On F-pure thresholds (Q1888814)

Positive characteristic methods have played a significant role in commutative algebra in the past 20 years or so. More recently, ring theoretical methods of this nature have led to important applications to birational geometry as well. The authors are among the researchers whose results shaped the current understanding of the relationship between singularities of pairs in algebraic geometry and the theory of tight closure in commutative algebra. The paper under review is another important contribution of the authors to this development. For a pair \((R, a)\), where \(R\) is a ring and \(a\) an ideal of \(R\), the F-pure threshold \(c(a)\) is defined. The concept is based on previous work by \textit{S. Takagi} [Invent. Math. 157 (1), 123-146 (2004; Zbl 1121.13008)] on the F-purity and strong F-regularity of pairs \((R, a)\). The definition works in the equal characteristic case by reduction to positive characteristic. The authors develop the fundamental properties of the F-pure threshold in positive characteristic and give a number of interesting examples. In characteristic zero, the paper examines the relationship between the F-pure threshold and the log canonical threshold. Notably, for a \(\mathbb Q\)-Gorenstein normal ring essentially of finite type over a field of characteristic zero that is of dense F-pure type, we have that the F-pure threshold \(c(a)\) of \(a\) equals its log canonical threshold lc\((a)\), if \(a\) is contained in the radical ideal defining the non-log-terminal locus of Spec\((R)\). The authors apply their theory to the study of three-dimensional \(\mathbb Q\)-Gorenstein rings with terminal singularities and provide a simpler proof of a result by Kakimi. The final section of the paper proves results regarding the F-pure threshold in positive characteristic that parallel characteristic zero results on the log canonical threshold.