Twisted modules for vertex algebras associated with vertex algebroids, , November 27, 2006

1 Introduction

For most of the important examples of vertex operator algebras

V = \prod_{n \in Z} V_{(n)}

graded by the

L (0)

-weight (see [FLM, FHL]), the

Z

-grading satisfies the condition that

V_{(n)} = 0

for

n < 0

and

V_{(0)} = C 1

where

1

is the vacuum vector. For a vertex operator algebra

V

with this special property, the homogeneous subspace

V_{(1)}

has a natural Lie algebra structure with

[u, v] = u_{0} v

for

u, v \in V_{(1)}

and the product

u_{1} v

(\in V_{(0)})

defines a symmetric invariant bilinear form on

V_{(1)}

In a series of study on Gerbs of chiral differential operators in [GMS] and on chiral de Rham complex in [MSV, MS1,2], Malikov and his coauthors investigated

N

-graded vertex algebras

V = \prod_{n \in N} V_{(n)}

with

V_{(0)}

not necessarily

1

-dimensional. In this case, the bilinear operations

(u, v) \mapsto u_{i} v

for

i \geq 0

are closed on

V_{(0)} \oplus V_{(1)}

u_{i} v \in V_{(0)} \oplus V_{(1)} for u, v \in V_{(0)} \oplus V_{(1)}, i \geq 0 .

The skew symmetry and the Jacobi identity for the vertex algebra

V

give rise to several compatibility relations. Such algebraic structures on

V_{(0)} \oplus V_{(1)}

are summarized in the notion of what was called a

1

-truncated conformal algebra.

Furthermore, the subspace

V_{(0)}

equipped with the product

(a, b) \mapsto a_{- 1} b

is a commutative associative algebra with the vacuum vector

1

as the identity and

V_{(0)}

as a nonassociative algebra acts on

V_{(1)}

a \cdot u = a_{- 1} u

for

a \in V_{(0)}

u \in V_{(1)}

. All these structures on

V_{(0)} \oplus V_{(1)}

are further summarized in the notion of what was called a vertex

A

-algebroid, where

A

is a (unital) commutative associative algebra. On the other hand, in [GMS], among other important results, Gorbounov, Malikov and Schechtman constructed an

N

-graded vertex algebra

V = \prod_{n \in N} V_{(n)}

from any vertex

A

-algebroid, such that

V_{(0)} = A

and the vertex

A

-algebroid

V_{(1)}

is isomorphic to the given one. All the constructed

N

-graded vertex algebras are generated by

V_{(0)} \oplus V_{(1)}

with a spanning property of PBW type. As it was demonstrated in [GMS], such

N

-graded vertex algebras are natural and important to study. For example, the vertex (operator) algebra associated with a

β γ

system, which plays a central role in free field realization of affine Lie algebras (see [W, FF1-3, FB]) is such an

N

-graded vertex algebra. The vertex (operator) algebras constructed from toroidal Lie algebras are also of this type (see [BBS, BDT]). In [LY] , we revisited those

N

-graded vertex algebras and we classified all the

N

-graded simple modules in terms of simple modules for certain Lie algebroids.

In the theory of vertex algebras, in addition to the notion of module we have the notion of twisted module and twisted modules play a very important role, especially in the study of the so-called orbifold theory. Certainly, twisted modules also play an important role in other studies. In this paper, we continue to study the twisted modules for the

N

-graded vertex algebras associated with vertex algebroids.

Let

B

be a vertex

A

-algebroid and let

V_{B}

be the associated

N

-graded vertex algebra. In this paper, we define a notion of automorphism of the vertex

A

-algebroid

B

and we prove that any automorphism of the vertex

A

-algebroid

B

can be extended uniquely to an automorphism of the

N

-graded vertex algebra

V_{B}

and that the full automorphism group of the

N

-graded vertex algebra

V_{B}

is naturally isomorphic to the full automorphism group of the vertex

A

-algebroid

B

. Let

g

be an automorphism of the vertex

A

-algebroid

B

of order

T

(finite). Then the

g

-fixed point

A^{0}

is a subalgebra of

A

and the

g

-fixed point

B^{0}

is a vertex

A^{0}

-algebroid. Furthermore,

B^{0} / A^{0} \partial A^{0}

is a Lie

A^{0}

-algebroid. It is proved that the category of

\frac{1}{T} N

-graded simple

g

-twisted

V_{B}

-modules is equivalent to a subcategory of simple modules for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

This paper is organized as follows: In Section 2, we review the construction of vertex algebras associated with vertex algebroids and we identify their automorphism groups with the automorphism groups of the vertex algebroids. In Section 3, we classify graded simple twisted modules.

2 Preliminaries

We recall the notions of

1

-truncated conformal algebra, vertex algebroid and Lie algebroid, and we review the construction of the

N

-graded vertex algebra

V_{B}

associated with a vertex

A

-algebroid

B

. We also define notions of (endomorphism) automorphism of a

1

-truncated conformal algebra and of a vertex

A

-algebroid

B

We then identify the group of grading-preserving automorphisms of

V_{B}

with the group of automorphisms of the vertex

A

-algebroid

B

First, we recall from [GMS] (cf. [Br1-2]) the notions of

1

-truncated conformal algebra, vertex algebroid and Lie algebroid.

Definition 2.1. A

1

-truncated conformal algebra is a graded vector space

C = C_{0} \oplus C_{1}

, equipped with a linear map

\partial : C_{0} \to C_{1}

and bilinear operations

(u, v) \mapsto u_{i} v

for

i = 0, 1

of degree

- i - 1

C

such that the following axioms hold:

1. (Derivation) for

a \in C_{0}

u \in C_{1}

\begin{matrix} (\partial a)_{0} = 0; (\partial a)_{1} = - a_{0}; \partial (u_{0} a) = u_{0} \partial a \end{matrix}

(2.1)

2. (Commutativity) for

a \in C_{0}

u, v \in C_{1}

\begin{matrix} u_{0} a = - a_{0} u; u_{0} v = - v_{0} u + \partial (v_{1} u); u_{1} v = v_{1} u \end{matrix}

(2.2)

3. (Associativity) for

α, β, γ \in C_{0} \oplus C_{1}

\begin{matrix} α_{0} β_{i} γ = β_{i} α_{0} γ + (α_{0} β)_{i} γ . \end{matrix}

(2.3)

Remark 2.2. Let

C = C_{0} \oplus C_{1}

be a

1

-truncated conformal algebra and let

ℓ

be any nonzero complex number. Set

C [ℓ] = C_{0} \oplus C_{1}

as a vector space.

We retain all the structures on

C

except that we change the bilinear operation

C_{1} \times C_{1} \to C_{0} : u \times v \mapsto u_{1} v

by multiplying

1 / ℓ

and change the linear operator

\partial

by multiplying

ℓ

. Then one can show that

C [ℓ]

is a

1

-truncated conformal algebra.

Definition 2.3. Let

A

be a unital commutative associative algebra over

C

. A vertex

A

-algebroid is a

C

-vector space

Γ

equipped with

1. a $C$ -bilinear map $A \times Γ \to Γ; (a, v) \mapsto a * v$ such that $1 * v = v$ for $v \in Γ$ .
2. a structure of a Leibniz $C$ -algebra $[\cdot, \cdot] : Γ \otimes_{C} Γ \to Γ$ .
3. a homomorphism of Leibniz $C$ -algebras $π : Γ \to D e r (A)$ .
4. a symmetric $C$ -bilinear pairing $〈 \cdot, \cdot 〉 : Γ \otimes_{C} Γ \to A$ .
5. a $C$ -linear map $\partial : A \to Γ$ such that $π \circ \partial = 0$ .

All the following conditions are assumed to hold:

\begin{matrix} a * (a^{'} * v) - (a a^{'}) * v & = & π (v) (a) * \partial (a^{'}) + π (v) (a^{'}) * \partial (a) \end{matrix}

\begin{matrix} [u, a * v] & = & π (u) (a) * v + a * [u, v] \end{matrix}

\begin{matrix} [u, v] + [v, u] & = & \partial (〈 u, v 〉) \end{matrix}

\begin{matrix} π (a * v) & = & a π (v) \end{matrix}

\begin{matrix} 〈 a * u, v 〉 & = & a 〈 u, v 〉 - π (u) (π (v) (a)) \end{matrix}

\begin{matrix} π (v) (〈 v_{1}, v_{2} 〉) & = & 〈 [v, v_{1}], v_{2} 〉 + 〈 v_{1}, [v, v_{2}] 〉 \end{matrix}

\begin{matrix} \partial (a a^{'}) & = & a * \partial (a^{'}) + a^{'} * \partial (a) \end{matrix}

\begin{matrix} [v, \partial (a)] & = & \partial (π (v) (a)) \end{matrix}

\begin{matrix} 〈 v, \partial (a) 〉 & = & π (v) (a) \end{matrix}

for

a, a^{'} \in A

u, v, v_{1}, v_{2} \in Γ .

The following was proved in [LY] :

Proposition 2.4. Let

A

be a unital commutative associative algebra and let

B

be a module for

A

as a nonassociative algebra. Then a vertex

A

-algebroid structure on

B

is equivalent to a

1

-truncated conformal algebra structure on

C = A \oplus B

with

\begin{matrix} a_{i} a^{'} = 0, \end{matrix}

(2.4)

\begin{matrix} u_{0} v = [u, v], u_{1} v = 〈 u, v 〉, \end{matrix}

(2.5)

\begin{matrix} u_{0} a = π (u) (a), a_{0} u = - u_{0} a = - π (u) (a) \end{matrix}

(2.6)

for

a, a^{'} \in A

u, v \in B

i = 0, 1

, such that

\begin{matrix} a (a^{'} u) - (a a^{'}) u = (u_{0} a) \partial a^{'} + (u_{0} a^{'}) \partial a, \end{matrix}

(2.7)

\begin{matrix} u_{0} (a v) - a (u_{0} v) = (u_{0} a) v, \end{matrix}

(2.8)

\begin{matrix} u_{0} (a a^{'}) = a (u_{0} a^{'}) + (u_{0} a) a^{'}, \end{matrix}

(2.9)

\begin{matrix} a_{0} (a^{'} v) = a^{'} (a_{0} v), \end{matrix}

(2.10)

\begin{matrix} (a u)_{1} v = a (u_{1} v) - u_{0} v_{0} a, \end{matrix}

(2.11)

\begin{matrix} \partial (a a^{'}) = a \partial (a^{'}) + a^{'} \partial (a) . \end{matrix}

(2.12)

Definition 2.5. Let

A

be a unital commutative associative algebra. A Lie

A

-algebroid is a Lie algebra

g

equipped with an

A

-module structure and a module action on

A

by derivation such that

\begin{matrix} [u, a v] & = & a [u, v] + (u a) v, \end{matrix}

(2.13)

\begin{matrix} a (u b) & = & (a u) b for u, v \in g, a, b \in A . \end{matrix}

(2.14)

A module for a Lie

A

-algebroid

g

is a vector space

W

equipped with a

g

-module structure and an

A

-module structure such that

\begin{matrix} u (a w) - a (u w) = (u a) w, \end{matrix}

(2.15)

\begin{matrix} a (u w) = (a u) w for a \in A, u \in g, w \in W . \end{matrix}

(2.16)

The following result was due to [Br2]:

Lemma 2.6. Let

A

be a unital commutative associative algebra (over

C

) and let

B

be a vertex

A

-algebroid. Then

B / A \partial A

is naturally a Lie

A

-algebroid.

Next, we recall the construction of vertex algebras associated with vertex algebroids, following the exposition of [LY] .

First, starting with a

1

-truncated conformal algebra

C = A \oplus B

we construct a Lie algebra. Set

\begin{matrix} L (A \oplus B) = (A \oplus B) \otimes C [t, t^{- 1}] . \end{matrix}

(2.17)

In the obvious way we define the subpaces

L (A)

and

L (B)

. Set

\begin{matrix} \hat{\partial} = \partial \otimes 1 + 1 \otimes d / d t : L (A) \to L (A \oplus B) . \end{matrix}

We define

\begin{matrix} deg (a \otimes t^{n}) & = & - n - 1 for a \in A, n \in Z, \end{matrix}

\begin{matrix} deg (b \otimes t^{n}) & = & - n for b \in B, n \in Z, \end{matrix}

making

L (A \oplus B)

Z

-graded vector space. The linear map

\hat{\partial}

is homogeneous of degree

1

. Set

\begin{matrix} ℒ = L (A \oplus B) / \hat{\partial} L (A) . \end{matrix}

(2.18)

Define a bilinear product

[\cdot, \cdot]

L (A \oplus B)

such that for

a, a^{'} \in A, b, b^{'} \in B, m, n \in Z

\begin{matrix} [a \otimes t^{m}, a^{'} \otimes t^{n}] = 0, \end{matrix}

(2.19)

\begin{matrix} [a \otimes t^{m}, b \otimes t^{n}] = a_{0} b \otimes t^{m + n}, \end{matrix}

(2.20)

\begin{matrix} [b \otimes t^{n}, a \otimes t^{m}] = b_{0} a \otimes t^{m + n}, \end{matrix}

(2.21)

\begin{matrix} [b \otimes t^{m}, b^{'} \otimes t^{n}] = b_{0} b^{'} \otimes t^{m + n} + m (b_{1} b^{'}) \otimes t^{m + n - 1} . \end{matrix}

(2.22)

The following result was established in [LY] :

Proposition 2.7. Let

C = A \oplus B

be a

1

-truncated conformal algebra. The subspace

\hat{\partial} L (A)

of the nonassociative algebra

(L (A \oplus B), [\cdot, \cdot])

is a two-sided ideal. Furthermore, the quotient nonassociative algebra

ℒ

is a

Z

-graded Lie algebra.

Let

ρ

be the projection map from

L (A \oplus B)

ℒ

. For

u \in A \oplus B, n \in Z

, we set

u (n) = ρ (u \otimes t^{n}) = u \otimes t^{n} + \hat{\partial} L (A) \in ℒ .

We have graded Lie subalgebras

\begin{matrix} ℒ^{\geq 0} & = & ρ ((A \oplus B) \otimes C [t]), \end{matrix}

\begin{matrix} ℒ^{< 0} & = & ρ ((A \oplus B) \otimes t^{- 1} C [t^{- 1}]) \end{matrix}

and we have

ℒ = ℒ^{\geq 0} \oplus ℒ^{< 0}

as a vector space.

Considering

C

as a trivial

ℒ^{\geq 0}

-module we form the induced module

V_{ℒ} = U (ℒ) \otimes_{U (ℒ^{\geq 0})} C .

We assign

deg C = 0

, making

V_{ℒ}

naturally an

N

-graded

ℒ

-module:

\begin{matrix} V_{ℒ} = \prod_{n \in N} (V_{ℒ})_{(n)} . \end{matrix}

(2.23)

Throughout this paper,

N

denotes the set of nonnegative integers. Set

1 = 1 \otimes 1 \in V_{ℒ} .

By the P-B-W theorem, we have

V_{ℒ} = U (ℒ^{< 0}) = S (ℒ^{< 0})

. In view of this, we can and we do consider

A \oplus B

as a subspace:

A \oplus B \to V_{ℒ}; u \mapsto u (- 1) 1 .

The following was proved in [LY] (cf. [DLM3] ):

Theorem 2.8. There exists a unique vertex algebra structure on

V_{ℒ}

with

1

as the vacuum vector and with

Y (u, x) = \sum_{n \in Z} u (n) x^{- n - 1}

for

u \in A \oplus B

Moreover, the vertex algebra

V_{ℒ}

is naturally an

N

-graded vertex algebra and is generated by the subspace

A \oplus B

with

A

of degree

0

and

B

of degree

1

Remark 2.9. For

n \in Z

, set

A (n) = {a (n) | a \in A}, B (n) = {b (n) | b \in B} \subset ℒ,

and we set

B (-) =^{\infty} \prod_{n = 1} B (- n) \subset ℒ .

Both

A (- 1)

and

B (-)

are Lie subalgebras of

ℒ^{< 0}

and we have

ℒ^{< 0} = A (- 1) \oplus B (-)

as a vector space. Then

\begin{matrix} V_{ℒ} = U (ℒ^{< 0}) = S (ℒ^{< 0}) = S (A (- 1) \oplus B (-)) = S (B (-)) \otimes S (A (- 1)) . \end{matrix}

(2.24)

Consequently,

(V_{ℒ})_{(n)} = S (B (-))_{(n)} \otimes S (A (- 1))

for

n \in N

. In particular,

(V_{ℒ})_{(0)} = S (A (- 1))

Now, we assume that

A

is a unital commutative associative algebra with the identity

e

and

B

is a vertex

A

-algebroid. In particular,

C = A \oplus B

is a

1

-truncated conformal algebra. We set

\begin{matrix} E & = & s p a n {e - 1, a (- 1) a^{'} - a a^{'}, a (- 1) b - a b | a, a^{'} \in A, b \in B} \subset V_{ℒ}, \end{matrix}

\begin{matrix} I_{B} & = & U (ℒ) C [D] E . \end{matrix}

It was proved in [LY] that the

ℒ

-submodule

I_{B}

V_{ℒ}

is a two-sided graded ideal of the

N

-graded vertex algebra

V_{ℒ}

. The

N

-graded vertex algebra

V_{B}

associated with the vertex

A

-algebroid

B

is defined to be the quotient vertex algebra

\begin{matrix} V_{B} = V_{ℒ} / I_{B} . \end{matrix}

(2.25)

We have (see [GMS], [LY] ):

Proposition 2.10. Let

A

be a unital commutative associative algebra with the identity

e

and

B

a vertex

A

-algebroid. Then

V_{B}

is an

N

-graded vertex algebra such that

(V_{B})_{(0)} = A

(V_{B})_{(1)} = B

and for

n \geq 1

\begin{matrix} (V_{B})_{(n)} \end{matrix}

\begin{matrix} = & s p a n {b_{1} (- n_{1}) \dots b_{k} (- n_{k}) 1 | b_{i} \in B, n_{1} \geq n_{2} \geq \dots \geq n_{k} \geq 1, n_{1} + \dots + n_{k} = n} . \end{matrix}

In particular,

V_{B}

is generated by the subspace

A \oplus B

Next, we discuss homomorphisms and automorphisms for

1

-truncated conformal algebras, vertex

A

-algebroids and for the

N

-graded vertex algebras

V_{B}

Definition 2.11. Let

C = A \oplus B

and

C^{'} = A^{'} \oplus B^{'}

1

-truncated conformal algebras. A homomorphism from

C

C^{'}

is a linear map

f : C \to C^{'}

such that

f (A) \subset A^{'}, f (B) \subset B^{'}

f \partial = \partial f

, and such that

f (u_{i} v) = f (u)_{i} f (v)

for

u, v \in C, i = 0, 1

Lemma 2.12. Let

f

be an endomorphism of a

1

-truncated conformal algebra

C = A \oplus B

. Then the linear endomorphism of

L (A \oplus B)

defined by

\begin{matrix} \hat{f} (u \otimes t^{n}) = f (u) \otimes t^{n} \end{matrix}

(2.26)

for

u \in A \oplus B, n \in Z

gives rise to an endomorphism of

ℒ

, which we denote by

\hat{f}

again. Furthermore,

\hat{f}

preserves the

Z

-grading of

ℒ

Proof. Using the property that $f \partial = \partial f$ , we have $\hat{f} \hat{\partial} = \hat{\partial} \hat{f}$ . For $u, v \in C = A \oplus B$ , as $f (u_{i} v) = f (u)_{i} f (v)$ for $i = 0, 1$ , from ( 2.19 )-( 2.22 ) we have $\hat{f} ([u \otimes t^{m}, v \otimes t^{n}]) = [f (u) \otimes t^{m}, f (v) \otimes t^{n}] = [\hat{f} (u \otimes t^{m}), \hat{f} (v \otimes t^{n})] .$ Thus $\hat{f}$ gives rise to an endomorphism of the Lie algebra $ℒ$ . It is clear that $\hat{f}$ preserves the $Z$ -grading.

Definition 2.13. Let

A

and

A^{'}

be unital commutative associative algebras and let

B

be a vertex

A

-algebroid,

B^{'}

a vertex

A^{'}

-algebroid. A vertex algebroid homomorphism from

B

B^{'}

is a linear map

f : A \oplus B \to A^{'} \oplus B^{'}

such that

f (A) \subset A^{'}, f (B) \subset B^{'}

and such that

1. $f |_{A}$ is an associative algebra homomorphism.
2. $f |_{B}$ is a Leibniz algebra homomorphism.
3. $f (a b) = f (a) f (b)$ for $a \in A, b \in B$ .
4. $〈 f (u), f (v) 〉 = f (〈 u, v 〉)$ for $u, v \in B$ .
5. $f \circ \partial = \partial \circ f$ .
6. $f (b_{0} a) = f (b)_{0} f (a)$ for $a \in A, b \in B$ .

An automorphism of a vertex

A

-algebroid

B

is a bijective vertex algebroid endomorphism of the vertex

A

-algebroid

B

Let

(V, Y, 1)

be a vertex algebra. An endomorphism of

V

is a linear map

g : V \to V

such that

\begin{matrix} g (1) & = & 1, \end{matrix}

(2.27)

\begin{matrix} g (Y (u, x) v) & = & Y (g (u), x) g (v) \end{matrix}

(2.28)

for

u, v \in V

. An automorphism of

V

is a bijective endomorphism of

V

. The group of automorphisms of

V

is denoted by

A u t (V)

. If

V = \prod_{m \in Z} V_{(m)}

is a

Z

(or

N

)-graded vertex algebra, we denote by

A u t^{o} (V)

the group of grading-preserving automorphisms of

V

Lemma 2.14. Let

B

be a vertex

A

-algebroid and let

g

be a grading-preserving automorphism of the vertex algebra

V_{B}

. Then

g

restricted to

A \oplus B

is an automorphism of the vertex

A

-algebroid

B

Proof. As

(V_{B})_{(0)} = A

and

(V_{B})_{(1)} = B

g

is a linear bijection on

A \oplus B

that preserves the subspaces

A

and

B

. For

a, a^{'} \in A, b, b^{'} \in B

, we have

\begin{matrix} g (a a^{'}) = g (a (- 1) a^{'}) = g (a)_{- 1} g (a^{'}) = g (a) g (a^{'}), \end{matrix}

\begin{matrix} g (a b) = g (a (- 1) b) = g (a)_{- 1} g (b) = g (a) g (b), \end{matrix}

\begin{matrix} g ([b, b^{'}]) = g (b_{0} b^{'}) = g (b)_{0} g (b^{'}) = [g (b), g (b^{'})], \end{matrix}

\begin{matrix} g (〈 b, b^{'} 〉) = g (b_{1} b^{'}) = g (b)_{1} g (b^{'}) = 〈 g (b), g (b^{'}) 〉, \end{matrix}

\begin{matrix} g (b_{0} a) = g (b)_{0} g (a), \end{matrix}

\begin{matrix} g (\partial (a)) = g (a (- 2) 1) = g (a)_{- 2} 1 = \partial (g (a)) . \end{matrix}

Thus

g

is an automorphism of vertex

A

-algebroid

B

On the other hand, we are going to prove that any automorphism of a vertex

A

-algebroid

B

extends canonically to an automorphism of the

N

-graded vertex algebra

V_{B}

. First we have:

Lemma 2.15. Let

C = A \oplus B

be a

1

-truncated conformal algebra and let

g

be an endomorphism of

C

. Then

g

extends uniquely to an endomorphism of the

N

-graded vertex algebra

V_{ℒ}

. Furthermore, if

g

is an automorphism, then the extension is an automorphism.

Proof. Since $A \oplus B$ generates $V_{ℒ}$ as a vertex algebra, the uniqueness is clear. It remains to prove the existence. By Lemma 2.12 , we have a grading-preserving endomorphism $\hat{g}$ of the Lie algebra $ℒ$ , hence a grading-preserving endomorphism of the universal enveloping algebra $U (ℒ)$ . Consequently, $\hat{g}$ preserves the Lie subalgebra $ℒ^{< 0}$ and its universal enveloping algebra $U (ℒ^{< 0})$ . It follows from the construction of $V_{ℒ}$ that there exists a linear endomorphism $\bar{g}$ of $V_{ℒ}$ such that $\bar{g} (1) = 1$ and $\bar{g} (u_{n} v) = g (u)_{n} \bar{g} (v)$ for $u \in A \oplus B$ , $v \in V_{ℒ}$ , $n \in Z$ . Since $V_{ℒ}$ is generated by $A \oplus B$ , it follows (cf. [LLi]) that $\bar{g}$ is an endomorphism of $V_{ℒ}$ . It is clear that $\bar{g}$ extends $g$ .
If $g$ is an automorphism of the $1$ -truncated conformal algebra $C = A \oplus B$ , from the first assertion we have vertex algebra endomorphisms $\bar{g}$ and $\bar{g^{- 1}}$ of $V_{ℒ}$ , extending $g$ and $g^{- 1}$ , respectively. Since $g g^{- 1} = g^{- 1} g = 1$ on $A \oplus B$ and since $A \oplus B$ generates $V_{ℒ}$ as a vertex algebra, we have $\bar{g} \bar{g^{- 1}} = \bar{g^{- 1}} \bar{g} = 1$ . Thus, $\bar{g}$ is an automorphism of $V_{ℒ}$ .

Proposition 2.16. Let

g

be an endomorphism of a vertex

A

-algebroid

B

. Then

g

extends uniquely to an endomorphism of

V_{B}

as an

N

-graded vertex algebra. Furthermore, if

g

is an automorphism, then the extension is an automorphism.

Proof. The uniqueness is clear, as

A \oplus B

generates

V_{B}

as a vertex algebra. For the the existence, first by Lemma 2.15 , we have a grading-preserving endomorphism

\bar{g}

of the vertex algebra

V_{ℒ}

, extending

g

. Now we show that

\bar{g}

reduces to an endomorphism of

V_{B}

. Recall that

V_{B} = V_{ℒ} / I_{B}

, where

I_{B}

is the two-sided ideal of

V_{ℒ}

, generated by

E = s p a n {e - 1, a (- 1) a^{'} - a a^{'}, a (- 1) b - a b | a, a^{'} \in A, b \in B} .

Now, we must prove

\bar{g} (I_{B}) \subset I_{B}

. As

E

generates

I_{B}

as a two-sided ideal, it suffices to prove that

\bar{g} (E) \subset E

. Let

a, a^{'} \in A, b \in B

. We have

\begin{matrix} \bar{g} (e - 1) = e - 1 \in E, \end{matrix}

\begin{matrix} \bar{g} (a (- 1) a^{'} - a a^{'}) = g (a) (- 1) g (a^{'}) - g (a) g (a^{'}) \in E, \end{matrix}

\begin{matrix} \bar{g} (a (- 1) b - a b) = g (a) (- 1) g (b) - g (a) g (b) \in E . \end{matrix}

This proves

\bar{g} (E) \subset E

. Therefore,

\bar{g}

reduces to an endomorphism of the

N

-graded vertex algebra

V_{B}

. The second assertion follows immediately from the proof of the second assertion of Lemma 2.15 .

Recall that

A u t^{o} (V_{B})

denotes the group of grading-preserving automorphisms of

V_{B}

, namely the group of automorphisms of

V_{B}

as an

N

-graded vertex algebra.

Combining Lemma 2.14 with Proposition 2.16 , we have:

Theorem 2.17. Let

A

be a unital commutative associative algebra and let

B

be a vertex

A

-algebroid. The group

A u t^{o} (V_{B})

of (grading-preserving) automorphisms of the

N

-graded vertex algebra

V_{B}

is isomorphic to the group of automorphisms of vertex

A

-algebroid

B

with the restriction map as an isomorphism.

3 Classification of graded simple twisted $V_{B}$ -modules

In this section we construct and classify graded simple twisted

V_{B}

-modules by exploiting a twisted analogue of the Lie algebra

ℒ

. First, we recall the definition of the notion of twisted module for a vertex algebra and we discuss several properties of twisted modules.

Let

V

be a vertex algebra and let

g

be an automorphism of

V

of order

T < \infty

Decompose

V

into eigenspaces of

g

V =^{T - 1} \prod_{r = 0} V^{r}, where V^{r} = {v \in V | g (v) = e^{2 r π \sqrt{- 1} / T} v} .

g

-twisted

V

-module (see [Le] , [FLM] , [FFR] , [D] ) is a vector space

M

equipped with a linear map

\begin{matrix} Y_{M} : V & \to & (E n d M) [[x^{\frac{1}{T}}, x^{- \frac{1}{T}}]] \end{matrix}

\begin{matrix} u & \mapsto & Y_{M} (u, x) = \sum_{n \in \frac{1}{T} Z} u_{n} x^{- n - 1} \end{matrix}

(3.1)

satisfying the following conditions:

1. For $u \in V, w \in M$ , $u_{n} w = 0$ for $n \in \frac{1}{T} Z$ sufficiently large.
2. $Y_{M} (1, x) = 1_{M}$ (the identity operator on $M$ ).

3. For

u \in V^{r}

with

0 \leq r \leq T - 1

\begin{matrix} Y_{M} (u, x) = \sum_{n \in \frac{r}{T} + Z} u_{n} x^{- n - 1} \in x^{- \frac{r}{T}} (E n d M) [[x, x^{- 1}]] . \end{matrix}

(3.2)

4. For

u \in V^{r}

with

0 \leq r \leq T - 1

v \in V

\begin{matrix} x_{0}^{- 1} δ (\frac{x_{1} - x_{2}}{x_{0}}) Y_{M} (u, x_{1}) Y_{M} (v, x_{2}) - x_{0}^{- 1} δ (\frac{x_{2} - x_{1}}{- x_{0}}) Y_{M} (v, x_{2}) Y_{M} (u, x_{1}) \end{matrix}

\begin{matrix} = x_{2}^{- 1} {(\frac{x_{1} - x_{0}}{x_{2}})}^{- r / T} δ (\frac{x_{1} - x_{0}}{x_{2}}) Y_{M} (Y (u, x_{0}) v, x_{2}) \end{matrix}

(3.3)

(the twisted Jacobi identity).

Remark 3.1. Let

(M, Y_{M})

be a

g

-twisted

V

-module and let

U

be any vertex subalgebra of

V^{0}

. Then

M

is a

U

-module. In particular, if

g

is taken to be the identity map, the notion of

g

-twisted

V

-module reduces to that of

V

-module while the twisted Jacobi identity reduces to the ordinary (untwisted) Jacobi identity.

The following was proved in [DLM2] (cf. [DLM1] ):

Lemma 3.2. Let

(M, Y_{M})

be a

g

-twisted

V

-module. Then

\begin{matrix} Y_{M} (D v, x) = \frac{d}{d x} Y_{M} (v, x) \end{matrix}

(3.4)

for

v \in V

, where

D v = v_{- 2} 1

Remark 3.3. For

v \in V

u \in V^{r}

p \in Z

and

s, t \in Q

, comparing the coefficients of

z_{0}^{- p - 1} z_{1}^{- s - 1} z_{2}^{- t - 1}

on the both sides of the twisted Jacobi identity ( 3.3 ) we get

\begin{matrix} \sum_{m \geq 0} (\binom{s}{m}) (u_{p + m} v)_{s + t - m} = \sum_{m \geq 0} (- 1)^{m} (\binom{p}{m}) {u_{p + s - m} v_{t + m} - (- 1)^{p} v_{p + t - m} u_{s + m}} . \end{matrix}

(3.5)

By taking

R e s_{x_{0}}

of ( 3.3 ), we obtain the twisted commutator formulae:

\begin{matrix} [Y_{M} (u, x_{1}), Y_{M} (v, x_{2})] \end{matrix}

\begin{matrix} = R e s_{x_{0}} x_{2}^{- 1} {(\frac{x_{1} - x_{0}}{x_{2}})}^{- r / T} δ (\frac{x_{1} - x_{0}}{x_{2}}) Y_{M} (Y (u, x_{0}) v, x_{2}) . \end{matrix}

(3.6)

Multiplying ( 3.3 ) by

{(\frac{x_{1} - x_{0}}{x_{2}})}^{\frac{r}{T}}

and then taking

R e s_{x_{1}}

, we obtain the twisted iterate formulae:

\begin{matrix} Y_{M} (Y (u, x_{0}) v, x_{2}) = R e s_{x_{1}} {(\frac{x_{1} - x_{0}}{x_{2}})}^{\frac{r}{T}} \cdot X \end{matrix}

(3.7)

where

X = x_{0}^{- 1} δ (\frac{x_{1} - x_{2}}{x_{0}}) Y_{M} (u, x_{1}) Y_{M} (v, x_{2}) - x_{0}^{- 1} δ (\frac{x_{2} - x_{1}}{- x_{0}}) Y_{M} (v, x_{2}) Y_{M} (u, x_{1}) .

From the twisted Jacobi identity one has the following twisted weak associativity: For

u \in V^{r}

with

0 \leq r \leq T - 1

and for

v \in V, w \in W

(x_{0} + x_{2})^{k + \frac{r}{T}} Y_{M} (u, x_{0} + x_{2}) Y_{M} (v, x_{2}) w = (x_{2} + x_{0})^{k + \frac{r}{T}} Y_{M} (Y (u, x_{0}) v, x_{2}) w

where

k

is a nonnegative integer such that

x^{k + \frac{r}{T}} Y_{M} (u, x) w \in M [[x]]

. One can prove (cf. [Li3] ; Lemma 2.8) that the twisted Jacobi identity is equivalent to the twisted commutator formulae and the twisted weak associativity.

Let

M

be a

g

-twisted

V

-module. For a subset

U

M

, denote the smallest

g

-twisted

V

-submodule containing

U

〈 U 〉

, which is called the

g

-twisted

V

-submodule generated by

U

. Just as with untwisted modules, from the twisted weak associativity, we have

〈 U 〉 = s p a n {v_{n} w | v \in V, n \in \frac{1}{T} Z, w \in U} .

We define

\begin{matrix} A n n_{V} (U) = {v \in V | Y (v, x) w = 0 for w \in U}, \end{matrix}

(3.8)

the annihilator of

U

V

Proposition 3.4. For any subset

U

of a

g

-twisted

V

-module

M

, the annihilator

A n n_{V} (U)

is an ideal of

V

. Moreover,

A n n_{V} (U) = A n n_{V} (〈 U 〉) .

Proof. It follows immediately from the proof of Proposition 4.5.11 in [LLi] with the weak associativity and the weak commutativity relations being replaced by the twisted associativity and the twisted commutativity relations, respectively.

Let

S

be a subset of

V

. Define

A n n_{M} (S) = {w \in M | Y_{M} (v, x) w = 0 for v \in S},

the annihilator of

S

M

. By suitably modifying the proof of Proposition 4.5.14 in [LLi] and replacing the weak commutativity and Proposition 4.5.11 (of [LLi]) in the proof of Proposition 4.5.14 in [LLi] by the twisted commutativity relation and Proposition 3.4 , respectively, we immediately have:

Proposition 3.5. For a subset

S

V

, the annihilator

A n n_{M} (S)

is a

g

-twisted

V

-submodule of

M

. Furthermore,

A n n_{M} (S) = A n n_{M} (〈 S 〉) .

Here,

〈 S 〉

is the ideal of

V

generated by

S

We shall use the following result of [Li3] (Lemma 2.11):

Lemma 3.6. Let

V

be a vertex algebra with an automorphism

g

of order

T

and let

a \in V^{k}, b, u^{0}, . . ., u^{r} \in V

with

0 \leq k \leq T - 1

. If

\begin{matrix} [Y (a, x_{1}), Y (b, x_{2})] = \sum_{j = 0}^{r} \frac{1}{j!} Y (u^{j}, x_{2}) {(\frac{\partial}{\partial x_{2}})}^{j} x_{1}^{- 1} δ (\frac{x_{2}}{x_{1}}) \end{matrix}

(3.9)

acting on

V

, then for any

g

-twisted

V

-module

(M, Y_{M})

we have

\begin{matrix} [Y_{M} (a, x_{1}), Y_{M} (b, x_{2})] = \sum_{j = 0}^{r} \frac{1}{j!} Y_{M} (u^{j}, x_{2}) {(\frac{\partial}{\partial x_{2}})}^{j} x_{1}^{- 1} δ (\frac{x_{2}}{x_{1}}) {(\frac{x_{2}}{x_{1}})}^{\frac{k}{T}} \end{matrix}

(3.10)

acting on

M

. On the other hand, the converse is also true for any faithful

g

-twisted

V

-module

(M, Y_{M})

Definition 3.7. Let

V = \prod_{m \in Z} V_{(m)}

be a

Z

-graded vertex algebra. A

\frac{1}{T} N

-graded

g

-twisted

V

-module is a

g

-twisted

V

-module

M

equipped with a

\frac{1}{T} N

-grading

M = \prod_{n \in \frac{1}{T} N} M (n)

such that

v_{m} M (n) \subset M (n + p - m - 1)

for

v \in V_{(p)}, m, n \in \frac{1}{T} Z

with

p \in Z

Next, we study

\frac{1}{T} N

-graded

g

-twisted modules for the

N

-graded vertex algebra

V_{B}

associated to a vertex

A

-algebroid

B

, where

g

is an automorphism of order

T < \infty

of the

N

-graded vertex algebra

V_{B}

and of the vertex

A

-algebroid

B

(cf. Theorem 2.17 ).

Noticing that

A \oplus B

is a

1

-truncated conformal algebra, we start with a general

1

-truncated conformal algebra

C = C_{0} \oplus C_{1}

with an automorphism

g

C

of order

T < \infty

. Associated with the

1

-truncated conformal algebra

C = C_{0} \oplus C_{1}

we have the Lie algebra

ℒ (C)

and the vertex algebra

V_{ℒ (C)}

with

C

as a generating subspace.

In view of Lemma 2.15

g

is an order-

T

automorphism of the vertex algebra

V_{ℒ (C)}

The following result can be found in [DLM2] (cf. [B] ):

Lemma 3.8. Let

V

be a vertex algebra and let

T

be a positive integer.

Set

\begin{matrix} L_{T} (V) = V \otimes C [t^{\frac{1}{T}}, t^{- \frac{1}{T}}], \end{matrix}

(3.11)

a vector space, and set

\hat{\partial} = D \otimes 1 + 1 \otimes \frac{d}{d t},

a linear operator on

L_{T} (V)

. Then the bilinear (multiplicative) operation on

L_{T} (V)

, defined by

\begin{matrix} [u \otimes t^{m}, v \otimes t^{n}] = \sum_{i \geq 0} (\binom{m}{i}) (u_{i} v \otimes t^{m + n - i}) \end{matrix}

(3.12)

for

u, v \in V, m, n \in \frac{1}{T} Z

, gives rise to a Lie algebra structure on

L_{T} (V) / \hat{\partial} L_{T} (V)

, which is denoted by

ℒ (V, T)

. Furthermore, any order-

T

automorphism

g

C

gives rise to an order-

T

automorphism, also denoted by

g

, of

ℒ (V, T)

, where

\begin{matrix} g (v \otimes t^{n}) = e^{- 2 n π \sqrt{- 1}} (g v \otimes t^{n}) \end{matrix}

(3.13)

for

v \in V, n \in \frac{1}{T} Z

Specializing Lemma 3.8 with

V = V_{ℒ (C)}

, we have a Lie algebra

ℒ (V_{ℒ (C)}, T)

and an automorphism

g

. For

u \in C, m \in \frac{1}{T} Z

, denote by

u (m)

the canonical image of

u \otimes t^{m}

ℒ (V_{ℒ (C)}, T)

. We have

\begin{matrix} (\partial a) (m) = - m a (m - 1), \end{matrix}

(3.14)

\begin{matrix} [u (m), v (n)] = \sum_{i = 0}^{1} (\binom{m}{i}) (u_{i} v) (m + n - i) \end{matrix}

(3.15)

for

a \in C_{0}, u, v \in C, m, n \in \frac{1}{T} Z

. Because

u_{i} v \in C

for

u, v \in C, i \geq 0

, we see that

u (m) u \in C, m \in \frac{1}{T} Z

span a Lie subalgebra

ℒ (C, T)

ℒ (V_{ℒ (C)}, T)

. Denote by

ℒ (C, g)

the

g

-fixed point Lie subalgebra:

\begin{matrix} ℒ (C, g) = ℒ (C, T)^{g} . \end{matrix}

(3.16)

Using Lemma 3.8 we immediately have:

Proposition 3.9. Let

C = C_{0} \oplus C_{1}

be a

1

-truncated conformal algebra and let

g

be an order-

T

automorphism of

C

. Then

\begin{matrix} ℒ (C, g) = L (C, g) / \hat{\partial} L (C_{0}, g), \end{matrix}

(3.17)

as a vector space, where

\begin{matrix} L (C, g) =^{T - 1} \prod_{r = 0} C^{r} \otimes t^{r / T} C [t, t^{- 1}], \end{matrix}

(3.18)

L (C_{0}, g)

is a subspace defined in the obvious way, and

\hat{\partial} = \partial \otimes 1 + 1 \otimes d / d t : L (C_{0}, g) \to L (C, g) .

For

u \in C^{r}

with

0 \leq r \leq T - 1

and for

n \in Z

, denote by

u (n + r / T)

the canonical image of

u \otimes t^{n + r / T}

ℒ (C, g)

. Then the following relations hold for

a \in C_{0}^{r}, a^{'} \in C_{0}^{r^{'}}, b \in C_{1}^{s}, b^{'} \in C_{1}^{s^{'}}, m, n \in Z

\begin{matrix} (\partial a) (m + r / T) = - (m + r / T) a (m - 1 + r / T), \end{matrix}

(3.19)

\begin{matrix} [a (m + r / T), a^{'} (n + r^{'} / T)] = 0, \end{matrix}

(3.20)

\begin{matrix} [a (m + r / T), b (n + s / T)] = (a_{0} b) (m + n + (r + s) / T), \end{matrix}

(3.21)

\begin{matrix} [b (m + s / T), b^{'} (n + s^{'} / T)] = (b_{0} b^{'}) (m + n + (s + s^{'}) / T) \end{matrix}

\begin{matrix} + (m + s / T) (b_{1} b^{'}) (m + n + (s + s^{'}) / T - 1) . \end{matrix}

(3.22)

We define

\begin{matrix} deg a (n + r / T) = - n - 1 for a \in C_{0}^{r}, n \in Z, \end{matrix}

\begin{matrix} deg b (n + r / T) = - n for b \in C_{1}^{r}, n \in Z, \end{matrix}

making

ℒ (C, g)

\frac{1}{T} Z

-graded Lie algebra. For

n \in \frac{1}{T} Z

, denote by

ℒ (C, g)_{(n)}

the degree-

n

subspace. We have the following triangular decomposition

ℒ (C, g) = ℒ (C, g)_{+} \oplus ℒ (C, g)_{(0)} \oplus ℒ (C, g)_{-},

where

ℒ (C, g)_{\pm} = \prod_{0 < n \in \frac{1}{T} Z} ℒ (C, g)_{(\pm n)} .

Notice that

ℒ (C, g)_{(0)}

is spanned by the elements

a (- 1), b (0)

for

a \in C_{0}^{0}, b \in C_{1}^{0}

For

u \in C^{r}

with

0 \leq r \leq T - 1

, form the generating function

\begin{matrix} u (x) = \sum_{n \in \frac{r}{T} + Z} u (n) x^{- n - 1} \in ℒ (C, g) [[x^{\frac{1}{T}}, x^{- \frac{1}{T}}]] . \end{matrix}

(3.23)

For any

ℒ (C, g)

-module

W

, we consider

u (x)

as an element of

(E n d W) [[x^{\frac{1}{T}}, x^{- \frac{1}{T}}]]

, which we denote by

u_{W} (x)

\begin{matrix} u_{W} (x) = u (x) = \sum_{n \in \frac{1}{T} Z} u (n) x^{- n - 1} \in (E n d W) [[x^{\frac{1}{T}}, x^{- \frac{1}{T}}]] . \end{matrix}

(3.24)

Lemma 3.10. The commutation relations ( 3.20 )–( 3.22 ) amount to the following relations in terms of generating functions:

\begin{matrix} [a (x_{1}), a^{'} (x_{2})] & = & 0, \end{matrix}

(3.25)

\begin{matrix} [a (x_{1}), b^{'} (x_{2})] & = & x_{2}^{- 1} {(\frac{x_{1}}{x_{2}})}^{- \frac{r}{T}} δ (\frac{x_{1}}{x_{2}}) (a_{0} b^{'}) (x_{2}), \end{matrix}

(3.26)

\begin{matrix} [b (x_{1}), b^{'} (x_{2})] & = & x_{2}^{- 1} {(\frac{x_{1}}{x_{2}})}^{- \frac{r}{T}} δ (\frac{x_{1}}{x_{2}}) (b_{0} b^{'}) (x_{2}) \end{matrix}

\begin{matrix} + (b_{1} b^{'}) (x_{2}) \frac{\partial}{\partial x_{2}} x_{2}^{- 1} {(\frac{x_{1}}{x_{2}})}^{- \frac{r}{T}} δ (\frac{x_{1}}{x_{2}}) \end{matrix}

(3.27)

for

a \in C_{0}^{r}, b \in C_{1}^{r}, a^{'} \in C_{0}

, and

b^{'} \in C_{1}

. Moreover, we have

\begin{matrix} (\partial a) (x) = \frac{d}{d x} a (x) for a \in C_{0} . \end{matrix}

(3.28)

From these relations we immediately have:

Corollary 3.11. For

a, a^{'} \in C_{0}, b, b^{'} \in C_{1}

\begin{matrix} [a (x_{1}), a^{'} (x_{2})] & = & 0, \end{matrix}

(3.29)

\begin{matrix} (x_{1} - x_{2}) [a (x_{1}), b (x_{2})] & = & 0, \end{matrix}

(3.30)

\begin{matrix} (x_{1} - x_{2})^{2} [b (x_{1}), b^{'} (x_{2})] & = & 0 . \end{matrix}

(3.31)

Definition 3.12. An

ℒ (C, g)

-module

W

is said to be restricted if for any

w \in W, u \in C^{r}

with

0 \leq r \leq T - 1

u (n + r / T) w = 0

for

n \in Z

sufficiently large, that is,

u_{W} (x) \in H o m (W, W ((x^{\frac{1}{T}})))

for

u \in C

The following result is analogous to a result of [Li3] for twisted affine Lie algebras:

Proposition 3.13. Let

C = C_{0} \oplus C_{1}

be a

1

-truncated conformal algebra and let

g

be an automorphism of

V_{ℒ (C)}

of order

T

, which is extended from an automorphism of

C

. Every

g

-twisted

V_{ℒ (C)}

-module

W

is naturally a restricted

ℒ (C, g)

-module with

u_{W} (x) = Y_{W} (u, x)

for

u \in C

. Moreover, the set of

g

-twisted

V_{ℒ (C)}

-submodules of

W

is precisely the set of

ℒ (C, g)

-submodules of

W

. On the other hand, for any restricted

ℒ (C, g)

-module

W

, there exists a unique

g

-twisted

V_{ℒ (C)}

-module structure

Y_{W}

W

such that

\begin{matrix} Y_{W} (u, x) = u_{W} (x) for u \in C = C_{0} \oplus C_{1} \subset V_{ℒ (C)} . \end{matrix}

(3.32)

Proof. On the vertex algebra

V_{ℒ (C)}

, the following relations hold for

a, a^{'} \in C_{0}, b, b^{'} \in C_{1}

\begin{matrix} [Y (a, x_{1}), Y (a^{'}, x_{2})] = 0, \end{matrix}

(3.33)

\begin{matrix} [Y (a, x_{1}), Y (b^{'}, x_{2})] = x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) Y (a_{0} b^{'}, x_{2}), \end{matrix}

(3.34)

\begin{matrix} [Y (b, x_{1}), Y (b^{'}, x_{2})] = x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) Y (b_{0} b^{'}, x_{2}) + Y (b_{1} b^{'}, x_{2}) \frac{\partial}{\partial x_{2}} x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) . \end{matrix}

(3.35)

From Lemmas 3.6 and 3.10 , every

g

-twisted

V_{ℒ (C)}

-module

W

is naturally a restricted

ℒ (C, g)

-module with

u_{W} (x) = Y_{W} (u, x)

for

u \in C

. As

C

generates

V_{ℒ (C)}

as a vertex algebra, the set of

g

-twisted

V_{ℒ (C)}

-submodules of

W

is precisely the set of

ℒ (C, g)

-submodules of

W

Let

S = s p a n {u_{W} (x) | u \in C}

. In view of Corollary 3.11 ,

S

is a local subspace of

H o m (W, W ((x^{\frac{1}{T}})))

. Note that

H o m (W, W ((x^{\frac{1}{T}})))

is naturally

Z / T Z

-graded:

H o m (W, W ((x^{\frac{1}{T}}))) =^{T - 1} \prod_{r = 0} x^{\frac{r}{T}} H o m (W, W ((x)))

and

S

is a graded subspace. Let

σ_{T}

be the linear automorphism of

H o m (W, W ((x^{\frac{1}{T}})))

defined by

σ_{T} (α (x)) = e^{- 2 r π \sqrt{- 1} / T} α (x)

for

α (x) \in x^{\frac{r}{T}} H o m (W, W ((x)))

with

0 \leq r \leq T - 1

(cf. ( 3.2 )).

From [Li3] ,

S

generates a vertex algebra

〈 S 〉

inside

H o m (W, W ((x^{\frac{1}{T}})))

with the identity operator

1_{W}

as the vacuum vector and with

σ_{T}

as an automorphism.

Furthermore,

W

is naturally a faithful

σ_{T}

-twisted

〈 S 〉

-module with

Y_{W} (α (x), x_{0}) = α (x_{0})

. With the relations ( 3.25 )-( 3.27 ), by Lemma 3.6 , we have

\begin{matrix} [Y (a_{W} (x), x_{1}), Y (a_{W}^{'} (x), x_{2})] = 0, \end{matrix}

(3.36)

\begin{matrix} [Y (a_{W} (x), x_{1}), Y (b_{W}^{'} (x), x_{2})] = x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) Y ((a_{0} b^{'})_{W} (x), x_{2}), \end{matrix}

(3.37)

\begin{matrix} [Y (b_{W} (x), x_{1}), Y (b_{W}^{'} (x), x_{2})] = x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) Y ((b_{0} b^{'})_{W} (x), x_{2}) \end{matrix}

\begin{matrix} + Y ((b_{1} b^{'})_{W} (x), x_{2}) \frac{\partial}{\partial x_{2}} x_{2}^{- 1} δ (\frac{x_{1}}{x_{2}}) \end{matrix}

(3.38)

for

a \in C_{0}^{r}, b \in C_{1}^{r}, a^{'} \in C_{0}

, and

b^{'} \in C_{1}

. We also have

Y ((\partial a)_{W} (x), x_{1}) = Y (\frac{d}{d x} a_{W} (x), x_{1}) = \frac{\partial}{\partial x_{1}} Y (a_{W} (x), x_{1})

for

a \in A

. By Lemmas 3.6 and 3.10 ,

〈 S 〉

is naturally an

ℒ (C)

-module with

u_{〈 S 〉} (x_{1}) = Y (u_{W} (x), x_{1})

for

u \in C

. Furthermore,

〈 S 〉

as an

ℒ (C)

-module is generated by

1_{W}

and we have

u_{W} (x)_{n} 1_{W} = 0

for

u \in C, n \geq 0

. From the construction of

V_{ℒ (C)}

as an

ℒ (C)

-module, there exists a unique

ℒ (C)

-module homomorphism

ψ

from

V_{ℒ (C)}

〈 S 〉

, sending

1

1_{W}

. As

V_{ℒ (C)}

as a vertex algebra is generated by

C

ψ

is a vertex algebra homomorphism. We have

ψ (u) = ψ (u (- 1) 1) = u_{W} (x)_{- 1} 1_{W} = u_{W} (x)

for

u \in C

. It is clear that

ψ (C^{r}) \subset S^{r}

for

0 \leq r \leq T - 1

. As

C

generates

V_{ℒ (C)}

as a vertex algebra,

ψ

preserves the

Z / T Z

-gradings, i.e.,

σ_{T} ψ = ψ g

. Consequently,

W

is a

g

-twisted

V_{ℒ (C)}

-module.

For the rest of this paper, we assume that

A

is a unital commutative associative algebra whose identity is denoted by

e

and

B

is a vertex

A

-algebroid and we assume that

g \in A u t^{o} (V_{B})

with

o (g) = T < \infty

. Recall that

C = A \oplus B

is naturally a

1

-truncated conformal algebra. An

ℒ (C, g)

-module of level

k \in C

is an

ℒ (C, g)

-module on which

e (- 1)

acts as scalar

k

Immediately from Proposition 3.13 we have:

Proposition 3.14. Every

g

-twisted

V_{B}

-module is naturally a restricted

ℒ (C, g)

-module of level

1

. Moreover, the set of

g

-twisted

V_{B}

-submodules is precisely the set of

ℒ (C, g)

-submodules.

We have the following decompositions into

g

-eigenspaces:

V_{B} =^{T - 1} \prod_{r = 0} V_{B}^{r}

and

\begin{matrix} A =^{T - 1} \prod_{r = 0} A^{r} and B =^{T - 1} \prod_{r = 0} B^{r} . \end{matrix}

Clearly,

A^{0}

is a subalgebra of

A

, containing the identity, and

B^{0}

is a vertex

A^{0}

-algebroid. Furthermore, by Lemma 2.6 ,

B^{0} / A^{0} \partial A^{0}

is a Lie

A^{0}

-algebroid. Set

\begin{matrix} I = \sum_{r = 1}^{T - 1} A^{r} \cdot A^{T - r} \subset A^{0} . \end{matrix}

(3.39)

It is clear that

I

is a two-sided ideal of

A^{0}

, so that

A^{0} / I

is a unital commutative associative algebra. Furthermore,

B^{0} / (I \cdot B^{0} + A^{0} \partial A^{0})

is a Lie

A^{0} / I

-algebroid.

Proposition 3.15. Let

M = \prod_{n \in \frac{1}{T} N} M (n)

be a

\frac{1}{T} N

-graded

g

-twisted

V_{B}

-module.

Then

M (0)

is a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

with

\begin{matrix} a \cdot w = a_{- 1} w for a \in A^{0}, w \in M (0), \end{matrix}

(3.40)

\begin{matrix} b \cdot w = b_{0} w for b \in B^{0}, w \in M (0) . \end{matrix}

(3.41)

Furthermore, for

a \in A^{r}

a^{'} \in A^{T - r}

b \in B^{T - r}

with

0 < r \leq T - 1

and for

w \in M (0)

, we have

(a a^{'}) \cdot w = 0

and

(a b) \cdot w = (1 - \frac{r}{T}) (a_{0} b) \cdot w

Proof. Let

U

be the vertex subalgebra of

V_{B}

generated by

A^{0} \oplus B^{0}

. As

A^{0} \oplus B^{0} \subset V_{B}^{0}

U

is actually a vertex subalgebra of

V_{B}^{0}

. From Remark 3.1 ,

M

is a

U

-module.

With

(V_{B})_{(0)} = A

and

(V_{B})_{(1)} = B

, we have

(V_{B}^{0})_{(0)} = A^{0}

and

(V_{B}^{0})_{(1)} = B^{0}

Consequently, we have

U_{(0)} = A^{0}

and

U_{(1)} = B^{0}

. It follows from the construction of

V_{B^{0}}

that

U

is a homomorphic image of the vertex algebra

V_{B^{0}}

, so that

W

is naturally a

V_{B^{0}}

-module. By [LY] (Proposition 4.8),

W (0)

is naturally a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

Let

a \in A^{r}, a^{'} \in A^{T - r}, b \in B^{T - r}, w \in M (0)

with

0 < r \leq T - 1

. By substituting

u = a

v = a^{'}

p = - 1

s = - 1 + \frac{r}{T}

t = - \frac{r}{T}

in ( 3.5 ), we get

\begin{matrix} (a a^{'}) \cdot w & = & (a a^{'})_{- 1} w \end{matrix}

\begin{matrix} = & (a (- 1) a^{'})_{- 1} w \end{matrix}

\begin{matrix} = & \sum_{m \geq 0} a_{- 2 + \frac{r}{T} - m} a_{- \frac{r}{T} + m}^{'} w + a_{- 1 - \frac{r}{T} - m}^{'} a_{- 1 + \frac{r}{T} + m} w \end{matrix}

\begin{matrix} = & 0 . \end{matrix}

Similarly, by substituting

u = a

v = b

p = - 1

s = - 1 + \frac{r}{T}

and

t = 1 - \frac{r}{T}

in ( 3.5 ), we have

\begin{matrix} (a b) \cdot w & = & (a b)_{0} w \end{matrix}

\begin{matrix} = & (a (- 1) b)_{0} w \end{matrix}

\begin{matrix} = & (1 - \frac{r}{T}) (a_{0} b)_{- 1} w + \sum_{m \geq 0} {a_{- 2 + \frac{r}{T} - m} b_{1 - \frac{r}{T} + m} + b_{- \frac{r}{T} - m} a_{- 1 + \frac{r}{T} + m}} w \end{matrix}

\begin{matrix} = & (1 - \frac{r}{T}) (a_{0} b) \cdot w, \end{matrix}

completing the proof.

Let

U

be a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

such that

(a a^{'}) \cdot u = 0

and

(a b) \cdot u = (1 - \frac{r}{T}) (a_{0} b) \cdot u

for

a \in A^{r}, a^{'} \in A^{T - r}, b \in B^{T - r}, u \in U, 0 < r \leq T - 1

We are going to construct a

\frac{1}{T} N

-graded

g

-twisted

V_{B}

-module

M = \prod_{n \in \frac{n}{T} N} M (n)

with

M (0) = U

as a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

First,

U

is a module for the Lie algebra

A^{0} \oplus B^{0} / \partial A^{0}

. Recall that

ℒ (C, g)_{(0)} = A^{0} \oplus B^{0} / \partial A^{0}

. For convenience, we set

ℒ (C, g)_{\leq 0} = ℒ (C, g)_{(0)} \oplus ℒ (g)_{-}

. Then

U

is an

ℒ (C, g)_{\leq 0}

-module under the following actions

\begin{matrix} a (n + \frac{r}{T} - 1) \cdot u & = & δ_{n + \frac{r}{T}, 0} a u, \end{matrix}

\begin{matrix} b (n + \frac{r}{T}) \cdot u & = & δ_{n + \frac{r}{T}, 0} b u \end{matrix}

for

a \in A^{r}, b \in B^{r}, n \geq 0

. Next, we form the induced

ℒ (C, g)

-module

\begin{matrix} M_{g} (U) = I n d_{ℒ (C, g)_{\leq 0}}^{ℒ (C, g)} U = U (ℒ (C, g)) \otimes_{U (ℒ (C, g)_{\leq 0})} U . \end{matrix}

(3.42)

We endow

U

with degree

0

, making

M_{g} (U)

\frac{1}{T} N

-graded restricted

ℒ (C, g)

-module.

By Proposition 3.13 ,

M_{g} (U)

is naturally a

g

-twisted

V_{ℒ (C)}

-module. In view of the P-B-W theorem, we may and we should consider

U

as the degree-zero subspace of

M_{g} (U)

We set

\begin{matrix} W_{g} (U) = s p a n {v_{n} u | v \in E, n \in \frac{1}{T} Z, u \in U} \subset M_{g} (U) \end{matrix}

(3.43)

and define

\begin{matrix} M_{B} (U) = M_{g} (U) / U (ℒ (C, g)) W_{g} (U) . \end{matrix}

(3.44)

Since

U (ℒ (C, g)) W_{g} (U)

is an

ℒ (C, g)

-submodule of

M_{g} (U)

, by Proposition 3.13

U (ℒ (C, g)) W_{g} (U)

is a

g

-twisted

V_{ℒ (C)}

-submodule. Then

M_{B} (U)

is a

g

-twisted

V_{ℒ (C)}

-module. Clearly,

M_{B} (U)

is generated by

\bar{U}

the image of

U

M_{B} (U)

. In fact,

M_{B} (U)

is a

g

-twisted

V_{B}

-module by the following:

Lemma 3.16. Let

(M, Y_{M})

be a

g

-twisted

V_{ℒ}

-module. Suppose that for

a \in A^{r}, a^{'} \in A, b \in B

with

0 \leq r \leq T - 1

\begin{matrix} Y_{M} (e, x) w & = & w, \end{matrix}

(3.45)

\begin{matrix} Y_{M} (a (- 1) a^{'}, x) w & = & Y_{M} (a a^{'}, x) w, \end{matrix}

(3.46)

\begin{matrix} Y_{M} (a (- 1) b, x) w & = & Y_{M} (a b, x) w \end{matrix}

(3.47)

for all

w \in K

, where

K

is a generating subspace of

M

. Then

M

is naturally a

g

-twisted

V_{B}

-module.

Proof. Recall that $E = s p a n {e - 1, a (- 1) a^{'} - a a^{'}, a (- 1) b - a b | a, a^{'} \in A, b \in B} \subset V_{ℒ (C)} .$ By ( 3.45 )-( 3.47 ), we have $K \subset A n n_{M} (E)$ . By using Propositions 3.5 , we have $A n n_{M} (I_{B}) = A n n_{M} (E)$ . Since $A n n_{M} (I_{B})$ is a $g$ -twisted $V_{ℒ (C)}$ submodule of $M$ and $M$ is generated by $K$ , we have $A n n_{M} (I_{B}) = M$ . This implies that $M$ is a $g$ -twisted $V_{B}$ -module.

One can see that Lemma 3.16 indeed implies that

M_{B} (U)

is naturally a

g

-twisted

V_{B}

-module. Furthermore we have:

Theorem 3.17. Let

U

be a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

such that

\begin{matrix} (a a^{'}) \cdot u = 0 and (a b) \cdot u = (1 - \frac{r}{T}) (a_{0} b) \cdot u \end{matrix}

(3.48)

for

a \in A^{r}, a^{'} \in A^{T - r}, b \in B^{T - r}, u \in U, r \neq 0

. Then

M_{B} (U)

is naturally a

g

-twisted

V_{B}

-module such that

M_{B} (U) (0) = U

Proof. To show that

M_{B} (U) (0) = U

, we must prove that

(U (ℒ (g)) W_{g} (U)) (0) = 0

First we show that

W_{g} (U) (0) = 0

. Notice that for

v \in (V_{ℒ (C)})_{(m)}^{r}

with

m \in Z

, we have

deg (v_{k + r / T}) = m - k - r / T - 1

for

k \in Z

. Then from the definition of

W_{g} (U)

W_{g} (U) (0)

is spanned by the vectors

(e - 1)_{- 1} u, (a (- 1) a^{'})_{- 1} u - (a a^{'})_{- 1} u, (a (- 1) b)_{0} u - (a b)_{0} u

for

u \in U

a \in A^{r}, a^{'} \in A^{T - r}

b \in B^{T - r}

with

0 \leq r \leq T - 1

. Since

e_{- 1}

acts as

e

(the identity of

A^{0}

) on

U

, we have

(e - 1)_{- 1} u = 0

for

u \in U

. If

r = 0

, by ( 3.5 ), we have

(a_{- 1} a^{'})_{- 1} u = a (- 1) a^{'} (- 1) u = a (a^{'} u) = (a a^{'}) u = (a a^{'})_{- 1} u,

and

(a (- 1) b)_{0} u = a (- 1) b (0) u = a (b u) = (a b) u = (a b)_{0} u .

Next, we assume that

r > 0

. By ( 3.5 ), we have

\begin{matrix} (a (- 1) a^{'})_{- 1} u & = & \sum_{i = 0}^{\infty} a (- 1 - i + \frac{r}{T}) a^{'} (i - 1 - \frac{r}{T}) u + \sum_{i = 0}^{\infty} a^{'} (- 2 - i - \frac{r}{T}) a (i + \frac{r}{T}) u \end{matrix}

\begin{matrix} = & a (- 1 + \frac{r}{T}) a^{'} (- 1 - \frac{r}{T}) u \end{matrix}

\begin{matrix} = & a^{'} (- 1 - \frac{r}{T}) a (- 1 - \frac{r}{T}) u \end{matrix}

\begin{matrix} = & 0 \end{matrix}

\begin{matrix} = & (a a^{'})_{- 1} u \end{matrix}

and

\begin{matrix} (a (- 1) b)_{0} u \end{matrix}

\begin{matrix} = & \sum_{i = 0}^{\infty} a (- 1 - i + \frac{r}{T}) b (i - \frac{r}{T}) u + \sum_{i = 0}^{\infty} b (- i - 1 - \frac{r}{T}) a (i + \frac{r}{T}) u - \frac{r}{T} (a_{0} b)_{- 1} u \end{matrix}

\begin{matrix} = & a (- 1 + \frac{r}{T}) b (- \frac{r}{T}) u - \frac{r}{T} (a_{0} b) \cdot u \end{matrix}

\begin{matrix} = & b (- \frac{r}{T}) a (- 1 + \frac{r}{T}) u + (a_{0} b)_{- 1} u - \frac{r}{T} (a_{0} b)_{- 1} u \end{matrix}

\begin{matrix} = & (1 - \frac{r}{T}) (a_{0} b) \cdot u \end{matrix}

\begin{matrix} = & (a b) \cdot u . \end{matrix}

Hence,

W_{g} (U) (0) = 0

Next, we show that

ℒ (C, g)_{\leq 0} W_{g} (U) \subset W_{g} (U)

. Recall from [LY] (Lemma 4.2) that

v_{i} E \subset E for v \in C = A \oplus B, i \geq 0 .

M_{g} (U)

is a

\frac{1}{T} N

-graded

ℒ (C, g)

-module with

U

as the degree-zero subspace, we have that

ℒ (C, g)_{\leq 0} U \subset U

. For

v \in C = A \oplus B, c \in E, m, t \in \frac{1}{T} Z, u \in U

, from the twisted commutator formula ( 3.6 ) (cf. ( 3.12 )), we have

v_{m} c_{t} u = c_{t} v_{m} u + \sum_{i \geq 0} (\binom{m}{i}) (v_{i} c)_{m - t - i} u .

These immediately imply that

ℒ (C, g)_{\leq 0} W_{g} (U) \subset W_{g} (U)

. Then

\begin{matrix} U (ℒ (C, g)) W_{g} (U) & = & U (ℒ (C, g)_{+}) U (ℒ (C, g)_{\leq 0}) W_{g} (U) \end{matrix}

\begin{matrix} = & U (ℒ (C, g)_{+}) W_{g} (U) \end{matrix}

\begin{matrix} = & W_{g} (U) + ℒ (C, g)_{+} U (ℒ (C, g)_{+}) W_{g} (U), \end{matrix}

which implies that

(U (ℒ (C, g)) W_{g} (U)) (0) = 0

. This completes the proof.

Next, we continue with Theorem 3.17 to construct and classify

\frac{1}{T} N

-graded simple

g

-twisted

V_{B}

-modules. Let

U

be a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

as in Theorem 3.17 . Let

J (U)

be the sum of all graded

ℒ (C, g)

-submodules of

M_{g} (U)

with trivial degree-zero subspaces. Then

J (U)

is the unique maximal graded

ℒ (C, g)

-submodule of

M_{g} (U)

with the property that

J (U) \cap U = 0

. Set

\begin{matrix} L_{g} (U) = M_{g} (U) / J (U), \end{matrix}

(3.49)

\frac{1}{T} N

-graded

g

-twisted

V_{B}

-module.

Lemma 3.18. Let

U

be a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

as in Theorem 3.17 . Then

L_{g} (U)

is a

\frac{1}{T} N

-graded

g

-twisted

V_{B}

-module such that

L_{g} (U) (0) = U

as a module for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

and such that for any nonzero graded submodule

W

L_{g} (U)

, we have

W (0) \neq 0

Furthermore, if

U

is a simple

B^{0} / A^{0} \partial A^{0}

-module,

L_{g} (U)

is a

\frac{1}{T} N

-graded simple

g

-twisted

V_{B}

-module.

Proof. It is similar to the proof of Theorem 4.12 in [LY] .

Lemma 3.19. Let

W = \prod_{n \in \frac{1}{T} Z} W (n)

be a

\frac{1}{T} N

-graded simple

g

-twisted

V_{B}

-module with

W (0) \neq 0

. Then

W \sim = L_{g} (W (0))

Proof. It is similar to the proof of Lemma 4.13 in [LY] .

To summarize we have:

Theorem 3.20. Let

H

be a complete set of equivalence class representatives of simple modules for the Lie

A^{0}

-algebroid

B^{0} / A^{0} \partial A^{0}

satisfying the condition that

(a a^{'}) U = 0, ((a b) - (1 - \frac{r}{T}) (a_{0} b)) U = 0

for

a \in A^{r}, a^{'} \in A^{T - r}, b \in B^{T - r}

with

0 < r < T

. Then

{L_{g} (U) | U \in H}

is a complete set of equivalence class representatives of

\frac{1}{T} N

-graded simple

g

-twisted

V_{B}

-modules.

Proof. It is similar to the proof of Theorem 4.14 in [LY] .

Remark 3.21. Taking

g = 1

the identity map of

V_{B}

, we recover Theorem 4.14 of [LY] : If

H

is a complete set of equivalence class representatives of simple modules for the Lie

A

-algebroid

B / A \partial A

, then

{L_{1} (U) | U \in H}

is a complete set of equivalence class representatives of

N

-graded simple

V_{B}

-modules.

References

S. Berman, Y. Billig and J. Szmigielski, Vertex operator algebras and the representation theory of toroidal algebras, in: Recent Developments in Infinite-Dimensional Lie Algebras and Conformal Field Theory (Charlottesville, VA, 2000), Contemporary Math. 297, Amer. Math. Soc., Providence, 2002, 1-26.
S. Berman, C. Dong and S. Tan, Representations of a class of lattice type vertex algebras, J. Pure Applied Algebra 176 (2002), 27-47.
R. E. Borcherds, Vertex algebras, Kac-Moody algebras, and the Monster, Proc. Natl. Acad. Sci. USA 83 (1986), 3068-3071.
P. Bressler, Vertex algebroids I, arXiv: math.AG/0202185.
P. Bressler, Vertex algebroids II, arXiv: math.AG/0304115.
C. Dong, Twisted modules for vertex algebras associated with even lattices, J. Algebra 165 (1994), 91-112.
C. Dong, H.-S. Li and G. Mason, Regularity of rational vertex operator algebras, Advances in Math. 132 (1997), 148-166.
C. Dong, H.-S. Li and G. Mason, Twisted representations of vertex operator algebras, Math. Annalen 310 (1998), 571-700.
C. Dong, H.-S. Li and G. Mason, Vertex Lie algebra, vertex Poisson algebras and vertex algebras, in: Recent Developments in Infinite-Dimensional Lie Algebras and Conformal Field Theory, Proceedings of an International Conference at University of Virginia, May 2000, Contemporary Math. 297 (2002), 69-96.
A. J. Feingold, I. B. Frenkel and J. F. X. Ries, Spinor Construction of Vertex Operator Algebras, Triality, and $E_{8}^{(1)}$ , Contemporary Math. 121, 1991.
E. Frenkel and D. Ben-Zvi, Vertex Algebras and Algebraic Curves, Mathematical Surveys and Monographs, Vol. 88, Amer. Math. Soc., Providence, 2001.
B. Feigin and E. Frenkel, A family of representations of affine Lie algebras, Russian Math. Surveys 43 (1988), 221-222.
B. Feigin and E. Frenkel, Affine Kac-Moody algebras and semi-infinite flag manifolds, Commun. Math. Phys. 128 (1990), 161-189.
B. Feigin and E. Frenkel, Representations of affine Kac-Moody Lie algebras and bosonization, in: Physics and Mathematics of Strings, World Scientific, 1990, 271-316.
I. B. Frenkel, J. Lepowsky and A. Meurman, Vertex Operator Algebras and the Monster, Pure and Applied Math., Vol. 134, Academic Press, Boston, 1988.
V. Gorbounov, F. Malikov and V. Schechtman, Gerbes of chiral differential operators, II, Vertex algebroids, Invent. Math. 155 (2004), 605-680.
J. Lepowsky, Calculus of twisted vertex operators, Proc. Natl. Acad. Sci. USA 82 (1985), 8295-8299.
J. Lepowsky and H.-S. Li, Introduction to Vertex Operator Algebras and Their Representations, Progress in Math. 227, Birkh $\ddot{a}$ user, Boston, 2003.
H.-S. Li, Representation theory and tensor product theory for vertex operator algebras, Ph.D. thesis, Rutgers University, 1994.
H.-S. Li, Local systems of vertex operators, vertex superalgebras and modules, J. Pure Appl. Alg. 109 (1996), 143-195; hep-th/9406185.
H.-S. Li, Local systems of twisted vertex operators, vertex operator superalgebras and twisted modules, in: Moonshine, the Monster and Related Topics, Proc. Joint Summer Research Conference, Mount Holyoke, 1994, ed. by C. Dong and G. Mason, Contemporary Math. 193, Amer. Math. Soc., Providence, 1996, 203-236.
H.-S. Li and G. Yamskulna, On certain vertex algebras and their modules associated with vertex algebroids, J. Algebra 283 (2005), 367-398.
F. Malikov and V. Schechtman, Chiral de Rham complex, II, axXiv: math.AG/9901065.
F. Malikov, V. Schechtman and A. Vaintrob, Chiral de Rham complex, axXiv: math.AG/9803041.
M. Wakimoto, Fock representations of affine Lie algebra $A_{1}^{(1)}$ , Commun. Math. Phys. 104 (1986), 605-609.