CFT PDF

Conformal Field Theory
A.N. Schellekens
[Word cloud by www.worldle.net]

Last modified 26 May 2016
1
Contents
1 Classical Conformal Invariance 7
1.1 Symmetries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.1 General coordinate invariance . . . . . . . . . . . . . . . . . . . . . 7
1.1.2 Weyl invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.1.3 Conformal invariance . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.2 Conformal transformations in d dimensions . . . . . . . . . . . . . . . . . 9
1.3 The conformal group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.4 Relation to stereographic projections . . . . . . . . . . . . . . . . . . . . . 12
1.5 The conserved current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 An example: the free boson . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.7 The conformal algebra in two dimensions . . . . . . . . . . . . . . . . . . 14
1.8 Complexification and Wick rotation . . . . . . . . . . . . . . . . . . . . . 16
1.9 The global subgroup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
1.10 Tensors in complex coordinates . . . . . . . . . . . . . . . . . . . . . . . . 18
1.11 Conformal fields in two dimensions . . . . . . . . . . . . . . . . . . . . . . 19
1.12 Relation to string theory . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
1.13 Free bosons in complex coordinates . . . . . . . . . . . . . . . . . . . . . . 20
2 Quantum Conformal Invariance 21

2.1 Radial quantization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.2 Radial ordering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 The generator of conformal transformations . . . . . . . . . . . . . . . . . 24
2.4 Quantization of the free boson . . . . . . . . . . . . . . . . . . . . . . . . 26
2.5 The free boson propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.6 The normally ordered energy momentum tensor . . . . . . . . . . . . . . . 30
2.7 Operator products for free bosons . . . . . . . . . . . . . . . . . . . . . . 30
2.8 The Virasoro algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3 Virasoro Representation Theory 32

3.1 Unitarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.2 Highest weight representations . . . . . . . . . . . . . . . . . . . . . . . . 33
3.3 The vacuum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.4 Positivity of c and h . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3.5 States and conformal fields . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.6 Descendant fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.7 The Kac determinant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
3.8 The Kac determinant at level 2 . . . . . . . . . . . . . . . . . . . . . . . . 38
3.9 The Kac determinant at level 3 and 4 . . . . . . . . . . . . . . . . . . . . 40
3.10 The discrete series of minimal unitary Virasoro models . . . . . . . . . . . 41
2
4 Correlation Functions 42
4.1 Correlation functions on the Riemann sphere . . . . . . . . . . . . . . . . 43
4.2 Two-point functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
4.3 Three-point functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
4.4 Four-point functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.5 Conformal Ward identities . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.6 Correlators of descendants . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.7 Null state decoupling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.8 Operator products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.9 Duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
5 Conformal Field Theory on a Torus 52

5.1 Parametrization of the torus . . . . . . . . . . . . . . . . . . . . . . . . . 52
5.2 The partition function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
5.3 The cylinder versus the Riemann sphere . . . . . . . . . . . . . . . . . . . 55
5.4 Virasoro characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.5 Modular invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
5.6 Modular transformations of the characters . . . . . . . . . . . . . . . . . . 59
5.7 Conditions for modular invariance . . . . . . . . . . . . . . . . . . . . . . 59
5.8 The diagonal invariant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.9 Integration over moduli . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
5.10 Operator products and locality . . . . . . . . . . . . . . . . . . . . . . . . 60
5.11 Fusion rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5.12 The Verlinde formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
5.13 Higher genus partition functions . . . . . . . . . . . . . . . . . . . . . . . 63
6 Extensions of the Virasoro Algebra 64

6.1 Rational conformal field theories . . . . . . . . . . . . . . . . . . . . . . . 64
6.2 Currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.3 Fermionic currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
6.4 Mode expansions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
6.5 Integer and half-integer modes . . . . . . . . . . . . . . . . . . . . . . . . 66
6.6 Types of chiral algebra extensions . . . . . . . . . . . . . . . . . . . . . . 67
6.7 Properties of extended Virasoro representations . . . . . . . . . . . . . . . 67
6.8 Charge conjugation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
6.9 Characters and modular transformations . . . . . . . . . . . . . . . . . . . 69
6.10 Virasoro tensor products . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
6.11 Extensions and off-diagonal partition functions . . . . . . . . . . . . . . . 71
6.12 The new S and T matrices . . . . . . . . . . . . . . . . . . . . . . . . . . 72
6.13 Extensions and automorphisms . . . . . . . . . . . . . . . . . . . . . . . . 73
6.14 Simple currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3
7 Free Fermions 74
7.1 The propagator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74
7.2 Energy-Momentum tensor and central charge . . . . . . . . . . . . . . . . 74
7.3 Mode expansion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
7.4 The spin field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
7.5 Free fermion characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
7.5.1 Neveu-Schwarz states . . . . . . . . . . . . . . . . . . . . . . . . . . 76
7.5.2 Neveu-Schwarz characters . . . . . . . . . . . . . . . . . . . . . . . 78
7.5.3 Ramond states . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
7.5.4 The Ramond ground state . . . . . . . . . . . . . . . . . . . . . . . 80
7.5.5 Ramond characters . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
7.7 Theta-functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
7.8 Modular transformations . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
7.9 The modular invariant partition function . . . . . . . . . . . . . . . . . . 83
7.10 Ising characters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
7.11 The matrix S and the fusion rules . . . . . . . . . . . . . . . . . . . . . . 84
7.12 Multi-fermion systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
8 Free Boson Partition Functions 84

8.1 The spectrum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
8.2 The characters and the diagonal invariant . . . . . . . . . . . . . . . . . . 85
8.3 Chiral bosons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
8.4 Further extensions of the chiral algebra . . . . . . . . . . . . . . . . . . . 87
8.5 Representations of the extended algebra . . . . . . . . . . . . . . . . . . . 88
8.6 The matrix S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.7 Relation with circle compactification . . . . . . . . . . . . . . . . . . . . . 91
8.8 R → 2/R duality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8.9 Rationality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
8.10 Theories with more than one free boson . . . . . . . . . . . . . . . . . . . 92
8.11 Orbifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
8.12 Orbifolds as singular manifolds . . . . . . . . . . . . . . . . . . . . . . . . 93
8.13 Orbifolds in conformal field theory . . . . . . . . . . . . . . . . . . . . . . 93
8.14 Orbifolds of the circle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8.14.1 The twist fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
8.15 Projecting on the invariant states . . . . . . . . . . . . . . . . . . . . . . . 95
8.17 The geometric description . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
8.18 The c = 1 conformal field theories . . . . . . . . . . . . . . . . . . . . . . 98
8.19 Moduli and marginal deformations . . . . . . . . . . . . . . . . . . . . . . 99
4
9 Kac-Moody Algebras 100
9.1 Spin one operator products . . . . . . . . . . . . . . . . . . . . . . . . . . 100
9.2 Intermezzo: some Lie algebra facts . . . . . . . . . . . . . . . . . . . . . . 100
9.3 The central term . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
9.4 Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
9.5 Twisted and untwisted affine algebras . . . . . . . . . . . . . . . . . . . . 103
9.6 Primary fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
9.7 The Sugawara tensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104
9.8 Highest weight representations . . . . . . . . . . . . . . . . . . . . . . . . 105
9.9 U (1) theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
9.10 The SU (2) Kac-Moody algebra . . . . . . . . . . . . . . . . . . . . . . . . 106
9.11 SU (2) at level 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
9.12 Generalization to other Kac-Moody algebras . . . . . . . . . . . . . . . . . 108
9.13 The Frenkel-Kac construction . . . . . . . . . . . . . . . . . . . . . . . . . 111
9.14 The WZW-model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
9.15 Modular transformation properties . . . . . . . . . . . . . . . . . . . . . . 112
9.16 Modular invariant partition functions for SU (2) . . . . . . . . . . . . . . . 113
9.17 Fusion rules and simple currents . . . . . . . . . . . . . . . . . . . . . . . 113
9.18 Modular invariant partition functions for other Kac-Moody algebras . . . 114
9.19 Coset conformal field theories . . . . . . . . . . . . . . . . . . . . . . . . . 114
9.20 The minimal discrete series as a coset theory . . . . . . . . . . . . . . . . 116
9.21 Field Identification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
9.22 Other coset models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10 Superconformal Algebras 118

10.1 The N = 1 algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
10.2 The N = 2 algebra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
5
Preface
Conformal field theory has been an important tool in theoretical physics during the
last decades. Its origins can be traced back on the one hand to statistical mechanics, and
on the other hand to string theory. Historically the most important impetus came from
statistical mechanics, where it described and classified critical phenomena. Mainly after
1984 the subject went through a period of rapid development because of its importance
for string theory. In addition there has been important input from mathematics, in
particular through the work of Kac and collaborators. One can distinguish yet another
separate origin of some ideas, namely from work on rigorous approaches to quantum field
theory.
At present the subject still continues to develop, and it is still important in all the fields
mentioned, plus a few additional branches of mathematics. These lectures are mainly on
two-dimensional CFT. Recently conformal field theory appeared in yet another context,
namely the “AdS/CFT-correspondence”, where also higher dimensional (super)conformal
invariance is relevant.
I tried to include references to most relevant papers, but the emphasis was on papers
I consider to be worth reading even today, and not on papers that are mainly of historical
interest. A more detailed account of the history may be found in [24], which was used
extensively for the preparation of these notes. In addition to the latter review, other
useful general references include the one by J. Cardy from the same proceedings [8].
Other sources I used are [36] and [49]. Some useful results can be found in books on
string theory, for example [31] and [5]. Standard reviews on Kac-Moody algebras are [29]
and [20]. Finally I should mention as a general reference the paper by Belavin, Polyakov
and Zamolodchikov [44], which is the starting point of many recent developments.
These notes were originally based on lectures given at the first “Saalburg” school for
graduate students, ”Grundlagen und neue Methoden der Theoretische Physik”, Saalburg,
Germany, 3-16 Sept. 1995, and at the Universidad Autonoma, Madrid, October-December
1995. I am grateful to the students of those classes for pointing out many errors and
misprints. The current version is an update prepared for the 21th “Saalburg” school,
“Foundations and New Methods in Theoretical Physics”, 31 Aug. - 11 Sept. 2015 in
Wolfersdorf, Thüringen. An earlier version was published in [47].
6
1 Classical Conformal Invariance
In this section we study classical field theories in an arbitrary number of dimensions. In
this space we have a metric gµν . Furthermore we define g = |det gµν |. We will work in
flat space, which means that the coordinates can be chosen in such a way that gµν = ηµν ,
where the latter has the form diag (−1, . . . , −1, +1, . . . + 1). The number of eigenvalues
−1 or +1 is q and p respectively. Our convention is to use −1 in the time direction. Hence
in practice q is either 0 (Euclidean space) or 1 (Minkowski space).
1.1 Symmetries
1.1.1 General coordinate invariance
Classical field theories may have a variety of symmetries. One symmetry that we will
assume them to have is general coordinate invariance. Using the action principle this can
be used to show that the energy momentum tensor is conserved. In general, this tensor is
defined in terms of the variation of the action S under changes of the space-time metric
gµν → gµν + δgµν . (1.1)
Then the definition of the energy momentum tensor is
√
Z
δS = 2 dd x g T µν δgµν .
1
(1.2)
If the theory is invariant under general coordinate transformations one can show that
(T νµ );ν = 0 . (1.3)
Here (as usual in general relativity) “;ν” denotes a covariant derivative. In flat coordinates
the condition reads ∂ν T νµ = 0.
1.1.2 Weyl invariance

We are not interested in general coordinate invariance, but in a different symmetry which
can also be formulated in terms of the metric and the energy momentum tensor. This
symmetry is called Weyl invariance. The transformation we consider is
gµν (x) → Ω(x)gµν (x) , (1.4)
or in infinitesimal form
gµν → gµν (x) + ω(x)gµν (x) . (1.5)
The condition for invariance of an action under such a symmetry can also be phrased in
terms of the energy momentum tensor. Substituting δgµν = ω(x)gµν (x) into (1.2) we find
√
Z
δS = 2 dd x g T µµ ω(x) .
1
(1.6)
Since this must be true for arbitrary functions ω we conclude that the condition for Weyl
invariance is
T µµ = 0 . (1.7)
7
1.1.3 Conformal invariance
A conformal transformation can now be defined as a coordinate transformation which acts
on the metric as a Weyl transformation. Consider a general coordinate transformation
x → x0 , such that xµ = f µ (x0ν ). This has the following effect on the metric
0 ∂f ρ ∂f σ
gµν (x) → gµν (x0 ) = gρσ (f (x0 )) . (1.8)
∂x0µ ∂x0ν
We are going to require that the left hand side is proportional to gµν . Rotations and
translations do not change the metric at all, and hence preserve all inner products
v · w ≡ v µ gµν wν . They are thus part of the group of conformal transformations. A co-
ordinate transformation satisfying Eqn (1.4) preserves all angles, √v·w
v 2 w2
(hence the name
‘conformal’). Later in this chapter we will determine all such transformations.
If a field theory has a conserved, traceless energy momentum tensor, it is invariant
both under general coordinate transformations and Weyl transformations. Suppose the
action has the form Z
S = dd xL(∂x , gµν (x), φ(x)) . (1.9)
Here φ denotes generically any field that might appear, except for the metric which we
have indicated separately since it plays a special rôle. We have also explicitly indicated
space-time derivatives. General coordinate invariance implies that
Z
S = S ≡ dd x0 L(∂x0 , gµν
0 0
(x0 ), φ0 (x0 )) (1.10)
0
Here gµν is as defined above, and the transformations of a field φ depends on its spin. If
it is a tensor of rank n one has
∂f ν1 ∂f νn
φ0µ1 ,...,µn (x0 ) = 0µ
. . . 0µn φν1 ,...,νn (f (x0 )) (1.11)
∂x 1 ∂x
In particular, for a scalar function φ(x) we find φ0 (x0 ) = φ(f (x0 ) and for the derivative of
a scalar function we get
∂ ∂ 0 0 ∂ 0 ∂f ν ∂
φ(x) → φ (x ) = φ(f (x )) = φ(f (x0 )) , (1.12)
∂xµ ∂x0µ ∂x0µ ∂x0µ ∂f ν
i.e. it transforms like a vector (note, however, that nth order ordinary derivatives do not
transform like a tensor of rank n; this is only true if one uses covariant derivatives). If the
coordinate transformation x → x0 is of the special type (1.4) we can use Weyl invariance
of the action to change the metric back into its original form. Then we have
Z Z
S = S ≡ d x L(∂x0 , gµν (f (x )), φ (x )) = dd x0 L(∂x0 , gµν (x), φ0 (x0 ))
00 d 0 0 0 0
(1.13)
This is the conformal symmetry of the action. Note that the metric now remains un-
changed if we start with a flat space metric gµν = ηµν . This means that we can define the
8
conformal transformation for theories in flat space that are not coupled to gravity. We
may then forget about general coordinate invariance and start with an action in which
no “dynamical” metric appears.
The statement of conformal invariance is then that the action of such a theory is
unchanged if we integrate the same Lagrangian (or other physical scalar) expressed in
terms of the new fields φ0 (x0 ) over the new coordinates x0 . If the relation between the old
coordinates x and the old ones x0 is x = f (x0 ), then the new fields are related to the old
fields as in Eqn. (1.11).
The restriction to flat space is not really a restriction if we are in two dimensions.
Then a general metric is given by three functions, g11 (x), g22 (x) and g12 (x) = g21 (x). A
general coordinate transformation allows us to change this using two functions, f 1 (x) and
f 2 (x), and we can – generically – use this freedom to set g12 (x) = 0 and g11 (x) = ±g22 (x)
(depending on the signature of the metric), so that the metric has the form g(x)ηµν . This
is called conformal gauge. Then, using a Weyl transformation, we can remove the function
g(x) and bring the metric to the form ηµν . In more than two dimensions we do not have
enough freedom to do this, and then the assumption made here is really a restriction to
non-gravitational theories in flat space.
On a given two-dimensional manifold the conformal gauge choice can be made locally,
but usually not globally. This means that we will be able to use conformal field theory
in some coordinate patch, but that additional data may be needed to describe the theory
globally.
Fields that transform like (1.11) under conformal transformations are called conformal
fields, or also primary fields.
1.2 Conformal transformations in d dimensions

In general the right hand side of (1.8) is of course not proportional to the original metric
gµν . To study when it is, consider the infinitesimal transformation x0µ = xµ + µ (x), or
rather its inverse, xµ = x0µ − µ (x0 ) + O(2 ). Then
∂xρ
= δµρ − ∂µ ρ , (1.14)
∂x0µ
and
δgµν = −∂µ ν − ∂ν µ (1.15)
This must be equal to ωgµν . Taking the trace we see then that ω = − d2 ∂ · (with
∂ · ≡ ∂ µ µ ), so that we get the following equation for
2
∂µ ν + ∂ν µ = gµν ∂ · (1.16)
d
For d = 1 this is satisfied for any . Let us now analyze the solutions to this condition for
d > 1. As a first step, we contract both sides with ∂ µ ∂ ν . This yields
1
(1 − ) ∂·=0 → ∂·=0 . (1.17)
d
9
Next we contract (1.16) with ∂ρ ∂ ν . This yields
2
∂ρ µ + (1 − )∂ρ ∂µ ∂ · = 0 . (1.18)
d
To this we add the same equation with ρ and µ interchanged, we use (1.16) once more
and finally (1.17). The result is
2
(1 − )∂ρ ∂µ ∂ · = 0 . (1.19)
d
We conclude that ∂ρ ∂µ ∂ · = 0 if d > 2. The third (and last) step is to take the uncon-
tracted derivative ∂ρ ∂σ of (1.16). Define ∆ρσµν ≡ ∂ρ ∂σ ∂µ ν . This function is manifestly
symmetric in the first three indices. Furthermore, by acting with ∂ρ ∂σ on (1.16) and using
(1.19) we find (for d > 2)
∆ρσµν = −∆ρσνµ . (1.20)
It is now easy to show that a tensor with these symmetries must vanish:
∆ρσµν = ∆ρµσν = −∆ρµνσ = −∆ρνµσ = ∆ρνσµ = ∆ρσνµ , (1.21)
which contradicts (1.20) unless ∆ρσµν = 0.
Hence we find that for d > 2 the full, uncontracted third order derivative of must
vanish, so that it can be of at most second order in x. Therefore we may write
µ (x) = αµ + βνµ xν + γνρ
µ ν ρ
x x . (1.22)
Substituting this into (1.16) and collecting the terms of the same order in x we find the
conditions
βµν + βνµ = d2 β ρρ gµν
γµνσ + γνµσ = d2 γρσ
ρ
gµν
The first one can be solved by splitting βµν into a symmetric and an anti-symmetric part,
βµν = ωµν + Sµν . (1.23)
There is no condition on the anti-symmetric part ωµν , whereas the symmetric part is
found to be proportional to gµν , Sµν = σgµν .
The equation for the quadratic part is somewhat harder to solve. Using the fact that
γµνσ is symmetric in the last two indices, we can derive
γµνσ = −γνµσ − 2bσ gµν = −γνσµ − 2bσ gµν
= γσνµ − 2bσ gµν + 2bµ gνσ = γσµν − 2bσ gµν + 2bµ gνσ
= −γµσν − 2bσ gµν + 2bµ gνσ − 2bν gµσ
= −γµνσ − 2bσ gµν + 2bµ gνσ − 2bν gµσ ,
where bµ = − d1 γ ρρµ . Therefore
γµνσ = −bσ gµν + bµ gνσ − bν gµσ , (1.24)
where bν is an arbitrary constant vector.
10
1.3 The conformal group
Most of the transformations we have obtained can easily be identified
• Translations: xµ → xµ + αµ
• (Lorentz) Rotations: xµ → xµ + ω µν xν
• Scale transformations: xµ → xµ + σxµ
The last transformation is perhaps less familiar and is called a
• Special conformal transformation: xµ → xµ + bµ x2 − 2xµ b · x
Note that ωµν is antisymmetric, and this might suggest that it is the parameter of a rota-
tion rather than a Lorentz transformation. However, the correct infinitesimal parameters
of the transformation are ω µν . Numerically (i.e. ignoring its tensor properties) this is
equal to the matrix Ω = g −1 ω, which satisfies ΩT = ω T g T −1 = −ωg −1 = −gΩg −1 . Hence
gΩ + ΩT g = 0, so that to first order in Ω, (1 + ΩT )g(1 + Ω) = g. Hence Ω is indeed an
infinitesimal Lorentz transformation.
These are all still in infinitesimal form, but it is fairly straightforward to write their global
version. In addition to translations and SO(p, q) Lorentz transformations (or rotations if
q = 0) one has the scale transformation x → x0 = λx. The global version of the special
conformal transformation has the form
x µ + bµ x 2
xµ → x0µ = . (1.25)
1 + 2b · x + b2 x2
The latter transformation can be made a little more intuitive by observing that it can
be obtained by the sequence x → I(x), x → x + b, x → I(x), where I(x) denotes
the space-time inversion xµ → xµ /x2 . The space-time inversion can be thought of as a
global conformal transformation. It preserves angles, but it obviously does not have an
infinitesimal form, and therefore there is no parameter µ corresponding to it, and we did
not find it in the previous analysis.
One can study the action of the infinitesimal conformal transformations on a space of
functions of x. For each transformation x → x0 = x + (x) one can define a differential
operator O so that the transformation of a function f (x) is f (x) → f (x) + O f (x).
The usual convention for defining these operators (including a factor i to make them
Hermitean) is
Pµ = −i∂µ
Mµν = i(xµ ∂ν − xν ∂µ )
D = −ixµ ∂µ
Kµ = i(x2 ∂µ − 2xµ xν )∂ν (1.26)
The Lorentz generators close into the algebra of SO(p, q):
[Mµν , Mρσ ] = i(ηνρ Mµσ + ηµσ Mνρ − ηµρ Mνσ − ηνσ Mµρ ) (1.27)
One can write down the commutators of the operators P, M, D and K, and check that
they form a closed algebra which is isomorphic to SO(p + 1, q + 1).
11
1.4 Relation to stereographic projections
Consider the n-dimensional plane Rn and a sphere Sn , embedded in Rn+1 and having an
n-dimensional surface. If we identify all points at infinity in Rn with a single point, the
two surfaces can be mapped into each other in the following way. Imagine the sphere
suspended above the plane in Rn . Now draw straight lines from the top of the sphere to
any point on the plane. A stereographic projection is the map between the point on the
plane, and the point where the line intersects the sphere. This is illustrated below for
n = 1; a point x is projected to a point P (x) on the plane.
x
y
P(x) P(z) P(y)
The reason such maps are of interest here is that they are conformal; it is indeed straight-
forward to check that they preserve angles. Hence now we can do the following. Map Rn
to Sn ; rotate or translate Sn within Rn+1 ; then map the result back to the plane. Since
both maps are conformal, and so are rotations and translations, the result is a conformal
transformation in Rn . Let us first check that the number of generators matches:
1 1
n(n + 1) [Rn+1 rotations] + (n + 1) [Rn+1 translations] = (n + 2)(n + 1)
2 2
The result is indeed the number of generators of the conformal group of Rn , namely
SO(n + 1, 1). The precise relation between the transformations is as follows. Obviously,
the SO(n + 1) rotations that act within Rn form the SO(n) the rotations of the con-
formal group, and the translations along a vector paralel to Rn are the translations of
the conformal group. The translation of the sphere orthogonal to Rn results in dilations.
The special conformal transformations are generated by rotations along axes paralel to
Rn . The last point requires a bit more care, since there are n axes paralel to Rn , and
the SO(n + 1) rotations leaving each axis invariant form a subgroup SO(n). But we need
only n special conformal transformations, one for each axis. This works fine for n = 2,
because SO(2) has just one generator. Indeed, the action of the conformal group can be
visualized beautifully for n = 2. For n > 2 there exist rotations along a vector paralel to
Rn that act entirely within Rn , and therefore have already been counted. There is just
one independent new rotation for each such axis, and this generates a special conformal
transformation.
12
Of course one might consider not only rotations and translations in Rn+1 , but all Rn+1
conformal transformations. However, this does not yield anything new. Furthermore,
the idea is to use the exact metric invariances (rotations and translations) to derive the
conformal ones; if the other conformal transformations in Rn+1 were needed it would not
be much of an explanation. Note that this procedure gives the generators of the conformal
group, but not how they commute. One cannot conclude that the conformal group of Rn
is the Poincaré group in n + 1 dimensions! Furthermore, visualizing the conformal group
in Rp,q along these lines is less intuitive, as it requires hyperbolic geometry.
1.5 The conserved current

Usually symmetries imply the existence of conserved currents. The current corresponding
to a conformal symmetry transformation µ satisfying the condition (1.16) is
Jµ () = Tµν ν (1.28)
This current is conserved because
∂ µ Jµ () = (∂ µ Tµν )ν + Tµν (∂ µ ν ) . (1.29)
The first term vanishes because the energy momentum tensor is conserved (note that
we are in flat space) and the second term vanishes because of (1.16) and because Tµν is
traceless.
1.6 An example: the free boson

A standard example is the free boson. The (Euclidean) action is
√
Z
S = 2 dd x gg µν ∂µ Φ(x)∂ν Φ(x) ,
1
(1.30)
where g ≡ det g. To compute the energy momentum tensor we need the variation of g µν
√
and g, given δgµν . To get the former, use 0 = δ(g µν gνρ ) to derive δg µν = −g µρ g νσ δgρσ .
The second variation is derived as follows
√ 1 √
δ g = δe 2 log g = 12 gδ log(g) (1.31)
with
δ log g = log det (gµν + δgµν ) − log det gµν

= log det [δρµ + g µν δgνρ ] = Trg µν δgνρ + O(δ 2 ) ,
where in the last step the identity det A = exp Tr log A was used, and the log was
expanded in g µν δgνρ . In the first and second line the arguments of “det” and “Tr” are
13
matrices with indices (µ, ν) and (µ, ρ) respectively. Putting this all together, and using
(1.2) we get
Tµν = −∂µ Φ∂ν Φ + 12 gµν g ρσ ∂ρ Φ∂σ Φ . (1.32)
It is straightforward to check that ∂ µ Tµν = 0 and that T µµ ∝ (1 − d2 ), so that the
theory is conformally invariant if (and only if) d = 2. Note that to prove ∂ µ Tµν = 0 one
has to use equation of motion Φ = 0, whereas tracelessness for d = 2 holds also if the
equation of motion is not satisfied.
A theory with classical conformal invariance in four dimensions is Yang-Mills theory
(both abelian and non-abelian). The verification is left as an exercise.
One may also directly check Weyl invariance of the free bosonic theory. If we trans-
form gµν to Ω(x)gµν , the square root of the determinant aquires a factor Ωd/2 . This
precisely cancels the factor Ω−1 from the transformation of g µν (the inverse of gµν ), in two
dimensions.
By contrast, a conformal transformation does not act on the metric, but changes the
integration variables and the derivatives. The simplest non-trivial example is a scale
transformation x = f (x0 ) = λx0 . Then dd x0 = λ−d dx0 and the new field is ∂x0 Φ0 (x0 ) =
(∂f /∂x0 )∂f Φ(f (x0 )) = λ∂x Φ(x), where Φ0 (x0 ) ≡ Φ(f (x0 )) = Φ(λx0 ) = Φ(x). For d = 2 the
explicit λ-dependence cancels out, and this demonstrates part of the conformal symmetry
of the action. R
It should be clear that adding a mass term d2 x m2 Φ2 to the theory breaks con-
formal invariance. Quantum effects also tend to spoil conformal invariance. Generically
they introduce renormalization scale dependence of physical parameters (such as coupling
constants) which destroys invariance under scale transformations q → λq in momentum
space. This does indeed happen for Yang-Mills theories, except when the β function van-
ishes so that the coupling constant is scale independent. The latter occurs for N = 4
super-Yang-Mills theory in four dimensions.
1.7 The conformal algebra in two dimensions

In two dimensions the restriction that (x) is of at most second order in x does not apply.
One can analyze (1.16) directly by writing it out in components. If one does that in
Euclidean space, gµν = δµν , one finds
∂1 1 = ∂2 2 , ∂1 2 = −∂2 1 . (1.33)
Going to complex variables,
= 1 − i2 , ¯ = 1 + i2
z = x1 − ix2 , z̄ = x1 + ix2 (1.34)
with
x1 = 21 (z + z̄) x2 = 2i (z − z̄)
∂1 = ∂z + ∂z̄ , ∂2 = −i(∂z − ∂z̄ )
14
we find
∂z ¯(z, z̄) = 0; ∂z̄ (z, z̄) = 0 , (1.35)
∂
with ∂z ≡ ∂z and analogously for z̄. The general solution to these conditions is that is
an arbitrary function of z (which does not depend on z̄) and ¯ an arbitrary function of z̄.
The corresponding global transformation is
z → f (z), z̄ → f¯(z̄) , (1.36)
where f (z) is an arbitrary function of z.

Analogously to (1.26) one may consider infinitesimal versions of these transformations
and define differential operators that generate them; the generator
Ln = −z n+1 ∂z (1.37)
generates the transformation

z → z 0 = z − z n+1 , (1.38)
and satisfies the commutation relation
[Ln , Lm ] = (n − m)Lm+n . (1.39)
The same holds for the barred quantities, and furthermore one has then
[Ln , L̄m ] = 0 . (1.40)
The resulting infinitesimal transformations are the most general ones that are analytic
near the point z = 0. They may introduce poles at z = 0, but not branch cuts. We will see
later that we will often need contour integrals around z = 0, and this is the justification
for this restriction.
The generator of an arbitrary conformal transformation is thus
X
n Ln + ¯n L̄n . (1.41)
n
This operator generates conformal transformations of functions f (z, z̄). If we want this
transformation to respect complex conjugation of z, we must require that ¯n is the complex
conjugate of n . In that case we can rewrite Eqn. (1.41) as
X
1

2
Re n (Ln + L̄n ) + Im n i(Ln − L̄n ) (1.42)
n
This is in fact the algebra written in terms of the original real coordinates x1 and x2 .
15
1.8 Complexification and Wick rotation
Usually this reality condition is dropped, and one treats and ¯ as independent complex
parameters. Then the algebra does not map (x1 , x2 ) ∈ R2 to another point in R2 , but it
is a well-defined map on C2 . This is justified if we define our field theory on a complex
instead of a real space-time. This allows us to treat the two commuting algebras generated
by Ln and L̄n independently (i.e. we may now set ¯n = 0, n 6= 0 or vice-versa). Even if
we are ultimately only interested in the restriction to a real vector space, we can always
impose the reality condition at the end.
Note that the distinction between Euclidean space and Minkowski space become
irrelevant if we complexify the coordinates. The complex coordinate transformation
x0 = −ix2 changes ηµν to δµν [Our convention is to use indices (0, . . . d − 1) in d di-
mensional Minkowski space, with x0 as the time coordinate, and (1, . . . , d) in Euclidean
space, with x2 = ix0 (= −ix0 ). Consequently the indices on δ and η have a different
range.] This is known as a Wick rotation. The sign of the Wick rotation is determined by
requiring that in Minkowski time evolution towards the future the time evolution operator
becomes a negative exponential exp (−H∆t), dominated by the lowest states. See also
chapter 2.
We are usually interested in conformal field theories in Minkowski space, but it is
convenient to make use of the powerful theorems that are available for complex functions.
For that reason one usually makes a Wick rotation to Euclidean space, which in its turn is
mapped to the complex plane. This is not an obviously innocuous transformation though.
The Wick rotation changes (in fact, improves) the convergence properties of quantities
such as the path integral or the propagator in the quantum theory, which is why it is often
used in field theory in four dimensions as well. One has to assume or, if possible, prove
that the relevant quantities can indeed by analytically continued to Euclidean space, and
if there are singularities one has to find a way to avoid them.
1.9 The global subgroup

An interesting subalgebra of the algebra is the one generated by L−1 , L0 , L1 and their
conjugates. This algebra – or rather its restriction to real generators, as discussed above
– is isomorphic to SO(3, 1), which is precisely the naively expected conformal group
SO(p + 1, q + 1), if one extrapolates from arbitrary d to d = 2 (in Euclidean space, with
p = 2 and q = 0).
The precise identification can easily be derived by transforming back to the standard
16
Euclidean coordinates x1 , x2 . The precise relation with the operators defined in (1.26) is
P1 = −i(∂z + ∂z̄ ) = i(L−1 + L̄−1 )

P2 = −(∂z − ∂z̄ ) = (L−1 − L̄−1 )
M = z∂z − z̄∂z̄ = −L0 + L̄0
D = −i(z∂z + z̄∂z̄ ) = i(L0 + L̄0 )
K1 = −i(z 2 ∂z − z̄ 2 ∂z̄ ) = i(L1 + L̄1 )
K2 = −(z 2 ∂z − z̄ 2 ∂z̄ ) = −(L1 − L̄1 ) (1.43)
The algebra satisfied by the holomorphic generators is
[L0 , L−1 ] = L−1

[L0 , L1 ] = −L1
[L1 , L−1 ] = 2L0
This is precisely the SU (2) rotation algebra if we identify L0 with Jz , iL1 with J − =
Jx − iJy and iL−1 with J + = Jx + iJy . The factor i is essential to compensate the sign in
[J − , J + ] = −2J 0 .
The SO(3, 1) generators are the only ones that are globally defined on the complex
plane including ∞ (this is called the Riemann sphere). Clearly the generator −z n+1 ∂z is
non-singular at z = 0 for n ≥ −1. To investigate the behavior at infinity it is convenient
to make a conformal mapping that interchanges the points z = 0 and z = ∞. A conformal
map that does this is z = w1 . Under this transformation the generator Ln transforms to
−1
n+1 −(n+1) dz
−z ∂z → −w ∂w = +w1−n ∂w (1.44)
dw
This operator is non-singular for n ≤ 1, which combined with the range obtained above
leaves −1 ≤ n ≤ 1. For these values of n the generators are defined on the Riemann
sphere. The infinitesimal and global forms of the transformations are as shown in this
table.
generator local transformation global transformation

−L−1 z →z+ z →z+α
−L0 z → z + z z → λz
z
−L1 z → z + z 2 z→ 1−βz
Combining these transformations we get

az + b
z→ , with ad − bc = 1 (1.45)
cz + d
17
Note that there are only three independent transformations, and hence there should be
only three parameters. This is indeed true because of the condition ad − bc = 1.∗ Doing
it this way and taking the parameters ∈ C, we see that the action is that of the group
SL2 (C)/Z2 .
The group SL2 (C) is the set of 2 × 2 complex matrices with determinant 1. The most
general such matrix is !
a b
with ad − bc = 1 . (1.46)
c d
To get
the transformation shown above we make it act on complex two-dimensional
z1
vectors z2 , with vectors related by an overall complex scale identified. In this space
only the ratio z = z1 /z2 is a free parameter, and that parameter is easily seen to transform
as in (1.45). The transformation (1.45) is clearly unchanged if we multiply the matrix
by an overall factor. This freedom is fixed by the determinant condition, except for an
overall sign. Therefore the correct group action is SL2 (C)/Z2 rather than SL2 (C).
In combination with the transformation of the anti-holomorphic sector we get then
the group of transformations SL2 (C) × SL2 (C). This contains as a subgroup SO(3, 1),
the expected global conformal group, but in terms of real generators SL2 (C) × SL2 (C)
is twice as large as SO(3, 1). The reason is of course that we allow the two SL2 (C)
transformations to act independently on z and z̄. If we impose a reality condition (i.e. z̄
is the complex conjugate of z) we reduce the number of generators to that of SO(3, 1).
1.10 Tensors in complex coordinates

Tensors can be transformed to complex coordinates using the transformation formula
(1.11). One should be careful with factors of two in these transformations. A potential
source of confusion is the fact that often the same notation is used for coordinates and
indices: z ≡ xz ; z̄ ≡ xz̄ . The transformation is
z ≡ xz = x1 − ix2 ; z̄ ≡ xz̄ = x1 + ix2 (1.47)
The inverse is then
z ≡ x1 = 21 (xz + xz̄ ) = 21 (z + z̄) ; x2 = 12 i(xz − xz̄ ) = 21 i(z − z̄) (1.48)
The transformation of a vector to complex coordinates goes as follows
∂x1 ∂x2
Vz = V1 + V2 = 12 (V1 + iV2 ); Vz̄ = 21 (V1 − iV2 ) (1.49)
∂xz ∂xz
The generalization to higher rank tensors is obvious.
∗
√ √
Note that the scale transformation is given by a = λ, d = 1/ λ, b = c = 1.
18
The metric gµν = δµν transforms to gzz = 14 (g11 + ig12 + ig21 − g22 ) = 0 = gz̄z̄ ,
gzz̄ = 41 (g11 + ig12 − ig21 + g22 ) = 12 = gz̄z . Hence g zz̄ = g z̄z = 2. This leads to the
akward-looking relation ∂ z̄ = 2∂z ∗.
The same kind of relation holds for the energy-momentum tensor. Since T is traceless
we have T11 + T22 = 0, and hence Tzz̄ = 0.
The metric allows us to convert every upper index to a lower one (or vice-versa) at the
expense of a simple numerical factor, 2 or 12 . One can make use of this freedom to avoid
counter-intuitive quantities such as xz or ∂ z . From now on all tensors and derivatives will
be written with lower indices, and all coordinates with upper indices. The latter will be
denoted as z or z̄.
Conservation of the energy momentum tensor now reads (since Tzz̄ = 0)
∂z̄ Tzz = ∂z Tz̄z̄ = 0 , (1.50)
which implies that Tzz is holomorphic and Tz̄z̄ anti-holomorphic.†

The infinitesimal parameter for the conformal transformations, , has been transformed
to complex components in (1.34). This definition also requires a bit of care. If we use
(1.49) we get
z = 12 (1 + i2 ) = 21 ¯ , (1.51)
where in the last step the definition (1.34) was used. The “bar” on the right-hand side
may look out of place, but the notation was chosen in (1.34) because ¯ is a function only
of z̄.
The conserved current of conformal symmetry is defined analogously to (1.28)
Jµ () = Tµν ν =→ Jz = 2Tzz z̄ = Tzz (z); Jz̄ = Tz̄z̄ ¯(z̄) . (1.52)
Since Jz is holomorphic and Jz̄ anti-holomorphic, this current is manifestly conserved:

∂z̄ Jz = ∂z Jz̄ = 0
1.11 Conformal fields in two dimensions

The components of a tensor φ of rank n are of the form φz...z,z̄...z̄ (z, z̄). It is easy to see
that under conformal transformations this transforms into
p ¯ q
∂f (z) ∂ f (z̄)
φz...z,z̄...z̄ (f (z), f¯(z)) , (1.53)
∂z ∂ z̄
where p is the number of indices ‘z’, and q = n − p the number of indices z̄. A field that
transforms in this way is called a conformal field of weight (p, q).
∗
One should keep in mind here that ∂z can have two meanings, namely the “derivative with respect
to the coordinate xz ” (∂/∂xz ) or the “derivative with respect to the variable z” (∂/∂z.) Fortunately
these two meanings are the same. On the other hand ∂ z̄ can only mean “derivative with respect to the
coordinate xz̄ ”, and has nothing to do with a derivative with respect to z̄. † The word “holomorphic”
has become standard terminology for “depending only on z, not on z̄”. It does not imply absence of
singularities. Mathematicians might prefer the word “meromorphic”.
19
This rule was derived here for a tensor field, and one may think that p and q should
be integers. However, any real value is in fact (a priori) allowed. Usually the conformal
weight is denoted as (h, h̄), where the bar does not mean complex conjugation (both
numbers are real), but only serves to distinguish the two numbers. Sometimes h + h̄ is
called the scaling weight, and h − h̄ the conformal spin. As we will see later, these are
in fact the eigenvalues of (minus) the dilation operator −D = L0 + L̄0 and the SO(2)
rotation operator −iM = L0 − L̄0 .
1.12 Relation to string theory

Closed bosonic strings are described by means of the bosonic action (in Minkowski space,
with g00 = −1 and g11 = 1)
√
Z
1
S=− 0
d2 σ −gg αβ ∂α X µ ∂β Xµ , (1.54)
4πα
defined on a two-dimensional surface with the topology of a cylinder (for the non-interacting,
closed string, at least). Here X µ (σ0 , σ1 ) is a map from two dimensional space (called the
“world-sheet”) to space-time (often called “target space”), and α0 is the “Regge-slope
parameter”. This function defines the embedding of the string in space time, as a func-
tion of the proper time σ0 , i.e. is specifies where a point σ1 along the string is located at
proper time σ0 . If we take a flat two-dimensional metric g αβ and a Euclidean flat metric in
target space this action is nothing but the action of a free boson in two dimensions. The
conformal invariance of that action plays an important rôle in the proper quantization of
string theory in Minkowski space (note that X 0 appears then with the “wrong” sign in
the two-dimensional action). Furthermore conformal field theory has been used to find
alternatives to the free boson action that can be interpreted as consistent string theories.
1.13 Free bosons in complex coordinates

In complex coordinates the free boson action takes the form
Z
1
S= dzdz̄ ∂z Φ(z, z̄)∂z̄ Φ(z, z̄) (1.55)
2πα0
Our convention is that dzdz̄ = 2dx1 dx2 (which is indeed what one gets from the Jacobian,
but some authors prefer to omit the factor “2”). This complex form of the action is derived
from the Euclidean action (1.30) multiplied with an additional factor 1/2πα0 , so that one
gets the Euclidean action corresponding to Eqn. (1.54). The factor in front is conventional
in string theory.
According to the definition of conformal fields given earlier, ∂Φ is a conformal field.
In complex coordinates its transformation properties are
∂f (z)
∂z Φ(z, z̄) → ∂f Φ(f (z), f¯(z̄)) . (1.56)
∂z
20
For the energy momentum tensor we find, in complex coordinates
Tzz = − 12 (∂z Φ(z))2 (1.57)
and analogously for Tz̄z̄ . The normalization contains some conventions originating from
string theory, namely
1
• The factor 2πα0
, as above
• An extra factor 2π multiplying the definition of T µν (cf. footnote on page 24 of part

1 of [31]).
• The convention α0 = 2.
2 Quantum Conformal Invariance

As discussed in the previous chapter, the theories we consider are defined in Euclidean
space, usually obtained after a Wick rotation from Minkowski space. For computational
convenience (in particular because some quantities separate into holomorphic and anti-
holomorphic parts) we then go to the complex plane. It turns out that to simplify things
even more it is convenient to make yet another map, this time a conformal transformation
of the complex plane itself.
2.1 Radial quantization

Symmetries in the quantum theory are usually generated by charges, which are space
integrals of the zeroth component of a conserved current J µ , ∂µ J µ = 0. The definition of
a charge in a d-dimensional theory is then
Z
Q = dd−1 xJ 0 (x, t) . (2.1)
In the two-dimensional analog of this just one has a one-dimensional integration over x1 .
It is convenient to make the space direction finite, by imposing periodic boundary
conditions in the x1 direction. This is like regulating a quantum system by putting it
in a finite box in space. In this case the size of the box will be fixed for convenience
to the value 2π, but since the theory is scale invariant that is irrelevant. The Euclidean
coordinates (x1 , x2 ) = (x1 , ix0 ) can then be thought of coordinates on a cylinder. In this
situation we get then the following expression for the charge (conveniently normalized to
the length of the interval)
Z 2π Z 2π
1 1 0 1
Q= dx J = dx1 (−iJ2 ) . (2.2)
2π 0 2π 0
21
Now we introduce a complex coordinate z = x1 − ix2 as discussed before. From (1.49) we
find that J2 = −i(Jz − Jz̄ ). The charge becomes now
I I
1 cyl cyl
Q=− dzJz (z, z̄) − dz̄Jz̄ (z, z̄) (2.3)
2π
Here the integration is along a closed contour that encircles the cylinder. For convenience
we choose z̄ rather than z as the integration variable in the second term. Since we only
integrate
H overH Rez this makes no difference. The orientation of these contours is such
that dz = dz̄ = 2π. The superscripts “cyl” are added to remind ourselves that the
currents are defined on the cylinder.
It turns out to be convenient to perform a conformal transformation
2 +ix1
w = ex = eiz . (2.4)
Then the surface at the Euclidean time coordinate x2 = −∞ is mapped to w = 0, and
the surface at x2 = +∞ is mapped to the infinite circle at |w| = ∞.
22
To go to the new coordinates we first make a change of integration variables. Then
I I I I
dw dw̄
dz = ; dz̄ = . (2.5)
iw iw̄
It is clear from the picture that in the new coordinates the dz integration becomes a
contour integration around the origin. The definition of the integration volume for dw̄
implies a choice of orientation for the w̄ contour. Rather than remembering the direction
of the contour, it is easier to remember the corresponding results for the Cauchy integrals
I I
1 dw 1 dw̄
= =1 (2.6)
2πi w 2πi w̄
I I
1 dw cyl dw̄ cyl
Q=− J (z(w), z̄(w̄)) − J (z(w), z̄(w̄)) (2.7)
2π iw z iw̄ z̄
This formula has the disadvantage that Q is still expressed in terms of operators de-
fined on the cylinder. These operators are related to those on the plane by a conformal
transformation: h h̄
plane ∂z ∂ z̄
φ (w, w̄) = φcyl (z(w), z̄(w̄)) (2.8)
∂w ∂ w̄
For the transformation considered here this implies
φcyl (z(w), z̄(w̄)) = (iw)h (−iw̄)h̄ φplane (w, w̄) (2.9)
The current components considered here, Jz and Jz̄ transform as vectors; therefore they
have conformal weights (h, h̄) = (1, 0) and (h, h̄) = (0, 1) respectively. For the charge we
find then
I I
1 h−1 plane h̄−1 plane
Q=− dw(iw) Jw (w, w̄) + dw(−iw̄) Jw̄ (ww̄) (2.10)
2π
Here h = h̄ = 1, but we left these parameters in the formula for future purposes.
Usually the current splits into holomorphic and anti-holomorphic parts, so that we
¯
may write Jz (z, z̄) ≡ J(z) and Jz̄ (z, z̄) ≡ J(z̄). If a vector current has that property, then
it is automatically conserved:
∂ µ Jµ = 2 [∂z Jz̄ (z̄) + ∂z̄ Jz (z)] = 0 . (2.11)
The result of the contour integration depends, obviously, on the poles inside the con-
tour. Such poles can arise in the quantum theory when one considers the product of two
or more operators.
23
2.2 Radial ordering
Products of operators only make sense if they are radially ordered. This is the analogue
of time ordering for field theory on the cylinder. In the classical theory the ordering of
fields or charges in a product is of course irrelevant. In the quantum theory they become
operators and we have to specify an ordering. The product of two operators A(xa , ta ) and
B(xb , tb ) can be written, with the help of the Hamiltonian H of the system as
A(xa , ta )B(xb , tb ) = eiHta A(xa , 0)e−iHta eiHtb A(xb , 0)e−iHtb (2.12)
The factor e−iH(ta −tb ) becomes e−H(τa −τb ) when we Wick-rotate (here t corresponds to x0 ,
τ to x2 ). Usually the Hamiltonian is bounded from below, but not from above. Then
if τa < τb the exponential can take arbitrarily large values, and expectation values of the
operator product are then not defined. Hence in operator products one always imposes
time ordering, usually denoted as
(
A(ta )B(tb ) for ta > tb
T A(ta )B(tb ) = (2.13)
B(tb )A(ta ) for ta < tb
After mapping from the cylinder to the plane, the Euclidean time coordinate is mapped
to the radial coordinate, and time ordering becomes radial ordering
(
A(z, z̄)B(w, w̄) for |z| > |w|
RA(z, z̄)B(w, w̄) = . (2.14)
B(w, w̄)A(z, z̄) for |z| < |w|
A correlation function in field theory on the cylinder has the form
h0| T (A1 (t1 ) . . . An (tn )) |0i (2.15)
where |0i and h0| are “in” and “out” states at t = −∞ and t = +∞ respectively. After
the conformal mapping, the correlation functions are
h0| R (A1 (z1 , z̄1 ) . . . An (zn , z̄n )) |0i (2.16)
where |0i and h0| are states at z = 0 and z = ∞ respectively.
2.3 The generator of conformal transformations

Returning now to charge operators, let us consider the generator of the conformal trans-
formations. As we have seen in the previous chapter, the current for an infinitesimal
transformation is T (z)(z). For the corresponding charge we may then write
I I
1 1
Q = dz(z)T (z) + dz̄(z̄)T̄ (z̄) (2.17)
2πi 2πi
From now on the variables z and w are always coordinates on the complex plane, and
should not be confused with the complexified cylinder coordinates. Eqn. (2.17) defines Q
24
as an operator directly on the plane, and the foregoing serves here only as an inspiration.
We will worry later about the precise relation with operators on the cylinder. Finally, the
anti-holomorphic component of T will be omitted from here on.
We would expect Q to generate the conformal transformation with the global form
h
0 ∂f (w)
φ(w, w̄) → φ (w, w̄) = φ(f (w), w̄) , (2.18)
∂w
with f (w) = w + (w). Note that any field Φ will in general depend on both w and
w̄, but that we are treating w and w̄ as independent variables, which can therefore be
transformed independently. The infinitesimal form of this transformation is
δ Φ(w, w̄) = h∂w (w)Φ(w, w̄) + (w)∂w Φ(w, w̄) (2.19)
Consider now the quantum version of this transformation. We may expect the follow-
ing relation to hold
δ φ(w, w̄) = [Q , φ(w, w̄)] (2.20)
Let us try to evaluate the commutator on the right hand side. Naively we have
I
1
[Q , φ(w, w̄)] = dz(z) [T (z)φ(w, w̄) − φ(w, w̄)T (z)] . (2.21)
2πi
But we have just seen that the first term is defined only if |z| > |w|, whereas the second one
requires |z| < |w|. Note however that z is an integration variable, and that the definition
of Q did not include any prescription for the precise contours to be used. Classically Q
is in fact independent of the contour due to Cauchy’s theorem, because the integrand is a
holomorphic function. On the cylinder this can be interpreted as charge conservation, i.e.
evaluating Q at two different times gives the same answer. Classically the factor φ(w, w̄)
is irrelevant for the evaluation of the integral, and in fact classically the commutator
vanishes. In the quantum theory we have to be more careful. As one usually does, we use
the freedom we have in the classical theory in order to write the quantity of interest in
such a way that it is well-defined after quantization. Nothing forbids us to use different
contours in (2.21), so that we get
I I
1 1
[Q , φ(w, w̄)] = dz(z)T (z)φ(w, w̄) − dz(z)φ(w, w̄)T (z) . (2.22)
2πi |z|>|w| 2πi |z|<|w|
This can be written as

I I
1
[Q , φ(w, w̄)] = − dz(z)R(T (z)φ(w, w̄)) . (2.23)
2πi |z|>|w| |z|<|w|
Now we deform the contours as indicated in the following figure
25
z z
z
w
w w
0
- 0 = 0
The result is I
1
[Q , φ(w, w̄)] = dz(z)R(T (z)φ(w, w̄)) , (2.24)
2πi
where the integration contour encircles the point w. Clearly the integration only makes
sense if the radially ordered product is analytic in the neighborhood of the point w. We
may thus assume that it can be expanded it in a Laurent series around w
X
R(T (z)φ(w, w̄)) = (z − w)n On (w, w̄) , (2.25)
n
where the coefficients On are operators. It is now easy to verify that the contour integral
will produce the desired result if (and only if) the radially ordered product equals
h 1
R[T (z)φ(w, w̄)] = φ(w, w̄) + ∂w φ(w, w̄) + power series in (z − w) (2.26)
(z − w) 2 z−w
The last terms are free of poles at z = w, and hence do not contribute to the integral.
A few points should be noted in this computation. The integrand may have any
number of singularities, as long as they are isolated poles. In that case the two contours
can always be chosen in such a way that their difference only includes the singularity at
w, which is necessarily included because of radial ordering. If the integrand has branch
cuts the computation does not make sense. Branch cuts (for example logarithmic or
fractional power behavior) in operator products can in general not be tolerated in a
sensible conformal field theory. Later we will see how one can restrict the set of operators
in such a way that all operator products are well behaved. Note that the contour around
w crosses the circle |z| = |w|, where radial ordering cannot be imposed. This should be
understood in terms of analytic continuation of the integrand; after having used radial
ordering to define it properly, we find that in fact it is defined everywhere except in the
point z = w. This is the only point any contour should avoid.
The property (2.26) (plus the corresponding one for the anti-holomorphic quantities)
defines what we mean by a conformal field.
Often one simplifies the notation by omitting the radial ordering symbol (which how-
ever is always implied) and dropping the finite terms.
2.4 Quantization of the free boson

Now we return to the free boson, in order to look at a few concrete examples of the
foregoing, rather abstract discussion. On the cylinder, and in Minkowski space-time, the
26
action is Z
1 X
S= d2 x ∂α Φi ∂ α Φi (2.27)
8π i
1 0
(the normalization factor is 4πα0 with α = 2). The fact that the fields Φ live on a cylinder
implies that they satisfy periodic boundary conditions:

Φi (x0 , 0) = Φi (x0 , 2π) (2.28)
Any field satisfying these boundary condition can be Fourier expanded
∞
1
X
Φi (x0 , x1 ) = einx fni (x0 ) (2.29)
n=−∞
The classical equation of motion for Φ is

[∂02 − ∂12 ]Φi (x0 , x1 ) = 0 (2.30)
For the Fourier modes of a classical solution this implies
∂02 fni (x0 ) = −n2 fni (x0 ) . (2.31)
The solution is
0 0
fni (x0 ) = ain einx + bin e−inx , n 6= 0 (2.32)
and
f0i (x0 ) = pi x0 + q i . (2.33)
Putting all this together, and introducing a few convenient factors, we may write the
result as
X1
i 0 1 i i 0 i −in(x0 +x1 ) i −in(x0 −x1 )
Φ (x , x ) = q + 2p x + i (αn e + α̃n e ) , (2.34)
n6=0
n
Now we quantize this field using canonical quantization. The canonical momentum is
1
Πi = ∂0 Φi (2.35)
4π
We impose on it the following commutation relations
i 0 1
Φ (x , x ), Πj (x0 , y 1 ) = iδ ij δ(x1 − y 1 )

i 0 1
Φ (x , x ), Φj (x0 , y 1 ) = 0

i 0 1
Π (x , x ), Πj (x0 , y 1 ) = 0 .

(2.36)
These relations can be expressed in terms of modes by taking Fourier moments of these
conditions with respect to x1 and x01 . The result is
i j i j
αk , αl = α̃k , α̃l = kδ ij δk+l,0
i j
αk , α̃l = 0 (2.37)
27
and i j
q , p = iδ ij (2.38)
Now transform φ to the complex plane. Then we get
X1
Φi (z, z̄) = q i − i(pi log(z) + pi log(z̄)) + i αni z −n + α̃ni z̄ −n

(2.39)
n6=0
n
Note that Φi can almost be written as the sum of holomorphic and anti-holomorphic
functions. Almost, because log(z) is not a holomorphic function. However, if we take a
derivative of Φi we do get purely holomorphic functions, namely ∂z Φi (z) and ∂z̄ Φi (z̄).
2.5 The free boson propagator

The first quantity of interest is the product Φi (z)Φj (w). The quantum equivalent of
this classical product is R(Φi (z)Φj (w)), and we would like to know the behavior of this
product as z approaches w. A simple way to probe the short distance behavior is to
take expectation values of the operator between two states, for example the vacuum. To
compute this we first have to know how the various mode operators act on the vacuum.
Classically the field Φi is real. Consequently, in (2.34) q i and pi are real, and αni =
(α−n )∗ . The quantum equivalent of this statement is that q i and pi are represented by Her-
i
i
mitean operators, whereas α−n is the Hermitean conjugate of αni . Reality conditions are
best imposed on the cylinder and in terms of Minkowski coordinates. The complex Wick
rotation and subsequent conformal mappings can make reality properties less manifest.
We see now that the commutation relation [αki , αlj ] = kδ ij δk+l,0 , k > 0 is the same as
that for a set of harmonic oscillators, apart from the factor k, which can be absorbed in
the normalization of the operators (note that the commutator for k < 0 contains no new
information). Indeed, apart from the “zero mode” q i , pi the free boson is nothing but an
infinite set of harmonic oscillators.
By the usual reasoning for harmonic oscillators, the vacuum satisfies
αki |0i = 0 for k > 0 (2.40)
The algebra of the operators pi and q i is also a well-known one, namely the Heisenberg
algebra. Hence the vacuum must satisfy
pi |0i = 0 (2.41)
This is all we need to compute the vacuum expectation value. A convenient technique
for computing vacuum expectation values is normal ordering. We reorder the oscillators
in such a way, using the commutators, that creation operators are always to the left of
annihilation operators. Then the vacuum expectation value of normally ordered terms
always vanishes for every term that contains at least one harmonic oscillator, and we only
have to take into account the contributions picked up from the commutators.
28
Normally ordered products of oscillators are denoted as
αk1 . . . αkn :≡ Πi αkπ(i) , (2.42)
where π(i) is a permutation of the labels such that kπ(i) < kπ(j) if i < j (re-ordering
positive and negative labels among each other has no effect, but does not hurt either).
Note that oscillators within the normal ordering signs behave as if they are classical. They
can be written in any order, since the right hand side is always the same.
The only terms in the product Φi (z)Φj (w) that does not contain oscillators is the
zero mode contribution. These terms require some special attention. We define normal
ordering of pi and q i in such a way that pi is always to the right of q i . Using these rules
we get, when |z| > |w|,
R(Φi (z, z̄)Φj (w, w̄)) = : Φi (z, z̄)Φj (w, w̄) : −i[pi , q j ](log z + log z̄)
(" # )
X1 X 1
+ i αni z −n , i j
αm w−m + anti-holomorphic terms
n>0
n m<0
m
The commutator yields

X 1
− nz −n w−m δn+m,0 , (2.43)
n>0,m<0
nm
so that we get
( )
X 1 w n X 1 w̄ n
: Φi (z, z̄)Φj (w, w̄) : + δ ij −log(z z̄) + + . (2.44)
n>0
n z n>0
n z̄
The sum converges for |z| > |w|. Since the product was radially ordered, this is satisfied.
The result is
R(Φi (z, z̄)Φj (w, w̄)) =: Φi (z, z̄)Φj (w, w̄) : −δ ij [log(z − w) + log(z̄ − w̄)] . (2.45)
This is not quite what one usually gets when evaluating an operator product. Normally
the result consists of holomorphic and anti-holomorphic parts, whereas here there is a log-
arithmic singularity. A more standard result is the operator product ∂Φi (z, z̄)∂Φj (w, w̄),
which can be obtained from the above by differentiation (the notation ∂ is short-hand for
either ∂z or ∂w , depending on what it acts on).
1
R(∂Φi (z, z̄)∂Φj (w, w̄)) = −δ ij + : ∂Φi (z, z̄)∂Φj (w, w̄) : (2.46)
(z − w)2
since ∂z Φi (z, z̄) depends only on z we usually omit the second argument. Furthermore ∂z
is usually written as just ∂, if no confusion is possible. Furthermore the radial ordering is
usually not explicitly written, and the finite terms are usually omitted as well. Since the
objects within normal ordering signs behave as classical quantities, these are in particular
finite as z approaches w. Using all this short-hand notation, the result is then written as
δ ij
∂Φi (z)∂Φj (w) = − . (2.47)
(z − w)2
29
2.6 The normally ordered energy momentum tensor
This result shows that we have to be careful with the Pdefinition of the quantum en-
1
ergy momentum tensor, which classically is T (z) = − 2 i ∂z Φ (z)∂z Φi (z), plus the anti-
i
holomorphic term. If we naively quantize Φi the product of the two operators is singular.
For this reason one defines
X X δ ii
T (z) ≡ − 12 : ∂Φi (z)∂Φi (z) := − 21 lim [∂Φi (z)∂Φi (w) + ]. (2.48)
i i
z→w (z − w)2
This amounts to subtracting an infinite constant from the energy momentum tensor. This
sets the energy of the vacuum to zero.
2.7 Operator products for free bosons

We are now ready to compute the operator product of the energy momentum tensor
with various operators in the theory. Let us first consider T (z)∂Φi (w). To compute this
operator product we normal order all harmonic oscillators and the zero-mode operators
q i and pi . The operators within T (z) are already normal ordered, and hence the only
ordering to worry about is between T (z) and ∂Φi (w). We may write this as
δ ij

1 i i j 1 i i j i
− 2 : ∂Φ (z)∂Φ (z): ∂Φ (w) = − 2 : ∂Φ (z)∂Φ (z)∂Φ (w): −∂Φ (z) − (2.49)
(z − w)2
Note the factor of two in the last term, because there are two factors ∂Φi (z) to order with
respect to ∂Φj (w).
To get the operator product in the desired form we wish to express the remaining
factor ∂Φj (z) in terms of ∂Φj (w). This is simply a Taylor expansion, ∂Φj (z) = ∂Φj (w) +
(z − w)∂ 2 Φj (w) + 12 (z − w)2 ∂ 3 Φj (w) + . . .. The final result may thus be written as
1 1
T (z)∂Φj (w) = ∂Φj (w) + ∂ 2 Φj (w) , (2.50)
(z − w) 2 z−w
where as usual we drop all finite terms, and all operators appearing on the right hand
side are normally ordered. Hence ∂Φj (w) is a conformal field with conformal weight 1. In
a similar way one may check that ∂ 2 Φi (w) is not a conformal field. This is not a surprise,
because we have seen before that it is not a conformal field classically.
Now consider the energy momentum tensor itself. It is a simple exercise to compute
c/2 2 1
T (z)T (w) = + T (w) + ∂w T (w) (2.51)
(z − w)4 (z − w) 2 z−w
Here c equals the number of bosons Φi . If the first term were absent, T (z) would be a
conformal field of weight 2, the classical value. In this case quantum effects yield an extra
term, an anomaly. This is called the conformal anomaly.
30
2.8 The Virasoro algebra
The operator product (2.51), derived here for free bosons, has a completely general va-
lidity. Under quite general assumptions, one may show that the operator product of two
energy momentum tensors of a conformal field theory must have the form (2.51).
In (1.52) a current for conformal symmetry was introduced, J (z) = T (z)(z). Since
(z) is an arbitrary holomorphic function, it is natural to expand it in modes. The precise
mode expansion one uses depends on the surface one is working on. On the Riemann
sphere we require fields and transformations to be continuous on contours around the
origin. This was also the surface for which the classical mode expansion (1.37) was
written down. We expect thus that J (z) generates the transformation z → z 0 = z − z n+1
if we choose (z) = z n+1 . We then get an infinite series of currents J n (z) = T (z)z n+1 .
The correctly normalized operators are in fact
I
1
Ln = dzz n+1 T (z) . (2.52)
2πi
This relation can be inverted:
X
T (z) = z −n−2 Ln . (2.53)
n
To check that the normalization and the sign are correct one may compare the quantum
algebra with the classical algebra. The commutator of Ln and Lm can be evaluated using
contour integrals, as was already done earlier. One finds then the Virasoro algebra (the
paper by Virasoro [56], to which this algebra owes its name, contains the generators of
the algebra as “constraints”, but not the algebra itself)
c
[Ln , Lm ] = (n − m)Lm+n + n(n2 − 1)δn,−m (2.54)
12
The details of this derivation are as follows
I I
dz dw dw dz
[Ln , Lm ] = − z n+1 T (z)wm+1 T (w) . (2.55)
2πi 2πi 2πi 2πi
Now we deform the contour as indicated earlier, and substitute the operator product.
Now we have
I
dw dz n+1 m+1 c/2 2 1
z w + T (w) + ∂w T (w) , (2.56)
2πi 2πi (z − w)4 (z − w)2 z−w
where the z-contour goes around the singularity at z = w. Performing this integral using
Cauchy’s theorem, we get
I
dw h c n−2 m+1 n m+1 n+1 m+1
i
(n + 1)n(n − 1)w w + 2(n + 1)w w T (w) + w w ∂w T (w) .
2πi 12
Now one can integrate by parts in the last term, and apply Cauchy’s theorem once more,
this time around w = 0. This yields the right-hand side of (2.54). Not surprisingly, a
31
term proportional to c appears. If that term were absent, the quantum algebra would be
identical to the classical one. Strictly speaking, such a constant term is not allowed in an
algebra. The commutator of any two elements of the algebra must again be an element
of the algebra. We are thus forced to view c not as a number, but as an operator which
commutes with any element of the algebra. It follows then that on any representation
of the algebra this operator has a constant value, which is also denoted by c, just as the
operator itself. Such operators that appear only on the right hand side of commutators
are usually called central charges.
Note that the SL(2, C) subalgebra generated by L1 , L0 and L−1 is not affected by the
extra term. It remains thus meaningful to speak of the conformal weight of T (z).
Because of the central term the classical symmetry is not preserved in quantum me-
chanics. In particular, the central term prohibits the vacuum to have the full symmetry,
because we cannot impose the condition Ln |0i = 0 for all n, without getting a contra-
diction with the algebra. This is analogous to the position and momentum operators in
quantum mechanics, which also cannot simultaneously annihilate the vacuum.
Nevertheless we still have all the generators of the Virasoro algebra at our disposal,
and they still play a useful rôle. In those cases where conformal invariance is really crucial
this is not sufficient though. Presumably this is true in string theory, although there have
been attempts to make sense of it without conformal invariance. The simplest string
theory, the bosonic string, is constructed out of D free bosons, where D is the number of
space-time dimensions. One might think that this is always anomalous, because c = D
in this case. However, there is an additional ghost contribution (the ghost is related to
gauge fixing for two-dimensional gravity) of −26. This leads to the well-known concept
of a critical dimension D = 26.
3 Virasoro Representation Theory

Given any algebra, it is usually important to try and find its representations. The best
known example is probably the angular momentum algebra.
In that case all finite dimensional unitary representations are labelled by an integer
or half-integer j. The algebra consists of three generators, J − , J + and J3 . All states in
a representation are labelled by a J3 eigenvalue, which is lowered by J − and increased
by J + . The representation can be built up by starting with the state with maximal J3
eigenvalue, which is therefore annihilated by J + . Mathematicians call such a state the
highest weight state. The other states are obtained by acting on the highest weight state
(denoted |ji) with J − . This can only be done a finite number of times if j is integer or
half-inter, because one finds that the norm of the state (J − )2j+1 |ji is zero. Such states
are called null states or null vectors. The representation space is defined by setting such
states equal to zero.
This is the procedure we wish to mimic for the Virasoro algebra. In general, one starts
with determining a (preferably maximal) set of commuting operators (like J 2 and J3 for
angular momentum). A convenient choice is L0 and the central charge, c.
32
The Virasoro algebra has many more representations than will be considered here.
As is the case for SU (2), the representations of interest are those satisfying a number of
physically motived conditions. The ones we will consider here are the so-called unitary
highest weight representations.
3.1 Unitarity
A representation of the Virasoro algebra is called unitary if all generators Ln are realized
as operators acting on a Hilbert space, with the condition that L†n = L−n . The latter
condition implies in particular that T (z) is a Hermitean operator. This is most easy to
see on the cylinder, where we have, classically
X
1
2
[T11 (x0 , x1 ) + T00 (x0 , x1 )] = Ln e−in(x0 +x1 ) + L̄n e−in(x0 −x1 ) , (3.1)
n
and X
1
2
[T12 (x0 , x1 ) + T21 (x0 , x1 )] = Ln e−in(x0 +x1 ) − L̄n e−in(x0 −x1 ) , (3.2)
n
Reality of Tµν leads to the requirement that L∗n = L−n (and L̄∗n = L̄−n ), which naturally
leads to the quantum condition given above. On the complex plane the hermiticity con-
dition looks less natural, because the “in” and “out” states play an asymmetric rôle, and
also because we have complexified the coordinates.
In the following we will consider unitary representations. Non-unitary representations
have also been studied, in particular in statistical mechanics. Such representations still
consist of states in a Hilbert space (in particular having positive norm), but the require-
ment L†n = L−n is dropped.
3.2 Highest weight representations

By definition, a highest weight representation is a representation containing a state with
a smallest value of L0 . Not all representations have that property, but it is reasonable
to expect this in a physical theory, since L0 + L̄0 is the Hamiltonian, which is usually
bounded from below. The term “highest weight” for a state with lowest energy is perhaps
somewhat confusing, but has become standard terminology.
It follows from the structure of the algebra that Ln decreases the eigenvalue of L0 by
n,
L0 Ln |ψi = (Ln L0 − nLn ) |ψi = (h − n)Ln |ψi , (3.3)
if L0 |ψi = h |ψi.
If |hi is a highest weight state, then obviously |hi is annihilated by all generators Ln
with n > 0:
Ln |hi = 0, for n ≥ 1 (3.4)
Suppose the operator L0 acting on the highest weight state |hi creates a state |hi0 . Then
the Virasoro algebra tells us that Ln |hi0 = 0 for n ≥ 1, i.e. L0 maps highest weight
33
states to highest weight states. Since L0 is hermitean we can diagonalize it on the highest
weight states, so that we may assume that L0 |hi = h |hi (labelling the state only by its L0
eigenvalue is inadequate in case of degeneracies, but we will not worry about that now).
The negative modes Ln , n < 0 can be used to generate other states in the representa-
tion. Usually such states are referred to as descendants.
We have in fact already seen an example of a representation that is not a highest
weight representations, namely the adjoint representation, defined by the action of the
algebra on itself. The commutator
[L0 , Ln ] = −nLn (3.5)
tell us that in this representation the eigenvalue of L0 can take any integer value, whereas
[c, Ln ] = 0 tells us that the adjoint representation has central charge 0. It is in fact a
unitary representation.
3.3 The vacuum

The vacuum of the theory can be defined by the condition that it respects the maximum
number of symmetries. This means that it must be annihilated by the maximum number
of conserved charges. In the present context this means that we would like it to satisfy
Ln |0i = 0 for all n, but because of the central term that is obviously not possible. For
example, if L2 , L−2 as well as L0 annihilate the vacuum, so does the commutator of L2
and L−2 . But this is only consistent with the algebra if c = 0. We will soon see that
unitary conformal field theories with c = 0 are trivial.
The maximal symmetry we can have is
Ln |0i = 0 , for n ≥ −1 . (3.6)
Of course we could also have imposed this for n ≤ 1, but then |0i is a state with maximal
eigenvalue of L0 rather than one with minimal eigenvalue (a highest weight state).
Because of the commutator [L1 , L−1 ] = 2L0 any highest weight state which is anni-
hilated by L0 must be annihilated by L1 and L−1 (and vice-versa). It will always be
assumed that there is precisely one state in the theory that has these properties.
We also define its Hermitean conjugate h0|. It satisfies h0| Ln = 0 for n ≤ 1.
3.4 Positivity of c and h

The unitary highest weight representations are labelled by two real numbers, h and c.
Since all generators commute with c, it has a constant value on a representation. On
the other hand L0 does not have a constant value, but we can define h uniquely as its
eigenvalue on the highest weight state. With these two numbers given, we know the
Virasoro representation completely, since all states can be created by the action of the
Virasoro generators on the ground states, and since the norm of any state can be expressed
completely in terms of c and h. Hence any negative or zero norm condition depends only
34
on c and h. It follows that two representations with the same value of c and h are
equivalent as Virasoro representations.
In the rest of this chapter we will derive restrictions on c and h by requiring absence of
negative norm states. As a modest start we will show that these number are non-negative.
Note first of all the following commutator
c
[Ln , L−n ] = (n3 − n) + 2nL0 (3.7)
12
Hence we have
c
kL−n |0i k = h0| (L†−n L−n |0i = h0| Ln L−n |0i = h0| [Ln , L−n ] |0i = (n3 − n) (3.8)
12
For n ≥ 2 this implies that c ≥ 0 (this follows from the requirement that we work in
a Hilbert space, so that all states must have non-negative norm; furthermore zero norm
implies that the state vanishes.) Using the algebra just like we did for the vacuum we
find for any other highest weight state
c
kL−n |hi k = hh| Ln L−n |hi = hh| [Ln , L−n ] |hi = (n3 − n) + 2nh hh|hi (3.9)
12
We may assume that hh|hi = 6 0, since otherwise we would not consider |hi a state in our
theory. If the norm of the highest weight state does not vanish this tells us once again
that c ≥ 0 (since the first term dominates the second for large n), while for n = 1 we see
that either h > 0 or h = 0 and kL−1 |hi k = 0, i.e. |hi = |0i.
If c = 0 in a unitary theory the vacuum representation contains just one state, |0i
itself, since the foregoing argument shows that L−n |0i = 0 for n > 0. (To rule out non-
trivial representations with c = 0 and h > 0 requires a more sophisticated argument,
which we will present later.)
3.5 States and conformal fields

There is a simple connection between highest weight states and conformal fields. Consider
a conformal field φ(z, z̄) with weights h and h̄. Now define

h, h̄ = φ(0, 0) |0i , (3.10)
of course it is assumed that φ(z, z̄) |0i is well-behaved at the origin. Now compute
where
Ln h, h̄ . We find
I
1
[Ln , φ(w, w̄)] = z n+1 T (z)φ(w, w̄)
2πi
= h(n + 1)wn φ(w, w̄) + wn+1 ∂w φ(w, w̄) , (3.11)
which vanishes
if w = 0 and n > 0. Hence Ln , n > 0 commutes with φ(0, 0) and it follows
that h, h̄ is a highest weight state. It is also a highest weight state with respect to the

anti-holomorphic sector.
For n = 0 we find [L0 , φ(0, 0)] = hφ(0, 0), so that the state h, h̄ indeed has the
L0 -eigenvalue h, as the notation suggests.
35
3.6 Descendant fields
Ground states of Virasoro representations are generated from the vacuum by conformal
fields, which are also known as (Virasoro) primary fields. The addition “Virasoro” is
usually added in more general contexts, where other algebras are being considered. One
can then have a distinction between Virasoro primaries and primaries with respect to
other algebras. The name “conformal field” will be used here only in the strict sense of
Virasoro primary, which is equivalent to (3.11) being satisfied.
One can also consider fields that generate descendant states from the vacuum. They
are, quite naturally, called descendant fields. They can be defined by means of the operator
product with the energy momentum tensor
X
T (z)φ(w, w̄) = (z − w)k−2 φ(−k) (w, w̄) . (3.12)
k≥0
We may project out a term from this sum by

I
(−k) dz 1
φ (w, w̄) = T (z)φ(w, w̄) (3.13)
2πi (z − w)k−1
Clearly I
(−k) dz 1
φ (0, 0) |0i = T (z)φ(0, 0) |0i = L−k φ(0, 0) |0i , (3.14)
2πi (z)k−1

so that φ(−k) does indeed generated the L−k descendant of h, h̄ . To get descendant
states obtained by two Virasoro generators one has to consider operator products of T (z)
with φ(−k) , etc.
3.7 The Kac determinant

So far we have derived some necessary conditions for the positivity of norms of states.
But we have only looked at norms of states L−n |hi. At a given excitation level (i.e. at a
given L0 eigenvalue n + h ) there are in general many other descendants, which are linear
combinations of states X
L−n1 . . . L−nk |hi , ni = n . (3.15)
i
Because of the commutation relations of the Virasoro algebra we may in fact assume that
the generators are ordered, ni ≥ nj if i < j, since any incorrectly ordered product can be
expressed in terms of ordered ones. The collection of states (3.15) for all n ≥ 0 is called
the Verma module of |hi. Its definition does not make use of any norm on the space of
states. If one does have a norm, one can ask whether all states in the Verma module (i.e.
all linear combinations of the states (3.15)) have positive norm. In general that will not
be the case. Note that the set of states in the Verma module is closed with respect to the
action of the full set of Virasoro generators, i.e. acting with any Virasoro generator on
any state in the set produces a linear combination of states in the set. There is no need
36
to include positively moded Virasoro generators, since they can always be commuted to
the right where they annihilate |hi.
At the first excited level, the only state one can have is L−1 |hi. We have already seen
that this state has positive norm for h > 0 and norm zero for h = 0.
At the second level one can have L−2 |hi and (L−1 )2 |hi. It is not sufficient to check
whether each of these states separately has positive norm, because there could be linear
combinations that have zero or negative norm. To deal with this problem in general we
consider the matrix
!
hh| L†−2 L−2 |hi hh| L†−2 L−1 L−1 |hi
K2 = (3.16)
hh| (L−1 L−1 )† L−2 |hi hh| (L−1 L−1 )† L−1 L−1 |hi
This matrix is clearly Hermitean. Suppose it has negative or zero determinant. Then
there exists an eigenvector ~v = (α, β) with zero or negative eigenvalue, i.e. ~v K2~v T ≤ 0.
The left hand side is equal to the kαL−2 |hi + βL−1 L−1 |hi k, and we conclude that this
quantity is not positive.
At the nth level there is an analogous matrix Kn . The determinant of the matrix Kn
is called the Kac determinant. Of course it does not tell us precisely how many positive,
zero and negative eigenvalues there are. Even if det K > 0 there could be an even number
of negative eigenvalues. Usually one studies the behavior of the Kac determinant as a
function of parameters (such as h and c), starting in an asymptotic region where we know
that all eigenvalues are positive.
Of special interest are the null vectors, the eigenvectors of zero norm. The vanishing
of the norm corresponds to the equality vKn v T = 0, where ~v is a set of coefficients of the
basis states at level n. But if v is a vector in the zero eigenspace of the Hermitean matrix
Kn , it is clearly also true that wKn v T = 0 for any vector w, not just w = v. It follows
that the state defined by the vector v is orthogonal to any state at level n. Furthermore,
since the L0 eigenspaces are all orthogonal, it follows that a null state is orthogonal to any
other state in the Verma module. Then in particular, if |xi is a null state kLn |xi k = 0
for all n, since this relation can be interpreted as the orthogonality relation between |xi
and the Verma module state L−n Ln |xi. Thus the Virasoro generators take null states
to null states or annihilate them. If we act with positively moded Virasoro generators
it must happen that after a finite number of steps we encounter a state |xs i which is
annihilated by all positive Ln ’s, since |hi has positive norm. The state |xs i is at the same
time a descendant of |hi (as are all states in the Verma module) as a primary, since it is
annihilated by all positive Ln ’s. Such states are called singular states or (more frequently)
singular vectors. (Note that the definition of a singular vector (in contrast to a null vector)
does not require a norm.)
Consider now the states obtained by action with all Virasoro generators on the singular
state |xs i. Clearly they form a closed subset of the states in the Verma module of |hi.
This implies that if we remove all these states we will still have a non-trivial representation
of the Virasoro algebra. In other words, suppose that |xs i is generated from the ground
state |hi by a combination of Virasoro generators L, |xs i = L |hi. Then we can define
37
a representation of the Virasoro algebra on the subspace of the Verma module obtained
by removing |xs i (and all its descendants) by imposing the condition L |hi = 0. This
corresponds in SU (2) to the condition J − |−ji = 0 in a representation with spin j.
Obviously we can only remove a descendant if it has zero norm; otherwise the norms of
the left-hand side and the right-hand side of L |hi = 0 contradict each other.
We see thus that we can systematically remove all null states from the Verma module
by removing all sub-representations whose highest weight states are the singular vectors.
On the other hand, negative norm states cannot be removed. If we wish to obtain unitary
representations, we are obliged to consider only ground states |hi for which no negative
norm states appear at all. This turns out to be very restrictive, at least for c < 1.
3.8 The Kac determinant at level 2

The first evidence for that is seen at the second level. The explicit expression for K2 is
!
4h + 12 c 6h
K2 = hh|hi (3.17)
6h 4h + 8h2
For large values of c and h the diagonal terms dominate, and the eigenvalues are positive.
The determinant is
det K2 = 2 16h3 − 10h2 + 2h2 c + hc hh|hi2 .

(3.18)
This can be written as
det K2 = 32(h − h11 )(h − h12 )(h − h21 )hh|hi2 , (3.19)
where we introduce for future purposes
[(m + 1)p − mq]2 − 1

hpq = , (3.20)
4m(m + 1)
with r
1 1 25 − c
m=− ± (3.21)
2 2 1−c
Note that choosing the + or the − sign has the effect of interchanging m with −m − 1,
which amounts to interchanging p and q.
The determinant is proportional to h (in the second form this is slightly less manifest,
but note that h11 = 0). This is due to that fact that the norm of L−1 |hi is proportional
to h. Any state built on top of L−1 |hi will have a norm proportional to the norm of
L−1 |hi, and hence det K2 is also proportional to h.
The vanishing lines in the (h, c) plane are shown in fig. 1 Note that the branches h1,2
and h2,1 join smoothly at c = 1. The branch h11 coincides with the c-axis, as explained
above. Since there are two positive eigenvalues for large c and h, we move into a region
38
h
1.2
1.1
0.8
0.6
(2,1)
0.4
0.2
(1,2)
c
0 0.2 0.4 0.6 0.8
(1,1) 11.
Figure 1: Vanishing lines for level 2.
with precisely one negative eigenvalue when we cross one of the lines. Apart from the
region h < 0, which was already ruled out, this eliminates the dashed area to the left of
the curve. If h and c are within that area the corresponding representation of the Virasoro
algebra has negative norm states. Points on the border of the two regions are acceptable
since we can remove the zero norm states present there.
From the second row of K2 (see (3.17)) we can read off that the null state – if it exists
– is of the form
3 2
|Ψnull i = L−2 − L |hi . (3.22)
2(2h + 1) −1
Indeed,
3
hh| L21 |Ψnull i = 6h − 2
(4h + 8h ) hh|hi = 0 . (3.23)
2(2h + 1)
and
1 3
hh| L2 |Ψnull i = (4h + c) − 6h hh|hi . (3.24)
2 2(2h + 1)
The latter gives us an expression for c in terms of h: c = 2h(5 − 8h)/(2h + 1). This is
precisely the null-vector curve in the c − h plane shown above.
39
3.9 The Kac determinant at level 3 and 4
At the third level we have to consider the states L−3 |hi, L−2 L−1 |hi and (L−1 )3 |hi, etc.
A general formula for the Kac determinant can be derived [?, 53], namely
det Kn = αn Πpq≤n (h − hpq )P (n−pq) hh|hin , (3.25)
The function P (N ) gives the number of partitions of N , i.e. the number of ways of
writing N as a sum of integers. For example P (0) = 1, P (1) = 1, P (2) = 2, P (3) = 3 and
P (4) = 5, etc. This is equal to the number of states at level N , including null states.
Figure 2 shows the curves for the third level, together with those for the second one.
At level 2 we had ruled out regions B and C. The Kac determinant at level 3 tell us that
22
2.
(3,1)
1.5
AA
11
1.
(2,1) (1,3)
0.5 B
C
C
(1,2)
0 0.2 0.4
central charge 0.6 0.8 1.
11
Figure 2: Vanishing lines for level 3.

1
regions A and C are ruled out. It says nothing about region B, since we have to pass two
vanishing curves to get there, so that the determinant is positive there (but of course it
is a fair bet that there are in fact two negative eigenvalues in this region). But region
B was already ruled out. The entire area to the left of the two curves contains negative
norm states, and is thus ruled out.
If we include higher levels we get the lines shown in fig. 3. Now an even bigger region
gets ruled out, but it should also be clear that the picture at level n always contains all
vanishing lines from lower levels. Exactly what is ruled out cannot be decided from these
pictures alone. One really has to analyse this level-by-level.
40
2.5
1.5
0.5
0 0.2 0.4 0.6 0.8 1
Figure 3: Vanishing lines for levels up to 30. Only the bottom part is shown. The red dots show the
minimal model values of c and h.
3.10 The discrete series of minimal unitary Virasoro models

Strictly speaking one can never exclude these lines by looking at the determinant alone.
A more detailed argument [18, 19], shows that of the entire region 0 < c ≤ 1, h ≥ 0 only
a discrete set of points remains. These points are at the following c and h values:
6
c=1− , m≥3 (3.26)
m(m + 1)
with
[(m + 1)p − mq]2 − 1
h= , p = 1, . . . , m − 1, 1 ≤ q ≤ p . (3.27)
4m(m + 1)
The last formula looks quite similar to that for the vanishing curves, whereas the inverse
of the first formula gives m in terms of c exactly as in (3.21). The main differences are
that m is now restricted to integer values and that the range of p and q is limited. This
result implies that these values of h and c occur on an infinite number of vanishing lines,
i.e. they are intersection points of an infinite number of lines. The first such intersections,
occurring for c = 21 and h = 16 1
and h = 21 , can be seen at level 3 and 4. In fig. 3 many
other values are shown.
41
For c = 0 most h-values are now eliminated, except for a few discrete points where
the vanishing lines reach the h axis. These can be taken care of by considering the set of
states L−2n |hi and L2−n |hi for sufficiently large n [30].
Arguments of this kind can of course only rule out points. To show that conformal field
theories with these representations actually exist, the easiest thing to do is to construct
examples. We will return to this later.
For c ≥ 1 unitary representations exist for any positive value of h. For integer values
of c, c = N , it is quite easy to construct such representations explicitly using free bosons.
4 Correlation Functions
The objects we want to calculate in field theory in general, and in conformal field theory in
particular are the correlation functions (it is common practice to use statistical mechanics
terminology here; in field theory language we would speak of Green’s functions). If we
know all correlation functions, we can say that we have completely solved the theory; we
are then able to compute any scattering amplitude. In four-dimensional field theory this is
a very hard problem that we can only address in perturbation theory. In two-dimensional
field theory we can go much further.
In path-integral formulation we are interested in expressions of the form
Z
Dφ O1 (φ(z1 )) . . . On (φ(zn ))e−S(φ) , (4.1)
where φ stands generically for any field in the theory (possibly including ghosts), Oi is
some function of the fields, and S is the action, which one continues to Euclidean space
to improve convergence.
When computing such an integral one has to specify the two-dimensional surface on
which the fields φ live. This can be the plane, but it can be any other two-dimensional
surface as well. Using conformal transformations, we can transform the metric of any
such surface to δij in a finite neighborhood of any point. However, in general this does
not work globally because a surface can have a non-trivial topoogy. In two dimensions,
there is a complete classification of the different topologies one can have, the theory of
Riemann surfaces. They are classified in terms of a single number, the Euler index, or the
genus. The genus g simply counts the number of handles on the surface, with the sphere
having g = 0, the torus g = 1, etc (the Euler index χ is equal to 2(1 − g)). The cylinder
is a surface with boundaries, but if we make it infinitely long and at the points at ∞
and −∞ we may think of it topologically as a sphere (with two special points). Similarly
the complex plane is topologically a sphere, if we add the point |z| = ∞. If one tried to
do any of this in four dimensions one would quickly be lost, since there does not exist a
corresponding classification of four-manifolds. Nearly all four-dimensional field theory is
done on the four-dimensional plane. The possible rôle of other topologies and even how
to take them into account properly is still very poorly understood.
In statistical mechanics one usually considers only correlation functions on the plane
and the torus. In any case the two-dimensional topology is fixed, and determined by the
42
problem one is studying. If one imposes periodic boundary conditions in space and time
directions, one works on the torus. The correlation functions one computes are directly
related to quantities one measures in experiments.
In string theory the computation of two-dimensional correlation functions is part of the
computation of scattering amplitudes in space-time. The prescription (due to Polyakov
[43]) is to sum over all two-dimensional surfaces that satisfy given boundary conditions.
These boundary conditions are a consequence of the external particles for which one wants
to compute the scattering amplitude.∗ The surfaces of interest have a certain number of
handles, with tubes sticking out that correspond to the external particles. If we propagate
these particles to infinity, and project on a single particle, we may replace these external
lines by single points, just as we did in mapping the cylinder to the sphere. The process
of interest is then described by (4.1), where the functions “Oi ” operators describe the
emission of a certain particle state from the point zi on the surface. The corresponding
operators are known as vertex operators.
The topology of the surface corresponds to the order of string perturbation theory.
The sphere gives us all tree diagrams, the torus all one-loop diagrams, etc. Note that
there is only one diagram for each order of perturbation theory. To get the full space-time
scattering amplitude to arbitrary order in perturbation theory, we have to sum first over
all topologies, and then integrate over all different surfaces of given topology, as well as
over the points zi . These integration variables are called the moduli of the surface.
4.1 Correlation functions on the Riemann sphere

Now we turn to the simplest surface, namely the sphere. As before, we represent it as
the complex plane, with infinity added as a single point. This is known as the Riemann
sphere.
In this case the path-integral can be expressed as a vacuum-to-vacuum amplitude, or
vacuum expectation value,
h0| O1 (φ(z1 , z̄1 )) . . . On (φ(zn , z̄n )) |0i , (4.2)
where Oi are the quantum mechanical operators representing the functions in (4.1), and
radial ordering is implicitly understood. The relation between (4.1) and (4.2) is completely
analogous to the more familiar relation in field theory between the path integral and time-
ordered perturbation theory.
Conformal invariance puts strong constraints on correlators. Let us first consider
correlation functions of primary fields (here we omit for simplicity the dependence on z̄i )
h0| φ(z1 ) . . . φ(zn ) |0i , (4.3)

∗
There are also open string theories, which have non-trivial boundary conditions in the spatial direction
in two dimensions. They are thus defined on two-dimensional strips instead of the cylinder. When they
interact they may form, under certain circumstances, non-orientable surfaces like Möbius strips. These
theories are not considered here; we restrict ourselves to closed, orientable manifolds.
43
and investigate the consequence of invariance under the SL2 (C) subgroup of the conformal
group. We focus on this subgroup first to distinguish the extra information we get in two
dimensions from that of conformal invariance in arbitrary dimensions. We have
h0| Li = h0| L†i = h0| L−i = 0, for i = 0, ±1 (4.4)
Therefore we can derive (for i = 0, ±1)
0 = h0| Li φ(z1 ) . . . φ(zn ) |0i

X
= h0| φ(z1 ) . . . φ(zj−1 ) [Li , φ(zj )] φ(zj+1 ) . . . φ(zn ) |0i + h0| φ(z1 ) . . . φ(zn )Li |0i
j
The last term vanishes, and the commutator with Li generates the infinitesimal conformal
transformation δi . Hence we get
X
h0| φ(z1 ) . . . φ(zj−1 )δi φ(zj )φ(zj+1 ) . . . φ(zn ) |0i = 0 (4.5)
j
Now we may use the fact that the fields φi are conformal fields. Then
δ φ = ∂φ + h∂φ (4.6)
If we restrict Li to SL2 (C), can be 1, z or z 2 .
4.2 Two-point functions

Consider for example the two-point function (propagator)
G(z1 , z2 ) = hφ1 (z1 )φ2 (z2 )i (4.7)
We find that this function satisfies the differential equation
[(z1 )∂1 + h1 ∂(z1 ) + (z2 )∂2 + h2 ∂(z2 )] G(z1 , z2 ) = 0 , (4.8)
with as above. Let us look at each of these choices. The case i = −1 ( = 1) yields the
equation
(∂1 + ∂2 )G(z1 , z2 ) = 0 , (4.9)
which implies that G depends only on the difference of z1 and z2 , and not on the sum.
Hence G(z1 , z2 ) ≡ G(x), x = z1 − z2 . Then the equation with i = 0 can be written as
[x∂x + h1 + h2 ] G(x) = 0 . (4.10)
The solution is (up to a normalization)
G(x) = x−h1 −h2 . (4.11)
44
Finally we can substitute this solution in the equation for i = 1. This yields
(h1 − h2 )(z1 − z2 )G(z1 , z2 ) = 0 , (4.12)
so that h1 must be equal to h2 , or else the propagator vanishes. The final result is thus
that
G(z1 , z2 ) = C(z1 − z2 )−2h , h = h1 = h2 . (4.13)
As usual, we have dropped the anti-holomorphic part, which would have given rise to an
additional factor (z̄1 − z̄2 )−2h̄ . The coefficient C has no physical relevance, as it can be
set to 1 by changing the normalization of the primary fields.
4.3 Three-point functions

To get the three-point function
(3)
Gijk = h0| φi (z1 )φj (z2 )φk (z3 ) |0i (4.14)
we argue in a similar way. Translation invariance shows that it must be a function of the
differences zij = zi − zj (this holds in fact for an arbitrary n-point function. Rotation
(L0 ) invariance leads to the equation
[z1 ∂1 + z2 ∂2 + z3 ∂3 + h1 + h2 + h3 ] G(3) (z12 , z23 ) = 0 (4.15)
The correlator is a function of two independent variables instead of three, since z13 =
z12 + z23 . If we write the solution as
X
G(3) (z12 , z23 ) = a b
Dab z12 z23 , (4.16)
ab
we find the condition a + b = −h1 − h2 − h3 . Finally we use L1 . The solution is

(3) h3 −h1 −h2 h1 −h2 −h3 h2 −h3 −h1
Gijk (z12 , z23 , z13 ) = Cijk z12 z23 z31 , (4.17)
where we have introduced the redundant variable z13 to get the solution in a more sym-
metric form. The coefficients Cijk depend on the normalization of the two-point function,
and we will assume that the latter has been set equal to 1.
In more detail: the equation is
[a(z1 + z2 ) + b(z2 + z3 ) + h1 z1 + h2 z2 + h3 z3 ] G(3) (z12 , z23 ) = 0 (4.18)
Substituting b = −h1 − h2 − h3 − a, and using the ansatz given above we find (with
h = h1 + h2 + h3 and Da = Dab ):
X
a −h−a
G(3) (z12 , z23 ) = Da z12 z23 , (4.19)
a
45
with coefficients that must satisfy
X
a −h−a
[(a + 2h1 )z12 + (a + h1 + h2 − h3 )z23 ] Da z12 z23 =0 (4.20)
a
so that we need
X
a+1 −h−a
[(a + 2h1 )Da + (a + 1 + h1 + h2 − h3 )Da+1 ] z12 z23 =0, (4.21)
a
after shifting the sum from a to a+1 in the second term. This leads to a recursion relation
for Da :
a + 2h1
Da+1 = Da (4.22)
a + 1 + h1 + h2 − h3
Again one should multiply this expression with the anti-holomorphic factors.
The foregoing results can be understood as follows. We have three complex trans-
formations at our disposal. Using translations, we can move one of three variables z1 , z2
and z3 to any desired point in the complex plane, for example z1 = 0. Then, keeping
this point fixed we can use the second symmetry (scalings plus rotations, generated by
L0 and L̄0 ) to move z2 to any desired point, and finally we can do the same with z3 using
L1 and L̄1 , the special conformal transformation. Actually one can do this separately
for the holomorphic and the anti-holomorphic variables if one allows separate complex
transformations for each. Then it is simply a matter of requiring that
az1 + b az2 + b az3 + b
= α1 , = α2 , = α3 , (4.23)
cz1 + d cz2 + d cz3 + d
where α1 , α2 and α3 are three fixed points in the complex plane. Often one chooses
z = 0, z = 1 and z = ∞ for these points. These three equations for the four complex
variables a, b, c, d subject to the determinant condition ad − bc = 1 have a solution if all zi
are different. Hence the entire answer is determined if we know the three point function
in just three points.
4.4 Four-point functions

This tells us immediately that it cannot work for the four-point function. Indeed, the
best one can do in that case is
Y −h −h +h/3 −h̄ −h̄ +h̄/3
G(4) (zi , z̄i ) = f (x, x̄) zij i j z̄ij i j , (4.24)
i<j
where f (x) is a function of

z12 z34
x= , (4.25)
z13 z24
which is the only independent ratio which is invariant under the conformal group.
46
4.5 Conformal Ward identities
So far we have concentrated on the subgroup SL2 (C). The generalization of the foregoing
discussion
H is obtained by inserting the generator of infinitesimal conformal transformations
dz(z)T (z) into the correlation function:
I
dz
h0| (z)T (z)φ(w1 ) . . . φ(wn ) |0i , (4.26)
2πi
where the contour encircles all the points wi . By a suitable analytic continuation the
contour can be chosen in such a way that the point z = 0 is avoided.
w1 w1
w3 w3
w2 w2
We can deform the contour to encircle each wi separately; then we get

I
X dz
h0| φ(w1 ) . . . (z)T (z)φ(wi ) . . . φ(wn ) |0i
i
2πi
X
= h0| φ(w1 ) . . . δ φ(wi ) . . . φ(wn ) |0i ,
i
where the variation of each field φ(wi ) with respect to is represented by a contour integral
around wi (cf. (2.20), 2.24)).
Since the foregoing holds for any we may omit the integral; then we get
h0| T (z)φ(w1 ) . . . φ(wn ) |0i

X hi 1 ∂

= + h0| φ(w1 ) . . . φ(wn ) |0i .
i
(z − wi )2 z − wi ∂wi
4.6 Correlators of descendants

Consider the correlator
h0| φ1 (w1 ) . . . φn−1 (wn−1 )φ(−k)

n (wn ) |0i , (4.27)
47
where φ(−k) is the k th descendant of φ. Inserting the definition (3.14) and using the
conformal Ward identity we may rewrite the result as
I
dz 1
h0| T (z)φ1 (w1 ) . . . φn (wn ) |0i
2πi (z − wn )k−1
n−1
X hi 1 ∂
− + h0| φ(w1 ) . . . φ(wn ) |0i
i=1
(z − w i )2 z − w i ∂w i
In the first term the contour runs outside all points wi , and hence we can deform it to
infinity. Then the integral can be transformed toP a contour integral around the point
z = ∞, To do this explicitly, we can write T (z) = n Ln z −n−2 . Then we transform z to
w = z1 . This yields
I
dz 1
Ln z −n−2
2πi (z − wn )k−1
wk−1
I
dw 1
= − 2 Ln wn+2
2πi w (1 − wwn ) k−1
I
dw 1
= − Ln wn+k−1
2πi (1 − wwn ) k−1
Since h0| Ln = 0 for n ≤ 1 this integral has no singularity inside its contour, and hence it
vanishes. In other words, the contour can be pulled off the back of the Riemann sphere.
The other terms can be evaluated as follows. The term
I
i dz 1 hi 1 ∂
L−k ≡ − + (4.28)
w 2πi (z − wn )
k−1 (z − w )2 z − wi ∂wi
i
is a contour integral around the point wi . Using the standard formula

I
dz 1 1
f (z) = ∂ n−1 f (w) (4.29)
w 2πi (z − w) (n − 1)!
n
we get
(1 − k)hi 1 ∂
Li−k = − + . (4.30)
(wi − wn )k (wi − wn )k−1 ∂wi
Introducing the operator
n−1
X
L−k ≡ Li−k (4.31)
i=1
we can write the correlation function as
h0| φ1 (w1 ) . . . φn−1 (wn−1 )φ(−k)

n (wn ) |0i = L−k h0| φ1 (w1 ) . . . φn−1 (wn−1 )φn (wn ) |0i
In other words, the correlator of descendant fields can be expressed entirely in terms of
the correlator of primaries. The same is true for more complicated descendants or for cor-
relators involving more than one descendant, but the formulas become more complicated.
48
This is the extra information we get from the additional conformal symmetry specific to
two dimensions. As we have seen, SL2 (C) determines the two-point-functions completely,
the three-point functions up to constants Cijk , the four-point functions up to a function,
etc. In higher-dimensional conformal field theories one has an analogous symmetry, and
analogous restrictions. If we had only SL2 (C) at our disposal, there would still be some
relations between correlation functions, but clearly far more limited, since we could only
use the Virasoro generators L0 , L±1 . Sometimes field that transform like conformal fields
under SL(2, C) (or the conformal group in higher dimensions) are called quasi-primary.
Descendant fields can sometimes be quasi-primary. An example is the energy-momentum
tensor. Due to the conformal anomaly it is not a primary field, but since the anomaly
does not affect SL2 (C) it is a quasi-primary field.
The extra Virasoro generators beyond SL2 (C) allow us to organize the fields in the
theory into much larger sets, each consisting out of one primary field and in infinite
number of descendants. The task of computing the correlation functions reduces to doing
so for correlation functions of primary fields.
4.7 Null state decoupling

There are further constraints on the correlation functions, but they are dependent on the
representation we are considering, and do not hold generically.
The foregoing results have important implications in the special case that a descendant
of a primary field is a vector of norm zero, a null state. Since this is not a state in the
theory, it must decouple from all physical amplitudes. For example, we have found in
the previous chapter that certain Virasoro representations – namely those whose c and h
values fall on the second level vanishing curve – have a second level null state

3 2
L−2 − L |hi = 0 . (4.32)
2(2h + 1) −1
It follows that any amplitudes involving the corresponding descendant field must vanish.
The descendant field is (here φ is a conformal field with weight h)
3
φ(−2) − ∂ 2φ , (4.33)
2(2h + 1)
where we have made use of the fact that the first descendant of a field is simply the
derivative. Hence we get
X
hj 1 3 2
+ ∂j − ∂ h0| φ1 (w1 ) . . . φn (wn ) |0i = 0 ,
j6=i
(wi − wj )2 wi − wj 2(2hi + 1) i
where the first two terms come from L−2 , and where we are assuming that the field φi
has a second level null vector. The third term may be expressed in terms of (L−1 )2 using
the results of the previous subsection, but that does not simplify the answer. This is a
differential equation that any n-point function involving φi has to satisfy. Obviously there
49
are higher order differential equations for any other singular state. Thus we see that all
correlation functions satisfy a huge number of differential equations.
It is instructive to verify that the two-point and three point functions do indeed satisfy
the equation that follows from the second level descendants.
Note that we have these strong constraints only if there are null states in the Virasoro
representations i.e. only if 0 < c ≤ 1.∗
4.8 Operator products

(3)
Consider the three-point function Gijk (z1 , z2 , z3 ) in the limit z1 → z2 . The leading term
is
(3)
h0| φi (z1 )φj (z2 )φk (z3 ) |0i = Gijk (z1 , z2 , z3 ) ≈ Cijk (z1 − z2 )h3 −h1 −h2 (z1 − z3 )−2h3 (4.34)
The last term looks like the propagator of the field φ3 , and the expression suggests that
the two primary fields φi and φj contain in their product the field φ3 , with strength Cijk .
The precise statement of this fact is the operator product expansion, which says that the
product of two operators Oi (x) and Oj (y) in field theory can be expanded in a complete
set of operators Ok (y) (without loss of generality we can choose our basis of operators at
the point y). X
Oi (x)Oj (y) = Cijk (x, y)Ok (y) . (4.35)
k
By translation invariance Cijk (x, y) = Cijk (x − y).
In conformal field theory we can take as the basis all primaries and a complete set of
descendants. Then the operator product expansion has the form
X
φi (z, z̄)φj (w, w̄) = Cijk (z − w)hk −hi −hj (z̄ − w̄)h̄k −h̄i −h̄j φk (w, w̄) . (4.36)
k
The sum on the right-hand side contains both primaries and descendants, and hk must
be interpreted as the ground state conformal weight plus the excitation level of the de-
scendant. If i, j, k are primaries, this expression agrees with the result of the three-point
function, and tells us in particular that the coefficients Cijk appearing in the operator
product and the three-point function must be the same. In fact, the operator product
holds also if all three fields are descendants. The behavior of the coefficients as a func-
tion of z and w is completely fixed by translation invariance and the scaling properties
(“dimensions”) of the fields involved. Both are good symmetries, even for descendants.
We have already encountered the operator product of the energy momentum tensor (a
descendant) with other fields.
Since correlation functions involving descendants are related to those of primaries, it
should not be surprising that one can do the same for operator products. Indeed, the
operator product of two primary fields may be written as
X pp̄
φi (z, z̄)φj (w, w̄) = Cijk (z − w)hk +np −hi −hj (z̄ − w̄)h̄k +np̄ −h̄i −h̄j φpp̄
k (w, w̄) , (4.37)
k,p,p̄
∗
The only exception is the null state L−1 |0i.
50
where φpp̄
k is a descendant of φk at level (np , np̄ ). One can show that the operator product
pp̄
coefficients Cijk can be expressed in terms of those of the primary fields,
pp̄ k(p) k(p̄)

Cijk = Cijk βij β̄ij , (4.38)
where the coefficients β are determined completely by conformal invariance.
4.9 Duality
The operator product is a useful tool for the computation of four-point functions. How-
ever, if we consider the correlator
h0| φi (z1 , z̄1 )φj (z2 , z̄2 )φk (z3 , z̄3 )φl (z4 , z̄4 ) |0i (4.39)
it is not obvious for which pairs of fields we should compute the operator product first.
If we combine the fields in pairs there are three alternatives, namely (i, j)(k, l), (i, k)(j, l)
or (i, l)(j, k). The result should not depend on how we perform the calculation. This is
known as duality (as are many other symmetries). Diagrammatically we may represent
this as
j k
j k j k
m
m m
i l i l
i l
where in all three cases a sum over the intermediate states is understood. To write the
equations belonging to this picture one defines the conformal blocks. First we simplify the
four-point function by fixing all but one coordinate using SL2 (C)
Fijkl (z, z̄) = h0| φi (z, z̄)φj (0, 0)φk (1, 1)φl (∞, ∞) |0i . (4.40)
If we compute this correlator by making the contractions (i, j)(k, l) we can write the result
as X
m m
Fijkl (z, z̄) = Cijm Cmkl Fijkl (z)F̄ijkl (z̄) , (4.41)
m
where the sum is over all primary fields. All descendant contributions are taken into
account by the functions F, called the conformal blocks. Since the Virasoro algebra acts
chirally (i.e. splits completely in holomorphic and anti-holomorphic generators) we can
factorize the contribution of each primary into a holomorphic and an anti-holomorphic
contribution, as shown.
Duality imposes the following condition:
51
X
m m
Cijm Cmkl Fijkl (z)F̄ijkl (z̄)
m
X
m m
= Cikm Cmjl Fikjl (1 − z)F̄ikjl (1 − z̄)
m
m 1 m 1
X
= z −2hi z̄ −2h̄i Cilm Cmkj Filkj ( )F̄ilkj ( )
m
z z
Note that to go from the first to the second (third) line one has to interchange j and k
(l). This implies that the choice of the fixed points 0, 1 and ∞ does not agree anymore
with the convention chosen in (4.40), and we have to make a conformal transformation
to get back to our conventional coordinates. In going from the first to the second line
that transformation is z → 1 − z, which interchanges 0 and 1 while leaving ∞ fixed.
In going from the first to the third we use z → z1 , which interchanges 0 and ∞ while
leaving 1 fixed. Inevitably these transformations also act on the fourth argument, z, and
furthermore they introduce some conformal factors.
Once the conformal blocks are known, these equations impose strong constraints on
the operator product coefficients Cijk . The conformal blocks are constraint by decoupling
equations for null vectors (assuming there are null vectors). In principle it may be possible
to use these constraints to determine the conformal blocks as well as all the operator
product coefficients completely. In practice this is quite hard, although the computation
can be carried out completely for the simplest conformal field theory, the one at c = 12
(the critical Ising model).
5 Conformal Field Theory on a Torus

Up to now we have seen that a conformal field theory is described algebraically by a set
of ground states
h, h̄ = φh,h̄ (0, 0) |0i
on which the left- and right Virasoro algebra acts.
The question which we want to address now is: which combinations of ground states
can actually occur in a conformal field theory. Once we know that we have specified the
set of physical states in the theory completely: they consist of the ground states plus
all the descendants generated by the generators of the Virasoro algebra, minus all null
vectors.
5.1 Parametrization of the torus

Intuitively one would expect all the states in a theory to contribute to loop diagrams.
Hence loop diagrams should be a useful tool to answer this question. For this reason we
are going to study conformal field theory on the simplest loop diagram, the torus.
52
The torus is a cylinder whose ends have been sewn together, as shown in the “artists
impression” above. The most convenient mathematical description of the torus is in terms
of the complex plane modulo a lattice, as shown below
Im
0 1 Re
The meaning of this picture is that all points in the complex plane that differ by a linear
combination of the two basic lattice vectors are considered identical. Identification along
the real axis has the effect of rolling up the complex plane to a cylinder; then identification
along the vector labelled τ rolls up the cylinder to a torus.
A lot of symmetries have been taken into account already in arriving at the picture.
First of all two-dimensional general coordinate invariance (also called reparametrization
invariance or diffeomorphism invariance) has been used to “straighten” the coordinates
so that we get a lattice; rotational invariance has been used to make one direction point
along the real axis; translation invariance to put a point of the lattice at the origin and
global scale invariance to make the horizontal lattice spacing equal to 1. This means that
finally the entire lattice is described by one complex number, τ , which can be chosen in
the upper half plane.
5.2 The partition function

We would like to compute the path integral
Z
Dφe−SE (φ) , (5.1)
53
where SE is the Euclidean action of a given field configuration on the torus, and the
integral is over all field configurations. This notation is only symbolic. If we have a
Lagrangian description of a conformal field theory, φ stands for all fields in the theory.
For example, we could consider free bosonic theories, and in that case the integral is over
the bosonic fields Φi . However, in many cases such a Lagrangian formulation is either
not available, or not practically usable. It is therefore more convenient to express the
path-integral in terms of the Hamiltonian of the theory. The Hamiltonian is (related to)
L0 + L̄0 , and this is a quantity we always have at our disposal in a conformal field theory.
In ordinary quantum dynamics (in 0+1 dimensions) one can derive the formula
Z
Dqe−SE (q) = Tre−βH . (5.2)
PBC
Here SE is some Euclidean action that satisfies certain conditions; those conditions are
satisfied for example for the harmonic oscillator. The integral is over all paths q(t) that
start at t = 0 and end at t = β, subject to the periodic boundary condition (“PBC”)
q(0) = q(β). The derivation can be found in many text books on path integrals.
The path integral we are considering is not in 0+1 dimensions, but in 1+1 dimensions.
Hence the integration variables φ have an extra continuous label x1 . However this is merely
a generalization of (5.2) from one degree of freedom (q) to infinitely many (φ(x1 )).
In the lattice description of the torus we regard the real axis as the x1 direction, and
the imaginary axis as the Euclidean time direction. If Reτ = 0 we would find for the path
integral (5.1), as a straightforward generalization of (5.2)
Z
Dφe−SE (φ) = Tre−2πIm τ H .
The only small subtlety here is the factor 2π. It appears because the torus as depicted
earlier has a periodicity 1 rather than 2π along the x1 direction. To make contact with
earlier conventions we had to scale up the entire lattice by a factor of 2π, so that we get
2πτ instead of τ .
What happens if Reτ 6= 0? In that case we have to twist the torus before gluing it
together again, and the periodic boundary conditions in the Euclidean time direction are
defined including such a shift. The operator performing such a shift in the x1 direction is
the momentum operator P . A shift by an amount 2πReτ is achieved by the operator
eiP (2πRe τ )
The correct result is obtained by inserting this shift operator in the trace, so that we get
Z
Dφe−SE (φ) = Tre−2πIm τ H eiP (2πRe τ ) . (5.3)
The operators H and P are the time and space translation operators on the cylinder, and
they can be derived from the energy-momentum tensor.
54
5.3 The cylinder versus the Riemann sphere
Up to now we have defined quantum conformal field theory, and in particular the energy-
momentum tensor, on the complex plane (Riemann sphere), related to the cylinder via
a conformal mapping. We now want to examine carefully how the energy-momentum
tensor defined on the complex plane is related to that on the cylinder.
To do so we need the transformation of the energy momentum tensor itself under
conformal transformations. Since it is not a conformal field, we have to work this out
explicitly. Infinitesimal conformal transformations are generated by
I
1
Q = dz(z)T (z) ,
2πi
as we have seen before. The infinitesimal conformal transformation of the energy momen-
tum tensor is thus
I
1
δ T (w) = [Q , T (w)] = dz(z)T (z)T (w) .
2πi
Into this expression we insert the operator product of T (z) and T (w). The result is
c 3
δ T (w) = (w)∂T (w) + 2∂(w)T (w) + ∂ (w)
12
The global form of this transformation is
c
T (w) → (∂f )2 T (f (w)) + S(f, w) (5.4)
12
c
The factor multiplying 12
is known as the Schwartzian derivative,
∂f ∂ 3 f − 23 (∂ 2 f )2
S(f, z) = .
(∂f )2
It is easy to see that its infinitesimal form is indeed ∂ 3 (just substitute f (w) = w + (w)),
but the exact expression is a bit harder to understand. Note however the following. If we
apply a second conformal transformation, w → g(w), we get in the first step (5.4), and
after the second step
h c c i
(∂w g(w))2 (∂g f (g))2 T (f (g(w))) + S(f (g), g)] + S(g(w), w)
12 12
On the other hand, defining the function h = f ◦ g (which means h(w) ≡ f (g(w))) we
would expect to get exactly (5.4) with f replaced by h. For the terms involving T this
is manifestly true, since ∂w g(w)∂g f (g) = ∂w f (g(w) = ∂w h(w), but for the constant terms
we get the non-trivial condition
S(f ◦ g, w) = (∂w g(w))2 S(f (g), g) + S(g(w), w)
55
One may check that the Schwartzian derivative does indeed satisfy this condition.
We are now ready to apply this to the map from the plane to the cylinder. The map
we will use is w = eiz , as in Eqn. (2.4). Hence w is the plane coordinate and z the cylinder
coordinate. We get
c h ci
Tcyl (z) = [∂z w(z)]2 T (w(z) + S(w, z) = − w2 T (w) − (5.5)
12 24
Now we can substitute the mode expansion for T (w) on the plane
X
T (w) = w−n−2 Ln .
n
Then we find " #

X c
Tcyl (z) = − e−inz (Ln ) − ,
n
24
and analogously for the anti-holomorphic component.
Our next task is to find the precise definition of H and P . They are derived from the
energy-momentum tensor on the cylinder in the following way∗
Z Z Z
1 1 M 1 1 E 1
H= dx T00 = dx T22 = − dRez[T cyl (z) + T cyl (z̄)]
2π 2π 2π
and Z Z Z
1 1 M 1 1 E 1
P = dx T01 = dx (−iT21 ) =− dRez[T cyl (z) − T cyl (z̄)]
2π 2π 2π
Hence we find
c c̄
H = L0 −
+ L̄0 − ,
24 24
where we have allowed for the possibility that the holomorphic and anti-holomorphic
components have different central charges. For the momentum we find
c c̄
P = (L0 − ) − (L̄0 − ) .
24 24
Note that H and P do not generate space and time translations on the plane. The latter
would in fact be generated by L−1 and L̄−1 . What we have found is that H is proportional
to the dilatation operator on the plane, while P is proportional to the rotation operator.
This makes perfect sense, because D corresponds to radial time, whereas rotations around
z = 0 correspond to translations on the cylinder.
Substituting this into (5.3) we get
Z
c c̄
Dφe−SE (φ) = Tre2πiτ (L0 − 24 ) e−2πiτ̄ (L̄0 − 24 ) ≡ P (τ, τ̄ ) .
∗
For a generic rank 2 tensor X00 = −X22 and X01 = iX21 . Here there is an additional − sign because
the Euclidean action is defined with an extra − sign w.r.t. the Minkowski action. The Euclidean and
Minkowski energy momentum tensors T E and T M are both derived from these respective actions using
formula (1.2).
56
The right hand side of this expression will serve as the definition of the partition function
for general conformal field theories. As promised, this expression does not require a
Lagrangian formulation, and only uses the Virasoro generators themselves. The trace is
over all states in the Hilbert space (i.e. not including zero norm states).
5.4 Virasoro characters

The partition function can be expressed in terms of primary fields and descendants, in
the following way X
P (τ, τ̄ ) = Mij Xi (τ )Xj (τ̄ ) . (5.6)
i,j
Here i and j label a certain highest weight states |i, ji. The label i is used for representa-
tions of the holomorphic algebra, and j for the anti-holomorphic algebra. The multiplicity
of such a state is Mij , a non-negative integer. The functions X are the (Virasoro) char-
acters of the representation. They are defined as
c
Xi (τ ) = Trdescendants of i e2πiτ (L0 − 24 )
Thus the trace is over all (positive norm) states in the highest weight representation
labelled i. If we know the content of the representation, the missing information is thus
contained in the integers Mij .
5.5 Modular invariance

Up to now we had defined the torus in terms of a lattice. This lattice was defined by two
basis vectors, corresponding to the points “1” and “τ ” in the complex plane. However,
the same lattice – and hence the same torus – can be described just as well by choosing
different basis vectors. For example the choice “1”, “τ + 1” clearly describes the same
lattice
Im
τ τ+1
0 1 Re
57
This can be generalized further. One should keep in mind that the torus was defined by
rotating one basis vector along the real axis in the complex plane, and scaling it to 1.
The choice of this basis vector is free; we can also choose the direction “τ ”. This has the
effect of interchanging the two basis vectors. This rotation, combined with a rescaling
the new basis vector along the real axis, has the effect of replacing τ by − τ1 . This is
most easily illustrated by taking τ purely imaginary, as shown in fig. 4. The set of such
⌧
Rotated
1
⌧ Rescaled
Figure 4: The transformation τ → − τ1 .
transformations of the torus forms a group, called the modular group. We have identified
two elements of that group, namely
T : τ →τ +1
1
S : τ →−
τ
It turns out that these two transformations generate the entire group. The most general
modular transformation has the form
aτ + b
τ→ , a, b, c, d ∈ Z; ad − bc = 1 .
cτ + d
This group is isomorphic to SL2 (Z)/Z2 . The group SL2 can be defined by the set of 2 × 2
matrices !
a b
(5.7)
c d
with determinant 1 (we have already encountered the group SL2 (C), where the matrix
elements are complex numbers). The group SL2 contains the element −1. In the modular
transformation this is indistinguishable from the identity, and for this reason the modular
group is actually isomorphic to SL2 (Z)/Z2 rather than SL2 (Z). One may check that the
modular transformations satisfy
(ST )3 = S 2 = 1 .
58
It is now natural to ask how the partition function behaves under transformations
in the modular group. If we start with a well-defined two-dimensional theory on the
torus, in which all fields are periodic along all cycles around the torus, the result of
the path-integral should not depend on how that torus was parametrized. Hence the
partition function should be invariant under modular transformations. For example, if
we compute the path integral for free bosons on the torus, we will automatically get a
modular invariant partition function.
5.6 Modular transformations of the characters

On the other hand, if we just choose some multiplicities in (5.6) we will in general not get
a modular invariant partition function. To verify if a partition function written in terms
of characters is modular invariant we need to know how the characters transform. This
is easy for the transformation T :
c
Xi (τ + 1) = e2πi(hi − 24 ) Xi (τ )
This is often written in matrix form,

X
Xi (τ + 1) = Tij Xj (τ ) ,
j
where T is a diagonal matrix of phases. The transformation S is much harder to compute.

However, it is clear that such a matrix must exist, because we are simply writing the theory
in terms of a different basis. The result is
1 X
Xi (− ) = Sij Xj (τ ) ,
τ j
where S is a unitary and symmetric matrix.
5.7 Conditions for modular invariance

The conditions for modular invariance of the partition function can now be phrased in
the following simple matrix form
[M, T ] = [M, S] = 0 ,
with
Mij ∈ Z, Mij ≥ 0 .
Furthermore this is usually supplemented with the additional physical requirement that
the vacuum is present in the theory, and is unique. If we label the vacuum by “0”, we get
thus the condition
M00 = 1 .
59
5.8 The diagonal invariant
These conditions have a trivial solution
Mij = δij .
This is called the diagonal invariant. Consider for example a conformal field theory with
central charge 12 . We have seen that at this value for c there are just three Virasoro
representations, with h = 0, 21 and 16 1
. A modular invariant partition
function of this
system is thus obtained by choosing the three ground states |0, 0i , 12 , 12 and 16
1 1

, 16 .
Corresponding to these ground states there are three primary fields that create the ground
states from the vacuum, often denoted as 1, ψ and σ respectively. We have now specified
the theory completely. It is known as the Ising model.
5.9 Integration over moduli

As remarked above, it is not hard to construct partition functions that are not modular
invariant. Usually these are rejected. We have already argued that they can not cor-
respond to well-defined two-dimensional theory on the torus. In string theory there is
another reason, namely that one has to integrate over the parameter τ . String pertur-
bation theory is a summation over all two-dimensional surfaces. This sum splits in a
sum over all different topologies, and integrals over all different shapes of surfaces with a
given topology, the moduli. In two dimensions the topology can be described by a single
parameter, the number of handles, or the genus. The genus is 0 for the sphere, 1 for the
torus, etc. At genus 1 there is one complex modulus, the parameter τ . The integral over
τ is not over the full positive upper half plane, but should be restricted to a region that
covers the set of distinct tori just once. An example of such a region is shown in fig. 5.
The entire upper half plane is covered with an infinite number of such regions of
different shapes and sizes. For example, the lower part of the strip − 12 ≤ Re τ < 12
contains an infinite number of such regions. The integral over τ should not depend on the
choice of the region, or otherwise the theory is not well-defined. If the theory is modular
invariant, this problem does not arise.
5.10 Operator products and locality

An important consequence of modular invariance is that operator products are local, i.e.
that it has no branch cuts as a function of z − w. This is true because one always finds
combinations like (z − w)h (z̄ − w̄)h̄ = |z − w|2h̄ (z − w)h−h̄ , with h − h̄ an integer.
This implies that integrals over the positions z and w around the cylinder (or along
cycles of the torus or higher Riemann surfaces) are well-defined. Such integral occur
always in the application to string theory, where one has to integrate not only over the
moduli of the Riemann surface, but also over the positions of the vertex operators. If the
operator product were not local, there would inevitably exist correlation functionsH which√
have branch cuts in some of their variables, and one would encounter integrals like dz z
60
-1 - 12 1
2
1
Figure 5: Modular group domains in the complex upper half plane.
around z = 0. Since the integrand is not periodic around the origin, the integral depends
on the choice of the beginning and the end of the interval, and is thus not defined.
In applications to statistical mechanics the existence of such branch cuts is less obvi-
ously fatal, and indeed even in the application to string theory it can be useful to drop
the requirement of locality (and modular invariance) in intermediate results.
5.11 Fusion rules

The coefficients Cijk in three-point functions or in the operator product satisfy certain
selection rules. Note that i, j and k label fields φi (z, z̄) etc. The label “i” stands thus for
some combination of holomorphic and anti-holomorphic representations of the Virasoro
algebra. The selection rules depend on those representations. To discuss them it is thus
better to label the fields as φi,ı̄ .
The selection rules imposed by the Virasoro algebra are called fusion rules [44]. They
are written as follows X
[i] × [j] = Nij k [k] .
k
Here Nijk is a set of non-negative integers, and [i] . . . label representations of the Virasoro
algebra. The notation with raised indices is introduced for future purposes (when we
61
consider extended algebras), and has no relevance here.
If the fusion coefficient Nij k vanishes, this means that the OPE-coefficients Ciı̄,j̄,kk̄
vanish. If the coefficient does not vanish, the corresponding OPE-coefficient is allowed,
and usually it is then indeed non-zero.
From this discussion one might think that Nij k must be either zero or 1, but actually
it can be any non-negative integers. Values higher than 1 indicate that there exists more
than one way of coupling the fields. This should be compared to the tensor product rules
in group theory; for example in SU (3) there are two distinct ways of coupling two 8’s to
a third 8. The interpretation in terms of three point functions tells us that Nijk must be
symmetric in all three indices.
Although the fusion rules are similar to rules for tensor products in several respects, it
would not be correct to refer to them as the tensor product rules of the Virasoro algebra.
If one tensors two Virasoro representations, one would have to add up the central charge
and the conformal weight, which is not the case for the fusion rules.
5.12 The Verlinde formula

The reason fusion rules are discussed in this chapter is that there turns out to be a relation
between the fusion rule coefficients and the matrix S, discovered by E. Verlinde [55]. His
formula is∗ X Sin Sjn Skn
Nijk = . (5.8)
n
S 0n
It is remarkable that such a bizarre-looking expression involving matrix elements of a

unitary matrix actually produces non-negative integers, but it does!
Another useful way of looking at this formula is to regard Nijk as a collection of
matrices (Ni )jk = Nijk . Then (5.8) can be rewritten as

† Siq
(S Ni S)pq = δpq .
S0q
In words, the matrix S simultaneously diagonalizes the fusion rules for all fields i. In
deriving this relation we have used the fact that the matrix S on Virasoro representations
is real.
The ratios
(n) Sin
λi ≡
S0n
are sometimes called the (generalized) quantum dimensions of the field i (there is a more
(0)
restricted definition where one only calls λi the quantum dimension). Yet another way
of writing the Verlinde formula is as
(n) (n)
X (n)
λi λj = Nijk λk ,
k
∗
We are restricting ourselves here to theories with trivial charge conjugation, which is equivalent to S
being real. This includes the minimal series of Virasoro theories. The general case will be discussed later.
62
with no sum over n. This equation states that the quantum dimensions for each value of
n form a one-dimensional representation of the fusion rules.
Formula (5.8) was conjectured by Verlinde, and proved by Moore and Seiberg [38, 39].
5.13 Higher genus partition functions

The basic ideas discussed here have generalizations to higher genus. On a surface with n
handles one can define a basis of homology cycles ai , bi , i = 1, . . . n as shown in the fig.
6. One can choose a set of n holomorphic 1-forms on the surface, and normalize them so
b1 b2 b3
a1 a2 a3
Figure 6: Modular parameters of the triple torus.
that Z
ωj = δij
ai
Then the integral along the b cycles defines the period matrix Ωij
Z
ωj = Ωij
bi
This is the higher genus generalization of τ . In the lattice picture of the torus, the a-
cycle is the line from 0 to 1, and the b cycle the line from 0 to τ . The holomorphic
1-form isR dz. Since it is constant it is periodic, hence well-definedR on theR torus. Clearly
R 1 τ
a
dz = 0
dz = 1 so that it is properly normalized. Then Ω11 = b dz = 0 dz = τ . The
modular transformation S corresponds to the mapping a → −b and b → a, while the
transformation T corresponds to replacing b by a + b, without changing a.
The transformation a → b and b → a is not a modular transformation. The reason is
that part of the definition of such transformations is that they should leave the intersection
matrix of the cycles invariant. This matrix counts the number of intersections for each
pair of cycles, with their signs. Note that if one follows the a-cycle along the direction
of the arrow, at the intersection point the b-cycle has a definite direction. This direction
is flipped by the mapping a → b, b → a, but not by a → −b, b → a. Note also that
the square of the latter transformation is a → −a, b → −b. On the complex upper
half-plane in which τ lives, this map is represented by the identity. Hence the modular
transformations represented as τ → aτ +b
cτ +d
do not form a faithful representation of the full
modular group of the torus, Sp2 (Z).
63
The higher genus generalization of the modular transformation is
Ω → (AΩ + D)(CΩ + D)−1 ,
where the n × n matrices A, B, C, D must be such that

!
A B
∈ Sp2n (Z) (5.9)
C D
For n = 1 one has the isomorphism Sp2n ∼ SL2 .

The condition for higher genus modular invariance is thus simply invariance of the
partition function P (Ω, Ω̄) under this group. The computation of the partition function
is much harder, and in practice has only been done for theories of free bosons and fermions,
plus a few isolated other cases.
The most important constraint come from the partition function at genus 1. The
genus-2 partition function adds some further restrictions, in particular to quantities which
vanish at genus 1. There are believed to be no further constraints from genus three and
higher.
The definition of the modular domain which contains all inequivalent surfaces exactly
once becomes much harder than it is for the torus.
6 Extensions of the Virasoro Algebra

In most applications one is dealing with theories that have more symmetries than just the
Virasoro algebra. Such symmetries are called extensions of the Virasoro algebra. These
theories have a larger algebra that contains the Virasoro algebra as a subalgebra. These
generalized algebras are often referred to as chiral algebras, since they are generated by
currents that are holomorphic or anti-holomorphic, as we will see. When going back to the
cylinder and to Minkowski space, holomorphic dependence on z translates to dependence
on x + t, so that the corresponding modes are purely left-moving; hence the word chiral.
The chiral algebra may in fact be different in the holomorphic and the anti-holomorphic
sector; then one speaks sometimes of a heterotic theory.
Since the algebra contains the Virasoro algebra, all representations can be decomposed
into Virasoro representations. In most cases the number of Virasoro representations con-
tained in an extended algebra representation is infinite.
6.1 Rational conformal field theories

We have seen that the Virasoro algebra with 0 < c < 1 has a finite number of unitary
representations. In general, a conformal field theory with an extended algebra with a
finite number of unitary representations is called a rational conformal field theory. One
can show that the values of c and h in such a theory must be rational numbers [1, 54].
64
For c ≥ 1 the number of Virasoro representations is infinite. This implies that a
diagonal modular invariant partition function necessarily contains an infinite number of
terms.
Since extended algebras can group an infinite number of Virasoro characters into
extended algebra representations, it may happen that a conformal field theory with c > 1,
which has an infinite number of Virasoro representations, becomes a rational conformal
field theory with respect to an extended algebra.
This clearly has positive consequences for the solvability of the theory. First of all
one would expect that there are fewer distinct correlation functions to be calculated.
Just as for the Virasoro algebra one would expect that it should be sufficient to know
the correlation functions for just one (highest weight) state in each representation. In
a rational conformal field theory the number of distinct n-point functions is then finite.
This implies also that the number of operator product coefficients is finite, so that one has
a much better chance of determining them all from duality arguments. But perhaps the
most important consequence of additional symmetry is the appearance of additional null
vectors. This implies more constraints on correlation functions than the Virasoro algebra
gives by itself, so that correlators have to satisfy additional differential equations.
6.2 Currents
The Virasoro algebra is generated by operators Ln , which are modes of a current T (z),
which has conformal spin 2. Extended algebras are generated by modes of other currents.
These currents may have integer spin (bosons), half-integer spin (fermions) or fractional
(rational, but not half-integer) spin (para-fermions [16]). An important difference between
bosonic currents and (para)-fermionic ones is that the former can satisfy the condition
h − h̄ = 0 mod 1. Hence the currents J(z) can appear as conformal fields in a modular
invariant theory. The converse is also true. Suppose a conformal field theory contains a
conformal field J(z, z̄) with h ∈ Z and h̄ = 0. Then we know that the state J(0, 0) |0i
satisfies L̄−1 J(0, 0) |0i = 0, since this would-be descendant has zero norm. Hence the
corresponding descendant field must vanish. This field is ∂z̄ J(z, z̄). Since it must vanish,
J(z, z̄) must be holomorphic.
Strictly speaking this argument only shows that the derivative must vanish at the
origin. However, we want the descendant not only to vanish as an “in” state, but we also
want all correlators involving this state to vanish. This implies that the descendant field
must vanish for all values of z.
6.3 Fermionic currents

Fermionic currents are often discussed in a similar way, as if they were holomorphic
operators. However, they can never appear like that in a modular invariant theory. One
way of thinking about them is in terms of a theory that is conformally invariant only
under a maximal subgroup of the modular group that does not contain the element T but
does contain T 2 . Then fermionic operators are allowed, but correlation functions on the
65
cylinder may have branch cuts. We will soon see the consequences of this fact. Similar
remarks apply to para-fermionic theories, but will not be discussed here.
6.4 Mode expansions

Modes of the currents are defined as follows
I
dz r+h−1
Jr = z J(z)
2πi
The inverse relation is X
J(z) = z −r−h Jr (6.1)
r
Since the currents are conformal fields, it is straightforward to compute their commutator
with the Virasoro generators,
[Ln , Jr ] = (n(h − 1) − r)Jr+n , (6.2)
where h is the conformal weight of J(z). It follows that acting with Jr decreases the
conformal weight of a state by r. The commutators of the current modes themselves
define the extended algebra. To compute them requires more detailed knowledge, namely
the operator product of two currents with each other.
6.5 Integer and half-integer modes

The parameter r that defines the modes of a current is not necessarily an integer. Suppose
on some primary field φ(w, w̄)
J(z)φ(w, w̄) = (z − w)α φ0 (w, w̄) + . . .
Then we would like to define the J(z)-charges of the state created by φ from the vacuum
by means of the contour integral
I I
dzz r+h−1
J(z)φ(0, 0) |0i = dzz r+h−1 z α φ0 (0, 0) |0i .
writing z = eiθ , dz = eiθ idθ we get a phase integral

Z 2π
idθei(r+h+α)θ
0
This integral is well-defined (i.e. independent of the choice of the θ interval) only if
r + h + α ∈ Z. In modular invariant theories α is always an integer, and it follows then
that h + r must be an integer as well. In theories with fermionic currents, α can have both
integer and half-integer values, and hence we must choose integer or half-integer modes
for the currents, depending on which representation they act. Note that the right-hand
66
side of an operator product does not contain just one term, but in general an infinite
number of terms. A mode expansion can only be defined if the fractional parts of the
exponents α are the same for all terms. One calls such a universal phase exp 2πiα the
monodromy of J around φ.
Note that the periodicity changes if we go from the complex plane to the cylinder.
Because of the factor ( ∂f
∂z
)h in the conformal transformation currents on the plane and
the cylinder are related as
J cyl (w) = z h J(z) ,
where z = ew . If h is half-integer, the periodicity changes.
For historical reasons [45, 42] representations on which J is half-integer moded are
called Neveu-Schwarz representations. On the plane, the current acting on such ground
states is periodic around the origin (cf. (6.1)), but it is anti-periodic around the cylin-
der. Representations which allow integer modes are called Ramond representations. The
current is anti-periodic on the plane, but periodic on the cylinder. This is summarized in
the following table
r∈ Plane Cylinder
1
Neveu-Schwarz Z+ 2
periodic anti-periodic
Ramond Z anti-periodic periodic
6.6 Types of chiral algebra extensions

A rough classification of the possible extensions of the Virasoro algebra is in terms of the
conformal spin of the currents. The following possibilities are of interest
1
• h= 2
Free fermions
• h = 1 Affine Lie algebras (also called Kac-Moody algebras in the physics literature)
3
• h= 2
Superconformal algebras
• h = 2 Virasoro tensor products
• h > 2 W-algebras.
6.7 Properties of extended Virasoro representations

In the foregoing five lectures we have extensively studied Virasoro representations. For-
tunately most of what we learned remains valid for the extended theories.
Extended symmetries affect the representation theory in two ways. First the number
of ground states is typically reduced, because one imposes extra conditions on them.
These conditions usually take the form Jn |φi = 0 for positive modes of the extra currents
67
J(z). The rationale behind this should be clear: just as we did for the Virasoro algebra
we would like to build highest weight representations. Since Jn , n > 0 decreases the
conformal weight, any state not annihilated by Jn , n > 0 is obviously not a highest weight
state. Having determined the highest weight states, one uses the negative modes of the
extra currents in addition to those of the Virasoro algebra to build representations. Since
extra currents are used, the representations can only get larger, even though additional
null-states appear.
There is thus a new notion of primary field required. Primary fields not only have
operator products of a prescribed form with T (z) (namely (2.26)), but there are additional
operator products with the currents J(z) that must have a certain form. Furthermore
descendant fields are now not only created by T (z) but also by all currents J(z) in the
chiral algebra. These operator products are equivalent to the aforementioned requirements
on highest weight states, since that latter are still created from the vacuum by φ(0) |0i,
if φ(z, z̄) is an (extended algebra) primary field.
Explicitly these primary field conditions take the following form for integer-moded
currents of integer spin h
J(z)φ(w, w̄) ∝ (z − w)−h φ0 (w, w̄) + higher order in z − w , (6.3)
This means that the operator product cannot be more singular than indicated. If it is
more singular φ(w, w̄) is a descendant field. However, the operator product can be less
singular than indicated in (6.3). For example, if the zero-mode J0 annihilates the state,
the leading power is −h + 1 or less; if J−1 also annihilates the states it is −h + 2 or less,
etc. Usually there is only one state that is annihilated by J0 , namely the vacuum, created
by the operator φ(z, z̄) = 1. The leading power is then in fact 0, and for example if
we consider the Virasoro algebra this implies that both L0 and L−1 must annihilate the
vacuum. Fields which have powers of (z − w)−1 larger than h in their operator products
are descendants; the corresponding states are not annihilated by Jn , n > 0.
For half-integer spin algebras we have to distinguish half-integer moded (Neveu-Schwarz)
and integer moded (Ramond) operators. In the former case the operator product of a
primary field φ(w) is
1
J(z)φ(w, w̄) = (z − w)−h+ 2 φ0 (w, w̄) + higher order in z − w ,
where φ0 is a descendant that has a conformal weight that is 12 larger than that of φ. It
may happen that there is no such field. Then J− 1 annihilates the ground state φ(0) |0i,
2
and the leading power in the operator product is lower. But in any case fields with a
power higher than h are descendants.
In the Ramond sector one has
J(z)φ(w, w̄) ∝ (z − w)−h φ0 (w, w̄) + higher order in z − w .
Only states that are annihilated by the zero mode generators do not satisfy this formula.
They have a leading power (z − w)−h+1 (or less). There can be arbitrarily many such
states.
68
Note that the zero-modes in of integer-moded operators either annihilate a state, or
they transform ground states of a given value of h into each other. The ground states
form in this way a representation of an algebra generated by the zero-mode generators.
All these notions have been developed explicitly for Kac-Moody algebras, free fermions
and superconformal algebras. In the application to W algebras there are several footnotes
to be added to this general picture.
6.8 Charge conjugation

Everything discussed in the section 4 and 5 goes through for extended algebras, apart from
one important difference. We have seen that two-point functions of Virasoro primaries
are diagonal in the sense that
h0| φh (z)φh0 (w) |0i = δh,h0 (z − w)−2h (6.4)
for the holomorphic part. Since Virasoro representations are uniquely determined by
c and h, the Kronecker δ implies that only identical representations have a non-trivial
propagator connecting them.
In the extended case (6.4) still holds, but now there can be representations with
identical values of h that are different with respect to other generators of the algebra. In
particular it may happen that the propagator does not act diagonally within each set of
h values. One can always choose a basis of fields so that they come in pairs connected by
the propagator. The two members of such a pair are called each others charge conjugates.
Charge conjugation thus defines a matrix C which is symmetric, whose entries are 0
or 1, and which satisfies C 2 = 1. It either takes a field into itself (such a field is called
self-conjugate), or to its charge conjugate. The vacuum is necessarily self-conjugate, since
it is non-degenerate.
If charge conjugation is non-trivial, the duality diagrams of the previous section must
be modified by assigning arrows to each line.
6.9 Characters and modular transformations

Virasoro characters of extended algebras are defined exactly as for Virasoro representa-
tions. One can generalize the notion of the character by inserting exponentials of zero-
modes of other currents into the trace, but we will not consider that here.
The matrix T requires no further discussion. There also is a matrix S with the property
X
1
χi − = Sij χj (τ ) .
τ j
Note however that this transformation as it stands does not always determine S com-
pletely, because it is now possible that several representations j have the same character.
This was excluded for Virasoro representations because all representations have different
conformal weights. One can nevertheless define S completely by taking into account extra
69
variables in the characters (as mentioned above) and by requiring it to be a unitary and
symmetric matrix.
The relation among the generators in the general case is
(ST )3 = S 2 = C with C 2 = 1
Note that the modular transformation S acting on the variable τ (τ → − τ1 ) squares to 1.

However the transformation on the a and b cycles is
a → −b; b→a
and squares to −1. It is thus a double cover of the transformation on the positive upper
half plane in which τ is defined. The transformation S 2 , i.e. a → −a, b → −b is non-
trivial, but it acts trivially on the “period matrix” τ . Intuitively S 2 flips the time (and
the space) direction on the torus, and this is why a field goes into its charge conjugate
rather than itself.
Because S is still unitary and symmetric, we have
S = CS † = S † C = S ∗ C = CS ∗
so that in particular reality of S is equivalent to C being equal to the identity.

The Verlinde formula in its general form reads
X Sin Sjn S †
k nk
Nij = .
n
S0n
Here the raised index indicates charge conjugation. We may also define
X X Sin Sjn Snl
Nijk = Nij l Clk =
l n
S0n
Because S is symmetric, Nijk is symmetric in all its indices. This is the quantity that
counts the number of couplings in the three point vertex. In other words, if Nij k does
not vanish, [i] × [j] contains the representation [k]. Hence they can be coupled to the
representation [k ∗ ] to form a non-vanishing three point coupling, by insertion of the k − k ∗
propagator. This is illustrated below.
j j
= k
k* k
i i
Nij k
Nijk
70
6.10 Virasoro tensor products
A simple example of an extended chiral algebra is obtained by taking the tensor product
of two Virasoro representations, with central charges c1 and c2 . The resulting theory has a
(1) (2)
Virasoro algebra generated by Ln + Ln with central charge c1 + c2 . The representations
are simply all pairs of representations of the two algebras, and have conformal weights
(1) (2)
hi + hi . It is easy to check that in such a theory there is a conformal field with weights
(2, 0), namely
J(z) = c2 T (1) (z) − c1 T (2) (z) .
This is the current of the extended symmetry.
The simplest modular invariant partition function of such a system is the diagonal one,
which is the product of the diagonal invariant of the two systems. However, in principle
there can be many additional modular invariants, and in general there are.
Of course one can also consider tensor products of representations of other extended
algebras.
6.11 Extensions and off-diagonal partition functions

Sometimes possible extensions of the chiral algebra can be read off directly from the
existence of modular invariant partition functions. A typical such partition function has
the form of a sum of squares,
M N
X X 2
P (τ, τ̄ ) = χl,a (6.5)
l=1 a=1
Such an expression can often be interpreted in terms of an extension of the original chiral
algebra (which itself may be an extension of the Virasoro algebra), in such a way that the
characters of the new algebra are equal to sums of characters of the original algebra
X
χnew
l = χl,a . (6.6)
a
The new theory as M characters, whereas the original one had at least N M characters.
In fact it always has more, because a general feature of a partition function of the form
(6.5) is that certain representations of the original algebra are “projected out”, i.e. they
do not appear at all in the off-diagonal partition function.
Of special interest is the identity character. If it is a sum of several characters of the
old theory, then the extra terms imply the existence of matrix elements Mi0 6= 0, where
M is the multiplicity matrix in the modular invariant. The corresponding primary fields
have h = hi 6= 0, h̄ = 0. They can thus be interpreted as currents, and they are in fact
precisely the currents that extend the chiral algebra.
Although in practice one only deals with explicit modular invariant partition functions
for extensions from one rational conformal field theory to another one, conceptually the
chiral algebra extension that make a non-rational conformal field theory rational work in
the same way. In that case M is finite and N is infinite.
71
6.12 The new S and T matrices
Once we have a new, smaller set of characters, one expects to have a new set of modular
transformation matrices S and T . The new T matrix is trivial to get, since by modular
invariance all terms in (6.6) have the same T -eigenvalue. The new matrix S new can be
obtained easily from the original one, S old , in the simplest case, where all linear combina-
tions (6.6) have the same number of terms (each with coefficient 1). It is then not hard
to show that the matrix
new 1 X old
Sl,m = S
N a,b (l,a)(m,b)
transforms the new characters if S old is the transformation matrix for the original one.
The proof goes as follows. Suppose the matrix M defining the modular invariant
partition function has the form X
Mij = via vja ,
a
a
where v is a set of orthogonal vectors
via vib = Na δ ab
In typical cases each v a consists of zeroes and one’s, and the position of the one’s defines
the blocks, but there is no need to be specific. The condition for S-invariance for such a
matrix yields X X
Sil vla vja = via vm
a
Smj
a a
Contracting both sides with vic yields

X
c
Nc vm Smj = vic Sil vla vja
a
The extended characters are X

χnew
a = via χi
i
Their transformation is

1 1
χnew
a − a
= vi χi −
τ τ
a
= vi Sij χj (τ )
1 a X b b
= v Sil vl vj χj (τ )
Na i b
X 1
via Sil vlb χnew

= b (τ )
b
Na
From this we read off that

new 1 a
Sab = vi Sil vlb
Na
72
Note that Sab is symmetric if and only if Na = Nb for all a, b that are connected via S. If
v a has its typical form, the result reduces to the one above.
Many off-diagonal invariant invariants have a more complicated from. For example, it
may happen (although it rarely does) that the linear combinations in (6.6) have coefficients
larger than 1. A more serious complication occurs when the linear combinations have
different lengths. The typical form of such a partition function – in this example with
linear combinations of either N terms or 1 – is something like
M N Nf
X X 2 X
P (τ, τ̄ ) = χl,a +
N |χf |2
l=1 a=1 f =1
This partition function can – usually – be interpreted in terms of a new, extended algebra
with M + Nf × N representations. Note that the last Nf × N representations have
characters that are identical in groups of N . This means that it is not obvious which
matrix S new to use for the transformations among these characters. Indeed, since they are
identical in groups of N the transformation τ → − τ1 does not determine S new completely.
This problem can be solved by imposing unitarity as well as the modular group property
(ST )3 = S 2 on S new , but it turns out that in this case the matrix elements of S new are
not simply linear combinations of those of S old .
6.13 Extensions and automorphisms

The matrices M that define a modular invariant partition function can be divided into
two main groups: those with M0i = Mi0 = 0 for i 6= 0, and all others. It can be shown
that in the former case there is no extension of the chiral algebra, and that all characters
must appear in the partition function. However, they may appear non-diagonally, as
X
χl Πlm χ̄m
l
where Π is a permutation of the labels. It is not hard to see that Π is then an auto-
morphism of the fusion rules, i.e. the fusion coefficients Nijk are invariant when Π acts
simultaneously on all labels.
If some matrix elements M0i or Mi0 are non-zero, the modular invariant can always be
interpreted as an extension of the chiral algebra. If one re-writes it in terms of characters
of the new algebra one either gets the diagonal invariant of the new algebra, or a fusion
rule automorphism of the new algebra.
6.14 Simple currents

Many conformal field theories have representations [J] with the property that
[J] × [i] = [i0 ]
73
for all other representations [i]. The special property is thus that there is just one term
on the right hand side. Then [J] is referred to as a simple current [48].
The word “current” anticipates the fact that it may be used to extend the chiral
algebra, or at least plays the rôle of a (para)fermionic current.
Simple currents organize the fields in a conformal field theory in an obvious way
into orbits, and one can in an equally obvious way assign an order N to them. Among
themselves they generate an abelian group called the center of the conformal field theory.
Simple currents can always be used to extend the chiral algebra. In the simplest cases
– N prime – it is furthermore true that currents of fractional spin N` generate fusion rule
automorphisms. In more complicated cases one gets combinations of automorphisms and
extensions.
The number of simple current invariants of a given conformal field theory grows very
rapidly with the number of abelian factors of the center, but all solutions have now been
classified.
It seems that most modular invariant partition functions can be described in terms of
simple currents, but there are exceptions. These are called, quite naturally, exceptional
invariants.
7 Free Fermions
Free fermions are described by the two-dimensional action
Z
1
d2 z ψ∂z̄ ψ + ψ̄∂z ψ̄ ,

S=
8π
where we have already switched to Euclidean space and to complex coordinates. We will
focus on the fields ψ from here on; all equations that follow are also valid with bars on
all relevant quantities.
7.1 The propagator

The equations of motion for ψ(z, z̄) are ∂z̄ ψ(z, z̄) = 0, so that we may write ψ(z) instead
of ψ(z, z̄). The operator product of two fermions is
1
ψ(z)ψ(w) =
z−w
7.2 Energy-Momentum tensor and central charge

The energy-momentum tensor is
T (z) = − 21 : ψ(z)∂z ψ(z) : , (7.1)
where as usual normal ordering implies that the vacuum expectation value of T (z) is zero.
This requires the subtraction of the singular terms in the operator product.
74
It is a simple exercise to verify that the central charge is equal to 12 , and that ψ(z)
has conformal weight ( 21 , 0). This is an interesting result in view of the classification of
Virasoro representations. We have seen that for c = 21 three representations exist: with
h = 0, h = 21 and h = 16 1
. The conformal field ψ(z) clearly creates an h = 21 state from
the vacuum: 1
= ψ(z) |0i
2
7.3 Mode expansion

The free fermion can be expanded in modes. On the complex plane the mode expansion
is X 1
ψ(z) = bn z −n− 2 , (7.2)
n
which can be inverted in the usual way. When going to the cylinder the free fermion picks
1
∂z 2
up a conformal factor ∂w . Hence we get
1 X 1 X
ψ cyl (w) = z 2 bn z −n− 2 = bn e−nw .
n n
Here we see explicitly the aforementioned periodicity flip.
7.4 The spin field

The field ψ(z) has local operator products with all primary fields we have seen so far
(namely ψ(z) itself and the identity). We expect there to exist also fields with which it
has a square root branch cut, so that ψ(z) is realized à la Ramond.
1
Furthermore we expect fields with conformal weight 16 , since that is another allowed
Virasoro representation at c = 21 . Indeed, we will see that modular invariance forces such
fields to exist.
1
Let us therefore introduce a field σ(z, z̄) with h = h̄ = 16 . Its operator product with
ψ(z) has the form
1 1
ψ(z)σ(w, w̄) = (z − w)hµ − 2 − 16 µ(w, w̄) ,
where µ is some other field in the theory. Since we know all Virasoro representations
its conformal weight hµ can only be 0, 21 or 16 1
, perhaps up to integers if we allow µ to
1
be a descendant. Clearly only the choice 16 leads to an acceptable branch cut, since for
fermions only square root branch cuts (or no cuts at all) are allowed. We find thus that
the field σ does indeed introduce the expected branch cut. This field is often referred to
as a spin field. 1 1
Acting on the vacuum the field σ produces a state 16 , 16 = σ(0, 0) |0i. The field
ψ(z), acting on such a state is anti-periodic on the plane, and hence integer-moded.
It is tempting to argue that µ and σ are in fact one and the same field. In the present
context that is in fact not quite correct, since we are not dealing with a modular invariant
75
partition function. When we make the theory modular invariant, ψ(z) and either σ or µ
are removed from the spectrum (i.e. all the states they create are removed). The primary
fields in the modular invariant theory are 1, ψ(z)ψ̄(z̄) and
1 1 σ(z,
z̄). 1Each creates one state
1
from the vacuum, namely the vacuum itself, the state 2 , 2 and 16 , 16 . On these states

one builds Virasoro representations. The ground states are non-degenerate, i.e. there is
just one state with the corresponding values of h, h̄.
The partially modular invariant theory has in addition the primary fields ψ(z), ψ̄(z̄)
and µ(z, z̄). The operator products are now non-local. In addition to the ones already
mentioned one has
σ(z, z̄)σ(w, w̄) → 1, ψ(w)ψ̄(w̄) (7.3)
µ(z, z̄)µ(w, w̄) → 1, ψ(w)ψ̄(w̄) (7.4)
σ(z, z̄)µ(w, w̄) → ψ(w), ψ̄(w̄) (7.5)
If one substitutes the conformal weight factors (z − w)hk −hi −hj one finds that the last
operator product is non-local, indicating that one cannot have both σ and µ in the same
modular invariant theory. Removing the free fermions and either µ or σ solves the non-
locality problem in a consistent way, i.e. the operator product closes after this truncation.
This will be made explicit later in this chapter.
7.5 Free fermion characters

One advantage of the free fermion formulation of the c = 21 theory is that it is straightfor-
ward to compute the characters. The reason why this is not straightforward for Virasoro
representations is the existence of null vectors. Let us compare the lowest lying states in
the three representations.
7.5.1 Neveu-Schwarz states

Consider first the Neveu-Schwarz sector. Using the operator product of two free fermions
and the mode-expansion (7.2), one can easily derive that the modes satisfy the following
anti-commutator
{br , bs } = δr+s,0 ,
where r and s are half-integers. Clearly we cannot impose br |0i = 0 for all r, and hence
we only do so for r > 0. This is also the natural definition for highest weight modules,
since the positively moded br ’s decrease the L0 eigenvalue.
The Virasoro generator L0 can be expressed in terms of the fermionic oscillators.
Classically, the result is X
L0 = 12 rb−r br .
r
Quantum mechanically we have to be more careful, since the operators br and b−r do not
commute. Changing their order only affects L0 by a constant, so that we get
X
L0 = rb−r br + constant .
r>0
76
Here we have normal ordered the fermionic oscillators. Since we have already defined
normal ordering in (7.1), the constant is not a free parameter. It must be chosen in such
a way that h0| L0 |0i = 0, i.e. the constant must be zero. It follows immediately that
L0 |hi = 0 if |hi is a highest weight state of the fermionic algebra, i.e. if br |hi = 0 for
positive r. Hence the fermionic algebra can have just one representation in the Neveu-
Schwarz sector, namely the one built on the vacuum.
At the first few levels, this representation contains the following states:
h=0 |0i
h = 21 b− 1 |0i
2
h=1 none
h = 23 b− 3 |0i
2
h=2 b− 3 b− 1 |0i
2 2
5
h= 2
b− 5 |0i
2
h=3 b− 5 b− 1 |0i
2 2
7
h= 2
b− 7 |0i
2
h=4 b− 7 b− 1 |0i ; b− 5 b− 3 |0i
2 2 2 2
Note that fermionic oscillators must satisfy the Pauli exclusion principle, so that for
example b1/2 b1/2 is zero. For this reason there is no state at level h = 1, and we have to
go to h = 4 to find more than one state.
An important question is whether all these states have positive norm. Due to the
simplicity of the free fermion algebra it is not hard to show that indeed the norm of every
state is exactly 1, and that all distinct states are orthogonal. The fact that ψ(z) as a field
on the cylinder is real implies that b†r = b−r . It is then trivial to prove that the states are
indeed orthonormal.
Hence we may expect them to fit exactly into one or more Virasoro representations.
The relevant Virasoro representations are, for the ground state representation
h=0 |0i
h=1 L−1 |0i
h=2 L−2 |0i , (L−1 )2 |0i
h=3 L−3 |0i , L−1 L−2 |0i , (L−1 )3 |0i
h=4 L−4 |0i , L−3 L−1 |0i , L−2 L−2 |0i , (L−1 )4 |0i , L−2 (L−1 )2 |0i
For the representation with ground state weight h = 21 we find exactly the same result,
1

with |0i replaced by 2 , and all conformal weights shifted up by half a unit. However,
we have already seen that not all these states have positive norm. The ground state
representation has a null state at its first excited level (which propagates trough to all
77
higher levels), while the h = 12 representation has a null state at its second level. This
agrees precisely with the assumption that the fermionic representation is the sum of the
two Virasoro representations, and also gives us a quick way of counting the number of
Virasoro null states at higher levels.
7.5.2 Neveu-Schwarz characters

It is straightforward to compute the character for the fermionic representation, since there
are no null states to be taken into account. Each oscillator b−r can act once or zero times
on the ground state. If there were just one oscillator b−r there would just be two states,
|0i and b−r |0i with h = 0 and h = r. The character is thus Trq L0 = 1 + q r . All oscillators
with different modes acts independently, and it is easy to see that each contributes via
additional factors of this form. Furthermore we have to take into account the subtraction
−c/24. The result is thus
∞
c 1
Y
L0 − 24 − 48
χ0 + χ 1 = Trq =q (1 + q r ) . (7.6)
2 1
r=
2
This gives us the sum of the characters of two Virasoro representations. Their difference
is also easy to compute. Just observe that states created by an odd number of fermions
contribute to the spin- 21 representation, and the remaining ones to the vacuum represen-
tation. Hence we can get the difference by changing the sign of the contribution of each
single fermion to the trace,
∞
1
Y
− 48
χ0 − χ 1 = q (1 − q r ) . (7.7)
2 1
r=
2
This expression can also be written as a trace over the fermion representation, namely as
c
χ0 − χ 1 = Tr(−1)F q L0 − 24 . (7.8)
2
Here F is the fermion number operator. We have now succeeded in computing both the
h = 0 and h = 21 character at c = 12 .
7.5.3 Ramond states

In the Ramond sector the fermionic oscillators are integer moded, which has the interesting
consequence that there exists a zero mode oscillator b0 , which satisfies {b0 , b0 } = 1. The
expression for L0 is X
L0 = nb−n bn + constant .
n>0
Obviously highest weight states |hi must satisfy bn |hi = 0 for n > 0, and this implies
that all highest weight representations in the Ramond sector must have the same highest
78
weight, namely “constant”. Since we know that the Ramond sector is realized on states
1 1
created by the field σ with h = 16 , it follows that “constant” must be equal to 16 in this
case.
Here we were making use of the fact that we knew the ground state energy from
Virasoro representation theory. It should be possible to derive this directly from the
properties of the free fermion system. This can indeed be done, and works as follows.
Consider the computation of the fermion operator product
X −r− 1 X 1
ψ(z)ψ(w) = br z 2 bs w−s− 2
r s
The moding depends on the state on which this operator acts. If that state is the vacuum
we are in the Neveu-Schwarz sector and the sum is over half-integers; we can normal order
the fermionic oscillators to get
∞
X 1 1 1
z −r− 2 wr− 2 + n.o. = + n.o. ,
1
z−w
r=
2
where “n.o” stands for normal ordered terms. The expectation value of those terms
vanishes for any highest weight state. The result of this computation is different in the
Ramond sector; now we get
∞
(p )
z
pw
X 1 1 1 +
z −n− 2 wn− 2 + 12 √ + n.o. = 12 w z
+ n.o.
n=1
zw z − w
The zero modes have been taken into account by using 1 = {b0 , b0 } = 2b20 . Notice that
the singularity for z → w is the same in both cases, but the Ramond sector propagator
has branch cuts in z and w. The energy-momentum tensor is defined as

1 1
T (z) = − 2 lim ψ(z)∂w ψ(w) −
w→z (z − w)2
This yields a purely normally ordered result in the Neveu-Schwarz sector, but in the
Ramond sector we get
√
1 1 1 z 1
T (z)R = n.o + lim √ √ −2 √ = n.o + .
w→z 4 z w3 2
(z + w) w 16z 2
Now suppose that the Ramond ground state on which T (z) acts is created from the
vacuum by a field σ(w) (ignoring any w̄-dependence). Then this result can be interpreted
in terms of the energy momentum tensor acting on σ(0) |0i,
hσ
T (z)σ(0) |0i ≈ σ(0) |0i ,
z2
1
provided that hσ = 16
.
79
7.5.4 The Ramond ground state
It is fairly obvious how to build up the representation, the only slight problem being the
action of the operator b0 . This operator changes the fermion number of the state it acts
on. To realize this we need thus two states, one with (−1)F = + and one with (−1)F = −.
Denoting these states as |+i and |−i we have thus
1 1
b0 |+i = √ |−i , b0 |−i = √ |+i ,
2 2
so that b20 = 21 . Of course we can realize this operator algebra on even more states, but
two is the minimum required.
7.5.5 Ramond characters

Having done this, we get for the character
∞
c 1
Y
L0 − 24
χ 1 ∝ Trq = 2q 24 (1 + q n ) ,
16
n=1
by exactly the same arguments as used above. The correct normalization will be discussed
in a moment. In principle the ground state of a Virasoro (or extended Virasoro) algebra
can be degenerate, so in principle it could be possible that the factor 2 should be absorbed
into the character itself. It is also possible to define this trace with a factor (−1)F , but
it is clear that the result is then zero: the operator b0 maps any state into a degenerate
state, while flipping the fermion number.

Let us now assemble the partition function, by combining it with the anti-holomorphic
fields. This is trivial in the Neveu-Schwarz sector. The oscillators b̄r contribute addition
factors (1 + q̄ r ) so that we get
∞
1
Y
PNS = (q q̄)− 48 (1 + q r )(1 + q̄ s ) .
1
r,s=
2
Now consider the Ramond sector. Here some further thought is needed. Do we again
double the ground state to deal with the action of b̄0 (in other words, do we take the
absolute value squared of (7.5.5))? Clearly this is not needed, because we already have
two ground states, and that is sufficient to realize simultaneously the b0 and the b̄0 algebra.
If we define the fermion number operator F to count the total fermion number (for ψ(z)
as well as ψ̄(z̄), and we choose two ground states |+i and |−i with opposite total fermion
numbers, everything will work automatically. Hence we define
∞
1
Y
PR = 2(q q̄) 24 (1 + q n )(1 + q̄ m ) .
n,m=1
80
7.7 Theta-functions
Altogether we have now defined four kinds of partition functions on the torus: with
R or NS boundary conditions along the space direction, and with or without (−1)F
operator inserted. This latter operator can be interpreted in terms of periodicity along
the Euclidean time direction of the torus. The normal trace corresponds to a fermion path
integral with anti-periodic boundary conditions (this boundary condition has the same
origin as the usual − sign in fermion loops; it can be computed by repeating the calculation
that yields (5.2) for fermions). The insertion of (−1)F gives an extra − sign for every
fermion in the loop, so it flips the boundary condition to periodic. Hence we have Here the

c c̄
TrNS q L0 − 24 q̄ L̄0 − 24
θ3
AA η

c c̄
TrNS (−1)F q L0 − 24 q̄ L̄0 − 24
θ4
AP η

c c̄
TrR q L0 − 24 q̄ L̄0 − 24
θ2
PA η

c c̄
TrR (−1)F q L0 − 24 q̄ L̄0 − 24
θ1
PP η
letters “AP” indicate anti-periodicity along the “space” direction and periodicity along
the “time” direction on the torus, etc. It turns out that these four partition functions can
be expressed in terms of standard mathematical functions, namely the Jacobi θ-functions
and the Dedekind η function. These functions are defined as follows
hai X 2
θ (z|τ ) = eiπ[(n+a) τ +2(n+a)(z+b)]
b n
with the additional definitions

1/2 1/2 0 0
θ1 = θ ; θ2 = θ ; θ3 = θ ; θ4 = θ
1/2 0 0 1/2
and ∞
1
Y
η(q) = q 24 (1 − q n )
n=1
2πiτ
with q = e . The last column above indicates the identification of each partition
function with ratios of θ and η functions. The Jacobi θ-functions have two arguments,
but we are only using them at z = 0 here. The function θ1 (z|τ ) vanishes for z = 0,
as does the partition function in the PP sector, but it can be made plausible that the
identification given here is the correct one. The fact that these functions are identical is
far from obvious, but is one of many remarkable identities that modular functions enjoy.
81
7.8 Modular transformations
Finally we discuss modular invariance. Clearly modular transformations change the
fermion boundary conditions. For example, the transformation S interchanges the two
cycles (“space” and “time”) on the torus, and hence it interchanges AP and PA. The
transformation T maps XY to X(XY) as shown in the figure, where X and Y stand
for A or P, and the multiplication rule is AA=P, AP=A and PP=P. In other words, it
interchanges AA and AP. Since S and T generate the modular group we generate all
permutations of AA, AP and PA, whereas PP transforms into itself.
τ τ+1
Y XY
These transformations are clearly sensitive to the correct normalization of the partition
functions. They can be computed explicitly for the θ and η functions, and one find
1 √ √
θ1 (− ) = −i −iτ θ1 (τ ); θ2 (− τ1 ) = −iτ θ4 (τ );
τ
1 √ √
θ3 (− ) = −iτ θ3 (τ ); θ4 (− τ1 ) = −iτ θ2 (τ )
τ
θ1 (τ + 1) = eiπ/4 θ1 (τ ); θ2 (τ + 1) = eiπ/4 θ2 (τ );
θ3 (τ + 1) = θ4 (τ ); θ4 (τ + 1) = θ3 (τ )
1 √
η(− ) = −iτ η(τ ); η(τ + 1) = eiπ/12 η(τ )
τ

It follows that the partition function PR + PNS = θη3 + θη2 is not modular invariant,

as expected. It is in fact invariant under a subgroup of the modular group generated by
T ST and T 2 . This is clearly a subgroup of order 2, since by adding the element T we
get the full modular group. This shows in particular that we have chosen the correct
normalization for the ground state in the Ramond sector.
82
7.9 The modular invariant partition function
It is also clear that the following partition function is fully modular invariant

1
θ3 θ4 θ2 θ1
+ + ±
2 η η η η
The factor 21 was added to make sure that the vacuum appears with the correct multi-
plicity, namely 1. The last term can be added with any factor, since it is (a) modular
invariant by itself and (b) zero. However, consistency of higher loop diagrams as well as
one-loop diagrams with external legs force this term to appear exactly as it does. The
two signs have a simple interpretation: the Ramond ground state appears in the partition
function with a factor 12 (1 ± (−1)F ) so that depending on the sign either the ground state
with positive or the one with negative fermion number survives. Note that the modular
invariant partition function has just one Ramond ground state. This is no problem, since
the operators b0 and b̄0 (zero modes of ψ(z) and ψ̄(z̄) are not in the theory anymore. The
first two terms only have contributions from the state ψ(0)ψ̄(0̄) |0i and its descendants,
and this operator does not change fermion number by an odd amount. The partially
modular invariant partition function has two Ramond ground states, corresponding to
the fields σ and µ. Depending on the sign choice, either one of these is projected out.
This sort of operation (for going from a partially modular invariant partition function to
a modular invariant one) is sometimes called a GSO-projection (GSO stands for Gliozzi,
Scherk and Olive, whose paper [26] was the starting point of superstring theory).
7.10 Ising characters

We can write the modular invariant partition function as
2 2 2
χ 0 + χ 1 + χ 1
2 16
by making the identifications

s s !
1 θ3 θ4
χ0 = +
2 η η
s s !
1 θ3 θ4
χ1 = −
2 2 η η
s
1 θ2
χ1 = √
16
2 η
The first two equations follow already from (7.6) and (7.8).
83
7.11 The matrix S and the fusion rules
Using the transformation properties of the θ functions and the η function it is now easy
to get the matrix S for the c = 21 system. On the basis (1, ψ, σ) the result is
√
 
1 1 1
 2 2 2
2
√ 
 
S= (7.9)
 1 1 1
 2 2
− 2 2 
 √ √ 
1 1
2
2 − 2
2 0
Using this matrix and the Verlinde formula we can compute the fusion rules:
[1] × [1] = [1]

[1] × [ψ] = [ψ]
[1] × [σ] = [σ]
[ψ] × [ψ] = [1]
[ψ] × [σ] = [σ]
[σ] × [σ] = [1] + [ψ]
This result should be compared with (7.5).
7.12 Multi-fermion systems

It can be shown that if there is more than one fermionic current with spin 12 , then a
corresponding part of the theory can be described as a free fermionic theory with c = N/2,
where N is the number of fermions. Of course this c = N/2 theory can appear as part
of a tensor product with other (extended) Virasoro representations, but at least the free
fermion part is easy to describe, and exactly solvable.
In such a free fermion theory each fermion can have its own boundary conditions on
the torus and higher Riemann surfaces, but there are constraints from modular invariance.
These constraints have been solved in general when the number of fermions is even, but
there is still some controversy regarding the odd fermion number case.
The number of modular invariant partition functions one can write down grows ex-
tremely rapidly with N . Systematic studies of conformal field theories built out of free
fermions (in the context of heterotic string construction) were presented in [35, 2].
8 Free Boson Partition Functions

The free boson provides another simple example of a theory with an extended algebra.
In this case the current has spin 1,
J(z) = ∂Φ(z) .
84
The mode expansion has already been discussed before.
As we will see later, such a current can be interpreted as a generator of a U (1) sym-
metry, with the momenta as charges.
8.1 The spectrum

The discussion of the spectrum is quite similar to that of the free fermion. The ground
states are defined by the condition
αn |xi = 0, n>0
The representations are built up by acting with the negatively moded oscillators. It is
not hard to see that any state gotten this way has positive norm.
The Virasoro generators are dependent on the bosonic oscillators,
X
Ln = 21 αn−m αm ,
m
where α0 = p, and the sum is over all integers. For the Virasoro zero mode we get thus
X
L0 = 21 p2 + α−m αm .
m>0
In principle we would have to worry about normal ordering, but since we know that
L0 |0i = 0 we see immediately that there is no additional constant.
The ground state |xi is completely determined by the action of the zero-mode generator
p. Once this is fixed, we know the entire representation, and the action of the Virasoro
generators. Hence we define
|pi : pop |pi = p |pi ,
where p on the left-hand side is the operator, and on the right hand side the eigenvalue.
Note that there is no separate holomorphic and anti-holomorphic zero-mode algebra:
α0 = ᾱ0 = p.
8.2 The characters and the diagonal invariant

It is straightforward to derive the character formula, since just as for the free fermion all
oscillators act independently and without generating null vectors. The result is
∞ 1 2
1 2 1 Y 1 q 2p
χp (q) = q 2 p − 24 = .
n=1
(1 − q n) η(q)
Note that the expansion of (1 − q n )−1 yields exactly one contribution at any level that is
a multiple of n. Thus each such factor describes the contribution of one bosonic oscillator
α−n acting any number of times on |pi.
85
Since any real value of p is allowed, there exists an infinite number of characters. The
diagonal partition function is therefore not a sum, but an integral
Z ∞ 2 2 √
eiπτ p e−iπτ̄ p Imτ
P (τ, τ̄ ) = dp ∝ .
0 η(τ ) η(τ̄ ) η(τ )η(τ̄ )
The proper derivation requires of course a discussion of the measure and the normalization,
but the result is correct. This factor appears in the partition function of the bosonic string,
which is described by a tensor product of 26 free bosonic theories (plus ghosts).
Note that in this partition function we are exactly using all the ground states we have
at our disposal. Although the algebra is extended by ∂Φ we do not get a finite number of
primary fields, i.e. a rational conformal field theory. In many of the representations the
extension does not even make any difference. If there are no null vectors in a Virasoro
representation, the Virasoro algebra acts just like a free bosonic oscillator, and one gets
a partition function
∞
c
h− 24
Y 1
q ,
n=1
(1 − q n )
where now every factor represents a single Virasoro generator L−n instead of a free bosons
α−n . Hence if on |pi the Virasoro algebra has no null vectors, the Virasoro representation
is equal to the “Virasoro+∂Φ”- representation.
The existence of Virasoro null vectors follows from the same curves we used for c < 1.
These curves hit the line c = 1 at several values of h, and only for those values the
Virasoro representation has a null state. From (3.20) and (3.21) we see that this happens
for m → ∞,
1
h = (p − q)2
4
For example for h = 0, L−1 |0i is a null state. The state α−1 |0i is of course not null, so
that the identity representation is indeed non-trivially extended by ∂Φ.
8.3 Chiral bosons

The free boson mode expansion can be generalized by adding separate momenta for the
holomorphic and anti-holomorphic terms:
X1
αn z −n + α̃n z̄ −n

Φ(z, z̄) = q − i(pL log(z) + pR log(z̄)) + i
n6=0
n
We have denoted these momenta as “L” (left) and “R” (right) because z and z̄ originate
from left- and right-moving modes on the cylinder. We may straightforwardly split also
q in left- and rightmoving operators by writing q = qL + qR . Furthermore we define the
canonical commutators [qL , pL ] = [qR , pR ] = i, while left and right operators commute. If
we identify pL = pR = p this leads again to old commutator [q, p] = i. Having done this
we can now split the boson completely in left and right components
Φ(z, z̄) = ΦL (z) + ΦR (z̄)
86
with X1
ΦL (z) = qL − ipL log(z) + i αn z −n ,
n6=0
n
and analogously for ΦR .

These manipulations do not influence any previous results that depend only on ∂Φ or
¯ but we can now give meaning to chiral (holomorphic) objects like
∂Φ,
eiλΦR (z)
2
It may be checked that this is a conformal field of weight 12 ~λ .
To see what the meaning is of the separate left and right momenta we can express the
field back into cylinder coordinates. Then we get
Φ(x0 , x1 ) = q + 2px0 + Lx1 + oscillators ,
where
pL = p + 12 L , pR = p − 12 L (8.1)
Previously we did not have the extra x1 term because we required Φ to be periodic,
Φ(x0 , x1 ) = Φ(x0 , x1 + 2π). The extra term destroys the periodicity unless we impose it
as a symmetry on the field Φ: Φ = Φ + 2πL. This must hold for any eigenvalue that
the operator L can have, and obviously also for all integer linear combinations of those
eigenvalues. If we want Φ to have a non-trivial dependence on x1 , the only possibility is
then that the L eigenvalues are quantized on a lattice of dimension equal to the number
of free bosons.
This has a natural interpretation in closed string theory, where Φ is viewed as the
coordinate of a space in which the string is embedded (this space is called target space).
The existence of a lattice means that the space is compactified on a torus (a D-dimensional
torus can be defined as D-dimensional Euclidean space in which points differing by vectors
on a lattice are identified). If L is a non-trivial lattice vector this means that the string is
not closed in the Euclidean space, but it is closed on the torus, i.e. the string winds around
a couple of times around the torus and ends in a point identified with its beginning.
8.4 Further extensions of the chiral algebra

We arrive at the same lattice description naturally by extending the chiral algebra further.
In addition to ∂Φ we add integer spin currents
Vλ = eiλΦ(z) , λ2 ∈ 2Z (8.2)
to both the left and the right algebra. Note that such a current corresponds to momenta
(λ, 0), so that it is only after introducing separate pL and pR that we have this possibility.
It is easy to check that Vλ satisfies the operator product
0
Vλ (z)Vλ0 (w) = (z − w)λλ Vλ+λ0 (w) + . . . (8.3)
87
Therefore, closure of the operator product requires V2λ (z) to be an operator in the theory
if Vλ (z) is. More generally we see that the set of λ’s such that Vλ is in the chiral algebra
must close under addition. It forms thus a one-dimensional even lattice, which we will
call Λ. Note that the operator product (8.3) is automatically local if the lattice is even.
8.5 Representations of the extended algebra

Since the chiral algebra contains in any case the Virasoro algebra and the operator ∂X,
any other states in the theory are built on ground states |pL , pR i. We have to restrict
this set by imposing on it highest weight conditions with respect to the extended algebra.
The field creating these states from the vacuum are
VpL pR (z, z̄) = eipL ΦL (z) eipL ΦR (z̄) ,
because
eipL ΦL (0) eipL ΦR (0̄) |0i = eipL qL +ipR qR |0i = |pL , pR i .
Locality with respect to the left and right chiral algebra requires that λpL ∈ Z and
λpR ∈ Z. This immediately restricts the set of left and right momenta that we can ever
encounter to the set
pL ∈ Λ∗ ,
where Λ∗ is the dual (or reciprocal) lattice of λ,
Λ∗ = {µ ∈ R|µλ ∈ Z, ∀λ ∈ Λ}
The lattice Λ is necessarily of the form nR, n ∈ Z and R2 even. The lattice Λ∗ has the
form m/R, m ∈ Z. For example, if Λ is the set of even integers, Λ∗ is the set of integer
and half-integers. In this description R denotes the smallest positive value of λ on the
lattice.
Now let us try to find which fields are primary with respect to the full extended
algebra. As we have seen in (6.3), any field with a singularity stronger than (z − w)−h
in its operator product with a current of spin h is a descendant. A field VpL pR (z, z̄) has
singularity (z − w)pL λ with Vλ . Hence we find the condition
pL λ ≤ 21 λ2 , ∀λ∈Λ (8.4)
and the same for pR . The vectors on Λ∗ satisfying the highest weight condition (8.4) are
thus those with
− 12 R2 ≤ m ≤ 12 R2
We see thus that there is – in both the left as the right chiral algebra – only a finite
number of highest weight representations. Hence the theory we are constructing is a
rational conformal field theory.
Note that each highest weight completely fixes the corresponding representation, since
it determines completely how all the oscillators and the operators p and q act on a state.
88
Now we can build these representations by acting with all negative modes of ∂Φ and
VλΦ . Doing this in an unrestricted way would certainly lead to null states, since there
is an infinite number of chiral algebra generators. However, writing everything in terms
of oscillators and momenta, one sees that the only states one can ever get starting from
|pL , pR i are of the form
(oscillators) |pL + λ, pR + λ0 i , λ, λ0 ∈ Λ .
Furthermore any state of this form is indeed generated by the chiral algebra.
Note that the highest weight condition (6.3) is saturated only for m = ± 12 R2 , and
furthermore this only happens for V∓R , not for any other operators in the chiral algebra.
These are therefore the only highest weights which are not annihilated by the zero mode
of V∓R . The modes of V∓R are defined in the usual way
I I
1 2
Vn,∓R = dzz h+n−1
V∓R (z) = dzz 2 R +n−1 V∓R (z) (8.5)
It is easy to verify that

Z
1 2
dzz 2 R −1 V∓R (z) ± 21 R
1
V0,∓R ± 2 R =
Z
1 2
dzz 2 R −1 V∓R (z)V± 1 R (0) |0i = ∓ 12 R

=
2
so that these two highest weight states are actually in the same representation of the
horizontal algebra.
This brings us then finally to the following characterization of the representations of
the chiral algebra. If the algebra is specified by a lattice Λ with spacing R, satisfying
R2 = 2N , then the representations are labelled by the integers m, −N < m ≤ N , and
have characters
1 X 1(m 2
χm (q) = q 2 R +nR) (8.6)
η(q) n
Note that we may define m modulo 2N , since a shift m → m + 2N = m + R2 can be
cancelled by a shift in the summation index n. It is sometimes convenient to choose m
in the range 0 ≤ m < 2N . There are in total 2N representations. The ground state
multiplicity for each of them except one is 1, the ground states being |pi = |m/Ri with
− 21 R2 < m < 12 R2 . The exception is the representation labelled
1 by m = N (∼ −N ). Here
the ground state multiplicity is two, because the states ± 2 R are degenerate.
8.6 The matrix S

This is the condition for T invariance. To examine S invariance we have to determine
first how the characters transform. We know this already for the η-function. To deal with
the infinite sum one can use a trick called Poisson resummation. Define
X 1 (µ +λ+z)2
fi (q, z) = q2 i , (8.7)
λ∈Λ
89
where µi = Ri , i = 0, . . . 2N −1 (Although these are not highest weights, this is a convenient
way of labeling all representations uniquely.) For z = 0 these functions are equal to the
numerators of the characters. This function is manifestly periodic under z → z + R since
one can shift the sum. Hence one can define its Fourier transform (we drop the argument
q here for simplicity)
1 R
Z
∗
fi (w) = dye2πiwy fi (y) (8.8)
R 0
with the inverse X
fi (z) = e−2πiρz fi∗ (ρ). (8.9)
ρ∈Λ∗
Now we substitute into (8.8) the function (8.7). The essential step is now to combine the
integral in (8.8) with the infinite sum in (8.7) to get an integral over the real line:
1 ∞
Z
∗ 2
fi (w) = dyeπiwy eπiτ (µi +y)
R −∞
This is a standard Gaussian integral, and yields
1 w2
fi∗ (w) = √ e−2πiµi w−πi τ
R −iτ
This can be substituted in (8.9) to get an expression for fi (τ, 0). From the latter we derive
immediately
1 √ X 1 2
fi (− , 0) = −iτ e−2πiµi ρ eπiρ τ
τ ρ∈Λ∗
R
The sum over ρ can be split in a sum over a set ρi = Ri , i = 0, . . . 2N − 1 and the lattice
Λ:
2N −1 X
1 √ X 1 −2πiµi ρj πi(ρj +λ)2 τ
fi (− , 0) = −iτ e e
τ j=0 λ∈Λ∗
R
Taking into account the η function we get finally
1 X
χi (− ) = Sij χi (τ )
τ j
with
1 −2πiµi µj 1 ij
Sij = e = √ e−2πi 2N
R 2N
This is a unitary, symmetric 2N × 2N matrix. It is not real, a reflection of the fact that
the theory does not have charge conjugation symmetry. Indeed, only the representations
i = 0 and i = N are self-conjugate.
Since the characters do indeed transform into each other, the diagonal partition func-
tion is indeed a modular invariant. Another modular invariant is defined by the charge
conjugation matrix C, which always commutes with S and T .
For a given R there are usually many more modular invariant partition functions.
90
8.7 Relation with circle compactification
The modular invariant partition functions we have found (without claiming uniqueness)
can be described most conveniently by introducing a new lattice Γ with momenta (pL , pR ).
This lattice contains all combinations of pL and pR that occur, and once we know it, we
know the full partition function:
1 X 2 2
P (τ, τ̄ ) = e2πiτ pL e−2πiτ̄ pR .
η(τ )η(τ̄ ) p ,p ∈Γ
L R
where the sum is over all vectors in the two-dimensional lattice. It is easy to show (again
using Poisson resummation) that this partition function is modular invariant if and only
if Γ is an even self-dual lattice with respect to the Lorentzian metric (−, +). Here “even”
means of course that for all lattice vectors p2L − p2R must be an even integer, and self-dual
means that Γ = Γ∗ (but with duality defined using the Lorentzian metric). One of the
conditions for modular invariance is locality. It is easy to verify that
0 0
VpL pR (z, z̄)Vp0L p0R (w, w̄) = (z − w)pL pL (z̄ − w̄)pR pR VpL +p0L ,pR +p0R + . . . ,
so that locality clearly requires that pL p0L − pR p0R ∈ Z. This follows indeed from the
condition that the lattice is Lorentzian even, by considering the vector (p − p0 ).
The momenta occurring in our partition functions are

i i i i
+ nR, + mR ; + nR, − + mR
R R R R
for the diagonal and charge conjugation invariant respectively. Here i lies in the range
0, 2N − 1 and n, m are arbitrary integers. It may be verified that this defines an even
self-dual Lorentzian lattice. One can also characterize these partition functions by two
unrestricted integers as
n n n n
, + mR ; , − + mR
R R R R
To make the result look more symmetric one can subtract mR/2 from both pL and pR
(i.e. one writes n = n0 − 12 R2 m) to get
n n n n
− 21 mR, + 12 mR ; + 12 mR, − + 12 mR ,
R R R R
where in the second term the lattice vector (mR, 0) was added. If we compare this to
(8.1) we see that the first partition function can be identified with it if L takes the values
mR. Because of the interpretation of L this implies that we are in a compact space with
radius R, defined by the lattice Λ. The momenta p in such a space must be such that
exp (ipx) respects the periodicity x → x + nR of that space, and this implies that p must
lie on the dual lattice.
91
8.8 R → 2/R duality
In interesting feature of these partition functions is duality. If one replaces R by 2/R and
interchanges the variables n and m (which are summed over in the partition function),
the two partition function (diagonal and charge conjugation) are switched. However these
two partition functions are indistinguishable, since charge conjugation does not change
the conformal weight, it only flips the U (1) charge. But our choice for the left and right
U (1) generator is just a convention. One arrives thus at the surprising conclusion that
two theories that are priori distinct are in fact indistinguishable.
8.9 Rationality
Note that earlier in this chapter we had found that R2 should be an even integer. However,
from the point of view of circle compactification it does not make any difference what R
is. There is an interesting subset of values of R for which the conformal field theory is
rational. This happens if the lattice contains vectors (pL , 0) or (0, pR ), which correspond
to operators in the chiral algebra. The condition for rationality is thus
n 1
+ mR = 0 ,
R 2
for at least one non-trivial set of integers. The most general solution is R2 = 2p/q, where
p/q can be any rational number.
Although we only constructed the special cases q = 1 and p = 1 (the latter is obtained
from duality) explicitly, all other cases can be obtained by constructing all other modular
invariant partition functions out of the characters.
Note that there is an infinite number of irrational values. Nevertheless, as far as exact
solvability is concerned these values are not worse than the rational ones.
The rational theories can all be obtained as modular invariant partition functions of
theories with the extended algebras of the form (8.2). The generators of this algebra are
thus
∂X; einRΦ n ∈ Z, n 6= 0 (8.10)
This are clearly the only operators we have at our disposal. These extended algebras are
characterized by a number R with R2 an even integer. It follows that if we allow rational
values of R it cannot be true that one should substitute those values in (8.10). This would
lead to non-integer conformal weights for the extended algebra generators. Instead, the
theory for other rational R values is realized as a non-diagonal modular invariant of a
theory satisfying R2 ∈ 2Z.
8.10 Theories with more than one free boson

All the foregoing results have a simple generalization to theories with more than one free
boson. The most general modular invariant partition function is described by a Lorentzian
even self-dual lattice ΓN,N with metric ((−)N , (+)N ) (this is called a Narain lattice [41]).
92
To get the most general theory of this kind from a compactification on on N -dimensional
torus requires the addition of a term to the Lagrangian, namely
Z
∼ d2 xBij αβ ∂α Φi ∂β Φj
where Bij is an arbitrary set of constants.

It is not hard to write down partition functions for these theories at arbitrary genus,
and check modular invariance.
8.11 Orbifolds
There are still more conformal field theories one can construct with one boson. From the
point of view of the target space interpretation the additional freedom consists of another
choice for the “manifold”. Most manifolds are unsuitable since the resulting theory would
not be conformally invariant. The torus is always a solution to these conditions, since it
is flat and affects the theory only via boundary conditions. In one dimension there is not
much choice, and the only proper manifold one can use is the circle. However it turns out
that one can still get sensible conformal field theories (and string theories) using spaces
that are not proper manifolds, but manifolds with singularities called orbifolds. This
notion was first used in heterotic string compactification [14, 15], but rapidly acquired a
much more general significance.
8.12 Orbifolds as singular manifolds

The definition of an orbifold is as follows. Consider a manifold which has a discrete
symmetry. Such a symmetry is said to act freely if it moves every point to a different
point. Now we define a new “manifold” by regarding points related to each other by the
symmetry as identical. If one uses a symmetry that does not act freely then the fixed
points of that symmetry introduce conical singularities. This object is not a manifold,
but is called an “orbifold”.
8.13 Orbifolds in conformal field theory

In conformal field theory the name “orbifold construction” is often used in a more general
sense for a method that allows one to modify conformal field theories by adding new
fields, while removing some others. In some cases this procedure has an interpretation
in terms of manifolds. There is no need to distinguish freely acting or non-freely acting
symmetries, although the latter are usually more difficult to deal with.
Intuitively the orbifold procedure implies the following changes to the theory
• Some states do not respect the discrete symmetry. They have to be removed from
the theory (they are “projected out”).
93
• Since some points are identified one can relax the boundary conditions of the boson.
Rather than Φi (x1 + 2π) = Φi (x1 ) (for the uncompactified boson), or Φi (x1 + 2π) =
Φi (x1 ) + 2πLi (for the boson on the torus), one must now also allow Φi (x1 + 2π) =
Gij Φj (x1 ) + 2πLi , where Gij is a matrix representing the symmetry. This implies
that new states are added to the theory. This new set of states is called the twisted
sector.
The two items mentioned above are closely related. Roughly speaking, a modular
invariant theory contains the maximal set of mutually local fields. They must be mutually
local to have T -invariance, and maximal for S-invariance. This same structure is seen in
the requirements “even” and “self-dual” that a modular invariant torus compactification
must satisfy. Thus if we remove some fields from a modular invariant theory, we can
only maintain modular invariance by adding some other fields. Such fields are called twist
fields [13, 32].
8.14 Orbifolds of the circle

In one dimension we have to consider the discrete symmetries of the circle. There are two
obvious Z2 symmetries, namely the “anti-podal map” and the reflection with respect to
some axis. The anti-podal map is a special case of an infinite series of ZN symmetries,
R
which can be realized by shifts Φ → Φ + 2π` N . These maps do not have fixed points. One
can use them in an orbifold construction, but one finds that they simply lead to a theory
on a circle with a different radius, and not to anything new. The reflection corresponds
to the symmetry Φ → −Φ. This map has two fixed points, Φ = 0 and Φ = πR (note that
−πR = πR because of the lattice identification), and does lead to a new series of theories.
We will discuss these theories here starting from the diagonal partition function of a
circle theory. Hence we will assume that R2 is an even integer. Other radii can then be
obtained by means of non-diagonal modular invariants.
8.14.1 The twist fields

It follows from the general reasoning that the twist fields must be non-local with respect
to the fields that are projected out.
The discrete symmetry acts by taking the conformal field ∂Φ to −∂Φ. Thus this
operator must be removed. This is done by introducing a twist field with respect to
which ∂Φ is non-local:
∂Φ(z)σ(w, w̄) = (z − w)∆h σ 0 (w, w̄) + . . . , (8.11)
where ∆h is non-integer. The branch cut must be such that it still respects the peri-
odicities of the new manifold. This means that when ∂Φ(z) is moved once around w it
can only change sign. Note that moving ∂Φ(z) around the origin on the complex plane
is related by a conformal mapping to moving ∂Φ once around the cylinder. In the latter
case ∂Φ can return to itself with or without a sign change. Hence we require that ∂Φ(z) in
94
the complex plane also changes by at most a sign when carried along a circle around the
origin. Whether or not there is a sign change depends on the state inserted at the origin
(the initial state in the cylinder picture). Since we want ∂Φ(z) to be non-local w.r.t. σ
we require that ∆h must be half-integer.
Just as we did for the free fermion, we can describe the construction in terms of an
intermediate partially modular invariant invariant theory, in which ∂Φ and the twist field
can co-exist. In this theory, if we require that σ is primary with respect to ∂Φ we find
that ∆h must in fact be − 21 (if it were not primary, we expect another operator to exist
which is primary, and which we would use instead.) One can then show (see [13]) that
1 9
hσ = 16 . Consequently hσ0 = 16 .
We will assume that there exists a modular invariant diagonal theory in which all
ground states have equal holomorphic and anti-holomorphic conformal weights. That
theory will ultimately be obtained by making a projection on our T 2 -invariant, but not
T -invariant theory. Such a theory must contain an operator σ whose anti-holomorphic
1
conformal weight is also 16 .
1
Note however that this implies that h̄σ0 = 16 , since ∂Φ(z) has h̄=0, and it cannot
introduce an anti-holomorphic branch cut. Hence the operator σ 0 has conformal weights
9 1
( 16 , 16 ) and must disappear in the modular invariant theory. This is consistent with
the operator product (8.11) since also ∂Φ(z) will not be an operator in the final theory,
because it is odd under the orbifold symmetry.
Now we also need a field σ̃ with the operator product
1
¯
∂Φ(z̄)σ̃(w, w̄) = (z̄ − w̄)− 2 σ̃ 0 (w, w̄) + . . . .
The conformal weights of these operators must, by arguments similar to the foregoing
1 1
ones, be ( 16 , 16 ) for σ 0 and ( 16
1 9
, 16 ) for σ̃ 0 . Clearly σ 0 6= σ̃ 0 , so that we clearly need at least
one new twist field. In fact it turns out we need two: both σ̃ and σ̃ 0 must be new fields.
8.15 Projecting on the invariant states

The partially modular invariant theory is now obtained by acting on these twisted ground
states with all combinations of oscillators (note that the oscillators are half-integer moded
in the twisted sector), and including all states in the untwisted sector.
Now we remove all states that are not invariant under the symmetry Φ → −Φ, and
the corresponding operators. In particular this removes the operator ∂Φ and hence the
branch cut (8.11) causes no problems anymore. Note that this symmetry changes the sign
of all the oscillators as well as the momentum operator.
We do not only wish to remove all states that are odd under the discrete symmetry, but
also organize the remaining ones into representations of the chiral algebra of the orbifold
theory. In particular this means that we write the new partition function in the standard
diagonal form. The chiral algebra of the orbifold theory does not contain the current ∂Φ,
but it does contain some other operators, namely the symmetric combinations
einRΦ + e−inRΦ , n>0.
95
√
The operator of lowest conformal weight in this set has conformal weight 21 R2 . For R = 2
this current has spin 1.
The ground states in the untwisted sector transform as follows. We start from the
diagonal partition function of the circle theory, which is created by oscillators acting on
the states |m, mi with −N < m < N , as well as the states |±N, ±N i. The latter four
come from the terms |χN |2 in the diagonal partition function. Here the notation is as in
X.X, i.e. m denotes a representation with ground state momentum m/R.
A state |m, mi (m 6= 0, m 6= ±N ) is mapped to |−m, −mi, so that only the linear
combination |m, mi + |−m, −mi is left in the orbifold theory. At the first excited level
there were four states, α−1 |m, mi ᾱ−1 |m, mi, α−1 |−m, −mi and ᾱ−1 |−m, −mi. In this
case the linear combinations α−1 |m, mi − α−1 |−m, −mi and ᾱ−1 |m, mi − ᾱ−1 |−m, −mi
survive the projection. These two states are created from the ground state by the mode
L−1 of the energy-momentum tensor − 21 (∂Φ)2 and its anti-holomorphic partner.∗ Thus
we see that the structure of the lowest lying states is consistent with a contribution to
the partition function of the form |q h−c/24 (1 + q + . . .)|2 , the square of a single character.
With some more work one can show that this structure persists to higher excitation levels.
Thus for each value of m in the range 0 < m < N we find precisely one representation of
the orbifold chiral algebra.
The states with charges ±N are slightly more subtle. Of the four states |±N, ±N i
two linear combinations survive, namely |+i = |N, N i + |−N, −N i and |−i = |N, −N i +
|−N, N i. These two states are mapped into each other by the operator J = eiRΦ + e−iRΦ ,
but the linear combinations |+i±|−i are eigenstates of J. They form two separate ground
states, each of one representation. This is as it should be: in the diagonal partition func-
tion ground states are represented by the square of a character, and hence the multiplicity
of any ground state |h, hi must be a square. If one finds a ground state with multiplicity
2, it must be obtained as 1 + 1, since 2 is not a square.
The vacuum sector also requires more attention. Here we have to distinguish two
cases. For R2 > 2 the first excited states are are α−1 |0, 0i and ᾱ−1 |0, 0i. They are both
odd under the symmetry Φ → −Φ and disappear. However, the symmetric excitation
α−1 ᾱ−1 |0, 0i does survive. This contribution to the diagonal partition function starts
thus as [q q̄]−c/24 (1 + q q̄), and does not factorize (the circle partition function has as its
−c/24 2
leading terms |q (1 + q)| .) Hence we are forced to introduce a new ground state,
¯
denoted as ∂Φ, ∂ Φ̄ , that corresponds to the circle state α−1 ᾱ−1 |0, 0i. It may then be
shown that all further excitations factorize in a sum of two terms, one corresponding the
vacuum representation and one to the representation built on the ground state ∂Φ, ∂¯Φ̄ .

√
If R = 2 the circle chiral algebra contains 3 spin 1 currents, ∂Φ and exp (±iRΦ), and
hence the leading terms in the circle partition function are [q q̄]−c/24 (1 + 3q + 3q̄ + 9q q̄).
Only one of the three currents survive the projection, and of the nine current-current
states five survive. Hence the orbifold partition function starts with
[q q̄]−c/24 (1 + q + q̄ + 5q̄) = [q q̄]−c/24 |1 + q + . . . |2 + [q q̄]1−c/24 |2 + . . . |2

∗
Note that L−1 contains a term ∝ α0 α−1 , and that α0 |m, m0 i = m |m, m0 i.
96
In this case the ground state of the representation denoted “∂Φ, ∂¯Φ̄ ” has thus multi-

plicity 2, and contributes to the full partition function with multiplicity 4.

In the twisted sector we have to define the action of the symmetry on the ground
states; then the rest is fixed. The unprimed and primed twist fields must transform with
1 1
an opposite sign, as is clear from (8.11). Since σ(0) |0, 0i = 16 , 16 is a desirable state
0 9 1

and σ (0) |0, 0i = 16 , 16 is not, we choose σ (as well as σ̃) to transform with a + sign.

Then the state σ 0 (0) |0, 0i transforms with a − sign and is removed, while for example
ᾱ−1/2 σ 0 (0) |0, 0i with conformal weight h = h̄ = 16
9
remains. Since we have removed σ 0 as
9 9
well as ∂φ we need a new operator that creates the state 16 , 16 from the vacuum. We
will call this operator once again σ 0 .

To summarize, we find thus the following partition function
X
Porb = |χr |2
r
where the label r stands for the following representations. The notation is inspired by the
foregoing discussion in an obvious way, but note that here we are only considering one
chiral sector.
r h r h
0 0 ∂Φ 1
1 1
σ 16
σ̃ 16
σ0 9
16
σ̃ 0 9
16
N N
N(1) 4
N(2) 4
m2
m (0 < m < N ) 4N
There are in total 1 + 1 + 4 + 2 + N − 1 = N + 7 representations. Each has ground

state multiplicity 1, except r = ∂Φ for R2 = 2, as noted above.
8.17 The geometric description

Although the presentation given above was a bit intuitive and not completely rigorous, it
is not hard to show that it actually leads to a modular invariant partition function. The
partition function we were constructing can be summarized as follows
Porb = 21 PBPP + PBPA + PBAA + PBAP

Here PB represents the free boson path integral on the (world-sheet) torus, with boundary
conditions as indicated. The sum over boundary conditions is as for the free fermion, and
is modular invariant for the same reason.
97
The term PBPP is the circle partition function. The second term is anti-periodic in the
time direction, which means that odd numbers of bosons contribute with a − sign. The
third and fourth term are anti-periodic in the space direction, and represent the twisted
sector. The combination of terms projects out the unwanted states in that sector. The
first term and the sum over the last three terms are separately modular invariant. The
precise combination of these two modular invariant sets is dictated by the requirement
of having a unique vacuum and positive integral multiplicities for all other states. In
particular a relative − sign between these modular invariant sets (which was allowed for
the free fermion) is not allowed here because it would project out the vacuum.
The partition function may also be written as
Porb = TrP 21 (1 + g)q L0 −c/24 q̄ L̄0 −c̄/24 + TrA 12 (1 + g)q L0 −c/24 q̄ L̄0 −c̄/24 ,
where g represents the non-trivial Z2 element that sends Φ to −Φ. This formula has an
immediate generalization to arbitrary discrete abelian groups, often written suggestively
as
1 X
Porb = g ,
|G| g,h∈G
h
where |G| is the number of elements in the group G. The sum over h is over all possible
twisted sectors, whereas the sum over g performs the projections. Modular invariance of
this expression is intuitively clear.
The advantage of this formulation applied to the c = 1 orbifolds is that it works
immediately for arbitrary (even non-rational) R. The disadvantages is that it does not
give direct information on the chiral algebra and the representations.
8.18 The c = 1 conformal field theories

We have now identified two sets of c = 1 conformal field theories, each parametrized by
a real number R. Furthermore there is a duality in both spectra, since√ R and 2/R are
giving rise to the same spectrum. The self-dual point occurs at R = 2.
One may think that the orbifold and the circle theories√are all different, but in fact
they are
√ not. It can be shown that the orbifold of the R = 2 theory and the circle with
R = 2 2 describe
√ one and the same theory. This cannot happen at any other point, since
only for R = 2 the orbifold theory has a spin-1 current. It is easy to verify that the
spectra of the orbifold and circle theories are indeed the same, and not much harder to
show that they are in fact the same theory.
Hence the two lines are not separate, but connected, as shown in fig. 7 [12, 4]. Note
that only the topology of the picture matters, not the geometry. The dashed lines indicate
values of R that have already been taken into account because of R → 2/R duality. The
orbifold radius is denoted by Ro . Apart from this continuum there also exist three√isolated
theories. They can be obtained by an orbifold procedure applied to the R = 2 circle
theory [25]. With these points included the picture is conjectured to be complete.
98
Orbifold
Isolated
CFT’s
Ro
p p
Ro = 2 Rc = 2 2
Circle
Rc
p
Rc = 2
(Self-dual point)
Figure 7: Moduli space of the c = 1 conformal field theories.
8.19 Moduli and marginal deformations

This picture provides the simplest example of moduli in conformal field theory. Moduli
are free parameters which can be varied continuously without affecting conformal invari-
ance. Apart from the three isolated points, every point on the diagram corresponds to a
conformal field theory with moduli. The point where the circle and the orbifold meet is
characterized by the existence of an additional modulus.
One can detect the existence of such conformal invariant deformations within a given
theory by looking for conformal fields of dimension (1, 1), called marginal operators. Such
operators have precisely the correct weights to yield a conformal invariant result when
integrated over dz and dz̄.∗ This implies that they can be added as a perturbation to the
action, Z
δS ∝ dzzdz̄V1,1 (z, z̄) ,
where V1,1 is a marginal operator.

In the circle and orbifold theories, this operator is ∂Φ(z)∂Φ(z̄). The additional
marginal operator in the √ meeting point of the lines is due to combinations with the
additional spin-1 field ei 2Φ .
∗
Marginal operators must satisfy additional constraints not explained here.
99
9 Kac-Moody Algebras
In this chapter we consider extensions of the chiral algebra by a set of (anti)-holomorphic
spin-1 currents.
9.1 Spin one operator products

These currents are conformal fields with respect to the Virasoro algebra. The operator
product of two such currents in a modular invariant conformal field theory must be local.
Since the currents are holomorphic their operator product is holomorphic as well. Hence
it is an expansion of integer powers of (z − w) multiplied by integer spin operators. Since
the lowest spin an operator in a unitary conformal field theory can have is zero (the
identity), the leading term is a constant times (z − w)−2 . The next one is (z − w)−1 times
a holomorphic spin-1 operator, which must therefore be one of the currents. Hence we
get
dab if abc J c
J a (z)J b (w) = + + ... (9.1)
(z − w)2 z−w
Note that the next term has spin 2, and hence is a candidate for a Virasoro generator; we
will return to it later.
Since integer spin currents are bosons the left-hand side is symmetric under interchange
(z, a) ←→ (w, b). It follows that dab must be symmetric and f abc anti-symmetric in a and
b. Since f abc appears in the three-point function it must then be anti-symmetric in all
three indices (this is true provided a Hermitean basis is chosen). Using duality relations
one can then show that the coefficients f abc must satisfy Jacobi identities. It follows then
that they are structure constants of a Lie algebra. This Lie algebra must be a direct
product of some simple Lie algebras and optionally some U (1) factors. The arguments
given aboves are due to A. Zamolodchikov [60]; we refer to this paper for further details.
9.2 Intermezzo: some Lie algebra facts

We will fix some standard normalizations for simple Lie algebras. The algebra is
a b
T , T = if abc T c
It is satisfied in particular by the matrices

a
(Tadj )bc = −if abc ,
which are the generators of the adjoint representation. Their commutator is in fact nothing
but the Jacobi identity. The generators in the adjoint representation act on the algebra
via the commutator.
The root system is defined by selecting out of the generators T a the maximally com-
muting set H i , the Cartan subalgebra. The number of such generators is called the rank
of the algebra. Hence the label i takes values i = 1, . . . , N . In the adjoint representation
100
we may simultaneously diagonalize the Cartan subalgebra acting on the remaining ones,
so that i α~
H , E = αi E α~ (9.2)
The set of vectors α
~ is called the root system of the algebra. If we denote the dimension
of the algebra as D, then the number of roots α ~ is D − r. In a compact∗ Lie algebra, a
basis can be chosen so that
a b
Tr Tadj Tadj = N δ ab , (9.3)
where N is a normalization, to be fixed in a moment. The left-hand side is called the
Killing metric of the Lie algebra. It will be assumed that the Cartan subalgebra generators
are elements of the basis. The basis then consists of r generators H i and D − r generators
E α~ .
Given a root system we can choose a plane which divides the roots into positive and
negative ones. This plane must be chosen in such a way that none of the roots lies in
it. Then one defines a set of simple roots which form a basis of the root system with the
property that all positive roots are linear combinations with positive coefficients of the
simple roots. One also defines a highest root ψ as the unique positive root from which all
other roots can be obtained by subtracting simple roots. Now we define the dual Coxeter
number g:
~ + 2~
(ψ ~
ρ) · ψ
g= ,
~·ψ
ψ ~
where ρ~ is half the sum of the positive roots. Note that this definition is independent of
the normalization of the inner product. The values of the dual Coxeter number for all
simple Lie algebras are listed in the following table:
Typerank Algebra Value of g Adjoint dimension

AN −1 SU (N ) N N2 − 1
1
B N −1 SO(N ), N > 4, odd N −2 2
N (N − 1)
2
CN Sp(2N ) N +1 N (2N + 1)
1
DN SO(N ), N > 3, even N −2 2
N (N − 1)
2
G2 4 14
F4 9 52
E6 12 78
E7 18 133
E8 30 248
∗
By a “compact Lie-algebra” we mean the Lie algebra of a compact Lie group, whose group manifold is
a compact space. Most simple Lie algebras also have non-compact forms, related to the compact forms
by multiplying a suitable subset of the generators by i. Examples are the compact Lie algebra SO(N ),
and its non-compact forms SO(p + q), p > 0, q > 0; p + q = N . In compact Lia algebras all generators
can be chosen Hermitean.
101
The first column gives the Dynkin classification, while the second one gives the identifi-
cation with the perhaps more familiar classical Lie algebras.
We now fix the normalization of the generators by requiring that N = 2g in (9.3).
This normalization implies that the highest root has norm 2. To see why, note that the
value of N can be related to the norms of the roots. First we go to a Hermitean basis by
defining E α~ (1) = E α~ + E −~α , E α~ (2) = −i(E α~ − E −~α ). There are two such generators for
~
each positive root. Now (9.2) defines the structure constants f i,~α(1),β(2) = αi δα~ β~ , where
we label the generators by labels i = 1, . . . , r and α ~ (1), α~ (2). These constants are indeed
anti-symmetric. Now we can compute
~ ~
X X
Tr T i (adj)T j (adj) = −f i,~α(1),β(2) f j,β(2),~α(1)
α β
positive positive
~ ~
−f i,~α(2),β(1) f j,β(1),~α(2)
X
= 2 αi αj = N δ ij
α
positive
Contracting this with δ ij we find

1X 2
N= α
~ ,
r α
where the sum is now over all (positive and negative) roots. Clearly this fixes the overall
scale in the root system. It can be shown (for example by explicit computation) that
X
~2 .
~ 2 = rg ψ
α
α
so that N = 2g implies that ψ 2 = 2.

To conclude this section we write down the remaining commutators among the gener-
ators in this basis. For the commutator between the root generators one has
h i
α
~ ~
β ~ α~ +β~ ,
E , E = (~α, β)E
~ + β~ is a root, and
if α
~ ,
α~ −~α
E ,E ~ ·H
=α
and zero in all other cases. The coefficients (~ ~ are non-zero real numbers.
α, β)
9.3 The central term

The first tensor in (9.1) must be symmetric in a and b, and furthermore the Lie algebra
structure we have just identified requires it to be an invariant tensor of the Lie algebra.
Hence it must be proportional to the Killing form, which in our conventions means it
102
is proportional to δ ab . Since we have already fixed the normalization of the structure
constants, the normalization of the first term is fixed. Note that the first term determines
the current-current propagator, and that this has a positive residue only if the Lie-algebra
is compact (if it is not compact the Killing form has negative eigenvalues). If the prop-
agator had a wrong-sign residue this would violate unitarity. Thus in unitary conformal
field theories the Lie algebra must be compact.
If the Lie algebra is semi-simple the term dab takes the form k a δ ab , where k a is constant
on each simple factor. From now on we will focus on simple Lie algebras; the index a on
k a can then be dropped.
9.4 Modes
The mode expansion of the currents is as discussed in general in chapter 6. It is straight-
forward to derive the algebra in terms of modes
a b
Jm , Jn = if abc Jm+n
c
+ kmδ ab δm+n,0 (9.4)
Note that for m = n = 0 one obtains a subalgebra which is a simple Lie-algebra. Since
the modes with m = n = 0 do not alter the conformal weight, this algebra takes the states
of a given weight into each other. It is usually referred to as the horizontal algebra.
If k = 0 the algebra is referred to as the loop algebra. If k 6= 0 one gets strictly
speaking only an algebra if we consider k as the eigenvalue of an operator K, which is
called the central extension of the loop algebra. This operator commutes with all others.
This is analogous to the central extension of the Virasoro algebra.
The algebra (9.4) is called a centrally extended loop algebra, or current algebra. It
is often also referred to as an affine Lie algebra, or a Kac-Moody algebra. This is not
quite correct. The mathematical definition of an affine Lie-algebra includes in addition
to the operators appearing in (9.4) still one more operator called the derivation D. This
operator satisfies [D, Jna ] = nJna , and [K, D] = 0. Comparing the first expression with
(6.2), one finds that it is satisfied by D = −L0 ; because of (9.4), L0 commutes with K
and hence the second commutator is also satisfied. Since we will only consider the spin-1
current algebras in combination with a Virasoro algebra, the distinction between the two
definitions is not essential for us. Note that the current algebra is unaffected if we omit D,
since it never appears on the right hand side of a commutator, but from the mathematical
point of view it is convenient to introduce it in order to define an invertible Killing form.
The mathematical definition of a Kac-Moody algebra is much more general, and includes
ordinary as well as affine Lie algebras, and many more. We will nevertheless use the term
“Kac-Moody” algebra from here on in a restricted sense, to refer to (9.4).
9.5 Twisted and untwisted affine algebras

Since the current has integral spin, the “natural” mode expansion is in terms of integer
modes. One can however also consider fractionally moded operators by introducing twist
fields. One find that in many cases the fractionally mode algebras are isomorphic to the
103
integrally moded ones. There is a set of algebras and twistings (related to so-called outer
automorphisms of the horizontal Lie-algebra) for which this is not the case. They are
known as twisted affine algebras. In these lectures we will only encounter the untwisted
ones.
9.6 Primary fields

Primary fields are defined by the condition that they should be Virasoro primary fields,
and in addition satisfy
a i
Tija
J (z)Φ (w, w̄) = Φj (w, w̄) + . . .
z−w
The leading pole is determined as in the general arguments given in chapter 6. Since the
field appearing on the right hand side has the same conformal weight as Φ, one can label
all the fields with that conformal weights by a label i, and then the operator product
inevitably looks like the one above.
This implies that the ground states |ri are rotated into each other by the horizontal
algebra, which acts via the matrices Tija :
J0a |ri i = Tija (r) |rj i , (9.5)
where
|ri i = Φi (0) |0i
The matrices Tija (r) can be shown to satisfy the commutation relations of the horizontal
algebra, a b
T , T = if abc T c ,
by acting with a second generator J0b . They are the representation matrices of the hori-
zontal algebra in some representation r determined by Φi .
Note that the current itself is not a Kac-Moody primary field, just as the energy
momentum tensor is not a conformal field.
9.7 The Sugawara tensor

In addition to the current modes the algebra under consideration consists of Virasoro gen-
erators, with definite commutation relations with themselves and the currents. Actually,
there is one as yet unknown quantity in the Virasoro algebra, namely its central charge.
It turns out that the Virasoro generators can be expressed in terms of the currents in the
following way:
1 X
T (z) = : J a (z)J a (z) : , (9.6)
2(k + g) a
where the sum is over all generators of the horizontal algebra. This is called the Sug-
awara energy-momentum tensor [52]. As usual, normal ordering means subtraction of the
104
singular terms,
" #
X X kdim (adj)
: J a (z)J a (z) :≡ lim J a (z)J a (w) −
a
w→z
a
(z − w)2
For U (1) algebras dim (adj) should be interpreted as the number of U (1) generators. To
verify that this is indeed the Virasoro generator, we have to check the operator product
with the current, and with T (w). The requirement that J a (z) is a conformal field of
weight 1 fixes the normalization in (9.6). In the computation one uses the relation
−f acd f bdc = Tr Tadj

a b
Tadj = 2gδ ab
Then the computation of T (z)T (w) serves as a check, but in addition determines the
central charge:
kdim (adj)
c=
k+g
The Virasoro generators can be expressed in terms of the modes of the currents:
∞
1 X
Ln = : Ja Ja : ,
2(k + g) m=−∞ m+n −m
where normal ordering means that positive modes should appear to the right of negative
ones.
9.8 Highest weight representations

Highest weight representations are characterized by a ground state |ri that is annihilated
by all positive modes of Jn . This implies automatically that it is annihilated by all
positive modes of the (Sugawara) energy-momentum tensor, i.e. that it is a Virasoro
highest weight.
The only remaining freedom we have in characterizing representations is the action of
the zero-mode generator J0a . We have already seen before that the ground states form a
representation r of the horizontal algebra generated by the zero-modes. Representations
of simple Lie algebras are themselves generated by step operators acting on highest weight
vectors. This implies that any irreducible unitary representation of a Kac-Moody algebra
is completely characterized by a highest weight vector of the horizontal algebra and the
eigenvalue of the operator K, called the level (k). Completely, because once we know the
horizontal algebra highest weight and k we know the action of all current modes and the
Virasoro generators.
In particular we know the conformal weight of the ground state:
hr| J0a J0a |ri

P
h= a (9.7)
2(k + g)
105
The expectation value can be computed using (9.5):
X X
hri | J0a J0a |rj i = [T a (r)T a (r)]ij = C2 (r)δij .
a a
Here i and j label the components of the representation r, and C2 (r) is the quadratic
Casimir operator. The result is thus
1
C (r)
2 2
hr =
k+g
Note that our normalization is such that in the adjoint representation C2 (adj) = 2g.
The representation r must be an irreducible highest weight representation of the hori-
zontal algebra. What remains to be done is to determine which representations and which
values of k are allowed. Rather than attempting to solve this directly in general, we start
by looking at the simplest theories.
9.9 U (1) theories

If all structure constants f abc vanish one obtains a product of one or more U (1) factors.
Their currents can always be written in terms of free bosons, J i = i∂Φi . They satisfy
the operator product (9.1) with k = 1. We have already studied this case in detail, and
discuss it only here to show how it fits in.
Since f abc = 0, the dual Coxeter number g vanishes. Then the energy-momentum
tensor has the standard form for free bosons, T (z) = − 12 (∂Φ(z))2 . The central charge is
equal to the number of free bosons, as expected.
The representations are labelled by the zero-mode momenta of the ground states,
usually referred to as charges. The ground states satisfy thus
J0 |qi = q |qi ,
and they are in fact uniquely labelled by q. Their conformal weight is 21 q 2 . Note that
J0 = p, the momentum operator.
9.10 The SU (2) Kac-Moody algebra

The root system of SU (2) has just one simple root α. This is also the only positive root,
and is also equal to the highest root. The Weyl vector ρ is equal to half the sum of the
positive roots, and is thus equal to 21 α. The dual Coxeter number is easily computed to
be 2.
The algebra is generated by three currents J a , a = 1 . . . 3. The structure constants
are proportional to abc . The proportionality constant can be determined by (9.3), which
reads
(−ixacd )(−ixbdc ) = 2gδ ab = 4δ ab ,
106
√
where x is the proportionality constant. We find thus that x = 2. This is a disadvantage
of this normalization: SU (2) generators are not normalized
√ i in the familiar way. Similarly
1
the generators in the spinor representation are 2 2τ , where τ i are the Pauli matrices.
[An advantage of our normalization is that for any algebra and any representation the
quantity I2 (R), defined by TrT a T b = I2 (r)δ ab is an integer.]
Highest weight representations of the SU (2) Kac-Moody algebra are characterized by
SU (2) Lie-algebra representations and the level k; hence they are characterized by k and
the SU (2) spin j. A ground state has 2j + 1 components |j, mi. Its conformal weight is
j(j + 1)
h=
k+2
Here we recognize the SU (2) Casimir eigenvalue j(j + 1).
The following argument restricts the values of k. The algebra (9.4) has several inter-
esting sub-algebras. One is the zero-mode algebra,
a b √
J0 , J0 = i 2abc J0c .
Apart from the normalization this is a standard SU (2) algebra. Since we want to use
results from SU (2) representation theory, we have to change the normalization of the
generators. Furthermore we go to a basis of raising/lowering operators. Hence we define
1 1
I ± = √ (J01 ± iJ02 ); I3 = √ J3 ,
2 2
so that [I + , I − ] = 2I 3 . Standard results in SU (2) unitary representation theory tell us now
that the eigenvalues of I 3 must be (half)-integers. It is easy to check that the following
generators satisfy the same commutation relations:
I˜+ = √1 (J 1 − iJ 2 ); I˜− = √1 (J 1 +
2 +1 +1 2 −1
2
iJ−1 );
I˜3 = 2 k − √2 J 3 = 2 k − I 3 .
1 1 1
Hence we conclude that the eigenvalues of I˜3 must also be (half)-integers, and furthermore
since I 3 and I˜3 commute we can diagonalize them simultaneously. This is only consistent
if k is an integer. Furthermore unitarity (positivity of the residue of the propagator)
requires it to be a positive integer.
Now we can directly get a further constraint by computing the norm of the state
˜ −
I |j, mi, where |j, mi is one of the components of the ground state
0 ≤ hj, m| I˜+ I˜− |j, mi

h i
= hj, m| I˜+ ,̃I − |j, mi
= hj, m| 2I˜3 |j, mi
= hj, m| (k − 2I 3 ) |j, mi
= hj, m| (k − 2m) |j, mi
107
Here we used the requirement of unitarity (positivity of the norm), the highest weight
a
property of |j, mi, which implies that J+1 |j, mi = 0, and the SU (2) commutator [I + , I − ] =
2I 3 . Clearly m cannot be larger than 21 k, and the same follows then for j. It is convenient
to label the representations by integers l = 2j. They are thus restricted to the values
0 ≤ l ≤ k.
9.11 SU (2) at level 1.

For k = 1 there are thus precisely 2 representations, with ground state spins j = 0 and
1
2
. We have already seen a realization of this theory, namely in the self-dual point of√the
c = 1 circle theory. At this point there are three spin-1 fields, namely ∂Φ and e±i 2Φ .
Their operator products are (singular terms only)
√
√ ∓i 2 ±i√2Φ
±i 2Φ
∂Φ(z)e (w) = e (w)
(z − w)
1
∂Φ(z)∂Φ(w) = −
(z − w)2
√ √
ei 2Φ(z) i 2Φ(w)
e = non-singular
and √
√ √ 1 i 2
i 2Φ(z) −i 2Φ(w)
e e = + ∂Φ(w)
(z − w)2 z − w
These is precisely equal to (9.1) with k = 1 provided we define
√ √ √ √ √ √
J 1 (z) = 12 2(ei 2Φ + e−i 2Φ ) , J 2 (z) = − 12 i 2(ei 2Φ − e−i 2Φ ) ,
J 3 = i∂Φ .
Thus we see that this algebra can be realized with a single free boson. We have already
seen in the previous chapter that for R2 = 2N the bosonic theories have 2N characters
m2
with conformal weights 4N , −N < m < N . For N = 1 this agrees with the SU (2) level-1
description of the same theory.
The primary field corresponding to the only non-identity
√ representation can also be
written in terms of the free boson, namely as exp (i 12 2Φ(z). Unfortunately things are
less simple at higher levels.
9.12 Generalization to other Kac-Moody algebras

The foregoing results on SU (2) have an immediate generalization to other algebras. This
generalization works exactly like the reasoning one follows to derive the Lie-algebra rep-
resentations from the representation theory of SU (2). The results for SU (2) are valid
for any SU (2) sub-algebra of some Kac-Moody algebra, and now it is simply a matter of
finding the most suitable one.
108
Let us first find a suitable basis for the current modes Jna . For the zero modes there
is a standard basis, the one introduced in section 9.2.
To generalize this to Kac-Moody algebras one simply attaches an extra index n to all
operators, and includes the central term. The result is
i
[Hm , H j ] = mδm+n,0 δ ij
i n α~
Hm , En = αi Em+n α
~
h i ~
α
~ ~ (~
~ n+mα+β)
Em , Enβ = (~ α, β)E
~ n+m + Knδn+m,0 .
α~ −~α
En , Em = α ~ ·H
It is easy to see that any root

qα ~ defines a conventionally
q normalized SU (2) subalgebra,
whose generators are I˜+ = 2 α ˜−
−~
2 En , I
α
~
= 2 α
~ ˜
2 E−n and I3 =
α
~
1
2 (Kn − α
α
~
~ · H0 ). The
normalization of this SU (2) is the traditional one, i.e. [I + , I − ] = 2I 3 etc. By arguments
similar to the ones used for SU (2) we conclude that the quantity 2nK/(~ α2 ) must have
integer eigenvalues, for any n and α ~ . Obviously the strongest constraint comes from
n = 1. If we have normalized our root system in the canonical way, i.e. ψ ~ 2 = 2, there
is always a root with norm 2, and we find that K must have integer eigenvalues k. The
norms of other roots that can occur in simple Lie algebras are 1 or 32 , in the canonical
normalization. This does not impose additional constraints.
One can use the same subalgebra to find constraints on the ground states. We know
already that the ground states are representations of the horizontal algebra, and are
characterized by a highest weight ~λ. The ground state has then dim (r~λ ) components,
where r~λ indicates the representation with highest weight ~λ.
Take any component |µi, where µ is any weight in r~λ . By requiring positivity of the
norm of I˜− |µi we get now the condition
~ ·µ
α ~
2 ≤k.
~ ·α
α ~
This condition is most restrictive if we take µ equal to the highest weight of the ground
state representation ~λ, and ψ equal to the highest root. In the canonical normalization
we get then
ψ~ · ~λ ≤ k
Just as for SU (2) the number of representations satisfying this condition is finite. Fig. 8
shows the allowed highest weights for the algebra A2 at various levels
a
The negatively moded currents J−n act on these ground states and create the full
Kac-Moody representation. Since they are in the adjoint representation of the horizontal
algebra, they also change the representation that one finds at higher excitation levels. The
excitation level, also called grade is defined as the conformal weight of a descendant minus
that of the ground state. It should not be confused with the level of the algebra. Naively
109
k=5
k=4
k=3
k=2
k=1
Figure 8: Highest weights of the algebra A2 .
the representation content at the higher excitation levels can be obtained by selecting all
combinations of current modes that produce the desired excitation level, and tensoring
the ground state representation with the adjoint representation as many times as required.
For example, one might expect the first excited level to contain all representations in the
tensor product rλ ⊗ rψ , the latter being the adjoint representation. However, the norms
of some of the representations in the tensor product might be 0, just as was the case for
Virasoro representations.
Zero-norm states are removed. Nothing in the previous arguments guarantees the
absence of negative norm states, which would make the representation non-unitary. The
conditions we have satisfied are necessary conditions for the absence of some potential
negative norm states, namely those occurring in certain SU (2) subalgebras. One way to
show that the representations are indeed unitary is to find an explicit realization of the
symmetries in some well-defined field theory.
110
9.13 The Frenkel-Kac construction
One such realization is the Frenkel-Kac construction [17, 51]. This is a generalization of
the level-1 construction which we gave for SU (2). It works for any Lie-algebra whose roots
have the same length, which is conveniently normalized to the value 2. Such a Lie-algebra
is called simply-laced, and the algebras enjoying this property are Ar , Dr and E6 , E7 and
E8 .
The generators of these algebras at level 1 can be written down explicitly in terms of
r free bosons, where r is the rank. One simply writes them as
E α~ (z) = ei~α·Φ (z)
and
H i (z) = i∂Φi (z) ,
and defines modes in the usual way. This yields the correct operator product for the
SU (2) sub-algebras associated to each of the roots (as one may check), and furthermore
one gets
~ ~
ei~α·Φ (z)eiβ·Φ (w) = (z − w)α·β ei(~α+β)·Φ (w) + . . .
Inner products between roots of simply laced algebras can be 2, 1, 0, −1 and −2. In the
first case α ~ and in the last case α
~ = β, ~ If α
~ = −β. ~ · β~ = −1 one finds that α~ + β~ is a root.
Precisely in that case the operator product has a pole, exactly as required by (9.1).
However, this is not quite the end of the story, because the coefficients (~ ~ can
α, β)
have signs. Although many of these signs are merely a conventions, some are essential.
To reproduce them one has to introduce so-called co-cycle factors in the definition of the
root generators, whose commutators produce the correct signs. We will not discuss this
further.
The Frenkel-Kac construction yields thus an explicit realization of level-1 simply laced
algebras in terms of free bosons. The lattice on which the momenta of these bosons are
quantized is the weight lattice of the simply-laced algebra, which is the dual of the root
lattice.
9.14 The WZW-model

Realizations of the other theories can be obtained from the so-called Wess-Zumino-Witten
models. These are conformal field theories with a two-dimensional action
S = k[SW Z + SW ]
The first term is due to Wess and Zumino [58], and has the form
Z
1
SW Z = d2 z Tr∂α g(z)∂ α g(z) ,
16π
where g(z) is a map from the two-dimensional surface to a group manifold. In other
words, for every point z on the manifold, g(z) is some element of the group G under
consideration. Here G can be any compact group belonging to a simple Lie algebra.
111
The second term was added by Witten [59], and has the bizarre form
Z
1
SW = d3 yαβγ Trg −1 (y)∂α g(y)g −1 (y)∂β g(y)g −1 (y)∂γ g(y)
24π
The strange feature is that the integral is over a three-dimensional surface. However
the integral is a total derivative, and hence it can be written as a surface integral over
the boundary of the three-surface. The boundary of a three-surface is a two-dimensional
manifold, for which we take the one used in the first term, with the boundary condition
g(y)|y=z = g(z). The extra term is required to make the theory conformally invariant.
Upon quantizing the theory one finds that k must be an integer for the integral to be
consistent.
The currents that generate the Kac-Moody algebra for this model are J(z) = ∂gg −1
¯ = g −1 ∂g.
and J(z̄) ¯
9.15 Modular transformation properties

Virasoro characters for representations of Kac-Moody algebras can be defined in the usual
way. It is however useful to define a more general quantity, namely
~ ~
~ = TrV e2πiτ (L0 −c/24) e2πiθ·H0 .
Xλ,k (τ, θ) λ,k
Here the trace is over all states in the representation with highest weight λ and level k.
If we put the variables θi to zero this reduces to the Virasoro character.
A general formula for the characters and their transformation properties was given by
Kac and Peterson [34]. The result is
~ = e2πi(hλ,k −ck /24) Xλ,k (τ, θ)

Xλ,k (τ + 1, θ) ~ , (9.8)
with h and c as defined earlier, and
1 θ~ ~2 X
−ikθ
k ~
Xλ,k (− , ) = e 4πτ Sλλ 0 Xλ0 ,k (τ, θ)
τ τ λ0
A very important feature is that different levels do not mix under modular transforma-
tions. This could have been expected on the basis of the WZW-model (which has a
definite level and can be defined on the torus).
Formulas for the matrix elements of S can be found in the literature. Many important
results on Kac-Moody algebras are due to V. Kac, in collaboration with various other
authors. These results are collected in a book [33], but this is not easily accessible. The
formulas for S can be found for example in [23] or [20].
To compute the matrix S numerically, one may use the program kac [46]. This program
can also be used to compute fusion rules, modular invariant partition functions (especially
simple current invariants) and matrices S of coset CFT’s.
112
9.16 Modular invariant partition functions for SU (2)
An as yet unsolved problem is that of finding all modular invariant partition functions
for all WZW-models. That is, one wants to find all non-negative integer matrices Mλ,λ0
that commute with S and T (the latter is implicitly defined in (9.8) and with M00 = 1,
so that the vacuum is unique.
The only horizontal algebras for which this problem has been solved completely are
SU (2) and SU (3). For SU (2) the solutions are divided into three types called A, D and
E:
• A: These are simply the diagonal invariants, which exist at any level, and for any
algebra.
• D: They occur at all even levels. If the level is a multiple of 4, they imply and
extension of the chiral algebra. For the other even levels they correspond to auto-
morphisms of the fusion rules.
• E: They occur for level 10, 16 and 28.
The notation is chosen because the solutions resemble the classification of the simply-
laced Lie-algebras. The resemblance is more precise than suggested here, but so far there
is no deep understanding of the mathematical structure (if any) behind this observation.
The A and D invariants are explicitly
k
X
|Xl |2
l=0
k/4−1
X
|X2m + Xk−2m |2 + 2|Xk/2 |2 (k = 0 mod 4)
m=0
k
X k
X
2 ∗
|Xl | + Xl Xk−l (k = 2 mod 4)
l=0,even l=0,odd
This is called the ADE-classification of the SU (2) modular invariants. It was obtained
and shown to be complete by Cappelli, Itzykson and Zuber [7].
9.17 Fusion rules and simple currents

The fusion rules can be derived using Verlinde’s formula. There is also a more direct
approach which is a modified version of the tensor product rules of the horizontal Lie
algebra.
Such a tensor product has the general form
X
ri ⊗ rj = Mij l rl ,
113
where Mij l gives the multiplicity of the representation rl in the tensor product of ri and
rj . For example, in SU (3) one has the rule
(8) ⊗ (8) = (1) + (10) + (10) + 2(8) + (27)
Here representations are indicated by their dimension, and the bar indicates the complex
conjugate. The coefficients Mij l are somewhat reminiscent of the fusion rule coefficients.
Indeed, it is true that Nij l ≤ Mij l , with equality in limit of infinite level (for fixed i, j
and k). For example, these are the results for SU (3) at various level, with [n] indicating
a Kac-Moody representations whose ground state is the Lie-algebra representation (n):
k=2: [8] × [8] = [1] + [8]
k=3: [8] × [8] = [1] + [10] + [10] + 2[8]
k=4: [8] × [8] = [1] + [10] + [10] + 2[8] + [27] .
For higher levels the result is as for k = 4. For k = 1 the ground state [8] does not exist.
One method for finding these results starts with the group theory tensor products, to
which certain level-dependent projections are applied.
Most Kac-Moody algebras have simple currents. They are the representations whose
ground state highest weight is k times a so-called co-minimal fundamental weights. The
only exception is E8 level 2, which has a simple currents even though it has no fundamental
weights at all.
For SU (2) the simple current is the representation with j = k. For SU (N ) they are
all N representations with Dynkin labels (0, . . . , 0, k, 0, . . . , 0), etc.
9.18 Modular invariant partition functions for other Kac-Moody

algebras
No complete classification exists, although it seems plausible that at least for simple
horizontal algebras the present list of solutions is close to complete. The majority of the
invariants on that list are simple current invariants. For example for SU (2) all D-type
invariants are simple current invariants. Only the three exceptional invariants remain
mysterious. This is also the pattern one observes for other algebras.
9.19 Coset conformal field theories

A huge class of rational conformal field theories can be obtained with the coset con-
struction [27, 28]. Consider a Kac-Moody algebra G and another Kac-Moody algebra H.
Suppose the horizontal sub-algebra of H can be embedded in that of G, Then one can
associate a conformal field theory with any such pair G and H. For simplicity we will
assume that both horizontal algebras are simple.
The embedding implies that one can write the currents of H in terms of those of G:
X
JHi (z) = Mai JGa (z)
a
114
Substituting this into the operator product (9.1) one finds
kMai Mbj δ ab iMai Mbj f abc c

JHi (z)JHj (w) = + JG (w)
(z − w)2 z−w
The fact that we have an embedding in the horizontal algebra implies that in the last
term the identity Mai Mbj f abc = f ijl Mcl can be used to get f ijl J l (w), and that in the first
term Mai Mbj δ ab ∝ δ ij . However, in general there is a proportionality coefficient, which is
called the Dynkin index of the embedding. This index, which we denote I(G, H) is an
integer. We find thus the following relation for the level of G and H:
kH = I(G, H)kG
If H is not simple, one simply attaches a label to H to indicate the simple factors; if
G is not simple one does the same, and includes on the right-hand side a sum over the
G-labels.
The energy-momentum tensor of the coset conformal field theory is TG (z) − TH (z),
where TG and TH are the Sugawara tensors for G and H, each at the appropriate level.
The currents of H are spin-1 conformal fields with respect to TH ; on the other hand, they
are linear combinations of currents of TG , and hence they are also spin-1 conformal fields
with respect to TG . But that implies that the operator product of TG (z) − TH (z) with JHi
is non-singular, since the singularities cancel. Furthermore, since the Sugawara tensor TH
is constructed completely out of the currents of H, it follows that the operator product
(TG (z) − TH (z))TH (w) is non-singular, or in other words TG (z)TH (w) = TH (z)TH (w) up
to non-singular terms. The same is true for TH (z)TG (w). Hence we get
(TG (z) − TH (z)) (TG (w) − TH (w))

= TG (z)TG (w) − TH (z)TH (w)
cG − cH TG (w) − TH (w) ∂w (TG (w) − TH (w))
= +2 + + ...
(z − w) 4 (z − w)2 (z − w)2
This tells us that TG − TH is a Virasoro generator with central charge cG − cH that is
“orthogonal” to TH in the sense that their operator product is non-singular. Hence the
original energy-momentum tensor TG has been decomposed into two orthogonal pieces
TG = TG/H + TH ,
with TG/H = TG − TH .
Given such a decomposition, any representation of G can be decomposed in terms of
H representations,
X
V (λG ) = ⊕λH VH (λH ) ⊗ VG/H (λG , λH ) .
Here λH labels all representations of the Kac-Moody algebra H, and V (λ) denotes a
representation space. Each single state in the G representation is a product of some state
115
in an H-representation times a state in a G/H representation. In this way we define the
representation spaces for the coset theory. Note that TG/H = TG − TH realizes a unitary
representation on this space. This follows from unitarity of the modes of TG and TH (in
the sense that L†n = L−n ) as well as the fact that the norm of states in G representations
are equal to products of norms of states in H and G/H representations. Hence the norms
of the latter cannot be negative.
Naively, we can explicitly construct the characters of the coset theory by decomposing
any G representation systematically into H representations. This corresponds to the
following relation X λ
XλG (τ ) = bλGH (τ )XλH (τ ) (9.9)
λH
The functions bλλGH (τ ) are called the branching functions of the embedding. They are
sometimes confused with the characters of the coset theory, but in general this is not
correct. The relation (9.9) does not give sufficient information to compute the branching
functions. To compute them one has to take into account not only the dependence on τ
of the characters, but use also the representation content with respect to the horizontal
algebra.
9.20 The minimal discrete series as a coset theory

An interesting example is the series
SU (2)1 × SU (2)k
SU (2)k+1
The central charge is

6
c=1− ,
(k + 2)(k + 3)
which corresponds precisely to the central charges of the minimal Virasoro models if we
make the identification m = k +2. Since the minimal models are the only unitary theories
with these central charges (apart from non-diagonal modular invariants of these theories)
the coset theories must form an explicit realization of the minimal models. This is quite
useful, because we had not proved that the minimal models are actually unitary, we just
had not been able to rule them out.
Let us compute some of the branching functions. The representations of G are labelled
by two integers 0 ≤ l1 ≤ 1 and 0 ≤ l2 ≤ k, and those of H by one integer 0 ≤ l3 ≤ k + 1.
Let us consider l1 = l2 = 0. The ground state of the G Kac-Moody representation is then
the Lie-algebra representation (0, 0). It decomposes to (0) of H. The branching function
a
starts thus at hG − hH = 0. At the next excitation level we encounter the states (J−1 )1 |0i
a
and (J−1 )2 |0i, generated from the vacuum by the currents of SU (2) × SU (2). There are
six states, and they transform in the representation (3) + (3) of H.
a
In the vacuum representation of H we will have also a set of states J−1 |0i, in the (3) of
SU (2). This removes one of the (3)’s we found. The other is not a singlet, and hence can
116
not contribute to the branching function b00,0 . It must thus be interpreted as the first term
in a new branching function b0,0
2 , where “2” denotes the representation (3) (in general the
dimension is (l + 1), since l denotes twice the usual SU (2) spin). The leading conformal
weight in that branching function is
1
l (1l
2 3 2 3
+ 1) 1
0+0+1− = 2
,
(k + 1) + 2
where the first two terms are the ground state weight in G, the third is the excitation
level, and the last is the contribution from the term −TH , with l3 = 2 and k = 1. This
branching function is seen to correspond to the h = 12 representation of the Ising model.
There is no contribution at the first excited level to b0,0
0 . This agrees with the fact that
L−1 |0i = 0 on the ground state.
9.21 Field Identification

The complications with interpreting the branching functions as characters start becom-
ing clear as soon as one observes that for example the branching function b10,0 is identi-
cally zero, since the G-representation contains only integer spin representations of SU (2).
Closely related, but less obvious, is the fact that several branching functions are in fact
identical. Something like this had to happen, since the total number of branching func-
tions one gets for the coset SU (2)1 ⊗ SU (2)1 /SU (2)2 is 2 × 2 × 3 = 12. This exceeds the
number of Ising model representations by a factor of 4.
The solution is that only the following branching functions are non-vanishing, and
that they are identical in pairs:
b0,0 1,1
0 = b2 h=0
b1,0 0,1
1 = b1
1
h = 16
b0,0 1,1
2 = b0 h= 1
2
This phenomenon is called field identification [40, 22].

In this case it is still true that the branching functions are equal to the characters.
However, in other cases it happens that the number of fields that is identified is not
always the same. In that case there are non-trivial problems [37]. The solution is beyond
the scope of these lectures, and partly beyond the scope of what is presently known.
However, it is certainly true that in these cases the characters are not simply equal
to the branching functions. This problem occurs frequently, for example in the cosets
SU (2)k ⊗SU (2)l /SU (2)k+l whenever k and l are both even. For a more detailed discussion
of this problem see [48] and [21].
9.22 Other coset models

It should be clear that the set of coset models is huge. Most of them have a central charge
larger than 1, and are example of rational conformal field theories with an extended
117
algebra. For example, it was shown that the series
SU (3)1 × SU (3)k
SU (3)k+1
has a chiral algebra with currents of spin 3, and forms the minimal series of the W3 algebra
(which will not be discussed here further).
The number of coset models is so large that it has even been suggested that in combi-
nation with free bosonic theories and orbifolds and perhaps some other ideas it exhausts
the set of rational conformal field theories. Unfortunately this “conjecture” has never
been made sufficiently precise to disprove it. Any claims that rational conformal field
theories have in some – usually vague – sense been classified should be regarded with a
great amount of suspicion. In fact even rational conformal field theories with a single
primary field are essentially unclassifiable.
10 Superconformal Algebras
There is yet another important class of extensions of the chiral algebra, namely by currents
of spin 32 . Since these are half-integer spin currents, many of the remarks we made in the
section on fermionic currents are valid here as well. In particular there are two sectors,
Neveu-Schwarz and Ramond, and there may be square root branch cuts in operator
products.
The name “superconformal” refers to the fact that a spin- 32 current can be put in a
supermultiplet together with the energy-momentum tensor. The currents of this algebra
generate the so-called superconformal transformations, a supersymmetric generalization
of conformal transformations. Indeed, one can describe the entire algebra in a manifestly
supersymmetric way, but we will write it in terms of components.
10.1 The N = 1 algebra

The simplest superconformal algebra is generated by a single spin- 32 current TF (z) in
addition to the Virasoro generator. This is the N = 1 superconformal algebra. The
complete set of operator products is
1
2
c 2 1
T (z)T (w) = + T (w) + ∂T (w)
(z − w)4 (z − w) 2 z−w
3
2 1
T (z)TF (w) = TF (w) + ∂TF (w)
(z − w) 2 z−w
1 1
6
c
TF (z)TF (w) = + 2 ∂T (w)
(z − w) 3 z−w
The first two operator products simply state that T (z) is the energy-momentum tensor
and TF (z) a spin 32 conformal field.
118
Modes are defined as in section 6. The modes of the supercurrent are traditionally
called Gn . The algebra in terms of modes looks like this
c
[Lm , Ln ] = (m − n)Lm+n + (m3 − m)δm+n,0
12
[Lm , Gr ] = ( 12 m − r)Gm+r
{Gr , Gs } = 2Lr+s + 31 c(r2 − 14 )δr+s,0
In the last term we find an anti-commutator because the left-hand side of the correspond-
ing operator product is odd under the exchange z ↔ w. This is also exactly like the free
fermion.
The fermionic currents G can be half-integer-moded (Neveu-Schwarz) or integer moded
(Ramond). To emphasize this we have used indices r and s for this current. A new feature,
in comparison with the free fermion, is that it is now possible that a Ramond ground state
is annihilated by G0 (because of the anti-commutator {b0 , b0 } = 1 this is impossible for
c
the free fermion). Because of the last relation, this implies immediately that h = 24 for
such a state. These states are often called chiral states. Furthermore any state which is
c
not annihilated by G0 must have h > 24 . Note that the latter ground states necessarily
come in pairs of opposite fermion number, related by G0 , whereas the ones annihilated
by G0 are unpaired. This also implies that in superconformal theories the trace in the
Ramond sector with (−1)F projection may be non-zero, unlike the free fermion case. In
fact this trace clearly receives contributions only from the chiral states with h − c/24 = 0,
so that the corresponding terms in the partition function are constants. This implies
in particular that this contribution to the partition function (which corresponds to the
PP-sector) is modular invariant by itself.
In the Neveu-Schwarz sector one should note the relation
{Gr , G−r } = 2L0 + 31 c(r2 − 41 )
Since r2 ≥ 41 the left-hand side is positive or zero, with the latter value occurring only for
r = 12 and h = 0. If the left-hand side is positive we have
|G−r |xi |2 > 0
for ground states. Hence the excitations have positive norm. There is a unique ground
state with the property G−1/2 |xi = 0, namely the vacuum (note that ground states in
any case satisfy Gr |xi = 0, r ≥ 21 ).
The unitary representations of this algebra form a discrete series for 0 ≤ c < 3/2,
whereas for larger values of c there are infinitely many representations, just as for the
Virasoro algebra. The c-values for this series are

3 8
c= 2 1− , m = 3, 4, . . .
m(m + 2)
The m = 3 value is c = 7/10, and coincides with a member of the minimal Virasoro series.
Obviously superconformal representations are in particular representations of the Virasoro
algebra. The second member is on the c = 1 boundary of the Virasoro representations.
119
A concrete realization of this series is given by the coset models
SU (2)2 × SU (2)k
.
SU (2)k+2
10.2 The N = 2 algebra

There are also superconformal models with extended supersymmetry [57, 11, 9, 10, 6, 50].
The case of most interest is N = 2, since it occurs in supersymmetric string theories. In
these theories there are two supercurrents. Note that just having two supercurrents is not
yet enough, since a tensor product of two N = 1 models would also have that property,
and one would not expect it to have extended supersymmetry. To get an N = 2 algebra
the currents need to satisfy a set of operator products. Furthermore it turns out that the
algebra must contain one additional current J of spin 1. This current generates a U (1)
algebra.
The full algebra is, in terms of modes:
c
[Lm , Ln ] = (m − n)Lm+n + (m3 − m)δm+n,0
12
Lm , G± 1 ±

= ( m − r)G
− +r 2 m+r
Gr , Gs = 2Lr+s − (r − s)Jr+s + 13 c(r2 − 41 )δr+s,0
[Lm , Jn ] = −nJm+n
[J , J ] = 31 cmδm+n,0
m ±n
Jm , Gr = ±G± m+r
This algebra also has a discrete series, with central charges

2
c = 3(1 − ), m = 3, 4, . . .
m
The first member of this series has c = 1. It is also in the N = 1 series, and can be
realized as a circle compactification of a single free boson (with R2 = 2N = 12). The
central charges turn out to be identical to those of the SU (2) Kac-Moody algebras, if one
substitutes m = k + 2. This is related to the fact that the minimal series can be obtained
from the following cosets
SU (2) × SO(2)
.
U (1)
Ground states are characterized by a conformal weight h and a U (1) charge q. In
addition one can have chiral states both in the Neveu-Schwarz and in the Ramond sector.
In the Neveu-Schwarz they have the special property
G+
−1/2 |φi = 0 or G−
−1/2 |φi = 0
and are called respectively chiral or anti-chiral states. Primary states satisfy the condition
−
G+r |φi = Gr |φi = 0 for r > 0. Chiral primary (or anti-chiral primary) states satisfy the
corresponding combination of these conditions.
120
Using the algebra (as for N = 1 above) it is easy to deduce that for chiral primaries
h = 21 q, and for anti-chiral primaries h = − 12 q. The only state that is both chiral and
anti-chiral primary is thus the vacuum. Furthermore it can be shown that any other state
in the theory has h > 12 |q|, and that the conformal weights of chiral primaries satisfies
h ≤ c/6.
An interesting consequence of the relation between charges and conformal weights is
that within the set of chiral primary states conformal weights are “conserved” in operator
products just like charges. Consider the operator product of two chiral primary fields φ1
and φ2 (ignoring anti-holomorphic components). Then
φ1 (z)φ2 (w) = (z − w)h3 −h1 −h2 φ3 (w) + less singular terms.
The charge of φ3 is q1 + q2 , and therefore h3 ≥ 21 (q1 + q2 ) = h1 + h2 (note that chiral

primaries have positive charges). Hence the operator product is non-singular. Therefore
we can define
φ1×2 (z) = lim φ1 (z)φ2 (w) .
w→z
This limit is zero if φ3 is not a chiral primary state, and is equal to φ3 if it is a chiral
primary. Hence this defines a closed operation on the chiral primary states. This is called
the chiral ring [37]. There is of course also an anti-chiral ring.
In the Ramond sector one defines chiral states as those which are annihilated by both
−
G+0 and G0 . From the anti-commutator of these two operators one learns that those are
precisely the states with h − c/24 = 0.
An important property of N = 2 algebras is spectral flow. This means that there
exists an operator Uθ that maps the entire algebra to an isomorphic one. It acts on the
generators by conjugation, and the mapping has the following effect
c
Uθ Ln Uθ−1 = Ln + θJn + θ2 δn,0
6
−1 c
Uθ Jn Uθ = Jn + θδn,0
3
± −1 ±
Uθ Gr Uθ = Gr±θ
The interesting feature of this map is that it changes the mode of the supercurrent. Closer
inspection shows that for θ = 12 it maps the Neveu-Schwarz moded algebra to the Ramond
moded algebra. It is not difficult to show that chiral primary states are mapped to the
chiral Ramond grounds states, while the latter are mapped to the anti-chiral states by the
same map. This shows in particular that there is a one-to-one correspondence between
chiral Ramond ground states and (anti)-chiral states in the Neveu-Schwarz sector. In
string theory this is related to space-time supersymmetry, as the Neveu-Schwarz sector
yields space-time bosons and the Ramond sector space-time fermions; one of the conditions
for having space-time supersymmetry is N = 2 supersymmetry in two dimensions.
121
References
[1] G. Anderson and G. W. Moore. Rationality in Conformal Field Theory. Commun.
Math. Phys., 117:441, 1988.
[2] I. Antoniadis, C. Bachas, and C. Kounnas. Four-Dimensional Superstrings.

Nucl.Phys., B289:87, 1987.
[3] D. F. B. Feigin. Anal. Pril., 16:47, 1982.
[4] K. Bardakci, E. Rabinovici, and B. Saering. String Models with c < 1 Components.
Nucl. Phys., B299:151, 1988.
[5] R. Blumenhagen, D. Lüst, and S. Theisen. Basic concepts of string theory. Springer
Verlag, 2013.
[6] W. Boucher, D. Friedan, and A. Kent. Determinant Formulae and Unitarity for
the N=2 Superconformal Algebras in Two-Dimensions or Exact Results on String
Compactification. Phys. Lett., B172:316, 1986.
[7] A. Cappelli, C. Itzykson, and J. B. Zuber. Modular Invariant Partition Functions in

Two-Dimensions. Nucl. Phys., B280:445–465, 1987.
[8] J. L. Cardy. Conformal Invariance And Statistical Mechanics. In Les Houches Sum-
mer School in Theoretical Physics: Fields, Strings, Critical Phenomena Les Houches,
France, June 28-August 5, 1988, pages 0169–246, 1989.
[9] P. Di Vecchia, J. L. Petersen, and M. Yu. On the Unitary Representations of N=2

Superconformal Theory. Phys. Lett., B172:211, 1986.
[10] P. Di Vecchia, J. L. Petersen, M. Yu, and H. B. Zheng. Explicit Construction of

Unitary Representations of the N=2 Superconformal Algebra. Phys. Lett., B174:280,
1986.
[11] P. Di Vecchia, J. L. Petersen, and H. B. Zheng. N=2 Extended Superconformal

Theories in Two-Dimensions. Phys. Lett., B162:327, 1985.
[12] R. Dijkgraaf, E. P. Verlinde, and H. L. Verlinde. C = 1 Conformal Field Theories on

Riemann Surfaces. Commun. Math. Phys., 115:649–690, 1988.
[13] L. J. Dixon, D. Friedan, E. J. Martinec, and S. H. Shenker. The Conformal Field

Theory of Orbifolds. Nucl. Phys., B282:13–73, 1987.
[14] L. J. Dixon, J. A. Harvey, C. Vafa, and E. Witten. Strings on Orbifolds. Nucl.Phys.,

B261:678–686, 1985.
[15] L. J. Dixon, J. A. Harvey, C. Vafa, and E. Witten. Strings on Orbifolds. 2. Nucl.Phys.,

B274:285–314, 1986.
122
[16] V. A. Fateev and A. B. Zamolodchikov. Parafermionic Currents in the Two-
Dimensional Conformal Quantum Field Theory and Selfdual Critical Points in Z(n)
Invariant Statistical Systems. Sov. Phys. JETP, 62:215–225, 1985. [Zh. Eksp. Teor.
Fiz.89,380(1985)].
[17] I. B. Frenkel and V. G. Kac. Basic Representations of Affine Lie Algebras and Dual
Resonance Models. Invent. Math., 62:23–66, 1980.
[18] D. Friedan, Z.-a. Qiu, and S. H. Shenker. Conformal Invariance, Unitarity and Two-
Dimensional Critical Exponents. Phys. Rev. Lett., 52:1575–1578, 1984.
[19] D. Friedan, S. H. Shenker, and Z.-a. Qiu. Details of the Nonunitarity Proof for
Highest Weight Representations of the Virasoro Algebra. Commun. Math. Phys.,
107:535, 1986.
[20] J. Fuchs. Affine Lie algebras and quantum groups: An Introduction, with applications
in conformal field theory. 1992.
[21] J. Fuchs, B. Schellekens, and C. Schweigert. The resolution of field identification

fixed points in diagonal coset theories. Nucl.Phys., B461:371–406, 1996.
[22] D. Gepner. Field Identification in Coset Conformal Field Theories. Phys. Lett.,
B222:207, 1989.
[23] D. Gepner and E. Witten. String Theory on Group Manifolds. Nucl. Phys., B278:493,
1986.
[24] P. H. Ginsparg. Applied Conformal Field Theory. In Les Houches Summer School in
Theoretical Physics: Fields, Strings, Critical Phenomena Les Houches, France, June
28-August 5, 1988, 1988.
[25] P. H. Ginsparg. Curiosities at c = 1. Nucl. Phys., B295:153–170, 1988.
[26] F. Gliozzi, J. Scherk, and D. I. Olive. Supergravity and the Spinor Dual Model.
Phys.Lett., B65:282, 1976.
[27] P. Goddard, A. Kent, and D. I. Olive. Virasoro Algebras and Coset Space Models.
Phys. Lett., B152:88, 1985.
[28] P. Goddard, A. Kent, and D. I. Olive. Unitary Representations of the Virasoro and
Supervirasoro Algebras. Commun. Math. Phys., 103:105–119, 1986.
[29] P. Goddard and D. I. Olive. Kac-Moody and Virasoro Algebras in Relation to Quan-
tum Physics. Int. J. Mod. Phys., A1:303, 1986.
[30] J. F. Gomes. The Triviality of Representations of the Virasoro Algebra With Van-
ishing Central Element and L(0) Positive. Phys. Lett., B171:75, 1986.
123
[31] M. B. Green, J. Schwarz, and E. Witten. Superstring Theory, Vol. 1 and 2. Cambridge
University Press, 1987.
[32] S. Hamidi and C. Vafa. Interactions on Orbifolds. Nucl. Phys., B279:465, 1987.
[33] V. Kac. Infinite Dimensional Lie Algebras. Cambridge University Press, 1985.
[34] V. G. Kac and D. H. Peterson. 112 Constructions Of The Basic Representation Of

The Loop Group Of E(8). In Proceedings, Anomalies, Geometry, Topology*, 276-298,
1986.
[35] H. Kawai, D. C. Lewellen, and S. H. Tye. Construction of Four-Dimensional

Fermionic String Models. Phys.Rev.Lett., 57:1832, 1986.
[36] W. Lerche, A. Schellekens, and N. Warner. Lattices and Strings. Phys.Rept., 177:1,
1989.
[37] W. Lerche, C. Vafa, and N. P. Warner. Chiral Rings in N=2 Superconformal Theories.
Nucl.Phys., B324:427, 1989.
[38] G. W. Moore and N. Seiberg. Polynomial Equations for Rational Conformal Field
Theories. Phys. Lett., B212:451, 1988.
[39] G. W. Moore and N. Seiberg. Naturality in Conformal Field Theory. Nucl. Phys.,
B313:16, 1989.
[40] G. W. Moore and N. Seiberg. Taming the Conformal Zoo. Phys. Lett., B220:422,
1989.
[41] K. Narain. New Heterotic String Theories in Uncompactified Dimensions. Phys.Lett.,

B169:41, 1986.
[42] A. Neveu and J. H. Schwarz. Factorizable dual model of pions. Nucl. Phys., B31:86–
112, 1971.
[43] A. M. Polyakov. Quantum Geometry of Bosonic Strings. Phys. Lett., B103:207–210,

1981.
[44] A. M. Polyakov, A. A. Belavin, and A. B. Zamolodchikov. Infinite Conformal Sym-

metry of Critical Fluctuations in Two-Dimensions. J. Statist. Phys., 34:763, 1984.
[45] P. Ramond. Dual Theory for Free Fermions. Phys. Rev., D3:2415–2418, 1971.
[46] A. Schellekens. The program Kac; http://www.nikhef.nl/∼t58/Site/Kac.html.
[47] A. Schellekens. Introduction to conformal field theory. Fortsch.Phys., 44:605–705,

1996.
124
[48] A. Schellekens and S. Yankielowicz. Simple Currents, Modular Invariants And Fixed
Points. Int.J.Mod.Phys., A5:2903–2952, 1990.
[49] B. Schellekens, editor. Superstring construction. 1989.
[50] A. Schwimmer and N. Seiberg. Comments on the N=2, N=3, N=4 Superconformal
Algebras in Two-Dimensions. Phys. Lett., B184:191, 1987.
[51] G. Segal. Unitarity Representations of Some Infinite Dimensional Groups. Commun.

Math. Phys., 80:301–342, 1981.
[52] H. Sugawara. A Field theory of currents. Phys. Rev., 170:1659–1662, 1968.
[53] C. B. Thorn. Computing the Kac Determinant Using Dual Model Techniques and
More About the No - Ghost Theorem. Nucl. Phys., B248:551, 1984.
[54] C. Vafa. Toward Classification of Conformal Theories. Phys. Lett., B206:421, 1988.
[55] E. P. Verlinde. Fusion Rules and Modular Transformations in 2D Conformal Field

Theory. Nucl. Phys., B300:360, 1988.
[56] M. S. Virasoro. Subsidiary conditions and ghosts in dual resonance models. Phys.
Rev., D1:2933–2936, 1970.
[57] G. Waterson. Bosonic Construction Of An N=2 Extended Superconformal Theory

In Two-Dimensions. Phys.Lett., B171:77, 1986.
[58] J. Wess and B. Zumino. Consequences of anomalous Ward identities. Phys. Lett.,
B37:95, 1971.
[59] E. Witten. Nonabelian Bosonization in Two-Dimensions. Commun. Math. Phys.,

92:455–472, 1984.
[60] A. B. Zamolodchikov. Infinite Additional Symmetries in Two-Dimensional Confor-

mal Quantum Field Theory. Theor. Math. Phys., 65:1205–1213, 1985. [Teor. Mat.
Fiz.65,347(1985)].
125

CFT PDF

Uploaded by

CFT PDF

Uploaded by

Conformal Field Theory

[Word cloud by www.worldle.net]

2 Quantum Conformal Invariance 21

3 Virasoro Representation Theory 32

5 Conformal Field Theory on a Torus 52

6 Extensions of the Virasoro Algebra 64

8 Free Boson Partition Functions 84

10 Superconformal Algebras 118

1.1.2 Weyl invariance

1.2 Conformal transformations in d dimensions

P(x) P(z) P(y)

1.5 The conserved current

Jµ () = Tµν ν (1.28)

This current is conserved because

∂ µ Jµ () = (∂ µ Tµν )ν + Tµν (∂ µ ν ) . (1.29)

1.6 An example: the free boson

δ log g = log det (gµν + δgµν ) − log det gµν

1.7 The conformal algebra in two dimensions

Going to complex variables,

z → f (z), z̄ → f¯(z̄) , (1.36)

where f (z) is an arbitrary function of z.

generates the transformation

[Ln , Lm ] = (n − m)Lm+n . (1.39)

[Ln , L̄m ] = 0 . (1.40)

1.9 The global subgroup

P1 = −i(∂z + ∂z̄ ) = i(L−1 + L̄−1 )

The algebra satisfied by the holomorphic generators is

[L0 , L−1 ] = L−1

generator local transformation global transformation

Combining these transformations we get

1.10 Tensors in complex coordinates

z ≡ xz = x1 − ix2 ; z̄ ≡ xz̄ = x1 + ix2 (1.47)

The inverse is then

z ≡ x1 = 21 (xz + xz̄ ) = 21 (z + z̄) ; x2 = 12 i(xz − xz̄ ) = 21 i(z − z̄) (1.48)

The transformation of a vector to complex coordinates goes as follows

∂z̄ Tzz = ∂z Tz̄z̄ = 0 , (1.50)

which implies that Tzz is holomorphic and Tz̄z̄ anti-holomorphic.†

Since Jz is holomorphic and Jz̄ anti-holomorphic, this current is manifestly conserved:

1.11 Conformal fields in two dimensions

1.12 Relation to string theory

1.13 Free bosons in complex coordinates

Tzz = − 12 (∂z Φ(z))2 (1.57)

• An extra factor 2π multiplying the definition of T µν (cf. footnote on page 24 of part

2 Quantum Conformal Invariance

2.1 Radial quantization

φcyl (z(w), z̄(w̄)) = (iw)h (−iw̄)h̄ φplane (w, w̄) (2.9)

∂ µ Jµ = 2 [∂z Jz̄ (z̄) + ∂z̄ Jz (z)] = 0 . (2.11)

A(xa , ta )B(xb , tb ) = eiHta A(xa , 0)e−iHta eiHtb A(xb , 0)e−iHtb (2.12)

A correlation function in field theory on the cylinder has the form

h0| T (A1 (t1 ) . . . An (tn )) |0i (2.15)

h0| R (A1 (z1 , z̄1 ) . . . An (zn , z̄n )) |0i (2.16)

where |0i and h0| are states at z = 0 and z = ∞ respectively.

2.3 The generator of conformal transformations

δ Φ(w, w̄) = h∂w (w)Φ(w, w̄) + (w)∂w Φ(w, w̄) (2.19)

This can be written as

Now we deform the contours as indicated in the following figure

2.4 Quantization of the free boson

implies that they satisfy periodic boundary conditions:

The classical equation of motion for Φ is

2.5 The free boson propagator

αki |0i = 0 for k > 0 (2.40)

The commutator yields

2.7 Operator products for free bosons

3 Virasoro Representation Theory

3.2 Highest weight representations

[L0 , Ln ] = −nLn (3.5)

3.3 The vacuum

Ln |0i = 0 , for n ≥ −1 . (3.6)

3.4 Positivity of c and h

3.5 States and conformal fields

We may project out a term from this sum by

3.7 The Kac determinant

3.8 The Kac determinant at level 2

det K2 = 2 16h3 − 10h2 + 2h2 c + hc hh|hi2 .

This can be written as

det K2 = 32(h − h11 )(h − h12 )(h − h21 )hh|hi2 , (3.19)

Jµ () = Tµν ν (1.28)

∂ µ Jµ () = (∂ µ Tµν )ν + Tµν (∂ µ ν ) . (1.29)

δ Φ(w, w̄) = h∂w (w)Φ(w, w̄) + (w)∂w Φ(w, w̄) (2.19)

δ φ = ∂φ + h∂φ (4.6)

If we restrict Li to SL2 (C), can be 1, z or z 2 .

[(z1 )∂1 + h1 ∂(z1 ) + (z2 )∂2 + h2 ∂(z2 )] G(z1 , z2 ) = 0 , (4.8)