The geometric Breuil-Mézard conjecture for two-dimensional potentially Barsotti-Tate Galois representations
Ana Caraiani, Matthew Emerton, Toby Gee, David Savitt
TL;DR
This work develops a geometric framework for the Breuil–Mézard conjecture and the weight part of Serre's conjecture in dimension two by organizing moduli of mod $p$ Galois representations into stacks built from Breuil–Kisin modules and étale $\varphi$-modules. It introduces cycles $Z^{\sigma}$ on irreducible components labeled by Serre weights and proves that a representation lies in the weight set $W(\overline{r})$ precisely when it meets the support of the corresponding cycle, after passing through tamely potentially Barsotti–Tate deformation rings and related stacks $\mathcal{Z}^{\mathrm{dd}}$ and $\mathcal{Z}^{\tau}$. The paper establishes generic reducedness of the relevant deformation rings, provides a BM decomposition for the cycles on tamely BT stacks, and then transfers these results to the EGmoduli stack $\mathcal{X}_{2,\mathrm{red}}$, where non-Steinberg weights correspond to irreducible components and Steinberg weights decompose in a controlled way with multiplicity one. Collectively, these results realize a geometric Breuil–Mézard conjecture for the stacks $Z^{\mathrm{dd},1}$ and, via EGmoduli, for $X_{2,\mathrm{red}}$, offering a robust geometric lens on the $p$-adic Langlands program for $\mathrm{GL}_2$ over $p$-adic fields and advancing modularity-lifting techniques. The methods hinge on a blend of Breuil–Kisin module theory, $\varphi$-module techniques, versal deformation rings, and the interplay between local deformation data and global Serre weight information.
Abstract
We establish a geometrisation of the Breuil-Mézard conjecture for potentially Barsotti-Tate representations, as well as of the weight part of Serre's conjecture, for moduli stacks of two-dimensional mod p representations of the absolute Galois group of a p-adic local field.
