N N
m(t) N 1 C(t, t ) N 1 h si (t)si (t ) i ,
X X
h si (t) i , (5.1)
i=1 i=1

where the angular brackets indicate an average over many independent runs
(i.e. random numbers) starting from identical initial conditions and/or aver-
ages over different initial configurations.
In critical quenches, patches with equilibrium critical fluctuations grow
in time but their linear extent never reaches the equilibrium correlation length
that diverges. Clusters of neighbouring spins pointing the same direction of
many sizes are visible in the figures and the structure is quite intricate with
clusters within clusters and so on and so forth. The interfaces look pretty
rough too.
In quenches into the ordered phase through a second order phase
transition the ferromagnetic interactions tend to align the neighbouring
spins in parallel direction and in the course of time domains of the two ordered
phases form and grow, see Fig. 22. At any finite time the configuration is such
that both types of domains exist. If one examines the configurations in more
detail one reckons that there are some spins reversed within the domains.
These errors are due to thermal fluctuations and are responsible of the fact
that the magnetization of a given configuration within the domains is smaller
than one and close to the equilibrium value at the working temperature
(apart from fluctuations due to the finite size of the domains). The total
magnetization, computed over the full system, is zero (up to fluctuating
time-dependent corrections that scale with the square root of the inverse

112
 200 200 200
data data data

150 150 150

100 100 100

50 50 50

0 0 0
0 50 100 150 200 0 50 100 150 200 0 50 100 150 200

Figure 21: Monte Carlo simulations. Three snapshots at t = 1, 3105 , 3106

MCs after a quench to 0.01 Tc . There is almost perfect order within the
domains (meq 1).

system size). The thermal averaged spin, hsi (t)i vanishes for all i and all
finite t, see below for a more detailed discussion of the time-dependence. As
time passes the typical size of the domains increases and the interfaces get
flatter in a way that we shall also discuss below.
Quenches across first order phase transitions will be discussed sep-
arately below.
In order to treat phase-transitions and the coarsening process analyti-
cally it is preferable to introduce a coarse-grained description in terms of a
continuous coarse-grained field,
1 X
(~x, t) si (t) , (5.2)
V iV~x

the fluctuating magnetization density. In a first approximation a Landau-

Ginzburg free-energy functional is introduced
c
Z  
d
F [] = d x [(~x, t)]2 + V [(~x, t)] . (5.3)
2
With the choice of the potential one distinguishes between a second order
and a first order phase transition. In the former case, the typical form is the
4 form:
V () = a4 + b(g)2 . (5.4)
The first term in eq. (5.3) represents the energy cost to create a domain wall
or the elasticity of an interface. The second term depends on a parameter,
g, and changes sign from positive at g > gc to negative at g < gc . Above the

113
Figure 22: Snapshot of the 2d Ising model at a number of Monte Carlo steps
after a quench from infinite to a subcritical temperature. Left: the up and
down spins on the square lattice are represented with black and white sites.
Right: the domain walls are shown in black.

critical point determined by b(gc ) = 0 it has a single minimum at = 0, at gc

it is flat at = 0 and below gc it has a double well structure with two minima,
= [b(g)/(2a)]1/2 = hieq (g), that correspond to the equilibrium states
in the ordered phase. Equation (5.3) is exact for a fully connected Ising
model where V () arises from the multiplicity of spin configurations that
contribute to the same (~x) = m. The order-parameter dependent free-
energy density reads f (m) = Jm2 hm + kB T {(1 + m)/2 ln[(1 + m)/2] +
(1 m)/2 ln[(1 m)/2] that close to the critical point where m 0 becomes
f (m) (kB T 2J)/2 m2 hm + kB T /12 m4 demonstrating the passage
from a harmonic form at kB T > kB Tc = 2J, to a quartic well at T = Tc , and
finally to a double-well structure at T < Tc .
Exercise. Prove the above.
With a six-order potential one can mimic the situation in the right panel
of Fig. 18.
When discussing dynamics one should write down the stochastic evolu-
tion of the individual spins and compute time-dependent averaged quantities
as the ones in (5.1). This is the procedure used in numerical simulations.
Analytically it is more convenient to work with a field-theory and an evolu-
tion equation of Langevin-type. This is the motivation for the introduction

114
of continuous field equations that regulate the time-evolution of the coarse-
grained order parameter. Ideally these equations should be derived from the
spin stochastic dynamics but in practice they are introduced phenomenolog-
ically. In the magnetic case as well as in many cases of interest, the domain

wall and interface dynamics can be argued to be overdamped (i.e. t tr ).
Two very similar approaches are used. Assuming T is only relevant to
determine the equilibrium coarse-grained field one uses the phenomenologi-
cal zero-temperature time-dependent Ginzburg-Landau equation or
model A in the classification of Hohenberg-Halperin deterministic equation
(~x, t) F []
= (5.5)
t (~x, t)
(the friction coefficient has been absorbed in a redefinition of time). Initial
conditions are usually chosen to be random with short-range correlations
[ (~x, 0)(~x, 0) ]ic = (~x ~x ) (5.6)
thus mimicking the high-temperature configuration ([. . .]ic represent the av-
erage over its probability distribution). The numeric solution to this equation
with the quartic potential and b < 0 shows that such a random initial condi-
tion evolves into a field configuration with patches of ordered region in which
the field takes one of the two values [b/(2a)]1/2 separated by sharp walls. It
ignores temperature fluctuations within the domains meaning that the field
is fully saturated within the domains and, consequently, one has access to the
aging part of the correlations only, see e.g. eq. (5.20). The phase transition
is controlled by the parameter b in the potential.
Another, similar approach, is to add a thermal noise to the former
(~x, t) F []
= + (~x, t) . (5.7)
t (~x, t)
This is the field-theoretical extension of the Langevin equation in which
the potential is replaced by the order-parameter-dependent funcitonal free-
energy in eq. (5.3) with a potential form with fixed parameters (independent
of T ). is a noise taken to be Gaussian distributed with zero mean and
correlations
h(~x, t)(~x , t )i = 2kB T d (~x ~x )(t t ) . (5.8)
The friction coefficient has been absorbed in a redefinition of time. For a
quartic potential a dynamic phase transition arises at a critical Tc ; above

115
Tc the system freely moves above the two minima and basically ignores the
double well structure while below Tc this is important. Within the growing
domains the field fluctuates about its mean also given by [b/(2a)]1/2 and
the fluctuations are determined by T . One can describe the rapid relaxation
at ties such that the domain walls do not move with this approach. This
formulation is better suited to treat critical and sub-critical dynamics in the
same field-theoretical framework.
These equations do not conserve the order parameter neither locally nor
globally. Extensions for cases in which it is conserved exist (model B). Cases
with vectorial or even tensorial order parameters can be treated similarly and
are also of experimental relevance, notably for vectorial magnets or liquid
crystals.

5.2 Relaxation and equilibration time

We wish to distinguish the relaxation time, tr , defined as the time
needed for a given initial condition to reach equilibrium in one of the (possibly
many equivalent) phases, from the decorrelation time, td , defined as the
time needed for a given configuration to decorrelate from itself. To lighten
the notation we do not signal out the variable that we study to study these
typical times (as we did with the velocity and position in the examples of
Sect. 4.3). We further define the reversal time, tR , as the time needed to
go from one to another of the equivalent equilibrium phases. We focus on
second-order phase transitions here.

5.2.1 Quench from T Tc to T > Tc

If one quenches the system to T > Tc the relaxation time, tr , needed
to reach configurations sampled by the Boltzmann measure depends on the
systems parameters but not on its size. Hence it is finite even for an infinite-
size system. Once a short transient overcome, the average of a local spin
approaches the limit given by the Boltzmann measure, hsi (t)i hsi ieq =
m = 0, for all i and all other more complex observables satisfy equilibrium
laws. The relaxation time is estimated to behave as |T Tc |zeq close to Tc ,
with the critical exponent characterizing the divergence of the equilibrium
correlation length, (T Tc ) , and zeq the equilibrium exponent that
links times and lengths, t1/zeq .

116
 The relaxation of the two-time self-correlation at T > Tc , when the time

t is chosen to be longer than tr , decays exponentially

lim

hsi (t)si (t )i e(tt )/td (5.9)
t tr

with a decorrelation time that increases with decreasing temperature and

close to (but still above) Tc diverges as the power law, td (T Tc )zeq .
The divergence of td is the manifestation of critical slowing down. The
asympotic value verifies

lim hsi (t)si (t )i = lim hsi (t)i lim hsi (t )i = hsi ieq hsi ieq = m2 = 0 ,
tt t tr ttr
t tr
(5.10)
cfr. eq. (4.56).

5.2.2 Quench from T Tc to T Tc

At or below Tc , coarsening from an initial condition that is not corre-
lated with the equilibrium state and with no bias field does not take
the system to equilibrium in finite times with respect to a function of the
systems linear size, L. More explicitly, if the growth law is a power law [see
eq. (5.28)] one needs times of the order of Lzeq (critical) or Lzd (subcrititcal)
to grow a domain of the size of the system. This gives a rough idea of the
time needed to take the system to one of the two equilibrium states. For
any shorter time, domains of the two types exist and the system is out of
equilibrium.
The self-correlation of such an initial state evolving at T Tc involves
power laws or logarithms and although one cannot associate to it a decay
time as one does to an exponential, one can still define a characteristic time
that, quite generally, turns out to be related to the age of the system, td tw
[see eq. (5.26)].
In contrast, the relaxation time of an equilibrium magnetized config-
uration at temperature T vanishes since the system is already equilibrated
while the decorrelation time td is a finite function of T .
The relaxation of the two-time self-correlation at T < Tc , when the time t
is chosen to be longer than tr , that is to say, once the system has thermalized
in one of the two equilibrium states, decays exponentially

hsi (t)si (t )i e(tt )/td (5.11)

117
with a decorrelation time that decreases with decreasing temperature and
close to Tc (but below it) also diverges as a power law, td (T Tc )zeq .
The asympotic value verifies

lim hsi (t)si (t )i = lim hsi (t)i lim hsi (t )i = hsi ieq hsi ieq = m2 0 ,
tt t tr ttr
t tr
(5.12)
cfr. eqs. (4.56) and (5.10), depending on T = Tc or T > Tc .

5.2.3 Summary
The lesson to learn from this comparison is that the relaxation time and
the decorrelation time not only depend upon the working temperature but
they also depend upon the initial condition. Moreover, in all critical or low-
temperature cases we shall study the relaxation time depends on (L, T )
and diverges in the infinite size limit while the decorrelation time depends
on (T, tw ). For a random initial condition and an infinite system one has


finite T > Tc ,
>
tr |T Tc |zeq T Tc ,

T Tc

while for a finite system

(
Lzeq T = Tc ,
tr
Lzd T < Tc .

Still another time scale is given by the time needed to reverse an equi-
librium configuration in the low-T phase. This one is expected to be given
by an Arrhenius law, with the height of the barrier being determined by
the extensive free-energy barrier between the two minima, i.e. F Ld f ,
therefore,
tR eL
df
Reversal time-scale . (5.13)
The Ginzburg-Landau description allows for a pictorial interpretation of
these results. The dynamics of the full system is visualized as the motion of
its representative point in the Ginzburg-Landau potential. At high T the po-
tential is harmonic in the deterministic Allen-Cahn equation, or the double-
well structure in the time-dependent stochastic Ginzburg-Landau equation
is completely ignored. The relaxation is similar to the one of a particle in a

118
harmonic potential studied in Sect. 4.3.2. At low T , the initial position in
the double-well potential depends on the type of initial condition (~x, 0) = 0
or (~x, 0) 6= 0. In the first case, the point sits on top of the central barrier
and it does not detach from it in finite times with respect to a function of
L. In the second case, the point starts from within one well and it simply
rolls to the bottom of the well. This relaxation is similar to the one in the
harmonic case. To reverse the configuration from, say, positive to negative
magnetization the point needs to jump over the barrier in the double well
potential and it does via thermal activation ruled by the Arrhenius law.
Note however that the phase-space of the system is actually N-dimensional
while the description that is given here is projected onto one single coordi-
nate, the one of the order-parameter. This reduction might lead to some
misunderstandings and one should be very careful with it.

5.3 Short-time dynamics

Take an initial configuration (~x, t) = 0 = 0 on average with small fluc-
tuations, as in equilibrium at very high temperature, and quench the system.
At very short time one can expand the non-linear potential and the GL equa-
~
tion (5.5), for the Fourier components, (~k, t) = Ld/2 dd x (~x, t)eik~x with
R

~k = 2/L (n1 , . . . , nd ) and nk integer, reads

(~k, t)
= [k 2 V (0)](~k, t) + (~k, t) . (5.14)
t
If V (0) > 0 all modes decay exponentially and no order develops. If V (0)
0 instead modes with k 2 V (0) > 0 are unstable and grow exponentially
until a time t 1/V (0) when the small expansion ceases to be justified.
The instability of the small wave-vector modes indicates that the system
tends to order. To go beyond this analysis one needs to consider the full
non-linear equation.

5.4 Growing length and dynamic scaling

In usual coarsening systems the averaged space-time correlation function
NC(r, t) = rj |=r hsi (t)sj (t)i
P
ri ~
ij/|~

119
allows for the identification of a growing length from, for example,
Z Z
d a+1
Ra (T, t) d rr C(r, t)/ dd r r a C(r, t) (5.15)

(a is a parameter chosen to weight preferentially short or long distances; the

time-dependence of Ra (t) should not depend on a.) Here and in the following
h. . .i stands for an average over different realizations of thermal histories at
heat-bath temperature T and/or initial conditions. In presence of quenched
disorder one adds an average over it and denotes it [. . .]. The stochastic time-
dependent function N 1 ij/|~ri~rj |=r si (t)sj (t) after a quench from a random
P

initial condition does not fluctuate in the thermodynamic limit. Therefore,

the averages are not really necessary but they are usually written down. In
spin-glasses and glasses this observable does not yield information on the
existence of any growing length as we shall discuss below.
The spherically averaged structure factor S(k, t) the Fourier transform
of C(r, t) can be measured experimentally with small-angle scattering of
neutrons, x-rays or light and from it Ra (T, t) can be extracted.
The ordering process is characterized by the growth of a typical length,
R(T, t). The growth regimes are summarized in the following equation and
in Fig. 23:

(T )

T > Tc saturation,
R(T, t) Rc (t) (T ) T Tc critical coarsening, (5.16)

R(T, t) > (T ) T < Tc sub-critical coarsening.
After a quench to the high temperature phase T > Tc the system first grows
equilibrium regions until reaching the correlation length and next relaxes in
equilibrium as explained in the previous section. The correlation length could
be very short and the transient non-equilibrium regime be quite irrelevant
(T Tc ). In the critical region, instead, the correlation length grows and it
becomes important. In a critical quench the system never orders sufficiently
and R(Tc , t) < for all finite times. Finally, a quench into the subcritical
region is characterized by two growth regimes: a first one in which the critical
point dominates and the growth is as in a critical quench; a second one in
which the proper sub-critical ordering is at work. The time-dependence of
the growth law is different in these two regimes as we shall see below. (Note
that below Tc does not measure the size of ordered regions but the typical
distance until which a fluctuation has an effect.)
In the asymptotic time domain, when R(T, t) has grown much larger than
any microscopic length in the system, a dynamic scaling symmetry sets

120
 R(T, t)

Tc

Figure 23: Sketch of the growth process in a second-order phase transition.

The thick line is the equilibrium correlation length |T Tc | . The thin
arrows (solid red and dashed black) indicate the growing length R in the
critical coarsening and sub-critical coarsening regimes.

in, similarly to the usual scaling symmetry observed in equilibrium critical

phenomena. According to this hypothesis, the growth of R(T, t) is the only
relevant process and the whole time-dependence enters only through R(T, t).

5.5 Critical coarsening

The scaling behavior of binary systems quenched to the critical point
is quite well understood. It can be addressed with scaling arguments and
renormalization group approaches [27] which give explicit expressions for
many of the quantities of interest up to two loops order. Numerical simula-
tions confirm the analytic results and probe exponents and scaling functions
beyond the available perturbative orders. In this case the system builds cor-
related critical clusters with fractal dimension D = (d + 2 )/2, where
is the usual static critical exponent, in regions growing algebraically as

121
R(Tc , t) R(t) t1/zeq ; henceforth we simplify the notation and avoid writ-
ing Tc within R.
In the asymptotic time regime the space-time correlation function has the
scaling form
 
r
C(r, t) = R(t)2(dD) f R(t)
Multiplicative separation. (5.17)

The pre-factor R(t)2(dD) takes into account that the growing domains have
a fractal nature (hence their density decreases as their size grows) and the
dependence on r/R(t) in f (x) expresses the similarity of configurations at
different times once lengths are measured in units of R(t). At distances
and times such that r/R(t) 1 the equilibrium power-law decay, Ceq (r)
r 2d , should be recovered, thus f (x) x2(dD) at x 0. f (x) falls off
rapidly for x 1 to ensure that spins are uncorrelated at distances larger
than R(t).
For two-time quantities, when t is sufficiently large one has
 
R(t)
C(t, t ) = Cst (t t ) fc R(t )
Multiplicative separation. (5.18)

Here Cst (tt ) R(tt )2(dD) = R(tt )2d . The scaling function fc (x)
describes the non-equilibrium behavior. It satisfies fc (1) = 1 and fc (x
) = 0, see the sketch in Fig. 24 (a). In the scaling forms the equilibrium
and non-equilibrium contributions enter in a multiplicative structure. Non-
equilibrium effects are taken into account by taking ratios between the sizes
of the correlated domains at the observation times t and t in the scaling
functions.

5.6 Sub-critical coarsening

5.6.1 Dynamic scaling hypothesis

The dynamic scaling hypothesis states that at late times and in the
scaling limit

r (g) , R(g, t) (g) , r/R(g, t) arbitrary , (5.19)

where r is the distance between two points in the sample, r |~x ~x |, and
(g) is the equilibrium correlation length that depends on all parameters

122
 100 stationary Cst
qea
aging
t Cag
C101
(a) (b)

102 1
10 101 103 105 101 101 103 105
t tw t tw

Figure 24: Sketch of the decay of the two-time correlation at Tc (a) and
T < Tc (b) for different values of the waiting-time, increasing from left to
right.

1 1
t=4 t=4
0.8 8 0.8 8
16 16
0.6 32 0.6 32
64 64
C(r,t)

C(r,t)

128 128
0.4 256 0.4 256
512 512
0.2 1024 0.2 1024
0 0

-0.2 -0.2
0 20 40 60 80 100 120 0 1 2 3 4 5 6
r 1/2
r/t

Figure 25: The equal-time correlation as a function of distance in the 2dIM

quenched below Tc . Raw (left) and scaled (right) data.

(T and possibly others) collected in g, there exists a single characteris-

tic length, R(g, t), such that the domain structure is, in statistical sense,
independent of time when lengths are scaled by R(g, t). Time, denoted by
t, is typically measured from the instant when the critical point is crossed.
In the following we ease the notation and write only the time-dependence
in R. This hypothesis has been proved analytically in very simple models
only, such as the one dimensional Ising chain with Glauber dynamics or the
Langevin dynamics of the d-dimensional O(N) model in the large N limit
(see Sect. 5.9).
The late stage of phase-ordering in binary systems is characterized by a

123
patchwork of large domains the interior of which is basically thermalized in
one of the two equilibrium phases while their boundaries are slowly moving.
This picture suggests the splitting of the degrees of freedom (spins) into two
categories, providing statistically independent contributions to observables
such as correlation or response functions. More precisely, a quasi-equilibrium
stationary contribution arises as due to bulk spins, while boundaries account
for the non-equilibrium part. Then asymptotically one has

C(r, t) Cst (r) + Cag (r, t) Additive separation. (5.20)

The first term describes the equilibrium fluctuations in the low temperature
broken symmetry pure states
!
r
Cst (r) = (1 hsi i2eq ) g , (5.21)

where hsi ieq is the equilibrium expectation value of the local spin, and g(x)
is a function with the limiting values g(0) = 1, limx g(x) = 0. The second
term takes into account the motion of the domain walls through
!
r
Cag (r, t) = hsi i2eq f , (5.22)
R(t)

with f (1) = 1 and limx f (x) = 0. Both Cst and Cag obey (separately)
scaling forms with respect to the equilibrium and the non-equilibrium lengths
, R(t). In particular, eq. (5.22) expresses the fact that system configurations
at different times are statistically similar provided that lengths are measured
in units of R(t), namely the very essence of dynamical scaling.
Monte Carlo simulations of the Ising model and other systems quenched
below criticality and undergoing domain growth demonstrate that in the long
waiting-time limit t t0 , the spin self-correlation hsi (t)si (t )i separates into
two additive terms

C(t, t ) Cst (t t ) + Cag (t, t ) Additive separation (5.23)

see Fig. 26, with the first one describing equilibrium thermal fluctuations
within the domains,
(
1 hsi i2eq = 1 m2 , t t = 0 ,
Cst (t t ) = (5.24)
0, t t ,

124
and the second one describing the motion of the domain walls
! (
R(t) hsi i2eq , t t ,
Cag (t, t ) = hsi i2eq fc = (5.25)
R(t ) 0, t t .

To ease the notation we have not written the explicit T -dependence in R

that, as we shall see below, is less relevant than t. Note that by adding the
two contributions one recovers C(t, t) = 1 as expected and C(t, t ) 0 when
t t . The first term is identical to the one of a system in equilibrium in one
of the two ordered states, see eq. (5.12) for its asymptotic t t t limit;
the second one is inherent to the out of equilibrium situation and existence
and motion of domain walls. They vary in completely different two-time
scales. The first one changes when the second one is fixed to hsi i2eq , at times
such that R(t)/R(t ) 1. The second one varies when the first one decayed
to zero. The mere existence of the second term is the essence of the aging
phenomenon with older systems (longer t ) having a slower relaxation than
younger ones (shorter t ). The scaling of the second term as the ratio between
two lengths is a first manifestation of dynamic scaling.
A decorrelation time can also be defined in this case by expandind the
argument of the scaling function around t t. Indeed, calling t t t
one has R(t)/R(t ) R(t + t)/R(t ) [R(t ) + R (t )t]/R(t ) 1 +
t/[d ln R(t )/dt ]1 and one identifies a t -dependent decorrelation time

td [d ln R(t )/dt ]1 decorrelation time (5.26)

which is, in general, a growing function of t .

In order to fully characterise the correlation functions one then has to
determine the typical growing length, R, and the scaling functions, g, f ,
fc , etc. It turns out that the former can be determined with semi-analytic
arguments and the predictions are well verified numerically at least for
clean system. The latter, instead, are harder to obtain. We shall give a very
brief state of the art report in Sect. 5.7. For a much more detailed discussion
of these methods see the review articles in [26].
The time-dependent typical domain length, R(t), is determined numeri-
cally by using several indirect criteria or analytically within certain approxi-
mations. The most common ways of measuring R are with numerical simula-
tions of lattice models or the numerical integration of the continuous partial
differential equation for the evolution of the order parameter. In both cases
one

125
 C(t,tw) 1 1

C(t,tw)
tw=2 tw=2
4 4
8 8
0.1 16 0.1 16
32 32
64 64
128 128
256 256
512 512
1 10 100 1000 1 10 100 1000
t-tw t-tw

Figure 26: The two-time self-correlation in the 2dIM with non-conserved

order parameter dynamics at several waiting-times given in the key at tem-
perature T = 0.5 (left) and T = 2 (right). Data obtained with Monte Carlo
simulations. Note that the plateau is located at a lower level in the figure
on the right consistently with the fact that hieq decreases with increasing
temperature.

Computes the inverse perimeter density R(t) = hHieq /[hH(t)i

hHieq ] with hH(t)i the time-dependent averaged energy and hHieq the equi-
librium energy both measured at the working temperature T .
Puts the dynamic scaling hypothesis to the test and extracts R from
the analysis.

5.6.2 R(t) in clean one dimensional cases with non-conserved order

parameter
In one dimension, a space-time graph allows one to view coarsening as the
diffusion and annhilitation upon collision of point-like particles that represent
the domain walls. In the Glauber Ising chain with non-conserved dynamics
one finds that the typical domain length grows as t1/2 while in the continuous
case the growth is only logarithmic, ln t.

5.6.3 R(t) in non-conserved curvature driven dynamics (d > 2)

The time-dependent Ginzburg-Landau model allows us to gain some in-
sight on the mechanism driving the domain growth and the direct computa-
tion of the averaged domain length. In clean systems temperature does not
play a very important role in the domain-growth process, it just adds some

126
 t1 t2

y x
t1 g
t2
g
o< 0 o> 0 o< 0 o> 0

Figure 27: Left: domain wall profile. Right: view from the top. (g is n.)

thermal fluctuations within the domains, as long as it is smaller than Tc . In

dirty cases instead temperature triggers thermal activation.
We focus first on the clean cases at T = 0 and only later we discuss
thermal effects. Equation (5.5) for = 0 is just a gradient descent in the
energy landscape F . Two terms contribute to F : the bulk-energy term that
is minimized by = 0 and the elastic energy ()2 which is minimized by
flat walls if present. As a consequence the minimization process implies that
regions of constant field, (~x, t) = 0 , grow and they separated by flatter
and flatter walls.
Take a flat domain wall separating regions where the configuration is
the one of the two equilibrium states, (~x, t) = 0 + (~x, t). Linearizing
eq. (5.5) around 0 and looking for static configurations, i.e. (~x, t) =
(~x) = (n) where n is the distance from the wall along the normal direc-
2 2
d (n)/dn = V (0 )(n). This equation has the solution
tion one finds
(n) e V (0 )n where n is the perpendicular distance to the wall . The
order parameter approaches 0 on both sides of the wall very rapidly. This
means that the free-energy of a configuration with an interface (sum of the
elastic and potential terms) is concentrated in a very narrow region close to
it. In consequence, the domain-wall curvature is the driving force for domain
growth.

Allen and Cahn showed that the local wall velocity is proportional to the
local curvature working with the GL equation at = 0. The proof goes as
follows. Take the GL equation and trasform the derivatives to apply in the

127
direction normal to the wall:
  
(~x, t) (~x, t)  n  ~ x, t) = (~x, t)  n

=  , (~ ,
t n t t  n t

 
2 2 (~x, t)  (~x, t)  ~
(~x, t) = +  n

n2 t n t


where the subscripts mean that the derivatives are taken at t or fixed.
2 (~
Using now n x,t)
2 |t = V () (note that the derivative is taken at fixed t) in
the GL equation one finds the Allen-Cahn result
~ n
v t n| = (5.27)

valid in all d with the geodesic curvature.

Equation (5.27) allows one to get an intuition about the typical growth
law in such processes. Take a spherical wall in any dimension. The local
curvature is constant and = (d 1)/R where R is the radius of the pshere
within the hull. Equation (5.27) is recast as dR/dt = (d1)/R that implies
R2 (t) = R2 (0) 2(d 1)t.
A closer look at the 2d equation allows one to go beyond and prove, in
this case, that all areas enclosed by domain walls irrespective of their being
other structures within (the so-called hull-enclosed areas) tend to diminish at
constant rate dA/dt = . This, of course, does not mean that all domains
reduce their area since a domain can gain area from the disappearance of
an internal domain of the opposite sign, for instance. The proof is simple
and just uses the Gauss-Bonnet theorem: dA =
H
~
v d~ = H vd. The local
dt
wall-velocity, ~v , is proportionalHto the local geodesic curvature, , and the
Gauss-Bonnet theorem implies d = 2 for a planar 2d manifold with no
holes. Therefore, the hull-enclosed area decreases with constant velocity for
any geometry.
There are a number of ways to find the growth law

R(t) = t1/zd (5.28)

with zd the dynamic exponent, in pure and isotropic systems (see [26]).
The effects of temperature enter only in the parameter and, for clean
systems, growth is slowed down by an increasing temperature since thermal
fluctuation tend to roughen the interfaces thus opposing the curvature driven
mechanism. We estimate the T dependence of in Sect. 5.6.5.

128
 In curvature driven Ising or Potts cases with non-conserved order pa-
rameter the domains are sharp and zd = 2 with a weakly T -dependent
coefficient. For systems with continuous variables such as rotors or XY mod-
els and the same type of dynamics, a number of computer simulations have
shown that domain walls are thicker and zd = 4.

5.6.4 R(t) in conserved order parameter and the role of bulk

diffusion
A different type of dynamics occurs in the case of phase separation (the
water and oil mixture ignoring hydrodynamic interactions or a binary allow).
In this case, the material is locally conserved, i.e. water does not transform
into oil but they just separate. The main mechanism for the evolution is
diffusion of material through the bulk of the opposite phase. After some
discussion, it was established, as late as in the early 90s, that for scalar
systems with conserved order parameter zd = 3.

5.6.5 Crossover between critical and sub-critical coarsening: (T )

In the case of non-conserved scalar order-parameter dynamics behaves as

R(t) t1/zeq (5.29)

with zeq the equilibrium dynamics exponent (note that zeq is different from
zd ). We shall not discuss critical dynamics in detail; this problem is treated
analytically with dynamic renormalization group techniques and it is very
well discussed in the literature [27]. In short, the exponent zeq is given by [9]
N +2 4 1 2
 
zeq = 2 + 2
3 ln + O(3 ) (5.30)
(N + 8) 3 2
where N is the dimension of the possibly vector field, N = 1 for a scalar one,
and = 4 d with d the dimension of space. Note that zeq is larger than 2
for all finite N and it approaches 2 in the large N limit (at least up to this
order in perturbation theory). In particular, one finds


2.0538 d=2
zeq 2.0134 d=3 (5.31)
2 d=4

129
for N = 1. Numerical simulations indicate zeq 2.13 in d = 2.
Matching critical coarsening with sub-critical one allows one to find the
T -dependent prefactor [12]. The argument goes as follows. The out of
equilibrium growth at criticality and in the ordered phase are given by
(
t1/zeq at T = Tc ,
R(t) (5.32)
((T )t)1/zd at T < Tc .

zeq is the equilibrium dynamic critical exponent and zd the out of equilibrium
growth exponent. Close but below criticality one should have an interpolating
expression of the kind
!
a 1/zd t
R(t) t f at T = Tc (5.33)
zeq

with the T -dependent equilibrium correlation length, (T ) (Tc T ) .

The last factor tends to one, f (x ) 1, when R(t) , that is
to say when the argument diverges and the system enters the sub-critical
coarsening regime. It is however non-trivial when R(t) , the argument is
finite and critical coarsening must be described. In particular, we determine
its behaviour for x = O(1) by requiring that eq. (5.33) matches the subcritical
growing length which is achieved by (i) recovering the correct t dependence,
(ii) cancelling the factor. (i) implies

f (x) x1/zd +1/zeq for x = O(1) . (5.34)

Then eq. (5.33) becomes

R(t) a+zeq /zd 1 t1/zeq (5.35)

and to eliminate we need

a = zeq /zd 1 . (5.36)

Comparing now the subcritical growing length and (5.33) in the very long
times limit such that R(t) and f (x ) 1:

[(T )]1/zd a (Tc T )(zeq zd )/zd . (5.37)

130
5.6.6 The 2d xy model
The critical dynamics of an infinite system prepared in a non-equilibrium
initial condition occurs, necessarily, out of equilibrium. Indeed, since the
equilibrium correlation length, (Tc ), diverges, the longest relaxation time,
tr (Tc ), and hence the equilibration time also diverges. At criticality one
observes critical slowing down or, in other words, a slow dynamics associ-
ated to the decay of correlations as power laws of time. An infinite critical
system evolves towards equilibrium without ever reaching it through a
non-equilibrium scaling regime characterized by a single length scale (t)
(< (Tc ) ). This dynamic correlation length can be extracted from the
decay of the equal-time two-point correlation function that, according to the
conventional scaling theory of non-equilibrium critical dynamics [?], is given
by eq. (xxx) above.
The xy model in d = 2 is quite special since it is critical at all tem-
peratures below Tkt . It is then worth analyzing this special case in detail.
Moreover, it has topological defects and the rate of approach to the equilib-
rium state is affected by them.
The model is fully solvable in the spin-wave approximation in which the
field is supposed to vary smoothly in space and, hence, vortices are neglected.
The functional Langevin equation acting on the angle between the local spins
and a chosen axis is linear in Fourier space and it can be readily solved. The
angle correlation functions in equilibrium are
kB T
h((r) (0))2 i = ln r/a (5.38)
J
leading to
 kB T /J
a a (T )
 
C(r) = hs(r)s(0)h= = (5.39)
r r
The equilibrium correlation length is (T ) = a/ ln(kB T /J) that tends to
zero only at T and diverges at T 0.
Spin-waves are non-local and extensive while vortices are local and in-
tensive. The latter cannot be eliminated by simple perturbations but they
annihilate.
C(t, t ) = V 1 d2 x h s(~x, t)
R
The global correlation and linear response, 
x,t) i 
s(~x, t ) i and R(t, t ) = V 1 d2 x hh(~
s(~
R
x,t )
 take the following scaling forms
h=0

131
in the limit t t 2 :
!
1 R(t)
C(t, t )
(5.40)
(t t ) (T )/2 R(t )
!
1 R(t)
R(t, t ) (5.41)
4(T )(t t )1+(T )/2 R(t )

with a scaling function and R(t) the growing correlation length (that should
not be confused with the linear response). The first remarkable property of
these functions is that they are both decomposed in the product of a function
of the time-difference t t and a function of the ratio R(t )/R(t), like
in the general critical coarsening case. When t t R(t ) and 1, the
decay is stationary

C(t, t ) (t t )(T )/2 , R(t, t ) (t t )1(T )/2

and the fdr equals one. This limit defines a quasi-equilibrium regime. When
the time difference increases and becomes smaller than one the relaxation
enters an aging regime in which the decay of the correlation and response
depends on the waiting-time t . The behaviour in the aging regime depends
on the initial conditions as discussed below.
Uniform initial conditions.
The uniform initial condition contains no free vortices and none are gener-
ated by thermal fluctuations at any T < Tkt . The evolution is well captured
by the simple spin-wave approximation and after a simple calculation one
finds
! " #(T )/4
(t) (1 + )
= , R(t) = t . (5.42)
(t ) 4

Beyond the crossover time tt t , when C(2t , t ) t(T )/2 and becomes
smaller than one, the correlation and response decay to zero as power laws
of the waiting-time t . There is no clear-cut separation of time-scales char-
acterised by the correlation reaching a constant value independently of the
waiting-times but only a t dependent pseudo-plateau where the behaviour of
the two-time correlation changes. This is to be confronted to the behaviour
of ferromagnetic coarsening systems quenched to the low-temperature phase
for which the crossover occurs at C(2t , t ) = m2eq . Above this plateau, the

132
relaxation corresponds to the equilibrium fluctuations of short wave-length
while below the plateau the decorrelation is due to the domain-wall motion
that manifests into a scaling in terms of = t /t only. In the 2d xy case the
order parameter vanishes and there is no plateau at any finite value of C.
In the aging regime the fluctuation dissipation ratio is larger than one.
This a priori surprising result can be understood when interpreted in terms
of the effective temperature. The completely order configuration is the
equilibrium state at zero temperature. The evolution of this initial state at
finite temperature can be thought of as representing a sudden inverse quench
of the system from T = 0 to T > 0. If the fdr is related to a remembrance
of the temperature of the initial condition, in this case this is lower than the
working temperature T and thus, the effective temperature also turns out to
be lower than T .
Random initial conditions.
When random initial conditions with only short-ranged spatial correla-
tions are considered, free vortices and antivortices are present initially. The
relaxation occurs vis the annihilation of vortex-antivortex pairs and this
coarsening process is much slower than the relaxation of spin-waves. The
simple Gaussian theory is no more suited to describe this dynamics and a
full analytic treatment is too hard to implement. With scaling and numeric
analysis the dynamic correlation length has been estimated to be [26]

R(t) (t/ ln t)1/2 .

The numerical simulations of Berthier, Holdsworth and Sellitto have proven

that the two-time correlation and response are correctly described by the
scaling form (5.40) and (5.41) with this length scale and the full decay looks
like the one shown in the sketch above. The fdr is rather different from
the one following the evolution of a uniform initial condition. The non-
equilibrium susceptibility is now smaller than the equilibrium value, and in
terms of the effective temperature [10] this means that the fluctuations of the
wave-lengths longer than R(t) occur at a Tef f > T and hence keep a memory
of the initial temperature T = .

5.6.7 Role of disorder: thermal activation

The situation becomes much less clear when there is quenched disorder
in the form of non-magnetic impurities in a magnetic sample, lattice dislo-

133
cations, residual stress, etc. Qualitatively, the dynamics are expected to be
slower than in the pure cases since disorder pins the interfaces. In general,
based on an argument due to Larkin (and in different form to Imry-Ma) one
expects that in d < 4 the late epochs and large scale evolution is no longer
curvature driven but controlled by disorder. Indeed, within a phase space
view disorder generates metastable states that trap the system and thus slow
down the relaxation.
A hand-waving argument to estimate the growth law in dirty systems is
the following. Take a system in one equilibrium state with a domain of linear
size R of the opposite equilibrium state within it. This configuration could
be the one of an excited state with respect to the fully ordered one with
absolute minimum free-energy. Call F (R) the free-energy barrier between
the excited and equilibrium states. The thermal activation argument (see
Sect. 3.7.3) yields the activation time scale for the decay of the excited state
(i.e. erasing the domain wall)

tA eF (R)/(kB T ) . (5.43)

For a barrier growing as a power of R, F (R) (T, J)R (where J rep-

resents the disorder) one inverts (5.43) to find the linear size of the domains
still existing at time t, that is to say, the growth law
 1/
kB T
R(t) (T )
ln t . (5.44)

All smaller fluctuation would have disappeared at t while typically one would
find objects of this size. The exponent is expected to depend on the
dimensionality of space but not on temperature. In normal systems it is
expected to be just d 1 the surface of the domain but in spin-glass
problems, it might be smaller than d1 due to the presumed fractal nature of
the walls. The prefactor is expected to be weakly temperature dependent.
One assumes that the same argument applies out of equilibrium to the
reconformations of a portion of any domain wall or interface where R is the
observation scale.
However, already for the (relatively easy) random ferromagnet there is no
consensus about the actual growth law. In these problems there is a compe-
tition between the pure part of the Hamiltonian, that tries to minimize the
total (d 1) dimensional area of the domain wall, and the impurity part
that makes the wall deviate from flatness and pass through the locations of

134
lowest local energy (think of Jij = J + Jij with J and Jij contributing to
the pure and impurity parts of the Hamiltonian, respectively). The activa-
tion argument in eq. (5.43) together with the power-law growth of barriers
in F (R) (T, J)R imply a logarithmic growth of R(t). Simulations, in-
stead, suggest a power law with a temperature dependent exponent. Whether
the latter is a pre-asymptotic result and the trully asymptotic one is hidden
by the premature pinning of domain walls or it is a genuine behaviour inval-
idating F (R) (T, J)R or even eq. (5.43) is still an open problem. See
the discussion below for a plausible explanation of the numerical data that
does not invalidate the theoretical expectations.
In the 3d RFIM the curvature-driven growth mechanism that leads to
(5.28) is impeded by the random field roughening of the domain walls. The
dependence on the parameters T and h has been estimated. In the early
stages of growth, one expects the zero-field result to hold with a reduction
in the amplitude R(t) (A Bh2 ) t1/2 . The time-window over which this
law is observed numerically decreases with increasing field strength. In the
late time regime, where pinning is effective Villain deduced a logarithmic
growth R(t) (T /h2 ) ln t/t0 by estimating the maximum barrier height
encountered by the domain wall and using the Arrhenius law to derive the
associated time-scale.
In the case of spin-glasses, if the mean-field picture with a large number
of equilibrium states is realized in finite dimensional models, the dynamics
would be one in which all these states grow in competition. If, instead, the
phenomenological droplet model applies, there would be two types of domains
growing and R(t) (ln t)1/ with the exponent satisfying 0 d 1.
Some refined arguments that we shall not discuss here indicate that the
dimension of the bulk of these domains should be compact but their surface
should be rough with fractal dimension ds > d 1.

5.6.8 Temperature-dependent effective exponents

The fact that numerical simulations of dirty systems tend to indicate
that the growing length is a power law with a T -dependent exponent can be
explained as due to the effect of a T -dependent cross-over length LT . Indeed,
if below LT T the growth process is as in the clean limit while above LT
quenched disorder is felt and the dynamics is thermally activated:
(
t1/zd for R(t) LT ,
R(t) (5.45)
(ln t)1/ for R(t) LT .

135
These growth-laws can be first inverted to get the time needed to grow a
given length and then combined into a single expression that interpolates
between the two regimes:

t(R) e(R/LT ) Rzd (5.46)
where the relevant T -dependent length-scale LT has been introduced. 
1 R (t)
Now, by simply setting t(R) Rz(T ) one finds z(T ) zd + ln R(t) L
  T

1/z(T ) z(T ) t/z(T )

that replacing R t becomes z(T ) zd + ln t L
. Using now
T
z(T ) zd in the correction term and focusing on times such that t/zd / ln t is
almost constant and equal to c one finds z(T ) zd c zd /LT . Similarly, by
equating t(R) exp(R(T ) /T ) one finds that (T ) is a decreasing function
of T approaching at high T .

5.7 Scaling functions for subcritical coarsening

Even though the qualitative behaviour of the solution to eq. (5.5) is easy
to grasp, it is still too difficult to solve analytically and not much is known
exactly on the scaling functions. A number of approximations have been
developed but none of them is fully satisfactorily (see [26] for a critical review
of this problem).
The super-universality hypothesis states that in cases in which tem-
perature and quenched disorder are irrelevant in the sense that they do not
modify the nature of the low-temperature phase (i.e. it remains ferromag-
netic in the case of ferromagnetic Ising models) the scaling functions are not
modified. Only the growing length changes from the, say, curvature driven
t1/2 law to a logarithmic one. This hypothesis has been verified in a number
of two and three dimensional models including the RBIM and the RFIM.

5.7.1 Breakdown of dynamic scaling

Some special cases in which dynamic scaling does not apply have also been
exhibited. Their common feature is the existence of two (or more) growing
lengths associated to different ordering mechanisms. An example is given by
the Heisenberg model at T 0 in which the two mechanisms are related to
the vectorial ordering within domains separated by couples of parallel spins
that annhilate in a way that is similar to domain-wall annihilation in the
Ising chain.

136
5.8 Annealing
There has been recent interest in understanding how a finite rate cooling
affects the defect density found right after the quench. A scaling form in-
volving equilibrium critical exponents was proposed by Zurek following work
by Kibble. The interest is triggered by the similarity with the problem of
quantum quenches in atomic gases, for instance. An interplay between crit-
ical coarsening (the dynamics that occurs close in the critical region) that
is usually ignored (!) and sub-critical coarsening (once the critical region is
left) is the mechanism determining the density of defects right after the end
of the cooling procedure.
The growing length satisfies a scaling law

R(t, (t)) (t) f [tzeq (t)] (t) = |T (t) Tc |

(
ct
x 1 Equilibrium at high T
f (x)
x x1 Coarsening at low T

with t measured from the instant when the critical point is crossed and
x (1, 1) is the critical region. A careful analysis of this problem can be
found in [15].

5.9 An instructive case: the large N approximation

A very useful approximation is to upgrade the scalar field to a vectorial
one with N components
~ x, t) = (1 (~x, t), . . . , N (~x, t)) ,
(~x, t) (~ (5.47)

and modify the free-energy

1 ~ 2 N 2
Z  
F = dd x () + (0 N 1 2 )2 , (5.48)
2 4
PN
with 2 = =1 2 and 0 finite. The T = 0 dynamic equation then becomes

t (~x, t) = 2 (~x, t) 4 (~x, t) [20 N 1 2 (~x, t)] (5.49)

and it is clearly isotropic in the N dimensional space implying

C (~x, t; ~x , t ) = C(~x, t; ~x , t ) (5.50)

137
 In the limit N while keeping the dimension of real space fixed to d,
the cubic term in the right-hand-side can be replaced by

(~x, t)N 1 2 (~x, t) (~x, t)N 1 [ 2 (~x, t) ]ic (~x, t) a(t) (5.51)

since N 1 2 (~x, t) does not fluctuate, it is equal to its average over the initial
conditions and it is therefore not expected to depend on the spatial position
if the initial conditions are chosen from a distribution that is statistically
translational invariant. For the scalar field theory the replacement (5.51) is
just the Hartree approximation. The dynamic equation is now linear in
the field (~x, t) that we rename (~x, t) (and it is now order 1):

t (~x, t) = [2 + a(t)](~x, t) , (5.52)

where the time-dependent harmonic constant a(t) = 20 [2 (~x, t)]ic = 20

[2 (~0, t)]ic has to be determined self-consistently. Equation (5.52) can be
Fourier transformed

t (~k, t) = [k 2 + a(t)](~k, t) , (5.53)

and it takes now the form of almost independent oscillators under different
time-dependent harmonic potentials coupled only through the self-consistent
condition on a(t). The stability properties of the oscillators depend on the
sign of the prefactor in the rhs. The solution is
2
Rt
dt a(t )
(~k, t) = (~k, 0) ek t+ 0 (5.54)

and the equation on a(t) reads:

d/2
2
Rt 
dt a(t )
a(t) = 20 e2 0 , (5.55)
4t

where one used [2 (~x, t)]ic = [2 (~0, t)]ic and a delta-correlated Gaussian dis-
tribution of initial conditions with strength . The self-consistency equation
is not singular at t = 0 since there is an underlying cut-off in the integration
over k corresponding to the inverse of the lattice spacing, this implies that
times should be translated as t t + 1/2 with = 1/a the lattice spacing.
Without giving all the details of the calculation, eq. (5.55), generalized
to the finite T case, can be solved at all temperatures [16]. One finds that
there exists a finite Tc (d) and

138
Upper-critical quench
a(t) 2 (5.56)
with the equilibrium correlation length, and the mass (in field theoretical
terms) or the harmonic constant saturates to a finite value: k 2 + a(t)
k 2 2.
Critical quench

a(t) w/(2t) with w = 0 for d > 4 and w = (d 4)/2 for d < 4 .

(5.57)
The dynamics is trivial for d > 4 but there is critical coarsening in d < 4. zeq
equals 2 in agreement with the result from the expansion once evaluated
at N .
Sub-critical coarsening
In the long times limit in which the system tends to decrease its elastic
and potential energies [ 2 (~x, t) ]ic must converge to 20 6= 0 below criticality
Rt
and this imposes 2 0 dt a(t ) d2 ln(t/t0 ) with t0 = /2 (/20 )2/d at large
times, i.e.
d
a(t) for t t0 (5.58)
4t
and the time-dependent contribution to the spring constant vanishes asymp-
totically. Knowing the long-time behaviour of a(t) implies that each mode
[(~k, t)]ic with ~k 6= 0 vanishes asymptotically but the ~k = 0 mode grows as
td/4 . The growth of the ~k = 0 reflects the domain growth process whereby
all modulations tend to disappear and the configuration gets more and more
uniform as time passes.
We focus now on two interesting cases: quenches to Tc and T < Tc . The
asymptotic behaviour of the space-time correlation function in the aging
regime is
#d/4
4tt (~x ~x )2
" " #

[ (~x, t)(~x , t ) ]ic = 20 exp , (5.59)
(t + t )2 4(t + t )

for t t for a quench to T < Tc and

(~x ~x )2
" #

[ (~x, t)(~x , t ) ]ic = 20 t1d/2 f (t/t ) exp , (5.60)
4(t + t )

139
for a quench to Tc . We focus on d < 4. These expressions capture the main
features of the domain growth process:
1. In Fourier space all k 6= 0 modes have an exponential decay while the
k = 0 one is fully massless asympotically and diffuses.
2. In sub-critical quenches, for any finite and fixed (~x ~x ), in the long
times limit the exponential factor approaches one and one obtaines a
function of t /t only.
3. In critical quenches the two-time dependent prefactor is of the form
expected from dynamic scaling.
4. Due to the exponential factor, for fixed but very large time t and t
the correlation falls off to zero over a distance |~x ~x | t + t . This
means that, at time t, the typical size of the regions in the states 0 is
R(t) t1/2 . This holds for critical and sub-critical quenches as well and
it is a peculiar property of the large N O(N) model that has zeq = zd .
5. For fixed |~x ~x |, the correlation always falls to zero over a time sepa-
ration tt which is larger than t . This means that the time it takes to
the system to decorrelate from its configuration at time t is of the order
of t itself, td t . The age of the system is the characteristic time-scale
for the dynamical evolution: the older is the system, the slower is its
dynamics. After a time of the order of the age of the system any point
~x will be swept by different domain walls and the correlation will be
lost.
6. In a critical quench the correlation always decays to zero due to the
prefactor that goes as t(2d)/2 and vanishes in all d > 2. The aging
curves have an envelope that approaches zero as a power law.
7. In a sub-critical quench, for any finite and fixed (~x ~x ), in the long
t and t limit such that t /t 1 the time dependence disappears and
the correlation between two points converges to 20 . This means that,
typically, if one looks at a finite spatial region on a finite time-scale this
region will be in one of the two states 0 , i.e. within a domain.

Note that we have obtained the field and then computed correlations
from the time-dependent configuration. We have not needed to compute the
linear response. We shall see later that in other more complex glassy systems
one cannot follow this simple route and one needs to know how the linear
response behave. We refer to the reviews in [29] for detailed accounts on the
behaviour of the linear response in critical dynamics.

140
5.10 Nucleation and growth
In a first-order phase transition the equilibrium state of the system
changes abruptly. Right at the transition the free-energies of the two states
involved are identical and the transition is driven by lowering the free-energy
as the new phase forms, see Fig. 18. The original phase remains meta-stable
close to the transition. The nucleation of a sufficiently large bubble of the
trully stable phase into the metastable one needs to be thermally activated
to trigger the growth process [28]. The rate of the process can be very low
or very fast depending on the height of the free-energy barrier between the
two states and the ambient temperature.
Two types of nucleation are usually distinguished: homogeneous (occury-
ing at the bulk of the meta-stable phase) and heterogeneous (driven by im-
purities or at the surface). The more intuitive examples of the former, on
which we focus here, are the condensation of liquid droplets from vapour and
the crystallization of a solid from the melt.
The classical theory of nucleation applies to cases in which the iden-
tification of the nucleous is easy. It is based on a number of assumptions that
we now list. First, one associates a number of particles to the nucleous (al-
though in some interesting cases this is not possible and a different approach
is needed). Second, one assumes that there is no memory for the evolution of
the number densities of clusters of a given size in time (concretely, a Markov
master equation is used). Third, one assumes that clusters grow or shrink by
attachment or loss of a single particle, that is to say, coallescence and fission
of clusters are neglected. Thus, the length-scale over which the slow part of
the dynamics takes place is the one of the critical droplet size, the first one to
nucleate. Fourth, the transition rates satisfy detail balance and are indepen-
dent of the droplet form. They just depend on the free-energy of the droplet
with two terms: a contribution proportional to the droplet volume and the
chemical potential difference between the stable and the metastable states,
f , and a contribution proportional to the bubble surface that is equal to
the surface area times the surface tension, , that is assumed to be the one
of coexisting states in equilibrium - that is to say the energy of a flat domain
wall induced by twisted boundary conditions. Fift, the bubble is taken to be

141
spherical and thus dependent of a single parameter, the radius. Thus

F [R] = d1 Rd1 |f | d Rd (5.61)

for d > 1. d is the volume of the unit sphere in d dimensions. For small
radii the surface term dominates and it is preferable to make the droplet
disappear. In contrast, for large radii the bulk term dominates and the
growth of the bubble is favoured by a decreasing free-energy. Thus the free-
energy difference has a maximum at

(d1) d1
R = d d |f |
|f |1 (5.62)

and the system has to thermally surmount the barrier F F [R ]. The

Kramers escape theory, see Sect. 3.7.3, implies that the nucleation rate or
the average number of nucleations per unit of volume and time is suppressed
by the Arrhenius factor
d
(d1)d1 d1 d
rA = t1
A e
F
with F = dd d1 |f |d1
d
(5.63)

As expected, F increases with increasing and/or |f |1 and r 1 vanishes

for T 0 when thermal agitation is switched off. The implicit assumption
is that the time to create randomly the critical droplet is much longer than
the time involved in the subsequent growth. The relaxation of the entire
system is thus expected to be given by the inverse probability of escape from
the metastable well. The determination of the prefactor [that is ignored in
eq. (5.63)] is a hard task.

142
6 Classical dynamic generating functional and
symmetries
In this Section we discuss some static and dynamic aspects of classical
statistical systems in the canonical ensemble. In this chapter we introduce
the classical path integral formalism. The symmetry arguments follow closely
the discussion in [?].

6.1 Classical statics: the reduced partition function

In order to analyze the statistical static properties of the classical coupled
system, we study the partition function or Gibbs functional, Ztot that reads
X
Ztot [] = exp(Htot x) (6.1)
conf osc
conf syst

where the sum represents an integration over the phase space of the full
system, the particles and the oscillators, and is a source. Having chosen
a quadratic bath and a linear coupling, the integration over the oscillators
coordinates and momenta can be easily performed. This yields the reduced
Gibbs functional

Nb
1X c2a
x2 . (6.2)
X
Zred [] exp Hsyst + Hcounter + x 2
conf syst 2 a=1 ma a

The counterterm Hcounter is chosen to cancel the last term in the exponential
and it avoids the renormalization of the particles mass (the coefficient of the
quadratic term in the potential) due to the coupling to the environment
that could have even destabilize the potential taking negative values. An
alternative way of curing this problem would be to take a vanishingly small
coupling to the bath in such a way that the last term must vanish by itself
(say, all ca 0). However, this might be problematic when dealing with the
stochastic dynamics since a very weak coupling to the bath implies also a
very slow relaxation. It is then conventional to include the counterterm to
cancel the mass renormalization. One then finds
X
Zred [] exp [ (Hsyst + x)] = Zsyst[] . (6.3)
conf syst

143
The interaction with the reservoir does not modify the statistical proper-
ties of the particle since Zred Zsyst . We shall see in Sect. 3.7.3 that
this does not happen quantum mechanically. (For a non-linear coupling
2
Hint = c q V(x) the counterterm is Hcounter = 12 mc2 [V(x)]2 .)
P P

6.2 Classical dynamics: generating functional

In Sect. 3.7.3 we showed a proof of the (generally non-Markov) Langevin
equation based on the integration over a large ensemble of harmonic oscilla-
tors that act as a bath with which the system is set in contact.
Observables which are functions of the solution to the Langevin equation
can also be computed using a dynamic generating functional that reads [42]
RT RT RT
2k 1
Z
dt dt (t )1 (t t )(t )+ dt (t )x (t )
Zd [] D dP (t0 ) e BT t0 t0 t0

x (t) is the solution to the Langevin equation with initial condition x0 =

0 ) at the initial time t0 . The factor dP (t0 ) is the measure
x(t0 ), x 0 = x(t
of the initial condition, dP (t0 ) dx0 dx 0 Pi [x0 , x 0 ]. The Gaussian factor is
proportional to P [] the functional probability measure of the noise. The
measure is D N k=0 d(tk ) with k = 0, . . . , N , tk = t0 + k(T t t0 )/N
Q

and N while (T t0 ) remains finite. The inverse kernel 1 is defined

within the interval [t0 , T ]: tT0 dt (t t )1 (t t ) = (t t ).
R

A very useful expression for Zd [], usually called the Martin-Siggia-Rose

generating functional (actually derived by Janssen [?]), is obtained by intro-
ducing the identity
Z T
Eq[x(t)] m
x(t) + dt (t t )x(t
) + V [x(t)] = (t) (6.4)
t0

valid at each time t, with the delta function

 
Z  Eq[x(t)] 
1= Dx [Eq[x(t)] (t)] det , (6.5)

x(t ) 



with Dx N k=1 dx(tk ). The factor |det . . . | is the determinant of the op-
Q

erator (t t ){mt2 + V [x(t)]} + (t t )t and ensures that the integral

equals one.9 The delta function can be exponentiated with an auxiliary field
9
Its origin is in the change of variables. In the same way as in the one dimensional
integral, dx [g(x)] = dz 1/|g [g 1 (z)]| (z) = 1/|g [gR1 (0)]|, to get 1 as a result one
R R
includes the inverse of the Jacobian within the integral: dx [g(x)] |g (x)| = 1.

144
i x= N
x (using the Fourier representation of the -function). Di 1
k=1 di
x(tk ).
Q

The determinant can be exponentiated with time-dependent anticommunt-

ing variables opening the way to the use of super-symmetry [23], a subject
that we shall not touch in these notes. However, since it does not yield a rel-
evant contribution to the kind of dynamics we are interested in, we forget it
(one can show that the determinant is a constant for Langevin processes with
coloured noise and/or inertia and that the discretization of an over-damped
Langevin equation with white-noise can also be chosen to set it to one It
convention, see App. ??). Zd reads
Z
Zd [, ] DDxDi
x dP (t0 )
RT h RT i
dt i
x(t ) m
x(t )+ dt (t t )x(t
)+V [x(t )](t )
e t0 t0
RT T
R RT
2k 1 dt dt (t )1 (t t )(t )+ dt [(t )x(t )+
(t )i
x(t )]
e BT t0 t0 t0

where we have introduced a new source (t), coupled to the auxiliary field
x(t). The integration over the noise (t) is Gaussian and it can be readily
i
done; it yields
kB T T T
Z Z
+ dt dt i
x(t ) (t t ) i
x(t ) (6.6)
2 t0 t0

and, for a coloured bath, the environment generates a retarded inter-

action in the effective action. In the usual white noise case, this term
reduces to, kB T 0 tT0 dt [i
x(t )]2 , a local expression. In the end, the gener-
R

ating function and resulting Martin-Siggia-Rose-Jaenssen-deDominicis

(MSRJD) action reads
Z
Zd [, ] x dP (t0) eS[x,ix,,]
DxDi
Z  Z 

x, , ] =
S[x, i dt i
x(t ) m
x(t ) + dt (t t )x(t
) + V [x(t )]
kB T
Z Z
+ dt dt i
x(t )(t t )i
x(t ) + sources . (6.7)
2
All integrals runs over [t0 , T ]. Causality in the integral over t is ensured by
the fact that is proportional to .
The MSRJD action has two kinds of contributions: the ones that depend
on the characteristics of the bath (through ) and the ones that do not. The

145
latter also exist in a functional representation of Newton dynamics and we call
them S det (for deterministic) while the former contain all information abour
thermal fluctuations and dissipation and we call it S diss (for dissipation):

S[x, i ] = S diss [x, i

x, , i x; ] + S det [x, i
x, , i
] . (6.8)

If the distribution of the initial condition were to be included in the action

x0 ], 6t would be part of S det .
as an additional term, ln Pi [x0 , i
Interestingly enough, the dynamic generating functional at zero
sources is identical to one for any model:

Zd [ = 0, = 0] = 1 (6.9)

as can be concluded from its very first definition. In particular, it does not
depend on the coupling constants of the chosen model. This property will be
utilized in disordered systems to render the dynamic calculations relatively
easier than the static ones.

6.3 Generic correlation and response.

The mean value at time t of a generic observable A is
Z
hA(t)i = x dP (t0 ) A[x(t)] eS[x,ix] ,
DxDi (6.10)

where S[x, ix] is ashort-hand notation for S[x, i

x, = 0, = 0]. The self-
correlation and linear response of x are given by
 
1 2 Zd [, ]  2 Zd [, ] 
C(t, t ) = hx(t)x(t )i = =  (6.11)
Zd [, ] (t)(t ) = =0 (t)(t ) = =0


hx(t)i  S[x, i
x; h]
R(t, t ) =
= hx(t)
x(t )i
i|h=0 = hx(t)i
h(t ) h=0 h(t )


 
1 2 Zd [, ]  2 Zd [, ] 
= = (6.12)
Zd [, ] (t) (t ) ==0 (t) (t ) ==0
 

with h(t ) a small field applied at time t that modifies the potential energy
according to V V h(t )x(t ). The i x auxiliary function is sometimes
called the response field since it allows one to compute the linear response

146
by taking its correlations with x. Similarly, we define the two-time correlation
between two generic observables A and B,
Z

CAB (t, t ) xdP (t0 ) A[x(t)]B[x(t )] eS[x,ix] = hA[x(t)]B[x(t )]i(6.13)
DxDi

and the linear response of A at time t to an infinitesimal perturbation linearly

applied to B at time t < t,

hA[x(t)]ifB 
RAB (t, t ) , (6.14)
fB (t ) fB =0


with V (x) 7 V (x) fB B(x). The function B(x) depends only on x (or on
an even number of time derivatives, that is to say, it is even with respect to
t t). By plugging eq. (6.10) in this definition we get the classical Kubo
formula for generic observables:

S[x, i
x; fB ] 
RAB (t, t ) = hA[x(t)] i x(t )B [x(t )]i (6.15)
= hA[x(t)]i
fB (t ) 
fB =0

with B [x(t )] = x B[x(t )]. This relation is also causal and hence propor-
tional to (t t ); it is valid in and out of equilibrium. For B[x] = x it
reduces to the correlation between x and i x.
If the system has quenched random exchanges or any kind of dis-
order, one may be interested in calculating the averaged correlations and
responses over different realizations of disorder. Surprisingly, this average is
very easy to perform in a dynamic calculation [43]. The normalization factors
1/Zd[, ] in (6.11) and (6.12) have to be evaluated at zero external sources
in which they are trivially independent of the random interactions. Hence,
it is sufficient to average Zd [, ] over disorder and then take variations with
respect to the sources to derive the thermal and disorder averaged two-point
functions. This property contrasts with an equilibrium calculation where the
expectation values are given by [hAi] = [1/Z conf A exp(H)], with [] de-
P

noting the disorder average. In this case, the partition function Z depends
explicitly on the random exchanges and one has to introduce replicas [24]
to deal with the normalization factor and do the averaging.
Having assumed the initial equilibration of the environment ensures that
a normal system will eventually equilibrate with it. The interaction with
the bath allows the system to dissipate its energy and to relax until ther-
malization is reached. However, in some interesting cases, as the dyamics

147
across phase transitions and glassy models, the time needed to equilibrate
is a fast growing function of N, the number of dynamic degrees of freedom.
Thus, the evolution of the system in the thermodynamic limit occurs out of
equilibrium. In real systems, a similar situation occurs when the equilibra-
tion time crosses over the observation time and falls out of the experimental
time-window.
A final interesting remark on the relevance of quenched disorder is the
following. When a system with quenched disorder evolves out of equilibrium
at finite temperature, the correlation function and the response function do
not depend on the realization of disorder if the size of the system is large
enough (the realization of disorder has to be a typical one). These quantities
are self-averaging. This statement is easily checked in a simulation. When
times get sufficiently long as to start seeing the approach to equilibrium,
dependencies on the realizations of disorder appear.

6.4 Time-reversal
Since it will be used in the rest of this chapter, we introduce the time-
reversed variable x by x(t) x(t) for all t. The time-reversed observable
is defined as

Ar ([x], t) A([
x], t). (6.16)

It has the effect of changing the sign of all odd time-derivatives in the expres-
sion of local observables, e.g. if A[x(t)] = t x(t) then Ar [x(t)] = t x(t).
As an example for non-local observables, the time-reversed Langevin equa-
tion reads
Z T
x(t) Fr ([x], t)
Eqr ([x], t) = m du (u t)x(u)
. (6.17)
T

Notice the change of sign in front of the friction term that is no longer
dissipative in this new equation.

6.5 An equilibrium symmetry

If the initial time t0 is set to t0 = T and the system is in equilibrium
at this instant, PT is given by the Maxwell-Boltzmann measure. One can
then check that the zero-source action, S[x, ix], is fully invariant under the

148
field transformation Tc defined as
(
xu
7 xu ,
Tc
xu
i 7 i
xu + du xu .

We introduced the notation xt = x(t) so as to make the notation more

compact. Note that du xu = du xu . This transformation does not involve
the kernel and it includes a time-reversal. The invariance is achieved
independently by the deterministic (S det ) and dissipative (S diss ) terms in
the action. The former includes the contribution from the initial condition,
ln PT . Moreover, the path-integral measure is invariant since the Jacobian
associated to this transformation is also block triangular with ones on the
diagonal. The proof goes as follows.

6.5.1 Invariance of the measure

The Jacobian Jc of the transformation Tc is the determinant of a trian-

gular matrix:
" xu #
(x, x) 0  2
Jc det = det1
uv
xv
x
u x
u = det1
uv [u+v ] =1
(Tc x, Tc x) xv x
v

and it is thus identical to one.

6.5.2 Invariance of the integration domain

Before and after the transformation, the functional integration on the

field x is performed for values of xt on the real axis. However, the new
domain of integration for the field x is complex. For all times, xt is now
integrated over the complex line with a constant imaginary part it xt .
One can return to an integration over the real axis by closing the contour
at both infinities. Indeed the integrand, eS , goes to zero sufficiently fast at
xt for neglecting the vertical ends of the contour thanks to the term
1 0 (i
xt )2 (in the white noise limit or the correspondong ones in colored
noise cases) in the action. Furthermore the new field is also integrated with
the boundary conditions x(T ) = x(T ) = 0.

149
6.5.3 Invariance of the action functional

The deterministic contribution satisfies

Z h i
S det
[Tc x, Tc x] = ln Pi (xT , x T ) xu + u xu ] mu2 xu + V (xu )
[i
Z u Z
= ln Pi (xT , x T ) i xu + V (xu )] +
xu [m xu + V (xu )]
x u [m
Zu Zu
= ln Pi (xT , x T ) i xu + V (xu )] +
xu [m u ln Pi (xu , x u )
u u
= S det [x, x] ,
 
= 21 mx 2 + V (x)
where we used the initial equilibrium measure ln Pi (x, x)
ln Z. In the first line we just applied the transformation, in the second line
we made the substitution u 7 u, in the third line we wrote the last inte-
grand as a total derivative the integral of which cancels the first term and
recreates the initial measure at T .
Secondly, we show that the dissipative contribution is also invariant under
Tc . We have
Z Z
S diss
[Tc x, Tc x] = [i
xu + u xu ] 1 uv i
xv
Zu Z v

= xu x u ]
[i vu 1 i
xv
u v
= S diss [x, x] .
In the first line we just applied the transformation, in the second line we made
the substitution u 7 u and v 7 v and in the last step we exchanged u
and v.

6.5.4 Invariance of the Jacobian (Grassmann variables)

Finally we show that the Jacobian term in the action is invariant once it
is expressed in terms of a Gaussian integral over conjugate Grassmann fields
(c and c ) and provided that the transformation Tc is extended to act on
these as follows10
(
xu
7 xu , cu 7 cu ,
Tc (6.18)
xu
i 7 i
xu + u xu , cu 7 cu .
10
More generally, the transformation on c and c is cu 7 cu and cu 7 1 cu
with C .

150
We start from
Z Z h i Z
J
S [c, c , x] = cu mu2 uv + u uv cv + cu V (xu )cu (6.19)
u v u

and we have

S J (Tc c, Tc c , Tc x)
Z Z h i Z
= cu mu2 uv + u uv cv + cu [V (xu )] cu
Z Zu v h i Z u

= cv mu2 vu u vu cu + cu V (xu )cu

u v u
= S J (c, c , x) .

In the first line we just applied the transformation, in the second line we
exchanged the anti-commuting Grassmann variables and made the substitu-
tions u 7 u and v 7 v, finally in the last step we used v vu = v uv
and we exchanged u and v. Finally the set of boundary conditions [ c(T ) =
c(T
) = c (T ) = c (T )] is left invariant.

6.6 Consequences of the transformation

We now use the transformation Tc to derive a number of exact results.

6.6.1 The fluctuation-dissipation theorem

This symmetry implies

hxt i xt iS[Tc x,Tc ix]

xt iS[x,ix] = hTc xt Tc i
= hxt ixt iS[x,ix] + hxt dt xt iS[x,ix] (6.20)

that, using time-translational invariance and t t , becomes

R( ) R( ) = d C( ) = d C( ) . (6.21)

For generic observables one can similarly apply the Tc transformation to

expression (6.15) of the linear response

RAB ( ) RAr Br ( ) = d CAB ( ) = d CAB ( ) . (6.22)

151
where we defined Ar and Br as

Ar ([x], t) A([
x], t) . (6.23)
R R
Take for instance a function A[x(t), t] = duf (x(u))(ut)+ R duf (x(u))(u

x(u))(ut)+. . . then Ar [x(t), t] = A[x(t), t] = duf (x(u))(u+
R
t)+ duf (
R R
t) + duf (x(u))(u
+ t) + duf (x(u))(u + t) + . . ..
Relations between higher order correlation functions evaluated at different
times t1 , t2 , . . . tn are easy to obtain within this formalism.

6.6.2 Fluctuation theorems

Let us assume that the system is initially prepared in thermal equilibrium

with respect to the potential V (x, T )11 . The expression for the determin-
istic part of the MSRJD action functional is

S det [x, x; , f] = H([xT ], T ) ln Z(T )

Z
i xu + V (xu , u ) fu [x]] ,
xu [m
u

where H([xt ], t ) 12 mx 2t +V (xt , t ) and f is a non-conservative force applied

on the system. The external work done on the system along a given trajectory
between times T and T is given by
Z u T Z u T Z u T Z
W [x; , f] dE = du V = u u V + x u x V
u u T u T u
Z uT Z u T
T u
= uTT u V (xu , u ) + x u fu [x] (6.24)
u T u T

where we take into account the time variation of the parameter .

Fluctuation Theorem 1.
The transformation Tc does not leave S det invariant but yields
Tc fr ] F W [x; ,
fr ]
S det [x, x; , f] 7 S det [x, x; , (6.25)
11
This is in fact a restriction on the initial velocities, x T , that are to be taken from the
Maxwell distribution with temperature 1 , independently of the positions xT . These
latter can be chosen from a generic distribution since the initial potential can be tailored
at will through the dependence of V .

152
 fr ] is the MSRJD action of the system that is prepared (in
where S det [x, x; ,
equilibrium) and evolves under the time-reversed protocol ((u) (u))
and external forces (fr ([x], u) f([ x], u)). F is the change in free en-
ergy: F = ln Z((T )) ln Z((T )) between the initial and the final
virtual equilibrium states. W is defined above. The dissipative part of the
action, S diss does not involve and it is still invariant under Tc . This means
that, contrary to the external forces, the interaction with the bath is not
time-reversed: the friction is still dissipative after the transformation. This
immediately yields

hA[x, x]iSc [x,x;,f] = eF hA[Tc x, Tc x]eW [x;,fr ] iSc [x,x;,f
r] (6.26)

for any functional A of x and x. In particular for a local functional of the

field, A[x(t)], it leads to the Crooks relation

hA[x(t)]iSc [x,x;,f] = eF hAr [x(t)]eW [x;,fr ] iSc [x,x;,f
r] , (6.27)

or also

hA[x(t)]B[x(t )]iSc [x,x;,f]

= eF hAr [x(t)]Br [x(t )]eW [x;,fr ] iSc [x,x;,f
r] . (6.28)

Setting A[x, x] = 1, we obtain the Jarzynski equality

1 = eF heW [x;,f]iSc [x,x;,f] . (6.29)

Setting A[x, x] = (W W [x; , f]) we obtain the transient fluctuation the-

orem

P (W ) = Pr (W ) e(W F ) , (6.30)

where P (W ) is the probability for the external work done between T and
T to be W given the protocol (t) and the non-conservative force f([x], t).
and time-
Pr (W ) is the same probability, given the time-reversed protocol
reversed force fr .
Fluctuation Theorem 2.
The result we prove in the following lines is not restricted to Langevin
processes with an equilibrium dissipative bath. It applies to generic clas-
sical equations of motion, with or without stochastic noise. In short, the

153
proof consists in applying time-reversal on the system and yields an equal-
ity between observables and their time-reversed counterparts in a so-called
backward (B) process in which the system is prepared in equilibrium with
respect to the final conditions of the forward process and is evolved according
to the time-reversed equations of motions and protocol. Let us rewrite the
action as
Z
Sc [x, x, ] = H(xT , x T , T ) i
xu Eq([xu ], u )
u
1
Z Z
+ xu 1 uv i
i xv ln Z(T ) ,
2 u v

and apply the following time-reversal of the fields

(
xu
7 xu ,
Ttr (6.31)
xu
i 7 i
xu .
This yields
Z
Sc [x, x, ]
7 T )
H([xT ], i u)
xu Eqr ([xu ],
u
1Z Z
+ xu 1 uv i
i xv ln Z(T )
2 u v
or, by introducing zeroes:
Z
T )
Wr H([xT ], i u)
xu Eqr ([xu ],
u
1
Z Z
+ xu 1 uv i
i T ) ,
xv F ln Z( (6.32)
2 u v

where F F (T ) F (T ) is the free-energy difference between the two

virtual equilibrium states corresponding to T and T . Wr H([xT ], T )
H([xT ], T ) is the work applied on the system that evolves with the time-
reversed equation of motion Eqr and under the time-reversed protocol . In
particular and contrary to the previous paragraph, the friction is no longer
dissipative after the transformation. This defines the backward (B) process.
Finally, for any observable A[x, x] we get the relation
hA[x, x]iF = eF hA[
x, x]eWr iB . (6.33)
In particular, for two-time correlations, it reads
hA[x(t)]B[x(t )]iF = eF hAr [x(t)]Br [x(t )]eWr iB . (6.34)

154
Setting A[x, x] = (W W [x; , f ]) we obtain the transient fluctuation the-
orem

PF (W ) = PB (W ) e(W F ) , (6.35)

where PF (W ) is the probability for the external work done between T and
T to be W in the forward process. PB (W ) is the same probability in the
backward process.

6.7 Equations on correlations and linear responses

Take any Langevin process in the MSRJD path-integral formalism. From
the following four identities
* + * + * + * +
i
x(t) x(t) x(t) i
x(t)
= = (t t ) , = = 0 , (6.36)
x(t )
i x(t ) x(t )
i x(t )

where the angular brackets indicate an average with the MSRJD weight, after
an integration by parts, one derives four equations
* + * +
S S
x(t) = (t t ) , i
x(t) = (t t ) ,(6.37)
x(t ) x(t )
i
* + * +
S S
x(t) =0, i
x(t) =0. (6.38)
x(t )
i x(t )

The second and third one read

  Z 

i
x(t) m
x(t ) + dt (t t ) x(t
) + V [x(t )]
Z
+kB T dt (t t ) hi x(t )i = (t t ) ,
x(t)i
  Z 

x(t) m
x(t ) + dt (t t ) x(t
) + V [x(t )]
Z
+kB T dt (t t ) hx(t)i
x(t )i = 0 , (6.39)

while the other ones, once causality is used (basically hx(t )i x(t)i = 0 for
t > t and hi x(t )i = 0) do not yield further information. All terms are
x(t)i
easily identified with the four types of two-time correlations apart from the
ones that involve the potential and are not necessarily quadratic in the fields.

155
The linear terms in two-time functions can be put together after identifying
the free-operator

d2
G1
0 (t , t ) = (t t )m

2 + (t t ) (6.40)
dt t
The non-linear terms can be approximated in a number of ways: perturbation
theory in a small parameter, Gaussian approximation of the MSRJD action,
self-consistent approximations, etc. The choice should be dictated by some
knowledge on the systems behaviour one wishes to reproduce. In short then
Z Z
0= dt G1
0 (t , t )C(t , t) + hx(t)V [x(t )]i + kB T dt (t t )R(t, t ) ,
Z

(t t ) = dt G1
x(t)V [x(t )]i .
0 (t , t )R(t , t) + hi (6.41)

156
6.8 Classical statics: the reduced partition function
In order to analyze the statistical static properties of the classical coupled
system, we study the partition function or Gibbs functional, Ztot that reads
X
Ztot [] = exp(Htot x) (6.42)
conf osc
conf syst

where the sum represents an integration over the phase space of the full
system, the particles and the oscillators, and is a source. Having chosen
a quadratic bath and a linear coupling, the integration over the oscillators
coordinates and momenta can be easily performed. This yields the reduced
Gibbs functional

Nb
1X c2a
x2 (6.43)
X
Zred [] exp Hsyst + Hcounter + x 2
.
conf syst
2 a=1 ma a

The counterterm Hcounter is chosen to cancel the last term in the exponential
and it avoids the renormalization of the particles mass (the coefficient of the
quadratic term in the potential) due to the coupling to the environment
that could have even destabilize the potential taking negative values. An
alternative way of curing this problem would be to take a vanishingly small
coupling to the bath in such a way that the last term must vanish by itself
(say, all ca 0). However, this might be problematic when dealing with the
stochastic dynamics since a very weak coupling to the bath implies also a
very slow relaxation. It is then conventional to include the counterterm to
cancel the mass renormalization. One then finds
X
Zred [] exp [ (Hsyst + x)] = Zsyst[] . (6.44)
conf syst

The interaction with the reservoir does not modify the statistical proper-
ties of the particle since Zred Zsyst . We shall see in Sect. 3.7.3 that
this does not happen quantum mechanically. (For a non-linear coupling
2
Hint = c q V(x) the counterterm is Hcounter = 21 mc2 [V(x)]2 .)
P P

157
7 Quantum formalism
We now consider a quantum system

[ ] = ih .
p, x (7.45)

The density operator evolves according to the equation

(t) i
= [H(t), (t)] (7.46)
t h

that is solved by

(t) = U (t, t0 ) t0 )] = U (t, t0 )

(t0 )[U(t, 0 , t)
(t0 )U(t (7.47)

with Rt
i )
dt H(t
t0 ) = T e h
U(t, t0
t > t0 . (7.48)
Averaged observables read
i i
n o n o
Tr A (t) Tr A T e h Ht 0 [T e h Ht ]

hA(t)i = = i i
n o
Tr
(t) Tr T e h Ht 0 [T e h Ht ]
i i
n o
Tr T [e h Ht ] A T e h Ht 0
= . (7.49)
Tr
0

(To ease the notation we do not write the possible time-dependence in the
Hamiltonian and the integral over t in the exponential.) We then use the
Heisenberg operators
i i
AH (t) [T e h Ht ] A T e h Ht (7.50)

The under-script H recalls that these are operators in the Heisenberg repre-
sentation but we shall drop it in the following.

7.1 Equilibrium

158
7.1.1 Feynman path integral
Feynmans path-integral represents the (forward) propagator K(x, t; x , t ) =

hx|eiH(tt )/h |x i as a sum over all paths leading from, in the Lagrangain rep-
resentation, x = x(t ) to x = x(t) with t > t weighted with the exponential
of the action (times i/h):
Z x(t)=x i
K(x, t; x , t ) = Dx e h S[x,x] (7.51)
x(t )=x
m 2
Z t  

S[x, x]
= dt x (t ) V [x(t )] . (7.52)
t 2
The formalism is well-suited to compute any type of time-ordered correlation
since the evolution is done towards the future. A concise review article on the
construction of path-integrals with some applications is [20]; several books
on the subject are [21].

7.1.2 The Matsubara imaginary-time formalism

The statistical properties of a canonical quantum system and, in par-
ticular, its equilibrium partition function, can also be obtained with the
path-integral method,
Z
Z = TreH = TreiH(t=ih)/h = dx K(x, ih; x, 0) . (7.53)

Having used an imaginary-time t = ih corresponds to a Wick-rotation.

The propagator K is now represented as a path-integral with periodic bound-
ary conditions to ensure the fact that one sums over the same initial and
finite state. Introducing = it, : 0 h (since t : 0 ih), defining
functions of , x( ) = x(t = i ) and dropping the tilde one has
Z
1 E [x,x]
K(x, ih; x, 0) = Dx e h S
(7.54)
x(0)=x(
h)

with the Euclidean action

h m 2
Z  
E
S [x, x]
= d x ( ) + V [x( )] . (7.55)
0 2

7.1.3 The equilibrium reduced density matrix

159
 We now follow the same route as in the derivation of the classical Langevin
equation by coupling the system to an ensemble of quantum harmonic oscil-
lators,
a , qb ] = ihab .
[ (7.56)
The equilibrium density matrix reads
1
tot (x , qa ; x , q a ) = hx , qa | e Htot |x , qa i , (7.57)
Ztot

with the partition function given by Ztot = Tre Htot and the trace taken
over all the states of the full system. As usual the density matrix can be
represented by a functional integral in imaginary time,
1
Z h)=x
x( Z h)=qa
qa ( 1 e
tot (x , qa ; x , q a ; ) = Dx Dqa e h Stot [] . (7.58)
Ztot x(0)=x qa (0)=qa

The Euclidean action S E has contributions from the system, the reservoir,
E E E E E
the interaction and the counterterm: Stot [] = Ssyst + Senv + Sint + Scounter +
R h
0 d ( )x( ) and we have included a source coupled to the particles
coordinate in order to use it to compute expectation values of the kind defined
in (7.49). The environment action is
N Z
m 2
( )
E

h m
[q ( )]2 + [q ( )]2
X
Senv = d , (7.59)
=1 0 2 2
that is to say, we choose an ensemble of independent oscillators. For simplic-
ity we take a linear coupling
Z h
N
E
X
Sint = d x( ) ca qa ( ) (7.60)
0 a=1

but others, as used in the derivation of the Langevin equation, are also possi-
ble. As in the classical case, the path integral over the oscillators coordinates
and positions is quadratic. The calculation of expectation values involves a
trace over states of the full system. For operators that depend only on
the particle, as A( x) above, the trace over the oscillators can be done ex-
plicitly. Hence, if one constructs the reduced equilibrium density operator
red = Trenv tot that acts on the systems Hilbert space, the expectation
value of the observables of the system is given by
Trsyst A(
x)red
hA(x)i = . (7.61)
Tr
tot

160
In the path-integral formalism this amounts to performing the functional
integral over periodic functions q ( ). From the point of view of the os-
cillators the systems coordinate is an external -dependent force. Using a
Fourier representation, q ( ) = n in
n= q e with n = 2n the Matsub-
P

h
ara frequencies, the integration over the q ( ) can be readily done [22]. A
long series of steps, very carefully explained in [19] allow one to obtain the
reduced density matrix:

red (x , x ; ) = Trenv tot (x , qa ; x , q a ; )

Z x(h)=x
1 1 E 1
R h R
= Dx e h Ssyst h 0 d 0 d x( )K( )x( ) (7.62)
Zred x(0)=x
where Zred is the partitition function of the reduced system, Zred = Ztot /Zenv
and Zenv the partition function of the isolated ensemble of oscillators. The
interaction with the reservoir generated a renormalization of the mass can-
celled by the counterterm but also a retarded interaction in the effective
action controled by the kernel
Z
2 X S() n2
K( ) = d ein , (7.63)
h n= 0 n2 + 2

with S() the spectral density of the bath, see eq. (4.20) for its definition.
The last retarded interaction in (7.62) remains. The imaginary time depen-
dence of K varies according to S(). Power laws in S lead to power-law
decays in K and thus to a long-range interaction in the imaginary-time di-
rection.
The effect of the quantum bath is more cumbersome than in the classical
case and it can lead to rather spectacular effects. A well-known example is
the localization transition, as function of the coupling strength to an Ohmic
bath in the behaviour of a quantum particle in a double well potential [39].
This means that quantum tunneling from the well in which the particle is ini-
tially located to the other one is completely suppressed by sufficiently strong
couplings to the bath. In the context of interacting macroscopic systems, e.g.
ferromagnets or spin-glasses, the locus of the transition between the ordered
and the disordered phase depends strongly on the coupling to the bath and
on the type of bath considered [?], as discussed in Section ??.

7.2 Quantum dynamics

161
 We shall see that the distinction between the effect of a reservoir on the
statistical properties of a classical and quantum system is absent from a fully
dynamic treatment. In both classical and quantum problems, the coupling
to an environment leads to a retarded interaction. In classical problems one
generally argues that the retarded interaction can be simply replaced by a
local one due to the very short correlation times involved in cases of interest,
i.e one uses white baths, but in the quantum problems one cannot do the
same.

7.2.1 Schwinger-Keldysh path integral

The Schwinger-Keldysh formalism [44, 45] allows one to analyse the real-
time dynamics of a quantum system. The starting point is the time depen-
dent density operator
i i
(t) = T e h Ht (0) T e h Ht . (7.64)

We have set the initial time to be to = 0. Introducing identities, an element

of the time-dependent density matrix reads

(x , x ; t) =
Z i
dXdX hx | T e h Ht |Xi hX| (0) |X i

i
hX | T e h Ht |x i . (7.65)

The first and third factors are the coordinate representation of the evolution

operators eiHt/h and eiHt/h , respectively, and they can be represented as
functional integrals:
i
Z x i +
h Ht
hx | T e |Xi = Dx+ e h S (7.66)
X
Z X
i i
hX | T e h Ht |x i = Dx e h S . (7.67)
x

Interestingly enough, the evolution operator in eq. (7.128) gives rise to a

path integral going backwards in time, from x (t) = x to x (0) = X . The
full time-integration can then be interpreted as being closed, going forwards
from t0 = 0 to t and then backwards from t to t0 = 0. This motivates
the name closed time-path formalism. A doubling of degrees of freedom
(x+ , x ) appeared and it is intimately linked to the introduction of Lagrange

162
multipliers in the functional representation of the Langevin dynamics in the
classical limit, see Sect. 6.2. The action of the system has two terms as in
(7.52) one evaluated in x+ and the other in x .
t m  2
Z  

Ssyst [ + ] = dt x (t ) V (x (t )) + (t )x (t ) (7.68)
0 2
where we have introduced two time-dependent sources (t), that appear in
the exponential with the sign indeicated.

7.2.2 Green functions

We define the position and fermionic Green functions

GB
ab (t, t ) ihTC xa (t)xb (t )i (7.69)
GFab (t, t ) ihTC a (t)b (t )i (7.70)

where a, b = and TC is time ordering on the close contour (see App. 8).
The definitions (7.69) and (7.70) and the fact that the insertion eval-
uated at the later time can be put on the upper (+) or lower () branch
indistinctively *** GIVE EXAMPLE WITH DRAWING *** yield the fol-
lowing relations between different Green functions

G++ (t, t ) = G+ (t, t )(t t ) + G+ (t, t )(t t) , (7.71)

G (t, t ) = G+ (t, t )(t t ) + G+ (t, t )(t t) , (7.72)

that hold for the bosonic and fermionic cases as well. We thus erase the
superscripts B, F that become superflous. Adding the last two identities one
finds

G++ + G G+ G+ = 0 for all t and t . (7.73)

In both cases one defines MSRDJ-like fields. For bosons these are

2 x(t) = x+ (t) + x (t) , 2h x(t) = x+ (t) x (t) , (7.74)

while for fermions they are

+ (t) + (t) + (t) (t)
2 (t) 2h (t)
2 (t) + (t) + (t) 2h (t) + (t) (t) .

163
One then constructs the Green functions
1 B 
GBxx (t, t ) ihTC x(t)x(t )i = G ++ + G B
+ G B
+ + G B
+
2
B
2i GK (t, t )
1  B 
GBxx (t, t ) ihTC x(t) x (t )i = G ++ G B
+ G B
+ G B
+
2h

GBR (t, t )
1  B 
GBx (t, t ) ihTC x
x (t)x(t )i = G++ GB B
G+ + G+
B
2h

GBA (t, t )
1  B 
GBx
x (t, t ) ihTC x
(t) x (t )i = G ++ + G B
G B
+ G B
+
2h2
B
G4 (t, t ) = 0 , (7.75)
and the fermionic ones
1 F 
GF (t, t ) ihTC (t) (t )i = G++ + GF + GF+ + GF+
2
2iGFK (t, t ) ,
1  F
GF(t, t ) ihTC (t) (t )i =

G++ GF + GF+ GF+
2h
GFR (t, t ) ,
1  F 
G (t, t ) ihTC (t) (t )i = G++ GF GF+ + GF+
2h
F
GA (t, t ) ,
(t )i = 1 GF + GF GF GF
 
GF(t, t ) ihTC (t) ++ + +
2h2
F
G4 (t, t ) = 0 , (7.76)
From the definition of GB
K it is obvious that
B
GB
K (t, t ) = GK (t , t) Re . (7.77)
Using eq. (7.73) the rotated Green functions are rewritten as
i i
GK = (G++ + G ) = (G+ + G+ ) , (7.78)
2 2
1 1
GR = (G++ G+ ) = (G G+ ) , (7.79)
h
h

1 1
GA = (G++ G+ ) = (G G+ ) . (7.80)
h
h

164
Using eqs. (7.71)-(7.72) one also finds
1
GR (t, t ) = [G+ (t, t ) G+ (t, t )] (t t ) , (7.81)
h

1
GA (t, t ) = [G+ (t, t ) G+ (t, t )] (t t) . (7.82)
h

which show explicitly the retarded and advanced character of GR and GA ,
respectively. Moreover,
B
GB
R (t, t ) = GA (t t) Re , (7.83)
GFR (t, t ) = [GFA (t , t)] , (7.84)
GFK (t, t ) = [GFK (t , t)] . . (7.85)

Inverting the above relations one has

iG++ = GK ih(GR + GA )/2 , iG+ = GK + ih(GR GA )/2 ,

(7.86)
iG+ = GK ih(GR GA )/2 , iG = GK + ih(GR + GA )/2 .

Going back to an operational formalism

2i
GB
R (t, t ) (t t ) h[x(t), x(t )]i , (7.87)
h


GB
K (t, t ) h{x(t), x(t )}i , (7.88)

for bosons (see App. 8), where one recognizes the Kubo formul in the first
two lines, and
2i
GFR (t, t ) (t t ) h{(t), (t )}i , (7.89)
h


GFK (t, t ) h[(t), (t )]i , (7.90)

for fermions.

7.2.3 Generic correlations

A generic two time correlation that depends only on the system is given
by
B(t
hA(t) )i = Z 1 (0) TrA(t)
B(t )
red (0) . (7.91)
red

165
 B(t
Clearly hA(t) )i = )A(t)i
6 hB(t and one can define symmetrized and anti-
symmetrized correlations:
B(t
C{A,B} (t, t ) = hA(t) ) + B(t
)A(t)i/2
, (7.92)
B(t
C[A,B] (t, t ) = hA(t) ) B(t
)A(t)i/2
, (7.93)

respectively.
Within the Keldysh path-integral representation these correlations can
be written as

C{A,B} (t, t ) = hA[x+ ](t){B[x+ ](t ) + B[x ](t )}i/2 ,

C[A,B] (t, t ) = hA[x+ ](t){B[x+ ](t ) B[x ](t )}i/2 , (7.94)

7.2.4 Linear response and Kubo relation

The linear response is defined as the variation of the averaged observable
A at time t due to a change in the Hamiltonian operated at time t in such
a way that H H fB B.
In linear-response theory, it can be expressed in
terms of the averaged commutator:


hA(t)i  2i )]i .
RAB (t, t )

= (t t )h[A(t), B(t (7.95)
fB (t ) fB =0 h




In the case A = x and B = x this implies Rxx (t, t ) = GB (t, t ). The path-
R
integral representation is in terms of the Keldysh fields x+ , x :
n o
RAB (t, t ) = ihA[x+ ](t) B[x+ ](t ) B[x ](t ) i/h . (7.96)

7.2.5 Quantum FDT

Proofs and descriptions of the quantum fdt can be found in several
textbooks [32, 46]. We first present a standard derivation that applies to
bosonic and fermionic Green functions in the canonical and grand-canonical
ensembles. We next show it for generic correlations and linear responses in
the bosonic case. In so doing we recall its expression in the time-domain and
in a mixed time-Fourier notation that gave us insight as to how to extend it
to the case of glassy non-equilibrium dynamics [33].

166
Canonical ensemble
Let us consider the canonical ensemble and let us write

iG+ (t, t ) = hTC + (t) (t )i (7.97)

where + = x+ and = x (bosons) or + = + and = (fermions).

This is equal to
iG+ (t, t ) = (1) h (t )+ (t)i (7.98)
with = 1 for fermions and = 0 for bosons. Using the analytic properties
of Green functions, we have

iG+ (t + ih, t ) = (1) h (t )+ (t + ih)i . (7.99)

In the canonical ensemble, h i Tr 0 Tr eH and after expanding

+ (t + ih) = 0 + (t)1
0 (in the Heisenberg representation) we get

Tr [ (t )0 + (t)]
iG+ (t + ih, t ) = (1) . (7.100)
Tr0
Using the cyclic property of the trace

iG+ (t + ih, t ) = (1) h+ (t) (t )i (7.101)

and we recognize

G+ (t + ih, t ) = (1) G+ (t, t ) (7.102)

If the systen has reached equilibrium, time translational invariance implies

G+ (t t + ih) = (1) G+ (t t ) . (7.103)

After Fourier transforming with respect to t t we get the KMS relation

G+ ()eh = (1) G+ () . (7.104)

Using eqs. (7.86), we have on the one hand

[GR () GA ()] = G+ () G+ () .
h (7.105)

Inserting the KMS relation (7.104) we get

[GR () GA ()] = G+ ()[1 (1) eh ]

h (7.106)

167
On the other hand,
i i
GK () = [G+ () + G+ ()] = G+ ()[1 + (1) eh ] .(7.107)
2 2
Combining this relation with eq. (7.104) we obtain the quantum FDTs

1 + (1) eh
" #
ih
GK () = [GR () GA ()] (7.108)
2 1 (1) eh

Using GR ()GA () = R()R () = 2iImGR () for bosons and fermions:

ih F h
GFK () = [GR () GFA ()] tanh (7.109)
2 2

ImGFR () tanh h2 ,
GFK () = h (7.110)

ih B h
GB
K () = [GR () GB
A ()] cotanh (7.111)
2 2

h
GB ImGB
K () = h R () cotanh 2 . (7.112)

In the case of bosons the FDT can be easily extended to

B h
ImRAB
C{A,B} () = h () cotanh . (7.113)
2

Grand canonical ensemble

In the grand-canonical ensemble, the proof remains essentially the same.
is the number
The initial density operator reads 0 e H+N , where N
operator which commutes with H (the number of particules is conserved
in non-relativistic quantum mechanics). For concreteness, let us focus on
fermions The three first steps in the proof presented above imply

GF+ (t + ih, t ) = Tr (t )e H + (t)eN /tr
h i
0 (7.114)

168
 and N
Since H commute and since for any operator f (N),
one has the
(N
property that f ) = f (N we have
+ 1),

+ (t)eN = e(N +1) + (t) (7.115)

and so
h i
GF+ (t + ih, t ) = e Tr (t )0 + (t) /Tr0 . (7.116)

At this point, we can follow the same steps as in the proof for the canonical
ensemble, namely use the cyclic property of the trace, time-translational
invariance and a Fourier transform, to obtain the fermionic KMS relation in
the grand-canonical ensemble

GF+ ()eh = e GF+ () . (7.117)

and the grand-canonical fermionic quantum FDT relation

!
ih
h
GFK () = [GFR () GFA ()] tanh (7.118)
2 2
or, equivalently
 
ImR() tanh h
GFK () = h 2
(7.119)

FDT for Generic observables in time-domain

If at time t the system is characterized by a density functional (t ), the
two-time correlation functions are given by eq. (7.91), (7.92) and (7.93). In
linear response theory RAB (t, t ) and the correlation C[A,B] (t, t ) are related
by the Kubo formula (7.95). If the system has reached equilibrium with a
heat-bath at temperature T at time t , the density functional syst (t ) is just
the Boltzmann factor exp(Hsyst )/Zsyst. It is then immediate to show that,
in equilibrium, time-translation invariance, CAB (t, t ) = CAB (t t ), and the
KMS properties

CAB (t, t ) = CBA (t , t + ih) = CBA (t ih, t ) (7.120)

hold. Using now these identities and assuming, for definiteness, that t > 0 it
is easy to verify the following equation
ih ih
C{A,B} (T ) + RAB (T ) = C{A,B} (T ) RAB (T ) , (7.121)
2 2
169
where T = t + ih/2. This is a way to express FDT through an analytic
continuation to complex times.
In terms of the Fourier transform defined in App. ?? the KMS relations

read CAB () = exp(h)CBA () and lead to 2C[A,B] () = 1 eh CAB (),
 
2C{A,B} () = 1 + eh CAB () and C[A,B] () = tanh(h/2) C{A,B}().
Back in the Kubo relation this implies
R
RAB (t t ) = hi (t t ) d
ei(tt ) tanh(h/2) C{A,B}() .(7.122)
R i
(recall C{A,B} () = C{A,B} ().) Using 0 dt exp(it) = lim0+ +i
=
() i P one has
1 d 1
Z
h
RAB () = lim+ tanh C{A,B} ( ) (7.123)
0 + i
h 2
from which we obtain the real and imaginary relations
1 h
ImRAB () = tanh C{A,B} ( ) ,
h
2
d h
Z !
1 1
ReRAB () = P
tanh C{A,B} ( )(7.124)
.
h
2
If h/2 1, tanh(h/2) h/2 and eq. (7.122) becomes the classical
FDT:
1 dCAB (t t )
RAB (t t ) = (t t ) . (7.125)
kB T dt

7.2.6 The influence functional

As in the derivation of the Langevin equation we model the environment
as an ensemble of many non-interaction variables that couple to the relevant
systems degrees of freedom in some convenient linear way. The choice of the
environment variables depends on the type of bath one intends to consider.
Phonons are typically modeled by independent quantum harmonic oscillators
that can be dealt with exactly. We then consider a system (
x, p) coupled to an
environment made of independent harmonic oscillators ( a ). An element
qa ,
of the total density function reads
(x , qa ; x , q a ; t) =
Z i
dXdX dQa dQa hx , qa | T e h Htot t |X, Qa i hX, Qa | tot (0) |X , Qa i

i
hX , Qa | T e h Htot t |x , qa i . (7.126)

170
The first and third factors are the coordinate representation of the evolution

operators eiHtot t/h and eiHtot t/h , respectively, and they can be represented as
functional integrals:

i
Z x Z qa i +
hx , qa| T e h Htot t |X, Qa i = Dx+ Dqa+ e h Stot (7.127)
X Qa
Z X Z Q
i a i
hX , Qa | T e h Htot t |x , qa i = Dx Dqa e h Stot . (7.128)
x qa

The action Stot has the usual four contributions, from the system, the
reservoir, the interaction and the counterterm.
As usual we are interested in the dynamics of the system under the effect
of the reservoir. Hence, we compute the reduced density matrix
Z
red (x , x ; t) = dqa hx , qa | tot (t) |x , qa i . (7.129)

7.2.7 Initial conditions

Factorization
The initial density operator tot (0) has the information about the initial
state of the whole system. If one assumes that the system and the bath are
set in contact at the initial time, the operator factorizes

(0) = syst (0)

env (0) . (7.130)

(Other initial preparations, where the factorization does not hold, can also
be considered and may be more realistic in certain cases [19] and see below.)
If the environment is initially in equilibrium at an inverse temperature ,
1 Henv
env (0) = Zenv e , (7.131)

the dependence on the bath variables is quadratic; they can be traced away
to yield the reduced density matrix:
Z Z Z x+ (t)=x Z x (t)=x
red (x , x ; t) = dX dX Dx+ Dx
x+ (0)=X x (0)=X
i
e h Sef f hX| syst (0) |X i (7.132)

171
 +
with the effective action Sef f = Ssyst Ssyst + Senv ef f . The last term has
been generated by the interaction with the environment and it reads [19]
i T [x+ (t ) x (t )]
T [x+ (t ) + x (t )]
Z Z
Senv = i dt dt 4(t t )
h
0 0 h
2
+ +
T [x (t ) x (t )] [x (t ) x (t )]
Z Z T
dt dt (t t )
h . (7.133)
0 0 h
h

The noise and dissipative kernels and are given by
Z 
1

(t) = d S() coth h cos(t) , (7.134)
0 2
d(t)
Z
(t) = (t) = (t) d S() sin(t) . (7.135)
dt 0

One can easily check that (t) = (t) and (t) is causal; moreover they
verify the bosonic FDT, as they should since the bath was assumed to be in
equilibrium. is like a response and is a correlation One can also write the
bath-generated action in the form
1Z Z
dtdt xa (t)ab (t, t )xb (t ) ,
X
SK,eff (7.136)
a,b= 2

with

++ = 2 i + = 2 + i , (7.137)
+ = 2 + i = 2 i (7.138)

and ++ + + + + + = 0. Although these relation resemble the

ones satisfied by the Gab s a more careful analysis shows that the matricial
is more like G1 than G (see the discussion on the classical dynamics of the
random manifold and the dependence on k; the relation between G and is
the same here).
In these equations, as in the classical case, S() is the spectral density of
the bath:
Nb
X c2a
S() = ( a ) , (7.139)
2 a=1 ma a
that can also be taken of the form in (4.20),



S() = 20
e/ . (7.140)

172
Figure 28: The kernel in the Ohmic = 1 case for different values of T
( = 5).

Figure 29: The kernel at T = 0 for different values of (h = 1 = 0 =

=1
and = 5) in linear (left) and logarithmic (right) scales.

As usual, a counterterm cancels the mass renormalization. The and

kernels are independent of T and h and are thus identical to the classical
ones, see eq. (4.21).The kernel does depend on T and h . After a change of
variables in the integral, in the cases in which S() ,
!
h
(t) = t(1+) g , t (7.141)
t
and
lim h(t) = 2(t) . (7.142)

h/t0

The classical limit is realized at high temperature and/or long times.

Figure 29-left shows the time-dependence of the kernels in the quantum
Ohmic case for different values of T . In the right panel of the same figure we
show the dependence on of the kernel .
Upper critical initial conditions for the system
Next, we have to choose an initial density matrix for the system. One
natural choice, having in mind the quenching experiments usually performed
in classical system, is the diagonal density matrix
hX| syst (0) |X i = (X X ) (7.143)
that corresponds to a random high-temperature situation and that simplifies
considerably the expression in (7.132). As in the classical stochastic problem,
Tr
red (0) = 1 and it is trivially independent of disorder. Hence, there is no
need to introduce replicas in such a quantum dynamic calculation.
Equilibrium initial conditions for the system
One might also be interested in using equilibrium initial conditions for
the isolated system
hX| syst (0) |X i = Zsys
1
hX| eHsyst |X i . (7.144)

173
This factor introduces interesting real-time imaginary-time correlations.

7.2.8 Transformation to MSR-like fields

Newton dynamics
We use x = x (h/2) x. We first study the kinetic terms:
!2 !2
iZ m 2 iZ m h
h

dt x + x x x

dt x + x 2 = (7.145)

h
2 h
2 2 2

expanding the squares the integrand equals 2h x x and integrating by parts

Z
dt i
x m
x. (7.146)

The potential term

The potential term is treated similarly
" ! !#
i i h
h

[V (x+ ) V (x )] = V x + x V x x (7.147)
h
h
2 2
Note that this expression is specially simple for quadratic and quartic poten-
tials:
V (y) = y 2

xi
x  quadratic
=  2 
x x2 20 + h2 (i
4xi x )2 quartic V (y) = (y 2 y02 )2

The noise terms

We introduce the variables x and i
x in the action terms generated by the
coupling to the environment and the classical limit of the kernel :
i T T d(t t )
Z Z
Senv = dt dt i
x(t ) 2(t t ) 2x(t )
h
0 0 dt
Z T Z T
+ dt dt i
x(t )h(t t )i
x(t )
0 0
Z T Z T

= 4 dt dt i
x(t ) (t t ) x(t
)
0 0
Z T Z T
+ dt dt i
x(t )h(t t )i
x(t ) . (7.148)
0 0

174
(Apart from a border term that can be treated more carefully and check that
it disappears.)

7.2.9 Classical limit

0, the Schwinger-Keldysh effective action re-
In the classical limit, h
duces to the one in the functional representation of a Newton classical me-
chanics (no bath) or a Langevin process with coloured noise (bath). The
kinetic term eq. (7.146) is already in the expected form with no nead to take
0. The potential term
h
i
x + O(h2 ) V (x)i
[V (x+ ) V (x )] = V (x)i x (7.149)
h

upto first order in h
. In the noise terms we just have to take the classical
limit of the kernel :
i
Z T Z T

Senv 4 dt dt i
x(t ) (t t ) x(t
)
h
0 0
Z T Z T
+2kB T dt dt i
x(t ) (t t ) i
x(t ) (7.150)
0 0

with O(h) corrections (apart from a border term that can be treated more
carefully and check that it disappears.)

7.2.10 A particle or manifold coupled to a fermionic reservoir

The coupling to leads are modeled by interactions to electron creation
and destruction operators in the form:
N
1 X
Vkk x Lk

Rk + Rk

 
Hint = Lk (7.151)
N k,k=1

L and R are left and right labels associated to two leads. L(R)k

and L(R)k
are creation and destruction operators, respectively, acting on the left (L)
and right (R) leads. k is the wave-vector of the electron. N is the number
of fermionic degrees of freedom that we take N . Vkk is the coupling
constant for the kk reservoir degrees of freedom. For simplicy we take Vkk =
c .
h

175
 The influence functional is
Z ()=+ ()
F [x+ , x ] = DD e h (Sint [x
i + ,+ ]S
int [x
, ]
)

e h (Senv [
i + ]S
env [
]
) h+ (0)| | (0)i
env

The coordinates represent the fermionic variables of the reservoirs:

(L , L , R , R ). We expand the reservoir-system coupling at second order
(first order gives no contribution) to exhibit an effective action SK,eff by the
following procedure
2
i 1 i 2

ienv e 2 h( h SK,int )
i 1 i
S ienv
he K,int
h ienv 1 + h SK,int ienv + h SK,int
h
2 h
i
SK,eff
e h
. (7.152)

We used the notation h...ienv to represent the functional integral over the
fermions evolution. The electron reservoirs contribution to the action is
N M 
1 X
Z 
a a a a
xa (t)
X X
SK,int a dt h
c Lk Rk + Rk Lk . (7.153)
a= i=1 M k=1

We have
1 2 1 X Z
2 hS i
K,int env = 2 ab dtdt (hc )2 xa (t)xb (t ) (7.154)
h
h
a,b=
M
1 X 
a a

b b

h Lk (t)Rk (t) + L R Lq (t )Rq (t ) + L R ienv .
M2 kk qq =1

Since left and right reservoirs are independant, operators of left reservoir
commute with the right ones and the averaging factorizes : hT R L L R i =
hT R R ihT L L i. Sites i and j 6= i live in different spaces so we have
hT i j i = ij hT i i i. Conservation of mometum implies hT k q i = kq hT k k i
and similarly hT k q i = k q hT k k i. Moreover, we use the fermionic prop-
erty hT i i i = hT i i i = 0 to finally simplify the expression (7.154) into
1 2 1 X
Z Z
hS i env = ab dtdt (hc )2 xa (t)xb (t ) (7.155)
2 K,int
h 2 a,b=
h
M M
!
1 X a 1 X b
hLk b
(t)Lk (t )ienv a
hRk (t)Rk (t )ienv + L R
M k=1 M k=1

176
Recalling the definition of the fermionic green functions
1 PM a b
Ga,b (t, t ) i M k=1 hTC k (t)k (t )i, and reconstructing an exponential
into an effective action SK,eff of the general form
1
Z Z
dtdt xa (t)ab (t, t )xb (t ) ,
X
SK,eff (7.156)
a,b= 2

we identify the self-energy components

h i
ab (t, t ) iabhc2 GR L
ab (t, t )Gba (t , t) + L R . (7.157)

0 , in equilibirum in the grand-canonical

For a single free fermion, H = h
ensemble, 0 e(HN ) , the Green function is
G+ ( ) = inF ei0 G+ ( ) = i(1 nF )ei0 (7.158)
where nF is the Fermi factor nF [1 + e(h0 ) ]1 . After the Keldysh
rotation we have
i
GK ( ) = 2 tanh[/2 (h0 )] ei0 , GR ( ) = ei0 ( ) = GA ((7.159)
).
h

The reservoir has a collection of free fermions (N ) with a distribution
of energy levels, or density of states, F (0 ). Thus,
Z
GK ( ) = 2 d0 (0 ) tanh[(h0 )/2] ei0
= 2 h tanh[(h0 )/2] ei0 i0
i i
Z
GR ( ) = d0 (0 ) ei0 ( ) = h ei0 i0 ( ) (7.160)
h
h

i i
Z
GA ( ) = d0 (0 ) ei0 ( ) = h ei0 i0 ( )
h
h

where we introduced a short-hand notation, the angular brackets, for the
integration over energy levels. In the end, the explicit calculation of the
kernels originating from the coupling to the bath yields
(hc )2
( )
L R
K ( ) = hh tanh[ (hL L )] tanh[ (hR R )] 1
2 2 2
cos[(L R ) ]iL iR
( )
21 L
R ( ) = (hc ) hh tanh[ (hL L )] tanh[R 2(hR R )]
h
2
sin[(L R ) ]iL iR ( ) (7.161)

177
 Two important energy scales characterize the density of states, (): their
mean-square-displacement or bandwidth, h F , and whether they have a
finite cut-off, h
cut or not. Some typical examples one can use are
q


F1 1 ( F )2 /F2 semi-circle h cut < ,
F , h

 2
21

() = F1 cut (7.162)
q
F
F
e Oscil. F < h
h ,

 (d2)/2
F1

Elect. gas F h
h cut .


F

Clearly, the maximal current one can apply to the system is eVmax = h cut
0 .
Some interesting limiting cases of the kernels are worth mentioning:
F h
For h and kB T . We obtain the Ohmic ( ) behaviour
1
ImR () 2(hc )2 2 (0 /h + eV /h) (7.163)
h

Interesting enough, this expression is independent of T . Simultane-
ously,

1 eV sinh(eV ) h
sinh(h)
K () 2(hc )2 2 (0 /h + eV /h)
h
cosh(eV ) cosh(h)
(7.164)

In equilibrium V = 0 one recovers the FDT relation

1
ImR () 2(hc )2 2 (0 /h) (7.165)
h

h

K () 2(hc )2 2 (0 /h) coth (7.166)
2
kB T h
eqs. (7.163) and (7.164). For h F
1
K () 4(hc )2 2 (0 /h) kB T . (7.167)
h

At T = 0 and eV h
F , one derives from eq. (7.164)
1
K ( = 0, T = 0, V ) 2(hc )2 2 (0 /h)|eV | (7.168)
h

178
 This last expression is of the same form as eq. (7.167) suggesting the
definition of an equivalent temperature

kB T |eV |/2 (7.169)

eV
A FDT like relation is verified up to the second order in F
h

eV 2 h
eV 1
K [T = 0, = 0, O( coth( )Im R [O()1 , O( (7.170)
) ]=h )]
h
2 h

7.2.11 Other baths

The electron-phonon interactions, in which the phonons are considered
the bath, was dealt with in [17]. Spin baths are reviewed in [18].

179
8 Quantum glassiness
In this section we discuss the dynamics of a number of quantum systems.
See the slides and a review article by L. F. Cugliandolo in the arXiv.

8.1 Quantum driven coarsening

Take the quantum O(N) model and couple it to two electronic reservoirs.
As we have seen this environment encompasses as a limit the oscillator case,
and allows for the application of a current through the system.
In MSR fields the action reads
i ~ ~
Z  
S = d
dtd x i(~x, t)[m(~ ~ x, t)]
x, t) + 2 (~
h
!2
4 ~ ~ x, t) 2 2 h
~
+ i(~x, t) (~ 0 i(~x, t)
N 2
~ ~ x, t) + i~(~
Z  
= + d
dtd x i(~x, t)4(t t )(~ x, t)h(t t )(~
~ x, t)

Apart from the last term in the potential part this action is like the MSR one
for a vector field theory in a strange bath. In the large N limit we assume
that the term between sqaure brackets can be replaced by its average (as
in the classical case) the value of which will be fixed self-consistently. The
contribution h(i) 2 i vanishes due to causality. We are then in presence of an
effective classical field theory in contact with an environment with quantum
character.
The main results are [13]
There is a coarsening phase in the (T, , eV ) phase diagram. It survives
the finite current imposed by the leads. Whether there is a finite critical
eVmax (at T = = 0) or the ordering phase extends to infinity in this
direction depends upon the type of electron bath used.

The coupling strength, say g, between the baths and the system plays a
very important role in the determination of the phase diagram. The ex-
tent of the sub-critical region in which the system will undergo coars-
ening after a quench (or annealing) increases for increasing coupling
strength although the classical critical temperature is not modified by
g. The effect is highly non-trivial in this very simple model: the action

180
 is almost classical, apart from the last term that arises from the cou-
pling to the bath and involves the kernel . Although at long-times this
kernel becomes classical one cannot simply argue that it will not have
an effect on the dynamics of the system; on the contrary, its short-
time effects alter the dynamics and make the phase transition occur
elsewhere.
This effect is similar to the Caldeira-Leggett localisation phenomenon.
A static calculation in which the partition function of the system and
the environment (in the case V = 0) gives the same critical line as the
dynamic one for this model. This is not necessarily the case, as we
shall see below with the random manifold.

In the ordering phase the correlations and linear responses present a

separation of two-time scales, similar to the one encountered in the
classical limit with a well-defined plateau in the correlation achieved at
a value that depends on the parameters (T, , eV ).

The relaxation in the stationary regime depends strongly on the param-

eters (T, , eV ). The correlations may ne non-monotonic and present
oscillations depending on the values of these parameters. The correla-
tions and linear-responses are related by the quantum FDT.

The study of the correlation functions in the aging regime yields a

growing correlation length R(t; T, , eV ) with a weak dependence upon
the parameters T, , eV and t1/2 growth in time as in the classical limit.

The scaling functions are the same as in the classical limit and do
not depend on the parameters (T, , eV ). This is an extension of the
super-universality hypothesis to the quantum case.

The relation between the linear-response and the correlation is not the
equilibrium FDT but a modified one with an effective temperature that
for diverges in the coarsening regime.

We conclude that the environment plays a dual rle: its quantum char-
acter basically determines the phase diagram but the coarsening process at
long times and large length-scales only feels a classical white bath at tem-
perature T . The two-time dependent decoherence phenomenon (absence of
oscillations, validity of a classical FDT when t/tw = O(1), etc.) is intimately

181
related to the development of a non-zero (actually infinite) effective temper-
ature, Tef f , of the system as defined from the deviation from the (quantum)
FDT. Tef f should be distinguished from T as the former is generated not
only by the environment but by the system interactions as well (Tef f > 0
even at T = 0). Moreover, we found an extension of the irrelevance of T
in classical ferromagnetic coarsening (T = 0 fixed-point scenario): after a
suitable normalization of the observables that takes into account all micro-
scopic fluctuations (e.g. qEA ) the scaling functions are independent of all
parameters including V and . Although we proved this result through a
map- ping to a Langevin equation that applies to quadratic models only, we
expect it to hold in all instances with the same type of ordered phase, say
ferromagnetic, and a long-time aging dynamics dominated by the slow mo-
tion of large domains. Thus, a large class of coarsening systems (classical,
quantum, pure and disordered) should be characterized by the same scaling
functions.

182
A Conventions

A.1 Fourier transform

The convention for the Fourier transform is
Z d i
f ( ) = e f () , (A.171)
2
Z
f () = d e+i f ( ) . (A.172)

The Fourier transform of the theta function reads

1
() = ivp + () . (A.173)

The convolution is
d
Z
[f g]() = f g() f ( )g( ) . (A.174)
2

A.2 Commutation relations

We defined the commutator and anticommutator: {A, B} = (AB+BA)/2
and [A, B] = (AB BA)/2.

A.3 Time ordering

We define the time odering operator acting on bosons as

q (t ) (t, t )
T q(t) q (t ) + (t , t)
q (t) q(t )
q (t) . (A.175)

For fermions, we define the time ordering operator as

(t
T (t) ) (t, t )(t)
(t ) (t , t)(t
)(t)
, (A.176)
(t ) (t, t )(t)
T (t) (t ) (t , t) (t )(t)
, (A.177)

In both cases (t, t ) is the Heaviside-function.

183
 We define the time-ordering operator TC on the Keldysh contour in such
a way that times are ordered along it:

TC x+ (t)x (t ) = x (t )x+ (t) TC x (t)x+ (t ) = x (t)x+ (t )

TC + (t) (t ) = (t )+ (t) TC (t)+ (t ) = (t)+ (t ) (A.178)

for all t and t .

B The instanton calculation

The path-integral formalism yields an alternative calculation of the Kramers
escape time, the Arrhenius exponential law and its prefactor that, in prin-
ciple, is easier to generalize to multidimensional cases. For the sake of sim-
plicity let us focus on the overdamped limit in which we neglect inertia. We
first rederive the Arrhenius exponential using a simplified saddle-point argu-
ment, and then show how Kramers calculation can be recovered by correctly
computing the fluctuations around this saddle point. Starting from the fol-
lowing representation of the probability to reach the top of the barrier from
the potential well:
*Z  !+
x(t)=xmax  eq[x](t) 
P (xmax , t|xmin ) = Dx ( eq[x]) det ,
 
x(t )

x(0)=xmin  

and neglecting the determinant (which is justified if one follows the It con-
vention), then, for a Gaussian white noise :
Z x(t)=xmax Rt 2
4k 1 dt (x+ dx )
dV
P (xmax , t|xmin ) = Dx e BT 0
x(0)=xmin

Expanding the square, we find a total derivative contribution to the integral

Rt 2
equal to 2[V (xmax ) V (xmin )], plus the sum of two squares: 0 dt [x +
(V (x))2 ]. For small T , the path, x , contributing most to the transition
probability is such that this integral is minimized. Using standard rules of
functional derivation one finds
d2 x
2
= V (x )V (x ) x = V (x ).
dt
In order to be compatible with the boundary conditions x (0) = xmin and
x(t) = xmax , the + solution must be chosen, corresponding to an overdamped

184
motion in the inverted potential V (x). The action of this trajectory is
Z t h i Z t
2
dt x + (V (x )) 2
=2 dt x V (x ) = 2[V (xmax ) V (xmin )],
0 0

that doubles the contribution of the total derivative above. Hence,

P (xmax , t|xmin ) e(V (xmax )V (xmin )) ,

independently of t, as in eq. (4.64). This type of calculation can be readily

extended to cases in which the noise has temporal correlations, or non
Gaussian tails, and to see how these effects change the Arrhenius result. The
calculation of the attempt frequency is done using the standard dilute gas
instanton approximation dveeloped by several authors [35] but we shall not
discuss it here.
The path-integral that we have just computed is a sum over the subset
of noise trajectories that lead from the initial condition to a particular final
condition that we imposed. Imposing a boundary condition in the future
destroys the causal character of the theory.
In a one dimensional problem as the one treated in this Section there is
only one possible reaction path. In a multidimensional problem, instead, a
system can transit from one state to another following different paths that
go through different saddle-points. The lowest saddle-point might not be
the most convenient way to go and which is the most favorable path is, in
general, difficult to established.

185
References
[1] P. G. Debenedetti, Metastable liquids (Princeton Univ. Press, 1997).
E. J. Donth, The glass transition: relaxation dynamics in liquids and
disordered materials (Springer, 2001). K. Binder and W. Kob, Glassy
Materials and Disordered Solids: An Introduction to their Statistical
Mechanics (World Scientific, 2005).

[2] K. H. Fischer and J. A. Hertz, Spin glasses (Cambridge Univ. Press,

1991). M. Mzard, G. Parisi, and M. A. Virasoro, Spin glass theory and
beyond (World Scientific, 1986).

[3] L. F. Cugliandolo, Dynamics of glassy systems, in Les Houuches 2002

(Springer, 2003).

[4] O.C. Martin, R. Monasson, R. Zecchina, Statistical mechanics methods

and phase transitions in optimization problems Theoretical Computer
Science 265 (2001) 3-67. M. Mzard and A. Montanari, Information,
Physics, and Computation, (Oxford Graduate Texts, 2009).

[5] D. J. Amit, Modeling Brain Function: The World Of Attractor Neural

Networks, (Cambridge Univ. Press, 1992). N. Brunel, Network models
of memory, in Les Houches 2003 (Elsevier, 2004).

[6] R. Zwanzig, J. Stat. Phys. 9, 215 (1973). S. Nordholm and R. Zwanzig,

J. Stat. Phys. 13, 347-371 (1975). K. Kawasaki, J. Phys. A 6, 1289-1295
(1973).

[7] U. Weiss, Quantum dissipative systems, Series in modern condensed

matter physics vol. 10, World Scientific (1999).

[8] J. Dunkel and P. Hnggi, Phys. Rep. 471, 1-73 (2009).

[9] R. Bausch, H. K. Janssen, and H. Wagner, Z. Phys. B 24, 113 (1976).

[10] L. F. Cugliandolo, J. Kurchan, and L. Peliti, Phys. Rev. E 55, 3898

(1997).

[11] P. Le Doussal, Exact results and open questions in first principle func-
tional RG arXiv:0809.1192

186
[12] A. Sicilia, J. J. Arenzon, A. J. Bray, and L. F. Cugliandolo, Phys. Rev.
E

[13] C. Aron, G. Biroli, and L. F. Cugliandolo, Phys. Rev. Lett. 102, 050404
(2009).

[14] L. F. Cugliandolo and P. Le Doussal, Phys. Rev. E L. F. Cugliandolo,

J. Kurchan, and P. Le Doussal, Phys. Rev. Lett.

[15] A. Sicilia, G. Biroli and L. F. Cugliandolo, Phys. Rev. E

[16] Phys. Rev. E 65, 046136 (2002).

[17] C. Caroli, R. Combescot, P. Nozieres and D. Saint-James, J. Phys. C 5,

21 (1972).

[18] Theory of the spin bath N. Prokofev and P. Stamp Rep. Prog. Phys. 63,
669 (2000).

[19] H. Grabert, P. Schramm and G.-L. Ingold, Phys. Rep. 168, 115 (1988).

[20] R. MacKenzie, Path integral methods and applications, VIth Vietnam

School of Physics, Vung Tau, Vietnam, 27 Dec. 1999 - 8 Jan. 2000;
arXiv:quantum-ph/0004090.

[21] R. P. Feynman and A. R. Hibbs, Quantum Mechanics and Path Inte-

grals (New York: McGraw-Hill, 1965). H. Kleinert, Path Integrals in
Quantum Mechanics, Statistics, Polymer Physics, and Financial Mar-
kets, 4th edition (World Scientific, Singapore, 2004) J. Zinn-Justin, Path
Integrals in Quantum Mechanics (Oxford University Press, 2004), R. J.
Rivers, Path Integrals Methods in Quantum Field Theory (Cambridge
University Press, 1987).

[22] R. P. Feynman, Statistical mechanics,

[23] J. Kurchan, Supersymmetry, replica and dynamic treatments of disor-

dered systems: a parallel presentation arXiv:cond-mat/0209399.

[24] M. Mzard, G. Parisi, and M. A. Virasoro, Spin glasses and beyond,

Fischer and J. Hertz,

187
[25] A. B. Kolton, D. R. Grempel and D. Domnguez, Phys. Rev. B 71,
024206 (2005)

[26] A. J. Bray, Adv. Phys. 43, 357 (1994). P. Sollich,

http://www.mth.kcl.ac.uk/ psollich/

[27] P. C. Hohenberg and B. I. Halperin, Rev. Mod. Phys. 49, 435 (1977).
P. Calabrese and A. Gambassi, J.Phys. A 38, R133 (2005).

[28] D. W. Oxtoby, Homogeneous nucleation: theory and experiment, J.

Phys.: Condens. Matter 4, 7626-7650 (1992). K. Binder, Theory of first-
order phase transitions, Rep. Prog. Phys. 50, 783-859 (1987).

[29] F. Corberi, E. Lippiello and M. Zannetti J. Stat. Mech. (2007) P07002.

[30] Viasnoff and Lequeux

[31] A. Cavagna, Supercooled Liquids for Pedestrians Physics Reports 476,

51 (2009).

[32] U. Weiss, Quantum dissipative systems, Series in modern condensed

matter physics vol. 10, World Scientific (1999).

[33] L. F. Cugliandolo and G. Lozano, Phys. Rev. Lett. 80, 4979 (1998) .
Phys. Rev. B59, 915 (1999).

[34] L. F. Cugliandolo, D. R. Grempel, G. Lozano and H. Lozza Phys. Rev.

B 70, 024422 (2004). L. F. Cugliandolo, D. R. Grempel, G. S. Lozano,
H. Lozza, C. A. da Silva Santos Phys. Rev. B 66, 014444 (2002).

[35] J. S. Langer, Statistical theory of decay of metastable states Ann. of

Phys. 54, 258 (1969). C. G. Callan and S. Coleman, Phys. Rev. D 16,
1762 (1977). B. Caroli, C. Caroli, and B. Roulet, Diffusion in a bistable
potential - systematic WKB treatment J. Stat. Phys. 21, 415 (1979). A.
M. Polyakov, Gauge fields and stings (Harwood Academic Publishers,
1994).

[36] R. P. Feynmann and F. L. Vernon, Jr, Ann. Phys. 24, 114 (1963).

[37] R. P. Feynmann and A. R. Hibbs, Quantum mechanics and path integrals

(Mc Graw-HIll, New York, 1965).

188
[38] A. J. Leggett, S. Chakravarty, A. T. Dorsey, M. P. A. Fisher, A. Garg,
W. Zwerner, Rev. Mod. Phys. 59, 1 (1987).

[39] A. J. Bray and M. A. Moore, Phys, Rev. Lett. 49, 1546 (1982).

[40] R. Zwanzig, J. Stat. Phys. 9, 215 (1973). Kawasaki,

[41] Hanggi on Langevins proof.

[42] C. P. Martin, E. Siggia and H. A. Rose, Phys. Rev. A8, 423 (1973), H. K.
Janssen, Z. Phys. B23, 377 (1976) and Dynamics of critical phenomena
and related topics, Lecture notes in physics 104, C. P. Enz ed. (Springer
Verlag, Berlin, 1979).

[43] C. de Dominicis, Phys. Rev. B18, 4913 (1978).

[44] J. Schwinger, J. Math. Phys. 2, 407 (1961). L. V. Keldysh, Zh. Eksp.

Teor. Fiz. 47, 1515 (1964), Sov. Phys. JETP 20 235 (1965).

[45] A. Kamenev,

[46] G. Parisi, Statistical Field Theory, Frontiers in Physics, Lecture Notes

Series, Addison-Wesley (1988).

189