UWThPh-1997-52

27. Oktober 1997

History and outlook of statistical physics 1

Dieter Flamm
Institut f
ur Theoretische Physik der Universitat Wien,
Boltzmanngasse 5, 1090 Vienna, Austria
Email: Flamm@Pap.UniVie.AC.AT

Abstract

This paper gives a short review of the history of statistical physics starting from D.
Bernoullis kinetic theory of gases in the 18th century until the recent new developments
in nonequilibrium kinetic theory in the last decades of this century. The most important
contributions of the great physicists Clausius, Maxwell and Boltzmann are sketched. It is
shown how the reversibility and the recurrence paradox are resolved within Boltzmanns
statistical interpretation of the second law of thermodynamics. An approach to classical
and quantum statistical mechanics is outlined. Finally the progress in nonequilibrium
kinetic theory in the second half of this century is sketched starting from the work
of N.N. Bogolyubov in 1946 up to the progress made recently in understanding the
diffusion processes in dense fluids using computer simulations and analytical methods.

1
Paper presented at the Conference on Creativity in Physics Education, on August 23, 1997, in
Sopron, Hungary.
 1

In the 17th century the physical nature of the air surrounding the earth was es-
tablished. This was a necessary prerequisite for the formulation of the gas laws. The
invention of the mercuri barometer by Evangelista Torricelli (160847) and the fact that
Robert Boyle (162791) introduced the pressure P as a new physical variable where im-
portant steps. Then BoyleMariottes law P V = const. for constant temperature, where
V is the volume, was formulated.

First Kinetic Theory of Gases

Daniel Bernoulli (170082) who had been born in Groningen in the Netherlands and
moved to Basle in Switzerland gave in 1738 in a treatise on hydrodynamics a derivation
of the gas laws from a billiard ball model. He assumed that the gas consists of a
very large number of small particles in rapid motion. He already identified heat with
kinetic energy (living force). He then derived BoyleMariottes law for the gaspressure
at constant temperature on a movable piston from the impact of the gas molecules
onto the piston just in the same way as it is still done today in elementary text books.
He also used the principle of conservation of mechanical energy and concluded that, if
the temperature changes, the pressure will also change so that it is proportional to the
square of the velocities of the gasparticles and thus, for constant volume, will rise with
rising temperature. At that time mechanical energy was still called vis viva in Latin or
living force in English. Bernoulli was, however, about a century ahead of his time with
his kinetic theory of gases. His model was almost forgotten. At that time only his Swiss
copatriots J.A. De Luc (17271817) and GeorgeLouis Le Sage (17241803) in Geneva
and M.V. Lomonossov (171165) in Russia mentioned it. As Stephen Brush points out
in his book on the kinetic theory The man who persuades the world to adopt a new
idea has accomplished as much as the man who conceived that idea.2 The reason why
Bernoullis kinetic theory received so little attention was that most scientists at that
time believed in the socalled caloric theory of heat.

The Caloric Theory

In the caloric theory heat was a substance called caloric. Caloric was considered to
be a fluid composed of particles which repel each other. Like the earth is surrounded by
its atmosphere each matter particle was thought to be surrounded by an atmosphere
of caloric whose density increases with temperature. Thus at small distances matter
particles repel each other due to the repulsion of their caloric atmospheres. In those days
one did not know that matter is held together by electrical forces but one thought that
the attractive forces between matter particles are of gravitational origin just as the forces
between the sun and the planets. At a certain distance between matter particles there
would be equilibrium between the caloric repulsion and the gravitational attraction. As
the temperature rises more caloric is added to each matter particle and consequently
the caloric repulsion increases shifting the equilibrium point outwards. In this way one
could explain the thermal expansion of matter including gases. It should be mentioned
that such reputed scientists as Pierre Simon, Marquis de Laplace (17491827) gave a
very sophisticated derivation of the gas laws within the caloric theory.
2
S.G. Brush: Kinetic Theory, vol. 1. The Nature of Gases and of Heat, Pergamon Press, London
1965, p. 9. Bernoullis paper and most other historic papers quoted below are reprinted in vol. 1 or in
vol.2: Irreversible Processes (1966) by S.G. Brush.
2

At the end of the 18th century various arguments against the caloric theory ap-
peared. For instance: Does caloric have weight? In 1798 Benjamin Thompson (1753
1814) could show that the expected additional weight when a body was heated could
not be detected. Another argument against the caloric theory was raised by Rumford
Humphry Davy (17781829). He remarked that an indefinite amount of heat can be
produced from matter by mechanical work, for instance by friction. If caloric was a
substance only a limited amount should be available in matter. But these arguments
did not really convince the fans of the caloric theory.
Now let me mention the attempts for a revival of the kinetic theory.

The Revival of the Kinetic Theory

In 1820 in England John Herapath, born 1790 at Bristol, critisized the derivation of the
gas laws by Laplace and gave an account of the kinetic theory, but his paper was not
accepted for publication in the Philosophical Transactions of the Royal Society because
it was considered to be too speculative. It finally appeared in 1821 in the Philosoph-
ical Transactions and would have been forgotten if James Prescott Joule (181889)
would not have been influenced by it and by Herapaths book on Mathematical Physics
published in 1847.
In 1847 Joule published a paper in the Manchester Courier with the title On Matter,
Living Force, and Heat where he stated the principle of conservation of energy. From
his experiments he concluded in this paper that heat is not a substance but a form of
energy: Experiment has shown that whenever living force [kinetic energy] is apparently
destroyed or absorbed, heat is produced. The most frequent way in which living force
is thus converted into heat is by means of friction. Furthermore he gave the amount of
heat equivalent to the converted kinetic energy. A year later he read another paper in
Manchester where he used Herapaths kinetic theory. It was not published until 1851
and at first got very little attention until Clausius quoted it in 1857.
The real breakthrough for the kinetic theory started in Germany after Karl Kronig
(182279) published a paper about it in Poggendorfs Annalen der Physik in 1856. Con-
trary to the widespread belief that molecules of gas merely oscillate around definite
positions of equilibrium he assumed that they move with constant velocity in straight
lines until they strike against other molecules, or against the surface of the container.
Kronig was actually a chemist but he had a great reputation because he was editor
of Fortschritte der Physik, an annual review of physics, and he had great influence in
the German Physical Society. Kronigs paper apparently motivated Rudolf Clausius
(182288) to publish on the kinetic theory. In fact Clausius writes that already before
his first paper on heat in 1850 he had a very similar conception of heat as Kronig but
in his former papers he intentionally avoided mentioning this conception, because his
conclusions were deducible from general principles and did not depend on these special
conceptions. Clausius was already well known from his papers on thermodynamics. In
1850 he had given his verbal formulation of the second law that there exists no thermo-
dynamic transformation whose sole effect is to extract a quantity of heat from a colder
reservoir and to deliver it to a hotter reservoir. In the years until 1854 he had worked
out its mathematical formulation. Since 1855 he was professor at the ETH in Z urich.
When Clausius started to work on the kinetic theory it became fashionable. In 1857 he
published his first paper on the kinetic theory with the title The Nature of the Motion
which we call Heat where he quoted the papers of Kronig and of Joule. The English
 3

translation of his paper appeared in the same year in the Philosophical Magazine. Two
important arguments against the kinetic theory of heat were the following:

How can heat traverse a vacuum if it is just irregular motion of matter particles?
There is no matter in the vacuum which could propagate heat while the particles of
caloric could easily penetrate through the vacuum.

C.H.D. BuysBallot (181790) argued that since gas particles in the kinetic theory
move with velocities of a few hundred meters per second one would expect that gases
diffuse and mix much more rapidly than observed.

In 1858 Clausius published a paper in which he could cope with the second of these
objections by introducing the mean free path of a gas molecule. Gas molecules move
at speeds of a few hundred meters an hour but they undergo collisions with other gas
molecules which change their direction after a very short time of flight. The actual
distance they can move on the average freely along a straight line in one direction is
the mean free path l given by
3 1 N
l= n= (1)
4 n 2 V
where n is the number density of gas molecules and is the diameter of the hard sphere
particles which approximate the gas molecules. For his estimate of the mean free path
Clausius made the drastic approximation that only one particle is moving and all others
are at rest. His result differs less than 10 per cent from the result in Eq. (3) obtained
by Maxwell one year later from a much more refined derivation. Clausius who in 1865
introduced the concept of entropy continued to work on the kinetic theory.
James Clerk Maxwell (183179), best known from his electromagnetic field theory
which he developed in the years from 1855 to 1873, read his first paper on the kinetic
theory in 1859 at a Meeting of the British Association at Aberdeen. With the title
Illustration of the Dynamical Theory of Gases it appeared 1860 in print in the Philo-
sophical Magazine. While in earlier treatments the absolute value of the velocities of
the molecules was considered to be rather uniform he was the first to assume a random
motion for the molecules. For thermal equilibrium he could then derive from symmetry
considerations his famous velocity distribution function which in modern notation is
given by
 3
m 2 m~v 2
f0 (~v ) = n exp (2)
2kT 2kT
where ~v is the velocity and n the density of the molecules, m their mass, k Boltzmanns
constant and T the absolute temperature. For the mean free path he then obtained
1 1
l= (3)
2 n 2
and for the viscosity of a dilute gas
!1
1 8kT 2
0 (T ) = nml
v with v = (4)
3 m
4

where v is the mean absolute value of the velocity. Inserting l into the last equation
he obtained a value independent of the density and because of v proportional to the
square root of the absolute temperature
1 m v
0 (T ) = . (5)
3 2 2
The density independence of the viscosity was quite unexpected since for a fluid the
viscosity in general increases with increasing density. After its experimental verification
this result served as a strong argument in favour of the kinetic theory.
The Boltzmann Equation
In 1872 Ludwig Boltzmann in Graz generalized Maxwells approach for the kinetic
theory of dilute gases to nonequilibrium processes, so that he could investigate the
transition from nonequilibrium to equilibrium. His nonequilibrium single particle dis-
.
tribution function f = f (~x, ~v , t) gives the average number of molecules in a dilute gas
at the position ~x with velocity ~v at time t. The temporal change of this distribution
function consists of two terms, a drift term due to the motion of the molecules and a
collision term due to collisions with other molecules. In the absence of an external field
of force this equation, which is now called Boltzmann equation, reads:
f f
= ~v + JB (f f ) . (6)
t ~x
Here JB (f f ) is the binary collision term which takes only two particle collisions into
account, a good approximation for a dilute gas. A further assumption in Boltzmanns
expression for the collision term is that the velocities of the colliding molecules must
be uncorrelated, which was later called the assumption of molecular chaos by Jeans.
Now Boltzmann introduced the funtional
Z Z
3
H[f ] = dx d3 vf (~x, ~v , t) log f (~x, ~v, t) (7)

for which he could show under very general assumptions for the intermolecular inter-
action that if f is a solution of Eq. (6) the time derivative of H is always smaller than
zero or at most zero:
dH[f ]
0. (8)
dt
Furthermore for an ideal gas in equilibrium he could show that the entropy S is up to
a sign proportional to H. For nonequilibrium this is a generalization of the thermody-
namic entropy now called Boltzmann entropy
S(t) = kH[f ] (9)
and Eq. (8) is nothing but the second law of themodynamics for a closed system
dS(t)
0. (10)
dt
This is Boltzmanns famous Htheorem.
The Htheorem and the Boltzmann equation met with violent objections from
physicists and from mathematicians. These objections can be formulated in the form of
paradoxes. The most important ones are the reversibility paradox formulated in 1876 by
Boltzmanns friend Josef Loschmidt (182195) and the recurrence paradox formulated
in 1896 by Ernst Zermelo (18711953).
 5

The Reversibility Paradox

In a paper in 1876 Loschmidt gave a recipe of how one can prepare an initial condition
with decreasing entropy for any system which follows a motion with increasing entropy.3
One need only reverse all its velocities at a certain instant of time: ~v(t) ~v (t).
This procedure is equivalent to time reversal and the argument is usually called the
reversibility paradox because the equations of classical mechanics are invariant under
time reversal while the Boltzmann equation is not. The procedure of time reversal vio-
lates the hypothesis of molecular chaos since it is like a film which runs backwards.
All molecules which have just had a collision will collide again and thus their velocities
are correlated. This paradox represented a severe objection to the mechanical interpre-
tation of the second law of thermodynamics. Apparently there are just as many initial
conditions which lead at least for a short time to a decrease in the entropy of the system
as there are initial conditions leading to an increase in the entropy. Why do we never
observe a decrease in entropy for large isolated systems? For very small systems one can
easily observe a decrease of entropy in the form of statistical fluctuations, e.g. density
fluctuations, local pressure fluctuations or local temperature fluctuations in very small
regions of a gas. For very large systems this is not the case because the small local
fluctuations average out. Statistical physicists have coined the term typicality for the
usual behaviour of macroscopic systems 4 . If one pours, for instance, a dye into a liquid
it will gradually spread through the whole liquid. This behaviour is typical whenever
you make such an experiment. It is easy to tell the sequence in which snapshots of a
spreading dye were taken, even after their original order has been deranged. How does
this unidirectional behaviour in time come about? For a very large system, by far the
largest number of states corresponds to equilibrium- and quasi-equilibrium-states. The
latter are states which differ very little from the equilibrium state with maximum en-
tropy and cannot be distinguished macroscopically from the equilibrium state. In our
example of the liquid containing a dye they correspond to a practically uniform distri-
bution of the dye through the whole liquid with very small local intensity fluctuations of
the dye. With increasing size of the system, the preponderance of the equilibrium- and
quasi-equilibrium-states becomes ever more overwhelming. If for every possible state of
a large system we put a marked sphere into an urn and afterwards drew spheres from
the urn indiscriminately, we would practically always draw an equilibrium- or quasi-
equilibrium-state. The transition from nonequilibrium to equilibrium thus corresponds
to a transition from exceptionally rare nonequilibrium-states to extremely probable
states. This is Boltzmanns statistical interpretation of the second law.
Statistical Mechanics
Loschmidts reversibility paradox led to a very fruitful discussion between Boltzmann
and Loschmidt about the second law and motivated Boltzmann to work out his sta-
tistical interpretation of the second law in detail. To handle the reversibility paradox
Boltzmann investigated the entire phase space of a dynamical system consisting of N
particles or molecules. He found that the volume in the 6N dimensional phase space of
the system which represents all possible values of the three coordinates and the three
momentum components of each particle can be subdivided into regions corresponding
to macroscopic states of the system which we shall call macrostates.
3
J. Loschmidt, Sitzungsber. Kais. Akad. Wiss. Wien, Math. Naturwiss. Classe 73, 128142 (1876).
4
J. L. Lebowitz: Boltzmanns Entropy and Times Arrow, Phys. Today 46 (9), 32 (1993).
6

In his paper of 1877 entitled On the relation between the second law of the me-
chanical theory of heat and the probability calculus with respect to the theorems on
thermal equilibrium5 , Boltzmann now presented a probabilistic expression for the en-
tropy. He could show that the entropy S is proportional to the 6Ndimensional phase
space volume occupied by the corresponding macrostate of an Nparticle system:

S log . (11)

It is now usually written in the notation of Max Planck

S = k log W (12)

where k is the Boltzmann constant and W is the number of microstates by which the
macrostate of the system can be realized. This relation has been called Boltzmanns
Principle by Albert Einstein (18791955) in 1905 since it can be used as the foundation
of statistical mechanics. It is not limited to gases as Eq. (9) but can also be applied to
liquids and solid states. It can be obtained from Eq. (11) by introducing cells of finite
volume in phase space as Boltzmann had already done in order to obtain a denumerable
set of microstates. It implies that the entropy is proportional to the logarithm of the so
called thermodynamic probability W of the macrostate which is just the corresponding
number of microstates. A macrostate is determined by a rather small number of macro-
scopic variables of the system such as volume, pressure and temperature. The latter two
correspond to averages over microscopic variables of the system. A microstate, on the
other hand, is specified by the coordinates and momenta of all molecules of the system.
Due to the large number of molecules there is a very large number of different choices
for the individual coordinates and momenta which lead to the same macrostate. It turns
out, that for a large system by far the largest number of microstates corresponds to
equilibrium and quasiequilibriumstates as we have already illustrated in the exam-
ple of the liquid containing a dye. The latter are states which differ very little from
the equilibrium state with maximum entropy and cannot be distinguished macroscopi-
cally from the equilibrium state. Thus this macrostate is the state of maximal entropy
and the transition from nonequilibrium to equilibrium corresponds to a transition from
exceptionally unprobable nonequilibrium-states to the extremely probable equilibrium
state. In Boltzmanns statistical interpretation the second law is thus not of absolute
but only of probabilistic nature. The appearance of socalled statistical fluctuations in
small subsystems was predicted by Boltzmann and he recognized Brownian motion as
such a phenomenon. The theory of Brownian motion has been worked out indepen-
dently by Albert Einstein in 1905 and by Marian von Smoluchowski. The experimental
verification of these theoretical results by Jean Baptiste Perrin was important evidence
for the existence of molecules.

The term Statistical Mechanics has actually been coined by the great American
physicist J. Willard Gibbs (18391903) at a meeting of the American Association for
the Advancement of Science in Philadelphia in 1884.6 This was one of the rare occa-
sions when Gibbs went to a meeting away from New Haven. He had been professor of
mathematical physics at Yale University since 1871 and had served nine years without
salary. Only in 1880, when he was on the verge of accepting a professoship at John
5
L. Boltzmann: Sitzungsber. Kais. Akad. Wiss. Wien, Math. Naturwiss. Classe 76 (1877) 373435.
6
M.J. Klein: The Physics of J. Willard Gibbs in his Time, Phys. Today, Sept. 1990, p. 40.
 7

Hopkins University, did his institution offer him a salary. He had realized that the pa-
pers of Maxwell and Boltzmann initiated a new discipline which could be applied to
bodies of arbitrary complexity moving according to the laws of mechanics which were
investigated statistically. In the years following 1884 he formulated a general framework
for Statistical Mechanics and in 1902 published his treatise.7
Gibbs started his consideration with the principle of conservation of the phase space
volume occupied by a statistical ensemble of mechanical systems. He considered three
types of ensembles.
The socalled microcanonical ensemble of Gibbs corresponds to an ensemble of
isolated systems which all have the same energy. Boltzmann called this ensemble
Ergoden.8 In this case each member of the ensemble coresponds to a different mi-
crostate and all microstates have the same probability.
The canonical ensemble of Gibbs corresponds to systems in contact with a heat bath.
In this case the energy of the individual systems is allowed to fluctuate around the mean
value E. If E is the energy of an individual system of the ensemble, its probability
  P
is proportional to an exponential function linear in the energy P exp kT which
E

is nowadays often called the Boltzmann factor.

For the grandcanonical ensemble of Gibbs not only the energy but also the number
of particles N of the individual systems is allowed to fluctuate around the mean value
N.
If we introduce the density in 6Ndimensional phase space for an ensemble of phys-
ical Nparticle systems
= (~x1 , p~1 , ~x2 , ~p2 , ...~pN , ~xN , t) (13)
the Gibbs entropy can be written in the form
Z Z
S = k 3 3
d x1 d x2 ... d3 p1 d3 p2 ...d3 pN log . (14)
V

Introducing finite cells in phase space the number of microstates becomes denumerable
and will be labelled by = 1, 2, ...W where W is the total number of microstates. The
expression for the entropy then becomes
X X
S = k P log P with P = 1 (15)

where P is the probability of the corresponding microstate. Eq. (15) has already the
same form as the corresponding expression for a quantum system with discrete energy
levels. We may thus use the procedure introduced by John von Neumann (19031957) in
1927 to determine the equilibrium distribution P . It can be found by demanding that
the entropy Eq. (15) becomes a maximum under certain subsidiary conditions which
implies that the variation of S with respect to the P vanishes.
For the microcanonical ensemble only the sum of all probabilities must be one and
if the total number of states is W one obtains the same probability P = W1 for all
microstates which implies that Eq. (15) reduces to Eq. (12).
7
J. W. Gibbs: Elementary Principles in Statistical Mechanics. Developed with Especial Reference
to the Foundation of Thermodynamics. Yale Univ. Press 1902.
8
L. Boltzmann: Uber die Eigenschaften monozyklischer und anderer damit verwandter Systeme.

Crelles Journal 98 (1884) p.6894, Uber die mechanischen Analogien des zweiten Hauptsatzes der
Thermodynamik. ibid 100 (1887) p. 201212 and Vorlesungen u ber Gastheorie, II. Teil, J.A. Barth,
Leipzig 1898, p. 89.
8

For the canonical ensemble the fluctuations of the energy E of the individual sys-
tems around the mean value E requires the subsidiary condition
X
P E = E . (16)

The maximum entropy principle then gives for the probabilities

 
exp E
X  
kT E
P = with Z= exp (17)
Z kT
where T is the absolute temperature and Z the canonical partition function.
For the grandcanonical ensemble the fluctuations of the number of particles N of
the individual systems around the mean value N require in addition to Eq. (16) the
subsidiary condition X
P N = N , (18)

which has to be multiplied by a Lagrange multiplier and added to the entropy when the
variation of P is performed. This way one obtains for the grand canonical probability
distribution
 
N E
X  
exp kT N E
P = with = exp (19)
kT
where is the chemical potential and the grand canonical partition function.
The quantum mechanical generalization of the framework of statistical mechanics
has already been given in 1927 by John von Neumann who was born in Hungary. He
introduced the density operator = (p1 , x
1 , ...
pN , x
N , t) where the hat denotes the
operator character. The quantum mechanical generalization of Eq. (14) is then given
by9
S = kTr( log ) (20)
where Tr denotes the trace. If one uses for the trace eigenstates of the Hamiltonian and
the energy spectrum is discrete, the expression in Eq. (20) reduces to the form given
in Eq. (15). An important difference between classical and quantum statistics is the
symmetry of the state under permutations of the particles. While for classical particles
permutations of the particles lead to a different state identical quantum mechanical
particles can only be in an antisymmetric state if they are fermions or in a symmetric
state if they are bosons. An important application of Eq. (20) is the treatment of
ferromagnetism.

Progress in Nonequilibrium Kinetic Theory

Efforts for a systematic derivation as well as a generalization of the Boltzmann equation
have been made by N.N. Bogolyubov in 1946.10 The starting point for the derivation
of the Boltzmann equation is the time reversal invariant Liouville equation for the
Nparticle phase space density in Eq. (13)

= {H, } (21)
t
9
J. v. Neumann: Gottinger Nachr. (1927) 273.
10
N.N. Bogolyubov: J. Phys. (USSR) 10 (1946) 256.
 9

where {, } is the Poisson bracket and H the Hamiltonian of the system. Eq. (21) follows
from the classical equations of motion and expresses the conservation of the probability
in phase space. To arrive at the Boltzmann equation which violates time reversal invari-
ance, because the direction of increasing entropy is singled out, some coarse graining is
necessary which is done by successively integrating over the coordinates and momenta
.
of N 1 particles until one arrives at the one particle distribution function f = f (~x, ~v, t)
which appeares in Eq. (6).11 This way one arrives at the socalled B.B.G.K.Y. chain
of equations which stands for the first letters of the physicists Bogolyubov, Born and
Green, Kirkwood, Yvon. Then one has to perform the limit of low density, the socalled
BoltzmannGradLimit and make the assumption of molecular chaos for the initial dis-
tribution which implies factorization of the reduced nparticle densities into products
of one particle densities. Furthermore one has to assume that the system is large enough
so that the influence of the walls of its container is negligible. There exist a number
of such derivations of the Boltzmann equation of which I would like to mention the
one by O.E. Lanford12 and one by an Italian group.13 While Lanfords derivation holds
only for a very short time interval, the Italian derivation implies so low densities that
nearly no collisions take place. These restrictions are not surprising since at higher den-
sities correlated collision sequences appear such as the ring collisions shown in Fig.
1 which introduce correlations between the colliding particles violating the assump-
tion of molecular chaos. In this way the relaxation to equilibrium is slowed down. In
computer simulations of a gas of hard spheres which are often called relaxation exper-
iments one can study the approach to the one particle equilibrium distribution. Alder
and Wainwright did pioneering experiments of this kind in 1958.14 Quite recently such
experiments have also been performed at Vienna University.15 Following the analytic
methods of Bogolyubov a modified Boltzmann equation of the type
f f
= ~v + J(f f ) + K(f f f ) + L(f f f f ) + ... (22)
t ~x
is obtained where J(f f ) containes the two particle collisions, K(f f f ) the three parti-
cle collisions, L(f f f f ) the four paticle collisions etc.. The solution of this generalized
Boltzmann equation with the ChapmanEnskog method then leads to a density or virial
expansion for the viscosity

(n, T ) = 0 (T ) + n1 (T ) + n2 2 (T ) + ... (23)

which is a power series expansion in the density n with temperature dependent co-
efficients. For higher densities, however, it turns out that especially due to the ring
collision terms the Bogolyubov collision integrals K(f f f ) and L(f f f f ) become di-
vergent and a cut off for the mean free path has to be introduced. The revised density
11
For a review see e.g. R. Jancel: Foundations of Classical and Quantum Statistical Mechanics,
English translation ed. D. Ter Haar, Pergamon Press, Oxford 1969.
12
O.E. Lanford: Time Evolution of Large Classical Systems, in Dynamical Systems, Theory and
Application, edited by J. Moser, Springer, Berlin, 1975.
13
R. Illner and M. Pulvirenti: Global Validity of the Boltzmann Equation for Two- and Three-
Dimensional Rare Gas in Vacuum: Erratum and Improved Result, Commun. Math. Phys. 121 (1989)
143146.
14
B.J. Alder and T. Wainwright, in Transport Processes in Statistical Mechanics, edited by I. Pri-
gogine, WileyInterscience, New York, 1958.
15
Ch. Dellago and H.A. Posch: Mixing, Lyapunov instability, and the approach to equilibrium in a
hard sphere gas, Phys. Rev. E 55 (1997) R9.
10

expansion for the viscosity now contains a logarithmic term in the density n:16

(n, T ) = 0 (T ) + n1 (T ) + n2 ln n 20 (T ) + n2 2 (T ) + ... (24)

Similar expressions result for the other transport coefficients.

Dense Fluids
With the development of more powerful computers extensive computer simulations
for dense hard sphere fluids have been performed in the last twenty years.17 In this
way, and from theoretical considerations, it has been possible to identify two especially
relevant collision sequences going on in a dense hard fluid. These are the cage diffusion
collision sequence in Fig. 2 and the vortex diffusion collision sequence in Fig. 3. In
cage diffusion the particle finds itself trapped in a cage made up by the surrounding
particles and it requires several collisions with its neighbours until it finds a hole to
sneek out. It leads to a significant change in the volume dependence of the viscosity
of a dense fluid as shown in Fig. 418 . Its importance was discussed by de Schepper
and Cohen in 1980.19 In vortex diffusion an energetic particle creates a vortex like a
macroscopic sphere moving through liquid, see Fig. 3. It was discovered around 1968
by Alder and Wainwright20 and affects the longtime behaviour at about twentyfife
mean free times. The above mentioned collision sequences constitute corrections to the
Boltzmann equation in dense fluids and provide an understanding of the behaviour of
dense fluids through collisions.

Acknowledgement
The author would like to thank Prof. H.A. Posch for critical reading of the manuscript.

16
For a review see e.g. E.G.D. Cohen: Kinetic Theory: Understanding Nature through Collisions,
in Thermodynamics and Statistical Physics Teaching Modern Physics, ed. M.G. Velarde and F.
Cuadros, World Scientific, 1995.
17
W.W. Wood, in Fundamental Problems in Statistical Mechanics, ed. E.G.D. Cohen, NorthHolland
Pub., Amsterdam 1975.
18
I.M. de Schepper, A.F.E.M. Haffmans and J.J. van Loef: J. Stat. Phys. 57 (1989) 631.
19
I.M. de Schepper and E.G.D. Cohen: Phys. Rev. A 22 (1980) 287; J. Stat. Phys. 27 (1982) 223.
20
B.J. Alder and T.E. Wainright: Phys. Rev. A1 (1979) 18.
 11

Figure 1: Examples of correlated four particle ring collision sequences.

Figure 2: Example of a correlated cage diffusion collision sequence.

12

Figure 3: Example of a correlated vortex diffusion collision sequence. (a) Two vortex
rings. (b) Left vortex ring in more detail.

1.0

0.5

0.0
1.0 1.5
V*

Figure 4: The reduced hardsphere fluidity = 0 , which corresponds to the inverse

of the viscosity multiplied with the low density limit 0 of the viscosity, is plotted as
a function of the reduced volume V which is the volume of the liquid divided by the
volume of the hard spheres. The dashed curve is according to the classical Enskog
theory and the solid curve according to the mode coupling theory which takes also
cage diffusion into account. The squares and crosses are from computer simulation
experiments.