Independent elements, unmixedness theorems, and asymptotic prime divisors (Q762221)

Let R denote a commutative Noetherian ring and \(I\subset R\) an ideal. Let \(x_ 1,...,x_ n\in I\) denote a d-sequence in the sense of \textit{C. Huneke} [Adv. Math. 46, 249-279 (1982; Zbl 0505.13004)] with \((x_ 1,...,x_{n-1})R:x_ n\subseteq I.\) Then \(\{x_ 1,...,x_{n-1},x^ 2_ n\}\) is an I-independent set of elements in R. Here elements \(a_ 1,...,a_ n\in I\) are called I-independent if every form \(F\in R[X_ 1,...,X_ n]\) with \(F(a_ 1,...,a_ n)=0\) has all its coefficients in I. This notion generalizes the notion of analytical independence. Put sup I for the maximal number of I-independent elements and \(\sup^{\infty}I=\min \{\sup I^ n: n\geq 1\}.\) The last notion was introduced by \textit{W. Bruns} [J. Lond. Math. Soc., II. Ser. 22, 57-62 (1980; Zbl 0452.13005)]. For two ideals I, J of R we define \(J:<I>=\cup_{n\geq 1}J:I^ n.\) Then it is studied when the topology defined by \(\{J^ n:<I>\}_{n\geq 1}\) is weaker than the I-adic topology, which is characterized in terms of associated prime divisors of \(R_ P\) for \(P\supset I\). Putting this together it yields some formulas for \(\sup^{\infty}I\) in terms of equivalent ideal topologies, resp. in terms of local invariants of \(R_ P\). As a first application there is a result describing the existence of a prime divisor p of the completion \(\hat R\) of a local ring (R,M) with \(\dim \hat R/p\leq r\) as equivalent to \(M\in Ass R/I\) for all ideals I contained in a sufficiently large power of M with \(\text{grad}e IR_ P\geq r\) for all \(P\neq M\). All of these results can be modified to the integral closures \((I^ n)_ a\), \(n\geq 1\), and the existence of certain minimal associated prime ideals. As a second application it yields a characterization of unmixed resp. quasi- unmixed local rings as those rings for which \(\{I^ n:<M>\}_{n\geq 1}\) resp. \(\{(I^ n)_ a:<M>\}_{n\geq 1}\) is equivalent to the I-adic topology, resp. to the topology defined by \(\{(I^ n)_ a\}_{n\geq 1}\) for every ideal of the principal class \(\dim R-1.\)