Vous êtes sur la page 1sur 14

Dahmane Hammaoui

The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
225
Received 31 May 2008; Accepted 11 November 2008
THE SMALLEST OCNEANU QUANTUM
GROUPO

ID OF SU(3) TYPE
ABSTRACT
Ocneanu quantum groupods are a type of weak Hopf algebras which can be dened for every Lie group, and
are at the core of the study of the quantum symmetries related to higher Coxeter graphs. In the present paper
I analyze the whole theory for this particular type of weak Hopf algebras or Ocneanu double triangle algebra
in a very simple case: the /
1
Ocneanu quantum groupod of SU(3) type. This small example is interesting to
study because it allows to explicitly describe the various structures of the theory.
MSC (2000): 81R50
Key words: Coxeter-Dynkin graphs, fusion and graph algebras, module graphs, modular invariants, quantum
symmetries, Ocneanu double triangle algebras

E-mail: d.hammaoui@sciences.univ-oujda.ac.ma or hammaouidahmane@yahoo.fr


:
Ocneanu Coxeter .
Hopf Lie . Hopf
Ocneanu : Ocneanu A
1
SU(3) .
.
Dahmane Hammaoui

Equipe de Physique Mathematique et Plasmas


Laboratoire de Physique Theorique
Physique des Particules et Modelisation (LPTPM)
Departement de Physique, Faculte des Sciences, Oujda, Morocco
and
Centre de Classes Preparatoires aux Grandes

Ecoles dIngenieurs -CPR- Tanger, Morocco
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
226
1. INTRODUCTION
Higher Coxeter graphs of SU(N) type were introduced by A. Ocneanu as a generalization to higher order of
the usual ADE Coxeter-Dynkin diagrams [1]. Their classication is an interesting problem both in physics and
in mathematics and is complete only for SU(N) when N = 2, 3, and 4. In mathematics these graphs appear in
various elds such as representation theory of quantum groups, (weak) Hopf algebras, classication of semi-simple
Lie algebras, operator algebras, nets of subfactors, category theory, and bimodules [2, 3, 4, 5, 6, 7, 8, 1]... In
physics, many models are based on this kind of graph, such as lattice integrable models in statistical mechanics
or models of quantum gravity in string theory and D-branes [9, 10, 11, 12, 13, 14]. From a conformal eld
theoretical viewpoint higher Coxeter graphs of the SU(N) type are related to the classication of modular
invariants of conformal models [15, 16, 17, 18] and enable us to construct twisted partition functions [19, 20, 21].
In the topological eld theoretical framework, these graphs provide the necessary data to compute the set of 3j
and 6j-symbols in many theories of strings [22].
Quantum geometry of ADE and generalized Coxeter-Dynkin graphs (introduced by A. Ocneanu [8]) describes
quantum symmetries of these graphs together with their corresponding Ocneanu graphs, which index the defect
lines of the CFT and brings out new structures of weak Hopf algebras called algebras of double triangles. An
algebraic method was adopted by R. Coquereaux [23] to build the Ocneanu algebra of quantum symmetries of a
given graph G of SU(N) type. Following this procedure, Ocneanu graphs and algebras of quantum symmetries
were either recovered (when they were already known) or established for all members of SU(2) type [19, 24, 25],
of SU(3) type [26, 27, 28] and exceptional cases of SU(4) system [29, 30]. In this way to every pair of graphs of
an SU(N) system with same generalized Coxeter number (namely a graph G and the corresponding graph /(G)
which describes its fusion algebra), one associates an algebra of quantum symmetries. Its elements are called
irreducible quantum symmetries that one can add and multiply like the irreducible representations of groups
and the Ocneanu graph encodes this algebraic structure. One can also start from a given modular invariant
and applying the modular splitting formula [8, 25, 28] to recover this algebra of quantum symmetries. The
fusion algebra and the Ocneanu algebra of quantum symmetries describe the theory of representations of the so-
called double triangle algebra (DTA). This later is a bialgebra (more precisely a weak Hopf algebra or Ocneanu
quantum groupod) which is related to the geometry of paths on the graph G. The paths are of two type: the
horizontal paths indexed by the irreducible representation (irreps) of the fusion algebra and the vertical ones
indexed by the irreducible quantum symmetries. This property of being a weak Hopf algebra characterizing
this Ocneanu bialgebra was given in [31] in a framework of quantum cohomology and was investigated in [32]
by considering solutions to the big pentagon equation. To ensure some consistency conditions for the products
and coproducts and the pairing between dual spaces one needs to introduce fundamental objects called Ocneanu
cells (a kind of generalized 6j-symbols). Because of the hard and technical task to compute these Ocneanu cells,
the full construction and study of the Ocneanu bialgebras are established only for the A
k
graphs belonging to
the ADE members of SU(2) system [33, 34, 35]. However, there is much information on quantum groupods
related to the Di Francesco-Zuber system of graphs associated with the SU(3) system in [26, 27]. Note that a
nice relation between these construction and face models in statistical mechanics can be found in [36].
The purpose of this paper is to give a brief description of the Ocneanu bialgebras where we focus more on the
algebra and coalgebra structures which are the special topics of this number of AJSE. The concrete results are
relative to the simplest (but not trivial) case of SU(3) system: the /
1
graph. Note that we need no knowledge
or use of Ocneanu cells since their values are all equal to 1 in this particular case. The article is organized as
follows: the rst part deals with the necessary background on graphs and their quantum symmetries where are
dened the fusion algebra /(G) of a given graph G, the associated Ocneanu algebra of quantum symmetries
Oc (G), and the properties of G as an /-module and as an Oc (G)-module and the connection with the /(G)
and Oc (G) algebras on one side and the DTA structure on the other side. The second part is devoted to the
Ocneanu bialgebras (there are two dual bialgebras denoted B and

B which are isomorphic) where are dened
horizontal and vertical double triangles and the associated algebras and coalgebras structures. As an explicit
application we give a full construction of the DTA related to the /
1
case.
THE SMALLEST OCNEANU QUANTUM GROUPO

ID OF SU(3) TYPE
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
227
2. BACKGROUND ON GRAPHS AND RELATED ALGEBRAIC STRUCTURES
Given a modular invariant / i.e. a matrix associated to a modular invariant partition function Z in RCFT
(see [15, 16] for the SU(2) series or [17, 18] for the SU(3) series but for higher SU(N) series there are no complete
classication), one can associate at least one higher Coxeter graph G of SU(N) type [1]. The graph G is a
module graph on the fusion algebra /(G) related to the graph /
k
with same generalized Coxeter number as
G. The /
k
graph has d
A
vertices denoted , , ..., while for the G graph there are d
G
vertices that we denotes
by a, b, c... The structure of /-module is encoded in non-negative integers F
b
a
which constitute a non-negative
integer valued-matrix representation (nim-rep) of dimension d
G
d
G
for the fusion algebra [23, 19, 18]. The
integer F
b
a
is just the number of horizontal essential paths on G, associated to the irreducible representation
of /
k
, starting from the vertex a and arriving at the vertex b. The graph G is also a module graph on its algebra
Oc(G) of quantum symmetries related to the so-called Ocneanu graph that we denote by (G). Vertices of the
(G) graph are denoted x, y, z.... The structure of Oc (G)-module is encoded in coecients S
b
xa
which constitute
a nim-rep of dimension d
(G)
d
(G)
for the Ocneanu algebra of quantum symmetries [23, 19, 24, 26, 27, 30].
The integer S
b
xa
is the number of vertical paths on G, associated to the irreducible quantum symmetry x of
Oc(G), starting from the vertex a and arriving at b.
2.1. The Fusion Algebra and the Graph of / Type
/
k
graphs are the Weyl alcoves of SU(N) type truncated at some level k and are characterized by the
(dual) Coxeter number = N +k. The vertices , , . . . denote irreducible representations (irreps) of quantum
sub-groups SU(N)
k
, at a root of unity q = e
i/
. The A
k
graphs encode the tensor product of irreps inherited
from fusion of primary elds in CFT [37]:
1
i
1
j
=
k
N
k
ij
1
k
,
where 1

are irreducible representations of the chiral algebra, an extension of the Virasoro algebra, of the RCFT.
For instance, the Weyl alcoves of SU(3) case are:
/
k
= = (
1
,
2
) =
1

1
+
2

2
[ 0
1
+
2
k,
1
,
2
N .
where
1
and
2
are the fundamental weights of the SU(3) Lie group and
1
,
2
are the corresponding Dynkin
labels. (0, 0) is the unit representation which index the unit vertex of /
k
and is related to the vacuum state,
(1, 0) is the fundamental generator (irrep) of SU(3)
k
and (0, 1) is its conjugate. The cardinality of /
k
is
d
A
k
= (k + 1)(k + 2)/2.
The nim-reps N
k
ij
are given by the Verlinde formula [38]:
N
k
ij
=

mA
k
SimSjmS

km
S0m
.
where S

are the entries of the symmetric, unitary Smatrix of the modular transformations of the charac-
ters
i
() = Tr e
2i(L0c/24)
such that

(1/) =

A
k
S

(). The other generator T of the modular


group is dened by its action on the characters by

( + 1) =

A
k
T

(). A d
A
k
d
A
k
-dimensional
matrix representation of the fusion algebra (also called the Verlinde algebra) is given by the following formula:
N

=

A
k
N

.
Each matrix N

is associated to a vertex = (
1
,
2
) of the graph /
k
. To determine the N

, it is convenient
to use the known recurrence relation for coupling of irreducible SU(3) representations: N
(0,0)
= I
d
A
k
, N
(1,0)
=
A(/
k
) is the adjacency matrix of the graph /
k
(the adjacency matrix A is reading directly from the graph,
conversely the knowledge of the A enable one to represent the graph itself), and
N
(,)
= N
(1,0)
N
(1,)
N
(1,1)
N
(2,+1)
if ,= 0
N
(,0)
= N
(1,0)
N
(1,0)
N
(2,1)
N
(0,)
= (N
(,0)
)
tr
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
228
2.2. Quantum Symmetries and Ocneanu Graphs
As we shall see in the second part, the Ocneanu bialgebra under study is a semi-simple graded algebra of
endomorphisms of essential paths. It has two diagonal block decompositions in simple blocks. The rst type
of blocks are labeled by Young diagrams (irreps) of the graph /
k
and the fusion algebra /
k
is interpreted
as the algebra spanned by minimal central projections related to these blocks. While the second type of blocks
are indexed by the irreducible quantum symmetries x of the Ocneanu graph (G) and the related algebra of
quantum symmetries Oc (G) is the algebra spanned by minimal central projections associated to these second
blocks. Another interpretation of quantum symmetries in the framework of twisted chiral conformal eld theory
is as follows: The vertices x, y, z . . . of the Ocneanu graph encode defect lines of the theory, and their fusion is
described by the Ocneanu algebra of quantum symmetries.
A technical method to determine the algebra of quantum symmetries of a graph is the modular splitting
formula based only on the knowledge of /(G) and the fusion matrices N

of /(G):

=

z
(W
0z
)

(W
z0
)

.
which enables us to compute all toric matrices
1
(of size d
A
d
A
) W
0z
or W
z0
with one defect line z. its
generalization gives a relation between twisted toric matrices with two defect lines.

z
(W
xz
)

(W
zy
)

(W
xy
)

.
The others W
xy
, which give the generalized twisted partition functions of the theory with two defect lines (say
x and y), are given by

z
(W
xz
)

W
z0
= N

W
x0
(N

)
tr
.
A faithful anti-representation of the Ocneanu algebra Oc(G) is carried by N-valued matrices called Ocneanu
matrices O
x
, of size d

, indexed by the vertices x of (G):


O
x
O
y
=

z
O
z
yx
O
z
.
To determine the adjacency matrices of the Ocneanu graph (G) related to the fundamental generators, it
is convenient to introduce the following matrices V

of size d

such that their entries satisfy to (V

)
xy
=
(W
xy
)

. These are the so-called double annular matrices and constitute a nim-rep for the double fusion algebra
/
k
/
k
V

What is special in these notions of double fusion algebra is that the fundamental generators of the Ocneanu
graph are related to double annular matrices: V

f
0
and V
0
f
where
f
are the fundamental irreps of the SU(N)
Lie group. For the SU (3) case, where
f
is either (1, 0) or its conjugate (0, 1), the Ocneanu graph (G) has
four chiral fundamental generators
2
associated to the following adjacency matrices:
V
(1,0)(0,0)
= O
1L
V
(0,0)(1,0)
= O
1R
V
(0,1)(0,0)
= O
1

L
V
(0,0)(0,1)
= O
1

R
and the unit vertex of (G) is associated to the identity matrix V
(0,0)(0,0)
= I
d
A
.
2.2.1. Module Structures Related to the Graph G
On the one hand, the graph G is a module over the fusion algebras /(G)
/G G
.a

b
F
b
a
b
1
The toric matrix associated to the unit vertex is the modular invariant itself: W
00
= M.
2
When the graph G is a graph of A type, one has only one fundamental generator and its conjugate. Furthermore the Ocneanu
graph (A) is the graph A itself and all types of the cited sets of matrices coincide with the set of fusion matrices N

.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
229
This action of /-module on vertices of G is encoded in the nim-reps (F

)
ab
which provide solutions to the Cardy
equation in boundary CFT [39], and where the boundary states of the theory are represented by the vertices
a, b, c . . . of G. The matrices F

are called the annular matrices such that F


(0,0)
= I
d
G
, F
(1,0)
= A(G) i.e. the
adjacency matrix of the graph G satises the same recurrence formula as the fusion matrices N

.
On the other hand, the graph G is also a module over the Ocneanu algebra of quantum symmetries Oc(G).
This structure is encoded in the set of nim-reps (S
x
)
ab
:
Oc(G) G G
x.a

b
S
b
xa
b
The set of matrices S
x
, called dual annular matrices, forms a new anti-representation of the Ocneanu algebra
satisfying to
S
x
S
y
=

z
O
z
yx
S
z
.
2.3. A Weak Hopf Algebra Structure: The Double Triangle Algebra
There are two type of paths on a given graph G, the essential horizontal paths indexed by Young frames of
/(G) and the vertical ones labeled by Ocneanu generators of Oc(G). Endomorphisms of paths dene two dual
graded algebras denoted B(G) and

B(G) and represented by horizontal and vertical double triangles respectively
(see [8] and [32]). These are both nite dimensional algebras with isomorphic underlying vector spaces and each
one of the two algebra structures can be traded against a coalgebra structure on its dual. Equipped with a
multiplication law, a unit, a coproduct, a counit, and an antipode, the bialgebra B(G) (and of course

B(G))
satises all the axioms of weak Hopf algebras cited in [2, 3] for instance. These are a particular kind of weak
Hopf algebras called Ocneanu quantum groupods. The representation theory of B(G) is encoded by the fusion
algebra /(G) and the representation theory of

B(G) is realized by the Ocneanu algebra of quantum symmetries
Oc(G). Many explicit examples are known only for the SU(2) case as in [33, 34, 35, 36], in the following we try
to treat in a full and explicit way these structures for the rst member of the SU(3) system.
3. THE OCNEANU BIALGEBRA STRUCTURES
In what follows, we give explicit results on the simplest (but not trivial) case of SU(3) system: the /
1
graph
3
.
The unit vertex is 1 (0, 0), the left generator is 3 (1, 0) and the right one is 3 (0, 1). The level of the
graph is k = 1 and its dual Coxeter number is = k + 3 = 4. The following gure (Figure 3) shows the graph
/
1
and its adjacency matrix A(/
1
).

d
d
d
1 3
3
E

s
t re
re
A(/
1
) =

0 1 0
0 0 1
1 0 0

Figure 1: The /
1
graph and its adjacency matrix
The fusion matrices of the Verlinde algebra /
1
are:
N
1
= I
3
=

1 . .
. 1 .
. . 1

, N
3
= A(/
1
) =

. 1 .
. . 1
1 . .

, N
3
= A(/
1
)
tr
=

. . 1
1 . .
. 1 .

.
For the /
k
cases, the graph /
k
and the Ocneanu graph (/
k
) coincide, and more especially all the following
matrices are the same: N
i
= F
i
= O
i
= S
i
.
3
For brevity, we denote the generators by their classical dimensions i.e. 1, 3, and 3.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
230
3.1. Essential Paths on G and the Double Triangle Algebras
3.1.1. General Data
Essential horizontal paths (or horizontal triangle)

ab
of type going from a to b span a vector space Hpaths(G)
graded by : Hpaths(G) =
A
H

with cardinality

A
d

where d

=

a,bG
(F

)
ab
.
Vertical paths (or vertical triangles)
ab
x
of type x going from a to b span a vector space 1paths(G) graded
by x: 1paths(G) =
x
1
x
. Its dimension is

x
d
x
where d
x
=

a,bG
(S
x
)
ab
.

ab
=
d d
a b

d
d
a b


x
ab
=

d d
a
b
x
=

d
d
a
b
x
E

Figure 2:
Now we consider the vector spaces of endomorphisms on essential paths:
B =

(G)

H

(G)

B =
x

B
x
=

x
1
x
(G)

1
x
(G).
Equipped with a product called vertical product dened as a composition of endomorphism of essential
paths, B is endowed with an algebra structure called algebra of double horizontal triangles. In the same way,

B is the algebra of double vertical triangles equipped by a product called horizontal product. Moreover the
product in

B allows us to dene a coproduct for its dual B, and vice-versa, so that we nally obtain two
dually paired bialgebras, which are both semi-simple and co semi-simple.
A basis for the algebra (B, ) is dened by e

where e

stands for

ab

cd
and are represented by matrix
units. For the dual algebra
_

B,
_
a basis is dened by
_
f

_
where f

means
ab
x

cd
x
. These are the algebras
of double triangles or by duality the algebras of diusion graphs.
e

=
d d
dd

a a

b b

d
d
d
d

a a

b b

x
=

d d
dd

x
c

c
d

d
d
d
d

x
c
c

d
d

'

Figure 3:
3.1.2. The A1 Case
In the /
1
case, the oriented horizontal essential paths
l
pq
, assigned to Young frame l, starting from a vertex
p and ending at a vertex q are denoted as

pq
l
or p
l
, and span the vector space H
l
. The graded vector space of
all horizontal essential paths is the direct sum: H =
lA1
H
l
.
H
1
=
_

1
11
,
1
33
,
1
33
_
=
_

11
1
,

33
1
,

33
1
_
=
_
1
1
, 3
1
, 3
1
_
H
3
=
_

3
13
,
3
33
,
3
31
_
=
_

13
3
,

33
3
,

31
3
_
=
_
1
3
, 3
3
, 3
3
_
H
3
=
_

3
13
,
3
31
,
3
33
_
=
_

13
3
,

31
3
,

33
3
_
=
_
1
3
, 3
3
, 3
3
_
The algebra of horizontal double triangles or by duality the algebra of vertical diusion graphs is:
B =
lA1
B
l
=
lA1
H
l


H
l
= (H
1


H
1
) (H
3


H
3
) (H
3


H
3
)
The elements e
l

for each B
l
are shorten written as e
l
pq
= p
l
q
l
.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
231
B
1
=
_
e
1
11
, e
1
13
, e
1
13
, e
1
31
, e
1
33
, e
1
33
, e
1
31
, e
1
33
, e
1
33
_
B
3
=
_
e
3
11
, e
3
13
, e
3
13
, e
3
31
, e
3
33
, e
3
33
, e
3
31
, e
3
33
, e
3
33
_
B
3
=
_
e
3
11
, e
3
13
, e
3
13
, e
3
31
, e
3
33
, e
3
33
, e
3
31
, e
3
33
, e
3
33
_
d d
dd
1
1 1
1 1

s
d d
dd
1
1 1
3 3

s
d d
dd
1
1 1
3 3

s
d d
dd
1
3 3
1 1

s
d d
dd
1
3 3
3 3

s
d d
dd
1
3 3
3 3

s
d d
dd
1
3 3
1 1

s
d d
dd
1
3 3
3 3

s
d d
dd
1
3 3
3 3

B =
d d
dd
3
1 3
1 3

s
d d
dd
3
1 3
3
3

s
d d
dd
3
1 3
3
1

s
d d
dd
3
3
3
1 3

s
d d
dd
3
3
3
3
3

s
d d
dd
3
3
3
3
1

s
d d
dd
3
3
1
1 3

s
d d
dd
3
3
1
3
3

s
d d
dd
3
3
1
3
1

d d
dd
3
1
3
1
3

s
d d
dd
3
1
3
3 1

s
d d
dd
3
1
3
3
3

s
d d
dd
3
3 1
1
3

s
d d
dd
3
3 1
3 1

s
d d
dd
3
3 1
3
3

s
d d
dd
3
3
3
1
3

s
d d
dd
3
3
3
3 1

s
d d
dd
3
3
3
3
3

s
Figure 4:
The dimension of each block B
l
of B is dimB
l
= d
2
l
=
_

m,n
(F
l
)
mn
_
2
and dimB =

l
d
2
l
= 27. A basis
for the algebra B is made of matrix units (pq
l
) of dimension 9 9 and take the value
4
1 at the indicated place
and 0 elsewhere:
_
e
l
pq
_
=

e
1
11
e
1
13
e
1
13
e
1
31
e
1
33
e
1
33
e
1
31
e
1
33
e
1
33

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33

The oriented vertical essential paths


pq
L
, associated to an Ocneanu vertex x, represented here by L, starting
from a vertex p and ending at a vertex q are denoted pq

L
or p
L
, and span the vector space 1
L
. The global
graded vector space of all vertical essential paths is the direct sum: 1 =
L(A1)
1
L
.
1
1
=
_

11
1
,
33
1
,
33
1
_
=
_
11

1
, 33

1
, 33

1
_
=
_

1
1
,

3
1
,

3
1
_
1
3
=
_

11
3
,
33
3
,
33
3
_
=
_
13

3
, 33

3
, 33

3
_
=
_

1
3
,

3
3
,

3
3
_
1
3
=
_

11
3
,
33
3
,
33
3
_
=
_
13

3
, 31

3
, 33

3
_
=
_

1
3
,

3
3
,

3
3
_
The algebra of vertical double triangles or by duality the algebra of horizontal diusion graphs

B is the dual
vector space of the vector space B, it is a graded algebra which is isomorphic
5
to the algebra B such that:

B =

B
1


B
3


B
3
.
4
For example the matrix unit (31
3
) is a square matrix of size 9 9, where the unique non-zero value is 1 at the 5
th
line and 4
th
column.
5
In fact B and

B describe the same algebra structure.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
232
An appropriate basis of

B is made of vertical double triangles represented by matrix units f
pq
L
= p
L
q
L
(f
pq
L
) =

f
11
1
f
13
1
f
13
1
f
31
1
f
33
1
f
33
1
f
31
1
f
33
1
f
33
1

f
11
3
f
13
3
f
13
3
f
31
3
f
33
3
f
33
3
f
31
3
f
33
3
f
33
3

f
11
3
f
13
3
f
13
3
f
31
3
f
33
3
f
33
3
f
31
3
f
33
3
f
33
3

3.2. A Bialgebra Structure


3.2.1. The Algebra (B, )
The explicit denition of the vertical products on B is written as: e

. In the /
1
case
we obtain the following results

e
1
11
e
1
13
e
1
13
e
1
31
e
1
33
e
1
33
e
1
31
e
1
33
e
1
33
e
1
11
e
1
11
e
1
13
e
1
13
. . . . . .
e
1
13
. . . e
1
11
e
1
13
e
1
13
. . .
e
1
13
. . . . . . e
1
11
e
1
13
e
1
13
e
1
31
e
1
31
e
1
33
e
1
33
. . . . . .
e
1
33
. . . e
1
31
e
1
33
e
1
33
. . .
e
1
33
. . . . . . e
1
31
e
1
33
e
1
33
e
1
31
e
1
31
e
1
33
e
1
33
. . . . . .
e
1
33
. . . e
1
31
e
1
33
e
1
33
. . .
e
1
33
. . . . . . e
1
31
e
1
33
e
1
33

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33
e
3
11
e
3
11
e
3
13
e
3
13
. . . . . .
e
3
13
. . . e
3
11
e
3
13
e
3
13
. . .
e
3
13
. . . . . . e
3
11
e
3
13
e
3
13
e
3
31
e
3
31
e
3
33
e
3
33
. . . . . .
e
3
33
. . . e
3
31
e
3
33
e
3
33
. . .
e
3
33
. . . . . . e
3
31
e
3
33
e
3
33
e
3
31
e
3
31
e
3
33
e
3
33
. . . . . .
e
3
33
. . . e
3
31
e
3
33
e
3
33
. . .
e
3
33
. . . . . . e
3
31
e
3
33
e
3
33

d d
dd
1
1
1
1
1

'
s


d d
dd
1
1
1
3
3

'
s


d d
dd
1
1
1
3
3

'
s


d d
dd
1
3
3
1
1

'
s


d d
dd
1
3
3
3
3

'
s


d d
dd
1
3
3
3
3

'
s


d d
dd
1
3
3
1
1

'
s


d d
dd
1
3
3
3
3

'
s


d d
dd
1
3
3
3
3

'
s

B =

d d
dd
3
3
1
3
1

'
s


d d
dd
3
3
1
3
3

'
s


d d
dd
3
3
1
1
3

'
s


d d
dd
3
3
3
3
1

'
s


d d
dd
3
3
3
3
3

'
s


d d
dd
3
3
3
1
3

'
s


d d
dd
3
1
3
3
1

'
s


d d
dd
3
1
3
3
3

'
s


d d
dd
3
1
3
1
3

'
s


d d
dd

3
3
1
3
1

'
s


d d
dd

3
3
1
1
3

'
s


d d
dd

3
3
1
3
3

'
s


d d
dd

3
1
3
3
1

'
s


d d
dd

3
1
3
1
3

'
s


d d
dd

3
1
3
3
3

'
s


d d
dd

3
3
3
3
1

'
s


d d
dd

3
3
3
1
3

'
s


d d
dd

3
3
3
3
3

'
s

Figure 5:
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
233

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33
e
3
11
e
3
11
e
3
13
e
3
13
. . . . . .
e
3
13
. . . e
3
11
e
3
13
e
3
13
. . .
e
3
13
. . . . . . e
3
11
e
3
13
e
3
13
e
3
31
e
3
31
e
3
33
e
3
33
. . . . . .
e
3
33
. . . e
3
31
e
3
33
e
3
33
. . .
e
3
33
. . . . . . e
3
31
e
3
33
e
3
33
e
3
31
e
3
31
e
3
33
e
3
33
. . . . . .
e
3
33
. . . e
3
31
e
3
33
e
3
33
. . .
e
3
33
. . . . . . e
3
31
e
3
33
e
3
33
The unit element of (B, ) is
I = e
1
11
+ e
1
33
+ e
1
33
+ e
3
11
+ e
3
33
+ e
3
33
+ e
3
11
+ e
3
33
+ e
3
33
,
which can be written as I =

l

l
, l = 1, 3 and 3, where the
i
are the minimal central projections of B on B
i
:

1
= e
1
11
+ e
1
33
+ e
1
33

3
= e
3
11
+ e
3
33
+ e
3
33
,
3
= e
3
11
+ e
3
33
+ e
3
33
such that
2
k
=
k
,

= , and
k
B
k
=
k
B = B
k
.
3.2.2. The Coalgebra (B, )
The existence of a product in

B allows to dene a coproduct : B B B in such a way to fulll the
compatibility condition
6
of a bialgebra structure between and

, e

_
=

f

, e

_
.
For the /
1
case of SU(3), a coproduct
7
is dened in such a way that it is a morphism for the product i.e.
(pq
l
rs
k
) = (pq
l
) (rs
k
). This procedure follows the denition involved in [34] for the A
2
case of SU(2)
type. From this obtained coproduct, we should deduce the denition of on the dual vector space. Explicitly
we obtain the following results
(e
1
11
) = e
1
11
e
1
11
+ e
3
11
e
3
33
+ e
3
11
e
3
33
(e
1
13
) = e
1
13
e
1
13
+ e
3
13
e
3
33
+ e
3
13
e
3
31
(e
1
13
) = e
1
13
e
1
13
+ e
3
13
e
3
31
+ e
3
13
e
3
33
(e
1
31
) = e
1
31
e
1
31
+ e
3
31
e
3
33
+ e
3
31
e
3
13
(e
1
33
) = e
1
33
e
1
33
+ e
3
33
e
3
33
+ e
3
33
e
3
11
(e
1
33
) = e
1
33
e
1
33
+ e
3
33
e
3
31
+ e
3
33
e
3
13
(e
1
31
) = e
1
31
e
1
31
+ e
3
31
e
3
13
+ e
3
31
e
3
33
(e
1
33
) = e
1
33
e
1
33
+ e
3
33
e
3
13
+ e
3
33
e
3
31
(e
1
33
) = e
1
33
e
1
33
+ e
3
33
e
3
11
+ e
3
33
e
3
11
(e
3
11
) = e
1
11
e
3
11
+ e
3
11
e
1
33
+ e
3
11
e
3
33
(e
3
13
) = e
1
13
e
3
13
+ e
3
13
e
1
33
+ e
3
13
e
3
31
(e
3
13
) = e
1
13
e
3
13
+ e
3
13
e
1
31
+ e
3
13
e
3
33
(e
3
31
) = e
1
31
e
3
31
+ e
3
31
e
1
33
+ e
3
31
e
3
13
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
1
33
+ e
3
33
e
3
11
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
1
31
+ e
3
33
e
3
13
(e
3
31
) = e
1
31
e
3
31
+ e
3
31
e
1
13
+ e
3
31
e
3
33
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
1
13
+ e
3
33
e
3
31
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
1
11
+ e
3
33
e
3
11
6
In general cases the values of cells appear in the pairing between both types of matrix units (respectively associated with the
products and )
7
The denition of this rst coproduct was given by E. Isasi [40].
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
234
(e
3
11
) = e
1
11
e
3
11
+ e
3
11
e
3
33
+ e
3
11
e
1
33
(e
3
131
) = e
1
13
e
3
13
+ e
3
13
e
3
33
+ e
3
13
e
1
31
(e
3
13
) = e
1
13
e
3
13
+ e
3
13
e
3
31
+ e
3
13
e
1
33
(e
3
31
) = e
1
31
e
3
31
+ e
3
31
e
3
33
+ e
3
31
e
1
13
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
3
33
+ e
3
33
e
1
11
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
3
31
+ e
3
33
e
1
13
(e
3
31
) = e
1
31
e
3
31
+ e
3
31
e
3
13
+ e
3
31
e
1
33
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
3
13
+ e
3
33
e
1
31
(e
3
33
) = e
1
33
e
3
33
+ e
3
33
e
3
11
+ e
3
33
e
1
11
Remark that the coproduct of the unit element of B is not I I, but of the form (I) =

I
(1)
I
(2)
as
(I) = (e
1
11
+ e
3
33
+ e
3
33
) (e
1
11
+ e
3
11
+ e
3
11
) + (e
1
33
+ e
3
11
+ e
3
33
) (e
1
33
+ e
3
33
+ e
3
33
)
+ (e
1
33
+ e
3
33
+ e
3
11
) (e
1
33
+ e
3
33
+ e
3
33
)
The counit of B can be dened as a C-valued linear map on B satisfying (e e

) =
_
eI
(1)
_

_
I
(2)
e

_
. For the
/
1
case we have
(e
l
pq
) =
_
1, if l = 1
0, if l = 3, 3
Note that the map is not an algebra homomorphism.
One can also dene an antipode S on B as an algebra anti-homomorphism like a conjugation of elements
of B in the following way: S
_
e

_
= ke

, where =

ab
=

ba
and k =
_
(a)(d)
(b)(c)
is a function of quantum
dimensions of the vertices of G. In the /
1
case, since k = 1 (all the Perron-Frobenius components of the /
1
graph are equal to 1) we get S
_
e
l
pq
_
= e
l
qp
.
3.3. The Algebra
_

B,
_
The explicit denition of the horizontal product on

B is written as: f

x
f

x
=
xx

. For the /
1
case it will involve in

B the dual basis ( e
l
pq
) of (e
l
pq
) instead of the basis of matrix units of the product: (f
pq
L
).
Before manipulating with this change of basis in

B we just mention the form of the unit element in terms of f
pq
L
:

I = f
11
1
+ f
33
1
+ f
33
1
+ f
11
3
+ f
33
3
+ f
33
3
+ f
11
3
+ f
33
3
+ f
33
3
Remark that

I =

L

L
, L = 1, 3 and 3, where
L
are the minimal central projections of

B, i.e.
L
=

L
=
2
L
and
L
(

B) =

B
L
.
To dene the product of

B for the /
1
case, we use the compatibility condition fullled by the product of

B and the coproduct on B in terms of the dual basis (e

ab
) and ( e

ab
):
e

ab
e

ab
, (e

ab
)) = e

ab
e

ab
, e

ab
)
So, from the known explicit action of on elements e
l
pq
of B (/
1
), we deduce the non-zero values of

e
1
11
e
1
13
e
1
13
e
1
31
e
1
33
e
1
33
e
1
31
e
1
33
e
1
33
e
1
11
e
1
11
. . . . . . . .
e
1
13
. e
1
13
. . . . . . .
e
1
13
. . e
1
13
. . . . . .
e
1
31
. . . e
1
31
. . . . .
e
1
33
. . . . e
1
33
. . . .
e
1
33
. . . . . e
1
33
. . .
e
1
31
. . . . . . e
1
31
. .
e
1
33
. . . . . . . e
1
33
.
e
1
33
. . . . . . . . e
1
33
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
235

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33
e
3
11
. . . . e
3
11
. . . .
e
3
13
. . . . . e
3
33
. . .
e
3
13
. . . e
3
13
. . . . .
e
3
31
. . . . . . . e
3
31
.
e
3
33
. . . . . . . . e
3
33
e
3
33
. . . . . . e
3
33
. .
e
3
31
. e
3
31
. . . . . . .
e
3
33
. . e
3
33
. . . . . .
e
3
33
e
3
33
. . . . . . . .

e
3
11
e
3
13
e
3
13
e
3
31
e
3
33
e
3
33
e
3
31
e
3
33
e
3
33
e
3
11
. . . . . . . . e
3
11
e
3
13
. . . . . . e
3
13
. .
e
3
13
. . . . . . . e
3
13
.
e
3
31
. . e
3
31
. . . . . .
e
3
33
e
3
33
. . . . . . . .
e
3
33
. e
3
33
. . . . . . .
e
3
31
. . . . . e
3
31
. . .
e
3
33
. . . e
3
33
. . . . .
e
3
33
. . . . e
3
33
. . . .
3.4. The Coalgebra
_

B,

_
In an analogous way a coproduct

is dened on

B by the use of the product in its dual B:
_

_
f

_
, e

_
=

f

, e

_
.
For the present example, an explicit action of

on the elements of the dual basis ( e
l
pq
), such that the
homomorphism algebra

( e
l
pq
e
l

q
) =

( e
l
pq
)

( e
l

q
) is fullled, reads as

( e
1
11
) = e
1
11
e
1
11

( e
1
13
) = e
1
13
e
1
13

( e
1
13
) = e
1
13
e
1
13

( e
1
31
) = e
1
31
e
1
31

( e
1
33
) = e
1
33
e
1
33

( e
1
33
) = e
1
33
e
1
33

( e
1
31
) = e
1
31
e
1
31

( e
1
33
) = e
1
33
e
1
33

( e
1
33
) = e
1
33
e
1
33

( e
3
11
) = e
3
11
e
3
33

( e
3
13
) = e
3
13
e
3
33

( e
3
13
) = e
3
13
e
3
31

( e
3
31
) = e
3
31
e
3
33

( e
3
33
) = e
3
33
e
3
33

( e
3
33
) = e
3
33
e
3
31

( e
3
31
) = e
3
31
e
3
13

( e
3
33
) = e
3
33
e
3
13

( e
3
33
) = e
3
33
e
3
11

( e
3
11
) = e
3
11
e
3
33

( e
3
13
) = e
3
13
e
3
33

( e
3
13
) = e
3
13
e
3
31

( e
3
31
) = e
3
31
e
3
33

( e
3
33
) = e
3
33
e
3
33

( e
3
33
) = e
3
33
e
3
31

( e
3
31
) = e
3
31
e
3
13

( e
3
33
) = e
3
33
e
3
13

( e
3
33
) = e
3
33
e
3
11
A counit of

B can be dened as a C-valued linear map on

B satisfying (f f

) =
_
f

I
(1)
_

I
(2)
f

_
, where

I
(1)
and

I
(2)
come from the right hand side of

acting on

I
(1)
as:

(

I =

I
(1)

I
(2)
. For the /
1
case, can
be dened on the dual basis as
( e
l
pq
) =
1 if l = 1
0 if l = 3, 3
.
Note that the map is not an algebra homomorphism.
One can also dene an antipode

S on

B as an algebra anti-homomorphism by

S (f

x
) = kf

x
. For the /
1
case, we have simply

S
_
e
l
pq
_
= e
l
qp
.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
236
4. CONCLUSION
The spaces B and

B are both semi-simple and co semi-simple bialgebras which are isomorphic: B (, I, , )
and

B
_
,

I,

,
_
. Moreover, equipped with the antipodes S and

S, B and

B are weak Hopf algebras [2, 3] and
describe the same algebra of double triangles [8]. In more general cases, for instance /
p
, p > 1, in order to
ensure the axioms of weak Hopf algebras, one might to introduce the values of cells generalizing the quantum 3j
and 6j-symbols. One type of these cells are called Ocneanu cells and gives the pairing
_
e

, f

x
_
between the
not dual sets of basis
_
e

_
in B and
_
f

x
_
in

B, the other type ensure the associativity and the actions of
and

[35, 32]. There are many technical diculties in calculation of the set of cells for higher level; the studied
example /
1
is the only case of SU(3) for which the quantum groupod structure is explicitly determined. A
good amount of information about all graphs of SU(3) type is available in [26, 27] and several examples of ADE
type are studied for instance in [34, 33, 41, 35].
ACKNOWLEDGMENTS
This work emanated from a question asked by Dr E. Isasi, I would like to thank him together with Professors
R. Coquereaux, E. H. Tahri and G. Schieber for many useful and enjoyable discussions. I would like also to
thank Professor J. Abuhlail for his kind invitation to me to contribute to the AJSE.
REFERENCES
[1] A. Ocneanu, The Classication of Subgroups of Quantum SU(N). Lectures at Bariloche summer school,
Quantum Symmetries in Theoretical Physics and Mathematics Argentina, January 2000, R. Coquereaux,
AMS Contemp. 294., A. Garca and R. Trinchero (eds.).
[2] G. Bohm and K. Szlach anyi, A Coassociative C

-Quantum Group with Non-Integral Dimensions, Lett.


Math. Phys. 200 (1996), pp. 437456.
[3] F. Nill, Axioms for Weak Bialgebras, Di. Geom. Quan. Phy. 334 (1998), pp. 148.
[4] J. Bockenhauer and D. E. Evans, Modular Invariants, Graphs and -Induction for Nets of Subfactors I,
Comm. Math. Phys. 197 (1998), pp. 361386; II, Comm. Math. Phys. 200 (1999), pp. 57103; III, Comm.
Math. Phys. 205 (1999), pp. 183228.
[5] J. Bockenhauer, D. E. Evans, and Y. Kawahigashi, On -Induction, Chiral Generators and Modular
Invariants for Subfactors, Comm. Math. Phys. 208 (1999), pp. 429487.
[6] R. Longo and K. -H. Rehren, Nets of Subfactors, Rev. Math. Phys. 7 (1995), pp. 567497.
[7] V. Ostrik, Module Categories, Weak Hopf Algebras and Modular Invariants, Transform. Groups. 8 (2003),
pp. 177206.
[8] A. Ocneanu, Paths on Coxeter Diagrams: From Platonic Solids and Singularities to Minimal Models and
Subfactors, (Notes by S. Goto) In: Lectures on operator theory. Fields Institute Monographs 13, R. Bhat,
G. A. Elliott, and P. A. Fillmore (eds.). AMS Providence, 1999.
[9] V. Pasquier, Ethiology of IRF Models, Comm. Math. Phys. 118 (1988), pp. 335364.
[10] V. Pasquier, Two-Dimensional Critical Systems Labelled by Dynkin Diagrams, Nucl. Phys. B. 285 [FS19]
(1987), pp. 162172.
[11] Ph. Roche, Ocneanu Cell Calculus and Integrable Lattice Models, Comm. Math. Phys. 127 (1990), pp.
395424.
[12] P. Di Francesco and J.-B. Zuber, SU(N) Lattice Integrable Models Associated with Graphs, Nucl. Phys.
B 338 (1990), pp. 602646.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
237
[13] P. A. Pearce and Y. K. Zhou, Intertwiners and A-D-E Lattice Models, Int. J. Mod. Phys. B 7 (1993),
pp. 36493705.
[14] N. Sochen, Integrable Models through Representations of the Hecke Algebras, Nucl. Phys. B 360 (1991),
pp. 613640.
[15] A. Cappelli, C. Itzykson, and J. -B. Zuber, Modular Invariant Partition Functions in Two Dimensions,
Nucl. Phys. B 280 (1987), pp. 445465.
[16] A. Cappelli, C. Itzykson, and J. -B. Zuber, The ADE Classication of Minimal and A
(1)
1
Conformal
Invariants Theories, Comm. Math. Phys. 113 (1987), pp. 120.
[17] T. Gannon, The Classication of Ane SU(3) Modular Invariants, Comm. Math. Phys. 161 (1994), pp.
233263.
[18] D. Hammaoui, G. Schieber, and E. H. Tahri, Higher Coxeter Graphs Associated to Ane su(3) Modular
Invariants, J. Phys. A: Math. Gen. 38 (2005), pp. 82598268.
[19] R. Coquereaux and G. Schieber, Twisted Partition Functions for ADE Boundary Conformal Field Theories
and Ocneanu Algebras of Quantum Symmetries, J. Geom. Phys. 781 (2002), pp. 143.
[20] R. Coquereaux and M. Huerta, Torus Structure on Graphs and Twisted Partition Functions for Minimal
and Ane Models, J. Geom. Phys. 48 (2003), pp. 580634.
[21] V. B. Petkova and J. -B. Zuber, Generalized Twisted Partition Functions, Phys. Lett. B 504 (2001), pp.
157164.
[22] J. Fuchs, I. Runkel, and C. Schweigret, TFT Construction of RCFT Correlators I: Partition Functions,
Nucl. Phys. B 646 (2002), pp. 353497; II: Unoriented World Sheets, Nucl. Phys. B 678 (2004), pp.
511637; III, Simple Currents, Nucl. Phys. B 694 (2004), pp. 277353; IV: Structure Constants and
Correlations Functions, Nucl. Phys. B 715 (2005), pp. 539638.
[23] R. Coquereaux, Notes on the Quantum Tetrahedron, Moscow Math. J. 2. (2002), pp. 4180; Notes on
the Classical Tetrahedron: a Fusion Graph Algebra Point of View, http://www.cpt.univ-mrs.fr/

coque/.
[24] R. Coquereaux and G. Schieber, Determination of Quantum Symmetries for Higher ADE Systems from
the Modular T Matrix, J. Math. Phys. 44 (2003), pp. 38093837.
[25] R. Coquereaux and E. Isasi, On Quantum Symmetries of the Non-ADE Graph F
4
, Adv. Theor. Math.
Phys. 8 (2004), pp. 955985.
[26] R. Coquereaux, D. Hammaoui, G. Schieber and E. H. Tahri, Comments About Quantum Symmetries of
SU(3) Graphs, J. Geom. Phys. 57 (2006), pp. 269292.
[27] D. Hammaoui Geometrie quantique dOcneanu des graphes de Di Francesco-Zuber associes aux mod`eles
conformes de type

su(3), Th`ese de Doctorat national, LPTPM, Oujda University, Morocco, January 2007.
[28] E. Isasi and G. Schieber, From Modular Invariants to Graphs: The Modular Splitting Method, J. Phys.
A: Math. Theor 40 (2007), pp. 6513-6537.
[29] R. Coquereaux and G. Schieber, From Conformal Embeddings to Quantum Symmetries: An Excep-
tional SU(4) Example, IOP, Journal of Physics, Conference Series: International Conference on Non-
Commutative Geometry and Physics, 23-27 April 2007, LPT, Universite Paris XI, France. Arxiv: math-
ph/07101397.
[30] R. Coquereaux and G. Schieber, Quantum Symmetries for Exceptional SU(4) Modular Invariants Associ-
ated with Conformal Embeddings, ArXiv:0805.4678.
[31] A. Ocneanu, Quantum Symmetries, Operator Algebras and Invariants of Manifolds, talk given at the
First Caribbean Spring School of Mathematical and Theoretical Physics, Saint-Fran cois, Guadeloupe, 1993.
Dahmane Hammaoui
The Arabian Journal for Science and Engineering, Volume 33, Number 2C December 2008
238
[32] V. B. Petkova and J. -B. Zuber, The Many Faces of Ocneanu Cells, Nucl. Phys. B 603 (2001), pp.
449496.
[33] R. Coquereaux and R. Trinchero, On Quantum Symmetries of ADE Graphs, Adv. Theor. Math. Phys. 8
(2004), pp. 189216.
[34] R. Coquereaux, The A
2
Ocneanu Quantum Groupod, in Algebraic structures and their representations.
Contemporary Mathematics 376 (2005), pp. 227247.
[35] R. Coquereaux, Racah-Wigner Quantum 6j Symbols, Ocneanu Cells for A
N
Diagrams and Quantum
Groupods, J. Geom. Phys. 57(2) (2007), pp. 387434.
[36] R. Trinchero, Symmetries of Faces Models and the Double Triangle Algebra, Adv. Theor. Math. Phys.
10 (2006), pp. 49-75.
[37] P. Di Francesco, P. Mathieu, and D. Senechal, Conformal Field Theory. New York: Springer-Verlag, 1997.
[38] E. Verlinde, Fusion Rules and Modular Transformations in 2D Conformal Field Theories, Nucl. Phys. B
300 (1988), pp. 360376.
[39] J. Cardy, Boundary Conditions, Fusion Rules and the Verlinde Formula, Nucl. Phys. B 324 (1989), pp.
581596.
[40] E. Isasi: Methode de scission modulaire et symetries quantiques des graphes non-simplement laces en
theorie des champs conforme, PhD Thesis, UP (Marseille), Oct. 2006.
[41] R. Coquereaux and A. Garcia, On Bialgebras Associated with Paths and Essential Paths on ADE Graphs,
Int. J. Geom. Meth. Mod. Phys. 2 (2005), pp. 441466.

Vous aimerez peut-être aussi