Analytic Spread of Filtrations and Symbolic Algebras

TL;DR

This paper introduces and investigates the concept of analytic spread for filtrations in local rings, extending classical ideal results to filtrations, especially symbolic and divisorial ones, and examines their properties and non-Noetherian conditions.

Contribution

It defines the analytic spread of filtrations, explores its properties for divisorial and symbolic filtrations, and highlights differences from ideal cases, including non-Noetherian conditions.

Findings

Analytic spread bounds extend to filtrations, with some differences.

For symbolic powers of a height two prime ideal, the spread can be 0, 1, or 2.

The symbolic algebra is non-Noetherian if the spread of symbolic powers is maximal.

Abstract

In this paper we define and explore the analytic spread $\ell(\mathcal I)$ of a filtration in a local ring. We show that, especially for divisorial and symbolic filtrations, some basic properties of the analytic spread of an ideal extend to filtrations, even when the filtration is non Noetherian. We also illustrate some significant differences between the analytic spread of a filtration and the analytic spread of an ideal with examples. In the case of an ideal $I$, we have the classical bounds $\mbox{ht}(I)\le\ell(I)\le \dim R$. The upper bound $\ell(\mathcal I)\le \dim R$ is true for filtrations $\mathcal I$, but the lower bound is not true for all filtrations. We show that for the filtration $\mathcal I$ of symbolic powers of a height two prime ideal $\mathfrak p$ in a regular local ring of dimension three (a space curve singularity), so that $\mbox{ht}(\mathcal I) =2$ and $\dim R=3$, we have that $0\le \ell(\mathcal I)\le 2$ and all values of 0,1 and 2 can occur. In the cases of analytic spread 0 and 1 the symbolic algebra is necessarily non-Noetherian. The symbolic algebra is non-Noetherian if and only if $\ell(\mathfrak p^{(n)})=3$ for all symbolic powers of $\mathfrak p$ and if and only if $\ell(\mathcal I_a)=3$ for all truncations $\mathcal I_a$ of $\mathcal I$.