Received 8 August 2015; accepted 24 April 2016; published 27 April 2016

1. Introduction
Weak Hopf algebras were introduced by G. Böhm and K. Szlachányi as a generalization of usual Hopf algebras and groupoid algebras [1] [2] . A weak Hopf algebra is a vector space that has both algebra and coalgebra structures related to each other in a certain self-dual fashion and possesses an analogue of the linearized inverse map [3] - [5] . The main difference between ordinary and weak Hopf algebras comes from the fact that the comultiplication of the latter is no longer required to preserve the unit (equivalently, the counit is not requires to be a homomorphism) and results in the existence of two canonical subalgebras playing the role of “non- commutative bases”.
Paper [6] was shown what is a weak Hopf algebra in the braided category of modules over a weak Hopf algebra. In [7] we prove a Fundamental Theorem of Hopf modules for the categorical weak Hopf algebra motivation to study quasitriangular weak Hopf algebras is the so-called biproduct construction and interpreted in the terms of braided categories. More precisely, we are interested in a specific type of quaitriangular weak Hopf algebras.
we prove the Fundamental Theorem for Hopf modules in the category of Yetter-Drinfeld modules according to the fact that the matrix R gives rise to a natural braiding for
and
. Furthermore
is also a right H-Hopf module in the category Yetter-Drinfeld modules. Using this result we obtain the existence and uniqueness of integrals for a finite dimensional weak Hopf algebra in
.
2. Preliminaries
Throughout this paper we use Sweedler’s notation for comultiplication, writing
. Let k be a fixed field and all weak Hopf algebras are finite dimensional.
Definition 1. A weak Hopf algebra is a vector space L with the structure of an associative unital algebra
with multiplication
and unit
and a coassociative coalgebra
with comultiplication
and counit
such that
1) The comultiplication
is a (not necessarily unit-preserving) homomorphism of algebras such that

2) The counit satisfies the following identity

3) There is a linear map
called an antipode, such that, for all ![]()
![]()
![]()
![]()
The linear map defined in the above equations are called target and source counital maps and denoted by
and
respectively:
![]()
![]()
For all
, we have
![]()
![]()
We will briefly recall the necessary definitions and notions on the weak Hopf algebras.
Definition 2. A quasitriangular weak Hopf algebra is a pair
where L is a weak Hopf algebra and
(called the R-matrix) satisfying the following conditions:
![]()
for all
, where
denotes the conditions apposite to
,
![]()
![]()
where
, etc. as usual, and such that there exits
with
where we write
. By [3] , we can obtain the following results.
Proposition 2.1. For any quasitriangular weak Hopf algebra
, we have
![]()
3. Weak Hopf Algebras in the Yetter-Drinfeld Module Category
Let L be a quasitriangular weak Hopf algebra with a bijective antipode
. Suppose H is a weak Hopf algebra in
. Paper [7] show that H is also a weak Hopf algebra in
with a left L-coaction via
. Bing-liang and Shuan-hong introduce the definition of Weak Hopf algebra in the braided monoidal category
in [6] . Moreover they have showed that if H is a finite-dimensional weak Hopf algebra in
, then its dual
is a weak Hopf algebra in
.
Definition 3. Let
be a quasitriangular weak Hopf algebra. An object
is called a weak bialgebra in this category if it is both an algebra and a coalgebra satisfying the following conditions:
1)
and
are not necessarily unit-preserving, such that
![]()
2) H is a left L-module algebra and left L-module coalgebra if H is a left L-module via
such that
![]()
![]()
3) H is a left L-comodule algebra and left L-comodule coalgebra if H is a left L-comodule via
such that
![]()
4) Furthermore, H is called a weak Hopf algebra in
if there exists an antipode
(here S is left L-linear and left L-colinear i.e., S is a morphism in the category of
) satisfying
![]()
Similar to the definition of weak Hopf algebra, we denote
If
one can obtain
. According to the definitions of
one obtains explicit expressions for these coproducts
![]()
Paper [7] give the following results:
Proposition 3.1. Suppose H is a weak Hopf algebra in
. For all
we have the identities
![]()
Since a weak Hopf algebra H in the weak Yetter-Drinfeld categories
is both algebra and coalgebra, one can consider modules and comodules over H. As in the theory of Hopf algebras, an H-Hopf module is an H-module which is also an H-comodule such that these two structures are compatible (the action “commutes” with coaction):
Definition 4. Let H be a weak Hopf algebra in
. A right H-Hopf module M in
is an object
such that it is both a right H-module
and a right H-comodule via
and the following equations hold for
:
1) ![]()
2) ![]()
3) ![]()
4) ![]()
5) ![]()
We remark that
is a right H-module by
and a right H-comodule
. The condition (1) means that the H-comodule structure
is H-linear, or equivalently the H-module structure map
is H- colinear. Also, (4) (resp. (2)) ![]()
is L-colinear (resp. L-linear); (3)(resp. (5)) ![]()
is L-colinear (resp. L-linear).
Example 3.2. H itself is a right H-Hopf module (in
) in the natural way. If V is an object in
, then so is
by
and
. It is also both a right H-module and a right H-comodule by
and
. One easily checks that
is an right H-Hopf module.
when H is a weak Hopf algebra in
and M a right H-Hopf module in
, we prove the Fundamental Theorem 3.3 [7] . Furthermore we will show
is a L-subcomodule of M.
Applying
we obtain
![]()
For
we do a calculation:
![]()
This implies that
. So
.
It is clearly to prove F is a left L-colinear by the following equation
![]()
Furthermore we can obtain the Structure Theorem for right H-Hopf modules in the category of Yetter- Drinfeld modules.
Theorem 3.3. If H is a weak Hopf algebra in
and M is a right H-Hopf module in
,
is defined as above. Then
1) Let
. Then
. If
and
, Then
and
.
2) The map
is an isomorphism of Hopf modules. The inverse map is given by
.
4. Fundamental Theorem for H* in ![]()
In [4]
has the contragredient left L-module structure by
![]()
Since H is a finite-dimensional left L-comodule,
has the transposed right L-comodule structure and so it becomes a left L-comodule via
![]()
i.e.
Now assume that H is finite-dimensional. We will show that
becomes a right H-Hopf module in
. First
is a right H-module by
![]()
Second,
is a right H-comodule using the identification
,
as follows:
![]()
That is
means
![]()
Proposition 4.1.
is a right H-comodule by
.
Proof. Now for
, we have
![]()
It implies that
.
Accord to
we have
. Applying the equality
we obtain
![]()
Hence
. Thus
becomes a right H-comodule.
Theorem 4.2. With the notation as above, then
is a right H-Hopf module in
. Moreover,
.
Proof. Now we prove that
is a right H-Hopf module. First we will show that
Since for
,
![]()
Next we want to check
for
. Since for ![]()
![]()
Applying the equality
for ![]()
![]()
It implies that
. Using the equality
we compute
![]()
Finally we show that
. Since for ![]()
![]()
From all above,
is a right H-Hopf module in
.
Applying Theorem 4.2 we can obtain the following result.
Corollary 4.3.
is defined a right H-Hopf module in
as above, then
.
5. Applications
As a consequence the space of coinvariants of the finite dimensional Hopf algebra is free of rank one. This is the case for the weak Hopf algebra in the category of the Yetter-Drinfeld modules.
Theorem 5.1. If H is a finite-dimensional weak Hopf algebra in
. Then
1)
.
.
2) The map
is an right H-module and an right H-comodules isomorphism. In particular H is a Frobenius weak Hopf algebra with Frobenius map
.
3) There exist a right integral t in H,
and a group-like elment
in L such that for all
![]()
a)
,
b)
,
c) ![]()
d)
, for all
.
4) The map
is a left L-semilinear and a left L-semicolinear in the sense that for all
, ![]()
Proof. 1) Since
is a right H-Hopf module in
, we have
,
. Since
, it follows that
.
2) Choose
. Then by (1)
is an right H-modules and an right H-comodules. Thus H is Frobenius weak Hopf algebra.
3) a) Since
, there is a unique element t in H such that
, i.e.
. For all
we have
It follows that
. So t is a right integral in H.
b) We remark that
for all
from Theorem 3.3. This implies
,
i.e.
for some
, by
.
c) From Theorem 3.3 we have
is a right L-comodule, i.e.
. By ![]()
we can obtain
for some group-like element
in L. This implies that
.
d) Applying
we have
![]()
This means
, for all
.
4) For all
we have
![]()
This implies
.
![]()
Acknowledgements
The author would like to thank the referee for many suggestions and comments, which have improved the overall presentations.
Funding
Research supported by the Project of Shandong Province Higher Educational Science and Technology Program
(J12LI07) and the Project of National Natural Science Foundation of China (51078225).