Universal formulas for counting nodal curves on surfaces (Q2902059)
From MaRDI portal
| This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use this page instead for the normal view: Universal formulas for counting nodal curves on surfaces |
scientific article; zbMATH DE number 6066963
| Language | Label | Description | Also known as |
|---|---|---|---|
| English | Universal formulas for counting nodal curves on surfaces |
scientific article; zbMATH DE number 6066963 |
Statements
17 August 2012
0 references
Göttsche's conjecture
0 references
Göttsche--Yau--Zaslow formula
0 references
algebraic cobordism
0 references
Hilbert scheme
0 references
degeneration formula
0 references
Universal formulas for counting nodal curves on surfaces (English)
0 references
Let \(S\) be any complex projective smooth surface with canonical bundle \(K\), and let \(r\) be any nonnegative integer. For every line bundle \(L\) on \(S\), denote by \(t_r(S,L)\) the number of \(r\)-nodal curves in a generic \(r\)-dimensional sublinear system of \(|L|\). (Here an \(r\)-nodal curve is a nodal curve with exactly \(r\) nodes.) In [Commun. Math. Phys. 196, No. 3, 523--533 (1998; Zbl 0934.14038)], \textit{L. Göttsche} conjectured that the number \(t_r(S,L)\) is a universal polynomial of degree \(r\) in the numbers \(L^2\), \(LK\), \(c_1(S)^2\), and \(c_2(S)\) if \(L\) is sufficiently ample with respect to \(r\). Note that the polynomial is universal in the sense that its coefficients are independent of \(S\) and \(L\).NEWLINENEWLINESuppose \(S^{[3r]}\) is the Hilbert scheme of \(3r\) points on \(S\), and \(S^r_{2,0}\) is the locally closed subset of \(S^{[3r]}\) which parametrizes subschemes of the form \(\coprod_{i=1}^r \mathrm{Spec} (\mathcal{O}_{S, x_i}/m^2_{S, x_i})\), where \(x_1, \ldots, x_r\) are distinct closed points on \(S\). Let \(S^r_2 \subset S^{[3r]}\) be the closure (with the reduced induced structure) of \(S^r_{2,0}\), and let \(L^{[3r]}\) be the rank \(3r\) tautological bundle on \(S^{[3r]}\) associated to \(L\), which is constructed from the universal subscheme of \(S \times S^{[3r]}\) and the obvious projections. \textit{L. Göttsche} introduced the intersection number NEWLINE\[NEWLINE d_r(S,L):=\int_{S^r_2} c_{2r}(L^{[3r]}) NEWLINE\]NEWLINE and showed that it coincides with the number \(t_r(S,L)\) if \(L\) is \((5r-1)\)-very ample (\(5\)-very ample if \(r=1\)), i.e., if the natural map \(H^0(S,L) \to H^0(\xi, L\otimes \mathcal{O}_\xi)\) is surjective for all zero-dimensional subschemes \(\xi \subset S\) of length \(5r\) (\(6\) if \(r=1\)).NEWLINENEWLINEIn the paper under review, the author studies the generating function \(\phi(S,L):=\sum_{r=0}^\infty d_r(S,L) \, x^r\) for every line bundle \(L\) on \(S\). Using algebraic cobordism theory and deriving a degeneration formula for \(d_r(S,L)\), she proves that \(\phi\) gives a homomorphism from the algebraic cobordism group \(\omega_{2,1}\), which is a four-dimensional vector space over \(\mathbb{Q}\) spanned by all pairs of surfaces and line bundles modulo extended double point relations, to the group \(\mathbb{Q}[[x]]^{\times}\) of units of \(\mathbb{Q}[[x]]\). (More details are provided in her paper [J. Differ. Geom. 90, No. 3, 439--472 (2012; Zbl 1253.14054)]). As a consequence, \(\phi(S,L)\) can be expressed as the product NEWLINE\[NEWLINE A_1^{L^2}A_2^{LK}A_3^{c_1(S)^2}A_4^{c_2(S)} NEWLINE\]NEWLINE for some power series \(A_1, A_2, A_3, A_4 \in \mathbb{Q}[[x]]^{\times}\) whose coefficients are independent of \(S\) and \(L\). (So far, there are no closed-form expressions for the coefficients of \(A_i\)'s, even though these coefficients are determined by the recursive formulas for \(\mathbb{P}^2\) and \(\mathbb{P}^1 \times \mathbb{P}^1\) due to \textit{L. Caporaso} and \textit{J. Harris} [Invent. Math. 131, No. 2, 345--392 (1998; Zbl 0934.14040)] and \textit{R. Vakil} [Manuscr. Math. 102, No. 1, 53--84 (2000; Zbl 0967.14036)] respectively.)NEWLINENEWLINEBy extracting the coefficient of \(x^r\) in \(A_1^{L^2}A_2^{LK}A_3^{c_1(S)^2}A_4^{c_2(S)}\), the author deduces that \(d_r(S,L)\) is a universal polynomial \(T_r(L^2, LK, c_1(S)^2, c_2(S))\) of degree \(r\) in \(L^2\), \(LK\), \(c_1(S)^2\), and \(c_2(S)\). This proves Göttsche's conjecture if \(L\) is \((5r-1)\)-very ample (\(5\)-very ample for \(r=1\)). Note that \textit{A.-K. Liu} proved the conjecture several years ago using symplectic approach (see [J. Differ. Geom. 56, No. 3, 381--579 (2000; Zbl 1036.14014); ``The algebraic proof of the Universality Theorem'', \url{arXiv:math/0402045}]). It is also worth mentioning that a different algebro-geometric proof of the conjecture was obtained by \textit{M. Kool, V. Shende}, and \textit{R. Thomas} under the weaker condition that \(L\) is \(r\)-very ample [Geom. Topol. 15, No. 1, 397--406 (2011; Zbl 1210.14011)].NEWLINENEWLINELet \(G_2=-\frac{1}{24}+\sum_{n=1}^\infty (\sum_{d | n} d) \, q^n\) (the second Eisenstein series), \(\Delta=q\prod_{k =1}^\infty (1-q^k)^{24}\), and \({D= q \frac{d}{d q}}\). The author also shows that there exist power series \(B_1\) and \(B_2\) in \(q\) whose coefficients are independent of \(S\) and \(L\) such that NEWLINE\[NEWLINE \sum_{r=0}^\infty T_r(L^2, LK, c_1(S)^2, c_2(S))(DG_2)^r=\frac{(DG_2/q)^{\chi(L)}B_1^{K^2}B_2^{LK}}{(\Delta D^2G_2/q^2)^{\chi(\mathcal{O}_S)/2}}, NEWLINE\]NEWLINE which is referred to as the Göttsche--Yau--Zaslow formula. In fact, the formula follows from the aforementioned results and the corresponding formula for generic \(K3\) surfaces due to \textit{J. Bryan} and \textit{N. C. Leung} [J. Am. Math. Soc. 13, No. 2, 371--410 (2000; Zbl 0963.14031)].NEWLINENEWLINEFor the entire collection see [Zbl 1245.00031].
0 references