Absolute integral closures are big Cohen-Macaulay algebras in characteristic $P$
Absolute integral closures are big Cohen-Macaulay algebras in characteristic $P$
复制标题
绝对积分闭包是具有 $P$ 特征的大 Cohen-Macaulay 代数
DOI:
10.1090/s0273-0979-1991-15970-7
复制
发表时间:
1991
影响因子:
1.3
通讯作者:
C. Huneke
中科院分区:
文献类型:
--
作者:
M. Hochster;C. Huneke
Throughout this paper "ring" means commutative ring with identity and modules are unital. Our main interest is in local rings, i.e., Noetherian rings (R, m) with a unique maximal ideal m . In such a ring, xx, . . . , xn e m is a system of parameters if m has a power in the ideal (x{, . . . , xn)R and n is the Krull dimension of R. When R is complete and contains a field, this means that R is module-finite over a formal power series subring K[[xx, . . . , xn]] = A . i? is called Cohen-Macaulay if some (equivalently, every) system of parameters is a regular sequence in i?, which means that every xi+x is a nonzerodivisor on R/{xx, . . . , xt)R, for 0 < i < n 1. In the case where R is module-finite over the formal power series subring A , this means that R is a free ^4-module. For many theorems of commutative algebra and of algebraic geometry, the Cohen-Macaulay condition (possibly on the local rings of a variety) is just what is needed to make the theory work. Our main result, which is given below in an algebraic form in Theorem 1 and in a geometric form in Theorem 2, asserts that, under mild conditions on a local ring R of positive prime characteristic p, one can "correct" the failure of the Cohen-Macaulay condition in R itself by passing to a very large integral extension of R. It is worth emphasizing that this is quite false in characteristic 0, and comes as a surprise, we believe, in characteristic p. One can get an idea of how far these theorems are from the truth over a field of characteristic 0 from the following observation: In the situation of Theorem 2, when X is projectively normal, the maps of cohomology are always injective, by an easy trace argument. Despite the fact that both Theorem 1 and Theorem 2