Integral closure and operations of algebraic tori of complexity 1 in affine varieties (Q367142)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Integral closure and operations of algebraic tori of complexity 1 in affine varieties
scientific article

    Statements

    Integral closure and operations of algebraic tori of complexity 1 in affine varieties (English)
    0 references
    0 references
    26 September 2013
    0 references
    In this paper, the author examines integral closures of multigraded affine algebras of complexity 1 and homogeneous ideals therein. Let \(\mathbf{k}\) be an algebraically closed field of characteristic 0 and \(X\) be an affine variety over \(\mathbf{k}\) equipped with an effective action of an algebraic torus \(\mathbb{T}=(\mathbf{k}^{\star})^n\). The coordinate algebra \(A=\mathbf{k}[X]\) is then multigraded by the weight lattice \(M=\mathbb{Z}^n\). Suppose that the complexity of \(X\) (=the codimension of general \(\mathbb{T}\)-orbits) is 1. If \(X\) is normal, then there exists a smooth curve \(C\), a solid polyhedral cone \(\omega\subset M\otimes\mathbb{Q}\) and a polyhedral divisor \(\mathfrak{D}=\sum_{z\in C}\Delta_z\cdot z\) (where \(\Delta_z\) are polyhedra in \(\text{Hom}(M,\mathbb{Q})\) with tail cone \(\omega^{\vee}\) satisfying some extra conditions) such that \(A=A(C,\mathfrak{D}):=\bigoplus_{m\in M}H^0(C,\mathcal{O}(\mathfrak{D}(m)))\cdot t^m\subset\mathbf{k}(C)[\mathbb{T}]\), where \(\mathfrak{D}(m)=\sum_{z\in C} \min_{v\in\Delta_z}v(m)\cdot z\) and \(t^m\) is the Laurent monomial on \(\mathbb{T}\) of multidegree \(m\) [\textit{K. Altmann} and \textit{J. Hausen}, Math. Ann. 334, No. 3, 557--607 (2006; Zbl 1193.14060); \textit{D. Timashev}, Contemp. Math. 460, 349--364 (2008; Zbl 1151.14037)]. For any \(X\), its normalization is again an affine \(\mathbb{T}\)-variety of complexity 1, whose coordinate algebra \(\widetilde{A}=A(C,\mathfrak{D})\) is the integral closure of \(A\). It follows that \(A=\mathbf{k}[C][f_1t^{m_1},\dots,f_rt^{m_r}]\subset\mathbf{k}(C)[\mathbb{T}]\) for some \(f_i\in\mathbf{k}(C)\), \(m_i\in M\). Conversely, given such a multigraded algebra \(A\) with quotient field \(\mathbf{k}(C\times\mathbb{T})\), its integral closure is given by \(\widetilde{A}=A(C,\mathfrak{D})\), where \(\omega\) is the convex cone generated by \(m_1,\dots,m_r\) and \(\Delta_z=\{v\mid v(m_i)\geq-\text{ord}_zf_i,\;i=1,\dots,r\}\) for any \(z\in C\). This is the first main result of the paper. Suppose now that \(A\) is integrally closed and let \(I\subset A\) be a homogeneous ideal generated by \(g_1t^{l_1},\dots,g_st^{l_s}\). Consider a polyhedron \(P=\text{conv}(l_1,\dots,l_s)+\omega\) and a polyhedral divisor \(\widetilde{\mathfrak{D}}=\sum_{z\in C}\widetilde\Delta_z\cdot z\), where \(\widetilde\Delta_z=\{(v,p)\in\Delta_z\times\mathbb{Q}\mid v(l_j)+p\geq-\text{ord}_zg_j,\;j=1,\dots,r\}\). It is proved that \(A[C,\widetilde{\mathfrak{D}}]\) is the integral closure of the Rees algebra \(\widehat{A}=\bigoplus_{k\geq0}I^k\). In particular, the integral closure of \(I\) is determined by \(P\) and \(\widetilde{\mathfrak{D}}\). In fact, there is a bijection between integrally closed ideals \(I\subset A\) and pairs \((P,\widetilde{\mathfrak{D}})\), where \(P\) is a polyhedron with tail cone \(\omega\) and \(\widetilde{\mathfrak{D}}\) is a polyhedral divisor satisfying some extra conditions, given by \(I=\bigoplus_{m\in P\cap M}H^0(C,\mathcal{O}(\widetilde{\mathfrak{D}}(m,1)))\cdot t^m\). This is the second main result. If \(C\) is affine, then there is a combinatorial sufficient condition in terms of geometry of \(P\) and \(\widetilde{\mathfrak{D}}\) for an integrally closed ideal \(I\) to be normal (which is equivalent to \(\widehat{A}\) be integrally closed). This yields a simple combinatorial proof for normality of integrally closed ideals on non-elliptic \(\mathbf{k}^{\star}\)-surfaces. Another corollary is the normality of any homogeneous ideal \(I\subset\mathbf{k}[x_0,\dots,x_n]\) (where the grading is with respect to the variables \(x_1,\dots,x_n\)) such that \(I^k\) is integrally closed whenever \(k\leq n\). It is worth noting that the results of the paper extend well-known results for toric varieties and monomial algebras and ideals [\textit{M. Hochster}, Ann. Math. (2) 96, 318--337 (1972; Zbl 0237.14019); \textit{G. Kempf} et al., Toroidal embeddings. I. Lecture Notes in Mathematics. 339. Berlin-Heidelberg-New York: Springer-Verlag (1973; Zbl 0271.14017); \textit{L. Reid} et al., Commun. Algebra 31, No. 9, 4485--4506 (2003; Zbl 1021.13008)], and for \(\mathbf{k}^{\star}\)-surfaces [\textit{H. Flenner} and \textit{M. Zaidenberg}, Osaka J. Math. 40, No. 4, 981--1009 (2003; Zbl 1093.14084)] to the general complexity 1 case.
    0 references
    algebraic torus
    0 references
    complexity
    0 references
    polyhedral divisor
    0 references
    integral closure
    0 references
    normality
    0 references
    Rees algebra
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references