Gorensteinness in Rees algebras of powers of parameter ideals (Q6114014): Difference between revisions

Let \(A\) be a commutative Noetherian ring and \(I\) an ideal. The Rees algebra of \(I\) is defined to be \[ \mathcal R(I) = \bigoplus_{n \geq 0} I^n t^n = A[It] \subset A[t] \] where \(t\) is an indeterminate. The Rees algebra often has a better property than \(A\). For example, very many ring \(A\) has a Cohen-Macaulay Rees algebra. In the present paper, authors studied local rings having a Gorenstein Rees algebra. Let \((A, \mathfrak m)\) be a local ring having finite local cohomology and \(d = \dim A \geq 2\). Then \(\mathcal R(\mathfrak q^n)\) is Cohen-Macaulay for any ideal \(\mathfrak q\) generated by a standard system of parameters \(a_1\), \dots, \(a_d\) and integer \(n \gg 0\). \textit{M. Herrmann} et al. [J. Algebra 200, No. 1, 279--311 (1998; Zbl 0933.13005)] showed that \(A\) is Gorenstein if \(\operatorname{depth} A \geq 2\) and \(\mathcal R(\mathfrak q^n)\) is Gorenstein. Therefore authors were interested in the case of \(\operatorname{depth} A = 1\). They showed that \(n = d\) and \(H_{\mathfrak m}^i(A) = 0\) (\(i \ne 1\), \(d\)) if \(\mathcal R(\mathfrak q^n)\) is Gorenstein. The main theorem of this paper is as follows: Assume \(d \geq 0\) and \(H_{\mathfrak m}^i(A) = 0\) (\(i \ne 1\), \(d\)). Then the followins conditions are equivalent. \begin{itemize} \item[1.] \(\mathcal R(\mathfrak q^d)\) is Gorenstein; \item[2.] \(H_{\mathfrak m}^1(A) \ne 0\), \(\ell(\operatorname{Hom}_A(A/\mathfrak m, H_{\mathfrak m}^1(A))) = 1\) and \(\mathfrak c :=\)annihilator of \(H_{\mathfrak m}^1(A)\) equals to \[ \sum_{i=1}^d (a_1, \dots, a_{i-1}, a_{i+1}, \dots, a_d)A: a_i; \] \item[3.] \(\operatorname{depth} A = 1\), \(\ell(\operatorname{Ext}_A^1(A/\mathfrak m, A)) = 1\), \(e_{\mathfrak c}(A) = 2 \ell(A/\mathfrak c)\) and \(\mathfrak q\) is a reduction of \(\mathfrak c\). \end{itemize} When this is the case, the \((S_2)\)-fication \(\tilde A\) of \(A\) is Gorenstein, \(\mathfrak c = \mathfrak q \tilde A = A: \tilde A\) and \(\ell(\tilde A/\mathfrak c) = 2 \ell(A/\mathfrak c)\).