and
0 < d
0
< d
1
. We are thus faced with a system of hard-core segments
of diameter
d
0
, that interact only at distances smaller than
d
1
; van Hove’s
remaining assumption about the interaction is that
U is a continuous,
bounded below function.
The way he proves this result is, as he himself says, by reducing the
problem to an eigenvalue problem. He is able to write the partition func-
tion of the system in terms of a transfer operator, whose largest eigenvalue
gives the only relevant contribution to the free energy in the thermo-
dynamic limit. After transforming the operator into a more useful form,
van Hove resorts to the theory of Fredholm integral operators and other
theorems of functional analysis to show that this eigenvalue is an analytic
function of temperature and, consequently, that the system can not have
phase transitions, understood rigorously as nonanalyticities of the free
energy. The mathematical basis of this result will be made clear by the
theorem we will present later in this article, and therefore we do not need
to go into further detail at this point (other than enthusiastically referring
the interested reader to the original paper
(8)
). For the time being, suffice it
to say that the basic idea is an extension of the well-known Perron–Frobenius
theorem for non-negative matrices;
(11, 12)
we will come back to this theorem
when discussing our first example in the next section.
The key point we want to make here relates to the
hypotheses needed
to prove van Hove’s theorem, i.e., to the class of systems to which it applies.
Let us consider them separately:
Homogeneity.
First of all, the system has to be perfectly homoge-
neous, made up of
identical particles. This automatically excludes any
inhomogeneous model, where inhomogeneous means either aperiodic or
disordered. Periodic systems could in principle be included in the frame of
van Hove’s theorem by analyzing the transfer operator for a unit cell. This
is a very strong restriction, and it should be very clear that any degree of
inhomogeneity in the system makes it impossible to exclude phase transi-
tions on the ground of van Hove’s result.
Do'stlaringiz bilan baham: