Integral domains of finite \(t\)-character (Q405916): Difference between revisions

Let \(D\) be an integral domain. Then we say that \(D\) has \textit{finite character} if any nonzero element of \(D\) belongs only to finitely many maximal ideals of \(D\). The finite character of a domain is strictly connected to the multiplicative structure of the ideals of the domain. For example, it is known that, for a Prüfer domain \(D\), the finite character of \(D\) is equivalent to the \textit{LPI} condition on \(D\), that is, every nonzero locally principal ideal of \(D\) is invertible. A generalization of finite character is given in the star-operations setting. Let \(\star\) be a star operation on \(D\). An ideal \(\mathfrak a\) of \(D\) is a \textit{\(\star\)-ideal} if \(\mathfrak a^\star=\mathfrak a\). Let \({\star-\mathrm{Max}}(D)\) denote the set of all the \textit{maximal \(\star\)-ideals} of \(D\), i.e., the maximal elements of the set of all the proper \(\star\)-ideals, ordered by inclusion (it is well known that \({\star-\mathrm{Max}} (D)\subset \text{spec}(D)\)). The domain \(D\) is said to have \textit{the \(\star\)-finite character} if any nonzero element of \(D\) belongs only to finitely many maximal \(\star\)-ideals of \(D\). If \(\star\) is of finite type on \(D\), an ideal \(\mathfrak a\) of \(D\) is called to be \(\star\)-locally principal if \(\mathfrak aD_{\mathfrak m}\) is principal, for any maximal \(\star\)-ideal \(\mathfrak m\) of \(D\). The ideal \(\mathfrak a\) is \(\star\)-invertible if \((\mathfrak a(D:\mathfrak a))^\star=D\). The domain \(D\) is called to be \(\star\)-LPI if if every nonzero \(\star\)-locally principal ideal of \(D\) is \(\star\)-invertible. In the paper under review the authors study the \(\star\)-finite character property when \(\star\) is the so called \(t\)-operation, i.e., the star-operation defined by \[ E^t:=\bigcup\{(D:(D:F)):F\text{ is a finitely generated fractional subideal of }E \}, \] for any fractional ideal \(E\) of \(D\). For example, Noetherian domains and Krull domains have the \(t\)-finite character. The authors prove that if a domain \(D\) has the \(t\)-finite character, then any maximal \(t\)-ideal \(\mathfrak m\) of \(D\) is \textit{potent}, that is, \(\mathfrak m\) contains a nonzero finitely generated ideal that is not contained in any other maximal \(t\)-ideal of \(D\). A consequence of this fact is that any domain \(D\) with the \(t\)-finite character is \textit{conditionally well behaved}, i.e., for any maximal \(t\)-ideal \(\mathfrak m\) of \(D\), the extension \(\mathfrak mD_{\mathfrak m}\) of \(\mathfrak m\) is a \(t\)-ideal of \(D_{\mathfrak m}\). Furthermore, the authors give a precise characterization of when a domain of the type \(D+XD_S[X]\) (where \(D\) is a fixed integral domain, \(S\) is a multiplicative subset of \(D\) and \(X\) is an indeterminate over \(D\)) is a \textit{Krull-type domain}, i.e., a P\(v\)MD with the \(t\)-finite character. Suppose now that \(\star\) is a star operation of finite type on \(D\) such that \({\star-\mathrm{Max}}(D)\) is \textit{independent}, i.e., the intersection of any two distinct ideals in \({\star-\mathrm{Max}}(D)\) contains some nonzero prime ideal. Then, under the previous assumption, the authors show that \(D\) has the \(\star\)-finite character if and only if \(D\) is \(\star\)-LPI.