, , ,

This paper is a continuation of the authors' work [33] at which several coset subalgebras of the lattice VOA

V_{\sqrt{2} E_{8}}

were constructed and the relationship between such algebras with the famous McKay observation [2, 38] on the extended

E_{8}

diagram

\begin{matrix} \begin{matrix} 3 C \frac{1}{2^{8}} \\ \circ \\ | \\ | \\ \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ \\ 1 A 2 A 3 A 4 A 5 A 6 A 4 B 2 B \\ \frac{1}{4} \frac{1}{32} \frac{13}{2^{10}} \frac{1}{2^{7}} \frac{3}{2^{9}} \frac{5}{2^{10}} \frac{1}{2^{8}} 0 \end{matrix} \end{matrix}

(1.1)

and the Monster simple group were discussed.

In this article, we shall provide the technical details. We shall determine the structure of the coset subalgebras and show that they are all generated by two conformal vectors of central charge

1 / 2

. We also study the representation theory of these coset subalgebras and show that the product of two Miyamoto involutions is in the desired conjugacy class of the Monster simple group if a coset subalgebra

U

is actually contained in the Moonshine VOA

V^{♮}

. The existence of

U

inside the Moonshine VOA

V^{♮}

for the cases of

1 A, 2 A, 2 B

and

4 A

is also established. Moreover, the cases for

3 A

5 A

and

3 C

are discussed.

The organization of the article is as follows. In Section 2 we shall review some important notation and terminology from [33] . We review certain conformal vectors in the lattice VOA

V_{\sqrt{2} R}

, where

R

is a root lattice of type

A

D

, or

E

(cf. [7] ). We then consider the sublattice

L

E_{8}

and define the coset subalgebra

U

and two conformal vectors

\hat{e}

and

\hat{f}

of central charge

1 / 2

. A canonical automorphism

σ

of order

n = | E_{8} / L |

induced by the quotient group

E_{8} / L

is also discussed. In Section 3, we study the structure of

U

in each of the nine cases corresponding to the McKay's diagram. We also study the representation theory of these coset subalgebras and show that the product of two Miyamoto involutions is in the desired conjugacy class of the Monster simple group if a coset subalgebra

U

is actually contained in the Moonshine VOA

V^{♮}

. The existence of

U

inside the Moonshine VOA

V^{♮}

for the cases of

1 A, 2 A, 2 B

and

4 A

is also established. Moreover, the cases for

3 A

5 A

and

3 C

are discussed. Appendix contains the classification of conformal vectors in

U

, calculations of certain characters which are used in Section 3 and the classification of irreducible modules for

5 A

and

3 C

cases.

The authors thank Masahiko Miyamoto and Masaaki Kitazume for stimulating discussions and Kazuhiro Yokoyama for helping them to compute the conformal vectors for the cases of

5 A

and

6 A

by a computer algebra system Risa/Asir. In Appendix C we study an extension of a simple rational VOA by an irreducible module which is not a simple current module.

A similar extension is also considered in [46] . The authors thank Kenichiro Tanabe forvaluable discussions concerning it. Part of the work was done while the third author (H.

Yamauchi) was visiting the National Center for Theoretical Science of Taiwan in August, 2004. He thanks the center for the hospitality during the stay.

2 Preliminary

In this section, we shall recall the notation and the constructions of certain coset subalgebras of

V_{\sqrt{2} E_{8}}

and their automorphisms from [33] . We shall mainly deal with lattice VOAs introduced by [17] . Let

V_{N}

be a lattice VOA associated with any positive definite even lattice

N

. By [34,Theorem3.1] , there is a unique symmetric invariant bilinear form

〈 \cdot, \cdot 〉

V_{N}

such that

〈 1, 1 〉 = 1

. That the form

〈 \cdot, \cdot 〉

is invariant, i.e.,

〈 Y (u, z) v, w 〉 = 〈 v, Y (e^{z L (1)} (- z^{- 2})^{L (0)} u, z^{- 1}) w 〉 for u, v, w \in V_{N} .

For

u, v \in (V_{N})_{2}

with

L (1) u = 0

, the invariance of the form implies that

〈 u, v 〉 = 〈 1, u_{3} v 〉

which induces a bilinear form on

V_{2}

. It is also well known that

V_{N}

possesses a positive definite invariant hermitian form

(\cdot, \cdot)

. Indeed, let

V_{N, R}

be the

R

-form of

V_{N}

defined as in [17,Section12.4] . Then

V_{N, R}

is invariant under the automorphism

θ

, where

θ

is a lift of

- 1

isometry of the lattice

N

. Let

V_{N, R}^{\pm}

be the eigenspaces for

θ

with eigenvalues

\pm 1

. Then

〈 \cdot, \cdot 〉

is positive definite on

V_{N, R}^{+}

and negative definite on

V_{N, R}^{-}

. Moreover,

〈 V_{N, R}^{+}, V_{N, R}^{-} 〉 = 0

. Hence

〈 \cdot, \cdot 〉

is positive definite on the

R

-vector space

{\tilde{V}}_{N, R} = V_{N, R}^{+} + \sqrt{- 1} V_{N, R}^{-}

. Clearly,

V_{N} = C \otimes_{R} {\tilde{V}}_{N, R}

and so

{\tilde{V}}_{N, R}

is an

R

-form of

V_{N}

. Define a hermitian form

(\cdot, \cdot)

V_{N}

(λ u, μ v) = λ \bar{μ} (u, v)

for

λ, μ \in C

and

u, v \in {\tilde{V}}_{N, R}

. Then

(\cdot, \cdot)

is positive definite on

V_{N}

. Furthermore, it is

{\tilde{V}}_{N, R}

-invariant, that is,

(Y (u, z) v, w) = (v, Y (e^{z L (1)} (- z^{- 2})^{L (0)} u, z^{- 1}) w)

for

u \in {\tilde{V}}_{N, R}

and

v, w \in V_{N}

, where

L (0) = ω_{1}

and

L (1) = ω_{2}

with

ω

being the Virasoro element of

V_{N}

. These two forms

〈 \cdot, \cdot 〉

and

(\cdot, \cdot)

for the case

N = \sqrt{2} E_{8}

will be used in Section 2 and 3.

2.1 Conformal vectors

We shall now review the construction of certain conformal vectors in the lattice VOA

V_{\sqrt{2} R}

from [7] , where

R

is a root lattice of type

A_{n}

D_{n}

, or

E_{n}

Let

Φ

be the root system of

R

and

Φ^{+}

and

Φ^{-}

the set of all positive roots and negative roots, respectively. Then

Φ = Φ^{+} \cup Φ^{-} = Φ^{+} \cup (- Φ^{+})

. The Virasoro element

ω

V_{\sqrt{2} R}

is given by

ω = ω (Φ) = \frac{1}{2 h} \sum_{α \in Φ^{+}} α (- 1)^{2} \cdot 1,

where

h

is the Coxeter number of

Φ

. Now define

\begin{matrix} \begin{matrix} s = s (Φ) & = \frac{1}{2 (h + 2)} \sum_{α \in Φ^{+}} (α (- 1)^{2} \cdot 1 - 2 (e^{\sqrt{2} α} + e^{- \sqrt{2} α})), & \tilde{ω} = \tilde{ω} (Φ) & = ω - s . \end{matrix} \end{matrix}

(2.1)

It is shown in [7] that

\tilde{ω}

and

s

are mutually orthogonal conformal vectors and the central charge of

\tilde{ω}

2 n / (n + 3)

R

is of type

A_{n}

1

R

is of type

D_{n}

and

6 / 7, 7 / 10

and

1 / 2

R

is of type

E_{6}, E_{7}

and

E_{8}

, respectively. In this article, we denote by

V i r (x)

the Virasoro sub VOA generated by a conformal vector

x

V

Now let

A u t (Φ)

be the automorphism group of

Φ

. For any element

g \in A u t (Φ)

g

induces an automorphism on the lattice

R

and hence also defines an automorphism of the VOA

V_{\sqrt{2} R}

g (u \otimes e^{\sqrt{2} α}) = g u \otimes e^{\sqrt{2} g α} for u \otimes e^{\sqrt{2} α} \in M (1) \otimes e^{\sqrt{2} α} \subset V_{\sqrt{2} R} .

Note that both

s

and

\tilde{ω}

are fixed by

A u t (Φ)

and thus also fixed by the Weyl group

W (Φ)

Φ

Let

R^{*} = {α \in Q \otimes_{Z} R | 〈 α, R 〉 \subset Z}

be the dual lattice of

R

. Then we have the following Proposition.

Proposition 2.1 (cf. Section 2 of [33] ). Let

γ + R

be a coset of

R

R^{*}

and

k = min {〈 α, α 〉 | α \in γ + R}

. Define

v = \sum_{\binom{α \in γ + R}{〈 α, α 〉 = k}} e^{\sqrt{2} α} \in V_{\sqrt{2} (γ + R)} .

Then

v

is a highest weight vector of highest weight

(0, k)

V_{\sqrt{2} (γ + R)}

with respect to

V i r (s) \otimes V i r (\tilde{ω})

, that is,

s_{j} v = {\tilde{ω}}_{j} v = 0

for all

j \geq 2

s_{1} v = 0

, and

{\tilde{ω}}_{1} v = k v

. In other words, with respect to

s

, there is always a highest weight vector of weight

0

V_{\sqrt{2} (γ + R)}

2.2 Extended $E_{8}$ diagram and coset subalgebras of $V_{\sqrt{2} E_{8}}$

Next, we shall review the construction some coset VOAs

U

using the extended

E_{8}

diagram. In each case,

U

contains some conformal vectors of central charge

1 / 2

and the inner products among these conformal vectors are the same as the numbers given in the McKay diagram 1.1 . First, we shall consider certain sublattices of the root lattice

E_{8}

by using the extended

E_{8}

diagram

\begin{matrix} \begin{matrix} α_{8} \\ \circ \\ | \\ | \\ \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ - - - - - \circ \\ α_{0} α_{1} α_{2} α_{3} α_{4} α_{5} α_{6} α_{7} \end{matrix} \end{matrix}

(2.2)

where

α_{1}, α_{2}, . . ., α_{8}

are the simple roots of

E_{8}

and

\begin{matrix} α_{0} + 2 α_{1} + 3 α_{2} + 4 α_{3} + 5 α_{4} + 6 α_{5} + 4 α_{6} + 2 α_{7} + 3 α_{8} = 0 . \end{matrix}

(2.3)

Then

〈 α_{i}, α_{i} 〉 = 2

0 \leq i \leq 8

. Moreover, for

i \neq j

〈 α_{i}, α_{j} 〉 = - 1

if the nodes

α_{i}

and

α_{j}

are connected by an edge and

〈 α_{i}, α_{j} 〉 = 0

otherwise. Note that

- α_{0}

is the highest root.

For any

i = 0, 1, . . ., 8

, let

L (i)

be the sublattice generated by

α_{j}, 0 \leq j \leq 8, j \neq i

Then

L (i)

is a rank

8

sublattice of

E_{8}

. Note

L (i)

is the lattice associated with the Dynkin diagram obtained by removing the corresponding node

α_{i}

in the extended

E_{8}

diagram 2.2 and the index

| E_{8} / L (i) |

is equal to

n_{i}

, where

n_{i}

is the coefficient of

α_{i}

in the left hand side of 2.3 . Actually,

\begin{matrix} \begin{matrix} L (0) & \sim = E_{8}, \\ L (3) & \sim = A_{3} \oplus D_{5}, \\ L (6) & \sim = A_{7} \oplus A_{1}, \end{matrix} \begin{matrix} L (1) & \sim = A_{1} \oplus E_{7}, \\ L (4) & \sim = A_{4} \oplus A_{4}, \\ L (7) & \sim = D_{8}, \end{matrix} \begin{matrix} L (2) & \sim = A_{2} \oplus E_{6}, \\ L (5) & \sim = A_{5} \oplus A_{2} \oplus A_{1}, \\ L (8) & \sim = A_{8} . \end{matrix} \end{matrix}

(2.4)

Now let us explain the details of our construction. First, we fix

i \in {0, 1, . . ., 8}

and denote

L (i)

L

. In each case,

| E_{8} / L | = n_{i}

and

α_{i} + L

is a generator of the quotient group

E_{8} / L

. Hence we have

\begin{matrix} E_{8} = L \cup (α_{i} + L) \cup (2 α_{i} + L) \cup \dots \cup ((n_{i} - 1) α_{i} + L) . \end{matrix}

(2.5)

Let

λ = \sqrt{2} α_{i}

. Then

\sqrt{2} E_{8} = \sqrt{2} L \cup (λ + \sqrt{2} L) \cup (2 λ + \sqrt{2} L) \cup \dots \cup ((n_{i} - 1) λ + \sqrt{2} L)

and the lattice VOA

V_{\sqrt{2} E_{8}}

can be decomposed as

V_{\sqrt{2} E_{8}} = V_{\sqrt{2} L} \oplus V_{λ + \sqrt{2} L} \oplus \dots \oplus V_{(n_{i} - 1) λ + \sqrt{2} L},

where

V_{j λ + \sqrt{2} L}

j = 0, 1, . . ., n_{i} - 1

, are irreducible modules of

V_{\sqrt{2} L}

(cf. [4] ).

Remark 2.2. The abelian group

E_{8} / L

actually induces an automorphism

σ

V_{\sqrt{2} E_{8}}

such that

\begin{matrix} σ (u) = ξ^{j} u for any u \in V_{j λ + \sqrt{2} L}, \end{matrix}

(2.6)

where

ξ = e^{2 π \sqrt{- 1} / n_{i}}

is a primitive

n_{i}

-th root of unity. More precisely, let

\begin{matrix} a = {\begin{matrix} α_{1} & if i = 0, \\ - \frac{1}{i + 1} (α_{0} + 2 α_{1} + \dots + i α_{i - 1}) & if 1 \leq i \leq 5, \\ - \frac{1}{8} (α_{0} + 2 α_{1} + \dots + 6 α_{5} + 7 α_{8}) & if i = 6, \\ \frac{1}{2} (α_{6} + α_{8}) & if i = 7, \\ - \frac{1}{9} (α_{0} + 2 α_{1} + \dots + 8 α_{7}) & if i = 8 . \end{matrix} \end{matrix}

(2.7)

Then

〈 a, α_{j} 〉 \in Z

for

0 \leq j \leq 8

with

j \neq i

and

〈 a, α_{i} 〉 \equiv - 1 / n_{i} (m o d Z)

. The automorphism

σ : V_{\sqrt{2} E_{8}} \to V_{\sqrt{2} E_{8}}

is in fact defined by

\begin{matrix} σ = e^{- π \sqrt{- 1} β (0)} = exp (- \sqrt{- 2} π a (0)) with β = \sqrt{2} a . \end{matrix}

(2.8)

For

u \in M (1) \otimes e^{α} \subset V_{\sqrt{2} E_{8}}

, we have

σ (u) = e^{- π \sqrt{- 1} 〈 β, α 〉} u

. Note that

a + R

is a generator of the group

R^{*} / R

for the cases

i \neq 0, 7

, where

R

is an indecomposable component of the lattice

L

of type

A

For any lattice VOA

V_{N}

associated with a positive definite even lattice

N

, there is a natural involution

θ

induced by the isometry

α \to - α

for

α \in N

. If

N = \sqrt{2} E_{8}

, which is doubly even, we may define

θ : V_{\sqrt{2} E_{8}} \to V_{\sqrt{2} E_{8}}

α (- n) \to - α (- n) and e^{α} \to e^{- α}

for any

α \in \sqrt{2} E_{8}

(cf. [17] ). Then

θ σ θ = σ^{- 1}

and the group generated by

θ

and

σ

is isomorphic to a dihedral group of order

2 n_{i}

Next we shall recall the definition of certain coset subalgebras from [33] . Let

R_{1}, . . ., R_{l}

be the indecomposable components of the lattice

L

and

Φ_{1}, . . ., Φ_{l}

the corresponding root systems of

R_{1}, . . ., R_{l}

(cf. 2.4 ). Then

L = R_{1} \oplus \dots \oplus R_{l}

and

V_{\sqrt{2} L} \sim = V_{\sqrt{2} R_{1}} \otimes \dots \otimes V_{\sqrt{2} R_{l}} .

By 2.1 in Section 2, one obtains

2 l

mutually orthogonal conformal vectors

\begin{matrix} s^{k} = s (Φ_{k}), {\tilde{ω}}^{k} = \tilde{ω} (Φ_{k}), 1 \leq k \leq l \end{matrix}

(2.9)

such that the Virasoro element

ω

V_{\sqrt{2} L}

, which is also the Virasoro element of

V_{\sqrt{2} E_{8}}

, can be written as a sum of these conformal vectors

ω = s^{1} + \dots + s^{l} + {\tilde{ω}}^{1} + \dots + {\tilde{ω}}^{l} .

Now we define

U

to be a coset (or commutant) subalgebra

\begin{matrix} U = {v \in V_{\sqrt{2} E_{8}} | (s^{k})_{1} v = 0 for all k = 1, . . ., l} . \end{matrix}

(2.10)

Note that

U

is a VOA with the Virasoro element

ω^{'} = {\tilde{ω}}^{1} + \dots + {\tilde{ω}}^{l}

and the automorphism

σ

defined by ( 2.6 ) induces an automorphism of order

n_{i}

U

. By abuse of notation, we shall denote it by

σ

also.

Remark 2.3. In [32] , it is shown that

{v \in V_{\sqrt{2} A_{n}} | s (A_{n})_{1} v = 0}

is isomorphic to a parafermion algebra

W_{n + 1} (2 n / (n + 3))

of central charge

2 n / (n + 3)

. Thus, if

L

has some indecomposable component of type

A_{n}

, then

U

will contain some subalgebra isomorphic to a parafermion algebra. It is well known [49] that the parafermion algebra

W_{n + 1} (2 n / (n + 3))

processes a certain

Z_{n + 1}

symmetry among its irreducible modules. The automorphism

σ

is actually related to such a symmetry. More details about the relation between coset subalgebra

U

and the parafermion algebra

W_{n + 1} (2 n / (n + 3))

can be found in Appendix

B

Next we shall recall the definition of two conformal vectors of central charge

1 / 2

from [33] . Note that

\begin{matrix} \begin{matrix} \hat{e} & = \frac{1}{16} ω + \frac{1}{32} \sum_{α \in Φ^{+} (E_{8})} (e^{\sqrt{2} α} + e^{- \sqrt{2} α}) \end{matrix} \end{matrix}

(2.11)

is a conformal vector of central charge

1 / 2

. Let

σ

be the automorphism defined in Remark 2.2 . Then we have

Theorem 2.4 (cf. [33] ). Let

\hat{e}

be defined as above and

\hat{f} = σ \hat{e}

. Then

\hat{e}, \hat{f} \in U

and we have

\begin{matrix} 〈 \hat{e}, \hat{f} 〉 = {\begin{matrix} 1 / 4 & if i = 0, \\ 1 / 32 & if i = 1, \\ 13 / 2^{10} & if i = 2, \\ 1 / 2^{7} & if i = 3, \\ 3 / 2^{9} & if i = 4, \\ 5 / 2^{10} & if i = 5, \\ 1 / 2^{8} & if i = 6, \\ 0 & if i = 7, \\ 1 / 2^{8} & if i = 8 . \end{matrix} \end{matrix}

(2.12)

In other words, the values of

〈 \hat{e}, \hat{f} 〉

are exactly the values given in McKay's diagram 1.1 .

3 The coset subalgebra $U$ and Miyamoto's $τ$ -involutions

This section is the main part of this article. We shall study the structure of the coset subalgebra

U

defined by 2.10 for each of the nine cases. Except for the case of

4 A

, we shall show that the subalgebra

U

always contain a set of mutually orthogonal conformal vectors such that their sum is the Virasoro element of

U

and the central charges of these conformal vectors are all coming from the unitary series

\begin{matrix} c = c_{m} = 1 - \frac{6}{(m + 2) (m + 3)}, m = 1, 2, 3, . . . . \end{matrix}

(3.1)

Such a conformal vector generates a simple Virasoro VOA isomorphic to

L (c_{m}, 0)

inside

U

. The irreducible modules of

L (c_{m}, 0)

are of the form

L (c_{m}, h_{r, s}^{m})

, where

\begin{matrix} h_{r, s}^{m} = \frac{(r (m + 3) - s (m + 2))^{2} - 1}{4 (m + 2) (m + 3)}, 1 \leq r \leq m + 1, 1 \leq s \leq m + 2 . \end{matrix}

(3.2)

Note that

h_{r, s}^{m} = h_{m + 2 - r, m + 3 - s}^{m}

and that

L (c_{m}, h_{r, s}^{m})

1 \leq s \leq r \leq m + 1

are all the inequivalent irreducible

L (c_{m}, 0)

-modules.

For the

4 A

case, we shall show that

U

is isomorphic to the fixed point subalgebra

V_{N}^{+}

θ

for some rank two lattice

N

Furthermore, we shall discuss the relation between Miyamoto's

τ

-involutions and the structure of the coset subalgebra

U

. We shall show that for any VOA

V

which contains a subalgebra isomorphic to

U

, the product of the Miyamoto involutions

τ_{\hat{e}}

and

τ_{\hat{f}}

naturally defines an automorphism of order

n_{i}

n_{i} / 2

V

. If

U

is actually contained in the Moonshine VOA

V^{♮}

, then we shall show that

τ_{\hat{e}} τ_{\hat{f}}

is of the desired conjugacy class of the Monster simple group mentioned in the McKay diagram. The existence of

U

inside the Moonshine VOA

V^{♮}

will also be established for the cases

1 A, 2 A, 2 B, 3 A

, and

4 A

As in Section 2.2 ,

L = L (i), i = 0, 1, \dots, 8

denotes the lattice associated with the Dynkin diagram obtained by removing the

i

-th node

α_{i}

in the extended

E_{8}

diagram. The coset decomposition of

E_{8}

L

is given in 2.5 . For

j = 1, . . ., n_{i} - 1

, we define

\begin{matrix} X^{j} = \sum_{\binom{α \in j α_{i} + L}{〈 α, α 〉 = 2}} e^{\sqrt{2} α} . \end{matrix}

(3.3)

Clearly

X^{j}

is of weight

2

. By Proposition 2.1 , it is easy to see that

X^{j} \in U

for all

j = 1, . . ., n_{i} - 1

Recall the positive definite invariant hermitian form

(\cdot, \cdot)

on lattice VOAs mentioned in Introduction. We shall consider the form

(\cdot, \cdot)

for the lattice VOA

V_{\sqrt{2} E_{8}}

. Let

Y^{j, +} = \frac{1}{2} (X^{j} + X^{n_{i} - j}), Y^{j, -} = \frac{1}{2} \sqrt{- 1} (X^{j} - X^{n_{i} - j}), 1 \leq j \leq [n_{i} / 2]

and set

ℬ_{R} = s p a n_{R} {{\tilde{ω}}^{1}, \dots, {\tilde{ω}}^{l}, Y^{1, \pm}, \dots, Y^{[n_{i} / 2], \pm}}

. Then

ℬ_{R}

is contained in

{\tilde{V}}_{\sqrt{2} E_{8}, R}

and so the form

(\cdot, \cdot)

ℬ_{R}

-invariant. It is clear that the conformal vectors

\hat{e}

and

\hat{f}

definedby 2.11 are contained in

ℬ_{R}

. In the following argument we use the fact that

V_{\sqrt{2} E_{8}}

possesses a positive definite hermitian form which is

ℬ_{R}

-invariant.

Now let us study the structure of

U

and the Miyamoto involutions

τ_{\hat{e}}

and

τ_{\hat{f}}

associated with the conformal vectors

\hat{e}

and

\hat{f}

defined by 2.11 in each of the nine cases. First we shall note that

\begin{matrix} τ_{\hat{e}} τ_{\hat{f}} = e^{2 π \sqrt{- 1} β (0)} \end{matrix}

(3.4)

as an automorphism of

V_{\sqrt{2} E_{8}}

. (cf. [33,Section4] ).

3.1 $1 A$ case

In this case,

L = L (0) \sim = E_{8}

n_{0} = | E_{8} / L | = 1

, and

U \sim = L (1 / 2, 0)

. The conformal vector

{\tilde{ω}}^{1} \in V_{\sqrt{2} E_{8}}

defined by 2.9 is the only conformal vector in

U

. Its central charge is

1 / 2

. Moreover,

\hat{e} = \hat{f} = {\tilde{ω}}^{1}

and

U = V i r (\hat{e}) \subset V_{Λ}^{+} \subset V^{♮}

. Thus we have

τ_{\hat{e}} τ_{\hat{f}} = 1

3.2 $2 A$ case

In this case,

L = L (1) \sim = A_{1} \oplus E_{7}

n_{1} = | E_{8} / L | = 2

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{1}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} E_{7}}

defined by 2.9 are of central charge

1 / 2

and

7 / 10

, respectively.

Proposition 3.1. The vector

X^{1}

defined by 3.3 is a highest weight vector of highest weight

(1 / 2, 3 / 2)

with respect to

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

. Thus as a module of

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

U \sim = L (\frac{1}{2}, 0) \otimes L (\frac{7}{10}, 0) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{7}{10}, \frac{3}{2}) .

This VOA has been well studied in [24, 30] . In fact,

U

has exactly three conformal vectors of central charge 1/2, namely

\hat{e}

\hat{f}

and

w = {\tilde{ω}}^{1}

. The automorphism group

A u t U

U

is a symmetric group

S_{3}

of degree

3

. Note that

U

is generated by

\hat{e}

and

\hat{f}

and they are both fixed by

θ

. Thus,

U \subset V_{Λ}^{+} \subset V^{♮}

The VOA

U

is rational and it has exactly eight inequivalent irreducible modules

M^{j}, W^{j}, j \in {0, a, b, c} \sim = Z_{2} \times Z_{2}

. As

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

-modules, they are of the following form,

\begin{matrix} M^{0} & \sim = [0, 0] \oplus [\frac{1}{2}, \frac{3}{2}], & W^{0} & \sim = [0, \frac{3}{5}] \oplus [\frac{1}{2}, \frac{1}{10}], \end{matrix}

\begin{matrix} M^{a} & \sim = M^{b} \sim = [\frac{1}{16}, \frac{7}{16}], & W^{a} & \sim = W^{b} \sim = [\frac{1}{16}, \frac{3}{80}], \end{matrix}

\begin{matrix} M^{c} & \sim = [\frac{1}{2}, 0] \oplus [0, \frac{3}{2}], & W^{c} & \sim = [\frac{1}{2}, \frac{3}{5}] \oplus [0, \frac{1}{10}], \end{matrix}

\begin{matrix}  \end{matrix}

where

[h_{1}, h_{2}]

denotes

L (1 / 2, h_{1}) \otimes L (7 / 10, h_{2})

It is known that the fusion rules among irreducible

U

-modules have a symmetry of

Z_{2} \times Z_{2}

. For any VOA

V

containing a subalgebra isomorphic to

U

, there are three automorphisms of order

2

1

associated with

U

(cf. [30] ). They are given by

τ^{a} = {\begin{matrix} 1 & on U^{0}, U^{a}, \\ - 1 & on U^{b}, U^{c}, \end{matrix} τ^{b} = {\begin{matrix} 1 & on U^{0}, U^{b}, \\ - 1 & on U^{c}, U^{a}, \end{matrix} τ^{c} = {\begin{matrix} 1 & on U^{0}, U^{c}, \\ - 1 & on U^{a}, U^{b}, \end{matrix}

where

U^{j}

is the sum of all irreducible

U

-submodules of

V

which are isomorphic to either

M^{j}

W^{j}

for

j = 0, a, b, c

. Actually,

τ^{a} = τ_{w}, τ^{b} = τ_{\hat{e}}, τ^{c} = τ_{\hat{f}},

and we have

τ_{\hat{e}} τ_{\hat{f}} = τ^{b} τ^{c} = τ^{a} = τ_{w}

V

is the Moonshine VOA

V^{♮}

, then we have the following theorem.

Theorem 3.2. As automorphisms of

V^{♮}

τ_{\hat{e}} τ_{\hat{f}} = τ_{w}

and thus

τ_{\hat{e}} τ_{\hat{f}}

is of class

2 A

3.3 $3 A$ case

In this case,

L = L (2) \sim = A_{2} \oplus E_{6}

n_{2} = | E_{8} / L | = 3

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{2}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} E_{6}}

defined by 2.9 are of central charge

4 / 5

and

6 / 7

, respectively.

¿From [26, 31] , we know that

\begin{matrix} U^{1} & = {u \in V_{\sqrt{2} A_{2}} | (s^{1})_{1} u = 0} \sim = L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3), \end{matrix}

\begin{matrix} U^{2} & = {u \in V_{\sqrt{2} E_{6}} | (s^{2})_{1} u = 0} \sim = L (\frac{6}{7}, 0) \oplus L (\frac{6}{7}, 5) . \end{matrix}

\begin{matrix}  \end{matrix}

Hence,

U \supset U^{1} \otimes U^{2} \sim = (L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3)) \otimes (L (\frac{6}{7}, 0) \oplus L (\frac{6}{7}, 5)) .

Proposition 3.3. Both of the vectors

X^{1}

and

X^{2}

defined by 3.3 are highest weight vectors of highest weight

(2 / 3, 4 / 3)

with respect to

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

Proof. First, note that

X^{1}

and

X^{2}

are highest weight vectors of weight

2

with respect to the Virasoro element

ω^{'} = {\tilde{ω}}^{1} + {\tilde{ω}}^{2}

U

and that

\begin{matrix} {\tilde{ω}}^{1} & = \frac{1}{15} \sum_{α \in Φ^{+} (A_{2})} α (- 1)^{2} \cdot 1 + \frac{1}{5} \sum_{α \in Φ^{+} (A_{2})} (e^{\sqrt{2} α} + e^{- \sqrt{2} α}) & = \frac{2}{5} ω^{'} + \frac{1}{5} \sum_{α \in Φ^{+} (A_{2})} (e^{\sqrt{2} α} + e^{- \sqrt{2} α}) . \end{matrix}

Clearly,

({\tilde{ω}}^{1})_{3} X^{1} = ({\tilde{ω}}^{1})_{2} X^{1} = 0

and

\begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} & = {(\frac{2}{5} ω^{'} + \frac{1}{5} \sum_{α \in Φ^{+} (A_{2})} (e^{\sqrt{2} α} + e^{- \sqrt{2} α}))}_{1} X^{1} & = (\frac{2}{5} \times \frac{2}{3} + \frac{1}{5} \times 2) X^{1} = \frac{2}{3} X^{1} . \end{matrix}

Hence,

X^{1}

is a highest weight vector of highest weight

(2 / 3, 4 / 3)

with respect to

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

. Similarly,

X^{2}

is also a highest weight vector of highest weight

(2 / 3, 4 / 3)

with respect to

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

. □

Since

U^{1} \otimes U^{2}

and

L (4 / 5, 2 / 3)^{\pm} \otimes L (6 / 7, 4 / 3)^{\pm}

are the only irreducible modules of

U^{1} \otimes U^{2}

which have integral weights (cf. [26, 29, 31] ), by comparing dimensions of the homogeneous subspaces of small weights, we have the following proposition.

Proposition 3.4. As a module of

V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})

\begin{matrix} U \sim = & (L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3)) \otimes (L (\frac{6}{7}, 0) \oplus L (\frac{6}{7}, 5)) & \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3}) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3}) . \end{matrix}

By using Appendix A, we know that there are four distinct pairs of mutually orthogonal conformal vectors of central charge

4 / 5

and

6 / 7

, namely,

{{\tilde{ω}}^{1}, {\tilde{ω}}^{2}}

{x^{j}, y^{j}}

j = 0, 1, 2

, where

x^{j} = \frac{1}{16} {\tilde{ω}}^{1} + \frac{7}{8} {\tilde{ω}}^{2} - \frac{1}{48} ξ^{j} X^{1} - \frac{1}{48} ξ^{2 j} X^{2}, y^{j} = \frac{15}{16} {\tilde{ω}}^{1} + \frac{1}{8} {\tilde{ω}}^{2} + \frac{1}{48} ξ^{j} X^{1} + \frac{1}{48} ξ^{2 j} X^{2}

and

ξ = e^{2 π \sqrt{- 1} / 3}

is a primitive cubic root of unity.

Lemma 3.5. Let

u = 135 {\tilde{ω}}^{1} - 126 {\tilde{ω}}^{2} - 13 (X^{1} + X^{2})

and

v = X^{1} - X^{2}

. Then

u

and

v

are highest weight vectors of highest weight

(2 / 3, 4 / 3)

and

(13 / 8, 3 / 8)

with respect to

V i r (x^{0}) \otimes V i r (y^{0})

, respectively.

Proof. We have

\begin{matrix} (x^{0})_{1} X^{1} & = (\frac{1}{16} {\tilde{ω}}^{1} + \frac{7}{8} {\tilde{ω}}^{2} - \frac{1}{48} X^{1} - \frac{1}{48} X^{2})_{1} X^{1} & = (\frac{1}{16} \times \frac{2}{3} + \frac{7}{8} \times \frac{4}{3}) - \frac{20}{48} X^{2} - \frac{1}{48} (135 {\tilde{ω}}^{1} + 252 {\tilde{ω}}^{2}) & = \frac{29}{24} X^{1} - \frac{5}{12} X^{2} - \frac{1}{16} (45 {\tilde{ω}}^{1} + 84 {\tilde{ω}}^{2}) . \end{matrix}

Similarly,

\begin{matrix} (x^{0})_{1} X^{2} & = \frac{29}{24} X^{2} - \frac{5}{12} X^{1} - \frac{1}{16} (45 {\tilde{ω}}^{1} + 84 {\tilde{ω}}^{2}) . \end{matrix}

Thus

(x^{0})_{1} v = \frac{13}{8} v

and hence

v

is a highest weight vector of highest weight

(13 / 8, 3 / 8)

Furthermore,

\begin{matrix} (x^{0})_{1} u = & 135 (\frac{1}{8} {\tilde{ω}}^{1} - \frac{1}{48} \times \frac{2}{3} (X^{1} + X^{2})) - 126 (\frac{7}{4} {\tilde{ω}}^{2} - \frac{1}{48} \times \frac{4}{3} (X^{1} + X^{2})) & - 13 (\frac{19}{24} (X^{1} + X^{2}) - \frac{1}{8} (45 {\tilde{ω}}^{1} + 84 {\tilde{ω}}^{2})) & = & \frac{2}{3} (135 {\tilde{ω}}^{1} - 126 {\tilde{ω}}^{2} - 13 (X^{1} + X^{2})) = \frac{2}{3} u . \end{matrix}

Thus

u

is a highest weight vector of highest weight

(2 / 3, 4 / 3)

. □

Proposition 3.6. As a module of

V i r (x^{0}) \otimes V i r (y^{0})

\begin{matrix} U \sim = & L (\frac{4}{5}, 0) \otimes L (\frac{6}{7}, 0) \oplus L (\frac{4}{5}, 3) \otimes L (\frac{6}{7}, 5) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3}) & \oplus L (\frac{4}{5}, \frac{13}{8}) \otimes L (\frac{6}{7}, \frac{3}{8}) \oplus L (\frac{4}{5}, \frac{1}{8}) \otimes L (\frac{6}{7}, \frac{23}{8}) . \end{matrix}

Proof. The fixed point subalgebra $U^{+}$ of $θ$ contains $V i r (x^{0}) \otimes V i r (y^{0})$ . Moreover, $u \in U^{+}$ and so $U^{+}$ contains a submodule of the form $L (4 / 5, 2 / 3) \otimes L (6 / 7, 4 / 3)$ by Lemma 3.5 .Hence comparing the first several terms of the characters, we know that $U^{+} \sim = L (\frac{4}{5}, 0) \otimes L (\frac{6}{7}, 0) \oplus L (\frac{4}{5}, 3) \otimes L (\frac{6}{7}, 5) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3})$ as a module of $V i r (x^{0}) \otimes V i r (y^{0})$ . Note that $U^{+}$ has the same form as a module of $V i r ({\tilde{ω}}^{1}) \otimes V i r ({\tilde{ω}}^{2})$ . Lemma 3.5 also implies that $U^{-}$ contains a submodule of the form $L (4 / 5, 13 / 8) \otimes L (6 / 7, 3 / 8)$ since $v \in U^{-}$ . Thus $U^{-} \sim = L (\frac{4}{5}, \frac{13}{8}) \otimes L (\frac{6}{7}, \frac{3}{8}) \oplus L (\frac{4}{5}, \frac{1}{8}) \otimes L (\frac{6}{7}, \frac{23}{8})$ and we have the desired result. □

The VOA

U

has been constructed and studied by Sakuma and Yamauchi [44] (see also Miyamoto [42] ). It is known that the automorphism group

A u t U

U

is isomorphic to the symmetric group

S_{3}

and

U

is generated by two conformal vectors of central charge 1/2, namely,

\hat{e}

and

\hat{f}

There are exactly six irreducible modules of

U

(cf. [44] ), namely,

\begin{matrix} M^{0} \otimes A^{0} \oplus M^{1} \otimes A^{1} \oplus M^{2} \otimes A^{2}, & M^{0} \otimes B^{0} \oplus M^{1} \otimes B^{1} \oplus M^{2} \otimes B^{2}, \end{matrix}

\begin{matrix} M^{0} \otimes C^{0} \oplus M^{1} \otimes C^{1} \oplus M^{2} \otimes C^{2}, & W^{0} \otimes A^{0} \oplus W^{1} \otimes A^{1} \oplus W^{2} \otimes A^{2}, \end{matrix}

\begin{matrix} W^{0} \otimes B^{0} \oplus W^{1} \otimes B^{1} \oplus W^{2} \otimes B^{2}, & W^{0} \otimes C^{0} \oplus W^{1} \otimes C^{1} \oplus W^{2} \otimes C^{2}, \end{matrix}

\begin{matrix}  \end{matrix}

where

\begin{matrix} M^{0} \sim = L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3), & M^{1} \sim = L (\frac{4}{5}, \frac{2}{3})^{+}, & M^{2} \sim = L (\frac{4}{5}, \frac{2}{3})^{-}, \end{matrix}

\begin{matrix} W^{0} \sim = L (\frac{4}{5}, \frac{2}{5}) \oplus L (\frac{4}{5}, \frac{7}{5}), & W^{1} \sim = L (\frac{4}{5}, \frac{1}{15})^{+}, & W^{2} \sim = L (\frac{4}{5}, \frac{1}{15})^{-} \end{matrix}

\begin{matrix}  \end{matrix}

are the irreducible modules of

L (4 / 5, 0) \oplus L (4 / 5, 3)

and

\begin{matrix} A^{0} \sim = L (\frac{6}{7}, 0) \oplus L (\frac{6}{7}, 5), & A^{1} \sim = L (\frac{6}{7}, \frac{4}{3})^{+}, & A^{2} \sim = L (\frac{6}{7}, \frac{4}{3})^{-}, \end{matrix}

\begin{matrix} B^{0} \sim = L (\frac{6}{7}, \frac{1}{7}) \oplus L (\frac{6}{7}, \frac{22}{7}), & B^{1} \sim = L (\frac{6}{7}, \frac{10}{21})^{+}, & B^{2} \sim = L (\frac{6}{7}, \frac{10}{21})^{-}, \end{matrix}

\begin{matrix} C^{0} \sim = L (\frac{6}{7}, \frac{5}{7}) \oplus L (\frac{6}{7}, \frac{12}{7}), & C^{1} \sim = L (\frac{6}{7}, \frac{1}{21})^{+}, & C^{2} \sim = L (\frac{6}{7}, \frac{1}{21})^{-} \end{matrix}

\begin{matrix}  \end{matrix}

are the irreducible modules of

L (6 / 7, 0) \oplus L (6 / 7, 5)

. Both of

L (4 / 5, 0) \oplus L (4 / 5, 3)

and

L (6 / 7, 0) \oplus L (6 / 7, 5)

are rational VOAs and the fusion rules among their irreducible modules have been determined in [41] and [31] . There are two

Z_{3}

-symmetries given as follows.

τ_{1} = {\begin{matrix} 1 & on M^{0}, W^{0}, \\ e^{2 π \sqrt{- 1} / 3} & on M^{1}, W^{1}, \\ e^{4 π \sqrt{- 1} / 3} & on M^{2}, W^{2}, \end{matrix} τ_{2} = {\begin{matrix} 1 & on A^{0}, B^{0}, C^{0}, \\ e^{2 π \sqrt{- 1} / 3} & on A^{1}, B^{1}, C^{1}, \\ e^{4 π \sqrt{- 1} / 3} & on A^{2}, B^{2}, C^{2} . \end{matrix}

V

is a VOA which contains

U

as a subalgebra, then both

τ_{1}

and

τ_{2}

induce automorphisms of

V

. Moreover, as automorphisms of

V

τ_{1} = τ_{2}

and we have that

τ_{\hat{e}} τ_{\hat{f}} = τ_{1}

is of order

3

Remark 3.7. Recall that

\hat{e}

and

\hat{f}

are fixed by the Weyl group

W (Φ) = W (A_{2}) \times W (E_{6})

of the root system

Φ = A_{2} \oplus E_{6}

L

. Since

U

is generated by

\hat{e}

and

\hat{f}

W (Φ)

leaves every element of

U

invariant. Let

g

be an element of order

3

W (Φ)

which acts fixed-point-freely on

\sqrt{2} E_{8}

. Then

g

induces a fixed-point-free action on the Leech lattice

Λ

also (cf. [25] ). Since every element of

U

is fixed by

g

U

is actually contained in the Moonshine VOA

V^{♮}

by the

Z_{3}

-orbifold construction of

V^{♮}

given by Dong and Mason [11] .

Thus,

τ_{\hat{e}} τ_{\hat{f}} = τ_{1}

is of class

3 A

(cf. [25, 41] ).

In [42] , Miyamoto showed that for any two conformal vectors

e

and

f

of central charge

1 / 2

with

〈 e, f 〉 = 13 / 2^{10}

in the Moonshine VOA

V^{♮}

, the vertex subalgebra

W

generated by

e

and

f

must contain a subalgebra of the form

L (4 / 5, 0) \otimes L (6 / 7, 0)

. In fact, Sakuma and Yamauchi [44] showed that the algebra

W

must be isomorphic to

U

. This gives a proof that

U

is contained in

V^{♮}

3.4 $4 A$ case

In this case,

L = L (3) = A_{3} \oplus D_{5}

n_{3} = | E_{8} / L | = 4

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{3}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} D_{5}}

defined by 2.9 are both of central charge

1

. Let

ε_{1}, . . ., ε_{8} \in R^{8}

be such that

〈 ε_{i}, ε_{j} 〉 = 2 δ_{i j}

for any

i, j \in {1, \dots, 8}

. Then

\sqrt{2} E_{8} = {\sum_{i = 1}^{8} a_{i} ε_{i} | \binom{all a_{i} \in Z or all a_{i} \in \frac{1}{2} + Z}{with \sum_{i = 1}^{8} a_{i} \equiv 0 m o d 2}} .

Let

K = {\sum_{i = 1}^{8} a_{i} ε_{i} | all a_{i} \in Z or all a_{i} \in \frac{1}{2} + Z} .

Then

\sqrt{2} E_{8} \subset K

and

| K / \sqrt{2} E_{8} | = 2

Now we consider two automorphisms

θ

and

ψ

V_{K}

defined by

θ (α (- 1) \cdot 1) = - α (- 1) \cdot 1 and θ (e^{α}) = e^{- α}, ψ (u \otimes e^{β}) = (- 1)^{〈 ε_{1} + \dots + ε_{8}, β 〉 / 2} u \otimes e^{β}

for any

α, β \in K

and

u \in M (1)

. Note that

V_{\sqrt{2} E_{8}} = (V_{K})^{ψ}

. It is well known (cf. [17,Chapter10] ) that

θ

and

ψ

are conjugate in

A u t V_{K}

. Thus, we have

V_{\sqrt{2} E_{8}} = (V_{K})^{ψ} \sim = V_{K}^{+} .

By using the same argument as in Dong et al. [8, 9] , one can show that the VOA

V_{\sqrt{2} A_{3}}

contains a subalgebra isomorphic to

V_{\sqrt{2} A_{2}}^{+} \otimes V_{Z γ_{3}}^{+}

and the VOA

V_{\sqrt{2} D_{5}}

contains a subalgebra isomorphic to

V_{\sqrt{2} A_{4}}^{+} \otimes V_{Z γ_{5}}^{+}

where

\sqrt{2} A_{2} \sim = s p a n_{Z} {- ε_{1} + ε_{2}, - ε_{2} + ε_{3}}

γ_{3} = ε_{1} + ε_{2} + ε_{3}

\sqrt{2} A_{4} \sim = s p a n_{Z} {- ε_{4} + ε_{5}, - ε_{5} + ε_{6}, - ε_{6} + ε_{7}, - ε_{7} + ε_{8}}

and

γ_{5} = ε_{4} + ε_{5} + ε_{6} + ε_{7} + ε_{8}

Moreover,

U

contains a subalgebra isomorphic to

V_{Z γ_{3}}^{+} \otimes V_{Z γ_{5}}^{+}

. Define

N = {x \in K | 〈 x, - ε_{j} + ε_{j + 1} 〉 = 0 for j = 1, 2, 4, 5, 6, 7} .

Then

V_{N} = {v \in V_{K} | u_{n} v = 0 for any u \in V_{\sqrt{2} A_{2}} or V_{\sqrt{2} A_{4}} and n \geq 0} .

Since

U = (V_{N})^{ψ}

, it is now easy to see that

U \sim = V_{N}^{+}

. Note that

N = {\sum_{i = 1}^{8} a_{i} ε_{i} | \binom{all a_{i} \in Z or all a_{i} \in \frac{1}{2} + Z with}{a_{1} + a_{2} + a_{3} = a_{4} + a_{5} + a_{6} + a_{7} + a_{8} = 0}} .

It is of rank two and generated by the elements

ξ_{1} = \frac{1}{2} (γ_{3} + γ_{5}), ξ_{2} = \frac{1}{2} (- γ_{3} + γ_{5}) .

Proposition 3.8. The VOA

V_{N}^{+}

is generated by its weight

2

subspace

(V_{N}^{+})_{_{2}}

Proof. Recall that

V_{N}^{+}

contains a subalgebra isomorphic to

V_{Z γ_{3}}^{+} \otimes V_{Z γ_{5}}^{+}

. Moreover,

\begin{matrix} V_{N}^{+} = & (V_{Z γ_{3}}^{+} \otimes V_{Z γ_{5}}^{+}) \oplus (V_{Z γ_{3}}^{-} \otimes V_{Z γ_{5}}^{-}) & \oplus (V_{γ_{3} / 2 + Z γ_{3}}^{+} \otimes V_{γ_{5} / 2 + Z γ_{5}}^{+}) \oplus (V_{γ_{3} / 2 + Z γ_{3}}^{-} \otimes V_{γ_{5} / 2 + Z γ_{5}}^{-}) \end{matrix}

Note that

(V_{N}^{+})_{_{2}}

is of dimension

5

and it has a basis consisting of the following five elements.

β_{1} (- 1)^{2} \cdot 1, β_{2} (- 1)^{2} \cdot 1, β_{1} (- 1) β_{2} (- 1) \cdot 1, e^{ξ_{1}} + e^{- ξ_{1}}, e^{ξ_{2}} + e^{- ξ_{2}},

where

β_{1} = \frac{1}{\sqrt{6}} γ_{3}

and

β_{2} = \frac{1}{\sqrt{10}} γ_{5}

. Let

W

be the subalgebra generated by

(V_{N}^{+})_{_{2}}

. We want to show that

W = V_{N}^{+}

. Since the minimal weights of

V_{Z γ_{3}}^{-} \otimes V_{Z γ_{5}}^{-}

V_{γ_{3} / 2 + Z γ_{3}}^{+} \otimes V_{γ_{5} / 2 + Z γ_{5}}^{+}

, and

V_{γ_{3} / 2 + Z γ_{3}}^{-} \otimes V_{γ_{5} / 2 + Z γ_{5}}^{-}

are all equal to

2

and since they are irreducible

V_{Z γ_{3}}^{+} \otimes V_{Z γ_{5}}^{+}

-modules, it suffices to show that

V_{Z γ_{3}}^{+} \otimes V_{Z γ_{5}}^{+} \subset W

For any rank one even lattice

Z α

with

〈 α, α 〉 = 2 k

, it was shown in [5] that the VOA

V_{Z α}^{+}

is generated by three elements

\begin{matrix} \begin{matrix} ω = & \frac{1}{2} β (- 1)^{2} \cdot 1, & J = & β (- 1)^{4} \cdot 1 - 2 β (- 3) β (- 1) \cdot 1 + \frac{3}{2} β (- 2)^{2} \cdot 1, & E = & e^{α} + e^{- α}, \end{matrix} \end{matrix}

(3.5)

where

β = \frac{1}{\sqrt{2 k}} α

Let

J_{1}

and

J_{2}

be the elements obtained by replacing

β

with

β_{1}

and

β_{2}

in the element

J

of 3.5 , respectively. Likewise, let

E_{1} = e^{γ_{3}} + e^{- γ_{3}}

and

E_{2} = e^{γ_{5}} + e^{- γ_{5}}

. We only need to show that

W

contains

J_{1}, J_{2}, E_{1}

, and

E_{2}

By direct computation, we have

\begin{matrix} (β_{1} (- 1) β_{2} (- 1) \cdot 1)_{- 1} (β_{1} (- 1) β_{2} (- 1) \cdot 1) & = β_{1} (- 3) β_{1} (- 1) \cdot 1 + β_{2} (- 3) β_{2} (- 1) \cdot 1 + β_{1} (- 1)^{2} β_{2} (- 1)^{2} \cdot 1 \end{matrix}

and

(β_{1} (- 1)^{2} \cdot 1)_{- 1} (β_{2} (- 1)^{2} \cdot 1) = β_{1} (- 1)^{2} β_{2} (- 1)^{2} \cdot 1 .

Thus,

β_{1} (- 3) β_{1} (- 1) \cdot 1 + β_{2} (- 3) β_{2} (- 1) \cdot 1 \in W

and

β_{1} (- 3) β_{1} (- 1) \cdot 1 = \frac{1}{8} (β_{1} (- 1)^{2} \cdot 1)_{1} (β_{1} (- 3) β_{1} (- 1) \cdot 1 + β_{2} (- 3) β_{2} (- 1) \cdot 1)

is also contained in

W

. Moreover,

(β_{1} (- 1)^{2} \cdot 1)_{- 1} (β_{1} (- 1)^{2} \cdot 1) = β_{1} (- 1)^{4} \cdot 1 + 4 β_{1} (- 3) β_{1} (- 1) \cdot 1, ((β_{1} (- 1)^{2} \cdot 1)_{0})^{2} (β_{1} (- 1)^{2} \cdot 1) = 16 β_{1} (- 3) β_{1} (- 1) \cdot 1 + 8 β_{1} (- 2)^{2} \cdot 1 .

Hence, we have

β_{1} (- 1)^{4} \cdot 1, β_{1} (- 2)^{2} \cdot 1 \in W

and thus

J_{1} \in W

. Similarly,

J_{2} \in W

. Clearly,

W

contains

E_{1} = e^{γ_{3}} + e^{- γ_{3}} = (e^{ξ_{1}} + e^{- ξ_{1}})_{0} (e^{ξ_{2}} + e^{- ξ_{2}}) .

Furthermore,

W

contains the following three elements.

\begin{matrix} (γ_{3} (- 1)^{2} \cdot 1)_{- 1} E_{1} & = γ_{3} (- 1)^{2} E_{1} + 12 γ_{3} (- 2) (e^{γ_{3}} - e^{- γ_{3}}), \end{matrix}

\begin{matrix} (γ_{5} (- 1)^{2} \cdot 1)_{- 1} E_{1} & = γ_{5} (- 1)^{2} E_{1}, \end{matrix}

\begin{matrix} (γ_{3} (- 1) γ_{5} (- 1) \cdot 1)_{- 1} E_{1} & = γ_{3} (- 1) γ_{5} (- 1) E_{1} + 6 γ_{5} (- 2) (e^{γ_{3}} - e^{- γ_{3}}) . \end{matrix}

\begin{matrix}  \end{matrix}

Then

W

also contains

\begin{matrix} (γ_{3} (- 1) γ_{5} (- 1) \cdot 1)_{1} (γ_{3} (- 1) γ_{5} (- 1) E_{1} + 6 γ_{5} (- 2) (e^{γ_{3}} - e^{- γ_{3}})) & = 10 (γ_{3} (- 1)^{2} E_{1} + 12 γ_{3} (- 2) (e^{γ_{3}} - e^{- γ_{3}})) + 6 γ_{5} (- 1)^{2} E_{1} + 6 γ_{5} (- 2) (e^{γ_{3}} - e^{- γ_{3}}) . \end{matrix}

Hence, we have

γ_{5} (- 2) (e^{γ_{3}} - e^{- γ_{3}}) \in W

and so

γ_{3} (- 2) (e^{γ_{3}} - e^{- γ_{3}}) = \frac{1}{20} (γ_{3} (- 1) γ_{5} (- 1) \cdot 1)_{1} (γ_{5} (- 2) (e^{γ_{3}} - e^{- γ_{3}}))

is also contained in

W

. Therefore,

γ_{3} (- 1)^{2} E_{1}

γ_{5} (- 1)^{2} E_{1}

, and

γ_{5} (- 1) γ_{3} (- 1) E_{1}

are contained in

W

Finally,

W

contains

\begin{matrix} (e^{ξ_{1}} + e^{- ξ_{1}})_{- 2} (e^{ξ_{2}} + e^{- ξ_{2}}) & = E_{2} + \frac{1}{2} ξ_{1} (- 1)^{2} E_{1} + \frac{1}{2} ξ_{1} (- 2) (e^{γ_{3}} - e^{- γ_{3}}) & = E_{2} + \frac{1}{8} (γ_{3} + γ_{5}) (- 1)^{2} E_{1} + \frac{1}{4} (γ_{3} + γ_{5}) (- 2) (e^{γ_{3}} - e^{- γ_{3}}) \end{matrix}

and thus

E_{2} \in W

. □

Theorem 3.9.

(1)

The Griess algebra

U_{2}

U

is generated by

\hat{e}

and

\hat{f}

(2)

The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

Proof. We only need to show the first assertion. Note that

U_{2} = s p a n_{C} {{\tilde{ω}}^{1}, {\tilde{ω}}^{2}, X^{1}, X^{2}, X^{3}}

is of dimension

5

and that

\begin{matrix} \hat{e} & = {\tilde{ω}}^{1} + {\tilde{ω}}^{2} + X^{1} + X^{2} + X^{3}, \end{matrix}

\begin{matrix} \hat{f} & = {\tilde{ω}}^{1} + {\tilde{ω}}^{2} + \sqrt{- 1} X^{1} - X^{2} - \sqrt{- 1} X^{3} . \end{matrix}

\begin{matrix}  \end{matrix}

Let

G

be the Griess subalgebra generated by

\hat{e}

and

\hat{f}

. Then

{\hat{e}}_{1} \hat{f}

{\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})

, and

{\hat{f}}_{1} ({\hat{e}}_{1} \hat{f})

are also in

G

. By direct computation, it is easy to see that

\hat{e}

\hat{f}

{\hat{e}}_{1} \hat{f}

{\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})

, and

{\hat{f}}_{1} ({\hat{e}}_{1} \hat{f})

are linearly independent. Thus

G = U_{2}

. □

Theorem 3.10. The automorphism group

A u t U

U

is a dihedral group of order

8

Proof. There are exactly four conformal vectors of central $1 / 2$ in $U_{2}$ , namely, $e_{j} = σ^{j} \hat{e} = \frac{3}{16} {\tilde{ω}}^{1} + \frac{5}{16} {\tilde{ω}}^{2} + \frac{1}{32} ((\sqrt{- 1})^{j} X^{1} + (- 1)^{j} X^{2} + (- \sqrt{- 1})^{j} X^{3}), 0 \leq j \leq 3 .$ Since $U$ is generated by $\hat{e}$ and $\hat{f}$ , we can consider $A u t U$ as a subgroup of the permutation group on the set ${e_{0}, e_{1}, e_{2}, e_{3}}$ . Now let $g \in A u t U$ . Then $g$ also preserves the inner product and thus $〈 g e_{i}, g e_{j} 〉 = 〈 e_{i}, e_{j} 〉 = {\begin{matrix} 1 / 2^{7} & if i + j is odd, \\ 0 & if i + j is even . \end{matrix}$ Hence, $g$ will either keep ${e_{0}, e_{2}}$ and ${e_{1}, e_{3}}$ invariant or $g$ will map ${e_{0}, e_{2}}$ to ${e_{1}, e_{3}}$ .
Thus $| A u t U | \leq 8$ . On the other hand, $θ$ and $σ$ defined in Remark 2.2 generate a subgroup isomorphic to a dihedral group of order $8$ inside $A u t U$ . Hence the assertion holds. □

Note that

〈 ξ_{1}, ξ_{1} 〉 = 〈 ξ_{2}, ξ_{2} 〉 = 4

〈 ξ_{1}, ξ_{2} 〉 = - 1

and the Leech lattice

Λ

does contain some sublattice isomorphic to

N

. Therefore,

U \subset V_{Λ}^{+} \subset V^{♮}

In [1] , the fusion rules for

V_{Z γ_{3}}^{+}

and

V_{Z γ_{5}}^{+}

are determined. It is known that there are

Z_{4}

-symmetries among the irreducible modules of

V_{Z γ_{3}}^{+}

and also among the irreducible modules of

V_{Z γ_{5}}^{+}

. By direct computation, it is easy to verify that the automorphism

τ_{\hat{e}} τ_{\hat{f}}

agrees with the

Z_{4}

-symmetries of

V_{Z γ_{3}}^{+}

and

V_{Z γ_{5}}^{+}

and thus

τ_{\hat{e}} τ_{\hat{f}}

is of class

4 A

(cf. [37] ). An explicit construction of

4 A

-elements as automorphisms of

V^{♮}

has already been obtained by Shimakura [45] .

3.5 $5 A$ case.

In this case,

L = L (4) \sim = A_{4} \oplus A_{4}

n_{4} = | E_{8} / L | = 5

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{4}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{4}}

defined by 2.9 are both of central charge

8 / 7

Lemma 3.11 (Lemma A.7). The coset subalgebra

U

contains a set of three mutually orthogonal conformal vectors of central charge

1 / 2, 25 / 28

, and

25 / 28

, respectively, namely,

\begin{matrix} u = \hat{e} = \frac{7}{32} ({\tilde{ω}}^{1} + {\tilde{ω}}^{2}) + \frac{1}{32} (X^{1} + X^{2} + X^{3} + X^{4}), \end{matrix}

\begin{matrix} v = \frac{15}{64} {\tilde{ω}}^{1} + \frac{35}{64} {\tilde{ω}}^{2} - \frac{3}{64} (X^{1} + X^{4}) + \frac{1}{64} (X^{2} + X^{3}), \end{matrix}

\begin{matrix} w = \frac{35}{64} {\tilde{ω}}^{1} + \frac{15}{64} {\tilde{ω}}^{2} + \frac{1}{64} (X^{1} + X^{4}) - \frac{3}{64} (X^{2} + X^{3}) . \end{matrix}

\begin{matrix}  \end{matrix}

By the above lemma,

U

contains

V i r (u) \otimes V i r (v) \otimes V i r (w) \sim = L (1 / 2, 0) \otimes L (25 / 28, 0) \otimes L (25 / 28, 0)

. All irreducible modules of

V i r (u) \otimes V i r (v) \otimes V i r (w)

are known (cf. [13, 48] ).

They are of the form

L (1 / 2, h_{1}) \otimes L (25 / 28, h_{2}) \otimes L (25 / 28, h_{3})

. Among them, the highest weights

(h_{1}, h_{2}, h_{3})

of the irreducible modules which have integral weights are as follows.

\begin{matrix} (0, 0, 0), (\frac{1}{16}, \frac{5}{32}, \frac{57}{32}), (\frac{1}{16}, \frac{57}{32}, \frac{5}{32}), (\frac{1}{2}, \frac{3}{4}, \frac{3}{4}), (0, \frac{3}{4}, \frac{13}{4}), (0, \frac{13}{4}, \frac{3}{4}), \end{matrix}

\begin{matrix} (\frac{1}{16}, \frac{57}{32}, \frac{165}{32}), (\frac{1}{16}, \frac{165}{32}, \frac{57}{32}), (\frac{1}{2}, \frac{13}{4}, \frac{13}{4}), (\frac{1}{2}, 0, \frac{15}{2}), (\frac{1}{2}, \frac{15}{2}, 0), (0, \frac{15}{2}, \frac{15}{2}) . \end{matrix}

\begin{matrix}  \end{matrix}

The following lemma can be proved by direct computation.

Lemma 3.12. Let

\begin{matrix} a & = ({\tilde{ω}}^{1} - {\tilde{ω}}^{2}) - \frac{1}{35} (X^{1} + X^{4}) + \frac{1}{35} (X^{2} + X^{3}), \end{matrix}

\begin{matrix} b^{1} & = 3 (X^{1} - X^{4}) + 2 (X^{2} - X^{3}), \end{matrix}

\begin{matrix} b^{2} & = 2 (X^{1} - X^{4}) - 3 (X^{2} - X^{3}) . \end{matrix}

\begin{matrix}  \end{matrix}

Then

a

b^{1}

and

b^{2}

are highest weight vectors of highest weight

(\frac{1}{2}, \frac{3}{4}, \frac{3}{4})

(\frac{1}{16}, \frac{5}{32}, \frac{57}{32})

and

(\frac{1}{16}, \frac{57}{32}, \frac{5}{32})

with respect to

V i r (u) \otimes V i r (v) \otimes V i r (w)

, respectively.

We denote the irreducible module

L (1 / 2, h_{1}) \otimes L (25 / 28, h_{2}) \otimes L (25 / 28, h_{3})

[h_{1}, h_{2}, h_{3}]

for simplicity of notation. By using the theory of characters (cf. Appendix B), we actually have the following decomposition of

U

into a direct sum of

[h_{1}, h_{2}, h_{3}]

's.

Theorem 3.13 (Theorem B.7 ). As a module of

V i r (u) \otimes V i r (v) \otimes V i r (w)

\begin{matrix} U \sim = & [0, 0, 0] \oplus [\frac{1}{16}, \frac{5}{32}, \frac{57}{32}] \oplus [\frac{1}{16}, \frac{57}{32}, \frac{5}{32}] \oplus [\frac{1}{2}, \frac{3}{4}, \frac{3}{4}] \end{matrix}

\begin{matrix} \oplus [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] \oplus [\frac{1}{16}, \frac{57}{32}, \frac{165}{32}] \oplus [\frac{1}{16}, \frac{165}{32}, \frac{57}{32}] \end{matrix}

\begin{matrix} \oplus [\frac{1}{2}, \frac{13}{4}, \frac{13}{4}] \oplus [\frac{1}{2}, 0, \frac{15}{2}] \oplus [\frac{1}{2}, \frac{15}{2}, 0] \oplus [0, \frac{15}{2}, \frac{15}{2}] . \end{matrix}

\begin{matrix}  \end{matrix}

Next, we shall discuss the generators of

U

Theorem 3.14. The coset subalgebra

U

is generated by its weight

2

subspace

U_{2}

We shall divide the proof into several steps. By direct computation, we can verify the following lemma.

Lemma 3.15. Let

\begin{matrix} y^{1} & = a_{- 1} a - \frac{192}{343} u_{- 3} \cdot 1 - \frac{128}{343} u_{- 1} u - \frac{192}{455} v_{- 3} \cdot 1 - \frac{128}{325} v_{- 1} v - \frac{192}{455} w_{- 3} \cdot 1 - \frac{128}{325} w_{- 1} w, \end{matrix}

\begin{matrix} y^{2} & = (b^{1})_{- 1} b^{2} + 105 u_{- 1} a - \frac{735}{2} v_{- 1} a + \frac{5355}{8} v_{0} v_{0} a - \frac{735}{2} w_{- 1} a + \frac{1715}{8} w_{0} w_{0} a \end{matrix}

\begin{matrix} + \frac{665}{2} u_{0} v_{0} a - \frac{245}{2} u_{0} w_{0} a - \frac{4655}{12} v_{0} w_{0} a . \end{matrix}

\begin{matrix}  \end{matrix}

Then

y^{1}

and

y^{2}

are non-zero singular vectors for

V i r (u) \otimes V i r (v) \otimes V i r (w)

, that is,

u_{n} y^{j} = v_{n} y^{j} = w_{n} y^{j} = 0

j = 1, 2

, for any

n \geq 2

Now let

W

be the subalgebra of

U

generated by

U_{2}

Lemma 3.16. There are highest weight vectors of highest weight

(0, \frac{3}{4}, \frac{13}{4})

and

(0, \frac{13}{4}, \frac{3}{4})

with respect to

V i r (u) \otimes V i r (v) \otimes V i r (w)

W

Proof. By fusion rules, we know that
$a_{- 1} a \in [0, 0, 0]_{4} \oplus [0, \frac{3}{4}, \frac{13}{4}]_{4} \oplus [0, \frac{13}{4}, \frac{3}{4}]_{4}, (b^{1})_{- 1} b^{2} \in [\frac{1}{2}, \frac{3}{4}, \frac{3}{4}]_{4} \oplus [0, \frac{3}{4}, \frac{13}{4}]_{4} \oplus [0, \frac{13}{4}, \frac{3}{4}]_{4} .$
Since both $y^{1}$ and $y^{2}$ are singular vectors of weight $4$ , we have $y^{1}, y^{2} \in [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] .$ Moreover, by direct computation, one can show that $〈 y^{1}, y^{2} 〉 = 〈 a_{- 1} a, (b^{1})_{- 1} b^{2} 〉 = 0$ .
Note that $a, b^{1}$ , and $b^{2}$ are highest weight vectors and $〈 a, b^{1} 〉 = 〈 a, b^{2} 〉 = 〈 b^{1}, b^{2} 〉 = 0$ .
Hence $y^{1}$ and $y^{2}$ are linearly independent. Thus the assertion holds. □

Proof of Theorem 3.14 . First, we note that the coset subalgebra

U

is simple. Then the subalgebra

U^{τ_{u}}

consisting of the fixed points of the Miyamoto involution

τ_{u}

associated with

u

U

is also simple.

As a module of

V i r (u) \otimes V i r (v) \otimes V i r (w)

\begin{matrix} U^{τ_{u}} \sim = & [0, 0, 0] \oplus [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] \oplus [0, \frac{15}{2}, \frac{15}{2}] \oplus [\frac{1}{2}, \frac{3}{4}, \frac{3}{4}] \end{matrix}

\begin{matrix} \oplus [\frac{1}{2}, \frac{13}{4}, \frac{13}{4}] \oplus [\frac{1}{2}, 0, \frac{15}{2}] \oplus [\frac{1}{2}, \frac{15}{2}, 0] . \end{matrix}

\begin{matrix}  \end{matrix}

Moreover, we can define an automorphism

σ_{u}

U^{τ_{u}}

(cf. Miyamoto [39] ) by

σ_{u} = {\begin{matrix} 1 & on L (\frac{1}{2}, 0), \\ - 1 & on L (\frac{1}{2}, \frac{1}{2}) . \end{matrix}

Then the subalgebra

(U^{τ_{u}})^{σ_{u}}

consisting of the fixed points of

σ_{u}

U^{τ_{u}}

is again simple.

As a module of

V i r (u) \otimes V i r (v) \otimes V i r (w)

(U^{τ_{u}})^{σ_{u}} \sim = [0, 0, 0] \oplus [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] \oplus [0, \frac{15}{2}, \frac{15}{2}] .

By the previous lemma, we know that

W

contains

[0, 0, 0]

[0, 3 / 4, 13 / 4]

[0, 13 / 4, 3 / 4]

[1 / 2, 3 / 4, 3 / 4]

[1 / 16, 5 / 32, 57 / 32]

and

[1 / 16, 57 / 32, 5 / 32]

. Hence,

(W^{τ_{u}})^{σ_{u}}

must contain an irreducible module isomorphic to

[0, 15 / 2, 15 / 2]

; otherwise

(W^{τ_{u}})^{σ_{u}} \sim = [0, 0, 0] \oplus [0, 3 / 4, 13 / 4] \oplus [0, 13 / 4, 3 / 4]

and the orthogonal complement of

(W^{τ_{u}})^{σ_{u}}

(U^{τ_{u}})^{σ_{u}}

with respect to a positive definite invariant hermitian form is isomorphic to

[0, 15 / 2, 15 / 2]

. However, the orthogonal complement is a module for

(W^{τ_{u}})^{σ_{u}}

because of the invariance of the form, which is impossible by the fusion rules. Thus

(W^{τ_{u}})^{σ_{u}} = (U^{τ_{u}})^{σ_{u}}

. Then we also have

W^{τ_{u}} = U^{τ_{u}}

since

U^{τ_{u}}

is a direct sum of two irreducible

(W^{τ_{u}})^{σ_{u}}

-modules and

W^{τ_{u}} \neq (W^{τ_{u}})^{σ_{u}}

. Hence,

W

contains all the simple current

V i r (u) \otimes V i r (v) \otimes V i r (w)

-modules.

Now it is easy to see that

W = U

by the fusion rules. □

Theorem 3.17.

(1)

The Griess algebra

U_{2}

U

is generated by

\hat{e}

and

\hat{f}

(2)

The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

Proof. By Theorem 3.14 , it suffices to show the first assertion. Let $G$ be the Griess subalgebra generated by $\hat{e}$ and $\hat{f}$ . Then by direct computation, we can verify that $\hat{e}$ , $\hat{f}$ , ${\hat{e}}_{1} \hat{f}$ , ${\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})$ , ${\hat{f}}_{1} ({\hat{e}}_{1} ({\hat{e}}_{1} \hat{f}))$ , ${\hat{e}}_{1} ({\hat{f}}_{1} ({\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})))$ are linearly independent. Thus $G = U_{2}$ , since $dim U_{2} = 6$ . □

Theorem 3.18. The automorphism group

A u t U

U

is a dihedral group of order

10

Proof. Recall that $σ$ and $θ$ defined in Remark 2.2 generate a subgroup isomorphic to a dihedral group of order $10$ in $A u t U$ . By Lemma A.6 , there are exactly five conformal vectors of central charge $1 / 2$ in $U$ , namely, $e_{j} = σ^{j} \hat{e}, 0 \leq j \leq 4$ . Since $U$ is generated by $\hat{e}$ and $\hat{f} = σ \hat{e}$ , $A u t U$ can be considered as a subgroup of the permutation group on the five elements set ${e_{0}, e_{1}, e_{2}, e_{3}, e_{4}}$ . In fact, $U$ is generated by any $2$ distinct elements in the set ${e_{0}, e_{1}, e_{2}, e_{3}, e_{4}}$ . Thus $A u t U$ can not contain any $2$ -cycle nor $3$ -cycle since such automorphisms must fix at least two elements in the set and thus fix the whole $U$ . Hence there is no element of order $6$ either. Now let $(i_{1} i_{2} i_{3} i_{4})$ be a $4$ -cycle. Then $(i_{0} i_{1} i_{2} i_{3} i_{4}) (i_{1} i_{2} i_{3} i_{4}) = (i_{0} i_{1} i_{3}) (i_{2} i_{4}) .$ Since $(i_{0} i_{1} i_{2} i_{3} i_{4}) \in A u t U$ but $(i_{0} i_{1} i_{3}) (i_{2} i_{4}) / \in A u t U$ , there is no $4$ -cycle in $A u t U$ .
Therefore, the only possible elements are $5$ -cycles and the products of two disjoint $2$ -cycles.
Thus the assertion holds. □

Theorem 3.19. There are exactly nine irreducible modules

U (i, j), i, j = 1, 3, 5

for

U

. As

V i r (u) \otimes V i r (v) \otimes V i r (w)

-modules, they are of the following form.

\begin{matrix} U (i, j) & \sim = [0, h_{i, 1}, h_{j, 1}] \oplus [0, h_{i, 3}, h_{j, 5}] \oplus [0, h_{i, 5}, h_{j, 3}] \oplus [0, h_{i, 7}, h_{j, 7}] & \oplus [\frac{1}{2}, h_{i, 1}, h_{j, 7}] \oplus [\frac{1}{2}, h_{i, 3}, h_{j, 3}] \oplus [\frac{1}{2}, h_{i, 5}, h_{j, 5}] \oplus [\frac{1}{2}, h_{i, 7}, h_{j, 1}] & \oplus [\frac{1}{16}, h_{i, 2}, h_{j, 4}] \oplus [\frac{1}{16}, h_{i, 4}, h_{j, 2}] \oplus [\frac{1}{16}, h_{i, 6}, h_{j, 4}] \oplus [\frac{1}{16}, h_{i, 4}, h_{j, 6}], \end{matrix}

where

h_{r, s} = h_{r, s}^{5}

is defined by 3.2 .

The proof of this theorem will be given at the Appendix C. We shall first note that

U = U (1, 1)

and the lattice VOA

V_{\sqrt{2} E_{8}}

can be decomposed as follows (cf. Appendix B.2).

V_{\sqrt{2} E_{8}} \sim = \oplus_{\binom{1 \leq k_{j}, ℓ_{j} \leq j + 2}{\binom{k_{j}, ℓ_{j} \equiv 1 m o d 2}{j = 0, 1, \dots, 4}}} L (c_{1}, h_{k_{0}, k_{1}}^{1}) \otimes L (c_{1}, h_{ℓ_{0}, ℓ_{1}}^{1}) \otimes \dots \otimes L (c_{4}, h_{k_{3}, k_{4}}^{4}) \otimes L (c_{4}, h_{ℓ_{3}, ℓ_{4}}^{4}) \otimes U (k_{4}, ℓ_{4}),

where

c_{m}

and

h_{r, s}^{m}

are defined by 3.1 and 3.2 , respectively Recall that

τ_{\hat{e}} τ_{\hat{f}} = e^{2 π \sqrt{- 1} β (0)}

as automorphisms of

V_{\sqrt{2} E_{8}}

(cf. Eq. 3.4 ). It thus induces a natural action on each of the

U (i, j)

i, j = 1, 3, 5

. Hence if a VOA

V

contains a subalgebra isomorphic to

U

, then

τ_{\hat{e}} τ_{\hat{f}}

will define an automorphism of order

5

V

. The subalgebra

U^{τ_{\hat{e}} τ_{\hat{f}}}

consisting of the fixed points of

τ_{\hat{e}} τ_{\hat{f}}

U

is of the form

U^{τ_{\hat{e}} τ_{\hat{f}}} \sim = W_{5} (8 / 7) \otimes W_{5} (8 / 7),

where

W_{5} (8 / 7)

is a parafermion algebra of central charge

8 / 7

(cf. [32] ). It is well known that

W_{5} (8 / 7)

possesses a

Z_{5}

symmetry (cf. [6, 49] ). The automorphism

τ_{\hat{e}} τ_{\hat{f}}

in fact agrees with this symmetry.

Remark 3.20. Recall that

\hat{e}

and

\hat{f}

are fixed by the Weyl group

W (Φ) = W (A_{4}) \times W (A_{4})

of the root system

Φ = A_{4} \oplus A_{4}

L

. Since

U

is generated by

\hat{e}

and

\hat{f}

W (Φ)

acts trivially on

U

. There is an element

ψ

of order

5

W (Φ)

such that it induces a fixed-point-free action on

\sqrt{2} E_{8}

and on the Leech lattice

Λ

. Therefore, if the conjectured

Z_{5}

-orbifold construction of the Moonshine VOA

V^{♮}

holds, then one can prove that

U

is contained in

V^{♮}

by using

ψ

and that as an automorphism of

V^{♮}

τ_{\hat{e}} τ_{\hat{f}}

is of class

5 A

(cf. [37] ).

3.6 $6 A$ case.

In this case,

L = L (5) \sim = A_{2} \oplus A_{1} \oplus A_{5}

n_{5} = | E_{8} / L | = 6

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{2}}

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{1}}

, and

{\tilde{ω}}^{3} \in V_{\sqrt{2} A_{5}}

defined by 2.9 are of central charge

4 / 5

1 / 2

, and

5 / 4

, respectively. Let

K = s p a n_{Z} {α_{0}, α_{1}, α_{2}, α_{3}, α_{4}, α_{8}} \sim = A_{1} \oplus A_{5}

J = s p a n_{Z} {α_{6}, α_{7}} \sim = A_{2}

, and

\tilde{α} = α_{3} + 2 α_{4} + 3 α_{5} + 2 α_{6} + α_{7} + α_{8}

. Then

Z \tilde{α} + K = K \cup (\tilde{α} + K)

. We have

〈 α_{j}, \tilde{α} 〉 = 0

for

j = 0, 1, 3, 4, 6, 7

〈 α_{2}, \tilde{α} 〉 = 〈 α_{8}, \tilde{α} 〉 = - 1

〈 \tilde{α}, \tilde{α} 〉 = 2

, and

α_{0} + α_{4} + α_{8} + 2 α_{1} + 2 α_{3} + 2 \tilde{α} + 3 α_{2} = 0

. Hence

{α_{0}, α_{1}, α_{2}, α_{3}, α_{4}, \tilde{α}, α_{8}}

forms an extended

E_{6}

diagram and so

K \cup (\tilde{α} + K) \sim = E_{6}

. Moreover,

L = J \oplus K

and

〈 J, Z \tilde{α} + K 〉 = 0

. Therefore, we have isometric embeddings

A_{2} \oplus A_{1} \oplus A_{5} \subset A_{2} \oplus E_{6} \subset E_{8} .

Then we obtain a conformal vector

\tilde{ω} (E_{6}) = \frac{1}{168} \sum_{α \in Φ^{+} (E_{6})} α (- 1)^{2} \cdot 1 + \frac{1}{14} \sum_{α \in Φ^{+} (E_{6})} (e^{\sqrt{2} α} + e^{- \sqrt{2} α})

of central charge

6 / 7

V_{\sqrt{2} E_{8}}

. By our embeddings, we also know that

V_{\sqrt{2} E_{6}} \sim = {v \in V_{\sqrt{2} E_{8}} | (s^{1})_{1} v = ({\tilde{ω}}^{1})_{1} v = 0} .

α \in 3 α_{5} + L

satisfies

〈 α, α 〉 = 2

, then

α \in \tilde{α} + K

since

3 α_{5} + L = \tilde{α} + L

and

〈 J, \tilde{α} + K 〉 = 0

. Thus

Φ (E_{6}) = Φ (K) \cup {α \in 3 α_{5} + L | 〈 α, α 〉 = 2}

. Note also that the Virasoro element of

V_{\sqrt{2} E_{6}}

coincides with that of

V_{\sqrt{2} K}

. Now we can verify that

\tilde{ω} (E_{6}) = \frac{2}{7} {\tilde{ω}}^{2} + \frac{4}{7} {\tilde{ω}}^{3} + \frac{1}{14} X^{3} \in U .

Let

w^{1} = {\tilde{ω}}^{1}

w^{2} = \tilde{ω} (E_{6})

, and

w^{3} = {\tilde{ω}}^{2} + {\tilde{ω}}^{3} - w^{2}

. Then

{w^{1}, w^{2}, w^{3}}

is a set of mutually orthogonal conformal vectors of central charge

4 / 5, 6 / 7

, and

25 / 28

, respectively and the Virasoro element

ω^{'} = {\tilde{ω}}^{1} + {\tilde{ω}}^{2} + {\tilde{ω}}^{3}

U

is a sum of

w^{1}, w^{2}, w^{3}

. Note that

s (E_{6}) = \frac{1}{28} \sum_{α \in Φ^{+} (E_{6})} (α (- 1)^{2} \cdot 1 - 2 (e^{\sqrt{2} α} + e^{- \sqrt{2} α}))

is a linear combination of

s^{2}

s^{3}

, and

w^{3}

. Recall that

s^{1} = s (A_{2})

s^{2} = s (A_{1})

, and

s^{3} = s (A_{5})

are defined by 2.9 .

Now let

U^{'} = {v \in V_{\sqrt{2} E_{8}} | (s^{1})_{1} v = s (E_{6})_{1} v = 0}

. Then

\begin{matrix} U^{'} & = {v \in V_{\sqrt{2} E_{8}} | (s^{1})_{1} v = (s^{2})_{1} v = (s^{3})_{1} v = (w^{3})_{1} v = 0} & = {u \in U | (w^{3})_{1} u = 0} . \end{matrix}

Hence

U^{'} \subset U

and by using the results in the

3 A

case, we know that

\begin{matrix} U^{'} \sim = & (L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3)) \otimes (L (\frac{6}{7}, 0) \oplus L (\frac{6}{7}, 5)) & \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3}) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{4}{3}) . \end{matrix}

Note that

U^{'} \otimes L (25 / 28, 0)

U^{'} (5 / 7) \otimes L (25 / 28, 9 / 7)

, and

U^{'} (1 / 7) \otimes L (25 / 28, 34 / 7)

are the only irreducible

U^{'} \otimes L (25 / 28, 0)

-modules which have integral weights (cf. [44] ), where

\begin{matrix} U^{'} (\frac{1}{7}) \sim = & (L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3)) \otimes (L (\frac{6}{7}, \frac{1}{7}) \oplus L (\frac{6}{7}, \frac{22}{7})) \end{matrix}

\begin{matrix} \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{10}{21}) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{10}{21}), \end{matrix}

\begin{matrix} U^{'} (\frac{5}{7}) \sim = & (L (\frac{4}{5}, 0) \oplus L (\frac{4}{5}, 3)) \otimes (L (\frac{6}{7}, \frac{5}{7}) \oplus L (\frac{6}{7}, \frac{12}{7})) \end{matrix}

\begin{matrix} \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{1}{21}) \oplus L (\frac{4}{5}, \frac{2}{3}) \otimes L (\frac{6}{7}, \frac{1}{21}) . \end{matrix}

By direct computation, it is straightforward to verify that

U

is generated by its weight

2

subspace and it contains some highest weight vectors of weight

5 / 7

and

1 / 7

with respect to

U^{'}

(cf. see Appendix B.3.1 for details). Thus we have the following theorem.

Theorem 3.21. As a module of

U^{'} \otimes L (25 / 28, 0)

U \sim = U^{'} \otimes L (\frac{25}{28}, 0) \oplus U^{'} (\frac{5}{7}) \otimes L (\frac{25}{28}, \frac{9}{7}) \oplus U^{'} (\frac{1}{7}) \otimes L (\frac{25}{28}, \frac{34}{7}) .

Moreover,

U

is generated by its weight

2

subspace

U_{2}

Theorem 3.22.

(1)

The Griess algebra

U_{2}

U

is generated by

\hat{e}

and

\hat{f}

(2)

The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

Proof. Let $G$ be the Griess subalgebra generated by $\hat{e}$ and $\hat{f}$ . By A.7 , it is easy to verify that $\hat{e}$ , $\hat{f}$ , ${\hat{e}}_{1} \hat{f}$ , ${\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})$ , ${\hat{f}}_{1} ({\hat{e}}_{1} \hat{f})$ , ${\hat{f}}_{1} ({\hat{e}}_{1} ({\hat{e}}_{1} \hat{f}))$ , ${\hat{e}}_{1} ({\hat{f}}_{1} ({\hat{e}}_{1} \hat{f}))$ , ${\hat{e}}_{1} ({\hat{f}}_{1} ({\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})))$ are linearly independent. Thus $G = U_{2}$ , since $dim U_{2} = 8$ . The second assertion follows from the preceding theorem. □

Theorem 3.23. The automorphism group

A u t U

U

is a dihedral group of order

12

Proof. By Lemma A.9 , there are exactly seven conformal vectors of central charge

1 / 2

U

, namely,

{\tilde{ω}}^{2}

and

e_{j} = σ^{j} \hat{e}, 0 \leq j \leq 5

. Moreover, we have

〈 {\tilde{ω}}^{2}, e_{j} 〉 = 1 / 32

for any

0 \leq j \leq 5

and

\begin{matrix} 〈 e_{i}, e_{j} 〉 = {\begin{matrix} 1 / 4 & if i = j, \\ 1 / 32 & if i - j \equiv 3 m o d 6, \\ 13 / 2^{10} & if i - j \equiv \pm 2 m o d 6, \\ 5 / 2^{10} & if i - j \equiv \pm 1 m o d 6 . \end{matrix} \end{matrix}

(3.6)

Since

U

is generated by

\hat{e}

and

\hat{f} = σ \hat{e}

A u t U

can be considered as a subgroup of the permutation group on the set

{{\tilde{ω}}^{2}, e_{0}, e_{1}, \dots, e_{5}}

. Now let

g \in A u t U

. Then

g

must preserve the inner product and so

g

fixes

{\tilde{ω}}^{2}

. Let

g e_{j} = e_{ν (j)}, 0 \leq j \leq 5

. Then by 3.6 ,

ν (i) - ν (j) \equiv \pm (i - j) m o d 6

for any

i, j = 0, 1, . . ., 5

. Hence there are only

12

possiblechoices for

ν

. Thus the assertion holds since

σ

and

θ

in Remark 2.2 generate a subgroup of

A u t U

isomorphic to a dihedral group of order

12

. □

There is a

Z_{2}

symmetry among the irreducible modules of

L (25 / 28, 0)

. It is given by

ρ = {\begin{matrix} 1 & on L (25 / 28, h_{r, s}) for odd s, \\ - 1 & on L (25 / 28, h_{r, s}) for even s, \end{matrix}

where

h_{r, s} = h_{r, s}^{5}

is defined by 3.2 . On the other hand,

U^{'}

defines an automorphism

τ

of order

3

(cf. the

3 A

case). If

V

is a VOA which contains a subalgebra isomorphic to

U

, then there is an automorphism of order

6

defined by

τ ρ

. In fact,

τ_{\hat{e}} τ_{\hat{f}} = τ ρ

in this case.

Theorem 3.24. If the Moonshine VOA

V^{♮}

contains a subalgebra isomorphic to

U

, then as an automorphism of

V^{♮}

τ_{\hat{e}} τ_{\hat{f}}

is of class

6 A

Proof. Since $τ_{\hat{e}} τ_{\hat{f}} = τ ρ$ , we have $(τ_{\hat{e}} τ_{\hat{f}})^{2} = τ$ is of class $3 A$ and $(τ_{\hat{e}} τ_{\hat{f}})^{3} = ρ$ is of class $2 A$ .
Thus $τ_{\hat{e}} τ_{\hat{f}}$ is of class $6 A$ . □

3.7 $4 B$ case.

In this case,

L = L (6) \sim = A_{1} \oplus A_{7}

n_{6} = | E_{8} / L | = 4

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{1}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{7}}

defined by 2.9 are of central charge

1 / 2

and

7 / 5

, respectively.

Let

\tilde{α} = 2 α_{6}

. Then

E_{7} = A_{7} \cup (\tilde{α} + A_{7})

. Thus we obtain a conformal vector

\tilde{ω} (E_{7})

of central charge

7 / 10

. Let

w^{1} = {\tilde{ω}}^{1}

w^{2} = \tilde{ω} (E_{7})

, and

w^{3} = {\tilde{ω}}^{2} - w^{2}

. Then

{w^{1}, w^{2}, w^{3}}

is a set of mutually orthogonal conformal vectors of central charge

1 / 2

7 / 10

, and

7 / 10

, respectively. Actually,

w^{1} = {\tilde{ω}}^{1}, w^{2} = \frac{1}{2} {\tilde{ω}}^{2} + \frac{1}{20} X^{2}, w^{3} = \frac{1}{2} {\tilde{ω}}^{2} - \frac{1}{20} X^{2} .

Lemma 3.25. Let

u = X^{1} + X^{3}

and

v = X^{1} - X^{3}

. Then

u

and

v

are highest weight vectors of highest weight

(1 / 2, 3 / 2, 0)

and

(1 / 2, 0, 3 / 2)

with respect to

V i r (w^{1}) \otimes V i r (w^{2}) \otimes V i r (w^{3})

, respectively.

Proof. It follows from A.10 that $(w^{2})_{1} X^{1} = \frac{3}{4} X^{1} + \frac{3}{4} X^{3}, (w^{2})_{1} X^{3} = \frac{3}{4} X^{1} + \frac{3}{4} X^{3} .$ Hence we have $(w^{2})_{1} u = \frac{3}{2} u$ and $(w^{2})_{1} v = 0$ . Similarly, $(w^{3})_{1} u = 0$ and $(w^{3})_{1} v = \frac{3}{2} v$ . □

For simplicity of notation, we denote

L (1 / 2, h_{1}) \otimes L (7 / 10, h_{2}) \otimes L (7 / 10, h_{3})

[h_{1}, h_{2}, h_{3}]

The following proposition is an immediate consequence of the above lemma.

Proposition 3.26. As a module of

V i r (w^{1}) \otimes V i r (w^{2}) \otimes V i r (w^{3})

U \sim = [0, 0, 0] \oplus [\frac{1}{2}, \frac{3}{2}, 0] \oplus [\frac{1}{2}, 0, \frac{3}{2}] \oplus [0, \frac{3}{2}, \frac{3}{2}] .

The coset subalgebra

U

contains four more sets

{x^{j}, y^{j}, z^{j}}, 0 \leq j \leq 3

, of three mutually orthogonal conformal vectors of central charge

1 / 2

7 / 10

7 / 10

, respectively (cf. Appendix A). They are given by

\begin{matrix} x^{j} & = \frac{1}{8} {\tilde{ω}}^{1} + \frac{5}{16} {\tilde{ω}}^{2} + \frac{1}{32} ((\sqrt{- 1})^{j} X^{1} + (- 1)^{j} X^{2} + (- \sqrt{- 1})^{j} X^{3}), \end{matrix}

\begin{matrix} y^{j} & = \frac{1}{2} {\tilde{ω}}^{2} - (- 1)^{j} \frac{1}{20} X^{2}, \end{matrix}

\begin{matrix} z^{j} & = \frac{7}{8} {\tilde{ω}}^{1} + \frac{3}{16} {\tilde{ω}}^{2} - \frac{1}{32} ((\sqrt{- 1})^{j} X^{1} - \frac{3}{5} (- 1)^{j} X^{2} + (- \sqrt{- 1})^{j} X^{3}) . \end{matrix}

\begin{matrix}  \end{matrix}

Note that

x^{0} = \hat{e}

. The following proposition can be verified by direct computation.

Proposition 3.27. As a module of

V i r (x^{j}) \otimes V i r (y^{j}) \otimes V i r (z^{j})

U \sim = [0, 0, 0] \oplus [\frac{1}{2}, 0, \frac{3}{2}] \oplus [\frac{1}{16}, \frac{3}{2}, \frac{7}{16}]

j = 0, 2

and

U \sim = [0, 0, 0] \oplus [\frac{1}{2}, \frac{3}{2}, 0] \oplus [\frac{1}{16}, \frac{7}{16}, \frac{3}{2}]

j = 1, 3

Theorem 3.28.

(1)

The Griess algebra

U_{2}

U

is generated by

\hat{e}

and

\hat{f}

(2)

The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

Proof. Let $G$ be the Griess subalgebra generated by $\hat{e}$ and $\hat{f}$ . By A.10 , it is easy to verify that $\hat{e}$ , $\hat{f}$ , ${\hat{e}}_{1} \hat{f}$ , ${\hat{e}}_{1} ({\hat{e}}_{1} \hat{f})$ , ${\hat{f}}_{1} ({\hat{e}}_{1} \hat{f})$ are linearly independent. Thus $G = U_{2}$ , since $dim U_{2} = 5$ .
By the structure of $U$ , it is east to show that $U$ is generated by $U_{2}$ . Hence the second assertion holds. □

Theorem 3.29. The automorphism group

A u t U

U

is a dihedral group of order

8

Proof. There are exactly five conformal vectors of central charge $1 / 2$ in $U$ , namely, ${\tilde{ω}}^{1}$ and $e_{j} = σ^{j} \hat{e}$ , $0 \leq j \leq 3$ . Moreover, we have $〈 {\tilde{ω}}^{1}, e_{j} 〉 = 1 / 32$ for any $j = 0, 1, 2, 3$ , and $〈 e_{i}, e_{j} 〉 = {\begin{matrix} 1 / 32 & if i + j \equiv 0 m o d 2, \\ 1 / 2^{8} & if i + j \equiv 1 m o d 2 . \end{matrix}$ Since $U$ is generated by $\hat{e}$ and $\hat{f} = σ \hat{e}$ , $A u t U$ can be considered as a subgroup of the permutation group on the set ${{\tilde{ω}}^{1}, e_{0}, e_{1}, e_{2}, e_{3}}$ . Furthermore, $A u t U$ must preserve the inner product $〈 \cdot, \cdot 〉$ so that $A u t U$ fixes ${\tilde{ω}}^{1}$ and $| A u t U | \leq 8$ . Since $σ$ and $θ$ generate a subgroup isomorphic to a dihedral group of order $8$ in $A u t U$ , we have the assertion. □

The set of all irreducible modules of

U

can be classified easily by using the same method as in [27, 29, 47] . They are given by

\begin{matrix} [0, 0, 0] \oplus [0, \frac{3}{2}, \frac{3}{2}] \oplus [\frac{1}{2}, \frac{3}{2}, 0] \oplus [\frac{1}{2}, 0, \frac{3}{2}], & [0, 0, \frac{3}{2}] \oplus [0, \frac{3}{2}, 0] \oplus [\frac{1}{2}, 0, 0] \oplus [\frac{1}{2}, \frac{3}{2}, \frac{3}{2}], \end{matrix}

\begin{matrix} [0, 0, \frac{3}{5}] \oplus [0, \frac{3}{2}, \frac{1}{10}] \oplus [\frac{1}{2}, \frac{3}{2}, \frac{3}{5}] \oplus [\frac{1}{2}, 0, \frac{1}{10}], & [0, \frac{3}{5}, 0] \oplus [0, \frac{1}{10}, \frac{3}{2}] \oplus [\frac{1}{2}, \frac{3}{5}, \frac{3}{2}] \oplus [\frac{1}{2}, \frac{1}{10}, 0], \end{matrix}

\begin{matrix} [0, \frac{3}{5}, \frac{3}{5}] \oplus [0, \frac{1}{10}, \frac{1}{10}] \oplus [\frac{1}{2}, \frac{1}{10}, \frac{3}{5}] \oplus [\frac{1}{2}, \frac{3}{5}, \frac{1}{10}], & [0, \frac{3}{5}, \frac{1}{10}] \oplus [0, \frac{1}{10}, \frac{3}{5}] \oplus [\frac{1}{2}, \frac{1}{10}, \frac{1}{10}] \oplus [\frac{1}{2}, \frac{3}{5}, \frac{3}{5}], \end{matrix}

\begin{matrix} [0, 0, \frac{1}{10}] \oplus [0, \frac{3}{2}, \frac{3}{5}] \oplus [\frac{1}{2}, \frac{3}{2}, \frac{1}{10}] \oplus [\frac{1}{2}, 0, \frac{3}{5}], & [0, \frac{1}{10}, 0] \oplus [0, \frac{3}{5}, \frac{3}{2}] \oplus [\frac{1}{2}, \frac{3}{5}, 0] \oplus [\frac{1}{2}, \frac{1}{10}, \frac{3}{2}], \end{matrix}

and

\begin{matrix} [\frac{1}{16}, \frac{7}{16}, \frac{7}{16}] \otimes Q, & [\frac{1}{16}, \frac{7}{16}, \frac{3}{80}] \otimes Q, & [\frac{1}{16}, \frac{3}{80}, \frac{7}{16}] \otimes Q, & [\frac{1}{16}, \frac{3}{80}, \frac{3}{80}] \otimes Q, \end{matrix}

where

Q

is the unique

2

-dimensional irreducible module of the quaternion group of order

8

The fixed point subalgebra

U^{τ_{\hat{e}} τ_{\hat{f}}}

τ_{\hat{e}} τ_{\hat{f}}

U

is isomorphic to

[0, 0, 0] \oplus [0, 3 / 2, 3 / 2]

. The fusion rules among irreducible modules of

L (7 / 10, 0) \otimes L (7 / 10, 0) \oplus L (7 / 10, 3 / 2) \otimes L (7 / 10, 3 / 2)

can be computed easily. There is a

Z_{4}

symmetry given as follows.

τ = {\begin{matrix} 1 & on [0, 0] \oplus [\frac{3}{2}, \frac{3}{2}], [0, \frac{3}{5}] \oplus [\frac{3}{2}, \frac{1}{10}], \\ [\frac{3}{5}, \frac{3}{5}] \oplus [\frac{1}{10}, \frac{1}{10}], [\frac{1}{10}, \frac{3}{2}] \oplus [\frac{3}{5}, 0], \\ \sqrt{- 1} & on [\frac{7}{16}, \frac{7}{16}]^{+}, [\frac{7}{16}, \frac{3}{80}]^{+}, \\ [\frac{3}{80}, \frac{7}{16}]^{+}, [\frac{3}{80}, \frac{3}{80}]^{+}, \\ - 1 & on [0, \frac{3}{2}] \oplus [\frac{3}{2}, 0], [\frac{3}{2}, \frac{3}{5}] \oplus [0, \frac{1}{10}], \\ [\frac{3}{5}, \frac{1}{10}] \oplus [\frac{1}{10}, \frac{3}{5}], [\frac{1}{10}, 0] \oplus [\frac{3}{5}, \frac{3}{2}], \\ - \sqrt{- 1} & on [\frac{7}{16}, \frac{7}{16}]^{-}, [\frac{7}{16}, \frac{3}{80}]^{-}, \\ [\frac{3}{80}, \frac{7}{16}]^{-}, [\frac{3}{80}, \frac{3}{80}]^{-}, \end{matrix}

where

[h_{1}, h_{2}]

denotes the irreducible module

L (7 / 10, h_{1}) \otimes L (7 / 10, h_{2})

L (7 / 10, 0) \otimes L (7 / 10, 0)

. Suppose

U

is contained in a VOA

V

. Then all

τ, τ_{\hat{e}}

, and

τ_{\hat{f}}

are well defined automorphisms of

V

. In this case,

τ = τ_{\hat{e}} τ_{\hat{f}}

and

τ^{2} = τ_{w^{1}}

. Thus we have the following theorem.

Theorem 3.30. If the Moonshine VOA

V^{♮}

contains a subalgebra isomorphic to

U

, then as an automorphism of

V^{♮}

τ_{\hat{e}} τ_{\hat{f}}

is of class

4 B

3.8 $2 B$ case.

In this case,

L = L (7) \sim = D_{8}

n_{7} = | E_{8} / L | = 2

, and the conformal vector

{\tilde{ω}}^{1} \in V_{\sqrt{2} D_{8}}

defined by 2.9 is of central charge

1

. The Virasoro element

ω^{'}

U

is equal to

{\tilde{ω}}^{1}

. Let

ε_{1}, ε_{2}, . . ., ε_{8} \in R^{8}

be such that

〈 ε_{i}, ε_{j} 〉 = 2 δ_{i j}

for any

i, j

. Then

\sqrt{2} D_{8} = {a_{1} ε_{1} + \dots + a_{8} ε_{8} | a_{i} \in Z, a_{1} + \dots + a_{8} \equiv 0 m o d 2} .

Let

γ = \frac{1}{2} (ε_{1} + \dots + ε_{8})

. Then

\sqrt{2} E_{8} \sim = \sqrt{2} D_{8} \cup (γ + \sqrt{2} D_{8})

. Denote

U^{0} = {u \in V_{\sqrt{2} D_{8}} | s (D_{8})_{1} u = 0}, U^{1} = {u \in V_{γ + \sqrt{2} D_{8}} | s (D_{8})_{1} u = 0},

where

s (D_{8})

is defined as in 2.1 . By [8, 9] , we know that

U^{0} \sim = V_{Z 2 γ}^{+}

and

U^{1} \sim = V_{γ + Z 2 γ}^{+}

Note that

Z γ = Z 2 γ \cup (γ + Z 2 γ)

and hence we have

U = U^{0} \oplus U^{1} \sim = V_{Z γ}^{+}

. Since

〈 γ, γ 〉 = 4

, it is well know that

V_{Z γ}^{+} \sim = L (\frac{1}{2}, 0) \otimes L (\frac{1}{2}, 0) .

In fact,

\frac{1}{16} γ (- 1)^{2} \cdot 1 + \frac{1}{4} (e^{γ} + e^{- γ})

and

\frac{1}{16} γ (- 1)^{2} \cdot 1 - \frac{1}{4} (e^{γ} + e^{- γ})

are the two mutually orthogonal conformal vectors of central charge

1 / 2

V_{Z γ}^{+}

(cf. [12, 39, 40] ).

The following theorem is clear from the structure of

U

Theorem 3.31. The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

. Moreover, the automorphism group

A u t U

U

is of order

2

Both of

\hat{e}

and

\hat{f}

are fixed by

θ

and thus

U

is contained in

V_{Λ}^{+} \subset V^{♮}

. In this case,

\hat{e}

and

\hat{f}

are mutually orthogonal and so

τ_{\hat{e}}

and

τ_{\hat{f}}

are commutative. Hence,

| τ_{\hat{e}} τ_{\hat{f}} | = 2

and as an automorphism of

V^{♮}

τ_{\hat{e}} τ_{\hat{f}}

is of class

2 B

(see also Shimakura [45] ).

3.9 $3 C$ case.

In this case,

L = L (8) \sim = A_{8}

n_{8} = | E_{8} / L | = 3

, and the conformal vector

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{8}}

defined by 2.9 is of central charge

16 / 11

. The Virasoro element

ω^{'}

U

is equal to

{\tilde{ω}}^{1}

The following lemma can be easily verified.

Lemma 3.32. Let

x = \frac{11}{32} {\tilde{ω}}^{1} + \frac{1}{32} (X^{1} + X^{2}), y = \frac{21}{32} {\tilde{ω}}^{1} - \frac{1}{32} (X^{1} + X^{2}) .

Then

x

and

y

are mutually orthogonal conformal vectors of central charge

1 / 2

and

21 / 22

, respectively. Moreover,

{\tilde{ω}}^{1} = x + y

The above lemma implies that

U

contains

V i r (x) \otimes V i r (y) \sim = L (1 / 2, 0) \otimes L (21 / 22, 0)

Lemma 3.33. Let

v = X^{1} - X^{2}

. Then

v

is a highest weight vector of highest weight

(1 / 16, 31 / 16)

with respect to

V i r (x) \otimes V i r (y)

Proof. By A.12 , we have $x_{1} v = {(\frac{11}{32} {\tilde{ω}}^{1} + \frac{1}{32} (X^{1} + X^{2}))}_{1} (X^{1} - X^{2}) = \frac{1}{16} (X^{1} - X^{2}) .$ Hence the assertion holds. □

Similarly, we obtain the following lemma by direct computation.

Lemma 3.34. Let

u = 4 v_{- 1} v - 45 x_{- 3} \cdot 1 - 9 x_{- 1} x - 682 x_{- 1} y - 1089 y_{- 1} y - 165 y_{- 3} \cdot 1 .

Then

u

is a highest weight vector of highest weight

(1 / 2, 7 / 2)

with respect to

V i r (x) \otimes V i r (y)

and

〈 u, u 〉 = 1280076

Note that

L (1 / 2, h_{1}) \otimes L (21 / 22, h_{2})

with

(h_{1}, h_{2}) = (0, 0)

(0, 8)

(1 / 2, 7 / 2)

(1 / 2, 45 / 2)

(1 / 16, 31 / 16)

and

(1 / 16, 175 / 16)

are the only irreducible

L (1 / 2, 0) \otimes L (21 / 22, 0)

-modules which are integrally graded. For simplicity of notation, we shall denote

L (1 / 2, h_{1}) \otimes L (21 / 22, h_{2})

[h_{1}, h_{2}]

. The following theorem can be obtained by using the characters (cf. Appendix B).

Theorem 3.35 (Theorem B.5 ). As a module of

V i r (x) \otimes V i r (y)

U \sim = [0, 0] \oplus [0, 8] \oplus [\frac{1}{2}, \frac{7}{2}] \oplus [\frac{1}{2}, \frac{45}{2}] \oplus [\frac{1}{16}, \frac{31}{16}] \oplus [\frac{1}{16}, \frac{175}{16}] .

Theorem 3.36.

(1)

The Griess algebra

U_{2}

U

is generated by

\hat{e}

and

\hat{f}

(2)

The coset subalgebra

U

is generated by

\hat{e}

and

\hat{f}

Proof. The first assertion is clear since

dim U_{2} = 3

. Let

W

be the subalgebra generated by

U_{2}

. We want to show that

W = U

. By Lemmas 3.33 and 3.34 ,

W

contains submodules of the form

[0, 0]

[1 / 2, 7 / 2]

, and

[1 / 16, 31 / 16]

We shall show that there is a highest weight vector of highest weight

(0, 8)

with respect to

V i r (x) \otimes V i r (y)

W

. Suppose

W

does not contain a highest weight vector of highest weight

(0, 8)

. Then the fixed point subalgebra

W^{τ_{x}}

of the Miyamoto involution

τ_{x}

W

is of the form

W^{τ_{x}} \sim = [0, 0] \oplus [1 / 2, 7 / 2]

. Let

w

be a highest weight vector of highest weight

(1 / 2, 7 / 2)

. We normalize

w

so that

〈 w, w 〉 = 1

. For example, we may take

w = u / \sqrt{〈 u, u 〉}

Then by fusion rules,

w_{- 1} w \in [0, 0]

. It is well known (cf. [14] [15] ) that there is an explicit construction of the vertex operator superalgebra

L (1 / 2, 0) \oplus L (1 / 2, 1 / 2)

by using one free fermionic field and we can find an orthonormal basis of

L (1 / 2, 0) \oplus L (1 / 2, 1 / 2)

(cf.

[23] ). By the assumption

W^{τ_{x}} \sim = [0, 0] \oplus [1 / 2, 7 / 2]

Z_{2}

-graded and so we can use the orthonormal basis of

L (1 / 2, 0) \oplus L (1 / 2, 1 / 2)

to make a computation of the inner product

〈 w_{- 1} w, w_{- 1} w 〉

easier. As a result, we can obtain

〈 w_{- 1} w, w_{- 1} w 〉 = \frac{231091005602}{134258427} .

Note that this value is deduced based on the assumption that

[0, 8]

is not contained in

W

as a

V i r (x) \otimes V i r (y)

-submodule.

On the other hand, we can also compute

〈 w_{- 1} w, w_{- 1} w 〉

directly by using the definition of

w

and the Jacobi identity. In this case, we have

\begin{matrix} 〈 w_{- 1} w, w_{- 1} w 〉 & = 〈 w, w_{7} w_{- 1} w 〉 = 〈 w, (w_{- 1} w)_{7} w 〉 & = \sum_{n \geq 0} (- 1)^{n} (\binom{- 1}{n}) 〈 w, (w_{- 1 - n} w_{7 + n} + w_{6 - n} w_{n}) w 〉 & = 〈 w, w_{- 1} w_{7} w 〉 + \sum_{n = 0}^{7} 〈 w, w_{6 - n} w_{n} w 〉 & = 2 〈 w, w 〉^{2} + \sum_{n = 0}^{5} 〈 w_{n} w, w_{n} w 〉 . \end{matrix}

Then by computing

〈 w_{5 - m} w, w_{5 - m} w 〉

m = 0, 1, . . ., 5

, inductively, we obtain another value

〈 w_{- 1} w, w_{- 1} w 〉 = \frac{1036050}{589} (\neq \frac{231091005602}{134258427}) .

Note that this value is deduced from the structure of the Griess algebra

U_{2}

and independent of the shape of

W

as a

V i r (x) \otimes V i r (y)

-module. Thus this contradiction comes from the assumption that

W

does not contain a highest weight vector of highest weight

(0, 8)

. Hence we conclude that there is a highest weight vector of highest weight

(0, 8)

W

. By the computation above, we also note that

w_{- 1} w

contains a singular vector with highest weight

(0, 8)

as a non-trivial summand.

Then,

W

also contains a highest weight vector of highest weight

(1 / 2, 45 / 2)

; otherwise,

W^{τ_{x}} \sim = [0, 0] \oplus [1 / 2, 7 / 2] \oplus [0, 8]

and the orthogonal complement of

W^{τ_{x}}

U^{τ_{x}}

withrespect to a positive definite invariant hermitian form is isomorphic to

[1 / 2, 45 / 2]

. But the orthogonal complement must be a module for

W^{τ_{x}}

, which is impossible by the fusion rules. Thus

W^{τ_{x}} = U^{τ_{x}}

Since the fusion product of

[1 / 2, 45 / 2]

and

[1 / 16, 31 / 16]

[1 / 16, 175 / 16]

W

contains a highest weight vector of highest weight

(1 / 16, 175 / 16)

also, and hence

W = U

as desired. □

Theorem 3.37. The automorphism group

A u t U

U

is a symmetric group

S_{3}

of degree

3

Proof. By Lemma A.14 , there are exactly three conformal vectors of central charge $1 / 2$ in $U$ , namely, $σ^{j} \hat{e}$ , $j = 0, 1, 2$ . Thus $A u t U$ can be considered as a subgroup of $S_{3}$ . Since $σ$ and $θ$ already generate a subgroup of order $6$ in $A u t U$ , we conclude that $A u t U \sim = S_{3}$ . □

Theorem 3.38. There are exactly five irreducible

U

-modules

U (2 k), 0 \leq k \leq 4

In fact,

U (0) = U

and as

V i r (x) \otimes V i r (y)

-modules,

\begin{matrix} U (2) & \sim = [0, \frac{13}{11}] \oplus [0, \frac{35}{11}] \oplus [\frac{1}{2}, \frac{15}{22}] \oplus [\frac{1}{2}, \frac{301}{22}] \oplus [\frac{1}{16}, \frac{21}{176}] \oplus [\frac{1}{16}, \frac{901}{176}], \end{matrix}

\begin{matrix} U (4) & \sim = [0, \frac{6}{11}] \oplus [0, \frac{50}{11}] \oplus [\frac{1}{2}, \frac{1}{22}] \oplus [\frac{1}{2}, \frac{155}{22}] \oplus [\frac{1}{16}, \frac{85}{176}] \oplus [\frac{1}{16}, \frac{261}{176}], \end{matrix}

\begin{matrix} U (6) & \sim = [0, \frac{1}{11}] \oplus [0, \frac{111}{11}] \oplus [\frac{1}{2}, \frac{35}{22}] \oplus [\frac{1}{2}, \frac{57}{22}] \oplus [\frac{1}{16}, \frac{5}{176}] \oplus [\frac{1}{16}, \frac{533}{176}], \end{matrix}

\begin{matrix} U (8) & \sim = [0, \frac{20}{11}] \oplus [0, \frac{196}{11}] \oplus [\frac{1}{2}, \frac{7}{22}] \oplus [\frac{1}{2}, \frac{117}{22}] \oplus [\frac{1}{16}, \frac{133}{176}] \oplus [\frac{1}{16}, \frac{1365}{176}] . \end{matrix}

\begin{matrix}  \end{matrix}

We shall again give a proof at Appendix C. Note that the lattice VOA

V_{\sqrt{2} E_{8}}

can be decomposed as follows (cf. [28, 32] ).

\begin{matrix} V_{\sqrt{2} E_{8}} \sim = & \oplus_{\binom{0 \leq k_{j} \leq j + 1}{\binom{k_{j} \equiv 0 m o d 2}{j = 0, 1, . . ., 8}}} L (c_{1}, h_{k_{0} + 1, k_{1} + 1}^{1}) \otimes \dots L (c_{8}, h_{k_{7} + 1, k_{8} + 1}^{8}) \otimes U (k_{8}), \end{matrix}

(3.7)

\begin{matrix}  \end{matrix}

where

c_{m}

and

h_{r, s}^{m}

are given by 3.1 and 3.2 . Moreover, as an automorphism of

V_{\sqrt{2} E_{8}}

τ_{\hat{e}} τ_{\hat{f}} = e^{2 π \sqrt{- 1} β (0)}

is of order

3

(cf. Remark 3.4 ) and it induces a natural action on each of

U (2 k)

0 \leq k \leq 4

. Hence, for any VOA

V

which contains a subalgebra isomorphic to

U

τ_{\hat{e}} τ_{\hat{f}}

defines an automorphism of order

3

V

Remark 3.39. In Miyamoto[42] , it is shown that if

e

and

f

are two conformal vectors of central charge

1 / 2

in the Moonshine VOA

V^{♮}

such that

τ_{e} τ_{f}

is of order

3

and

〈 e, f 〉 = 1 / 2^{8}

, then the subalgebra

W

generated by

e

and

f

must contain a subalgebra isomorphic to

L (1 / 2, 0) \otimes L (21 / 22, 0)

and the weight

2

subspace

W_{2}

W

is of dimension

3

. In fact,

W \sim = U

L (1 / 2, 0) \otimes L (21 / 22, 0)

-modules and

τ_{e} τ_{f}

is of class

3 C

A Conformal Vectors in $U$

In this appendix, we shall compute the conformal vectors of the coset subalgebra

U

defined by 2.10 in each of the nine cases. Except for the cases of

5 A

and

6 A

, the computation was done by Maple 7. The cases for

5 A

and

6 A

were computed by Kazuhiro Yokoyama of Kyushu University using a computer algebra system Risa/Asir. As in Section 2.2 ,

L = L (i)

denotes the lattice associated with the Dynkin diagram obtained by removing the

i

-th node

α_{i}

in the extended

E_{8}

diagram 2.2 and

n_{i} = | E_{8} / L |

Let

ℬ = s p a n_{C} {{\tilde{ω}}^{1},, . . ., {\tilde{ω}}^{l}, X^{1}, . . ., X^{n_{i} - 1}},

where

{\tilde{ω}}^{j}

and

X^{j}

are defined by 2.9 and 3.3 . Then

dim ℬ = l + n_{i} - 1

. Actually,

ℬ

is equal to the Griess algebra

U_{2}

U

. Note that

\begin{matrix} \begin{matrix} ({\tilde{ω}}^{j})_{1} {\tilde{ω}}^{k} = 2 {\tilde{ω}}^{j} δ_{j, k}, 〈 {\tilde{ω}}^{j}, {\tilde{ω}}^{k} 〉 = c δ_{j, k} / 2, 〈 {\tilde{ω}}^{j}, X^{k} 〉 = 0, & 〈 X^{j}, X^{k} 〉 = 0 if j + k ≢ 0 m o d n_{i}, \end{matrix} \end{matrix}

(A.1)

where

c

denotes the central charge of

{\tilde{ω}}^{j}

1 A

case. In this case,

L = L (0) \sim = E_{8}

n_{0} = 1

, and

{\tilde{ω}}^{1} \in V_{\sqrt{2} E_{8}}

is the only conformal vector in

ℬ

, whose central charge is

1 / 2

2 A

case. In this case,

L = L (1) \sim = A_{1} \oplus E_{7}

n_{1} = 2

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{1}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} E_{7}}

are of central charge

1 / 2

and

7 / 10

, respectively. The product and the inner product in

ℬ

are given by A.1 and

({\tilde{ω}}^{1})_{1} X^{1} = \frac{1}{2} X^{1}, ({\tilde{ω}}^{2})_{1} X^{1} = \frac{3}{2} X^{1}, (X^{1})_{1} X^{1} = 224 {\tilde{ω}}^{1} + 480 {\tilde{ω}}^{2}, 〈 X^{1}, X^{1} 〉 = 112 .

Lemma A.1. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c X^{1}

be a conformal vector in

ℬ

. Then

(a, b, c)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 112 c^{2} - a = 0, & b^{2} + 240 c^{2} - b = 0, \\ a c + 3 b c - 2 c = 0 . \end{matrix} \end{matrix}

(A.2)

The central charge of

w

is given by

\frac{1}{2} a^{2} + \frac{7}{10} b^{2} + 224 c^{2}

The solutions

(a, b, c)

of Equation A.2 are as follows.

Central charge

1 / 2

(1, 0, 0), (\frac{1}{8}, \frac{5}{8}, \frac{1}{32}), (\frac{1}{8}, \frac{5}{8}, - \frac{1}{32})

Central charge

7 / 10

(0, 1, 0), (\frac{7}{8}, \frac{3}{8}, \frac{1}{32}), (\frac{7}{8}, \frac{3}{8}, - \frac{1}{32})

Central charge

6 / 5

(1, 1, 0)

3 A

case. In this case,

L = L (2) \sim = A_{2} \oplus E_{6}

n_{2} = 3

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{2}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} E_{6}}

are of central charge

4 / 5

and

6 / 7

, respectively. The productand the inner product in

ℬ

are given by A.1 and

\begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = \frac{2}{3} X^{1}, ({\tilde{ω}}^{1})_{1} X^{2} = \frac{2}{3} X^{2}, ({\tilde{ω}}^{2})_{1} X^{1} = \frac{4}{3} X^{1}, ({\tilde{ω}}^{2})_{1} X^{2} = \frac{4}{3} X^{2}, & (X^{1})_{1} X^{1} = 20 X^{2}, (X^{2})_{1} X^{2} = 20 X^{1}, (X^{1})_{1} X^{2} = 135 {\tilde{ω}}^{1} + 252 {\tilde{ω}}^{2}, & 〈 X^{1}, X^{2} 〉 = 81 . \end{matrix}

Lemma A.2. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c X^{1} + d X^{2}

be a conformal vector in

ℬ

. Then

(a, b, c, d)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 135 c d - a = 0, & b^{2} + 252 c d - b = 0, \\ 15 d^{2} + 2 a c + 4 b c - 3 c = 0, & 15 c^{2} + 2 a d + 4 b d - 3 d = 0 . \end{matrix} \end{matrix}

(A.3)

The central charge of

w

is given by

\frac{4}{5} a^{2} + \frac{6}{7} b^{2} + 324 c d

The solutions

(a, b, c, d)

of Equation ( A.3 ) are as follows, where

ξ = e^{2 π \sqrt{- 1} / 3}

is a primitive cubic root of unity.

Central charge

1 / 2

(\frac{5}{32}, \frac{7}{16}, \frac{1}{32} ξ^{j}, \frac{1}{32} ξ^{2 j})

j = 0, 1, 2

Central charge

4 / 5

(1, 0, 0, 0)

(\frac{1}{16}, \frac{7}{8}, - \frac{1}{48} ξ^{j}, - \frac{1}{48} ξ^{2 j})

j = 0, 1, 2

Central charge

6 / 7

(0, 1, 0, 0)

(\frac{15}{16}, \frac{1}{8}, \frac{1}{48} ξ^{j}, \frac{1}{48} ξ^{2 j})

j = 0, 1, 2

Central charge

81 / 70

(\frac{27}{32}, \frac{9}{16}, - \frac{1}{32} ξ^{j}, - \frac{1}{32} ξ^{2 j})

j = 0, 1, 2

Central charge

58 / 35

(1, 1, 0, 0)

4 A

case. In this case,

L = L (3) \sim = A_{3} \oplus D_{5}

n_{3} = 4

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{3}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} D_{5}}

are both of central charge

1

. The product and the inner product in

ℬ

are given by A.1 and

\begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = \frac{3}{4} X^{1}, & ({\tilde{ω}}^{1})_{1} X^{2} = X^{2}, & ({\tilde{ω}}^{1})_{1} X^{3} = \frac{3}{4} X^{3}, \end{matrix}

\begin{matrix} ({\tilde{ω}}^{2})_{1} X^{1} = \frac{5}{4} X^{1}, & ({\tilde{ω}}^{2})_{1} X^{2} = X^{2}, & ({\tilde{ω}}^{2})_{1} X^{3} = \frac{5}{4} X^{3}, \end{matrix}

\begin{matrix} (X^{1})_{1} X^{1} = 16 X^{2}, & (X^{1})_{1} X^{2} = 15 X^{3}, & (X^{1})_{1} X^{3} = 96 {\tilde{ω}}^{1} + 160 {\tilde{ω}}^{2} \end{matrix}

\begin{matrix} (X^{2})_{1} X^{2} = 120 {\tilde{ω}}^{1} + 120 {\tilde{ω}}^{2}, & (X^{2})_{1} X^{3} = 15 X^{1}, & (X^{3})_{1} X^{3} = 16 X^{2}, \end{matrix}

\begin{matrix} 〈 X^{1}, X^{3} 〉 = 64, & 〈 X^{2}, X^{2} 〉 = 60 . \end{matrix}

Lemma A.3. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c X^{1} + d X^{2} + e X^{3}

be a conformal vector in

ℬ

. Then

(a, b, c, d, e)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 60 d^{2} + 96 c e - a = 0, & b^{2} + 60 d^{2} + 160 c e - b = 0, \\ 3 a c + 5 b c + 60 d e - 4 c = 0, & 8 c^{2} + 8 e^{2} + a d + b d - d = 0, \\ 3 a e + 5 b e + 60 c d - 4 e = 0 . \end{matrix} \end{matrix}

(A.4)

The central charge of

w

is given by

a^{2} + b^{2} + 256 c e + 120 d^{2}

The solutions

(a, b, c, d, e)

of Equation A.4 are as follows, where

ξ = e^{π \sqrt{- 1} / 4}

is a primitive

8

-th root of unity.

Central charge

1 / 2

(\frac{3}{16}, \frac{5}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, \frac{1}{32} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central charge

1

\begin{matrix} (1, 0, 0, 0, 0), (0, 1, 0, 0, 0), \end{matrix}

\begin{matrix} ((1 + \sqrt{1 - 240 d^{2}}) / 2, (1 - \sqrt{1 - 240 d^{2}}) / 2, 0, d, 0), \end{matrix}

\begin{matrix} ((1 - \sqrt{1 - 240 d^{2}}) / 2, (1 + \sqrt{1 - 240 d^{2}}) / 2, 0, d, 0), w h e r e d \in C \ {0} . \end{matrix}

Central charge

3 / 2

(\frac{13}{16}, \frac{11}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, - \frac{1}{32} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central Charge

6 / 7

(\frac{1}{7}, \frac{5}{7}, \frac{1}{28} ξ^{2 j + 1}, 0, \frac{1}{28} ξ^{- (2 j + 1)})

0 \leq j \leq 3

Central Charge

8 / 7

(\frac{6}{7}, \frac{2}{7}, \frac{1}{28} ξ^{2 j + 1}, 0, \frac{1}{28} ξ^{- (2 j + 1)})

0 \leq j \leq 3

Central Charge

2

(1, 1, 0, 0, 0)

5 A

case. In this case,

L = L (4) \sim = A_{4} \oplus A_{4}

n_{4} = 5

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{4}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{4}}

are both of central charge

8 / 7

. The product and the inner product in

ℬ

are given by A.1 and

\begin{matrix} \begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = \frac{4}{5} X^{1}, & ({\tilde{ω}}^{2})_{1} X^{1} = \frac{6}{5} X^{1}, & ({\tilde{ω}}^{1})_{1} X^{2} = \frac{6}{5} X^{2}, & ({\tilde{ω}}^{2})_{1} X^{2} = \frac{4}{5} X^{2}, \\ ({\tilde{ω}}^{1})_{1} X^{3} = \frac{6}{5} X^{3}, & ({\tilde{ω}}^{2})_{1} X^{3} = \frac{4}{5} X^{3}, & ({\tilde{ω}}^{1})_{1} X^{4} = \frac{4}{5} X^{4}, & ({\tilde{ω}}^{2})_{1} X^{4} = \frac{6}{5} X^{4}, \end{matrix} \end{matrix}

\begin{matrix} (X^{i})_{1} (X^{j}) = {\begin{matrix} 12 X^{i + j} & if i + j ≢ 0 m o d 5, \\ 70 {\tilde{ω}}^{1} + 105 {\tilde{ω}}^{2} & if i = 5 - j = 1 or 4, \\ 105 {\tilde{ω}}^{1} + 70 {\tilde{ω}}^{2} & if i = 5 - j = 2 or 3, \end{matrix} \end{matrix}

\begin{matrix} 〈 X^{1}, X^{4} 〉 = 〈 X^{2}, X^{3} 〉 = 50 . \end{matrix}

Lemma A.4. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c X^{1} + d X^{2} + e X^{3} + f X^{4}

be a conformal vector in

ℬ

. Then

(a, b, c, d, e, f)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 70 c f + 105 d e - a = 0, & b^{2} + 105 c f + 70 d e - b = 0, \\ 4 a c + 6 b c + 30 e^{2} + 60 d f - 5 c = 0, & 6 a d + 4 b d + 30 c^{2} + 60 e f - 5 d = 0, \\ 6 a e + 4 b e + 30 f^{2} + 60 c d - 5 e = 0, & 4 a f + 6 b f + 30 d^{2} + 60 c e - 5 f = 0 . \end{matrix} \end{matrix}

(A.5)

The central charge of

w

is given by

\begin{matrix} \frac{8}{7} (a^{2} + b^{2}) + 200 (c f + d e) . \end{matrix}

(A.6)

In order to solve the above system of equations, we treat

a, b, c, d, e, f

as variables.

Let

C [a, b, c, d, e, f]

be the polynomial algebra with variables

a, b, c, d, e, f

and

ℐ

the ideal generated by the six polynomials which appear on the left hand side of ( A.5 ). We then compute the primary decomposition of the ideal

ℐ = \cap P_{i}

over the field

Q

of rational numbers and solve the system corresponding to each prime ideal

P_{i}

. By using the computer algebra system Risa/Asir, we found that

ℐ

is an intersection of

21

prime ideals and there are

63

different nontrivial solutions of ( A.5 ). Thus we have the following lemma.

Lemma A.5. There are exactly

63

conformal vectors in

ℬ

The central charges of those conformal vectors are easily calculated by A.6 . We verified that there are only

43

conformal vectors whose central charges are rational numbers.

Their central charges are shown in Table 1 .

Table 1 . Central charge (c.c.) and number of conformal vectors

c.c.	$1 / 2$	$25 / 28$	$8 / 7$	$16 / 7$	$25 / 14$	$39 / 28$
number	$5$	$10$	$12$	$1$	$5$	$10$

The following two lemmas were verified by computer.

Lemma A.6. There are exactly five conformal vectors of central charge

1 / 2

ℬ

, namely,

σ^{j} \hat{e}, 0 \leq j \leq 4

. Note that

σ^{j} \hat{e} = \frac{7}{32} ({\tilde{ω}}^{1} + {\tilde{ω}}^{2}) + \frac{1}{32} (ξ^{j} X^{1} + ξ^{2 j} X^{2} + ξ^{3 j} X^{3} + ξ^{4 j} X^{4}),

where

ξ = e^{2 π \sqrt{- 1} / 5}

is a primitive

5

-th root of unity. Moreover, the inner product is

〈 σ^{i} \hat{e}, σ^{j} \hat{e} 〉 = 3 / 512

for any

i \neq j

Lemma A.7. There is a triple

(u, v, w)

of mutually orthogonal conformal vectors in

ℬ

such that the central charges of

u, v, w

are

1 / 2, 25 / 28

25 / 28

, respectively and

u + v + w

is equal to the Virasoro element

{\tilde{ω}}^{1} + {\tilde{ω}}^{2}

U

For example,

\begin{matrix} u & = \hat{e} = \frac{7}{32} ({\tilde{ω}}^{1} + {\tilde{ω}}^{2}) + \frac{1}{32} (X^{1} + X^{2} + X^{3} + X^{4}), \end{matrix}

\begin{matrix} v & = \frac{15}{64} {\tilde{ω}}^{1} + \frac{35}{64} {\tilde{ω}}^{2} - \frac{3}{64} (X^{1} + X^{4}) + \frac{1}{64} (X^{2} + X^{3}), \end{matrix}

\begin{matrix} w & = \frac{35}{64} {\tilde{ω}}^{1} + \frac{15}{64} {\tilde{ω}}^{2} + \frac{1}{64} (X^{1} + X^{4}) - \frac{3}{64} (X^{2} + X^{3}) . \end{matrix}

\begin{matrix}  \end{matrix}

6 A

case. In this case,

L = L (5) \sim = A_{2} \oplus A_{1} \oplus A_{5}

n_{5} = 6

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{2}}

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{1}}

, and

{\tilde{ω}}^{3} \in V_{\sqrt{2} A_{5}}

are of central charge

4 / 5

1 / 2

, and

5 / 4

, respectively.

The product and the inner product in

ℬ

are given by A.1 and

\begin{matrix} \begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = \frac{2}{3} X^{1}, & ({\tilde{ω}}^{2})_{1} X^{1} = \frac{1}{2} X^{1}, & ({\tilde{ω}}^{3})_{1} X^{1} = \frac{5}{6} X^{1}, \\ ({\tilde{ω}}^{1})_{1} X^{2} = \frac{2}{3} X^{2}, & ({\tilde{ω}}^{2})_{1} X^{2} = 0, & ({\tilde{ω}}^{3})_{1} X^{2} = \frac{4}{3} X^{2}, \\ ({\tilde{ω}}^{1})_{1} X^{3} = 0, & ({\tilde{ω}}^{2})_{1} X^{3} = \frac{1}{2} X^{3}, & ({\tilde{ω}}^{3})_{1} X^{3} = \frac{3}{2} X^{3}, \\ ({\tilde{ω}}^{1})_{1} X^{4} = \frac{2}{3} X^{4}, & ({\tilde{ω}}^{2})_{1} X^{4} = 0, & ({\tilde{ω}}^{3})_{1} X^{4} = \frac{4}{3} X^{4}, \\ ({\tilde{ω}}^{1})_{1} X^{5} = \frac{2}{3} X^{5}, & ({\tilde{ω}}^{2})_{1} X^{5} = \frac{1}{2} X^{5}, & ({\tilde{ω}}^{3})_{1} X^{5} = \frac{5}{6} X^{5}, \\ (X^{1})_{1} X^{1} = 8 X^{2}, & (X^{1})_{1} X^{2} = 9 X^{3}, & (X^{1})_{1} X^{3} = 8 X^{4}, \\ (X^{1})_{1} X^{4} = 10 X^{5}, & (X^{1})_{1} X^{5} = 60 {\tilde{ω}}^{1} + 72 {\tilde{ω}}^{2} + 48 {\tilde{ω}}^{3}, & (X^{2})_{1} X^{2} = 12 X^{4}, \\ (X^{2})_{1} X^{3} = 10 X^{5}, & (X^{2})_{1} X^{4} = 75 {\tilde{ω}}^{1} + 96 {\tilde{ω}}^{3}, & (X^{2})_{1} X^{5} = 10 X^{1}, \\ (X^{3})_{1} X^{3} = 80 {\tilde{ω}}^{2} + 96 {\tilde{ω}}^{3}, & (X^{3})_{1} X^{4} = 10 X^{1}, & (X^{3})_{1} X^{5} = 8 X^{2}, \\ (X^{4})_{1} X^{4} = 12 X^{2}, & (X^{4})_{1} X^{5} = 9 X^{3}, & (X^{5})_{1} X^{5} = 8 X^{4}, \\ 〈 X^{1}, X^{5} 〉 = 36, & 〈 X^{2}, X^{4} 〉 = 45, & 〈 X^{3}, X^{3} 〉 = 40 . \end{matrix} \end{matrix}

(A.7)

Lemma A.8. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c {\tilde{ω}}^{3} + d X^{1} + e X^{2} + f X^{e} + g X^{4} + h X^{5}

be a conformal vectors in

ℬ

. Then

(a, b, c, d, e, f, g, h)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 60 d h + 75 e g - a = 0, \\ b^{2} + 40 f^{2} + 72 d h - b = 0, \\ c^{2} + 48 f^{2} + 48 d h + 96 e g - c = 0, \\ 4 a d + 3 b d + 5 c d + 60 e h + 60 f g - 6 d = 0, \\ 12 d^{2} + 18 g^{2} + 2 a e + 4 c e + 24 f h - 3 e = 0, \\ b f + 3 c f + 18 d e + 18 g h - 2 f = 0, \\ 18 e^{2} + 12 h^{2} + 2 a g + 4 c g + 24 d f - 3 g = 0, \\ 4 a h + 3 b h + 5 c h + 60 d g + 60 e f - 6 h = 0 . \end{matrix} \end{matrix}

(A.8)

The central charge of

w

is given by

\begin{matrix} \frac{4}{5} a^{2} + \frac{1}{2} b^{2} + \frac{5}{4} c^{2} + 80 f^{2} + 144 d h + 180 e g . \end{matrix}

(A.9)

Again we treat

a, b, c, d, e, f, g, h

as variables and let

ℐ

be the ideal in

C [a, b, c, d, e, f, g, h]

generated by the eight polynomials which appear on the left hand side of ( A.8 ). By using the computer algebra system Risa/Asir, we found that the ideal

ℐ

is an intersection of

112

prime ideals over the field

Q

of rational numbers and that there are totally

256

conformalvectors in

ℬ

. The number of conformal vectors whose central charges are rational numbers less than

1

are listed in Table 2.

Table 2 . Central charge (c.c.) and number of conformal vectors

c.c.	$1 / 2$	$7 / 10$	$4 / 5$	$6 / 7$	$25 / 28$	$11 / 12$	$14 / 15$	$21 / 22$
number	$7$	$9$	$7$	$14$	$5$	$6$	$6$	$6$

The following three lemmas were verified by computer.

Lemma A.9. There are exactly seven conformal vectors of central charge

1 / 2

ℬ

, namely,

{\tilde{ω}}^{2}

and

σ^{j} \hat{e}, 0 \leq j \leq 5

. Note that

σ^{j} \hat{e} = \frac{5}{32} {\tilde{ω}}^{1} + \frac{1}{8} {\tilde{ω}}^{2} + \frac{1}{4} {\tilde{ω}}^{3} + \frac{1}{32} (ξ^{j} X^{1} + ξ^{2 j} X^{2} + ξ^{3 j} X^{3} + ξ^{4 j} X^{4} + ξ^{5 j} X^{5}),

where

ξ = e^{π \sqrt{- 1} / 3}

is a primitive

6

-th root of unity. Moreover, the inner product among

σ^{j} \hat{e}, 0 \leq j \leq 5

are

〈 σ^{i} \hat{e}, σ^{j} \hat{e} 〉 = {\begin{matrix} 5 / 2^{10} & if i - j \equiv \pm 1 m o d 6, \\ 13 / 2^{10} & if i - j \equiv \pm 2 m o d 6, \\ 1 / 32 & if i - j \equiv 3 m o d 6 . \end{matrix}

We are mainly interested in mutually orthogonal conformal vectors whose sum is the Virasoro element

ω^{'} = {\tilde{ω}}^{1} + {\tilde{ω}}^{2} + {\tilde{ω}}^{3}

U

Lemma A.10. There are exactly

10

triples

(u, v, w)

of mutually orthogonal conformal vectors in

ℬ

such that the central charge of

u, v, w

are

4 / 5, 6 / 7, 25 / 28

, respectively and

u + v + w = ω^{'}

Lemma A.11. There are exactly

6

triples

(u, v, w)

of mutually orthogonal conformal vectors in

ℬ

such that the central charge of

u, v, w

are

7 / 10, 11 / 12, 14 / 15

, respectively and

u + v + w = ω^{'}

4 B

case. In this case,

L = L (6) \sim = A_{1} \oplus A_{7}

n_{6} = 4

, and the conformal vectors

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{1}}

and

{\tilde{ω}}^{2} \in V_{\sqrt{2} A_{7}}

are of central charge

1 / 2

and

7 / 5

respectively. The product and the inner product in

ℬ

are given by A.1 and

\begin{matrix} \begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = \frac{1}{2} X^{1}, & ({\tilde{ω}}^{1})_{1} X^{2} = 0, & ({\tilde{ω}}^{1})_{1} X^{3} = \frac{1}{2} X^{3}, \\ ({\tilde{ω}}^{2})_{1} X^{1} = \frac{3}{2} X^{1}, & ({\tilde{ω}}^{2})_{1} X^{2} = 2 X^{2}, & ({\tilde{ω}}^{2})_{1} X^{3} = \frac{3}{2} X^{3}, \\ (X^{1})_{1} X^{1} = 12 X^{2}, & (X^{1})_{1} X^{2} = 15 X^{3}, & (X^{1})_{1} X^{3} = 112 {\tilde{ω}}^{1} + 120 {\tilde{ω}}^{2}, \\ (X^{2})_{1} X^{2} = 200 {\tilde{ω}}^{2}, & (X^{2})_{1} X^{3} = 15 X^{1}, & (X^{3})_{1} X^{3} = 12 X^{2}, \\ 〈 X^{1}, X^{3} 〉 = 56, & 〈 X^{2}, X^{2} 〉 = 70 . \end{matrix} \end{matrix}

(A.10)

Lemma A.12. Let

w = a {\tilde{ω}}^{1} + b {\tilde{ω}}^{2} + c X^{1} + d X^{2} + e X^{3}

be a conformal vector in

ℬ

. Then

(a, b, c, d, e)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 112 c e - a = 0, & b^{2} + 100 d^{2} + 120 c e - b = 0, \\ a c + 3 b c + 30 d e - 2 c = 0, & 6 c^{2} + 6 e^{2} + 2 b d - d = 0, \\ a e + 3 b e + 30 c d - 2 e = 0 . \end{matrix} \end{matrix}

(A.11)

The central charge of

w

is given by

\frac{1}{2} a^{2} + \frac{7}{5} b^{2} + 224 c e + 140 d^{2}

The solutions

(a, b, c, d, e)

of Equation A.11 are as follows, where

ξ = e^{π \sqrt{- 1} / 4}

is a primitive

8

-th root of unity.

Central charge 1/2:

(1, 0, 0, 0, 0)

(\frac{1}{8}, \frac{5}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, \frac{1}{32} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central charge 7/10:

(0, \frac{1}{2}, 0, \pm \frac{1}{20}, 0)

(\frac{7}{8}, \frac{3}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, \frac{3}{160} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central charge

21 / 22

(\frac{7}{11}, \frac{5}{11}, \frac{1}{22} ξ^{2 j + 1}, 0, \frac{1}{22} ξ^{- (2 j + 1)})

0 \leq j \leq 3

Central charge

52 / 55

(\frac{4}{11}, \frac{6}{11}, \frac{1}{22} ξ^{2 j + 1}, 0, \frac{1}{22} ξ^{- (2 j + 1)})

0 \leq j \leq 3

Central charge

6 / 5

(1, \frac{1}{2}, 0, \pm \frac{1}{20}, 0)

(\frac{1}{8}, \frac{13}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, - \frac{3}{160} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central charge

7 / 5

(0, 1, 0, 0, 0)

(\frac{7}{8}, \frac{11}{16}, \frac{1}{32} (\sqrt{- 1})^{j}, - \frac{1}{32} (- 1)^{j}, \frac{1}{32} (- \sqrt{- 1})^{j})

0 \leq j \leq 3

Central charge

19 / 10

(1, 1, 0, 0, 0)

2 B

case. In this case,

L = L (7) \sim = D_{8}

n_{7} = 2

, and the conformal vector

{\tilde{ω}}^{1} \in V_{\sqrt{2} D_{8}}

is of central charge

1

. The product and the inner product in

ℬ

are given by A.1 and

({\tilde{ω}}^{1})_{1} X^{1} = 2 X^{1}, (X^{1})_{1} X^{1} = 512 {\tilde{ω}}^{1}, 〈 X^{1}, X^{1} 〉 = 128 .

Lemma A.13. Let

w = a {\tilde{ω}}^{1} + b X^{1}

be a conformal vector in

ℬ

. Then we have

a = 1 / 2

and

b = \pm 1 / 32

, and the central charge of

w

1 / 2

. In other words, there are exactly two conformal vectors in

ℬ

and both of them are of central charge

1 / 2

3 C

case. In this case,

L = L (8) \sim = A_{8}

n_{8} = 3

, and the conformal vector

{\tilde{ω}}^{1} \in V_{\sqrt{2} A_{8}}

is of central charge

16 / 11

. The product and the inner product in

ℬ

are given by A.1 and

\begin{matrix} \begin{matrix} ({\tilde{ω}}^{1})_{1} X^{1} = 2 X^{1}, & ({\tilde{ω}}^{1})_{1} X^{2} = 2 X^{2}, \\ (X^{1})_{1} X^{1} = 20 X^{2}, & (X^{2})_{1} X^{2} = 20 X^{1}, & (X^{1})_{1} X^{2} = 231 {\tilde{ω}}^{1}, \\ 〈 X^{1}, X^{2} 〉 = 84 . \end{matrix} \end{matrix}

(A.12)

Lemma A.14. Let

w = a {\tilde{ω}}^{1} + b X^{1} + c X^{2}

be a conformal vector in

ℬ

. Then

(a, b, c)

satisfies the following system of equations.

\begin{matrix} \begin{matrix} a^{2} + 924 b c - a = 0, & 2 a b + 10 c^{2} - b = 0, \\ 2 a c + 10 b^{2} - c = 0 . \end{matrix} \end{matrix}

(A.13)

The central charge of

w

is given by

\frac{16}{11} a^{2} + 336 b c

The solutions

(a, b, c)

of Equation A.13 are as follows, where

ξ = e^{2 π \sqrt{- 1} / 3}

is a primitive cubic root of unity.

Central charge

1 / 2

(\frac{11}{32}, \frac{1}{32} ξ^{j}, \frac{1}{32} ξ^{2 j})

j = 0, 1, 2

Central charge

21 / 22

(\frac{21}{32}, - \frac{1}{32} ξ^{j}, - \frac{1}{32} ξ^{2 j})

j = 0, 1, 2

Central charge

16 / 11

(1, 0, 0, 0)

B Characters of $W$ -algebras and the structure of $U$ for the cases of $3 C$ and $5 A$

In this appendix, we shall determine the structure of the coset subalgebra

U

defined by 2.10 for the cases of

3 C

and

5 A

. The main tool is the character of parafermion algebras.

First, let us recall a construction of parafermion algebras from [6] .

Let

Λ_{0}

and

Λ_{1}

be the fundamental weights of the affine Lie algebra

\hat{s} l_{2} (C)

. For any positive integer

ℓ

and

0 \leq j \leq ℓ

, let

ℒ (ℓ, j)

be the irreducible highest weight module of

\hat{s} l_{2} (C)

with the highest weight

(ℓ - j) Λ_{0} + j Λ_{1}

. Note that

ℒ (ℓ, 0)

has a natural VOA structure and

{ℒ (ℓ, j) | 0 \leq j \leq ℓ}

is the set of all inequivalent irreducible modules of

ℒ (ℓ, 0)

(cf. [18] ).

Now let

{A_{1}}^{ℓ} = Z ε_{1} \oplus \dots \oplus Z ε_{ℓ}

be an even lattice with

〈 ε_{i}, ε_{j} 〉 = 2 δ_{i, j}

and

V_{{A_{1}}^{ℓ}}

the lattice VOA associated with

{A_{1}}^{ℓ}

. Then

V_{{A_{1}}^{ℓ}} \sim = (V_{A_{1}})^{\otimes ℓ} \sim = ℒ (1, 0)^{\otimes ℓ}

Set

H^{(ℓ)} = ε_{1} (- 1) \cdot 1 + \dots + ε_{ℓ} (- 1) \cdot 1

E^{(ℓ)} = e^{ε_{1}} + \dots + e^{ε_{ℓ}}

, and

F^{(ℓ)} = e^{- ε_{1}} + \dots + e^{- ε_{ℓ}}

Then

C H^{(ℓ)} + C E^{(ℓ)} + C F^{(ℓ)}

forms a simple Lie algebra

s l_{2} (C)

inside the weight one subspace of

V_{{A_{1}}^{ℓ}}

. Moreover, the subVOA generated by

{H^{(ℓ)}, E^{(ℓ)}, F^{(ℓ)}}

is isomorphic to

ℒ (ℓ, 0)

(cf. [6] ).

Let

γ = ε_{1} + \dots + ε_{ℓ}

. Then

γ (- 1) \cdot 1 = H^{(ℓ)}

and it is easy to verify that

e^{γ} = \frac{1}{ℓ!} ((E^{(ℓ)})_{- 1})^{ℓ - 1} E^{(ℓ)} .

Thus

ℒ (ℓ, 0)

contains a subalgebra isomorphic to the lattice VOA

V_{Z γ}

Let

A = 〈 e^{γ} 〉

be the subgroup of

C {{A_{1}}^{ℓ}}

generated by

e^{γ}

and denote by

Ω_{ℒ (ℓ, j)}^{A}

the set of all highest weight vectors for

V_{Z γ}

ℒ (ℓ, j)

0 \leq j \leq ℓ

. It is shown in [6] that

Ω_{ℒ (ℓ, 0)}^{A} = \oplus_{k = 0}^{ℓ - 1} (Ω_{ℒ (ℓ, 0)}^{A})^{k}

is a generalized VOA and

Ω_{ℒ (ℓ, j)}^{A}

are irreducible

Ω_{ℒ (ℓ, 0)}^{A}

-modules.

Note that

(Ω_{ℒ (ℓ, 0)}^{A})^{0}

itself is a VOA, which we shall denote by

W_{ℓ}

Now let

ℒ (ℓ, j) =^{2 ℓ - 1} \oplus_{k = 0} V_{(k / 2 ℓ) γ + Z γ} \otimes W_{ℓ} (j, k)

be the decomposition of

ℒ (ℓ, j)

as a

V_{Z γ} \otimes W_{ℓ}

-module, where

W_{ℓ} (j, k)

is the multiplicity of

V_{k γ / 2 ℓ + Z γ}

ℒ (ℓ, j)

. By [6] ,

W_{ℓ} (j, k) = 0

j + k \equiv 1 m o d 2

and so

\begin{matrix} ℒ (ℓ, j) = {\begin{matrix} \oplus_{k = 0}^{ℓ - 1} V_{(k / ℓ) γ + Z γ} \otimes W_{ℓ} (j, 2 k) & if j is even, \\ \oplus_{k = 0}^{ℓ - 1} V_{((2 k + 1) / 2 ℓ) γ + Z γ} \otimes W_{ℓ} (j, 2 k + 1) & if j is odd . \end{matrix} \end{matrix}

(B.1)

Proposition B.1 (cf. Dong and Lepowsky [6] ). All

W_{ℓ} (j, k)

0 \leq j \leq ℓ

0 \leq k \leq 2 ℓ - 1

j \equiv k m o d 2

, are irreducible

W_{ℓ}

-modules.

By B.1 , we can actually compute the character of

W_{ℓ} (j, k)

by using the character of

ℒ (ℓ, j)

and

V_{(k / 2 ℓ) γ + Z γ}

. Recall that the character

c h_{M} (q)

of a module

M

of a VOA

V

is defined by

c h_{M} (q) = t r_{M} q^{L (0)}

, where

L (0) = ω_{1}

with

ω

being the Virasoro element of

V

.In Kac and Raina [23] , the following formula is proved.

\begin{matrix} t r_{ℒ (ℓ, j)} σ (- \frac{1}{2} z γ) q^{L (0) - ℓ / 8 (ℓ + 2)} = \frac{θ_{j + 1, ℓ + 2} (τ, z) - θ_{- j - 1, ℓ + 2} (τ, z)}{θ_{1, 2} (τ, z) - θ_{- 1, 2} (τ, z)}, \end{matrix}

(B.2)

where

q = e^{2 π \sqrt{- 1} τ}

σ (z γ) = e^{2 π \sqrt{- 1} z γ (0)}

for

z \in Q

, and

θ_{n, m} (τ, z) = \sum_{j \in n / 2 m + Z} e^{2 π \sqrt{- 1} m z j} q^{m j^{2}} .

By using the

q

-integers

[n]_{a} = (a^{n} - a^{- n}) / (a - a^{- 1}) = a^{n - 1} + a^{n - 3} + \dots + a^{- n + 3} + a^{- n + 1}

for

n \in Z

, we can rewrite the formula B.2 in the following form.

\begin{matrix} t r_{ℒ (ℓ, 0)} σ (- \frac{1}{2} z γ) q^{L (0)} = \frac{1 + \sum_{m \in Z \ {0}} [2 (ℓ + 2) m + 1]_{e^{π \sqrt{- 1} z}} \cdot q^{(ℓ + 2) m^{2} + m}}{1 + \sum_{n \in Z \ {0}} [4 n + 1]_{e^{π \sqrt{- 1} z}} \cdot q^{2 n^{2} + n}} . \end{matrix}

(B.3)

The character of

V_{(k / 2 ℓ) γ + Z γ}

is also well known (cf. [17] ). It is given by

\begin{matrix} c h_{V_{(k / 2 ℓ) γ + Z γ}} (q) = t r_{V_{(k / 2 ℓ) γ + Z γ}} q^{L (0)} = \frac{1}{\prod_{n = 1}^{\infty} (1 - q^{n})} θ_{(k / 2 ℓ) γ + Z γ} (q), \end{matrix}

(B.4)

where

θ_{(k / 2 ℓ) γ + Z γ} (q) = \sum_{α \in (k / 2 ℓ) γ + Z γ} q^{〈 α, α 〉 / 2}

Next, we consider another construction of the VOA

W_{ℓ} = W_{ℓ} (0, 0)

given in [32] by using the lattice VOA

V_{\sqrt{2} A_{ℓ - 1}}

, namely,

W_{ℓ} \sim = {u \in V_{\sqrt{2} A_{ℓ - 1}} | s (A_{ℓ - 1})_{1} u = 0},

where

s (A_{ℓ - 1}) = \frac{1}{2 (ℓ + 2)} \sum_{α \in Φ^{+} (A_{ℓ - 1})} (α (- 1)^{2} \cdot 1 - 2 (e^{\sqrt{2} α} + e^{- \sqrt{2} α}))

and

Φ^{+} (A_{ℓ - 1})

is the set of all positive roots of the lattice of type

A_{ℓ - 1}

(cf. 2.1 ).

Let

N = s p a n_{Z} {ε_{1} - ε_{2}, ε_{2} - ε_{3}, \dots, ε_{ℓ - 1} - ε_{ℓ}} \subset {A_{1}}^{ℓ}

. Then

N \sim = \sqrt{2} A_{ℓ - 1}

. Moreover,

{A_{1}}^{ℓ} =^{ℓ - 1} ⋃_{k = 0} (k ε_{ℓ} + N + Z γ) =^{ℓ - 1} ⋃_{k = 0} ((k η + N) + (\frac{k}{ℓ} γ + Z γ)),

where

γ = ε_{1} + \dots + ε_{ℓ}

and

η = \frac{1}{ℓ} (- ε_{1} - \dots - ε_{ℓ - 1} + (ℓ - 1) ε_{ℓ})

. Indeed,

| {A_{1}}^{ℓ} / (N + Z γ) | = ℓ

and

η + γ / ℓ = ε_{ℓ}

. Note that

〈 η, γ 〉 = 0

and

〈 N, γ 〉 = 0

By the above argument, we obtain the following proposition.

Proposition B.2. For any positive integer

ℓ

and

k = 0, . . ., ℓ - 1

W_{ℓ} (0, 2 k) \sim = {u \in V_{k η + N} | s (A_{ℓ - 1})_{1} u = 0} .

B.1 The case for $3 C$

We consider an embedding of a

\sqrt{2} E_{8}

lattice into

R^{9}

by using

N

and

η

for the case

ℓ = 9

. Indeed,

\sqrt{2} α_{j}

0 \leq j \leq 8

, where

α_{0}, α_{1}, \dots, α_{8}

are the nine nodes in the extended

E_{8}

diagram 2.2 , can be realized as

\sqrt{2} α_{j} = ε_{j + 1} - ε_{j + 2}

0 \leq j \leq 7

, and

\sqrt{2} α_{8} = 3 η + (ε_{7} - ε_{8}) + 2 (ε_{8} - ε_{9}) \in 3 η + N

. Then

N = \sqrt{2} L (8)

, where

L (8)

is the sublattice spanned by

α_{0}, α_{1}, \dots, α_{7}

(cf. Section 3). Hence

\sqrt{2} E_{8} = N \cup (3 η + N) \cup (6 η + N)

. This implies that

V_{\sqrt{2} E_{8}} = V_{\sqrt{2} A_{8}} \oplus V_{3 η + \sqrt{2} A_{8}} \oplus V_{6 η + \sqrt{2} A_{8}} .

Furthermore, by the definition 2.10 of the coset subalgebra,

U = {u \in V_{\sqrt{2} E_{8}} | s (A_{8})_{1} u = 0}

. So by Proposition B.2 , we have the following result.

Proposition B.3. As a module of

W_{9}

\begin{matrix} U \sim = W_{9} (0, 0) \oplus W_{9} (0, 6) \oplus W_{9} (0, 12) . \end{matrix}

(B.5)

Remark B.4. It is easy to see that

W_{9} (0, 6)

and

W_{9} (0, 12)

have the same character by their construction. In fact,

W_{9} (0, 6)

is the dual module of

W_{9} (0, 12)

. Thus by B.5 ,

c h_{U} (q) \equiv c h_{W_{9}} (q) m o d 2

As discussed in Subsection 3.9 , the coset subalgebra

U

can be decomposed into a direct sum of irreducible modules of

L (1 / 2, 0) \otimes L (21 / 22, 0)

in the following form.

\begin{matrix} \begin{matrix} U & \sim = m_{1} L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0) \oplus m_{2} L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 8) & \oplus m_{3} L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{7}{2}) \oplus m_{4} L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{45}{2}) & \oplus m_{5} L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{31}{16}) \oplus m_{6} L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{175}{16}), \end{matrix} \end{matrix}

(B.6)

where

m_{j} \in Z

denotes the multiplicity of each summand. We want to show that

m_{j} = 1

for

j = 1, \dots, 6

By B.1 ,

ℒ (9, 0) = \oplus_{k = 0}^{8} V_{(k / 9) γ + Z γ} \otimes W_{9} (0, 2 k)

. Let

z = j / 9

in B.2 . Then

σ (- \frac{j}{18} γ)

acts on

V_{(k / 9) γ + Z γ}

as a scalar

e^{- 2 π \sqrt{- 1} k j / 9}

and acts on

W_{9} (0, 2 k)

as the identity. Hence

t r_{ℒ (9, 0)} σ (- \frac{j}{18} γ) q^{L (0)} = \sum_{k = 0}^{8} e^{- 2 π \sqrt{- 1} k j / 9} c h_{V_{(k / 9) γ + Z γ}} (q) c h_{W_{9} (0, 2 k)} (q)

and thus

c h_{V_{(k / 9) γ + Z γ}} (q) c h_{W_{9} (0, 2 k)} (q) = \frac{1}{9} \sum_{j = 0}^{8} t r_{ℒ (9, 0)} σ (- \frac{j}{18} γ) q^{L (0)} e^{2 π \sqrt{- 1} k j / 9} .

Now, consider the case for

k = 0

3

, and

6

. By B.2 and B.4 , we can calculate that

\begin{matrix} \begin{matrix} c h_{W_{9} (0, 0)} (q) & = 1 + q^{2} + 2 q^{3} + 4 q^{4} + 6 q^{5} + 11 q^{6} + 16 q^{7} + 27 q^{8} \\ + 40 q^{9} + 62 q^{10} + 90 q^{11} + 137 q^{12} + 194 q^{13} + 284 q^{14} \\ + 400 q^{15} + 569 q^{16} + 788 q^{17} + 1102 q^{18} + 1504 q^{19} \\ + 2066 q^{20} + 2792 q^{21} + 3776 q^{22} + 5046 q^{23} + \dots, \end{matrix} \end{matrix}

\begin{matrix} \begin{matrix} c h_{W_{9} (0, 6)} (q) = c h_{W_{9} (0, 12)} (q) & = q^{2} + q^{3} + 3 q^{4} + 5 q^{5} + 9 q^{6} + 14 q^{7} \\ + 25 q^{8} + 36 q^{9} + 58 q^{10} + 86 q^{11} + \dots . \end{matrix} \end{matrix}

Let

[h_{1}, h_{2}]_{n}

be the weight

n

subspace of

L (1 / 2, h_{1}) \otimes L (21 / 22, h_{2})

. Its dimension for small

n

is as follows.

\begin{matrix} dim [0, 0]_{2} = 2, & dim [0, 0]_{4} = 5, & dim [\frac{1}{16}, \frac{31}{16}]_{4} = 4, \end{matrix}

\begin{matrix} dim [0, 0]_{8} = 27, & dim [\frac{1}{16}, \frac{31}{16}]_{8} = 36, & dim [\frac{1}{2}, \frac{7}{2}]_{8} = 13, \end{matrix}

\begin{matrix} dim [0, 0]_{11} = 75, & dim [\frac{1}{16}, \frac{31}{16}]_{11} = 130, & dim [\frac{1}{2}, \frac{7}{2}]_{11} = 51, \end{matrix}

\begin{matrix} dim [0, 8]_{11} = 5, & dim [0, 0]_{23} = 3073, & dim [\frac{1}{16}, \frac{31}{16}]_{23} = 7040, \end{matrix}

\begin{matrix} dim [\frac{1}{2}, \frac{7}{2}]_{23} = 3510, & dim [0, 8]_{23} = 946, & dim [\frac{1}{16}, \frac{175}{16}]_{23} = 490 . \end{matrix}

\begin{matrix}  \end{matrix}

Comparing the characters of B.5 and B.6 , we have

m_{1} = m_{2} = m_{3} = m_{5} = m_{6} = 1

and

m_{4} \equiv 1 m o d 2

. Moreover,

m_{4} \leq 1

since

L (1 / 2, 1 / 2) \otimes L (21 / 22, 45 / 2)

is a simple current module. Hence we obtain the following theorem.

Theorem B.5. As a module of

L (1 / 2, 0) \otimes L (21 / 22, 0)

\begin{matrix} U & \sim = L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 8) & \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{7}{2}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{45}{2}) & \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{31}{16}) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{175}{16}) . \end{matrix}

B.2 The case for $5 A$

In this case we consider an embedding of a

\sqrt{2} E_{8}

lattice into

R^{10}

by using two

N

's and two

η

's for the case

ℓ = 5

. Let

ε_{i}, ε_{i}^{'} \in R^{10}

1 \leq i \leq 5

, be such that

〈 ε_{i}, ε_{j} 〉 = 〈 ε_{i}^{'}, ε_{j}^{'} 〉 = 2 δ_{i j}

and

〈 ε_{i}, ε_{j}^{'} 〉 = 0

. Set

N = s p a n_{Z} {ε_{i} - ε_{i + 1}; 1 \leq i \leq 4}, N^{'} = s p a n_{Z} {ε_{i}^{'} - ε_{i + 1}^{'}; 1 \leq i \leq 4},

η = \frac{1}{5} (- ε_{1} - ε_{2} - ε_{3} - ε_{4} + 4 ε_{5}), η^{'} = \frac{1}{5} (- ε_{1}^{'} - ε_{2}^{'} - ε_{3}^{'} - ε_{4}^{'} + 4 ε_{5}^{'}) .

Let

\sqrt{2} α_{j} = ε_{j + 1} - ε_{j + 2}

0 \leq j \leq 3

\sqrt{2} α_{7} = ε_{1}^{'} - ε_{2}^{'}

\sqrt{2} α_{6} = ε_{2}^{'} - ε_{3}^{'}

\sqrt{2} α_{5} = ε_{3}^{'} - ε_{4}^{'}

\sqrt{2} α_{8} = ε_{4}^{'} - ε_{5}^{'}

\sqrt{2} α_{4} = η + 2 η^{'} + (ε_{4}^{'} - ε_{5}^{'}) \in η + 2 η^{'} + N^{'}

. Then

α_{0}, α_{1}, \dots, α_{8}

are the nine nodes in the extended

E_{8}

diagram 2.2 and

N \oplus N^{'} = \sqrt{2} L (4)

, where

L (4)

is the sublattice spanned by

α_{j}

0 \leq j \leq 8

j \neq 4

(cf. Section 3). Hence

| \sqrt{2} E_{8} / N \oplus N^{'} | = 5

and

\begin{matrix} \sqrt{2} E_{8} & = N \oplus N^{'} \cup (η + N) \oplus (2 η^{'} + N^{'}) \cup (2 η + N) \oplus (4 η^{'} + N^{'}) \end{matrix}

\begin{matrix} \cup (3 η + N) \oplus (η^{'} + N^{'}) \cup (4 η + N) \oplus (3 η^{'} + N^{'}) . \end{matrix}

\begin{matrix}  \end{matrix}

This implies that

\begin{matrix} V_{\sqrt{2} E_{8}} & = (V_{\sqrt{2} A_{4}} \otimes V_{\sqrt{2} A_{4}}) \oplus (V_{η + \sqrt{2} A_{4}} \otimes V_{2 η + \sqrt{2} A_{4}}) \oplus (V_{2 η + \sqrt{2} A_{4}} \otimes V_{4 η + \sqrt{2} A_{4}}) \end{matrix}

\begin{matrix} \oplus (V_{3 η + \sqrt{2} A_{4}} \otimes V_{η + \sqrt{2} A_{4}}) \oplus (V_{4 η + \sqrt{2} A_{4}} \otimes V_{3 η + \sqrt{2} A_{4}}) . \end{matrix}

\begin{matrix}  \end{matrix}

Furthermore, the following proposition holds by the definition 2.10 of the coset subalgebra

U

and Proposition B.2 .

Proposition B.6. As a module of

W_{5} \otimes W_{5}

\begin{matrix} \begin{matrix} U & \sim = (W_{5} (0, 0) \otimes W_{5} (0, 0)) \oplus (W_{5} (0, 2) \otimes W_{5} (0, 4)) \oplus (W_{5} (0, 4) \otimes W_{5} (0, 8)) & \oplus (W_{5} (0, 6) \otimes W_{5} (0, 2)) \oplus (W_{5} (0, 8) \otimes W_{5} (0, 6)) . \end{matrix} \end{matrix}

(B.7)

Note that

c h_{V_{η + \sqrt{2} A_{4}}} (q) = c h_{V_{4 η + \sqrt{2} A_{4}}} (q)

and

c h_{V_{2 η + \sqrt{2} A_{4}}} (q) = c h_{V_{3 η + \sqrt{2} A_{4}}} (q)

. Thus by the construction, it is also easy to see that

c h_{W_{5} (0, 2) \otimes W_{5} (0, 4)} (q) = c h_{W_{5} (0, 4) \otimes W_{5} (0, 8)} (q) = c h_{W_{5} (0, 6) \otimes W_{5} (0, 2)} (q) = c h_{W_{5} (0, 8) \otimes W_{5} (0, 6)} (q) .

Hence

\begin{matrix} c h_{U} (q) = c h_{W_{5} (0, 0)} (q)^{2} + 4 c h_{W_{5} (0, 2)} (q) c h_{W_{5} (0, 4)} (q) . \end{matrix}

(B.8)

By B.1 ,

ℒ (5, 0) = \oplus_{k = 0}^{4} V_{(k / 5) γ + Z γ} \otimes W_{5} (0, 2 k)

. Let

z = j / 5

in B.2 . Then

σ (- \frac{j}{10} γ)

acts on

V_{(k / 5) γ + Z γ}

as a scalar

e^{- 2 π \sqrt{- 1} k j / 5}

and acts on

W_{5} (0, 2 k)

as the identity. Then arguing as in the case for

3 C

, we get

c h_{V_{(k / 5) γ + Z γ}} (q) c h_{W_{5} (0, 2 k)} (q) = \frac{1}{5} \sum_{j = 0}^{4} e^{2 π \sqrt{- 1} k j / 5} t r_{ℒ (5, 0)} σ (- \frac{j}{10} γ) q^{L (0)} .

Now using B.2 and B.4 , we can calculate that

\begin{matrix} c h_{W_{5} (0, 0)} (τ) = & 1 + q^{2} + 2 q^{3} + 4 q^{4} + 6 q^{5} + 10 q^{6} + 14 q^{7} + 23 q^{8} + 32 q^{9} + 48 q^{10} \end{matrix}

\begin{matrix} + 66 q^{11} + 96 q^{12} + 130 q^{13} + 183 q^{14} + 246 q^{15} + \dots, \end{matrix}

\begin{matrix} c h_{W_{5} (0, 2)} (τ) = & q^{4 / 5} (1 + q + 2 q^{2} + 3 q^{3} + 6 q^{4} + 8 q^{5} + 14 q^{6} + 20 q^{7} + 31 q^{8} + 43 q^{9} \end{matrix}

\begin{matrix} + 64 q^{10} + 87 q^{11} + 125 q^{12} + 169 q^{13} + 234 q^{14} + 313 q^{15} + \dots), \end{matrix}

\begin{matrix} c h_{W_{5} (0, 4)} (τ) = & q^{6 / 5} (1 + q + 3 q^{2} + 4 q^{3} + 7 q^{4} + 10 q^{5} + 17 q^{6} + 23 q^{7} + 36 q^{8} + 50 q^{9} \end{matrix}

\begin{matrix} + 73 q^{10} + 100 q^{11} + 142 q^{12} + 191 q^{13} + 265 q^{14} + 353 q^{15} + \dots) . \end{matrix}

In Subsection 3.5 we have shown that

U

contains a subalgebra isomorphic to

L (1 / 2, 0) \otimes L (25 / 28, 0) \otimes L (25 / 28, 0)

. We also know all irreducible modules

L (1 / 2, h_{1}) \otimes L (25 / 28, h_{2}) \otimes L (25 / 28, h_{3})

with integral weights. Note that

U

is a direct sum of these irreducible modules. Comparing the characters of those irreducible modules and B.8 , we can verify the following theorem.

Theorem B.7. As a module of

L (1 / 2, 0) \otimes L (25 / 28, 0) \otimes L (25 / 28, 0)

\begin{matrix} U \sim = L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{5}{32}) \otimes L (\frac{25}{28}, \frac{57}{32}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{57}{32}) \otimes L (\frac{25}{28}, \frac{5}{32}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{3}{4}) \otimes L (\frac{25}{28}, \frac{3}{4}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{3}{4}) \otimes L (\frac{25}{28}, \frac{13}{4}) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{13}{4}) \otimes L (\frac{25}{28}, \frac{3}{4}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{57}{32}) \otimes L (\frac{25}{28}, \frac{165}{32}) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{165}{32}) \otimes L (\frac{25}{28}, \frac{57}{32}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{13}{4}) \otimes L (\frac{25}{28}, \frac{13}{4}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, \frac{15}{2}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{15}{2}) \otimes L (\frac{25}{28}, 0) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{15}{2}) \otimes L (\frac{25}{28}, \frac{15}{2}) . \end{matrix}

B.3 Character of the 6A-algebra

Let

U_{6 A}

be the 6A-algebra, the coset subalgebra of

V_{\sqrt{2} E_{8}}

constructed from the coset decomposition

E_{8} / (A_{5} \oplus A_{2} \oplus A_{1})

. We show here that

\begin{matrix} \begin{matrix} U_{6 A} ≃ U_{3 A} (0) \otimes L (25 / 28, 0) \oplus U_{3 A} (5 / 7) \otimes L (25 / 28, 9 / 7) \\ \oplus U_{3 A} (1 / 7) \otimes L (25 / 28, 34 / 7), \end{matrix} \end{matrix}

(B.9)

where

U_{3 A} (0)

denotes the 3A-algebra and

U_{3 A} (h)

h = 5 / 7, 1 / 7

, denote the irreducible

U_{3 A} (0)

-modules whose top weights are equal to

h

. We show the above isomorphism by computing the

q

-character of

U_{6 A}

. Recall the notion of

W_{ℓ}

-algebras in Appendix B. By considering an isometric embedding

\sqrt{2} A_{5} \oplus \sqrt{2} A_{2} \oplus \sqrt{2} A_{1} ↪ A_{1}^{6} \oplus A_{1}^{3} \oplus \sqrt{2} A_{1}

, one can easily verify that

\begin{matrix} U_{6 A} ≃ W_{6} (0, 0) \otimes W_{3} (0, 0) \otimes L (1 / 2, 0) \oplus W_{6} (0, 2) \otimes W_{3} (0, 4) \otimes L (1 / 2, 1 / 2) \end{matrix}

\begin{matrix} \oplus W_{6} (0, 4) \otimes W_{3} (0, 2) \otimes L (1 / 2, 0) \oplus W_{6} (0, 6) \otimes W_{3} (0, 0) \otimes L (1 / 2, 1 / 2) \end{matrix}

\begin{matrix} \oplus W_{6} (0, 8) \otimes W_{3} (0, 4) \otimes L (1 / 2, 0) \oplus W_{6} (0, 10) \otimes W_{3} (0, 2) \otimes L (1 / 2, 1 / 2), \end{matrix}

where we have used the fact that the lattice VOA

V_{\sqrt{2} A_{1}}

is isomorphic to the code VOA associated to a binary code

{(00), (11)}

. Since

W_{3} (0, 0) ≃ L (4 / 5, 0) \oplus L (4 / 5, 3)

and

W_{3} (0, 2) ≃ W_{3} (0, 4) ≃ L (4 / 5, 2 / 3)

L (4 / 5, 0)

-modules, we only need to compute the characters of

W_{6} (0, 2 s)

0 \leq s \leq 5

. However, by noticing the dual module relations, we know that

c h_{W_{6} (0, 2)} (q) = c h_{W_{6} (0, 10)} (q)

and

c h_{W_{6} (0, 4)} (q) = c h_{W_{6} (0, 8)} (q)

. Therefore, we should compute characters of

W_{6} (0, 2 s)

for

s = 0, 1, 2, 3

. By a method in the preceding, we can obtain the following results.

\begin{matrix} c h_{W_{6} (0, 0)} = 1 + q^{2} + 2 q^{3} + 4 q^{4} + 6 q^{5} + 11 q^{6} + \dots, \end{matrix}

\begin{matrix} c h_{W_{6} (0, 2)} = q^{5 / 6} + q^{11 / 6} + 2 q^{17 / 6} + 3 q^{23 / 6} + 6 q^{29 / 6} + 9 q^{35 / 6} + \dots, \end{matrix}

\begin{matrix} c h_{W_{6} (0, 4)} = q^{4 / 3} + q^{7 / 3} + 3 q^{10 / 3} + 4 q^{13 / 3} + 8 q^{16 / 3} + \dots, \end{matrix}

\begin{matrix} c h_{W_{6} (0, 6)} = q^{3 / 2} + q^{5 / 2} + 3 q^{7 / 2} + 5 q^{9 / 2} + 8 q^{11 / 2} + \dots . \end{matrix}

Then by comparing characters, we can establish the desired isomorphism B.9 .

B.3.1 Highest weight vector of weight $(0, 1 / 7, 34 / 7)$

In this section, we prove that the 6A-algebra is generated by its weight two subspace as a vertex operator algebra. We use the notation as in section 5.6 of our preprint. Let

w^{1} = {\tilde{ω}}^{1} = \tilde{ω} (A_{2})

w^{2} = \tilde{ω} (E_{6})

and

w^{3} = \tilde{ω} (A_{1}) + \tilde{ω} (A_{5}) - w^{2}

. Then

w^{1}, w^{2}, w^{3}

are mutually orthogonal conformal vector of central charges 4/5, 6/7 and 25/28, respectively, and the sum

w^{1} + w^{2} + w^{3}

is the Virasoro vector of

U_{6 A}

and the sum

w^{1} + w^{2}

is the Virasoro vector of

U_{3 A} (0)

in the decomposition B.9 . We can write down

w^{i}

explicitly as follows.

w^{1} = \tilde{ω} (A_{2}), w^{2} = \frac{2}{7} \tilde{ω} (A_{1}) + \frac{4}{7} \tilde{ω} (A_{5}) + \frac{1}{14} X^{3}, w^{3} = \frac{5}{7} \tilde{ω} (A_{1}) + \frac{3}{7} \tilde{ω} (A_{5}) - \frac{1}{14} X^{3} .

Let

W

be the subalgebra of

U_{6 A}

generated by its weight two subspace. It is clear that

W

contains both

U_{3 A} (0) \otimes L (25 / 28, 0)

and

U_{3 A} (5 / 7) \otimes L (25 / 28, 9 / 7)

. So we only have to show that

W

contains an irreducible

V i r (w^{1}) \otimes V i r (w^{2}) \otimes V i r (w^{3})

-module of highest weight

(0, 1 / 7, 34 / 7)

. Let

v

be a highest weight vector for

V i r (w^{1}) \otimes V i r (w^{2}) \otimes V i r (w^{3})

with highest weight

(0, 5 / 7, 9 / 7)

. Since such a vector is unique up to linearity, we may take

v

as follows.

v = 400 \tilde{ω} (A_{1}) - 16 \tilde{ω} (A_{5}) + X^{3} .

Then by the fusion rules for the unitary Virasoro VOAs, we know that

v_{- 2} v

is contained in the weight 5 subspace of

L (4 / 5, 0) \otimes L (6 / 7, 5 / 7) \otimes L (25 / 28, 9 / 7) \oplus L (4 / 5, 0) \otimes L (6 / 7, 1 / 7) \otimes L (25 / 28, 34 / 7) .

Assume that

W

is a proper subalgebra. Then

W

is isomorphic to

U_{3 A} (0) \otimes L (25 / 28, 0) \oplus U_{3 A} (5 / 7) \otimes L (25 / 28, 9 / 7)

and

v_{- 2} v

is in the weight 5 subspace of

L (4 / 5, 0) \otimes L (6 / 7, 5 / 7) \otimes L (25 / 28, 9 / 7) .

In this case, by computing inner products, we must have the following equality:

\begin{matrix} v_{- 2} v = & - \frac{132}{7} w_{- 1}^{2} w_{0}^{2} v + \frac{1870}{49} w_{- 2}^{2} v - \frac{80}{21} w_{- 1}^{3} w_{0}^{3} v + \frac{1128}{49} w_{- 2}^{3} v + \frac{288}{23} w_{0}^{2} w_{- 1}^{3} v \end{matrix}

\begin{matrix} - \frac{40}{23} w_{0}^{2} w_{0}^{3} w_{0}^{3} v + \frac{440}{19} w_{- 1}^{2} w_{0}^{3} v - \frac{198}{19} w_{0}^{2} w_{0}^{2} w_{0}^{3} v . \end{matrix}

However, the squared length of the right hand side is 1913313600/437, whereas that of the left hand side is 10209600. This means that there is a highest weight vector of weight

(0, 1 / 7, 34 / 7)

v_{- 2} v

, and we can understand that the difference

10209600 - 1913313600 / 437 = 2548281600 / 437

is the squared length of the highest weight vector. Thus the 6A-algebra

U_{6 A}

is generated by its weight two subspace.

C Classification for the irreducible modules of the $3 C$ and $5 A$ algebras

In this appendix, we shall give the classification of all the irreducible modules for the

3 C

and

5 A

algebras. For convenience, we introduce the following notation. Let

V

be a VOA and

M

its module. For subsets

A \subset V

and

B \subset M

, we set

A \cdot B = s p a n {a_{n} v | a \in A, v \in B, n \in Z} .

It is shown in Lemma 3.12 of [36] that for

a, b \in V

v \in M

and

p, q \in Z

, there are

m, n \geq 0

such that

\begin{matrix} a_{p} b_{q} v = \sum_{i = 0}^{m} \sum_{j = 0}^{n} (\binom{p - n}{i}) (\binom{n}{j}) (a_{p - n - i + j} b)_{q + n + i - j} v . \end{matrix}

(C.1)

In particular,

V \cdot (V \cdot v) = V \cdot v

for any

v \in M

so that

V \cdot v

is a submodule of

M

Recall the notion of the fusion products (cf. [22] [35] ). We denote by

M^{1} ⊠_{V} M^{2}

the fusion product of

V

-modules

M^{1}

and

M^{2}

. The basic result (loc. cit.) is that the fusion product exists if

V

is rational.

We shall study an extension of a rational VOA by an irreducible module which is not a simple current module. Our settings are as follows.

Let

V

be a simple rational VOA and

W

an irreducible

V

-module such that

\begin{matrix} dim I_{V} (\binom{M^{2}}{W M^{1}}) \leq 1 \end{matrix}

(C.2)

for any irreducible

V

-modules

M^{1}

and

M^{2}

, where

I_{V} (\binom{M^{2}}{W M^{1}})

denotes the space of

V

-intertwining operators of type

(\binom{M^{2}}{W M^{1}})

. Assume that the space

\tilde{V} = V \oplus W

has a simple VOA structure

(\tilde{V}, \tilde{Y} (\cdot, z))

which is an extension of

V

such that for any

u, v \in W

\tilde{Y} (u, z) v = (ℐ (u, z) + J (u, z)) v,

where

ℐ (\cdot, z) \in I_{V} (\binom{V}{W W}) and J (\cdot, z) \in I_{V} (\binom{W}{W W})

are non-zero intertwining operators. Note that the simplicity of

\tilde{V}

implies that

V

and

W

are inequivalent

V

-modules. For, if

V

and

W

are isomorphic, then the isomorphic image of the vacuum vector of

V

W

is a vacuum-like vector (cf. [34] ). Since the vertex operator of a vacuum-like vector commutes with all the vertex operators on

\tilde{V}

and

\tilde{V}

is simple, every vacuum-like vector is a scalar multiple of the vacuum vector. Thus

V

and

W

are inequivalent. Then it follows from C.1 that

\tilde{V} = W \cdot W

. By fixing one VOA structure on

\tilde{V}

, we shall show that the module structure of certain types of irreducible

\tilde{V}

-modules are uniquely determined by their

V

-module structures.

Lemma C.1. Let

M

be an irreducible

\tilde{V}

-module. Assume that

M

contains an irreducible

V

-submodule

M^{0}

such that

M

is a direct sum

n M^{0}

n

copies of

M^{0}

as a

V

-module. Then

M = M^{0}

, i.e.,

n = 1

Proof. By C.1 , we know that

W \cdot M^{0}

is a

V

-submodule of

M

which is a direct sum of some copies of

M^{0}

. Since

\tilde{V}

is simple, it is clear that

W \cdot M^{0} \neq 0

(cf. Proposition 11.9 of [6] ). By the universal property of the fusion product (cf. [22] [35] ), there exists a

V

-epimorphism from

W ⊠_{V} M^{0}

onto

W \cdot M^{0}

. Since

dim I_{V} (\binom{M^{0}}{W M^{0}}) = 1

by C.2 ,

W ⊠_{V} M^{0}

contains a

V

-submodule isomorphic to

M^{0}

with multiplicity one. Therefore,

W \cdot M^{0}

is an irreducible

V

-submodule isomorphic to

M^{0}

. Similarly,

W \cdot (W \cdot M^{0})

is also an irreducible

V

-submodule of

M

. If

W \cdot M^{0} = M^{0}

, then

\tilde{V} \cdot M^{0} = (V \cdot M^{0}) + (W \cdot M^{0}) = M^{0}

as a

V

-module so that

M = M^{0}

and we are done. Assume that

W \cdot M^{0} \neq M^{0}

. In this case

M^{0} + (W \cdot M^{0}) = M^{0} \oplus (W \cdot M^{0})

since both

M^{0}

and

W \cdot M^{0}

are irreducible

V

-submodules.

Then by the irreducibility we have

M = \tilde{V} \cdot M^{0} = M^{0} \oplus (W \cdot M^{0})

. Consider

(W \cdot W) \cdot M^{0}

As we have seen,

W \cdot W = V \oplus W = \tilde{V}

so that

(W \cdot W) \cdot M^{0} = M

. By the associativity formula

\begin{matrix} (a_{m} b)_{n} = \sum_{i = 0}^{\infty} (- 1)^{i} (\binom{m}{i}) {a_{m - i} b_{n + i} - (- 1)^{m} b_{m + n - i} a_{i}}, \end{matrix}

(C.3)

we see that

(W \cdot W) \cdot M^{0} \subset W \cdot (W \cdot M^{0})

. Since

W \cdot (W \cdot M^{0})

is an irreducible

V

-submodule, so is

(W \cdot W) \cdot M^{0} = \tilde{V} \cdot M^{0}

. But this is a contradiction. Hence

W \cdot M^{0} = M^{0}

and

M = M^{0}

. □

Lemma C.2. Let

M

be an irreducible

\tilde{V}

-module which is also irreducible as a

V

-module. Then there is exactly one irreducible

\tilde{V}

-module structure on

M

up to isomorphism.

Proof. Suppose that

(M, Y_{M} (\cdot, z))

and

(M, Y_{M}^{'} (\cdot, z))

are two irreducible

\tilde{V}

-module structures on an irreducible

V

-module

M

. Without loss, we may assume that

Y_{M} (u, z) = Y_{M}^{'} (u, z)

for all

u \in V

. Since

M

is irreducible and

dim I_{V} (\binom{M}{W M}) \leq 1

by C.2 , there exists

λ \neq 0

such that

Y_{M}^{'} (a, z) v = λ Y_{M} (a, z) v for any a \in W, v \in M .

By the associativity, for any

a, b \in W

and

v \in M

, there exists

k > 0

such that

\begin{matrix} \begin{matrix} (z_{0} + z_{2})^{k} Y_{M} (a, z_{0} + z_{2}) Y_{M} (b, z_{2}) v = & (z_{0} + z_{2})^{k} Y_{M} (Y_{\tilde{V}} (a, z_{0}) b, z_{2}) v & = & (z_{0} + z_{2})^{k} Y_{M} (ℐ (a, z_{0}) b + J (a, z_{0}) b, z_{2}) v . \end{matrix} \end{matrix}

(C.4)

Similarly, we have

\begin{matrix} (z_{0} + z_{2})^{k} Y_{M}^{'} (a, z_{0} + z_{2}) Y_{M}^{'} (b, z_{2}) v = & (z_{0} + z_{2})^{k} Y_{M}^{'} (Y_{\tilde{V}} (a, z_{0}) b, z_{2}) v & = & (z_{0} + z_{2})^{k} Y_{M}^{'} (ℐ (a, z_{0}) b + J (a, z_{0}) b, z_{2}) v \end{matrix}

and thus

\begin{matrix} \begin{matrix} λ^{2} (z_{0} + z_{2})^{k} Y_{M} (a, z_{0} + z_{2}) Y_{M} (b, z_{2}) v & = & (z_{0} + z_{2})^{k} (Y_{M} (ℐ (a, z_{0}) b, z_{2}) v + λ Y_{M} (J (a, z_{0}) b, z_{2}) v) . \end{matrix} \end{matrix}

(C.5)

By C.4 and C.5 , we have

\begin{matrix} (z_{0} + z_{2})^{k} Y_{M} ((λ^{2} - 1) ℐ (a, z_{0}) b + (λ^{2} - λ) J (a, z_{0}) b, z_{2}) v = 0 . \end{matrix}

(C.6)

We note that

Y_{M} ((λ^{2} - 1) ℐ (a, z_{0}) b + (λ^{2} - λ) J (a, z_{0}) b, z_{2}) v \in M ((z_{0})) [[z_{2}, z_{2}^{- 1}]] .

Thus C.6 implies that

(λ^{2} - 1) ℐ (a, z_{0}) b = 0 and (λ^{2} - λ) J (a, z_{0}) b = 0 .

Therefore, we have

λ^{2} = λ = 1

. Hence there is only one irreducible

\tilde{V}

-module structure on

M

. □

Lemma C.3. Let

M

be an irreducible

\tilde{V}

-module. Assume that there are two inequivalent irreducible

V

-submodules

M^{0}

and

M^{1}

M

such that

M

is isomorphic to

m M^{0} \oplus n M^{1}

with

m, n > 0

as a

V

-module. If

dim I_{V} (\binom{M^{0}}{W M^{0}}) = 0

, then

m = n = 1

and there is exactly one

\tilde{V}

-module structure on

M

up to isomorphism.

Proof. First, we shall show that

W \cdot M^{0} = M^{1}

and

M = W \cdot M^{1} = M^{0} \oplus M^{1}

. It is clear from C.1 that both

W \cdot M^{0}

and

W \cdot M^{1}

are

V

-submodules of

M

which are non-zero by Proposition 11.9 of [6] . By the universal property of the fusion product, we have a

V

-epimorphism from

W ⊠_{V} M^{0}

onto

W \cdot M^{0}

. By our setting C.2 and the assumption

dim I_{V} (\binom{M^{0}}{W M^{0}}) = 0

W \cdot M^{0}

is an irreducible

V

-submodule isomorphic to

M^{1}

. Thus

M^{0} + (W \cdot M^{0})

is a direct sum. Since

M

is an irreducible

\tilde{V}

-module,

M = \tilde{V} \cdot M^{0} = (V \cdot M^{0}) + (W \cdot M^{0}) = M^{0} \oplus (W \cdot M^{0})

. This proves

W \cdot M^{0} = M^{1}

and

M = M^{0} \oplus M^{1}

Then

W \cdot M^{1} = W \cdot (W \cdot M^{0}) = (W \cdot W) \cdot M^{0} = \tilde{V} \cdot M^{0} = M

by C.1 and C.3 .

Let us suppose that

(M, Y_{M} (\cdot, z))

and

(M, Y_{M}^{'} (\cdot, z))

are two irreducible

\tilde{V}

-module structures on the

V

-module

M = M^{0} \oplus M^{1}

. Without loss, we may assume that

Y_{M} (u, z) = Y_{M}^{'} (u, z)

for all

u \in V

. Let

a \in W

v^{0} \in M^{0}

and

v^{1} \in M^{1}

be arbitrary. As we haveshown,

W \cdot M^{0} = M^{1}

under both structures. Thus there is a non-zero

V

-intertwining operator

T (\cdot, z)

of type

(\binom{M^{1}}{W M^{0}})

and a scalar

γ \neq 0

such that

Y_{M} (a, z) v^{0} = T (a, z) v^{0} and Y_{M}^{'} (a, z) v^{0} = γ T (a, z) v^{0} .

We have also shown that

W \cdot M^{1} = M^{0} \oplus M^{1}

under both structures. Thus there exist non-zero

V

-intertwining operators

I (\cdot, z) \in I_{V} (\binom{M^{0}}{W M^{1}})

J (\cdot, z) \in I_{V} (\binom{M^{1}}{W M^{1}})

and scalars

μ, λ \neq 0

such that

Y_{M} (a, z) v^{1} = I (a, z) v^{1} + J (a, z) v^{1} and Y_{M}^{'} (a, z) v^{1} = μ I (a, z) v^{1} + λ J (a, z) v^{1} .

Then

Y_{M} (u + a, z) (v^{0} + v^{1}) = Y_{M} (u, z) v^{0} + I (a, z) v^{1} + T (a, z) v^{0} + (Y_{M} (u, z) + J (a, z)) v^{1},

where

u \in V

. Note that

Y_{M} (u, z) v^{0} \in M^{0} ((z))

I (a, z) v^{1} \in M^{0} ((z))

T (a, z) v^{0} \in M^{1} ((z))

and

(Y_{M} (u, z) + J (a, z)) v^{1} \in M^{1} ((z))

. Hence the vertex operator

Y_{M} (\cdot, z)

can be written as the following matrix form.

Y_{M} ([\begin{matrix} u \\ a \end{matrix}], z) [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] = [\begin{matrix} Y_{M} (u, z) & I (a, z) \\ T (a, z) & Y_{M} (u, z) + J (a, z) \end{matrix}] [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] .

Likewise,

Y_{M}^{'} ([\begin{matrix} u \\ a \end{matrix}], z) [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] = [\begin{matrix} Y_{M} (u, z) & μ I (a, z) \\ γ T (a, z) & Y_{M} (u, z) + λ J (a, z) \end{matrix}] [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] .

For simplicity, we use the following notation. For

f (z_{0}, z_{2}), g (z_{0}, z_{2}) \in M [[z_{0}, z_{0}^{- 1}, z_{2}, z_{2}^{- 1}]]

f (z_{0}, z_{2}) \sim g (z_{0}, z_{2})

means that there exists

k > 0

such that

(z_{0} + z_{2})^{k} f (z_{0}, z_{2}) = (z_{0} + z_{2})^{k} g (z_{0}, z_{2}) .

Then, by the associativity on

Y_{M}^{'} (\cdot, z)

, for any

a, b \in W

we have the following system of equations:

\begin{matrix} [\begin{matrix} Y_{M} (ℐ (a, z_{0}) b, z_{2}) & μ I (J (a, z_{0}) b, z_{2}) \\ γ T (J (a, z_{0}) b, z_{2}) & Y_{M} (ℐ (a, z_{0}) b, z_{2}) + λ J (J (a, z_{0}) b, z_{2}) \end{matrix}] & \sim & [\begin{matrix} μ γ I (a, z_{0} + z_{2}) T (b, z_{2}) & μ λ I (a, z_{0} + z_{2}) J (b, z_{2}) \\ λ γ J (a, z_{0} + z_{2}) T (b, z_{2}) & μ γ T (a, z_{0} + z_{2}) I (b, z_{2}) + λ^{2} J (a, z_{0} + z_{2}) J (b, z_{2}) \end{matrix}] \end{matrix}

Similarly, by the associativity on

Y_{M} (\cdot, z)

, we have

\begin{matrix} [\begin{matrix} Y_{M} (ℐ (a, z_{0}) b, z_{2}) & I (J (a, z_{0}) b, z_{2}) \\ T (J (a, z_{0}) b, z_{2}) & Y_{M} (ℐ (a, z_{0}) b, z_{2}) + J (J (a, z_{0}) b, z_{2}) \end{matrix}] & \sim & [\begin{matrix} I (a, z_{0} + z_{2}) T (b, z_{2}) & I (a, z_{0} + z_{2}) J (b, z_{2}) \\ J (a, z_{0} + z_{2}) T (b, z_{2}) & T (a, z_{0} + z_{2}) I (b, z_{2}) + J (a, z_{0} + z_{2}) J (b, z_{2}) \end{matrix}] . \end{matrix}

Then by a similar argument as in the proof of Lemma C.2 , we have

μ γ = 1

μ = μ λ

and

λ γ = γ

, from which we conclude that

λ = 1

and

μ = 1 / γ

. Thus we obtain the following relation:

Y_{M}^{'} ([\begin{matrix} u \\ a \end{matrix}], z) [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] = [\begin{matrix} Y_{M} (u, z) & \frac{1}{γ} I (a, z) \\ γ T (a, z) & Y_{M} (u, z) + J (a, z) \end{matrix}] [\begin{matrix} v^{0} \\ v^{1} \end{matrix}] .

Now define

ψ : M^{0} \oplus M^{1} \to M^{0} \oplus M^{1}

ψ (v^{0}) = \frac{1}{γ} v^{0} for v^{0} \in M^{0}, ψ (v^{1}) = v^{1} for v^{1} \in M^{1} .

Then

\begin{matrix} ψ (Y_{M} ([\begin{matrix} u \\ a \end{matrix}], z) [\begin{matrix} v^{0} \\ v^{1} \end{matrix}]) = & [\begin{matrix} \frac{1}{γ} (Y_{M} (u, z) v^{0} + I (a, z) v^{1}) \\ T (a, z) v^{0} + Y_{M} (u, z) v^{1} + J (a, z) v^{1} \end{matrix}] & = & [\begin{matrix} Y_{M} (u, z) ψ (v^{0}) + μ I (a, z) ψ (v^{1}) \\ γ T (a, z) ψ (v^{0}) + Y_{M} (u, z) ψ (v^{1}) + J (a, z) ψ (v^{1}) \end{matrix}] & = & Y_{M}^{'} ([\begin{matrix} u \\ a \end{matrix}], z) [\begin{matrix} ψ (v^{0}) \\ ψ (v^{1}) \end{matrix}] . \end{matrix}

Hence

ψ

induces a

\tilde{V}

-isomorphism from

(M, Y_{M} (\cdot, z))

(M, Y_{M}^{'} (\cdot, z))

. □

C.1 Irreducible modules for the $3 C$ algebra

First we shall classify all irreducible modules for the

3 C

-algebra

U_{3 C}

. Recall that

\begin{matrix} U_{3 C} & \sim = L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 8) & \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{7}{2}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{21}{22}, \frac{45}{2}) & \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{31}{16}) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{21}{22}, \frac{175}{16}) \end{matrix}

as a module of

L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)

. For simplicity, we shall use

[h_{1}, h_{2}]

to denote the module

L (\frac{1}{2}, h_{1}) \otimes L (\frac{21}{22}, h_{2})

Let

e \in U_{3 C}

be the Virasoro element of the VOA

L (\frac{1}{2}, 0)

and

τ_{e}

the corresponding Miyamoto involution. Then the corresponding fixed point subalgebra of

τ_{e}

U_{3 C}^{τ_{e}} = [0, 0] \oplus [0, 8] \oplus [\frac{1}{2}, \frac{7}{2}] \oplus [\frac{1}{2}, \frac{45}{2}] .

Set

V = [0, 0] \oplus [\frac{1}{2}, \frac{45}{2}]

and

W = [0, 8] \oplus [\frac{1}{2}, \frac{7}{2}]

. Then

V

is a subalgebra of

U_{3 C}^{τ_{e}}

and

W

is an irreducible

V

-submodule of

U_{3 C}^{τ_{e}}

. It is shown in the proof of Theorem 3.36 that

W \cdot W = V \oplus W = U_{3 C}^{τ_{e}}

. Therefore, we can use the representation theory for

V

to classify irreducible

U_{3 C}^{τ_{e}}

-modules. Note that

V

is a

Z_{2}

-graded simple current extension of

[0, 0]

so that

V

is rational. Moreover, irreducible

V

-modules and their fusion rules are easily determined (cf. [27, 29, 47] ).

Lemma C.4 (cf. [27, 29, 47] ). The irreducible modules for

V = [0, 0] \oplus [\frac{1}{2}, \frac{45}{2}]

are given as follows.

\begin{matrix} [0, h_{i, j}] \oplus [\frac{1}{2}, h_{i, 12 - j}], & [\frac{1}{16}, h_{i, k}] \oplus [\frac{1}{16}, h_{i, 12 - k}], & [\frac{1}{16}, h_{i, 6}]^{+} and [\frac{1}{16}, h_{i, 6}]^{-}, \end{matrix}

where

1 \leq i \leq 5

j = 1, 3, 5, 7, 9, 11

and

k = 2, 4

Note that all irreducible modules are integrally graded, i.e., if

M

is an irreducible module, then

M = \oplus_{n \in Z} M_{λ + n} with L (0) |_{M_{λ + n}} = λ + n

for some

λ \in C

. Thus, by using the fusion rules of

V

-modules and the integrally graded condition above, we can list up possible irreducible

U_{3 C}^{τ_{e}}

-modules as follows.

Lemma C.5. Let

M

be an irreducible

U_{3 C}^{τ_{e}}

-module. Assume that

M

does not contain

[\frac{1}{16}, h_{i, 6}]

1 \leq i \leq 5

, as

L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)

-submodules. Then as a module of

L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)

M

is isomorphic to one of the following.

\begin{matrix} [0, h_{i, 1}] \oplus [\frac{1}{2}, h_{i, 5}] \oplus [0, h_{i, 7}] \oplus [\frac{1}{2}, h_{i, 11}], \end{matrix}

\begin{matrix} [\frac{1}{2}, h_{i, 1}] \oplus [0, h_{i, 5}] \oplus [\frac{1}{2}, h_{i, 7}] \oplus [0, h_{i, 11}], \end{matrix}

\begin{matrix} [\frac{1}{16}, h_{i, 4}] \oplus [\frac{1}{16}, h_{i, 8}], \end{matrix}

where

1 \leq i \leq 5

. Moreover, the

U_{3 C}^{τ_{e}}

-module structure on

M

is uniquely determined by its

V

-module structure.

Proof. Let $M$ be an irreducible $U_{3 C}^{τ_{e}}$ -module. Since $U_{3 C}^{τ_{e}} = V \oplus W$ and $V$ is rational, $M$ is a direct sum of irreducible $V$ -modules. By the list of irreducible $V$ -modules shown in Lemma C.4 and the fusion rules of $V i r (e)$ -modules, we have the following two cases: As a $V i r (e)$ -module, $M$ is a sum of $L (\frac{1}{2}, \frac{1}{16})$ or $M$ does not contain $L (\frac{1}{2}, \frac{1}{16})$ . In the former case, by the fusion rules of $V$ -modules and the integrally graded condition, we see that all irreducible $V$ -submodules of $M$ are mutually isomorphic and they are isomorphic to one of $[\frac{1}{16}, h_{i, 4}] \oplus [\frac{1}{16}, h_{i, 8}]$ with $1 \leq i \leq 5$ . Then by Lemmas C.1 and C.2 , $M$ is in fact irreducible as a $V$ -module and its $U_{3 C}^{τ_{e}}$ -module structure is uniquely determined by its $V$ -module structure. Hence $M$ is as in the assertion.
If $M$ does not contain $L (\frac{1}{2}, \frac{1}{16})$ as $V i r (e)$ -submodules, then by the list of irreducible $V$ -modules shown in Lemma C.4 and the fusion rules of $V$ -modules, we are in the situation as in Lemma C.3 and the integrally graded condition leads to that $M$ is in the list of the assertion. In this case the uniqueness of $U_{3 C}^{τ_{e}}$ -module structure is already shown in Lemma C.3 . □

Finally, we have the following classification theorem.

Theorem C.6. There are exactly five irreducible

U_{3 C}

-modules. As

L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)

-modules, they are of the following form:

\begin{matrix} U (0) & \sim = [0, 0] \oplus [0, 8] \oplus [\frac{1}{2}, \frac{7}{2}] \oplus [\frac{1}{2}, \frac{45}{2}] \oplus [\frac{1}{16}, \frac{31}{16}] \oplus [\frac{1}{16}, \frac{175}{16}] (= U_{3 C}), \end{matrix}

\begin{matrix} U (2) & \sim = [0, \frac{13}{11}] \oplus [0, \frac{35}{11}] \oplus [\frac{1}{2}, \frac{15}{22}] \oplus [\frac{1}{2}, \frac{301}{22}] \oplus [\frac{1}{16}, \frac{21}{176}] \oplus [\frac{1}{16}, \frac{901}{176}], \end{matrix}

\begin{matrix} U (4) & \sim = [0, \frac{6}{11}] \oplus [0, \frac{50}{11}] \oplus [\frac{1}{2}, \frac{1}{22}] \oplus [\frac{1}{2}, \frac{155}{22}] \oplus [\frac{1}{16}, \frac{85}{176}] \oplus [\frac{1}{16}, \frac{261}{176}], \end{matrix}

\begin{matrix} U (6) & \sim = [0, \frac{1}{11}] \oplus [0, \frac{111}{11}] \oplus [\frac{1}{2}, \frac{35}{22}] \oplus [\frac{1}{2}, \frac{57}{22}] \oplus [\frac{1}{16}, \frac{5}{176}] \oplus [\frac{1}{16}, \frac{533}{176}], \end{matrix}

\begin{matrix} U (8) & \sim = [0, \frac{20}{11}] \oplus [0, \frac{196}{11}] \oplus [\frac{1}{2}, \frac{7}{22}] \oplus [\frac{1}{2}, \frac{117}{22}] \oplus [\frac{1}{16}, \frac{133}{176}] \oplus [\frac{1}{16}, \frac{1365}{176}] . \end{matrix}

\begin{matrix}  \end{matrix}

Proof. Set $U_{3 C}^{-} = [\frac{1}{16}, \frac{31}{16}] \oplus [\frac{1}{16}, \frac{175}{16}]$ . Then $τ_{e} \in A u t (U_{3 C})$ acts on $U_{3 C}^{-}$ as $- 1$ and $U_{3 C} = U_{3 C}^{τ_{e}} \oplus U_{3 C}^{-}$ is a $Z_{2}$ -graded extension of $U_{3 C}^{τ_{e}}$ . Let $M$ be an irreducible $U_{3 C}$ -module. Denote by $M^{0}$ the sum of irreducible $V i r (e)$ -submodules of $M$ isomorphic to $L (\frac{1}{2}, 0)$ or $L (\frac{1}{2}, \frac{1}{2})$ and by $M^{1}$ the sum of irreducible $V i r (e)$ -submodules of $M$ isomorphic to $L (\frac{1}{2}, \frac{1}{16})$ . Then $M = M^{0} \oplus M^{1}$ . By the fusion rules of $L (\frac{1}{2}, 0)$ -modules, $M^{0}$ and $M^{1}$ are inequivalent irreducible $U_{3 C}^{τ_{e}}$ -submodules and $M$ carries the following $Z_{2}$ -grading: $U_{3 C}^{-} \cdot M^{0} = M^{1}$ and $U_{3 C}^{-} \cdot M^{1} = M^{0}$ . In Lemma C.5 we have classified irreducible $U_{3 C}^{τ_{e}}$ -modules having no $V i r (e)$ -submodules isomorphic to $L (\frac{1}{2}, \frac{1}{16})$ . Then by the integrally graded condition, we see that $M$ cannot contain $[\frac{1}{16}, h_{i, 6}]$ , $1 \leq i \leq 5$ as $L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)$ -submodules.
Now thanks to Lemma C.5 we can classify the possible pairs $(M^{0}, M^{1})$ of irreducible $U_{3 C}^{τ_{e}}$ -modules such that $M^{0} \oplus M^{1}$ are integrally graded. As a result, we see that $M$ is isomorphic to one of $U (2 i)$ , $i = 0, 1, . . ., 4$ , as an $L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)$ -module. We already know that all $U (2 i)$ , $i = 0, 1, . . ., 4$ , appear as $U_{3 C}$ -submodules of $V_{\sqrt{2} E_{8}}$ . Thus there exist irreducible $U_{3 C}$ -modules of the form $U (2 i)$ . It remains to show that there is only one irreducible $U_{3 C}$ -module structure on each $U (2 i)$ . Let $U (2 i) = M^{0} \oplus M^{1}$ be the decomposition above. It is clear that both $M^{0}$ and $M^{1}$ are self-dual $U_{3 C}^{τ_{e}}$ -modules. By case by case verifications, we can deduce the fusion rule $dim I_{U_{3 C}^{τ_{e}}} (\binom{M^{1}}{U_{3 C}^{-} M^{0}}) = 1$ from the fusion rules of $L (\frac{1}{2}, 0) \otimes L (\frac{21}{22}, 0)$ -modules. Since both $M^{0}$ and $M^{1}$ are self-dual, we also have $dim I_{U_{3 C}^{τ_{e}}} (\binom{M^{0}}{U_{3 C}^{-} M^{1}}) = 1$ . Then by a standard argument as in [27, 29, 47] we can easily show that there exists only one irreducible $U_{3 C}$ -module structure on $U (2 i) = M^{0} \oplus M^{1}$ , since $U_{3 C} = U_{3 C}^{τ_{e}} \oplus U_{3 C}^{-}$ is a $Z_{2}$ -graded extension of $U_{3 C}^{τ_{e}}$ . □

C.2 Irreducible modules for the $5 A$ algebra

Next we shall classify the irreducible modules for the

5 A

-algebra

U_{5 A}

. The argument used here is exactly the same as in the case of

U_{3 C}

so that we omit details in this case.

Recall that

\begin{matrix} U_{5 A} \sim = L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{5}{32}) \otimes L (\frac{25}{28}, \frac{57}{32}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{57}{32}) \otimes L (\frac{25}{28}, \frac{5}{32}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{3}{4}) \otimes L (\frac{25}{28}, \frac{3}{4}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{3}{4}) \otimes L (\frac{25}{28}, \frac{13}{4}) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{13}{4}) \otimes L (\frac{25}{28}, \frac{3}{4}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{57}{32}) \otimes L (\frac{25}{28}, \frac{165}{32}) \oplus L (\frac{1}{2}, \frac{1}{16}) \otimes L (\frac{25}{28}, \frac{165}{32}) \otimes L (\frac{25}{28}, \frac{57}{32}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{13}{4}) \otimes L (\frac{25}{28}, \frac{13}{4}) \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, \frac{15}{2}) \end{matrix}

\begin{matrix} \oplus L (\frac{1}{2}, \frac{1}{2}) \otimes L (\frac{25}{28}, \frac{15}{2}) \otimes L (\frac{25}{28}, 0) \oplus L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, \frac{15}{2}) \otimes L (\frac{25}{28}, \frac{15}{2}) . \end{matrix}

Again, we shall use

e \in U_{5 A}

to denote the Virasoro element of the VOA

L (1 / 2, 0)

and

τ_{e}

the corresponding Miyamoto involution. We shall also use

[h_{1}, h_{2}, h_{3}]

to denote the module of the form

L (\frac{1}{2}, h_{1}) \otimes L (\frac{25}{28}, h_{2}) \otimes L (\frac{25}{28}, h_{3})

Note that the fixed point subalgebra of

τ_{e}

is as follows.

\begin{matrix} U_{5 A}^{τ_{e}} \sim = & [0, 0, 0] \oplus [0, \frac{15}{2}, \frac{15}{2}] \oplus [\frac{1}{2}, \frac{15}{2}, 0] \oplus [\frac{1}{2}, 0, \frac{15}{2}] & \oplus [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] \oplus [\frac{1}{2}, \frac{13}{4}, \frac{13}{4}] \oplus [\frac{1}{2}, \frac{3}{4}, \frac{3}{4}] . \end{matrix}

It contains a subalgebra

V = [0, 0, 0] \oplus [0, \frac{15}{2}, \frac{15}{2}] \oplus [\frac{1}{2}, \frac{15}{2}, 0] \oplus [\frac{1}{2}, 0, \frac{15}{2}],

which is a

(Z_{2} \times Z_{2})

-graded simple current extension of

[0, 0, 0]

. Set

W = [0, \frac{3}{4}, \frac{13}{4}] \oplus [0, \frac{13}{4}, \frac{3}{4}] \oplus [\frac{1}{2}, \frac{13}{4}, \frac{13}{4}] \oplus [\frac{1}{2}, \frac{3}{4}, \frac{3}{4}] .

Then

W

is an irreducible

V

-submodule of

U_{5 A}^{τ_{e}}

and we have a decomposition

U_{5 A}^{τ_{e}} = V \oplus W

It is shown in Lemma 3.16 that

W \cdot W = V \oplus W = U_{5 A}^{τ_{e}}

. Therefore, we can use the representation theory of

V

to classify irreducible

U_{5 A}

-modules.

Lemma C.7 (cf. [27, 29, 47] ). The irreducible modules for

V

are given as follows.

\begin{matrix} [0, h_{i, m}, h_{k, n}] \oplus [\frac{1}{2}, h_{i, 8 - m}, h_{k, n}] \oplus [\frac{1}{2}, h_{i, m}, h_{k, 8 - n}] \oplus [0, h_{i, 8 - m}, h_{k, 8 - n}], \end{matrix}

\begin{matrix} [\frac{1}{16}, h_{i, j}, h_{k, ℓ}] \oplus [\frac{1}{16}, h_{i, 8 - j}, h_{k, ℓ}] \oplus [\frac{1}{16}, h_{i, j}, h_{k, 8 - ℓ}] \oplus [\frac{1}{16}, h_{i, 8 - j}, h_{k, 8 - ℓ}], \end{matrix}

\begin{matrix} ([\frac{1}{16}, h_{i, 4}, h_{k, ℓ}] \oplus [\frac{1}{16}, h_{i, 4}, h_{k, 8 - ℓ}])^{\pm} or ([\frac{1}{16}, h_{i, j}, h_{k, 4}] \oplus [\frac{1}{16}, h_{i, 8 - j}, h_{k, 4}])^{\pm}, \end{matrix}

\begin{matrix} [\frac{1}{16}, h_{i, 4}, h_{k, 4}] \otimes Q, \end{matrix}

where

1 \leq i, k \leq 3

m, n = 1, 3, 5, 7

j, ℓ = 2, 6

and

Q

is the unique 2-dimensional irreducible module of the quaternion group of order 8.

Now by using the fusion rules and the fact that the weights of irreducible modules are integrally graded, we have

Lemma C.8. Let

M

be an irreducible module for

U_{5 A}^{τ_{e}}

. Assume that

M

does not contain

[\frac{1}{16}, h_{p, 4}, h_{q, 4}]

1 \leq p, q \leq 3

, as

L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0)

-submodules.

Then, as an module of

L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0)

M

is isomorphic to one of the following, where

i, k = 1, 2

3

\begin{matrix} [0, h_{i, 1}, h_{k, 1}] \oplus [\frac{1}{2}, h_{i, 7}, h_{k, 1}] \oplus [\frac{1}{2}, h_{i, 1}, h_{k, 7}] \oplus [0, h_{i, 7}, h_{k, 7}] \end{matrix}

\begin{matrix} \oplus [0, h_{i, 3}, h_{k, 5}] \oplus [\frac{1}{2}, h_{i, 3}, h_{k, 3}] \oplus [\frac{1}{2}, h_{i, 5}, h_{k, 5}] \oplus [0, h_{i, 5}, h_{k, 3}], \end{matrix}

\begin{matrix} [0, h_{i, 1}, h_{k, 7}] \oplus [\frac{1}{2}, h_{i, 1}, h_{k, 1}] \oplus [\frac{1}{2}, h_{i, 7}, h_{k, 7}] \oplus [0, h_{i, 7}, h_{k, 1}] \end{matrix}

\begin{matrix} \oplus [0, h_{i, 3}, h_{k, 3}] \oplus [\frac{1}{2}, h_{i, 3}, h_{k, 5}] \oplus [\frac{1}{2}, h_{i, 5}, h_{k, 3}] \oplus [0, h_{i, 5}, h_{k, 5}], \end{matrix}

\begin{matrix} [\frac{1}{16}, h_{i, 2}, h_{k, 6}] \oplus [\frac{1}{16}, h_{i, 6}, h_{k, 6}] \oplus [\frac{1}{16}, h_{i, 6}, h_{k, 2}] \oplus [\frac{1}{16}, h_{i, 2}, h_{k, 2}], \end{matrix}

\begin{matrix} [\frac{1}{16}, h_{i, 4}, h_{k, 2}] \oplus [\frac{1}{16}, h_{i, 4}, h_{k, 6}] \oplus [\frac{1}{16}, h_{i, 6}, h_{k, 4}] \oplus [\frac{1}{16}, h_{i, 2}, h_{k, 4}] . \end{matrix}

Moreover, the

U_{5 A}^{τ_{e}}

-module structure on

M

is uniquely determined by its

V

-module structure.

Finally by using the same method for the

3 C

case, we have the following theorem.

Theorem C.9. There are exactly nine irreducible modules

U (i, j), i, j = 1, 3, 5

for

U_{5 A}

. As

L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0)

-modules, they are of the following form.

\begin{matrix} U (i, j) & \sim = [0, h_{i, 1}, h_{j, 1}] \oplus [0, h_{i, 3}, h_{j, 5}] \oplus [0, h_{i, 5}, h_{j, 3}] \oplus [0, h_{i, 7}, h_{j, 7}] & \oplus [\frac{1}{2}, h_{i, 1}, h_{j, 7}] \oplus [\frac{1}{2}, h_{i, 3}, h_{j, 3}] \oplus [\frac{1}{2}, h_{i, 5}, h_{j, 5}] \oplus [\frac{1}{2}, h_{i, 7}, h_{j, 1}] & \oplus [\frac{1}{16}, h_{i, 2}, h_{j, 4}] \oplus [\frac{1}{16}, h_{i, 4}, h_{j, 2}] \oplus [\frac{1}{16}, h_{i, 6}, h_{j, 4}] \oplus [\frac{1}{16}, h_{i, 4}, h_{j, 6}] . \end{matrix}

Proof. Let $M$ be an irreducible $U_{5 A}$ -module. By the integrally graded condition, we know that $M$ cannot contain $[\frac{1}{16}, h_{i, 4}, h_{k, 4}]$ , $1 \leq i, k \leq 3$ , as $L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0)$ -submodules.
Then by Lemma C.8 and the integrally graded condition, $M$ is isomorphic to one of $U (i, j)$ , $i, j = 1, 3, 5$ , as an $L (\frac{1}{2}, 0) \otimes L (\frac{25}{28}, 0) \otimes L (\frac{25}{28}, 0)$ -module. We already know that all $U (i, j)$ , $i, j = 1, 3, 5$ , appear as $U_{5 A}$ -submodules of $V_{\sqrt{2} E_{8}}$ . Thus there exist irreducible $U_{5 A}$ -modules of the form $U (i, j)$ . It remains to show the uniqueness of the irreducible $U_{5 A}$ -modulestructure on each $U (i, j)$ . By a similar argument as in the case of $U_{3 C}$ -modules, we can establish the uniqueness. □

References

T. Abe, Fusion rules for the charge conjugate orbifold, J. Algebra 242 (2001), 624–655.
J. H. Conway, A simple construction for the Fisher-Griess Monster group, Invent. Math. 79 (1985), 513-540.
J. H. Conway and N.J.A. Sloane, Sphere Packings, Lattices and Groups, Springer-Verlag, 1988.
C. Dong, Vertex algebras associated with even lattices, J. Algebra 161 (1993), 245-265.
C. Dong and R. L. Griess Jr., Rank One Lattice Type Vertex Operator Algebras and Their Automorphism Groups, J. Algebra 208,1998, 262–275.
C. Dong and J. Lepowsky, Generalized vertex algebras and relative vertex operators, Progress in Math. Vol. 112, Birkhäuser, Boston 1993.
C. Dong, H. Li, G. Mason and S.P. Norton, Associative subalgebras of Griess algebra and related topics, Proc. of the Conference on the Monster and Lie algebra at the Ohio State University, May 1996, ed. by J. Ferrar and K. Harada, Walter de Gruyter, Berlin New York, 1998, 27–42.
C. Dong, C. Lam and H. Yamada, Decomposition of the vertex operator algebra $V_{\sqrt{2} A_{3}}$ , J. Algebra 222 (1999), 500-510.
C. Dong, C. Lam, and H. Yamada, Decomposition of the vertex operator algebra $V_{\sqrt{2} D_{l}}$ . Commun. Contemp. Math. 3 (2001), no. 1, 137–151.
C. Dong and G. Mason, On Quantum Galois Theory, Duke Math. J. 86 (1997), no. 2, 305-321.
C. Dong and G. Mason, The construction of the moonshine module as a $Z_{p}$ -orbifold. Mathematical aspects of conformal and topological field theories and quantum groups (South Hadley, MA, 1992), Contemp. Math., 175, Amer. Math. Soc., Providence, RI, 1994, 37–52.
C. Dong, G. Mason and Y. Zhu, Discrete series of the Virasoro algebra and the moonshine module, Pro. Symp. Pure. Math., American Math. Soc. 56 II (1994), 295-316.
B. L. Feigin and D. B. Fuchs, Verma modules over the Virasoro algebra, Topology (Leningrad, 1982), Lecture Notes in Math., 1060, 230–245, Springer, Berlin-New York, 1984.
A. J. Feingold, I. B. Frenkel and J. F.X. Ries, Spinor construction of vertex operator algebras, triality, and $E_{8}^{(1)}$ , Contemp. Math. 121 (1991).
A. J. Feingold, J. F.X. Ries and M. Weiner, Spinor construction of the $c = \frac{1}{2}$ minimal model. Moonshine, the Monster, and related topics (South Hadley, MA, 1994), Contemp. Math. 193, Amer. Math. Soc., Providence, RI, 1996, 45–92.
I. B. Frenkel, Y. Huang and J. Lepowsky, On axiomatic approaches to vertex operator algebras and modules, Mem. Amer. Math. Soc. 104, 1993.
I. B. Frenkel, J. Lepowsky, and A. Meurman,Vertex Operator Algebras and the Monster, Pure and Applied Math., Vol. 134, Academic Press, 1988.
I. B. Frenkel and Y. Zhu, Vertex operator algebras associated to representations of affine and Virasoro algebras, Duke Math. J. 66(1992), 123-168.
G. Glauberman and S. P. Norton, On McKay's connection between the affine $E_{8}$ diagram and the Monster, CRM Proceedings and Lecture Notes, Vol. 30, Amer. Math. Soc., Providence, 2001, 37–42.
R. Griess, The Friendly Giant, Invent. Math. 69 (1982), 1-102.
P. Goddard, A. Kent and D. Olive, Unitary representations of the Virasoro and super-Virasoro algebras, Comm. Math. Phys. 103 (1986), 105-119.
Y.-Z. Huang and J. Lepowsky, A theory of tensor products for module categories for a vertex operator algebra, III, J. Pure Appl. Alg. 100 (1995), 141-171.
V. Kac and A. K. Raina, Bombary Lectures on Highest weight representations of infinite dimensional Lie algebra, Adv. Ser. Math. Phys. Vol. 2, World Scientific, 1987.
M. Kitazume, C. Lam and H. Yamada, Decomposition of the Moonshine vertex operator algebra as Virasoro modules, J. Algebra, 226 (2000), 893-919.
M. Kitazume, C.H. Lam, and H. Yamada, $3$ -state Potts model, Moonshine vertex operator algebra and 3A-elements of the Monster group, IMRN, 2003, No. 23, 1269 1303.
M. Kitazume, M. Miyamoto and H. Yamada, Ternary codes and vertex operator algebras, J. Algebra, 223 (2000), 379-395.
C. Lam, Induced modules for orbifold vertex operator algebras, J. Math. Soc. of Japan, 53 (2001), no. 3, 541-557.
C. Lam, Lattice vertex operator algebra $V_{\sqrt{2} E_{8}}$ and an algebra of Miyamoto of central charge $\frac{1}{2} + \frac{21}{22}$ , RIMS Kokyuroku 1327, Kyoto University, Japan, June 2003, 159–169.
C.H. Lam, N. Lam, and H. Yamauchi, Extension of unitary Virasoro vertex operator algebra by a simple module, IMRN 2003, No. 11, 577 611.
C. Lam and H. Yamada, $Z_{2} \times Z_{2}$ codes and vertex operator algebras, J. Algebra 224 (2000), 268-291.
C. Lam and H. Yamada, Tricrtical $3$ -state Potts model and vertex operator algebras constructed from ternary codes, Comm. Algebra, 32 (2004), 4197-4220.
C. Lam and H. Yamada, Decomposition of the lattice vertex operator algebra $V_{\sqrt{2} A_{l}}$ , J. Algebra, 272 (2004), 614-624.
C. Lam, H. Yamada and H. Yamauchi, Vertex operator algebras, extended $E_{8}$ diagram, and McKay's observation on the Monster simple group, to appear in Trans. Amer. Math. Soc.
H. Li, Symmetric invariant bilinear forms on vertex operator algebras, J. Pure and Appl. Algbera 96 (1994), 279–297.
H. Li, An analogue of the Hom functor and a generalized nuclear democracy theorem, Duke Math. J. 93 (1998), 73–114.
H. Li, The regular representation, Zhu's $A (V)$ -theory and induced modules, J. Algebra 238 (2001), 159–193.
A. Matsuo, Norton's trace formulae for the Griess algebra of a vertex operator agebra with larger symmetry, Comm. Math. Phys. 224 (2001), 565–591.
J. McKay, Graphs, singularities, and finite groups, Proc. Symp. Pure Math., Vol. 37, Amer. Math. Soc., Providence, 1980, 183–186.
M. Miyamoto, Griess algebras and conformal vectors in vertex operator algebras, J. Algebra 179 (1996), 523-548.
M. Miyamoto, Binary codes and vertex operator (super)algebras, J. Algebra 181 (1996), 207-222.
M. Miyamoto, $3$ -state Potts model and automorphism of vertex operator algebra of order $3$ , J. Algebra 239 (2001), 56–76.
M. Miyamoto, VOAs generated by two conformal vectors whose $τ$ -involutions generate $S_{3}$ , J. Algebra, 268 (2003), no. 2, 653–671. .
M. Miyamoto, A new construction of the Moonshine vertex operator algebras over the real number field, Ann. of Math., 159 (2004), no. 2, 535–596.
S. Sakuma and H. Yamauchi, Vertex operator algebra with automorphism group $S_{3}$ , J. Algebra 267 (2003), 272–297.
H. Shimakura, Decompositions of the moonshine module with respect to subVOAs associated to codes over $Z_{2 k}$ , J. Algebra 251 (2002), 308–322.
K. Tanabe and H. Yamada, The fixed point subalgebra of a lattice vertex operator algebra by an automorphism of order three, preprint.
H. Yamauchi, Module category of simple current extensions of vertex operator algebras, J. Pure Appl. Algebra 189 (2004) 315–328.
W. Wang, Rationality of Virasoro vertex operator algebras, Duke Math. J. IMRN, Vol. 71, No. 1 (1993), 197-211.
A.B. Zamolodchikov and V.A. Fateev, Nonlocal (parafermion) currents in two dimensional conformal quantum field theory and self-dual critical points in $Z_{N}$ -symmetric statistical systems, Sov. Phys. JETP 62(1985), 215-225.

(C.H. Lam) Department of Mathematics, National Cheng Kung University, Tainan, Taiwan 701 E-mail address : chlam@mail.ncku.edu.tw (H. Yamada) Department of Mathematics, Hitotsubashi University, Kunitachi, Tokyo 186-8601, Japan E-mail address : yamada@math.hit-u.ac.jp (H. Yamauchi) Graduate School of Mathematical Sciences, The University of Tokyo, Tokyo, 153-8914, Japan E-mail address : yamauchi@ms.u-tokyo.ac.jp

Ching Hung Lam † , Hiromichi Yamada ‡ ,

Hiroshi Yamauchi §