Reduction modulo p of crystalline Galois representations via {\mu}_p-equivariance

Technical Summary: Reduction Modulo $p$ of Crystalline Galois Representations via $\mu_p$-Equivariance

Problem Statement
The paper addresses the "weight part" of Serre's conjecture, specifically the problem of determining the possible weights of mod $p$ modular forms (or more generally, automorphic forms) that give rise to a specific mod $p$ Galois representation. In the context of $p$-adic Hodge theory, this translates to Question 1.1: Given a continuous mod $p$ representation $r: \text{Gal}_{\mathbb{Q}_p} \to \text{GL}_n(\mathbb{F}_p)$, what are the possible Hodge–Tate weights of crystalline representations $\rho: \text{Gal}_{\mathbb{Q}_p} \to \text{GL}_n(\mathbb{Z}_p)$ such that the reduction $\bar{\rho} \simeq r$?

While the set of predicted weights is known to depend only on the restriction of $r$ to the inertia subgroup $I_{\mathbb{Q}_p}$, a precise characterization of which mod $p$ representations arise from crystalline representations with specific Hodge–Tate weights has remained incomplete, particularly for general reductive groups and without genericity assumptions.

Methodology
The authors employ a synthesis of modern $p$-adic Hodge theory, specifically prismatic cohomology, and geometric representation theory involving affine Grassmannians.

1. Prismatic Framework and $\mu_p$-Equivariance:
   The core technical innovation is the discovery that Breuil–Kisin modules associated with crystalline representations possess a natural $\mu_p$-equivariant structure.

   • The Breuil–Kisin ring $S = W(k)[[u]]$ carries an action of the group scheme $\mu_p$ (the $p$-th roots of unity) via "loop rotation" ($u \mapsto \zeta u$).
   • The Frobenius map $\phi: S \to S$ is $\mu_p$-equivariant (with trivial action on the target).
   • The authors prove that for a crystalline representation, the associated Breuil–Kisin module $M$ admits a "canonical modification" $N$ which is $\mu_p$-equivariant. Crucially, the reduction of this modification modulo $p$ is isomorphic to the reduction of the original module $M$ (Theorem 4.2.3).
   • This structure is established via two constructions: one using the prismatization of $\text{Spf}(W(k))$ (specifically a map $\rho^\dagger$ distinct from the standard prism map) and another using the syntomification and Nygaard stacks.

2. Geometry of Affine Grassmannians:
   The authors translate the algebraic properties of these modules into the geometry of the affine Grassmannian $\text{Gr}_G$.

   • They analyze the fixed-point locus $\text{Gr}_G^{\mu_p}$ under the loop rotation action.
   • They establish a relationship between the standard dominance order ($\leq$) on cocharacters and a new partial order ($\uparrow$) defined via the $p$-dilated affine Weyl group.
   • A key geometric result (Proposition 2.4.10) states that the closure relations in the $\mu_p$-fixed locus of the affine Grassmannian are governed by the $\uparrow$ order, rather than the standard Bruhat or dominance orders.

3. Synthesis:
   By combining the $\mu_p$-equivariant structure of the Breuil–Kisin modules with the geometry of the affine Grassmannian, the authors relate the Hodge–Tate weights (which determine the position of the module in the generic fiber) to the inertial weights of the mod $p$ reduction (which determine the position in the special fiber).

Key Contributions and Results

• Theorem 1.2 (Theorem 5.5.6): Let $\rho$ be a crystalline representation with Hodge–Tate weights $\mu$. Then the semisimplification of its mod $p$ reduction, restricted to inertia, is isomorphic to a specific semisimple representation $\tau(\lambda, w)$. Here, $\lambda$ and $\mu$ are related by the partial order $\lambda \uparrow \mu$, and $w$ is an element of the Weyl group.

  • This provides a necessary condition on the inertial weights of the mod $p$ reduction.
  • The relation $\uparrow$ is defined via a sequence of reflections in the $p$-dilated affine Weyl group, generalizing the standard dominance order.

• Explicit Serre Weight Conjecture (Conjecture 6.3.4):
  The authors formulate a conjecture for the set of Serre weights $W(\rho)$ associated with a semisimple L-parameter $\rho$. They define an explicit set of weights $W_{\text{explicit}}(\rho)$ based on the $\uparrow$ relation and prove the inclusion $W_{\text{cris}}(\rho) \subseteq W_{\text{explicit}}(\rho)$, where $W_{\text{cris}}(\rho)$ is the set of weights admitting a crystalline lift.

  • They conjecture that $W(\rho) = W_{\text{cris}}(\rho) = W_{\text{explicit}}(\rho)$, providing a uniform candidate set of weights for general unramified reductive groups without requiring genericity assumptions.

• Elimination Direction:
  The paper proves the "elimination direction" of the Serre weight conjecture: if a weight is not in the explicitly predicted set $W_{\text{explicit}}(\rho)$, it cannot arise from a crystalline representation with the given Hodge–Tate weights.

• Compatibility with Existing Literature:
  The authors demonstrate that their explicit set of weights coincides with the predicted set $W_{\text{Her}}(\rho)$ of Gee–Herzig–Savitt in the generic case (Proposition 6.4.6). They also note that the key structural theorem in the proof of the Buzzard–Diamond–Jarvis conjecture is an immediate corollary of their results.

Significance and Claims
The paper claims to provide a new, structural constraint on the reduction modulo $p$ of crystalline Galois representations. By utilizing prismatic techniques to reveal a hidden $\mu_p$-equivariant structure on Breuil–Kisin modules, the authors bridge the gap between the Hodge–Tate weights (characteristic 0 data) and the inertial weights of the mod $p$ reduction.

The significance lies in:

1. Generality: The results are formulated for $L$-parameters of unramified connected reductive groups over $\mathbb{Q}_p$, extending beyond the classical $\text{GL}_n$ case.
2. Explicitness: The formulation of the Serre weight conjecture is explicit and relies on the combinatorial $\uparrow$ order, removing the need for genericity assumptions that often plague previous formulations.
3. Methodological Shift: The proof relies on the geometry of affine Grassmannians and prismatic stacks rather than classical $p$-adic Hodge theory alone, offering a new perspective on the relationship between crystalline lifts and their reductions.

The authors are modest regarding the sufficiency of their condition; they state that while $\lambda \uparrow \mu$ is necessary, it is not generally sufficient to guarantee the existence of a crystalline lift, though they conjecture it becomes sufficient when the Hodge–Tate weights are $p$-restricted (the case relevant to Serre's conjecture).