MAMA/303, NST3AS/303, MAAS/303

MAT3
MATHEMATICAL TRIPOS Part III

Friday 7 June 2024 9:00 am to 11:00 am

PAPER 303

STATISTICAL FIELD THEORY

Before you begin please read these instructions carefully

Candidates have TWO HOURS to complete the written examination.

Attempt no more than TWO questions.

There are THREE questions in total.
The questions carry equal weight.

STATIONERY REQUIREMENTS SPECIAL REQUIREMENTS

Cover sheet None
Treasury tag
Script paper
Rough paper

You may not start to read the questions

printed on the subsequent pages until
instructed to do so by the Invigilator.
 2

1
(a) Consider a theory which has an effective free energy

f (T, m) = a4 (T ) m4 + a6 m6 − B m ,

in the mean field approximation, where m is the magnetisation, a6 > 0, a4 (T ) varies from
positive to negative as the temperature T is lowered and a4 (T ) ∼ (T − Tc ) close to T = Tc .
(i) For given a4 , a6 , B and T , how can the equilibrium value of m be
determined from f ?

(ii) If B = 0 show that there is a phase transition at T = Tc . Is this

a continuous or discontinuous phase transition? Compute the critical
exponents α, β, γ and δ at this phase transition. [You should compute γ for
both T → Tc+ and T → Tc− .] [Hint: Recall that close to the critical point
c ∼ |T − Tc |−α , m ∼ (Tc − T )β for T < Tc , m ∼ |B|1/δ and χ ∼ |T − Tc |−γ ,
where c is the heat capacity and χ is the magnetic susceptibility.]

(iii) Does the system possess another phase transition? If so, is it continuous
or discontinuous? Justify your answers.

(b) Consider a model defined on a square lattice in d dimensions. The N lattice

sites are labeled by i and the spin variable σ i = (cos θi , sin θi ) at site i is a 2-dimensional
vector that can point in p different directions, where θi = 2πni /p and ni = 0, 1, . . . p − 1.
The energy is given by X
E = −J σi · σj ,
⟨ij⟩

where J > 0 is a constant and ⟨ij⟩ means that the sum is over nearest neighbour pairs.
(i) For the model above with p = 4, use the mean field approach to show that
the effective free energy per unit site can be approximated by
F h
i
f= = A m2 − T ln C + D cosh βJqm) ,
N
where m = |m| is the magnitude of the magnetisation, β = 1/T , q is the
number of nearest neighbours of each site, and you should determine A, C
and D. [Hint: Use the approach where σ i is written as σ i = m + δσ i and
terms of order (δσi )2 and higher are neglected.]

(ii) From the expression for f in part (b)(i), find an implicit equation for the
equilibrium value of m. [You should keep p = 4 here and in all subsequent
parts of this question unless otherwise stated.]

(iii) Now, treat m2 as small and expand f as a power series in m2 , neglecting

m6 and higher terms. For what value of T is there a phase transition?

(iv) What would you expect the lower critical dimension for this model to be?
What happens to the model in the limit p → ∞ and what would you
expect the lower critical dimension to be in that case? Briefly justify your
answers.

Part III, Paper 303

 3

2
Consider a theory involving a real scalar field ϕ in d dimensions with a free energy
of the form Z h i
F [ϕ] = dd x 21 ∇ϕ · ∇ϕ + 21 µ20 ϕ2 + . . . . (∗)

(a) Describe the three steps of the renormalisation group procedure (in momentum
space) for such a free energy and explain how they result in a flow of the parameters in
the free energy. You should denote the original cutoff Λ and the new cutoff Λ/ζ.
(b) Calculate the naive (engineering) dimensions of ϕ and µ20 .
(c) Why can the scaling dimension ∆ϕ of the field ϕ differ from the engineering
dimension?
(d) Now suppose
R d that 2m
the free energy contains quadratic terms in ϕ, as in equation
(∗), along with ∼ d x α ∇ ϕ2n , where n > 1 and m ⩾ 0 are integers. Compute the
naive (engineering) dimension of the coupling α. Give conditions on d in terms of n and
m for the coupling to be relevant, marginal and irrelevant.
(e) Now instead suppose that the free energy contains quadratic terms in ϕ, as in
equation (∗), along with Z h i
∼ dd x g0 ϕ4 + γ0 ϕ5 + λ0 ϕ6 .

(i) Draw a Feynman diagram that represents a contribution that gives rise to a
non-zero anomalous dimension of ϕ at order g02 . Justify you answer. [You
are not required to compute the correction, but you should explain how the
diagram you have drawn gives a non-zero anomalous dimension.]

(ii) Ignoring other interactions, does the λ0 ϕ6 term alone give a correction to the
flow of the coupling γ(ζ). If so, draw a Feynman diagram that represents such
a contribution; if not, explain why.

(iii) Draw a Feynman diagram representing the correction to the flow of the
coupling g(ζ) from the λ0 ϕ6 term at order λ0 . Calculate this contribu-
tion. [Here and in all subsequent parts of this question you may assume
that ⟨ϕ+ ϕ+ ⟩ = (2π)d δ (d) (k + k′ )G0 (k), where G0 (k) = 1/(k 2 + µ20 ), for
k k′ +
appropriately defined ϕ+ and ⟨. . .⟩+ , you may use Wick’s theorem without
proof, and you may leave your final answer in integral form. You may ignore
other corrections, including those from the rescaling of the field.]

(iv) Calculate the contribution to g(ζ) at order λ20 represented by the following
Feynman diagram.

(v) Draw Feynman diagrams representing all other corrections to g(ζ) from the
λ0 ϕ6 term at order λ20 . [You do not need to calculate these contributions.]

Part III, Paper 303 [TURN OVER]

 4

3
Consider an O(N ) model involving an N -component real field ϕ(x) in d dimensions
with free energy,
Z h X i
F [ϕ] = dd x 21 (∂i ϕa )(∂i ϕa ) + 21 µ20,a ϕ2a + g0 (ϕ · ϕ)2 ,
a

where g0 > 0, ϕ · ϕ = ϕa ϕa , repeated indices are summed over, i = 1, 2, . . . , d and

a = 1, 2, . . . , N .
(a) Suppose that µ20,a = µ20 for all a and µ20 ∼ (T − Tc ).

(i) What is the symmetry of F ? What is the symmetry of the mean field
theory ground state when µ20 > 0 and when µ20 < 0?

(ii) Draw Feynman diagrams representing the leading order corrections to µ20
from the g0 term. Indicate which diagrams (if any) are proportional to N .

(iii) Draw Feynman diagrams representing the leading order corrections to g0

from the g0 term. Indicate which diagrams (if any) are proportional to N .

(iv) The leading order corrections to the couplings are

" Z Λ #
d dq 1
µ2 (ζ) = ζ A µ20 + 4(N + 2)g0 d 2 2 ,
Λ/ζ (2π) q + µ0
" Z Λ #
ddq 1
g(ζ) = ζ B g0 − 4(N + 8)g02 d 2 2 2 . (∗)
Λ/ζ (2π) (q + µ0 )

Determine the constants A and B.

2
(v) Derive the β functions, dµ dg
ds and ds , where s = ln ζ. [You may use Ωd−1 to
denote the area of the unit sphere S d−1 .]

(vi) Consider d = 4−ϵ and define g̃ = Λ−ϵ g. Working to leading order in ϵ, find
the fixed points and calculate the critical exponent ν at each fixed point.

(b) Now consider N = 2, and allow µ20,1 and µ20,2 to differ.

(i) Assuming mean field theory, describe the phase diagram in the µ20,1 , µ20,2
plane. Identify points/lines where phase transitions occur and describe the
nature of any phase transitions.

(ii) What are the equations for µ21 (ζ) and µ22 (ζ) in this theory analogous to the
equation for µ2 (ζ) in (∗)?

(iii) Is a term ∼ ϕ1 ϕ2 generated by interactions? Briefly justify your answer.

END OF PAPER

Part III, Paper 303