Boundary Conformal Field Theory∗

arXiv:hep-th/0411189v2 20 Feb 2008

John Cardy
Rudolf Peierls Centre for Theoretical Physics
1 Keble Road, Oxford OX1 3NP, U.K.
and All Souls College, Oxford.

February 21, 2008

Boundary conformal field theory (BCFT) is simply the study of conformal

field theory (CFT) in domains with a boundary. It gains its significance
because, in some ways, it is mathematically simpler: the algebraic and ge-
ometric structures of CFT appear in a more straightforward manner; and
because it has important applications: in string theory in the physics of
open strings and D-branes, and in condensed matter physics in boundary
critical behavior and quantum impurity models.
In this article, however, I describe the basic ideas from the point of view of
quantum field theory, without regard to particular applications nor to any
deeper mathematical formulations.
∗
To appear in Encyclopedia of Mathematical Physics, J.-P. Françoise, G. Naber and
T.S. Tsun, eds. (Elsevier, 2005.)

1
Boundary conformal field theory 2

1 Review of CFT

1.1 Stress tensor and Ward identities

Two-dimensional CFTs are massless, local, relativistic renormalized quantum
field theories. Usually they are considered in imaginary time, i.e. on two-
dimensional manifolds with euclidean signature. In this article the metric
is also taken to be euclidean, although the formulation of CFTs on general
Riemann surfaces is also of great interest, especially for string theory. For
the time being the domain is the entire complex plane.
Heuristically the correlation functions of such a field theory may be thought of
as being given by the euclidean path integral, that is, as expectation values of
products of local densities with respect to a Gibbs measure Z −1 e−SE ({ψ}) [dψ],
where the {ψ(x)} are some set of fundamental local fields, SE is the euclidean
action, and the normalization factor Z is the partition function. Of course,
such an object is not in general well-defined, and this picture should be seen
only as a guide to formulating the basic principles of CFT which can then
be developed into a mathematically consistent theory.
In two dimensions, it is useful to use so-called complex coordinates z =
x1 + ix2 , z̄ = x1 − ix2 . In CFT there are local densities φj (z, z̄), called
primary fields, whose correlation functions transform covariantly under con-
formal mappings z → z ′ = f (z):
′ h̄j
f ′ (zj )hj f (zj ) hφ1 (z1′ , z̄1′ )φ2 (z2′ , z̄2′ ) . . .i ,
Y
hφ1 (z1 , z̄1 )φ2 (z2 , z̄2 ) . . .i = (1)
i

where (hj , h̄j ) (usually real numbers, not complex conjugates of each other)
are called the conformal weights of φj . These local fields can in general be
normalized so that their two-point functions have the form
hφj (zj , z̄j )φk (zk , z̄k )i = δjk /(zj − zk )2hj (z̄j − z̄k )2h̄j . (2)
They satisfy an algebra known as the operator product expansion (OPE)

cijk (z1 −z2 )−hi −hj +hk (z̄1 −z̄2 )−h̄i −h̄j +h̄k φk (z1 , z̄1 )+· · · ,
X
φi (z1 , z̄1 )·φj (z2 , z̄2 ) =
k
(3)
which is supposed to be valid when inserted into higher-order correlation
functions in the limit when |z1 − z2 | is much less than the separations of all
Boundary conformal field theory 3

the other points. The ellipses denote the contributions of other non-primary
scaling fields to be described below. The structure constants cijk , along with
the conformal weights, characterize the particular CFT.
An essential role is played by the energy-momentum tensor, or, in euclidean
field theory language, the stress tensor T µν . Heuristically, it is defined as the
response of the partition function to a local change in the metric:
T µν (x) = −(2π) δ ln Z/δgµν (x) (4)
(the factor of 2π is included so that similar factors disappear in later equa-
tions).
The symmetry of the theory under translations and rotations implies that
T µν is conserved, ∂µ T µν = 0, and symmetric. Scale invariance implies that
it is also traceless Θ ≡ Tµµ = 0. It should be noted that the vanishing of
the trace of the stress tensor for a scale invariant classical field theory does
not usually survive when quantum corrections are taken into account: indeed
Θ ∝ β(g), the renormalization group (RG) beta-function. A quantum field
theory is thus only a CFT when this vanishes, that is at an RG fixed point.
In complex coordinates the components Tz z̄ = Tz̄z = 4Θ vanish, while the
conservation equations read
∂z̄ Tzz = ∂z Tz̄ z̄ = 0 . (5)
Thus correlators of T (z) ≡ Tzz are locally analytic (in fact, globally mero-
morphic) functions of z, while those of T (z̄) ≡ Tz̄ z̄ are anti-analytic. It is
this property of analyticity which makes CFT tractable in two dimensions.
Since an infinitesimal conformal transformation z → z + α(z) induces a
change in the metric, its effect on a correlation function of primary fields,
given by (1), may also be expressed through an appropriate integral involving
an insertion of the stress tensor. This leads to the conformal Ward identity:
Z
(hj α′ (zj ) + α(zj )(∂/∂zj )) h
Y X Y
hT (z) φj (zj , z̄j )i α(z)dz = φj (zj , z̄j )i ,
C j j j
(6)
where C is a contour encircling all the points {zj }. (A similar equation hold
for the insertion of T .) Using Cauchy’s theorem, this determines the first few
terms in the OPE of T with any primary density:
hj 1
T (z) · φj (zj , z̄j ) = φ(zj , z̄j ) + ∂z φ(zj , z̄j ) + O(1) . (7)
(z − zj ) 2 z − zj j
Boundary conformal field theory 4

The other, regular, terms in the OPE generate new scaling fields, which are
not in general primary, called descendants. One way of defining a density to
be primary is by the condition that the most singular term in its OPE with
T is a double pole.
The OPE of T with itself has the form
c/2 2
T (z) · T (z1 ) = 4
+ T (z1 ) + · · · . (8)
(z − z1 ) (z − z1 )2
The first term is present because hT (z)T (z1 )i is non-vanishing, and must
take the form shown, with c being some number (which cannot be scaled
to unity, since the normalization of T is fixed by its definition) which is a
property of the CFT. It is known as the conformal anomaly number or the
central charge. This term implies that T is not itself primary. In fact under
a finite conformal transformation z → z ′ = f (z)
T (z) → f ′ (z)2 T (z ′ ) + c
12
{z ′ , z} , (9)
where {z ′ , z} = (f ′′′ f ′ − 32 f ′′ 2 )/f ′2 is the Schwartzian derivative.

1.2 Virasoro algebra

As with any quantum field theory, the local fields can be realized as linear
operators acting on a Hilbert space. In ordinary QFT, it is customary to
quantize on a constant time hypersurface. The generator of infinitesimal
time translations is the hamiltonian Ĥ, which itself is independent of which
time slice is chosen, because of time translational symmetry. It is also given
by the integral over the hypersurface of the time-time component of the stress
tensor. In CFT, because of scale invariance, one may instead quantize on
fixed circle of a given radius. The analog of the hamiltonian is the dilatation
operator D̂, which generates scale transformations. Unlike Ĥ, the spectrum
of D̂ is usually discrete, even in an infinite system. It may also be expressed
as an integral over the radial component of the stress tensor
1 Z 2π 1 Z 1 Z
D̂ = r T̂rr rdθ = z T̂ (z)dz − ˆ , (10)
z̄ Tˆ (z̄)dz̄ ≡ L̂0 + L0
2π 0 2πi C 2πi C
where, because of analyticity, C can be any contour encircling the origin.
This suggests that one define other operators
1
Z
L̂n ≡ z n+1 T̂ (z)dz , (11)
2π C
Boundary conformal field theory 5

and similarly the L ˆ . From the OPE (8) then follows the Virasoro algebra V
n
c
[L̂n , L̂m ] = (n − m)L̂n+m + n(n2 − 1)δn+m,0 , (12)
12
with an isomorphic algebra V generated by the L ˆ .
n

In radial quantization there is a vacuum state |0i. Acting on this with the
operator corresponding to a scaling field gives a state |φj i ≡ φ̂j (0, 0)|0i which
is an eigenstate of D̂: in fact

L̂0 |φj i = hj |φj i , ˆ |φ i = h̄ |φ i .

L (13)
0 j j j

From the OPE (7) one sees that |Ln φj i ∝ L̂n |φj i, and, if φj is primary,
L̂n |φj i = 0 for all n ≥ 1.
The states corresponding to a given primary field, and those generated by
acting on these with all the L̂n with n < 0 an arbitrary number of times,
form a highest weight representation of V. However, this is not necessarily
irreducible. There may be null vectors, which are linear combinations of
states at a given level which are themselves annihilated by all the L̂n with
n > 0. They exist whenever h takes a value from the Kac table:
(r(m + 1) − sm)2 − 1
h = hr,s = , (14)
4m(m + 1)
with the central charge parametrized as c = 1 − 6/(m(m + 1)), and r, s are
non-negative integers. These null states should be projected out, giving an
irreducible representation Vh .
The full Hilbert space of the CFT is then
M
H= nh,h̄ Vh ⊗ V h̄ , (15)
h,h̄

where the non-negative integers nh,h̄ specify how many distinct primary fields
of weights (h, h̄) there are in the CFT.
The consistency of the OPE (3) with the existence of null vectors leads to
the fusion algebra of the CFT. This applies separately to the holomorphic
and antiholomorphic sectors, and determines how many copies of Vc occur in
the fusion of Va and Vb :
c
X
Va ⊙ Vb = Nab Vc , (16)
c
Boundary conformal field theory 6

c
where the Nab are non-negative integers.
A particularly important subset of all CFTs consists of the minimal models.
These have rational central charge c = 1 − 6(p − q)2 /pq, in which case the
fusion algebra closes with a finite number of possible values 1 ≤ r ≤ q,
1 ≤ s ≤ p in the Kac formula (14). For these models, the fusion algebra
takes the form ′ ′
r1 +r
X 2 −1 s1 +s
X 2 −1

Vr1 ,s1 ⊙ Vr2 ,s2 = Vr,s , (17)

r=|r1 −r2 | s=|s1 −s2 |

where the prime on the sums indicates that they are to be restricted to the
allowed intervals of r and s.
There is an important theorem which states that the only unitary CFTs with
c < 1 are the minimal models with p/q = (m + 1)/m, where m is an integer
≥ 3.

1.3 Modular Invariance

The fusion algebra limits which values of (h, h̄) might appear in a consistent
CFT, but not which ones actually occur, i.e. the values of the nh,h̄ . This
is answered by the requirement of modular invariance on the torus. First
consider the theory on an infinitely long cylinder, of unit circumference.
This is related to the (punctured) plane by the conformal mapping z →
(1/2π) ln z ≡ t + ix. The result is a QFT on the circle 0 ≤ x < 1, in
imaginary time t. The generator of infinitesimal time translations is related
to that for dilatations in the plane:
πc ˆ ) − πc ,
Ĥ = 2π D̂ − = 2π(L̂0 + L0 (18)
6 6
where the last term comes from the Schwartzian derivative in (9). Similarly,
the generator of translations in x, the total momentum operator, is P̂ =
ˆ ).
2π(L̂0 − L0

A general torus is, up to a scale transformation, a parallelogram with vertices

(0, 1, τ, 1 + τ ) in the complex plane, with the opposite edges identified. We
can make this by taking a cylinder of unit circumference and length Im, τ ,
twisting the ends by a relative amount Re τ , and sewing them together. This
Boundary conformal field theory 7

−1/ τ

0 1 0 1
Figure 1: Two equivalent parametrizations of the same torus.

means that the partition function of the CFT on the torus can be written as
ˆ
Z(τ, τ̄ ) = Tr e−(Im τ )Ĥ+i(Im τ )P̂ = Tr q L̂0 −c/24 q̄ L0 −c/24 , (19)
using the above expressions for Ĥ and P̂ and introducing q ≡ e2πiτ .
Through the decomposition (15) of H, the trace sum can be written as
X
Z(τ, τ̄ ) = nh,h̄ χh (q) χh̄ (q) , (20)
h,h̄

where
χh (q) ≡ TrVh q L̂0 −c/24 = dh (N) q h−(c/24)+N
X
(21)
N
is the character of the representation of highest weight h, which counts the
degeneracy dh (N) at level N. It is purely an algebraic property of the Vira-
soro algebra, and its explicit form is known in many cases.
All of this would be less interesting were it not for the observation that the
parametrization of the torus through τ is not unique. In fact the transfor-
mations S : τ → −1/τ and T : τ → τ + 1 give the same torus (see Fig. 1).
Together, these operations generate the modular group SL(2, Z), and the
partition function Z(τ, τ̄ ) should be invariant under them. T -invariance is
simply implemented by requiring that h−h̄ is an integer, but the S-invariance
of the right hand side of (20) places highly nontrivial constraints on the nh,h̄ .
That this can be satisfied at all relies on the remarkable property of the
characters that they transform linearly under S:
′
χh (e−2πi/τ ) = Shh χh′ (e2πiτ ) .
X
(22)
h′

This follows from applying the Poisson sum formula to the explicit expres-
sions for the characters, which are related to Jacobi theta-functions. In
Boundary conformal field theory 8

many cases (for example, the minimal models) this representation is finite-
dimensional, and the matrix S is symmetric and orthogonal. This means
that one can immediately obtain a modular invariant partition function by
forming the diagonal sum
X
Z= χh (q)χh (q̄) , (23)
h

so that nh,h̄ = δhh̄ . However, because of various symmetries of the characters,

other modular invariants are possible: for the minimal models (and some
others) these have been classified. Because of an analogy of the results with
the classification of semisimple Lie algebras, the diagonal invariants are called
the A-series.

2 Boundary CFT
In any field theory in a domain with a boundary, one needs to consider how
to impose a set of consistent boundary conditions. Since CFT is formulated
independently of a particular set of fundamental fields and a lagrangian, this
must be done in a more general manner. A natural requirement is that the
off-diagonal component Tk⊥ of the stress tensor parallel/perpendicular to the
boundary should vanish. This is called the conformal boundary condition. If
the boundary is parallel to the time axis, it implies that there is no momen-
tum flow across the boundary. Moreover, it can be argued that, under the
RG, any uniform boundary condition will flow into a conformally invariant
one. For a given bulk CFT, however, there may be many possible distinct
such boundary conditions, and it is one task of BCFT to classify these.
To begin with, take the domain to be the upper half plane, so that the
boundary is the real axis. The conformal boundary condition then implies
that T (z) = T (z̄) when z is on the real axis. This has the immediate conse-
quence that correlators of T are those of T , analytically continued into the
lower half plane. The conformal Ward identity, c.f. (7), now reads
Y X hj 1
hT (z) φj (zj , z̄j )i = + ∂z
j j (z − zj )2 z − zj j
!
h̄j 1 Y
+ + ∂z̄ j
h φj (zj , z̄j )i .(24)
(z̄ − z̄j )2 z̄ − z̄j j
Boundary conformal field theory 9

In radial quantization, in order that the Hilbert spaces defined on different

hypersurfaces be equivalent, one must choose semicircles centered on some
point on the boundary, conventionally the origin. The dilatation operator is
now
1 1
Z Z
D̂ = z T̂ (z)dz − z̄ Tˆ (z̄)dz̄ , (25)
2πi S 2πi S
where S is a semicircle. Using the conformal boundary condition, this can
also be written as
1
Z
D̂ = L̂0 = z T̂ (z)dz , (26)
2πi C
where C is a complete circle around the origin. As before, one may similarly
define the L̂n , and they satisfy a Virasoro algebra.
Note that there is now only one Virasoro algebra. This is related to the
fact that conformal mappings which preserve the real axis correspond to real
analytic functions. The eigenstates of L̂0 correspond to boundary operators
φ̂j (0) acting on the vacuum state |0i. It is well-known that in a renormaliz-
able QFT operators at the boundary require a different renormalization from
those in the bulk, and this will in general lead to a different set of confor-
mal weights. It is one of the tasks of BCFT to determine these, for a given
allowed boundary condition.
However, there is one feature unique to boundary CFT in two dimensions.
Radial quantization also makes sense, leading to the same form (26) for the
dilation operator, if the boundary conditions on the negative and positive real
axes are different. As far as the structure of BCFT goes, correlation functions
with this mixed boundary condition behave as though a local scaling field
were inserted at the origin. This has led to the term ‘boundary condition
changing (bcc) operator’, but it must be stressed that these are not local
operators in the conventional sense.

3 The annulus partition function

Just as consideration of the partition function on the torus illuminates the
bulk operator content nh,h̄ , it turns out that consistency on the annulus helps
classify both the allowed boundary conditions, and the boundary operator
content. To this end, consider a CFT in an annulus formed of a rectangle
of unit width and height δ, with the top and bottom edges identified (see
Boundary conformal field theory 10

a b δ

1
Figure 2: The annulus, with boundary conditions a and b on either boundary.

Fig. 2). The boundary conditions on the left and right edges, labelled by
a, b, . . ., may be different. The partition function with boundary conditions
a and b on either edge is denoted by Zab (δ).
One way to compute this is by first considering the CFT on an infinitely long
strip of unit width. This is conformally related to the upper half plane (with
an insertion of boundary condition changing operators at 0 and ∞ if a 6= b)
by the mapping z → (1/π) ln z. The generator of infinitesimal translations
along the strip is

Ĥab = π D̂ − πc/24 = π L̂0 − πc/24 . (27)

Thus for the annulus

Zab (δ) = Tr e−δ Ĥab = Tr q L̂0 −πc/24 , (28)

with q ≡ e−πδ . As before, this can be decomposed into characters

nab
X
Zab (δ) = h χh (q) , (29)
h

but note that now the expression is linear. The non-negative integers nhab give
the operator content with the boundary conditions (ab): the lowest value of
h with nhab > 0 gives the conformal weight of the bcc operator, and the others
Boundary conformal field theory 11

give conformal weights of the other allowed primary fields which may also sit
at this point.
On the other hand, the annulus partition function may be viewed, up to an
overall rescaling, as the path integral for a CFT on a circle of unit circum-
ference, being propagated for (imaginary) time δ −1 . From this point of view,
the partition function is no longer a trace, but rather the matrix element of
e−Ĥ/δ between boundary states:

Zab (δ) = ha|e−Ĥ/δ |bi . (30)

Note that Ĥ is the same hamiltonian that appears in (18), and the boundary
states lie in H, (15).
How are these boundary states to be characterized? Using the transformation
law (9) the conformal boundary condition applied to the circle implies that
Ln = L−n . This means that any boundary state |Bi lies in the subspace
satisfying
ˆ |Bi .
L̂n |Bi = L (31)
−n

Moreover, because of the decomposition (15) of H, |Bi is also some linear

superposition of states from Vh ⊗ V h̄ . This condition can therefore be applied
in each subspace. Taking n = 0 in (31) constrains h̄ = h. For simplicity,
consider only the diagonal CFTs with nh,h̄ = δh,h̄ . It can then be shown that
the solution of (31) is unique and has the following form. The subspace at
level N of Vh has dimension dh (N). Denote an orthonormal basis by |h, N; ji,
with 1 ≤ j ≤ dh (N), and the same basis for V h by |h, N; ji. The solution to
(31) in this subspace is then
∞ dh
X X (N )
|hii ≡ |h, N; ji ⊗ |h, N; ji . (32)
N =0 j=1

These are called Ishibashi states. Matrix elements of the translation operator
along the cylinder between them are simple:

hhh′ |e−Ĥ/δ |hii (33)

∞ dh′ (N ′ ) ∞ dh (N )
ˆ
hh′ , N ′ ; j ′ | ⊗ hh′ , N ′ ; j ′ |e−(2π/δ)(L̂0 +L0 −c/12)
X X X X
=
N ′ =0 j ′ =1 N =0 j=1

|h, N; ji ⊗ |h, N; ji (34)

Boundary conformal field theory 12

∞ dh (N )
e−(4π/δ)( h+N −(c/24)) = δh′ h χh (e−4π/δ ) .
X X
= δh′ h (35)
N =0 j=1

Note that the characters which appear are related to those in (29) by the
modular transformation S.
The physical boundary states satisfying (29), sometimes called the Cardy
states, are linear combinations of the Ishibashi states:
X
|ai = hhh|ai |hii . (36)
h

Equating the two different expressions (29,30) for Zab , and using the modular
transformation law (22) and the linear independence of the characters gives
the (equivalent) conditions:
nhab = Shh′ ha|h′ iihhh′ |bi ;
X
(37)
h′
′
ha|h′ iihhh′ |bi = Shh nhab .
X
(38)
h

These are called the Cardy conditions. The requirements that the right hand
side of (37) should give a non-negative integer, and that the right hand side
of (38) should factorize in a and b, give highly nontrivial constraints on the
allowed boundary states and their operator content.
For the diagonal CFTs considered here (and for the nondiagonal minimal
models) a complete solution is possible. It can be shown that the elements
S0h of S are all non-negative, so one may choose hhh|0̃i = (S0h )1/2 . This defines
a boundary state
(S0h )1/2 |hii ,
X
|0̃i ≡ (39)
h

and a corresponding boundary condition such that nh00 = δh0 . Then, for each
h′ 6= 0, one may define a boundary state
hhh|h̃′ i ≡ Shh′ /(S0h )1/2 . (40)
From (37), this gives nhh′ 0 = δh′ h . For each allowed h′ in the torus parti-
tion function, there is therefore a boundary state |h̃′ i satisfying the Cardy
conditions. However, there is a further requirement:
Sℓh Shℓ ′ Shℓ ′′
nhh′ h′′ =
X
(41)
ℓ
S0ℓ
Boundary conformal field theory 13

should be a non-negative integer. Remarkably, this combination of elements

of S occurs in the Verlinde formula, which follows from considering consis-
tency of the CFT on the torus. This states that the right hand side of (41)
is equal to the fusion algebra coefficient Nhh′ h′′ . Since these are non-negative
integers, the consistency of the above ansatz for the boundary states is con-
sistent.
We conclude that, at least for the diagonal models, there is a bijection be-
tween the allowed primary fields in the bulk CFT and the allowed conformally
invariant boundary conditions. For the minimal models, with a finite num-
ber of such primary fields, this correspondence has been followed through
explicitly.

3.0.1 Example

The simplest example is the diagonal c = 21 unitary CFT corresponding to

m = 3. The allowed values of the conformal weights are h = 0, 21 16
1
, and
1 1 √1
 
2 2 2
1 1
− √12
 
S=
 2 2

 , (42)
√1 − √12 0
2

from which one finds the allowed boundary states

|0̃i = √1 |0ii + √1 | 1 ii + 1
| 1 ii ; (43)
2 2 2 21/4 16

| 21̃ i = √1 |0ii
2
+ √1 | 1 ii
2 2
− 1
| 1 ii ;
21/4 16
(44)
1̃
| 16 i = |0ii − | 12 ii . (45)
The nontrivial part of the fusion algebra of this CFT is
V1 ⊙V1 = V0 + V 1 (46)
16 16 2

V 1 ⊙ V1 = V 1 (47)
16 2 16

V 1 ⊙ V 1 = V0 , (48)
2 2

(49)
from which can be read off the boundary operator content nhh̃ = 1 and n01̃ 1̃
=
16 16
1 1 1
n 21̃ 1̃
= n 21̃ 1̃
= n 16
1̃ 1̃
= 1.
16 16 16 16 2 16
Boundary conformal field theory 14

The c = 21 CFT is known to describe the continuum limit of the critical Ising
model, in which spins s = ±1 are localized on the sites of a regular lattice.
The above boundary conditions may be interpreted as the continuum limit
of the lattice boundary conditions s = 1, free and s = −1 respectively. Note
there is a symmetry of the fusion rules which means that one could equally
well have inverted the ordering of this correspondence.

4 Other topics

4.1 Boundary entropy

The partition function on annulus of length L and circumference β can be
thought of as the quantum statistical mechanics partition function for a 1d
QFT in an interval of length L, at temperature β −1 . It is interesting to
consider this in the thermodynamic limit when δ = L/β is large. In that
case, only the ground state of Ĥ contributes in (30), giving

Zab (L, β) ∼ ha|0ih0|bi eπcL/6β , (50)

from which the free energy Fab = −β −1 ln Zab and the entropy Sab =
−β 2 (∂Fab /∂β) can be obtained. The result is

Sab = (πc/3β)L + sa + sb + o(1) , (51)

where the first term is the usual extensive contribution. The other two pieces
sa ≡ ln(ha|0i) and sb ≡ ln(hb|0i) may be identified as the boundary entropy
associated with the corresponding boundary states. A similar definition may
be made in massive QFTs. It has been shown that, analogously to the
statement of Zamoldochikov’s c-theorem in the bulk, the boundary entropy
is a non-increasing function along boundary RG flows, and is stationary only
for conformal boundary states.

4.2 Bulk-boundary OPE

The boundary Ward identity (24) has the implication that, from the point
of view of the dependence of its correlators on zj and z̄j , a primary field
Boundary conformal field theory 15

φj (zj , z̄j ) may be thought of as the product of two local fields which are
holomorphic functions of zj and z̄j respectively. These will satisfy OPEs as
|zj − z̄j | → 0, with the appearance of primary fields on the right hand side
being governed by the fusion rules. These fields are localized on the real axis:
they are the boundary operators. There is therefore a kind of bulk-boundary
OPE:
djk (Im zj )−hj −h̄j +hk φbk (Re zj ) ,
X
φj (zj , z̄j ) = (52)
k

where the sum on the right hand side is in principle over all the boundary
fields consistent with the boundary condition, and the coefficients djk are
analogous to the OPE coefficients in the bulk. As before, they are non-
vanishing only if allowed by the fusion algebra: a boundary field of conformal
weight hk is allowed only if Nhhjkh̄j > 0.
For example, in the c = 12 CFT, the bulk operator with h = h̄ = 16
1
goes over
1
into the boundary operator with h = 0, or that with h = 2 , depending on
the boundary condition. The bulk operator with h = h̄ = 12 , however, can
only go over into the identity boundary operator with h = 0 (or a descendent
thereof.)
The fusion rules also apply to the boundary operators themselves. The con-
sistency of these with bulk-boundary and bulk-bulk fusion rules, as well as
the modular properties of partition functions, was examined by Lewellen.

4.3 Extended algebras

CFTs may contain other conserved currents apart from the stress tensor,
which generate algebras (Kac-Moody, superconformal, W-algebras) which
extend the Virasoro algebra. In BCFT, in addition to the conformal bound-
ary condition, it is possible (but not necessary) to impose further boundary
conditions relating the holomorphic and antiholomorphic parts of the other
currents on the boundary. It is believed that all rational CFTs can be ob-
tained from Kac-Moody algebras via the coset construction. The classifi-
cation of boundary conditions from this point of view is fruitful and also
important for applications, but is beyond the scope of this article.
Boundary conformal field theory 16

4.4 Stochastic Loewner evolution

In recent years, there has emerged a deep connection between BCFT and con-
formally invariant measures on curves in the plane which start at a boundary
of a domain. These arise naturally in the continuum limit of certain statisti-
cal mechanics models. The measure is constructed dynamically as the curve
is extended, using a sequence of random conformal mappings called stochas-
tic Loewner evolution (SLE). In CFT, the point where the curve begins can
be viewed as the insertion of a boundary operator. The requirement that
certain quantities should be conserved in mean under the stochastic process
is then equivalent to this operator having a null state at level two. Many of
the standard results of CFT correspond to an equivalent property of SLE.

Acknowledgments
This article was written while the author was a member of the Institute
for Advanced Study. He thanks the School of Mathematics and the School
of Natural Sciences for their hospitality. The work was supported by the
Ellentuck Fund.

References
[1] P. di Francesco, P. Mathieu and D. Senechal, Conformal Field Theory
(Springer, 1999.)

[2] J. Cardy, Conformal invariance and surface critical behavior, Nucl. Phys.
B 240, 514-532, 1984.

[3] J. Cardy, Boundary conditions, fusion rules and the Verlinde formula,
Nucl. Phys. B 324, 581, 1989.

[4] E. Verlinde, Fusion rules and modular transformations in 2D conformal

field theory, Nucl. Phys. B 300, 360, 1988.

[5] V.B. Petkova and J.-B. Zuber, Conformal boundary conditions and what
they teach us, lectures given at the Summer School and Conference on
Boundary conformal field theory 17

Nonperturbative Quantum Field Theoretic Methods and their Applica-

tions, August 2000, Budapest, Hungary, hep-th/0103007.

[6] I. Affleck, Boundary condition changing operators in conformal field

theory and condensed matter physics, Nucl.Phys.Proc.Suppl. 58, 35,
1997.

[7] D. Friedan, A. Konechny, On the boundary entropy of one-dimensional

quantum systems at low temperature, Phys. Rev. Lett. 93, 030402, 2004.

[8] D. C. Lewellen, Sewing constraints for conformal field theories on sur-

faces with boundaries, Nucl. Phys. B 372, 654, 1992.

[9] For reviews of SLE, see W. Werner, Random planar curves and
Schramm-Loewner evolutions, to appear (Springer Lecture Notes),
math.PR/0303354; G. Lawler, Conformally invariant processes in the
plane, in preparation, http://www.math.cornell.edu/l̃awler/book.ps.
For a review aimed at theoretical physicists, see W. Kager and B. Nien-
huis, A guide to stochastic Loewner evolution and its applications, J.
Stat. Phys. 115, 1149 (2004).