scientific article; zbMATH DE number 1444655

Publication date 30 October 2001

Full work available at URL https://arxiv.org/abs/math/9812162

\(K3\) surfaces Calabi-Yau manifolds mirror maps mirror-moonshine phenomenon modularity Picard-Fuchs equations Shioda-Inose structure

Mathematics Subject Classification ID

Structure of families (Picard-Lefschetz, monodromy, etc.) (14D05) (K3) surfaces and Enriques surfaces (14J28) (3)-folds (14J30) Calabi-Yau manifolds (algebro-geometric aspects) (14J32) Supervarieties (14M30)

Review text

\textit{B. H. Lian} and \textit{S.-T. Yau} [Commun. Math. Phys. 176, 163-191 (1996; Zbl 0867.14017)] conjectured that the mirror map is modular, i.e., Hauptmodul for a small class of \(K3\) surface pencils obtained via orbifolding construction. Later the so-called mirror-moonshine phenomenon for mirror maps was extended to families of elliptic curves with section and families of lattice polarized \(K3\) surfaces (and more generally to families of Calabi-Yau manifolds). In the paper under review, it is determined when the mirror map \(q\)-series of certain elliptic curve and \(K3\) surface families are Hauptmoduln (genus zero modular functions). The modularity of mirror maps are characterised in terms of orbifold uniformization properties of the associated Picard-Fuchs differential equations. The period of the holomorphic top form of a family of Calabi--Yau manifolds defines a multi-valued solution to the associated Picard-Fuchs differential equation. The mirror map is defined as a local inverse to the projectivized period mapping at a regular singular point with maximal unipotent monodromy. Indeed, the mirror map is completely determined by the projective normal form of the Picard-Fuchs differential equation. Suppose that the differential equation has order \(2\). Then the mirror map is given by the inverse of the ratio of two solutions. Then the modularity of mirror map can be characterized as follows: The mirror map is an automorphic function if and only if the projective normal form of the Picard-Fuchs differential equation is a Schwarzian equation (or (orbifold) uniformizing differential equation). Families of elliptic curves with section have order \(2\) Picard-Fuchs differential equations. The moduli space of such a family is the upper half complex plane \({\mathbb H}\). Let \({\mathcal{J}}\) be the Kodaira functional invariant, which is the composition of the period map and the elliptic modular \(j\)-function. Theorem 1: The projective normal form of the Picard-Fuchs differential equation of a Weierstrass elliptic surface will orbifold uniformize if and only if there are no apparent singularities and the function invariant \({\mathcal{J}}(t)\) has zeros to orders \(=1\) or \(\equiv 0\pmod 3\) and ones to orders \(=1\) or \(\equiv 0\pmod 2\). Under this situation, the mirror map is modular. Let \(M_n=U\perp U\perp (-E_8)\perp (-E_8)\perp \langle-2n\rangle\) where \(n\) is a positive integer \(\geq 1\). Consider one parameter families of \(M_n\)-polarized \(K3\) surfaces. Then the Picard-Fuchs differential of \(M_n\)-polarized \(K3\) surfaces has order \(3\). Furthermore, such a family of \(K3\) surfaces admits a Shioda-Inose structure so that the Picard-Fuchs differential equation is the symmetric square of an order \(2\) ordinary Fuchsian differential equation. This implies that the mirror map of a \(M_n\)-polarized family of \(K3\) surfaces can be determined essentially by the order \(2\) differential equation. Another crucial fact is that the coarse moduli space of \(K3\) surfaces polarized by \(M_n\) is isomorphic to \(X_0(n)^*:={\mathbb H}/\Gamma_0(n)^*\). (Here \(\Gamma_0(n):=\{\left(\begin{smallmatrix} a & b \\ c & d \\ \end{smallmatrix}\right) \mid c\equiv 0\pmod n\} \subset \text{PSL}(2,{\mathbb Z})\) and \(F_n:= \left(\begin{smallmatrix} 0 & -1/\sqrt{n} \\ \sqrt{n} & 0 \end{smallmatrix} \right)\) denotes the Fricke involution. \(\Gamma_0(n)^*\) is the subgroup of \(\text{PSL}(2,{\mathbb R})\) generated by \(\Gamma_0(n)\) and \(F_n\)). The natural questions answered in this paper are: (A) Commensurability: When is the mirror map commensurable to a McKay-Thompson series (arising from the Monster)? Answer: When \(n\neq 49, 50\). (B) Modularity: Can one characterize the families of \(M_n\)-polarized \(K3\) surfaces whose mirror maps are Hauptmoduln for genus zero subgroups of \(\Gamma_0(n)^*\)? Answer will be described in what follows. Let \({\mathcal{H}}_n(z)\) denote the generalized functional invariant which is the composition of the truncated period map \(\Omega\) and the (normalized) Hauptmodul \(H_n\) onto the coarse moduli space, so \({\mathcal{H}}_n:=H_n\circ\Omega\). Theorem 2: The square root of the projective normal form of the Picard-Fuchs differential equation of a one parameter family of \(M_n\)-polarized \(K3\) surfaces (with genus zero subgroup \(\Gamma_0(n)^*\)) will orbifold uniformize if and only if it has no apparent singularities, and for each elliptic point \(p\in X_0(n)^*\), \({\mathcal{H}}_n(z)-p\) vanishes to permissible orders (which are completely determined a priori). Under this situtation, the mirror map is modular. For the entire collection see [Zbl 0935.00034].

Recommendations

Cited in

(22)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4953652)