Cohomology of ideals in elliptic surface singularities (Q1670777)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Cohomology of ideals in elliptic surface singularities
scientific article

    Statements

    Cohomology of ideals in elliptic surface singularities (English)
    0 references
    0 references
    6 September 2018
    0 references
    In the paper under review, the author continues the study of normal surface singularities \((A,\mathfrak m)\) and \(p_g\)-ideals in \(A\) [\textit{T.\ Okuma}, Math. Z. 249, No. 1, 31--62 (2005; Zbl 1091.32010)], [\textit{T. Okuma} et al, J. Algebra 499, 450--468 (2018; Zbl 1390.14016); Manuscripta Math. 150, No. 3--4, 499--520 (2016; Zbl 1354.13011); Proc. Am. Math. Soc. 145, No. 1, 39--47 (2016; Zbl 1357.13011)]. In this review, let \((A,\mathfrak m,k)\) be a two-dimensional local ring which is a normal surface singularity, i.e., \(A\) is excellent and normal, and contains an algebraically closed field isomorphic to \( k\). The author proves his results using assumption 1.1; this assumption holds if \(k\) has characteristic \(0\). Let \(f: X\to \text{Spec}(A)\) be a resolution of singularity, \(E:=f^{-1}(\mathfrak m)\) the exceptional locus, and \(E=\bigcup E_i\) the decomposition of \(E\) into its irreducible components. The \textit{geometric genus} \(p_g(A)\) of \(A\) is defined by \(p_g(A)=\ell_A(H^1(X,\mathcal O_X))\); this does not depend on the choice of the resolution \(X\). A rational singularity is characterized by \(p_g(A)=0\) [\textit{M. Artin}, Am. J. Math. 88, 129--136 (1966; Zbl 0142.18602)]. A \textit{cycle} is a divisor on \(X\) supported in \(E\). For a cycle \(B>0\) let \(\chi(\mathcal O_B)\) be the Euler characteristic and set \(\chi(B):=\chi(\mathcal O_B)\). The fundamental cycle on \(\text{Supp}(B)\) is the minimal cycle \(Z_B\) such that \(\text{Supp}(Z_B)=\text{Supp}(B)\) and \(Z_BE_i\leq 0\) for for all \(E_i\) with \(E_i\leq B\) (for a divisor \(D\) on \(X\) let \(DE_i\) be the intersection number). The following conditions are equivalent: (1) \(\chi (D)\geq0\) for all cycles \(D>0\) and \(\chi(F)=0\) for some cycle \(F>0\); (2) \(\chi(Z_E)=0\). A normal surface singularity satisfying these conditions was called elliptic by \textit{P. Wagreich} [Am. J. Math. 92, 419--454 (1970; Zbl 0204.56404)]. Let \((A,\mathfrak m)\) be an elliptic singularity. Then there exits a unique cycle \(E_{\min}\) such that \(\chi(E_{\min})=0\) and \(\chi(D)>0\) for all cycles \(D\) such that \(0<D<E_{\min}\) [\textit{H. B. Laufer}, Am. J. Math. 99, No. 6, 1257--1295 (1977; Zbl 0384.32003)]. The cycle \(E_{\min}\) is called a minimally elliptic cycle. The singularity \((A,\mathfrak m)\) is said to be \textit{minimally elliptic} if the fundamental cycle is minimally elliptic on the minimal resolution. Let \(I\) be an integrally closed \(\mathfrak m\)-primary ideal in \(A\); then there exists a resolution \(X\to\text{Spec}(A)\) and a cycle \(Z>0\) on \(X\) such that \(I\mathcal O_X=\mathcal O_X(-Z)\). In this case the ideal \(I\) is denoted by \(I_Z\) and \(I\) is said to be represented on \(X\) by \(Z\); clearly \(I_Z=H^0(X,\mathcal O_X(-Z))\). The invariant \(q(I)\) is defined by \(q(I)=\ell_A(H^1(X,\mathcal O_X(-Z)))\); \(q(I)\) does not depend on the choice of the representation of the ideal \(I\) [3]. The ideal \(I\) is called a \(p_g\)-\textit{ideal} if \(q(I)=p_g(A)\), and a cycle \(Z>0\) is called a \(p_g\)-\textit{cycle} if \(\mathcal O_X(-Z)\) is generated by global sections and \(\ell_A(H^1(X,\mathcal O_X(-Z)))=p_g(A)\). \(p_g\)-ideals have many nice properties [Zbl 1354.13011]: Let \(I\), \(I'\) be integrally closed \(\mathfrak m\)-primary ideals. Then \(I\) and \(I'\) are \(p_g\)-ideals iff \(II'\) is a \(p_g\)-ideals. If \(Q\) is a minimal reduction of \(I\), then \(I^2=QI\) (These results are well-known if \(A\) is regular and \(I\), \(I'\) are integrally closed; cf. [\textit{J. Sally} and \textit{C. Huneke}, Integral closure of ideals, ring and modules. Cambridge: Cambridge University Press (2006; Zbl 1117.13001)]. They go back to Zariski. They holds also if \(A\) is rational [\textit{J. Lipman}, Publ. Math. Inst. Hautes Études Sci. 36, 195--279 (1969; Zbl 0181.48903)]). The ideal \(I\) is a \(p_g\)-ideal iff the Rees algebra \(\bigoplus_{n\geq0}I^nt^n\subset A[t]\) is a Cohen-Macaulay normal domain [Zbl 1357.13011]. There is also a criterion for a cycle to be a \(p_g\)-cycle. In [Zbl 1390.14016] it is shown: For \(h\in I\) there exists \(h'\in I\) such that the integral closure of the ideal \((h,h')\) is a \(p_g\)-ideal. If \(A\) is rational, i.e. \(p_g(A)=0\), then every integrally closed \(\mathfrak m\)-primary ideal is a \(p_g\)-ideal [Zbl 0181.48903]. Conversely, this property characterizes a rational singularity because there always exist integrally closed \(\mathfrak m\)-primary ideals \(I\) with \(q(I)=0\) [3]. Let \(I\) be an \(\mathfrak m\)-primary ideal in \(A\) and \(Q\) a minimal reduction of \(I\). The \textit{normal reduction number} \(\overline r(I)\) of \(I\) is defined by \( \overline r(I)=\min\{r\in\mathbb Z_{\geq0}\mid \overline {I^{n+1}}=Q\overline {I^n}\,\, \text{for all}\,\,\, n\geq r\}\); this number is independent of the choice of the minimal reduction \(Q\). Set \(\overline r(A)=\max\{\overline r(I)\mid I\subset A \,\, \text{integrally closed}\,\, \mathfrak m\text{-primary ideal}\}\). If \(A\) is rational then from results of Lipman and [\textit{S. D. Cutkosky}, Invent. Math. 102, No. 1, 157--177 (1990; Zbl 0718.14025)] it follows that \(\overline r(A)=0\) iff \(A\) is a rational singularity. Let \(Z>0\) be a cycle on \(X\). The author calls \(\mathcal O_X(-Z)\) \textit{to have no fixed component} if \(H^0(X,\mathcal O_X(-Z))\neq H^0(X,\mathcal O_X(-Z-E_i))\) for every \(E_i\); to such a cycle the author associates a nonnegative integer \(n_0(Z)\) [\,cf.\ Remark 3.5]. Note that \(\mathcal O_X(-Z)\) has no fixed component when \(I\) is represented by \(Z\); then \(n_0(I):=n_0(Z)\) is independent of the representation \(Z\) of \(I\). In Cor.\ 3.9 it is shown that \(\overline r(I)=n_0(I)+1\). From this the author gets one of his main results, Theorem 3.3: If \(A\) is an elliptic singularity, then \(\overline r(A)=2\) . The invariant \(q\) is a function on the set of integrally closed \(\mathfrak m\)-primary ideals in \(A\). The set \(\text{Im}_A(q)\subset \mathbb Z\) is defined by \(\text{Im}_A(q)=\{q(I)\mid I\subset A\,\,\text{is an integrally closed}\,\, \mathfrak m\text{-primary ideal}\}\). One has \(\text{Im}_A(q)\subset\{0,1,\ldots,p_g(A)\}\); and one has equality if \(A\) is an elliptic singularity by Cor.\ 3.13; the converse does not hold, cf.\ Example 3.15. In the last part the author is interested in singularities for which the maximal ideal \(\mathfrak m\) is a \(p_g\)-ideal. In Def.\ 4.7 it is defined when \(A\) is a maximal elliptic singularity. Such a singularity is Gorenstein by a result of [\textit{S. S. Yau}, Trans. Am. Math. Soc. 257, No. 2, 269--329 (1980; Zbl 0343.32009]. In Theorem 4.10 the author proves the following result: Assume that \(A\) is not rational. Then \(A\) is Gorenstein and \(\mathfrak m\) is a \(p_g\)-ideal iff \(A\) is a maximally elliptic singularity with \(-Z_E^2=1\) where \(Z_E\) is the fundamental cycle on \(E\).
    0 references
    0 references
    normal surface singularity
    0 references
    \(p_g\)-ideals
    0 references
    elliptic surface singularities
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references