Table of Contents
Fetching ...

Formal smoothness of the Artin-Mazur formal groups

Livia Grammatica

TL;DR

This work analyzes the Artin-Mazur formal groups Φ^{i}(X, G_m) of a smooth proper variety X in characteristic p, giving cohomological criteria for formal smoothness and linking these to crystalline and Witt vector cohomology. The authors introduce C-smoothness as a practical detector for formal smoothness and leverage Nygaard filtrations to connect fppf and crystalline invariants, obtaining a torsion-free sufficient condition and a torsion-based necessary condition. They construct Igusa-type, characteristic-2 examples showing that Φ^{i} can be formally smooth for i<d but fail at i=d, illustrating sharpness and guiding principles for non-smooth behavior. An equivalent formulation in terms of Witt vector cohomology provides a unifying perspective and a computable criterion, tying formal smoothness to the surjectivity of H^{i}(X, W) → H^{i}(X, O_X). The results deepen understanding of higher Artin–Mazur groups and offer concrete methods to diagnose and engineer formal smoothness via crystalline and Witt-theoretic data.

Abstract

Let $X$ be a smooth proper variety over an algebraically closed field of positive characteristic $p$. We find cohomological conditions for the Artin-Mazur formal group functors $Φ^{i}(X,\mathbb{G}_m)$ to be formally smooth. We show that if all crystalline cohomology groups of $X$ are torsion-free (e.g. if $X$ is an abelian variety) then all of the $Φ^{i}(X,\mathbb{G}_m)$ are representable and formally smooth. We then identify a necessary condition for formal smoothness, which we use to give examples, for any $d\ge2$, of varieties $X$ for which $Φ^{i}(X,\mathbb{G}_m)$ is formally smooth when $i<d$, whereas $Φ^{d}(X,\mathbb{G}_m)$ is not. The constructions are inspired by Igusa's surface with non-smooth Picard scheme. Finally, we give a condition equivalent to formal smoothness in terms of Serre's Witt vector cohomology. The strategy relies on the notion of $C$-smoothness - where $C$ is the group algebra of $\mathbb{Q}_p/\mathbb{Z}_p$ - which is a condition that detects when a formal group is formally smooth, and on the use of the Nygaard filtration to relate fppf cohomology to crystalline cohomology.

Formal smoothness of the Artin-Mazur formal groups

TL;DR

This work analyzes the Artin-Mazur formal groups Φ^{i}(X, G_m) of a smooth proper variety X in characteristic p, giving cohomological criteria for formal smoothness and linking these to crystalline and Witt vector cohomology. The authors introduce C-smoothness as a practical detector for formal smoothness and leverage Nygaard filtrations to connect fppf and crystalline invariants, obtaining a torsion-free sufficient condition and a torsion-based necessary condition. They construct Igusa-type, characteristic-2 examples showing that Φ^{i} can be formally smooth for i<d but fail at i=d, illustrating sharpness and guiding principles for non-smooth behavior. An equivalent formulation in terms of Witt vector cohomology provides a unifying perspective and a computable criterion, tying formal smoothness to the surjectivity of H^{i}(X, W) → H^{i}(X, O_X). The results deepen understanding of higher Artin–Mazur groups and offer concrete methods to diagnose and engineer formal smoothness via crystalline and Witt-theoretic data.

Abstract

Let be a smooth proper variety over an algebraically closed field of positive characteristic . We find cohomological conditions for the Artin-Mazur formal group functors to be formally smooth. We show that if all crystalline cohomology groups of are torsion-free (e.g. if is an abelian variety) then all of the are representable and formally smooth. We then identify a necessary condition for formal smoothness, which we use to give examples, for any , of varieties for which is formally smooth when , whereas is not. The constructions are inspired by Igusa's surface with non-smooth Picard scheme. Finally, we give a condition equivalent to formal smoothness in terms of Serre's Witt vector cohomology. The strategy relies on the notion of -smoothness - where is the group algebra of - which is a condition that detects when a formal group is formally smooth, and on the use of the Nygaard filtration to relate fppf cohomology to crystalline cohomology.

Paper Structure

This paper contains 12 sections, 40 theorems, 109 equations, 7 figures.

Key Result

Theorem 1.1

(P:esempietto, T:esempio) Suppose $p=2$ and let $d\ge2$ be an integer. There exists a smooth proper $k$-variety $Z$ such that When $d=2$ there is an explicit $Z$ satisfying these conditions.

Figures (7)

  • Figure 1: The region where the spectral sequences of $X$ and $\tilde{E}$ coincide
  • Figure 2: the $E_2$ page
  • Figure 3: the $E_3$ page
  • Figure 4: the $E_2$ page
  • Figure 5: the $E_3$ page
  • ...and 2 more figures

Theorems & Definitions (74)

  • Theorem 1.1
  • Theorem 1.2
  • Theorem 1.3
  • Theorem 1.4
  • Definition 2.1
  • Lemma 2.2
  • proof
  • Remark 2.3
  • Proposition 2.4
  • proof
  • ...and 64 more