Heights of CM points on complex affine curves (Q1601124)

The purpose of this note is to refine a result of Edixhoven concerning a special case of the André-Oort conjecture, which says roughly that the irreducible components of the Zariski closure of any set of CM points in a Shimura variety are Shimura subvarieties. A complex number \(x\in\mathbb{C}\) is said to be CM if it is the \(j\)-invariant of a CM elliptic curve. A point \((x_1,x_2)\in\mathbb{C}^2\) is defined to be CM if both \(x_1, x_2\) are CM numbers. Consider the usual map \(\pi_N\) from a modular curve \(Y_0(N)\) to \(\mathbb{C}^2\), sending an isogeny \(\phi: E_1 \rightarrow E_2\) to the pair \((j(E_1),j(E_2))\). Viewing \(\mathbb{C}^2\) as the Shimura variety which is the moduli space of pairs of elliptic curves, its irreducible Shimura subvarieties are the lines \(\mathbb{C}\times \{x_2\}\) with \(x_2\) a CM point, \(\{x_1\}\times \mathbb{C}\) with \(x_1\) a CM-point, or the image of a map \(\pi_N\) for some integer \(N\geq 1\). Edixhoven proved, assuming the generalized Riemann hypothesis (GRH) for quadratic imaginary fields, that an irreducible algebraic curve in \(\mathbb{C}^2\) containing infinitely many CM points and such that neither of their projections to \(\mathbb{C}\) is constant, must be \(\pi_N(Y_0(N))\) for some \(N\geq 1\). Later on, André proved this result unconditionally, by different methods. The paper under review contains an effective generalization of Edixhoven's work to the affine space \(\mathbb{C}^n\). Here a point \((x_1,\dots,x_n)\) is CM if each \(x_i\) is the \(j\)-invariant of a CM-elliptic curve. The author defines the modular curves in \(\mathbb{C}^n\) as follows: given \((\sigma_1, \dots,\sigma_n)\in \text{GL}_2(\mathbb{Q})^n\), form the group \(\Gamma:=\bigcap _{i=1}^n \sigma_i^{-1}\text{SL}_2(\mathbb{Z})\sigma_i\); the quotient \(\Gamma\setminus\mathcal{H}\) of the Poincaré upper half-plane by the action of \(\Gamma\) is mapped to \(\mathbb{C}^n\) through the map sending \(\tau\in\mathcal{H}\) to \((j(\sigma_1(\tau)),\dots, j(\sigma_n(\tau)))\); the image \(Y_\Gamma\) of this map is a modular curve in \(\mathbb{C}^n\). The main result of the paper is the following theorem: Theorem 1.1: Assume GRH for imaginary quadratic fields. Given positive integers \(d_1,\dots, d_n, m\), there exists an effectively computable constant \(B=B(d_1,\dots,d_n,m)\) such that the following holds. Let \(X\) be an irreducible algebraic curve in \(\mathbb{C}^n\) defined over an algebraic field of degree \(m\) over \(\mathbb{Q}\), such that the degrees of the standard projections \(X\rightarrow \mathbb{C}\) are \(d_1,\dots, d_n\) respectively. Then \(X\) is a modular curve \(Y_\Gamma\) if and only if \(X\) contains a CM point of height greater than \(B\). Apart from passing to \(\mathbb{C}^n\), the advantages of this theorem compared to Edixhoven's result, are the effectivity of the constant \(B(d_1,\dots,d_n,m)\) and the fact that it suffices to find a single CM point of sufficient height, rather than finding infinitely many CM points. To prove Theorem 1.1, the author shows how to reduce it to the case \(n=2\) and then applies Edixhoven's result. The same idea, combined with André's result provides an unconditional (but ineffective) version of the theorem. Finally, it is shown that bounding the height of the CM points on a variety is equivalent to bounding their number.