Trisections on certain rational elliptic surfaces and families of Zariski pairs degenerating to the same conic-line arrangement (Q2071814)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Trisections on certain rational elliptic surfaces and families of Zariski pairs degenerating to the same conic-line arrangement |
scientific article |
Statements
Trisections on certain rational elliptic surfaces and families of Zariski pairs degenerating to the same conic-line arrangement (English)
0 references
31 January 2022
0 references
Let \(\phi : S \rightarrow C\) be an elliptic surface over a smooth projective curve \(C\) satisfying the conditions as follows: i) The morphism \(\phi\) is relatively minimal. ii) There exists at least one section \(O: C \rightarrow S\). One identifies \(O\) with its image, i.e., a curve on \(S\) meeting any fibre of \(\phi\) at one point. iii) There exists at least one singular fibre. Under these assumptions, the Neron-Severi group \(\mathrm{NS}(X)\) of \(S\) is finitely generated and torsion-free. Let \(\mathrm{MW}(S)\) be the set of sections \(\phi: S \rightarrow C\). Note that for a section \(s : C \rightarrow S\) one identifies \(s\) with it image in \(S\). Let \(E_{S}\) be the generic fibre of \(\phi: S \rightarrow C\), which can be regarded as a curve of genus \(1\) over the rational function filed \(\mathbb{C}(C)\). In this setting, \(S\) is known as the Kodaira-Neron model of \(E_{S}\) and we can identify the set of \(\mathbb{C}(C)\)-rational points \(E_{S}(\mathbb{C}(C))\) of \(E_{S}\) with \(\mathrm{MW}(S)\). For \(P \in E_{S}(\mathbb{C}(C))\), we denote the corresponding section by \(s_{P}\) and for \(s \in \mathrm{MW}(S)\) we denote the corresponding rational point by \(P_{s}\). Let \(D\) be a divisor on \(S\) and let \(\mathfrak{d}_{D}\) be the restriction of \(D\) to \(E_{S}\). Then \(\mathfrak{d}_{D}\) is a divisor on \(E_{S}\) defined over \(\mathbb{C}(C)\). By Abel's theorem on \(E_{S}\), one obtains \(P_{D} \in E_{S}(\mathbb{C}(C))\) from \(\mathfrak{d}_{D}\) and the corresponding section \(S_{P_{D}}\) which is also denoted by \(s(D)\). For a divisor \(D\) on \(S\), we say that \(D\) is vertical if all irreducible components are contained in fibres. A divisor \(D\) is said to be horizontal if \(D\) contains no vertical components. Let us recall that a multisection \(\phi : S \rightarrow C\) is an effective horizontal divisor \(D\) on \(S\). Furthermore, \(D\) is called \(m\)-section if \(D\cdot F = m\), where \(F\) is the divisor class of a fiber. One refers to \(3\)-sections as trisections. If \(Q\) is a reduced quartic which is not the union of four concurrent lines and which has at most nodes as singularities and let \(z_{0}\) be a smooth point of \(Q\). One can associated with \(Q\) and \(z_{0}\) a rational elliptic surface \(S_{Q,z_{0}}\) satisfying the above three conditions, a precise construction is presented on page \(4\) therein. In the present article, the authors study trisections of \(S_{Q,z_{0}}\) obtained as follows. Let \(E\) be a smooth cubic such that the intersection number \(i_{p}(E,Q)\) is even for every \(p \in E\cap Q\) and also \(z_{0} \not\in E\). In \(S_{Q,z_{0}}\), the curve \(E\) gives rise to either an irreducible curve \(\tilde{E}\) which is a \(6\)-section, or two irreducible \(E^{\pm}\), both of which are trisections. The first result of the present note can be formulated as follows. Proposition 1. Let \(Q\) be a nodal quartic (i.e., a possibly reducible quartic curve whose singularities are at most nodes) which is not the union of four lines. Let \(z_{0} \in Q\) be a general point on a non-linear component of \(Q\). Let \(E^{\pm}\) be trisections obtained as above. Then the value of the height pairing of \(P_{E^{\pm}}\) on \(S_{Q, z_{0}}\) is given by \[\langle P_{E^{+}}, P_{E^{+}} \rangle = \langle P_{E^{-}}, P_{E^{-}} \rangle = \frac{3}{2} - \frac{r}{2},\] where \(r\) is the number of nodes of \(Q\) that \(E\) passes through. In particular, \(r\leq 3\). After applying some calcuations obtained by the auhtoirs, they construct a family of Zariski pairs. Let \(C_{1}, C_{2}\) be smooth conics meeting transversely at four points \(\{p_{1}, p_{2},p_{3},p_{4}\}\) and let \(Q= C_{1} + C_{2}\). Let \(L\) be a bitangent line to \(Q\) and \(L_{ij}\) be the line through \(p_{i}\), \(p_{j}\). Theorem 1. There exist two family of plane curves \(\{ \mathcal{B}_{i,s}\}_{s \in \triangle_{\epsilon}}\) for \(i \in \{1,2\}\) over \(\triangle_{\epsilon} = \{s \in \mathbb{C} : |s| < \epsilon\}\) for sufficiently small \(\epsilon >0\) such that i) For \(s \neq 0\), \(\mathcal{B}_{i,s} = Q + E_{i,s} + L\) for \(i \in \{1,2\}\), where \(E_{i,S}\) are smooth cubics that are tangent to \(Q\) at \(6\) distinct smooth points of \(Q\) and intersects \(L\) transversely. ii) For \(s=0\), both \(\mathcal{B}_{i,0}\) with \(i \in \{1,2\}\) are \(Q + L_{12} + L_{13} + L_{23} + L\). iii) The pairs \((\mathcal{B}_{1,s}, \mathcal{B}_{2,s})\) are Zariski pairs for all \(s\neq 0\). In particular, there exists a family of Zariski pairs which degenerate to the same conic-line arrangement.
0 references
elliptic surfaces
0 references
Mordell-Weil lattice
0 references
topology of reducible plane curves
0 references
multisections
0 references
0 references
0 references
0 references
0 references