1 Introduction

Let A be an integral domain with quotient field K. Let \(\mathcal {F}(A)\) be the set of nonzero fractional ideals of D. For an \(I\in \mathcal {F}(A),\) set \(I^{-1}=\{x\in K \; \mid \; xI\subseteq A\}.\) The mapping on \(\mathcal {F}(A)\) defined by \(I\mapsto I_v=(I^{-1})^{-1}\) is called the v-operation on A. A nonzero fractional ideal I is said to be a v-ideal or divisorial if \(I=I_v,\) and I is said to be v-invertible if \((II^{-1})_v=A.\) For properties of the v-operation the reader is referred to [8, Section 34]. However, we will be mostly interested in the t-operation defined on \(\mathcal {F}(A)\) by \(I\mapsto I_t=\cup \{J_v,\; J\) is a nonzero finitely generated fractional subideal of \(I\}.\) (For properties of the t-operation the reader may consult [1]). A fractional ideal I is called a t-ideal if \(I=I_t.\) A t-ideal (respectively, v-ideal) I has t- (respectively, v-) finite type if \(I=J_t\) (respectively, \(I=J_v)\) for some finitely generated fractional ideal J of A. The set of v-ideals may be a proper subset of the set of t-ideals. A fractional ideal I is said to be t-invertible if \((II^{-1})_t=A.\) The set T(A) of t-invertible fractional t-ideals of A is a group under the t-multiplication \(I\star J:=(IJ)_t,\) and the set P(A) of nonzero principal fractional ideals of A is a subgroup of T(A). Following [7], we define the t-class group of A, denoted Cl\(_{t}(A),\) to be the t-group of t-invertible fractional t-ideals of A under t-multiplication modulo its subgroup of principal fractional ideals that is, Cl\(_{t}(A)= T(A)/ P(A).\)The t-class group of an integral domain was studied by many authors ([7,8,9,10,11]).

Let \(A\subseteq B\) be an extension of integral domains. In [2], the authors study when the natural mapping \(\varphi : Cl_{t}(A) \rightarrow Cl_{t}(A+XB[X]);\) \([I] \mapsto [I(A+XB[X])]\) is an isomorphism. They showed that if B is an integrally closed domain and qf\((A)\subseteq B,\) then \(Cl_{t}(A) \cong Cl_{t}(A+XB[X])\) ([2, Theorem 4.7]). Also, the authors study the form of v-invertible (respectively, t-invertible) ideals of the polynomial ring of the form \(A+XB[X].\) Let \(A\subseteq B\) be an extension of integral domains such that B is an integrally closed domain and \(A+XB[X].\) The authors proved that if I is a fractional divisorial v-invertible ideal of R,  then \(I = u(J_1 + XJ_2[X])\) for some \(u\in qf(A+XB[X]),\) \(J_2\) an integral divisorial v-invertible ideal of B and \(J_1\subseteq J_2\) a nonzero ideal of A ([2, Theorem 2.3]). In this paper, we extend these results to the ring of formal power series of the form \(A+XB[\![X]\!].\) In particular, we give a relationship between v-invertible v-ideals of an integral domain and those of its power series ring of the form \(A + XB[\![X]\!].\)

Let \(A\subseteq B\) be an extension of integral domains, \(B[\![X]\!]\) be the power series ring over B,  and \(R=A + XB[\![X]\!].\) In the first part of this paper, we study when the natural mapping

$$\begin{array}{ccccc} \varphi &{} : &{} Cl_{t}(A) &{} \rightarrow &{} Cl_{t}(R) \\ &{} &{} [I] &{} \mapsto &{} [(IR)_t] \\ \end{array}$$

is an injective homomorphism. We show that if B is a flat A-module, then the mapping \(\varphi \) is an injective homomorphism. Also, we prove that the mapping \(\varphi \) is not surjective in general (Remark 2.7). In the second part of this paper, we give a complete description of v-invertible v-ideals (with nonzero trace in A) of \(A+XB[\![X]\!].\) First, we show that if \(A\subseteq B\) is an extension of integral domains such that B is completely integrally closed, then for each divisorial ideal I of \(R = A+XB[\![X]\!]\) such that \(I\cap A\ne (0),\) there exist a divisorial ideal J of B and a nonzero ideal \(H\subseteq J\) of A such that \(I= H+XJ[\![X]\!]\) (Proposition 3.2). Based on the above result, we prove that if I is a fractional divisorial v-invertible ideal of R such that \(I\cap A \ne (0),\) then \(I = u(J_1 + XJ_2[\![X]\!])\) for some \(u\in qf(R),\) \(J_2\) an integral divisorial v-invertible ideal of B and \(J_1\subseteq J_2\) a nonzero ideal of A,  where B satisfies \(\circledast .\)

2 The t-Class Group of A+XB[[X]]

Let A be an integral domain. A fractional ideal I of A is said to be v-invertible (respectively, t-invertible, invertible) if \((II^{-1})_v=A\) (respectively, \((II^{-1})_t=A\), \(II^{-1}=A)\). Following [7], we define the t-class group of A, denoted by \(Cl_t(A),\) to be the group T(A) of t-invertible fractional t-ideals of A under t-multiplication (i.e., \(I\star J:=(IJ)_t\)) modulo its subgroup P(A) of principal fractional ideals, that is, \(Cl_t(A)= T(A)/ P(A).\) When A is a Krull domain, then the t-class group and the divisor class group coincide. We denote by [I] the equivalence class of a t-invertible t-ideal I of A. Let \(A\subseteq B\) be an extension of integral domains and \(R = A+XB[\![X]\!].\) In this section we show that the natural mapping \(\varphi : Cl_{t}(A) \rightarrow Cl_{t}(R);\) \([I] \mapsto [(IR)_t]\) is an injective homomorphism. To prove it, we need the following lemmas.

Lemma 2.1

Let \(A\subseteq B\) be an extension of integral domains and \(R = A+XB[\![X]\!].\) Let \(F_1\) (respectively, \(F_2\)) be a fractional ideal of A (respectively, B) such that \(F_1\subseteq F_2.\) Then \(F_1 + XF_2[\![X]\!]\) is a fractional ideal of R and

$$(F_1 + XF_2[\![X]\!])^{-1} = F_1^{-1}\cap F_2^{-1} +XF_2^{-1}[\![X]\!].$$

Proof

Let \(I = F_1 + XF_2[\![X]\!].\) Since \(F_1 \subseteq I,\) we obtain \(I^{-1} =R:I \subseteq R:F_1,\) where \(R:F_1=\) \( \{g\in qf(R), gF_1 \subseteq R\}.\) This implies that \(I^{-1}\subseteq \) \(K[\![X]\!],\) where \(K=qf(B).\) Indeed, let \(u\in I^{-1}\) and \(\alpha \in F_1\backslash (0).\) Since \(uF_1 \subseteq R,\) \(\displaystyle {u =\frac{\alpha u}{\alpha } \in \frac{1}{\alpha }R}\) \(\subseteq K[\![X]\!].\)

Now we show that \(u \in I^{-1}\) if and only if \(u(0)F_1 \subseteq A\) and \(uF_2[\![X]\!]\subseteq B[\![X]\!].\)

\((\Rightarrow )\) Let \(u \in I^{-1}.\) Since \(uI \subseteq R,\) we get \(uF_1 + uXF_2[\![X]\!]\) \(\subseteq \) \(A + XB[\![X]\!].\) Chose \(X = 0,\) we obtain \(u(0)F_1 \subseteq A.\) Moreover, \(uF_2[\![X]\!]\subseteq B[\![X]\!].\)

\((\Leftarrow )\) Assume that \(u(0)F_1 \subseteq A\) and \(uF_2[\![X]\!]\subseteq B[\![X]\!].\) We prove that \(u\in I^{-1}.\)

As \(u\in K[\![X]\!],\) we can write \(u =\displaystyle \sum _{i=0}^\infty a_iX^i,\) where \(a_i \in K.\) It is clearly that

$$uI = uF_1 + XuF_2[\![X]\!]\subseteq u(0)F_1 + \displaystyle (\sum _{i=1}^\infty a_iX^i)F_1 + XuF_2[\![X]\!].$$

Moreover, \(\displaystyle (\sum _{i=1}^\infty a_iX^i)F_1\) \(=\) \((u - u(0))F_1\) \(\subseteq \) \(uF_1 + u(0)F_1.\) Then \(uF_1 + u(0)F_1\) \(\subseteq B[\![X]\!],\) because \(uF_1\) \(\subseteq \) \(uF_2 \subseteq B[\![X]\!]\) and \(u(0)F_1\) \(\subseteq \) \(A \subseteq B[\![X]\!].\) This implies that \((\displaystyle \sum _{i=1}^\infty a_iX^i)F_1\) \(\subseteq \) \(B[\![X]\!].\) Now let P be an element of \((\displaystyle \sum _{i=1}^\infty a_iX^i)F_1.\) Then there exists an element \(\alpha \) of \(F_1\) such that \(P = X(\displaystyle \sum _{i=1}^\infty a_iX^{i-1})\alpha .\) Since \((\displaystyle \sum _{i=1}^\infty a_iX^i)F_1\) \(\subseteq \) \(B[\![X]\!],\) \((\displaystyle \sum _{i=1}^\infty a_iX^{i-1})\alpha \) \(\in \) \(B[\![X]\!].\) Thus \(P\in XB[\![X]\!],\) and so \((\displaystyle \sum _{i=1}^\infty a_iX^i)F_1\) \(\subseteq \) \(XB[\![X]\!].\) This shows that

$$\begin{array}{ccl} uI &{}\subseteq &{} u(0)F_1 + XuF_2[\![X]\!]+ (\displaystyle \sum _{i=1}^\infty a_iX^i)F_1 \\ &{} \subseteq &{} A + XB[\![X]\!]\\ &{} = &{} R. \\ \end{array}$$

Hence \(u\in I^{-1}.\)

Now \(u \in I^{-1}\) if and only if \(u(0)F_1 \subseteq A\) and \(uF_2[\![X]\!]\) \(\subseteq B[\![X]\!]\) which equivalent to \(u(0)\in F_1^{-1}\) and \(u \in (F_2[\![X]\!])^{-1}.\) But \((F_2[\![X]\!])^{-1}\) \(=\) \(F_2^{-1}[\![X]\!].\) Hence \(u\in I^{-1}\) if and only if \(u \in F_1^{-1}\cap F_2^{-1} +XF_2^{-1}[\![X]\!].\) \(\square \)

Example 2.2

Let A \(= \mathbb {Z},\) B \(= \mathbb {Z}[i]\) and \(R=\mathbb {Z}+X\mathbb {Z}[i][\![X]\!].\) Let \(I=2\mathbb {Z}+(1+i)X\mathbb {Z}[i][\![X]\!].\) We show that I is a divisorial ideal of R,  i.e., \(I_{\upsilon }=I.\)

It is clear that I is an ideal of R. Now by Lemma 2.1,

$$\begin{array}{ccl} I^{-1} &{}=&{} \frac{1}{2}\mathbb {Z} \bigcap ((1+i)\mathbb {Z}[i])^{-1} + X((1+i)\mathbb {Z}[i])^{-1}[\![X]\!]\\ {} &{}=&{} \frac{1}{2} \mathbb {Z} \bigcap (1+i)^{-1}\mathbb {Z}[i] + (1+i)^{-1}X\mathbb {Z}[i][\![X]\!].\end{array}.$$

But if \(x\in \) \(\frac{1}{2}\mathbb {Z}\bigcap \frac{1}{1+i}\mathbb {Z}[i],\) then x \(= \frac{1}{2}r\) \(= \frac{1}{1+i}u,\) with \(r\in \) \(\mathbb {Z}\) and \(u\in \) \(\mathbb {Z}[i].\) This implies that \((1+i)r\) \(= 2u.\) Write u \(= \alpha + i\beta .\) Then \(2\alpha \) \(= r\) and \(2\beta \) \(=r\) thus 2 divided r,  and so x \(= \alpha \) \(\in \) \(\mathbb {Z}.\) Hence \(I^{-1}\) \(=\mathbb {Z}\) \(+ X\frac{1-i}{2}\mathbb {Z}[i][[\![X]\!].\) Again by Lemma 2.1,

$$\begin{array}{ccl} I_{\upsilon } &{}=&{} (I^{-1})^{-1} \\ &{}=&{} \mathbb {Z}\bigcap ((1+i)^{-1}\mathbb {Z}[i])^{-1} + X((1+i)^{-1}\mathbb {Z}[i])^{-1}[\![X]\!]\\ &{}=&{} \mathbb {Z}\bigcap (1+i)\mathbb {Z}[i]+ (1+i)X\mathbb {Z}[i][\![X]\!]\\ &{}=&{} 2\mathbb {Z} + (1+i)X\mathbb {Z}[i][\![X]\!]\\ &{}=&{} I.\end{array}.$$

This shows that I is a divisorial ideal of R.

Let \(A\subseteq B\) be an extension of integral domains. Following [3], we say that B is t-linked over A,  if for each finitely generated fractional ideal I of A with \(I^{-1}=A,\) we have \((IB)^{-1}=B.\)

Lemma 2.3

Let \(A\subseteq B\) be an extension of integral domains and \(R = A+XB[\![X]\!].\) If B is t-linked over A,  then the extension \(A\subseteq R\) is t-linked.

Proof

Let I be a finitely generated fractional ideal of A such that \(I^{-1}=A.\) Since \(IR\subseteq \) \(I+(IB)[\![X]\!],\) then by Lemma 2.1,

$$I^{-1}\cap (IB)^{-1}+X(IB)^{-1}[\![X]\!]=(I+(IB)[\![X]\!])^{-1}\subseteq (IR)^{-1}.$$

But B is t-linked over A,  then \(R=A+XB[\![X]\!]\) \(=I^{-1}\cap (IB)^{-1}+X(IB)^{-1}[\![X]\!]\) \(\subseteq \) \((IR)^{-1},\) and hence R \(\subseteq \) \((IR)^{-1}.\)

Now we will show that \((IR)^{-1}\subseteq R.\) Let u be an element of \((IR)^{-1}.\) It is easy to prove that \(u\in \) \(L+XK[\![X]\!],\) where \(L=qf(A)\) and \(K=qf(B).\) Put \(u=\displaystyle \sum _{i=0}^\infty a_iX^i\in L+XK[\![X]\!],\) and let \(\alpha \in I.\) Since \(\alpha u= \sum _{i=0}^\infty (\alpha a_i)X^i\in R,\) \(\alpha a_0\in A,\) and hence \(a_0\in I^{-1}.\) Moreover, if \(r\in IB,\) then \(urX\in \) \(u(IR)\subseteq R.\) This implies that for each \(i\ge 1,\) \(ra_i\in B.\) Therefore for each \(i\ge 1,\) \(a_i\in (IB)^{-1}.\) Hence \(u\in I^{-1}+ X(IB)^{-1}[\![X]\!]=A+ XB[\![X]\!]=R\) since B is t-linked over A. Hence \((IR)^{-1}= R.\) \(\square \)

Proposition 2.4

Let \(A\subseteq B\) be an extension of integral domains such that B is t-linked over A. Then the mapping

$$\begin{array}{ccccc} \varphi &{} : &{} Cl_{t}(A) &{} \rightarrow &{} Cl_{t}(R) \\ &{} &{} [I] &{} \mapsto &{} [(IR)_t] \\ \end{array}$$

is an homomorphism.

Proof

Follows from Lemma 2.3 and [3, Theorem 2.2]. \(\square \)

Let \(A\subseteq B\) be an extension of integral domains and I a finitely generated ideal of A. It well known that \(I.A[\![X]\!]=(IA)[\![X]\!]=I[\![X]\!].\) Using the same proof we can prove that \(I.B[\![X]\!]=(IB)[\![X]\!].\)

Lemma 2.5

Let \(A\subseteq B\) be an extension of integral domains such that B is a flat A-module, I an ideal of A and \(R=A+XB[\![X]\!].\) We assume that I and \(I^{-1}\) are v-ideals of finite type. Then \((IR)_v=I+X(IB)[\![X]\!].\)

Proof

Since I and \(I^{-1}\) are v-ideals of finite type, \(I=J_v\) and \(I^{-1}=L_v\) for some finitely generated ideals J and L of A. Since \(JR=J+X(JB)[\![X]\!],\) by Lemma 2.1,

$$\begin{array}{ccl} (JR)^{-1} &{}=&{} (J+X(JB)[\![X]\!])^{-1} \\ &{}=&{} J^{-1}\cap (JB)^{-1}+X(JB)^{-1}[\![X]\!]\\ &{}=&{} J^{-1}\cap J^{-1}B +X(J^{-1}B)[\![X]\!]\\ &{}=&{} J^{-1},\end{array}$$

where the third equality follows from the fact that B is a flat A-module.

Again apply Lemma 2.1, \((JR)_v=\) \(J_v\cap (J^{-1}B)^{-1}\) \(+X(J^{-1}B)^{-1}[\![X]\!].\)

Since \(L_v=I^{-1}=J^{-1},\)

$$(J^{-1}B)^{-1}=(L_vB)^{-1}=(LB)^{-1}=L^{-1}B=J_vB,$$

where the second equality follow from the proof of [5, Proposition 2.2]. So

$$\begin{array}{ccl} (JR)_v &{}=&{} J_v\cap (J_vB)+X(J_vB)[\![X]\!]\\ &{}=&{} J_v+X(J_vB)[\![X]\!]\\ &{}=&{}I+X(IB)[\![X]\!].\end{array}$$

This implies that \(I+X(IB)[\![X]\!]\subseteq (IR)_v.\) Now, using Lemma 2.1, we can prove that

$$(I+X(IB)[\![X]\!])_v=I+X(IB)[\![X]\!].$$

This shows that \((IR)_v\subseteq \) \(I+X(IB)[\![X]\!],\) and hence \((IR)_v=\) \(I+X(IB)[\![X]\!].\) \(\square \)

We are now ready to prove the main result of this section.

Theorem 2.6

Let \(A\subseteq B\) be an extension of integral domains such that B is a flat A-module. Then the mapping

$$\begin{array}{ccccc} \varphi &{} : &{} Cl_{t}(A) &{} \rightarrow &{} Cl_{t}(R) \\ &{} &{} [I] &{} \mapsto &{} [(IR)_t] \\ \end{array}$$

is an injective homomorphism.

Proof

Since B is a flat A-module, B is t-linked over A. So by Proposition 2.4, the mapping \(\varphi \) is an homomorphism. We show that \(\varphi \) is injective. Let I be a t-invertible t-ideal of A such that \((IR)_t\) is a principal ideal of R. We will prove that I is principal. Since \((IR)_t\) is principal, \((IR)_t=fR\) for some \(f\in (IR)_t.\)

Case 1: I is an integral ideal of A.

As \((IR)_t=fR,\) then \((IR)_v=fR.\) By Lemma 2.5, \((IR)_v=\) \(I+X(IB)[\![X]\!];\) so \(I=f(0)A\) is a principal ideal of A.

Case 2: I is a fractional ideal of A.

Let \(d\in A\backslash (0)\) such that \(dI\subseteq A.\) Put \(I^\prime =dI.\) Then \(I^\prime \) is an integral t-invertible t-ideal of A. Moreover, \((I^\prime R)_t=dfR\) is a principal ideal of R. By case 1, \(I^\prime \) is a principal ideal of A. So I is a principal ideal of A,  and hence \(\varphi \) is injective. \(\square \)

Remark 2.7

Let \(A\subseteq B\) be an extension of integral domains and let \(\varphi : Cl_{t}(A) \rightarrow Cl_{t}(R)\) be the natural mapping. Note that \(\varphi \) is not surjective in general. Indeed, let A \(= \mathbb {Z},\) B \(= \mathbb {Z}[i]\) and \(R=\mathbb {Z}+X\mathbb {Z}[i][\![X]\!].\) Assume that \(\varphi \) is surjective.

By [6, Chapter 1, Proposition 2], \(\mathbb {Z}[i]=\mathbb {Z}\oplus i\mathbb {Z}\) is a flat \(\mathbb {Z}\)-module; so by Theorem 2.6, \(\varphi \) is an injective homomorphism, and hence \(\varphi \) is an isomorphism. This implies that

$$Cl_{t}(\mathbb {Z}) \cong Cl_{t}(\mathbb {Z}+X\mathbb {Z}[i][\![X]\!]).$$

Since \(\mathbb {Z}\) is a PID (principal ideal domain), \(Cl_{t}(\mathbb {Z})=0\) which implies that \(Cl_{t}(\mathbb {Z}+X\mathbb {Z}[i][\![X]\!])=0.\) Now we prove that \(Cl_{t}(\mathbb {Z}+X\mathbb {Z}[i][\![X]\!])\ne 0,\) and hence we obtain a contradiction. Let \(I=2\mathbb {Z}+(1+i)X\mathbb {Z}[i][\![X]\!].\)

Claim 1: I and \(I^{-1}\) are ideals of R of v-finite type.

It is clear that \((2,(1+i)X)\) \(\subseteq \) I. Conversely, let f \(\in \) I. Then f \(= 2r + X(1+i)Q,\) for some \(r\in \mathbb {Z}\) and \(Q\in \mathbb {Z}[i][\![X]\!]\) \(= \mathbb {Z} + i\mathbb {Z} +X\mathbb {Z}[i][\![X]\!].\) So there exist \(s,t\in \mathbb {Z}\) and \(h\in \mathbb {Z}[i][\![X]\!]\) such that f \(= 2r + X(1+i)(s + it + Xh)\) \(= 2(r - tX) + (1+i)X(s + t + Xh)\) \(\in \) \((2,(1+i)X).\) Hence I \(= (2,(1+i)X).\) Now, by Example 2.2, \(I^{-1}\) \(=\mathbb {Z}\) \(+ X\frac{1-i}{2}\mathbb {Z}[i][\![X]\!].\) In the same way, we can show that \(I^{-1}\) \(= (1, \frac{1-i}{2}X).\)

Claim 2: I is a v-invertible ideal of R.

Note that

$$II^{-1}=(1, \frac{1-i}{2}X)(2,(1+i)X)=(2, (1+i)X, (1-i)X, X^2).$$

Let \(u\in \) qf(R) such that \((2, (1+i)X,\) \((1-i)X, X^2)\) \(\subseteq \) uR. Since \(2\in \) \((2, (1+i)X,\) \((1-i)X, X^2)\) \(\subseteq \) uR,  then u \(=\frac{2}{f},\) with \(f\in R\) and \(X^2\in \) \((2, (1+i)X,\) \((1-i)X, X^2)\) \(\subseteq \) uR \(= \frac{2}{f}R.\) Thus \(X^2f= 2\,g,\) for some \(g = a_0 + a_1X + \cdots +a_nX^n\) \(\in R.\) This implies that \( a_0 = a_1 = 0,\) and so g \(= X^2\,h,\) where h \(= (a_2 + \cdots +a_nX^{n-2})\) \(\in \mathbb {Z}[i][\![X]\!].\) Then f(0) \(= 2\,h(0)\in \) \(\mathbb {Z}.\) But \(\mathbb {Z}[i]\) \(= \mathbb {Z} + i\mathbb {Z},\) then h(0) \(= s + it\) \(\in \) \(\mathbb {Z} + i\mathbb {Z}.\) Since \(2\,h(0)\in \) \(\mathbb {Z},\) then h(0) \(\in \mathbb {Z},\) and so 1 \(= uh\) \(\in \) uR. Thus

$$(II^{-1})_v= (2, (1+i)X, (1-i)X, X^2)_{v}= R.$$

Using claim 1 and 2, it is easy to prove that I is a t-invertible t-ideal of R. This implies that \([I]\in \) \(Cl_{t}(R).\) Now we show that \([I]\ne 0\) which equivalent to I is not a principal ideal of R. Assume the contrary that I is principal. Then \(I=PR\) for some \(P\in R.\) Since \(2\in I,\) \(P(0)\ne 0.\) In fact \(P(0) \in \{\pm 1, \pm 2\}.\) Moreover, as \((1+i)X \in I,\) we obtain \(P(0)\in \{\pm 1\}\) which implies that P(0) is a unit in \(\mathbb {Z}.\) A routine calculation (by induction) shows that P is a unit in R. This implies that \(I=PR=R,\) a contradiction. Then \([I]\ne 0,\) and hence \(Cl_{t}(\mathbb {Z}+X\mathbb {Z}[i][\![X]\!])\ne 0.\)

3 v-Invertible v-Ideals of A+XB[[X]]

In this section, we investigate a relationship between v-invertible v-ideals of an integral domain and those of its power series ring of the form \(A + XB[\![X]\!],\) where \(A\subseteq B\) is an extension of integral domain. We begin this section by the following proposition.

Proposition 3.1

Let \(A\subseteq B\) be an extension of integral domain, J an ideal of A and \(R= A + XB[\![X]\!]\).

  1. (1)

    If \((JR)_{v}\) \(= R,\) then \(J_{v}\) \(= A.\)

  2. (2)

    If \((JR)_t\) \(= R,\) then \(J_t\) \(= A.\)

Proof

  1. (1).

    Assume that \((JR)_{v}\) \(= R\) and let u \(\in \) qf(A) such that J \(\subseteq \) uA. Then JR \(\subseteq \) uAR \(\subseteq \) uR which implies that R \(=(JR)_{v}\) \(\subseteq \) \((uR)_{v}\) \(= uR.\) Thus

    $$A \subseteq \bigcap _{u\in \hbox {qf}(A), J \subseteq uA}Au=J_v.$$

    This shows that \(A\subseteq J_{v}\subseteq A,\) and hence \(J_{v}\) \(= A.\)

  2. (2).

    Suppose that \((JR)_t\) \(= R.\) Then

    $$\begin{aligned} R =\bigcup \{(FR)_{v}, F\subseteq J \text { of } \text { finite type of } A\}. \end{aligned}$$

    Thus there exists a finitely generated ideal \(F_0\) of A such that \(F_0\subseteq \) J and 1 \(\in \) \((F_0R)_{v}.\) This implies that R \(= (F_0R)_{v}.\) Now, by (1), \((F_0)_{v}\) \(= A;\) so

    $$\begin{aligned} A \subseteq \bigcup \{F_{v}, F\subseteq J \text { of } \text { finite type of } A\} =J_t \subseteq A. \end{aligned}$$

    Hence A \(= J_t.\)

\(\square \)

Let A be an integral domain. According to [12, Theorem 2.11], A is completely integrally closed if and only if for each \(f,g\in A[\![X]\!],\) \((A_fA_g)_{v}=(A_{fg})_{v}.\) Using this result we prove a complete description of v-invertible v-ideals (with nonzero trace in A) of R. First we need to prove the following proposition.

Proposition 3.2

Let \(A\subseteq B\) be an extension of integral domains such that B completely integrally closed and \(R = A+XB[\![X]\!].\) Then for each divisorial ideal I of R such that \(I\cap A\ne (0),\) there exist a divisorial ideal J of B and a nonzero ideal \(H\subseteq J\) of A such that \(I= H+XJ[\![X]\!].\)

Proof

Let \(H=I\bigcap A\) and J the ideal of B generated the coefficients of all elements of I.

It is clear that \(H\subseteq J\) and \(H\subseteq I.\) We show that \(XJ_v[\![X]\!]\subseteq I.\) Let \(f,g\in R,\;g\ne 0\) such that \(I\subseteq \frac{f}{g}R.\) Let \(0 \ne a\in H.\) Since \(a\in H\subseteq \) \(I\subseteq \) \(\frac{f}{g}R,\) then there exists an \(r\in R{\setminus }(0)\) such that \(\frac{a}{r}=\frac{f}{g}.\) Let \(0 \ne h\in I\) \(\subseteq \frac{f}{g}R\) \(=\frac{a}{r}R.\) Then \(rh\in aR\) which implies that \(rh\in aB[\![X]\!].\) So \((A_{rh})_v\subseteq aB.\) By hypothesis B is a completely integrally closed domain, then \(A_{r}A_{h}\subseteq \) \((A_{r}A_{h})_v\) \(\subseteq \) aB. This implies that \(rA_h[\![X]\!]\subseteq aB[\![X]\!].\) Now we show that \(rJ[\![X]\!]\) \(\subseteq \) \(aB[\![X]\!].\) Indeed, if \(f\in rJ[\![X]\!],\) then \(f=rf_1\) for some \(f_1=\displaystyle \sum _{i=0}^{\infty }a_iX^i\in J[\![X]\!].\) Put \(r=\displaystyle \sum _{i=0}^{\infty }\beta _iX^i.\) Then \(f=\displaystyle \sum _{n=0}^{\infty }(\displaystyle \sum _{i=0}^na_i\beta _{n-i})X^n.\) But \(a_i=\displaystyle \sum _{k=0}^{m_i}\alpha _{i,k}t_{i,k}\) with \(t_{i,k}\) \(\in B,\) \(\alpha _{i,k}\) \(\in A_{f_{i,k}},\) then

$$a_i\beta _{n-i}= \displaystyle \sum _{k=0}^{n_i}\alpha _{i,k}t_{i,k}\beta _{n-i} \in A_{r}A_{f_{i,k}} \subseteq aB.$$

Which implies that \(rJ[\![X]\!]\subseteq aB[\![X]\!].\) So

$$r(J[\![X]\!])_v=(rJ[\![X]\!])_v\subseteq (aB[\![X]\!])_v = a(B[\![X]\!])_v = aB[\![X]\!].$$

Since \((J[\![X]\!])_v\) = \(J_v[\![X]\!],\) \(rJ_v[\![X]\!]\) \(\subseteq \) \(aB[\![X]\!].\) This implies that \(\frac{aX}{r}B[\![X]\!]\) \(\subseteq \) \(\frac{a}{r}R;\) so \(\frac{rXJ_v[\![X]\!]}{r}\) \(\subseteq \) \(\frac{aX}{r}B[\![X]\!]\) \(\subseteq \) \(\frac{a}{r}R\) =\(\frac{f}{g}R\) which implies that \(XJ_v[\![X]\!]\) \(\subseteq \) \(\frac{f}{g}R.\) Thus

$$XJ_v[\![X]\!]\subseteq \displaystyle \cap _{f,g\in R, I\subseteq \frac{f}{g}}\frac{f}{g}R =I_v=I,$$

and hence \(H+XJ[\![X]\!]\) \(\subseteq \) \(H+XJ_v[\![X]\!]\) \(\subseteq I.\) Now we will show that \(I\subseteq H+XJ[\![X]\!].\) Let \(f\in I.\) Then \(f=a_0+\displaystyle \sum _{i=1}^\infty a_iX^i,\) where \(a_0\in A\) and \(\;a_i\in B\) for each \(i\ge 1.\)

As \(J=<A_f,\;f\in I>,\) then for each \(i\ge 1,\) \(a_i\in J;\) so

$$\displaystyle \sum _{i=1}^\infty a_iX^i =X\displaystyle \sum _{i=1}^\infty a_iX^{i-1}\in XJ[\![X]\!]\subseteq XJ_v[\![X]\!].$$

Since \(XJ_v[\![X]\!]\) \(\subseteq I,\) \(\displaystyle \sum _{i=1}^\infty a_iX^i\) \(\in I.\) This implies that \(a_0=f-\displaystyle \sum _{i=1}^\infty a_iX^i\in I.\) Thus \(a_0\in A\cap I=H,\) and hence \(f\in H+XJ[\![X]\!].\) Now we have

$$H+XJ[\![X]\!]\subseteq H+XJ_v[\![X]\!]\subseteq I \subseteq H+XJ[\![X]\!].$$

Hence \(I=H+XJ[\![X]\!]\) and \(J_v=J.\) \(\square \)

Our next result give a complete description of v-invertible v-ideals of \(A+XB[\![X]\!]\) with nonzero trace in A.

Theorem 3.3

Let \(A\subseteq B\) be an extension of integral domains such that B is completely integrally closed and \(R = A+XB[\![X]\!].\) Let I be a fractional divisorial v-invertible ideal of R such that \(I\cap A \ne (0).\) Then \(I = u(J_1 + XJ_2[\![X]\!])\) for some \(u\in qf(R),\) \(J_2\) an integral divisorial v-invertible ideal of B and \(J_1\subseteq J_2\) a nonzero ideal of A.

Proof

Since I is a divisorial ideal of R and B is completely integrally closed, by Proposition 3.2, \(I=H+XJ[\![X]\!]\) for some divisorial ideal J of B and a nonzero ideal \(H\subseteq J\) of A. We show that there exists nonzero \(c\in K\) such that \(cH\subseteq A\) and \(cJ\subseteq B.\)

Let \(a\in H\) be a nonzero element. We have \(aI^{-1}\) is a divisorial ideal of R. Using Lemma 2.1, it is easy to prove that \(aI^{-1}\cap A\ne (0).\) Then by Proposition 3.2, \(aI^{-1}=H^\prime +XJ^\prime [\![X]\!]\) for some divisorial ideal \(J^\prime \) of B and a nonzero ideal \(H^\prime \subseteq J^\prime \) of A.

$$\begin{array}{ccl} aR &{}=&{} a(II^{-1})_v\\ {} &{}=&{} (a(II^{-1}))_v\\ &{}=&{} (I(aI^{-1}))_v\\ &{}=&{} ((H + XJ[\![X]\!])(H^{\prime } + XJ^{\prime }[\![X]\!]))_v. \end{array}.$$

So \((H + XJ[\![X]\!])(H^{\prime } + XJ^{\prime }[\![X]\!])\) \(\subseteq \) \(aR=aA+aXB[\![X]\!].\) Then \(HH^{\prime }\) \(\subseteq \) aA and \(JJ^{\prime }\) \(\subseteq \) aB. This implies that \(\frac{1}{a}HH^{\prime }\) \(\subseteq \) A and \(\frac{1}{a}JJ^{\prime }\) \(\subseteq \) B.

Let \(c\in \) \(\frac{1}{a}H^{\prime }\) be a nonzero element. Then \(J_1\) \(= cH\) \(\subseteq \) \(\frac{1}{a}HH^{\prime }\) \(\subseteq A\) and \(J_2\) \(= cJ\subseteq B.\) We have \(J_1 \ne (0)\) and \(J_2\) is a divisorial ideal of B.

Since \(I=H + XJ[\![X]\!],\) then

$$I= \frac{1}{c}(cH + XcJ[\![X]\!])= \frac{1}{c}(J_1 +XJ_2[\![X]\!])= u(J_1 + XJ_2[\![X]\!]),$$

where \(u = \frac{1}{c}\) \(\in qf(R).\) Now we will show that \(J_2\) is v-invertible. By Lemma 2.1, we have

$$\begin{aligned} I^{-1} = \frac{1}{u}(J_1^{-1} \cap J_2^{-1} + XJ_2^{-1}[\![X]\!]). \end{aligned}$$

Thus

$$\begin{array}{ccl} II^{-1} &{}\subseteq &{} J_1(J_1^{-1} \cap J_2^{-1})+XJ_2(J_2^{-1}[\![X]\!])\\ &{}\subseteq &{} J_1J_1^{-1} +X(J_2J_2^{-1})[\![X]\!]\\ &{}\subseteq &{} A + XB[\![X]\!]\\ &{}=&{} R.\end{array}$$

Since I is v-invertible, we get

$$\begin{aligned} R =(J_1(J_1^{-1}\cap J_2^{-1}) + X(J_2J_2^{-1})[\![X]\!])^{-1}.\end{aligned}$$

Again by Lemma 2.1, R \(= (J_1(J_1^{-1}\cap J_2^{-1}))^{-1} \cap (J_2J_2^{-1})^{-1}\) \(+\) \(X(J_2J_2^{-1})^{-1}[\![X]\!].\) Then \(B[\![X]\!]\) \(= (J_2J_2^{-1})^{-1}[\![X]\!],\) and this implies that B \(= (J_2J_2^{-1})^{-1}.\) Hence \(J_2\) is v-invertible. \(\square \)

Clearly that every Krull domain is completely integrally closed. Using Theorem 3.3, we obtain a new characterization of divisorial v-invertible ideals of the power series ring of the form \(A+XB[\![X]\!].\)

Corollary 3.4

Let I be a fractional divisorial v-invertible ideal of \(R=A+XB[\![X]\!]\) such that \(I\cap A \ne (0).\) Assume that B is a Krull domain. Then \(I = u(J_1 + XJ_2[\![X]\!])\) for some \(u\in qf(R),\) \(J_2\) an integral divisorial v-invertible ideal of B and \(J_1\subseteq J_2\) a nonzero ideal of A.

Recall from [4] that an integral domain A is called formally integrally closed if \((A_{fg})_{t}=(A_fA_g)_{t}\) for all \(f,g\in A[\![X]\!]\backslash (0).\) It was shown in [4] that if A is formally integrally closed, then A is completely integrally closed, but the converse is false in general ([4, Example 3.2]).

Proposition 3.5

[4, Proposition 3.6] Let A be a formally integrally closed domain. If I is a finite type v-ideal of \(A[\![X]\!]\) with \(J\cap A\ne 0,\) then \(I=J[\![X]\!]\) for some v-ideal J of A.

Note that in [4] Anderson and Kang characterized the v-ideals of finite type of the power series ring \(A[\![X]\!]\) with nonzero trace in A in the case when A is a formally integrally closed domain. Now, using Proposition 3.2, in the particular case when \(A=B,\) we obtain a new approach to characterize the divisorial ideals of the ring \(A[\![X]\!]\) with nonzero trace in A.

Proposition 3.6

Let A be a completely integrally closed domain and I a fractional divisorial ideal of \(A[\![X]\!]\) such that \(I\cap A \ne (0).\) Then \(I = J_1 + XJ_2[\![X]\!]\) for some nonzero ideal \(J_1\) of A and some divisorial ideal \(J_2\) of A such that \(J_1\subseteq J_2.\)