Formal groups, supersingular abelian varieties and tame ramification (Q652171)

Let \(\ell\) be a prime number. Let \(A/\mathbb{Q}_{\ell}\) be an abelian variety with good supersingular reduction. If \(A=E\) is an elliptic curve, then it is known that the field extension \(\mathbb{Q}_{\ell}(E[\ell])/\mathbb{Q}_{\ell}\) is tamely ramified; see \(\S 1\) of \textit{J. P. Serre} [Invent. Math. 15, 259--331 (1972; Zbl 0235.14012)]. One essential part in the proof is to show the following fact: the wild inertia group \(I_w\) acts trivially on the \(\ell\)-torsion points of \(E\) as a \(\text{Gal}(\overline{\mathbb{Q}}_{\ell}/\mathbb{Q}_{\ell})\)-module. The trick used here does not generalized to abelian variety case due to the fact that the dimension of the formal group law associated to the \(\ell\)-torsion points of \(A\) increases from \(1\) (in the elliptic curve case) to arbitrary \(n\) (in the general abelian variety case), which makes it impossible to do explicit computations. In this paper, the author provides conditions when \(I_w\) remains to act trivially on the \(\ell\)-torsion points of \(A\). First, the author proves a general result that works for general formal group law of dimension \(n\). Let \(K\) be a local field of characteristic zero and residual characteristic \(\ell\), \(\mathcal{O}\) the ring of integers, and \(\overline{\mathfrak{m}} \subset \overline{K}\) be the set of elements that has positive valuation. Let \(\mathbf{F}\) be a formal group law defined on \(\mathcal{O}\) and \(V\) be the \(\mathbb{F}_{\ell}\)-vector space of \(\ell\)-torsion points of \(\mathbf{F}(\overline{\mathfrak{m}})\). The author proved (Theorem 3.3) that under a certain hypothesis (Hypothesis 3.2.) on the \(\ell\)-adic valuation on \(V\), the wild inertia group \(I_w\) acts trivially on \(V\). Second, the author obtains more explicit sufficient conditions (Theorem 4.16) that ensure that Hypothesis 3.2 holds. Namely, if \(\mathbf{F}\) is a \(2\)-dimensional symmetric formal group law over \(\mathbb{Z}_{\ell}\) of height \(4\), and if the multiplication by \(\ell\) of the reduction of \(\mathbf{F}\) to \(\mathbb{F}_{\ell}\) satisfies that \(\overline{[\ell]}(Z_1, Z_2) = (\overline{f}_1(Z_1^{\ell^2}, Z_2^{\ell^2}), \overline{f}_2(Z_1^{\ell^2}, Z_2^{\ell^2}))\) for some \(\overline{f}_1, \overline{f}_2 \in \mathbb{F}_{\ell}[[Z_1, Z_2]]\), then the Hypothesis 3.2 holds. Third, the author presents two families of genus \(2\) curves \(C\) such that their Jacobians \(J(C)\) are abelian surfaces with good supersingular reduction, and moreover the corresponding formal group laws satisfy the hypotheses of Theorem 4.16; see Theorems 5.9 and 6.4. Thus the wild inertia group \(I_w\) acts trivially on \(V\), and hence the Galois extension \(\mathbb{Q}_{\ell}(J(C))/\mathbb{Q}_{\ell}\) is tamely ramified.