Introduction and preliminaries

Lau [10] introduced a wide class of Banach algebras, called F-algebras, and studied the notion of left amenability for these algebras. In [12], Nasr-Isfahani introduced the concept of inner amenability for Lau algebras. A Lau algebra A was said to be inner amenable if there exists a topological inner invariant mean on the \(W^*\)-algebra \(A^*\), that is, a positive linear functional m of norm 1 on \(A^*\), such that \(m(f.a) = m(a.f)\) for all \(f\in A^*\) and all \(a\in P_1(A) = \{a \in A : \Vert a\Vert = 1\}\) (or equivalently, for all \(a\in A\)). Commutative Lau algebras, such as the Fourier algebra A(G) of a locally compact group G, are examples of inner amenable algebras. In addition, the group algebra \(L^1(G)\) of any locally compact group G is inner amenable.

Recently, Jabbari et al. [8] have introduced the notion of \(\varphi\)-inner amenability for a Banach algebra A, where \(\varphi \in \Delta (A)\), the character space of A. A Banach algebra A was said to be \(\varphi\)-inner amenable if there exists a \(m\in A^{**}\) satisfying \(m(\varphi )=1\) and \(m(f.a)=m(a.f) (a\in A, f\in A^*)\). A is said to be character inner amenable if and only if A is \(\varphi\)-inner amenable for every \(\varphi \in \Delta (A)\).

In [6], Ebrahimi Vishki and Khoddami have investigated the character inner amenability for certain products of Banach algebras consist of projective tensor product \(A\widehat{\otimes }B\), Lau product \(A\times _{\theta }B\), where \(\theta \in \Delta (B)\) and the module extension \(A\oplus X\). For instance, they showed that the projective tensor product \(A\widehat{\otimes }B\) is character inner amenable if and only if both A and B are character inner amenable.

Let \(\mathfrak {A}\) and A be Banach algebras, such that A be a Banach \(\mathfrak {A}\)-bimodule with compatible actions

$$\begin{aligned} \alpha .(ab)=(\alpha .a)b , (ab).\alpha =a(b.\alpha ), \alpha .(\beta .a)=(\alpha \beta ).a,(a.\beta ).\alpha =a.(\beta \alpha ), \end{aligned}$$

for all \(a,b\in A\) and \(\alpha \in \mathfrak {A}\).

Let X be a Banach A-bimodule and a Banach \(\mathfrak {A}\)-bimodule with compatible left actions defined by

$$\begin{aligned} \alpha .(a.x)=(\alpha .a ).x,~ a.(\alpha .x)=(a.\alpha ).x, ~ (\alpha .x).a=\alpha .(x.a)~ (a\in A , \alpha \in \mathfrak {A}, x\in X), \end{aligned}$$

and similar for the right or two-sided actions. Then, we say that X is a Banach A-\(\mathfrak {A}\)-module.

Let \(A\widehat{\otimes }A\) be the projective tensor product of A and A which is a Banach A-bimodule and a Banach \(\mathfrak {A}\)-bimodule by the following actions:

$$\begin{aligned} \alpha .(a\otimes b)=(\alpha .a)\otimes b, c.(a\otimes b)=(ca)\otimes b (\alpha \in \mathfrak {A}, a,b,c\in A), \end{aligned}$$

similarly for the right actions. Let \(I_{A\widehat{\otimes }A}\) be the closed ideal of \(A\widehat{\otimes }A\) generated by elements of the form:

$$\begin{aligned} \{a.\alpha \otimes b-a\otimes \alpha .b |\alpha \in \mathfrak {A}, a,b\in A\}. \end{aligned}$$
(1)

Consider the map \(\omega _A\in \mathcal {L}(A\widehat{\otimes }A,A)\) defined by \(\omega _A(a\otimes b)=ab\) and extended by linearity and continuity. Let \(J_A\) be the closed ideal of A generated by

$$\begin{aligned} \omega (I_{A\widehat{\otimes }A})=\{(a.\alpha )b-a(\alpha .b)\quad |a,b\in A, \alpha \in \mathfrak {A}\}. \end{aligned}$$
(2)

Then, the module projective tensor product \(A\widehat{\otimes }_{\mathfrak {A}}A\), which is \((A\widehat{\otimes }A)/I_{A\widehat{\otimes }A}\) by [16], and the quotient Banach algebra \(A/J_A\) are both Banach A-bimodules and Banach \(\mathfrak {A}\)-bimodules. In addition, \(A/J_A\) is A-\(\mathfrak {A}\)-module with compatible actions when A acts on \(A/J_A\) canonically.

Let A and \(\mathfrak {A}\) be Banach algebras, such that A is a Banach \(\mathfrak {A}\)-bimodule with compatible actions. Let \(\varphi \in \Delta (\mathfrak {A})\cup \{0\}\) and consider the set \(\Omega _{A,\varphi }\) of linear continuous maps \(\phi :A\rightarrow \mathfrak {A}\), such that

$$\begin{aligned} \phi (ab)=\phi (a)\phi (b), \phi (\alpha .a)=\phi (a.\alpha )=\varphi (\alpha )\phi (a)(a,b\in A, \alpha \in \mathfrak {A}). \end{aligned}$$
(3)

The concept of module \((\phi ,\varphi )\)-amenability and module character amenability for Banach algebra A, where \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\) were introduced by Bodaghi and Amini in [4].

Our aim in this paper is to introduce and study module \((\phi ,\varphi )\)-inner amenability and module character inner amenability of Banach algebras. We characterize \((\phi ,\varphi )\)-inner amenability and prove some hereditary properties. Moreover, we investigate that module \((\phi ,\varphi )\)-inner amenability for certain class of Banach algebras consists of projective tensor product \(A\widehat{\otimes }B\), \(A\oplus _{\infty }B\), and \(A\oplus _p B\), the \(l^p\)-direct sum of A and B, where \(1\le p<\infty\).

Characterization and hereditary properties

We commence this section with the following definition.

Definition 2.1

Let A be a Banach \(\mathfrak {A}\)-bimodule and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). Then, A is called module \((\phi ,\varphi )\)-inner amenable if there exists \(m\in A^{**}\), such that \(m(\varphi \circ \phi )=1\), \(m(f.a)=m(a.f)\) and \(m(\alpha .f)=m(f.\alpha )\) for all \(a\in A, f\in A^*\) and \(\alpha \in \mathfrak {A}\). A Banach \(\mathfrak {A}\)-bimodule A is called module character inner amenable if it is module \((\phi ,\varphi )\)-inner amenable for each \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\).

We note that if \(\mathfrak {A}=\mathbb {C}\) and \(\varphi\) is the identity map, then the module \((\phi ,\varphi )\)-inner amenability and module character inner amenability coincide with \(\phi\)-inner amenability and character inner amenability (see [8] and [6]).

The next theorem characterizes module \((\phi ,\varphi )\)-inner amenability of Banach algebras that is analogue of Proposition 2.1 of [5] on module \((\phi ,\varphi )\)-amenable Banach algebras.

Theorem 2.2

Let A be a Banach \(\mathfrak {A}\)-bimodule and let \(\varphi \in \Delta ({\mathfrak {A}})\) and \(\phi \in \Omega _A\). Then, the following statements are equivalent:

  1. (i)

    A is module \((\phi ,\varphi )\)-inner amenable;

  2. (ii)

    There exists a bounded net \((a_i)_i\) in A such that \(\Vert aa_i-a_ia\Vert \longrightarrow 0, \Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0~(a\in A, \alpha \in \mathfrak {A})\) and \(\varphi \circ \phi (a_i)=1\) for all i;

  3. (iii)

    There exists a bounded net \((a_i)_i\) in A such that \(\Vert aa_i-a_ia\Vert \longrightarrow 0, \Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0~(a\in A, \alpha \in \mathfrak {A})\) and \(\varphi \circ \phi (a_i)\longrightarrow 1\).

Proof

(iii) \(\Rightarrow\)(i) Assume that a net \((a_i)_i\) exists. Let m be a \(w^*\)-cluster point of the net \((a_i)_i\) in \(A^{**}\). Then, \(\langle m,\varphi \circ \phi \rangle =\lim _i\langle \varphi \circ \phi ,a_i\rangle =1.\) For every \(a\in A\) and \(f\in A^*\), we have

$$\begin{aligned} \langle m,f.a\rangle&=\lim _i\langle f.a,a_i\rangle =\lim _i\langle f,aa_i\rangle \\&=\lim _i\langle f,aa_i-a_ia\rangle +\lim _i\langle f,a_ia\rangle \\&=\lim _i\langle a.f,a_i\rangle =\langle m,a.f\rangle , \end{aligned}$$

and similarly, we have \(\langle m,f.\alpha \rangle =\langle m,\alpha .f\rangle (\alpha \in {\mathfrak {A}})\). Therefore, A is module \((\phi ,\varphi )\)-inner amenable.

(i)\(\Rightarrow\) (ii) Suppose that A is module \((\phi ,\varphi )\)-inner amenable. Then, there exists \(m\in A^{**}\) such that \(m(\varphi \circ \phi )=1\), \(m(f.a)=m(a.f)\), and \(m(\alpha .f)=m(f.\alpha )\) for all \(a\in A, f\in A^*\) and \(\alpha \in \mathfrak {A}\). Choose a net \((u_{\beta })_{\beta }\) in A with \(u_{\beta }\longrightarrow m\) in the \(w^*\)-topology of \(A^{**}\) and \(\Vert u_{\beta }\Vert \le \Vert m\Vert\) for all \(\beta\). Since \(\langle \varphi \circ \phi ,u_{\beta }\rangle \longrightarrow \langle \varphi \circ \phi ,m\rangle =1\), passing to a subnet and replacing \(u_{\beta }\) by \(\big (1/\varphi \circ \phi (u_{\beta })\big )u_{\beta },\) we may assume that \(\varphi \circ \phi (u_{\beta })=1\) and \(\Vert u_{\beta }\Vert \le \Vert m\Vert +1\) for all \(\beta\). Consider the product space \(A^A\) endowed with the product of norm topological. Define a linear map \(T:A\rightarrow A^A\) by \(T(b)=\big (ab-ba+\alpha .b-b.\alpha \big )_{a\in A},\) for all \(b\in A\) and \(\alpha \in \mathfrak {A}\). Let

$$\begin{aligned} B=\{b\in A:\Vert b\Vert \le \Vert m\Vert +1~ \mathrm \; {and}\; ~ \varphi \circ \phi (b)=1\}\subseteq A. \end{aligned}$$

Clearly, B is convex and so T(B) is a convex subset of \(A^A\). For every \(f\in A^*\), we have

$$\begin{aligned} \langle f,au_{\beta }-u_{\beta }a+\alpha .u_{\beta }-u_{\beta }.\alpha \rangle&=\langle f,au_{\beta }\rangle -\langle f,u_{\beta }a\rangle +\langle f,\alpha .u_{\beta }\rangle +\langle f,u_{\beta }.\alpha \rangle \\&= \langle f.a,u_{\beta }\rangle -\langle a.f,u_{\beta }\rangle +\langle f.\alpha ,u_{\beta }\rangle +\langle \alpha .f,u_{\beta }\rangle \\&\rightarrow \langle m,f.a\rangle -\langle m,a.f\rangle +\langle m,f.\alpha \rangle -\langle m,\alpha .f \rangle \\&=0. \end{aligned}$$

This product of weak topologies coincide with topology on \(A^A\) (see Theorem 4.3 of [17]). By Mazur’s theorem, \(0\in \overline{T(B)}^w=\overline{T(B)}^{\Vert .\Vert }.\) Therefore, there exists a bounded net \((a_i)_i\) in A, such that \(\varphi \circ \phi (a_i)=1\) and

$$\begin{aligned} \Vert aa_i-a_ia\Vert \longrightarrow 0, \Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0 (a\in A, \alpha \in \mathfrak {A}). \end{aligned}$$

(ii)\(\Rightarrow\) (iii) It is clear. \(\square\)

Definition 2.3

We say that the Banach algebra \(\mathfrak {A}\) acts trivially on A from the left (right) if there is a multiplicative linear functional f on \(\mathfrak {A}\), such that \(\alpha .a=f(\alpha )a\) (resp. \(a.\alpha =f(\alpha )a\)) for all \(\alpha \in \mathfrak {A}\) and \(a\in A\).

For the proof of the following result, we refer to Lemma 3.13 of [1].

Lemma 2.4

Let \(\mathfrak {A}\) acts on A trivially from the left or right and \(A/J_A\) has a right bounded approximate identity, then for each \(\alpha \in \mathfrak {A}\) and \(a \in A\) we have \(f(\alpha )a-a.\alpha \in J_A\).

Let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). Clearly, \(\phi \big ((a.\alpha )b-a(\alpha .b)\big )=0(\alpha \in \mathfrak {A}, a,b\in A)\), and hence, \(\phi =0\) on \(J_A\) and \(\tilde{\phi }:A/J_A\longrightarrow \mathfrak {A}\) given by \(\tilde{\phi }(a+J_A)=\phi (a)\) is well defined. Then, \(\tilde{\phi }\in \Omega _{A/J_A}\).

Proposition 2.5

Let A be a Banach \(\mathfrak {A}\)-bimodule, and let \(\mathfrak {A}\) acts on A trivially from the left and \(A/J_{A}\) has a bounded approximate identity. Then \(A/J_{A}\) is \(\varphi \circ \tilde{\phi }\)-inner amenable for every \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\).

Proof

Let \((e_{\alpha }+J_A)_{\alpha }\) be a bounded approximate identity of \(A/J_{A}\) and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). Then, \(\varphi \circ \tilde{\phi }(e_{\alpha }+J_A)\longrightarrow 1\). Clearly

$$\begin{aligned} \Vert (a+J_A)(e_{\alpha }+J_A)-(e_{\alpha }+J_A)(a+J_A)\Vert \longrightarrow 0 \end{aligned}$$

for all \(a+J_A\in A/J_{A}\). By Proposition 2.2 of [6], \(A/J_{A}\) is \(\varphi \circ \tilde{\phi }\)-inner amenable. \(\square\)

Note that in the above proposition, both left and right actions of \(\mathfrak {A}\) on \(A/J_{A}\) are trivial, by Lemma 2.4. Therefore, for every \(\alpha \in \mathfrak {A}\), we have

$$\begin{aligned} \Vert \alpha .(e_{\alpha }+J_A)-(e_{\alpha }+J_A).\alpha \Vert&=\Vert \alpha .e_{\alpha }+J_A-e_{\alpha }.\alpha +J_A\Vert \\&=\Vert f(\alpha )(e_{\alpha }+J_A)-f(\alpha )(e_{\alpha }+J_A)\Vert =0. \end{aligned}$$

Then, the net \((e_{\alpha }+J_A)_{\alpha }\) satisfies condition (iii) of Theorem 2.2, and hence, \(A/J_{A}\) is module \((\tilde{\phi },\varphi )\)-inner amenable.

Remark 2.6

A inverse semigroup is a discrete semigroup S, such that for each \(s\in S\), there is a unique element \(s^*\in S\) with \(ss^*s=s\) and \(s^*ss^*=s^*\). An element \(e\in S\) is called an idempotent if \(e^2=e^*=e\). The set of idempotent elements of S is denoted by \(E_S\). Define the relation \(\le\) on \(E_S\) by \(e\le d \Leftrightarrow ed=e~ (e,d\in E_S).\) Then, \(E_S\) is a commutative subsemigroup of S, and \(l^1(E_S)\) may be regarded as a subalgebra of \(l^1(S)\).

Let s be an inverse semigroup with the set of idempotents \(E_S\). We let \(l^1(E_S)\) acts on \(l^1(S)\) by multiplication from the right and trivially from the left, that is

$$\begin{aligned} \delta _e.\delta _s=\delta _s~ \delta _s.\delta _e=\delta _{se}=\delta _s*\delta _e~ (e\in E_S, s\in S). \end{aligned}$$

By these actions, \(l_1(S)\) becomes a Banach \(l_1(E_S)\)-module. In this case

$$\begin{aligned} J_{l^1(S)}= \{\delta _{set}-\delta _{st} | e\in E_S , s,t\in S\}. \end{aligned}$$

We consider an equivalence relation on S as follows \(s\approx t \Leftrightarrow \delta _s-\delta _t\in J_{l^1(S)}~(s,t\in S).\) For inverse semigroup S, the quotient semigroup \(S/\approx\) is discrete group and so \(l^1(S/\approx )\) has an identity (see [2, 13]). Indeed, \(S/\approx\) is homomorphic to the maximal group homomorphic image \(G_S\) of S (see [11, 14]). It is also shown in Theorem 3.3 of [15] that \(l^1(S)/J_{l^1(S)}\cong l^1(S/\approx )=l^1(G_S)\) is a commutative \(l^1(E_S)\)-bimodule with the following actions:

$$\begin{aligned} \delta _e.\delta _{[s]}=\delta _{[s]} ,\delta _{[s]}.\delta _e=\delta _{[se]}(s\in S , e\in E_S), \end{aligned}$$

where [s] denotes the equivalence class of s in \(G_S\).

It is shown in [4] that the maps \(\varphi\) and \(\phi\) satisfying (3) exist for \(l^1(S).\)

Example 2.7

Let S be an inverse semigroup with the set of idempotents \(E_S\). Consider \(l^1(S)\) as a Banach module over \(l^1(E_S)\) with the trivial left action and natural right action. Then, by Proposition 2.5, \(l^1(G_S)\) is \(\varphi \circ \tilde{\phi }\)-inner amenable (module \((\tilde{\phi },\varphi )\)-inner amenable) for all \(\varphi \in \Delta (l^1(E_S))\) and \(\phi \in \Omega _{l^1(S)}\).

Example 2.8

Let A be a commutative Banach algebra and commutative \(\mathfrak {A}\)-bimodule (i.e., \(\alpha .a=a.\alpha ~(a\in A, \alpha \in \mathfrak {A}))\). Let \(\varphi \in \Delta (\mathfrak {A}), \phi \in \Omega _{A}\) and let \(a\in A\) be such that \(\varphi \circ \phi (a)=1\). put \(m=\widehat{a}.\) Then, \(m(\varphi \circ \phi )=\widehat{a}(\varphi \circ \phi )=\varphi \circ \phi (a)=1\) and clearly, \(m(f.a)=m(a.f)\) and \(m(\alpha .f)=m(f.\alpha )\) for all \(a\in A, \alpha \in \mathfrak {A}.\) Therefore, A is module \((\phi ,\varphi )\)-inner amenable. In particular, if S is a commutative inverse semigroup, then \(l^1(S)\) is commutative and commutative \(l^1(E_S)\)-bimodule. Therefore, \(l^1(S)\) is module \((\phi ,\varphi )\)-inner amenable for all \(\varphi \in \Delta (l^1(E_S))\) and \(\phi \in \Omega _{l^1(S)}\).

Example 2.9

Let \(S=(\mathbb {N},\wedge )\) be the inverse semigroup of positive integers with the minimum operation. Let \(A=l^1(S), \mathfrak {A}=l^1(E_S)\) and \(\mathfrak {A}\) acts on A by the following actions:

$$\begin{aligned} \delta _e.\delta _s= \delta _s.\delta _e=\delta _{se}\quad (e\in E_S, s\in S). \end{aligned}$$

A is module amenable (see page 42 of [3]). By Theorem 2.1 of [4], A is module character amenable. Let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _{A}\). Then, there exist \(m\in A^{**}\), such that \(m(f.a)=\varphi \circ \phi (a)m(f), m(f.\alpha )=\varphi (\alpha )m(f)\) and \(m(\varphi \circ \phi )=1\) for every \(f\in A^*, a\in A\) and \(\alpha \in \mathfrak {A}\). Let \((a_{\alpha })_{\alpha }\) be a net in A converging to m in the \(w^*\)-topology of \(A^{**}\). Since A is commutative and commutative \(\mathfrak {A}\)-bimodule, for every \(a\in A, f\in A^*\)

$$\begin{aligned} m(f.a)&=\lim _{\alpha }\langle f.a,a_{\alpha }\rangle =\lim _{\alpha }\langle f,aa_{\alpha }\rangle \\&=\lim _{\alpha }\langle f,a_{\alpha }a\rangle =\lim _{\alpha }\langle a.f,a_{\alpha }\rangle \\&=m(a.f). \end{aligned}$$

Similarly, for every \(\alpha \in \mathfrak {A}\) and \(f\in A^*\), we have

$$\begin{aligned} m(f.\alpha )=m(\alpha .f). \end{aligned}$$

Thus, A is module \((\phi ,\varphi )\)-inner amenable. Therefore, A is module character inner amenable.

The proof of the following proposition is adapted from that of Proposition 2.3 of [4].

Proposition 2.10

Let A and B be Banach \(\mathfrak {A}\)-bimodules and let h be an \(\mathfrak {A}\)-module homomorphism with dense range. If \(\phi \in \Omega _B, \varphi \in \Delta (\mathfrak {A})\) and A is module \((\phi \circ h,\varphi )\)-inner amenable, then B is module \((\phi ,\varphi )\)-inner amenable.

Proof

Let \(m\in A^{**}\) be such that \(m\big (\varphi \circ (\phi \circ h)\big )=1, m(f.a)=m(a.f)\) and \(m(f.\alpha )=m(\alpha .f)\) for all \(a\in A, f\in A^*\) and \(\alpha \in \mathfrak {A}\). Define \(m_B\in B^{**}\) by \(m_B(g)=m(g\circ h) (g\in B^*)\). We show that \(m_B\big (g.b\big )=m_B\big (b.g\big )~ (b\in B).\) For see this let \(b\in B\) be such that \(h(a)=b\). One can easily check that \((g.h(a))\circ h=(g\circ h).a\) and \((h(a).g)\circ h=a.(g\circ h).\) Hence for every \(g\in B^*\)

$$\begin{aligned} m_B(g.b)&=m_B\big (g.h(a)\big )=m\big ((g.h(a))\circ h\big )\\&=m\big ((g\circ h).a\big )=m\big (a.(g\circ h)\big )\\&=m\big ((h(a).g)\circ h\big )=m_B\big (h(a).g\big )\\&=m_B(b.g). \end{aligned}$$

By density of the range of h and the continuity of h, we conclude that \(m_B\big (g.b\big )=m_B\big (b.g\big ) (b\in B).\) In addition, for every \(\alpha \in \mathfrak {A}\), we have

$$\begin{aligned} m_B(g.\alpha )&=m\big ((g.\alpha )\circ h\big )=m\big ((g\circ h).\alpha \big )\\&=m\big (\alpha .(g\circ h)\big )=m\big ((\alpha .g)\circ h)\big )\\&=m_B(\alpha .g). \end{aligned}$$

Furthermore, \(m_B(\varphi \circ \phi )=m\big ((\varphi \circ \phi )\circ h\big )=m\big (\varphi \circ (\phi \circ h)\big )=1.\) Therefore, B is module \((\phi ,\varphi )\)-inner amenable.\(\square\)

Corollary 2.11

Let A and B be Banach \(\mathfrak {A}\)-bimodules and let h be an \(\mathfrak {A}\)-module homomorphism with dense range. Then the module character inner amenability of A implies the module character inner amenability of B. In particular, if A is module character inner amenable, then so is \(A/J_A\).

The proof idea of the following result is taken from the proof of Lemma 2.6 of [4].

Proposition 2.12

Let A be a Banach \(\mathfrak {A}\)-bimodule and I be a closed ideal and \(\mathfrak {A}\)-submodule of A, and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\) be such that \(\phi |_I\ne 0.\) If A is module \((\phi ,\varphi )\)-inner amenable, then I is module \((\phi |_I,\varphi )\)-inner amenable.

Proof

Let \(m\in A^{**}\) satisfy \(m(\varphi \circ \phi )=1\), \(m(f.a)=m(a.f)\) and \(m(\alpha .f)=m(f.\alpha )\) for all \(a\in A, f\in A^*\) and \(\alpha \in \mathfrak {A}\). By a similar argument as in the proof of Lemma 3.1 of [9], one can define a bounded linear functional n on \(I^*\) by, \(n(g)=m(f)\) for all \(g\in I^*\), where f is an arbitrary element of \(A^*\) extending g. Now, for every \(g\in I^*, a\in I\) and \(\alpha \in \mathfrak {A}\), we have

$$\begin{aligned} n(g.a)=m(f.a)=m(a.f)=n(a.g), \end{aligned}$$

and

$$\begin{aligned} n(g.\alpha )=m(f.\alpha )=m(\alpha .f)=n(\alpha .g). \end{aligned}$$

In addition, \(n(\varphi \circ \phi |_I)=m (\varphi \circ \phi )=1.\) Therefore, I is module \((\phi |_I,\varphi )\)-inner amenable. \(\square\)

We need to recall the following remark from [4] to give the next result:

Remark 2.13

Let \(\mathfrak {A}\) be a Banach algebra and \(\mathfrak {A}^{\#}\) be the unitization of \(\mathfrak {A}\) which is \(\mathfrak {A}^{\#}=\mathfrak {A}\oplus \mathbb {C}\) is a unital Banach algebra which contains \(\mathfrak {A}\) as a closed ideal. Let A be a Banach \(\mathfrak {A}\)-module. Then, A is a Banach \(\mathfrak {A}^{\#}\)-bimodule with the following module actions:

$$\begin{aligned} (\alpha ,\lambda ).a=\alpha .a+\lambda a,~~~a.(\alpha ,\lambda )=a.\alpha +\lambda a~~~~~(\lambda \in \mathbb {C}, \alpha \in \mathfrak {A}, a\in A). \end{aligned}$$

Let \(A^{\sharp }=(A\oplus \mathfrak {A}^{\#},\cdot )\), where the multiplication \(\cdot\) is defined through

$$\begin{aligned} (a,u)\cdot (b,v)=(ab+a.v+u.b,uv)~~~~~(a,b\in A, u,v\in \mathfrak {A}^{\#}). \end{aligned}$$

Then, with the actions defined by

$$\begin{aligned} u.(a,v)=(u.a,uv),~~~(a,v).u=(a.u,vu)~~~~~(a\in A, u,v\in \mathfrak {A}^{\#}), \end{aligned}$$

\(A^{\sharp }\) is a unital \(\mathfrak {A}^{\#}\)-module Banach algebra with the identity \(e_{A^{\sharp }}=(0,e_{\mathfrak {A}^{\#}})\) , where \(e_{\mathfrak {A}^{\#}}=(0,1).\) Now, suppose that \(\phi \in \Omega _A\) and \(\varphi ^{\#}\) is the extension of \(\varphi\) on \(\mathfrak {A}^{\#}\) defined by \(\varphi ^{\#}(\alpha ,\lambda )=\varphi (\alpha )+\lambda ~(a\in A, \alpha \in \mathfrak {A}, \lambda \in \mathbb {C}\). If \(u=(\alpha ,\lambda )\in \mathfrak {A}^{\#}\), it is easy to see that

$$\begin{aligned} \phi (a.u)=\phi (u.a)=\varphi ^{\#}(u)\phi (a)(a\in A). \end{aligned}$$
(4)

Define \(\phi ^{\sharp }:A^{\sharp }\longrightarrow \mathfrak {A}^{\#}\) by

$$\begin{aligned} \phi ^{\sharp }(a,u)=(\phi (a),\varphi ^{\#}(u))(a\in A, u\in \mathfrak {A}^{\#}). \end{aligned}$$
(5)

Using (4), one can show that \(\phi ^{\sharp }\) is multiplicative and

$$\begin{aligned} \phi ^{\sharp }(u.(a,v))=\phi ^{\sharp }((a,v).u)=\varphi ^{\#}(u)\phi ^{\sharp }(a,v)(a\in A, u,v\in \mathfrak {A}^{\#}). \end{aligned}$$

Therefore, \(\phi ^{\sharp }\) is an extension of \(\phi\), such that \(\phi ^{\sharp }(0,u)=\varphi ^{\#}(u)\) is the extension \(h_0=\tilde{0}\) of the zero function given by (5).

The proof of the following theorem is inspired by the proof of Proposition 2.7 of [4].

Proposition 2.14

Let A be a Banach \(\mathfrak {A}\)-bimodule, and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). Then, A is module \((\phi ,\varphi )\)-inner amenable if and only if \(A^{\sharp }\) is module \((\phi ^{\sharp },\varphi ^{\#})\)-inner amenable.

Proof

Let \(A^{\sharp }\) be module \((\phi ^{\sharp },\varphi ^{\#})\)-inner amenable. Since the image of \(\phi ^{\sharp }|_A\) is included in \(\mathfrak {A}\), by 2.12, we conclude that A is module \((\phi ,\varphi )\)-inner amenable.

Conversely, suppose that A is module \((\phi ,\varphi )\)-inner amenable. Then, there exists a \(m\in A^{**}\), such that \(m(\varphi \circ \phi )=1\), \(m(f.a)=m(a.f)\), and \(m(\alpha .f)=m(f.\alpha )\) for all \(a\in A, f\in A^*\) and \(\alpha \in \mathfrak {A}\). By Remark 2.13, we may identify the dual space \((A^{\sharp })^{*}\) with \(A^*\oplus \mathbb {C}h_0\), where \(h_0|_A=0\) and \(h_0(e_{A^{\sharp }})=1.\) Define \(n\in (A^{\sharp })^{**}\) by \(n(f)=m(f)~(f\in A^*)\) and \(n(h_0)=0\). Since A is an ideal and \(\mathfrak {A}\)-submodule of \(A^{\sharp }\), it follows that \(h_0.a=0\) and \(h_0.\alpha =0\) for all \(a\in A\) and \(\alpha \in \mathfrak {A}\). A simple computation shows that

$$\begin{aligned} n\big ((f+\lambda h_0).(a+\lambda 'e_{A^{\sharp }})\big )=n\big ((a+\lambda 'e_{A^{\sharp }}).(f+\lambda h_0)\big ), \end{aligned}$$

and

$$\begin{aligned} n\big ((f+\lambda h_0).u\big )=n\big (u.(f+\lambda h_0)\big ), \end{aligned}$$

for all \(f\in A^*, a\in A, u\in \mathfrak {A}^{\#}\) and \(\lambda ,\lambda '\in \mathbb {C}\). For \(f\in A^*\), consider the map \(\bar{f}:A^{\sharp }\longrightarrow \mathbb {C}\) defined by \(\bar{f}(a,u)=f(a)+\tilde{\varphi }(u)~(a\in A,u\in \mathfrak {A}).\) Thus, \(\varphi ^{\#}\circ \phi ^{\sharp }=\overline{\varphi \circ \phi }\), and hence

$$\begin{aligned} n(\varphi ^{\#}\circ \phi ^{\sharp })=n(\overline{\varphi \circ \phi })=n(\varphi \circ \phi +\lambda h_0)=m(\varphi \circ \phi )=1. \end{aligned}$$

Therefore, \(A^{\sharp }\) is module \((\phi ^{\sharp },\varphi ^{\#})\)-inner amenable.

Module inner amenability of certain Banach algebras

Let \(A\widehat{\otimes }B\) be the projective tensor product of two Banach algebras A and B. For every \(f\in A^*\) and \(g\in B^*\), let \(f\otimes g\) denote the element of \((A\widehat{\otimes }B)^*\) satisfying, \(f\otimes g(a\otimes b)=f(a)g(b)(a\in A, b\in B)\). In addition, note that \(A\widehat{\otimes }B\) is a Banach \(\mathfrak {A}\widehat{\otimes }\mathfrak {A}\)-bimodule with the following actions:

$$\begin{aligned} (\alpha \otimes \beta ).(a\otimes b)=(\alpha .a)\otimes (\beta .b) (a\in A, b\in B, \alpha ,\beta \in \mathfrak {A}), \end{aligned}$$

and similarly for right action. For \(\varphi _1,\varphi _2\in \Delta (\mathfrak {A}), \psi \in \Omega _A (=\Omega _{A,\varphi _1})\) and \(\phi \in \Omega _A(=\Omega _{A,\varphi _2})\), define \((\phi \otimes \psi ):A\widehat{\otimes }B\rightarrow \mathfrak {A}\widehat{\otimes }\mathfrak {A}\) by \((\phi \otimes \psi )(a\otimes b)=\phi (a)\otimes \psi (b) (a\in A, b\in B)\). Clearly, \(\phi \otimes \psi \in \Omega _{A\widehat{\otimes }B}(=\Omega _{A\widehat{\otimes }B,\varphi _1\otimes \varphi _2})\) and \(\varphi _1\otimes \varphi _2\in \Delta (\mathfrak {A}\widehat{\otimes }\mathfrak {A})\). In addition, if \(\overline{\varphi }\in \Delta (\mathfrak {A}\widehat{\otimes }\mathfrak {A})\), then \(\overline{\varphi }=\varphi _1\otimes \varphi _2\), where \(\varphi _1, \varphi _2\in \Delta (\mathfrak {A})\) (see [4]).

The technique of proof of the following theorem (one side) is similar to that of Theorem 2.8 of [4].

Theorem 3.1

Let A and B be Banach \(\mathfrak {A}\)-bimodules, and let \(\varphi _1, \varphi _2\in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A, \psi \in \Omega _B\). Then \(A\widehat{\otimes }B\) is module \((\phi \otimes \psi , \varphi _1\otimes \varphi _2)\)-inner amenable (as \(\mathfrak {A}\widehat{\otimes }\mathfrak {A}\) module) if and only if A is module \((\phi ,\varphi _1)\)-inner amenable and B is module \((\psi ,\varphi _2)\)-inner amenable.

Proof

Suppose that \(A\widehat{\otimes }B\) is module \((\phi \otimes \psi , \varphi _1\otimes \varphi _2)\)-inner amenable. Then, there exists \(m\in (A\widehat{\otimes }B)^{**}\), such that \(m\big ( (\varphi _1\otimes \varphi _2)\circ (\phi \otimes \psi )\big )=1\) and

$$\begin{aligned} m\big (f\otimes g.(a\otimes b)\big )=m\big ((a\otimes b).f\otimes g\big ), m\big (f\otimes g.(\alpha \otimes \beta )\big )=m\big ((\alpha \otimes \beta ).f\otimes g\big ), \end{aligned}$$

for all \(a\otimes b\in A\widehat{\otimes }B, f\in A^*, g\in B^*\) and \(\alpha \otimes \beta \in \mathfrak {A}\widehat{\otimes }\mathfrak {A}\). Define \(m_A:A^*\rightarrow \mathbb {C}\) by \(m_A(f)=m\big (f\otimes (\varphi _2\circ \psi )\big )(f\in A^*)\). Therefore, \(m_A(\varphi _1\circ \phi )=m\big ((\varphi _1\circ \phi ) \otimes (\varphi _2\circ \psi )\big )=m\big ( (\varphi _1\otimes \varphi _2)\circ (\phi \otimes \psi )\big )=1.\) Choose \(b_0\in A\), such that \(\varphi _2\circ \psi (b_0)=1\). Therefore, for every \(a\in A\) and \(f\in A^*\), we have

$$\begin{aligned} m_A(f.a)&=m\big ((f.a)\otimes (\varphi _2\circ \psi )\big )\\&=m\big ((f.a)\otimes (\varphi _2\circ \psi ).b_0\big )\\&=m\big (f\otimes (\varphi _2\circ \psi ).(a\otimes b_0)\big )\\&=m\big ((a\otimes b_0).f\otimes (\varphi _2\circ \psi )\big )\\&=m\big ((a.f)\otimes b_0.(\varphi _2\circ \psi )\big )\\&=m_A(a.f). \end{aligned}$$

Similarly, for every \(\alpha \in \mathfrak {A}\), if we take \(\beta \in \mathfrak {A}\), such that \(\varphi _2(\beta )=1\), then

$$\begin{aligned} m_A(f.\alpha )&=m\big ((f.\alpha )\otimes (\varphi _2\circ \psi )\big )\\&=m\big ((f.\alpha )\otimes (\varphi _2\circ \psi ).\beta \big )\\&=m\big (f\otimes (\varphi _2\circ \psi ).(\alpha \otimes \beta )\big )\\&=m\big ((\alpha \otimes \beta ).f\otimes (\varphi _2\circ \psi )\big )\\&=m\big ((\alpha .f)\otimes \beta .(\varphi _2\circ \psi )\big )\\&=m\big ((\alpha .f)\otimes (\varphi _2\circ \psi )\big )\\&=m_A(\alpha .f), \end{aligned}$$

for all \(f\in A^*\). Therefore, A is module \((\varphi ,\phi _1)\)-inner amenable. Similarly, one can prove that B is module \((\psi ,\phi _2)\)-inner amenable.

For the converse, let A is module \((\phi ,\varphi _1)\)-inner amenable and B is module \((\psi ,\varphi _2)\)-inner amenable. Then, by Theorem 2.2, there exist bounded nets \((a_i)_i\) in A and \((b_j)_j\) in B with bounds \(M_1\) and \(M_2\), respectively, such that

$$\begin{aligned} \varphi _1\circ \phi (a_i)=1, \Vert aa_i-a_ia\Vert \longrightarrow 0, \Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0(a\in A, \alpha \in \mathfrak {A}), \end{aligned}$$

and

$$\begin{aligned} \varphi _2\circ \psi (b_j)=1, \Vert bb_j-b_jb\Vert \longrightarrow 0, \Vert \alpha .b_j-b_j.\alpha \Vert \longrightarrow 0(b\in B, \alpha \in \mathfrak {A}). \end{aligned}$$

Consider the bounded net \((a_i\otimes b_j)_{(i,j)}\) in \(A\widehat{\otimes }B\). Therefore, \(\big ((\varphi _1\otimes \varphi _2)\circ (\phi \otimes \psi )\big )(a_i\otimes b_j)=\varphi _1\circ \phi (a_i)\varphi _2\circ \psi (b_j)=1\). Let \(\mathfrak {F} =\sum _{l=1}^N\alpha _l\otimes \beta _l\in \mathfrak {A}\widehat{\otimes }\mathfrak {A}\), then

$$\begin{aligned} \Vert \mathfrak {F}.&(a_i\otimes b_j)-(a_i\otimes b_j).\mathfrak {F}\Vert \\&=\Vert \sum _{l=1}^N\big [(\alpha _l.a_i-a_i.\alpha _l)\otimes \beta _l.b_j+a_i.\alpha _l\otimes (\beta _l.b_j-b_j.\beta _l)\big ]\Vert \\&\le \sum _{l=1}^N M_2\Vert \beta _l\Vert \Vert \alpha _l.a_i-a_i.\alpha _l\Vert +\sum _{l=1}^N M_1\Vert \alpha _l\Vert \Vert \beta _l.b_j-b_j.\beta _l\Vert \longrightarrow 0. \end{aligned}$$

Now, let \(\mathfrak {G}\in \mathfrak {A}\widehat{\otimes }\mathfrak {A}\), so there exist sequences \((\alpha _l)_l\subseteq \mathfrak {A}\) and \((\beta _l)_l\subseteq \mathfrak {A}\), such that \(\mathfrak {G}=\sum _{l=1}^{\infty }\alpha _l\otimes \beta _l\) with \(\sum _{l=1}^{\infty }\Vert \alpha _l\Vert \Vert \beta _l\Vert <\infty .\) By the same argument as in the proof of the Theorem 3.1 of [6], for every \(\mathfrak {G}\in \mathfrak {A}\widehat{\otimes }\mathfrak {A}\), one can show that \(\Vert \mathfrak {G}.(a_i\otimes b_j)-(a_i\otimes b_j).\mathfrak {G}\Vert \longrightarrow 0\). Similarly, we may show that \(\Vert G(a_i\otimes b_j)-(a_i\otimes b_j)G\Vert \longrightarrow 0\) for all \(G\in A\widehat{\otimes }B\). Therefore, Proposition 2.2 implies that \(A\widehat{\otimes }B\) is module \((\phi \otimes \psi , \varphi _1\otimes \varphi _2)\)-inner amenable. \(\square\)

Let A and B be Banach algebras, it is well known that \(A\oplus _{\infty }B\) and \(A\oplus _p B\), the \(l^p\)-direct sum of A and B, are Banach algebras with respect to the canonical multiplication defined by

$$\begin{aligned} (a,b)(c,d):= (ac,bd)\quad (a,c\in A, b,d\in B), \end{aligned}$$

and norms \(\Vert (a,b)\Vert =\max \{\Vert a\Vert ,\Vert b\Vert \}\) and \(\Vert (a,b)\Vert =(\Vert a\Vert ^p+\Vert b\Vert ^p)^{\frac{1}{p}}~(a\in A, b\in B).\) Furthermore, if A and B are two Banach \(\mathfrak {A}\)-bimodules, then \(A\oplus _{\infty }B\) and \(A\oplus _p B\) are Banach \(\mathfrak {A}\)-bimodules under the module actions

$$\begin{aligned} \alpha .(a,b)=(\alpha .a,\alpha .b),\quad (a,b).\alpha =(a.\alpha ,b.\alpha )\quad (a\in A, b\in B, \alpha \in \mathfrak {A}). \end{aligned}$$

Before stating the next theorem, we note that if for every \(\varphi \in \Delta (\mathfrak {A}), \phi \in \Omega _A\) and \(\psi \in \Omega _B\), we define \((0,\psi ):A\oplus _pB\rightarrow \mathfrak {A}\) and \((\phi ,0):A\oplus _pB\rightarrow \mathfrak {A}\) by

$$\begin{aligned} (0,\psi )(a,b)=\psi (b), (\phi ,0)(a,b)=\phi (a)(a\in A, b\in B), \end{aligned}$$

where \(1\le p\le \infty ,\) then \((0,\psi )\) and \((\phi ,0)\in \Omega _{A\oplus _pB}(=\Omega _{A\oplus _pB,\varphi })\).

Theorem 3.2

Let A and B be two \(\mathfrak {A}\)-bimodule Banach algebras, \(\varphi \in \Delta ({\mathfrak {A}}), \phi \in \Omega _A, \psi \in \Omega _B\) and \(1\le p\le \infty .\) Then the following statements are valid:

  1. (i)

    \(A\oplus _pB\) is module \(\big ((\phi ,0),\varphi \big )\)-inner amenable if and only if A is module \((\phi ,\varphi )\)-inner amenable.

  2. (ii)

    \(A\oplus _p B\) is module \(\big ((0,\psi ),\varphi \big )\)-inner amenable if and only if B is module \((\psi ,\varphi )\)-inner amenable.

Proof

(i) Assume that \(A\oplus _pB\) is module \(\big ((\phi ,0),\varphi \big )\)-inner amenable. By Theorem 2.2, there exists a net \((a_i,b_i)_i\) in \(A\oplus _pB\), such that \(\varphi \circ (\phi ,0)(a_i,b_i)\longrightarrow 1\) and

$$\begin{aligned} \Vert (a,b).(a_i,b_i)-(a_i,b_i).(a,b)\Vert \longrightarrow 0, \Vert \alpha .(a_i,b_i)-(a_i,b_i).\alpha \Vert \longrightarrow 0, \end{aligned}$$

for all \((a,b)\in A\oplus _pB\) and \(\alpha \in \mathfrak {A}\). Consider the bounded net \((a_i)_i\) in A. One can easily show that \(\Vert aa_i-a_ia\Vert \longrightarrow 0\) and \(\Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0\) for all \(a\in A, \alpha \in \mathfrak {A}\). In addition, it is clear that \(\varphi \circ \phi (a_i)\longrightarrow 1\). Therefore, Theorem 2.2 implies that A is module \((\phi ,\varphi )\)-inner amenable.

Conversely, suppose that A is module \((\phi ,\varphi )\)-inner amenable. Then, there exists a bounded net \((a_i)_i\) in A, such that \(\varphi \circ \phi (a_i)\longrightarrow 1\), \(\Vert aa_i-a_ia\Vert \longrightarrow 0\) and \(\Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0\) for all \(a\in A, \alpha \in \mathfrak {A}\). Clearly, the bounded net \((a_i,0)_i\subset A\oplus _p B\) satisfies in the condition (iii) of Theorem 2.2. Therefore, \(A\oplus _p B\) is module \(\big ((\phi ,0),\varphi \big )\)-inner amenable.

Similarly, we can prove (ii). \(\square\)

Corollary 3.3

Let A and B be two \(\mathfrak {A}\)-bimodule Banach algebras and \(1\le p\le \infty .\) Then \(A\oplus _pB\) is module \(\big ((\phi ,0),\varphi \big )\)-inner amenable and module \(\big ((0,\psi ),\varphi \big )\)-inner amenable for every \(\varphi \in \Delta ({\mathfrak {A}}), \phi \in \Omega _A\) and \(\psi \in \Omega _B\) if and only if both A and B are module character inner amenable.

We note that for two Banach algebras A and B, a direct verification shows that

$$\begin{aligned} \Delta (A\oplus _p B)=(\Delta (A)\times \{0\})\cup (\{0\}\times \Delta (B)), 1\le p\le \infty . \end{aligned}$$

Now, if we take \(\mathfrak {A}=\mathbb {C}\) and \(\varphi\) is the identity map in the above corollary, then we obtain that \(A\oplus _pB\) is character inner amenable if and only if both A and B are character inner amenable. Therefore, the above corollary generalizes Proposition 4.2 of [6].

Let A be a Banach algebra and X be a Banach A-bimodule. The \(l^1\)-direct sum of A and X, denoted by \(A\oplus _1X\), with the product defined by

$$\begin{aligned} (a,x)(a',x')=(aa',a.x'+x.a')(a,a'\in A, x,x'\in X), \end{aligned}$$

is a Banach algebra that is called the module extension Banach algebra of A and X.

If A is \(\mathfrak {A}\)-bimodule and X is a Banach A-\(\mathfrak {A}\)-module, then \(A\oplus _1 X\) is Banach \(\mathfrak {A}\)-bimodules under the module actions:

$$\begin{aligned} \alpha .(a,x)=(\alpha .a,\alpha .x),(a,x).\alpha =(a.\alpha ,x.\alpha )(a\in A, x\in X, \alpha \in \mathfrak {A}). \end{aligned}$$
(6)

Let A and B be Banach algebras and let X be a Banach AB-module; that is, a left A-module and a right B-module satisfying \(\Vert axb\Vert \le \Vert a\Vert \Vert x\Vert \Vert b\Vert ,~(a\in A, b\in B, x\in X).\) The corresponding triangular Banach algebra

$$\begin{aligned} \tau =\Big \{ \left( \begin{array}{ccc} a &{} x \\ 0 &{} b \\ \end{array} \right) :a\in A, x\in X, b\in B \Big \}, \end{aligned}$$

is equipped with the usual \(2\times 2\)-matrix operations and the norm

$$\begin{aligned}\Vert \left( \begin{array}{ccc} a &{} x \\ 0 &{} b \\ \end{array} \right) \Vert =\Vert a\Vert +\Vert x\Vert +\Vert b\Vert . \end{aligned}$$

This Banach algebra were introduced by Forrest and Marcoux in [7]. Note that \(\tau\) can be identified with the module extension \((A\oplus _1B)\oplus _1 X\), in which X is considered as a \(A\oplus _1B\)-module under the operations:

$$\begin{aligned} (a,b).x=ax, ~x.(a,b)=xb (a\in A, b\in B, x\in X). \end{aligned}$$

Furthermore, if A and B are two Banach \(\mathfrak {A}\)-bimodules and X is a Banach \(A\oplus _1B\)-\(\mathfrak {A}\)-module, then \(\tau\) is Banach \(\mathfrak {A}\)-bimodules under the module actions defined as (6).

Let \(\phi \in \Omega _A\) and define \(\tilde{\phi }:A\oplus _1X\longrightarrow \mathfrak {A}\) by \(\tilde{\phi }(a,x)=\phi (a)~(a\in A, x\in X)\). Then, \(\tilde{\phi }\in \Omega _{A\oplus _1X}\).

Using Theorem 2.2, we can routinely prove the following proposition and so we omit its proof.

Proposition 3.4

Let A be \(\mathfrak {A}\)-bimodule and X be a Banach A-\(\mathfrak {A}\)-module and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). Then \(A\oplus _1X\) is module \((\tilde{\phi },\varphi )\)-inner amenable if and only if there exists a bounded net \((a_i,x_i)_i\) in \(A\oplus _1X\) satisfying

  1. (i)

    \(\varphi \circ \phi (a_i)\longrightarrow 1\) and \(\Vert aa_i-a_ia\Vert \longrightarrow 0, \Vert \alpha .a_i-a_i.\alpha \Vert \longrightarrow 0\) and \(\Vert \alpha .x_i-x_i.\alpha \Vert \longrightarrow 0\) for all \(a\in A\) and \(\alpha \in \mathfrak {A},\)

  2. (ii)

    \(\Vert x.a_i-a_i.x\Vert \longrightarrow 0\) for all \(x\in X\), and

  3. (iii)

    \(\Vert a.x_i-x_i.a\Vert \longrightarrow 0\) for all \(a\in A\).

Corollary 3.5

Let A be \(\mathfrak {A}\)-bimodule and X be a Banach A-\(\mathfrak {A}\)-module and let \(\varphi \in \Delta (\mathfrak {A})\) and \(\phi \in \Omega _A\). If \(A\oplus _1X\) is module \((\tilde{\phi },\varphi )\)-inner amenable, then A is module \((\phi ,\varphi )\)-inner amenable.

Corollary 3.6

Let A and B be \(\mathfrak {A}\)-bimodules and X be a Banach \(A\oplus _1B\)-\(\mathfrak {A}\)-module. If \(\tau\) is module character inner amenable, then so are A and B.

Proof

Suppose that \(\tau\) is module character inner amenable. By Corollary 3.5, \(A\oplus _1B\) is module character inner amenable. Therefore, corollary 3.3 implies that A and B are module character inner amenable. \(\square\)