Moduli interpretation of the integral anticanonical tower

101 Views Asked by At

This question is related to my reading of On torsion in the cohomology of locally symmetric varieties by Scholze.

In chapter $3$, using the theory of canonical subgroup, he produces Frobenius maps defined over $\text{Spf(}\mathbb{Z}_p^{\text{cycl}})$ between the integral models of strict neighborhoods of the ordinary locus for the modular curve. Then, he computes the projective limit along Frobenius of these maps, and he gets a formal scheme over $\text{Spf}(\mathbb{Z}_p^{\text{cycl}})$, whose generic fiber is perfectoid and is called the anticanonical tower of modular curves. Later on in his exposition, he writes that the $(C,\mathcal{O}_C)$-points of this perfectoid space (where $C$ is the completion of an algebraic closure of $\mathbb{Q}_p$, and $\mathcal{O}_C$ is its ring of integers) parametrize elliptic curves over $C$ with a trivialization of their Tate module.

First, I do not see why this construction provides a unique elliptic curve! First, I would like to say that the construction provides a projective system of elliptic curves over $C$, where every elliptic curve has $p^n$-torsion trivialized (for $n$ becoming bigger along the tower), where the maps defining the projective system are quotient by the canonical subgroup. But why is such a kind of system the same as a unique elliptic curve with Tate module trivialized?

Second question, does a similar description hold for the integral anticanonical tower? Is it true that an $R$ point of the anticanonical tower, where $R$ is a complete and flat (maybe normal) $\mathbb{Z}_p^{\text{cycl}}$-algebra, gives a family of elliptic curves over $R$ with a trivialization (at least a generic trivialization) of its Tate module seen as a sheaf? Thank you for any kind of suggestion!