Chapter 3

The Scaling Hypothesis

Previously, we found that singular behaviour in the vicinity of a second order critical
point was characterised by a set of critical exponents {α, β, γ, δ, · · ·}. These power law
dependencies of thermodynamic quantities are a symptom of scaling behaviour. Mean-
field estimates of the critical exponents were found to be unreliable due to fluctuations.
However, since the various thermodynamic quantities are related, these exponents can not
be completely independent of each other. The aim of this chapter is to employ scaling
ideas to uncover relationships between them.

3.1 Homogeneity
The non-analytic structure of the Ginzburg-Landau model was found to be a coexistence
line for t < 0 and h = 0 that terminates at the critical point h = t = 0. Thermodynamic
quantities Q(t, h) in the vicinity of the critical point are characterised by various expo-
nents. In particular, within the saddle-point approximation we found that the free energy
density was given by
! " # 2
βF t 2 4 −t /u, h = 0, t < 0,
f≡ = minm m + um − h · m ∼ 4/3 1/2 (3.1)
V 2 −h /u h $= 0, t = 0.

In fact, the free energy can be described by a single homogeneous function1 in t and h

f (t, h) = t2 gf (h/t∆ ) (3.2)

where ∆ is known as the “Gap exponent”. Comparison with Eq. (3.1) shows that, if we
set ∆ = 3/2, the correct asymptotic behaviour of f is obtained,

1 t2
lim gf (x) ∼ − , f (t, h = 0) ∼ − ,
x→0 u u
4/3
$ %4/3
x 2 h
lim gf (x) ∼ , f (t = 0, h) ∼ t ∼ h4/3 .
x→∞ u1/3 t∆
1
A function f (x) is said to be homogeneous of degree k if it satisfies the relation f (x) = bk f (bx).

Phase Transitions and Collective Phenomena

38 CHAPTER 3. THE SCALING HYPOTHESIS

The assumption of homogeneity is that, on going beyond the saddle-point approxi-

mation, the singular form of the free energy (and of any other thermodynamic quantity)
retains a homogeneous form
$ %
2−α h
fsing. (t, h) = t gf (3.3)
t∆

where the actual exponents α and ∆ depend on the critical point being considered.
Heat Capacity: For example, the dependence on t is chosen to reproduce the heat
capacity singularity at h = 0. The singular part of the energy is obtained from
∂f
Esing. ∼ ∼ (2 − α)t1−α gf (h/t∆ ) − ∆ht1−α−∆ gf$ (h/t∆ ) ≡ t1−α gE (h/t∆ ),
∂t
where the prime denotes the derivative of the function with respect to the argument. Thus,
the derivative of one homogeneous function is another. Similarly, the second derivative
takes the form
∂2f
Csing. ∼ − ∼ t−α gC (h/t∆ ),
∂t2
reproducing the scaling Csing. ∼ t−α as h → 0.2
Magnetisation: Similarly the magnetisation is obtained from Eq. (3.3) using the
expression
∂f
m(t, h) ∼ ∼ t2−α−∆ gm (h/t∆ ).
∂h
In the limit x → 0, gm (x) is a constant, and m(t, h = 0) ∼ t2−α−∆ (i.e. β = 2 − α − ∆).
On the other hand, if x → ∞, gm (x) ∼ xp , and m(t = 0, h) ∼ t2−α−∆ (h/t∆ )p . Since this
limit is independent of t, we must have p∆ = 2 − α − ∆. Hence m(t = 0, h) ∼ h(2−α−∆)/∆
(i.e. δ = ∆/(2 − α − ∆) = ∆/β).
Susceptibility: Finally, calculating the susceptibility we obtain
∂m
χ(t, h) ∼ ∼ t2−α−2∆ gχ (h/t∆ ) ⇒ χ(t, h = 0) ∼ t2−α−2∆ ⇒ γ = 2∆ − 2 + α.
∂h
Thus the consequences of homogeneity are:

• The singular parts of all critical quantities, Q(t, h) are homogeneous, with the same
exponents above and below the transition.

• Because of the interconnections via thermodynamic derivatives, the same gap expo-
nent, ∆ occurs for all such quantities.

• All critical exponents can be obtained from only two independent ones, e.g. α, ∆.
2
It may appear that we have the freedom to postulate a more general form, C± = t−α± g± (h/t∆ ±)
with different functions for t > 0 and t < 0 that match at t = 0. However, this can be ruled out by the
condition that the free energy is analytic everywhere except on the coexistence line h = 0 and t < 0.

Phase Transitions and Collective Phenomena

3.2. HYPERSCALING AND THE CORRELATION LENGTH 39

• As a result of above, one obtains a number of exponent identities:

α + 2β + γ = 2 (Rushbrooke$ s Identity)

δ − 1 = γ/β (Widom$ s Identity)

These identities can be checked against the following table of critical exponents. The first
three rows are based on a number of theoretical estimates in d = 3; the last row comes
from an exact solution in d = 2. The exponent identities are approximately consistent
with these values, as well as with all reliable experimental data.

α β γ δ ν η
d=3 n=1 Ising 0.12 0.31 1.25 5 0.64 0.05
n=2 XY-spin 0.00 0.33 1.33 5 0.66 0.00
n=3 Heisenberg −0.14 0.35 1.4 5 0.7 0.04
d=2 n=1 Ising 0 1/8 7/4 15 1 1/4

3.2 Hyperscaling and the Correlation Length

The homogeneity assumption relates to the free energy and quantities derived from it.
It says nothing about correlation functions. An important property of a critical point
is the divergence of the correlation length, which is responsible for (and can be deduced
from) the divergence of response functions. In order to obtain an identity involving the
exponent ν describing the divergence of the correlation length, we replace the homogeneity
assumption for the free energy with the following two conditions:

1. The correlation length has a homogeneous form,

$ %
−ν h
ξ(t, h) ∼ t gξ
t∆

For t = 0, ξ diverges as h−νh with νh = ν/∆.

2. Close to criticality, the correlation length ξ is the most important length scale, and
is solely responsible for singular contributions to thermodynamic quantities.

The second condition determines the singular part of the free energy. Since ln Z(t, h)
is dimensionless and extensive (i.e. scales in proportion with the volume Ld ), it must take
the form
$ %d $ %d
L L
ln Z = × gs + × ga ,
ξ a

Phase Transitions and Collective Phenomena

40 CHAPTER 3. THE SCALING HYPOTHESIS

Spins correlated
in each box

L ξ

Figure 3.1: Within each cell of size ξ spins are correlated.

where gs and ga are non-singular functions of dimensionless parameters (a is an appropri-

ate microscopic length). The singular part of the free energy comes from the first term
and behaves as
ln Z
fsing. (t, h) ∼ d ∼ ξ −d ∼ tdν gf (t/h∆ ). (3.4)
L
A simple interpretation of this result is obtained by dividing the system into units of
the size of the correlation length (Fig. 3.2). Each unit is then regarded as an independent
random variable, contributing a constant factor to the critical free energy. The number
of units grows as (L/ξ)d .
The consequences of Eq. (3.4) are:
• Homogeneity of fsing. emerges naturally.
• We obtain the additional exponent relation

2 − α = dν (Josephson$ s Idenitity)

Identities obtained from the generalised homogeneity assumption involve the space
dimension d, and are known as hyperscaling relations. The relation between α and ν
is consistent with the exponents in the table above. However, it does not agree with the
mean-field values, α = 0 and ν = 1/2, which are valid for d > 4. Any theory of critical
behaviour must therefore account for the validity of this relation in low dimensions, and
its breakdown in d > 4.

3.3 Correlation Functions and Self-Similarity

So far we have not accounted for the exponent η which describes the decay of correlation
functions at criticality. Exactly at the critical point the correlation length is infinite,
and there is no other length scale to cut-off the decay of correlation functions. Thus all
correlations decay as a power of the separation. As discussed in the previous chapter, the
magnetisation falls off as
1
Gc (x) ≡ )m(x) · m(0)* − )m2 * ∼
|x|d−2+η

Phase Transitions and Collective Phenomena

3.3. CORRELATION FUNCTIONS AND SELF-SIMILARITY 41

where η was deduced from the form factor.

Away from criticality, the power laws are cut-off for distances |x| + ξ. As the response
functions can be obtained from integrating the connected correlation functions, there are
additional exponent identities such as Fisher’s identity
& & ξ
d dd x
χ ∼ d x Gc (x) ∼ ∼ ξ 2−η ∼ t−ν(2−η) =⇒ γ = (2 − η)ν
|x|d−2+η

Therefore, two independent exponents are sufficient to describe all singular critical be-
haviour.
An important consequence of these scaling ideas is that the critical system has an
additional dilation symmetry. Under a change of scale, the critical correlation functions
behave as

Gcritical (λx) = λp Gcritical (x)

This implies a scale invariance or self-similarity: If a snapshot of the critical system

is enlarged by a factor of λ, apart from a change of contrast (λp ), the resulting snapshot
is statistically similar to the original. Such statistical self-similarity is the hallmark of
fractal geometry. The Ginzburg-Landau functional was constructed on the basis of local
symmetries such as rotational invariance. If we could add to the list of constraints the
requirement of dilation symmetry, the resulting probability would indeed describe the
critical point. Unfortunately, it is not in general possible to see directly how such a
requirement constrains the effective Hamiltonian.3 We shall instead prescribe a less direct
route by following the effects of the dilation operation on the effective energy; a procedure
known as the renormalisation group.

3
One notable exception is in d = 2, where dilation symmetry implies conformal symmetry.

Phase Transitions and Collective Phenomena

42 CHAPTER 3. THE SCALING HYPOTHESIS

Phase Transitions and Collective Phenomena