Asymptotic growth of saturated powers and epsilon multiplicity (Q2428843)

In the paper under review, the author studies the growth of saturated powers of modules. Let \(\lambda(M)\) denote the length of an \(R\)-module \(M\). The author proves the following result: Theorem. Suppose that \((R,\mathfrak{m}\)) is a noetherian local domain of of dimension \(d\geq 1\) and \(\mathrm{depth}\geq 2\) which is essentially of finite type over a field \(K\) of characteristic zero (or over a perfect field \(K\) such that \(R/\mathfrak{m}\) is algebraic over \(K\)). Suppose that \(E\) is a rank \(e\) submodule of a finitely generated free \(R\)-module \(F\). Then the limit \[ \lim_{k\to\infty} \lambda(E^k:_{F^k}\mathfrak{m}^\infty/E^k)/k^{d+e-1}\in \mathbb R \] exists. When \(E=I\) is a homogeneous ideal and \(R\) is a standard graded normal \(K\)-algebra, this theorem was proved in [\textit{S. D. Cutkosky} et al. Can. J. Math. 57, No. 6, 1178--1192 (2005; Zbl 1095.13015]. When \(R\) is regular, \(E=I\) is an ideal and the singular locus of \(\mathrm{Spec}(R/I)\) is \(\mathfrak{m}\), the theorem was proved in [\textit{S. D. Cutkosky} et al. Math. Proc. Camb. Philos. Soc. 148, No. 1, 55--72 (2010; Zbl 1200.13010)]. In addition, Kleiman (communication to the author) has also proved the above theorem when \(E\) is a direct summand of \(F\) locally at every non-maximal prime of \(R\). The author obtains the following corollary to the above theorem, which shows that the epsilon multiplicity \(\varepsilon(E)\) of a module, defined as a limsup by \textit{B. Ulrich} and \textit{J. Validashti} see [Math. Proc. Camb. Philos. Soc. 151, No. 1, 95--102 (2011; Zbl 1220.13006)], exists as a limit. Corollary. Suppose that \((R,\mathfrak{m}\)) is a noetherian local domain of of dimension \(d\geq 1\) and \(\mathrm{depth}\geq 2\) which is essentially of finite type over a field \(K\) of characteristic zero (or over a perfect field \(K\) such that \(R/\mathfrak{m}\) is algebraic over \(K\)). Suppose that \(E\) is a rank \(e\) submodule of a finitely generated free \(R\)-module \(F\). Then the limit \[ \lim_{k\to\infty}(d+e-1)!/k^{d+e-1}\lambda(H^0_{\mathfrak{m}}(F^k/E^k))\in \mathbb R \] exists. Thus the epsilon multiplicity \(\varepsilon(E)\) of \(E\) exists as a limit.