Do there exist uncountably many abelian surfaces with good reduction over $\mathbb{Q}_p$ with pairwise non-isomorphic rational $p$-adic Tate modules?

If we took $l$-adic Tate modules there would be countably many at most by a result of Kisin and the fact that the moduli of ppav is of finite type.

This answer would suggest that there is only countably many $p$-adic Tate modules for elliptic curves with good reduction.

Is it at least true that there are uncountably many non-isomorphic crystalline representations of $\mathrm{Gal}(\overline{\mathbb{Q}_p}/\mathbb{Q}_p)$ coming from geometry?

  • 1
    $\begingroup$ Given an abelian variety $A$ over a finite field $k$ its lifts to abelian schemes over $W(k)$ are in bijection with rank $g$ saturated submodules of $H^1_{cris}(A/W(k))$ that reduce mod $p$ to the 1st step of the Hodge filtration and are isotropic for the cup product so they are parametrized by (the $W(k)[1/p]$-points of) a residue disk in the isotropic Grassmanian. The isomorphism classes of Galois representations are then parametrized by the quotient of this disk by $Aut_{\varphi} H^1_{cris}(A)[1/p]$. $\endgroup$ – SashaP Aug 9 '20 at 20:30
  • 1
    $\begingroup$ If the dimension of this algebraic group is less than $g(g+1)/2$ then the quotient has uncountably many points. This happens e.g. if $A$ is the product of $4$ pairwise non-isogenous elliptic curves, for in that case the automorphisms of a $\varphi$-module is an $8$-dimensional group and $g(g+1)/2=10$. Maybe this happens in smaller dimensions as well because the action on the Grassmanian is not free, but I'm not sure. $\endgroup$ – SashaP Aug 9 '20 at 20:30
  • 1
    $\begingroup$ *4 pairwise non-isogenous ordinary elliptic curves $\endgroup$ – SashaP Aug 9 '20 at 20:47
  • 1
    $\begingroup$ @MikhailBondarko Sorry, I forgot to mention that we should pick a polarization on $A$ of degree prime to $p$ (let's assume that $A$ has such, if it doesn't then the question of existence of algebraizable deformations becomes subtler), it induces a perfect pairing on $H^1_{cris}$, and the claim is that lifts of the Hodge filtration isotropic for this pairing give lifts of the abelian variety that admit a lift of the polarizing line bundle. $\endgroup$ – SashaP Aug 12 '20 at 22:01
  • 1
    $\begingroup$ Deformations of abelian schemes are unobstructed and one can use that to show that any lift of the Hodge filtration is realized by a lift of $A$ to a formal abelian scheme (for a reference, see e.g. Messing's "Crystals associated to Barsotti-Tate group", though there the approach is slightly different). The chosen polarization lifts to a given formal lift if and only if the lift of the Hodge filtration is isotropic. this is because the obstruction to lifting a line bundle onto $\mathfrak{A}$ is the image of its crystalline Chern class in the group $H^2(\mathfrak{A},\mathcal{O})$. $\endgroup$ – SashaP Aug 12 '20 at 22:01

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy