Multiplier ideal sheaves associated with weights of log canonical threshold one (Q317294): Difference between revisions

Given any plurisubharmonic function \(u\) on \(\Omega \subset \mathbb{C}^n\), following [\textit{A. M. Nadel}, Proc. Natl. Acad. Sci. USA 86, No. 19, 7299--7300 (1989; Zbl 0711.53056); Ann. Math. (2) 132, No. 3, 549--596 (1990; Zbl 0731.53063)] one defines the \textit{multiplier ideal sheaf} \(\mathcal{I}(u)\) as the sheaf of germs of holomorphic functions \(f\) such that \(|f|^2e^{-2u}\) is locally integrable. Then \(\mathcal{I}(u)\) is a coherent algebraic sheaf and the zero variety \( V(\mathcal{I}(u))\) is the set of points in a neighborhood of which \(e^{-2u}\) is not integrable. Such points occur only if \(u\) has logarithmic poles. The ideal \(\mathcal{I}(u)\) plays a very important role in the dictionary between analytic geometry and algebraic geometry since it converts an analytic object into an algebraic one and encodes information on the singularities of \(u\). When the Lelong number of \(u\) at \(x\in \Omega\), denoted by \(\nu(u,x)\), is less than \(1\), \textit{H. Skoda} showed that \(e^{-2u}\) is integrable near \(x\) and in particular that \(\mathcal{I}(u)_x\) is \(\mathcal{O}_x\), the ring of germs of analytic function at \(x\) [Bull. Soc. Math. Fr. 100, 353--408 (1972; Zbl 0246.32009)]. In [Adv. Math. 285, 1688--1705 (2015; Zbl 1343.32025)], \textit{Q. Guan} and \textit{X. Zhou} characterized the structure of \(\mathcal{I}(u)\) where the weight \(u\) has Lelong number one at \(x\), proving the so-called \textit{Demailly's openness conjecture}. In particular, they proved that \(\mathcal{I}(u)_x=\mathcal{O}_x\) or \(\mathcal{I}(u)_x=\mathcal{I}(\log|h|)\) where \(h\) is a defining function of a germ of a regular complex hypersurface through the point \(x\). The assumption \(\nu(u,x)=1\) is equivalent to the existence of a complex line \(L\) through \(x=(x', x'')\in \Omega \subset \mathbb{C}\times \mathbb{C}^{n-1}\) such that the \textit{log canonical threshold } of \(u|_L\) at \(x' \), denoted by \(c_{x'} (u|_L)\), is \(1\). Motivated by this, the authors considered the case when there exists a \(2\)-dimensional complex plane \(H\) such that \(c_{x'} (u|_H)=1\), where here \(x=(x', x'')\in \Omega \subset \mathbb{C}^2\times \mathbb{C}^{n-2}\). In this paper they give a precise description of \(\mathcal{I}(u)\) in this more general case. In particular, considering \(H=\{z_3=\cdots=z_n=0\}\), they prove that there exist new local coordinates \((w_1, w_2, z_3,..., z_n)\) near \(x\) such that \(\mathcal{I}(u)_x\) is equal to one of the following possible ideals: \(\mathcal{O}_x\), \((w_1, w_2) \cdot \mathcal{O}_x, (w_1)\cdot \mathcal{O}_x \) or \((w_1^2+w_2^2+f(z_3, \dots z_n)) \mathcal{O}_x\) where \(f\) belongs to the maximal ideal of \(\mathcal{O}_x\).