Bogomolov–Tian–Todorov theorems via crystalline periods of differential forms

Haitao Zou Fakultät für Mathematik, Universität Bielefeld, 33501 Bielefeld, Germany hzou@math.uni-bielefeld.de
(Date: November 3, 2026 (draft))
Abstract.

We prove Bogomolov–Tian–Todorov theorems for smooth proper varieties over perfect fields of positive characteristic using differential forms whose contractions jointly detect obstructions. Numerical Hodge–de Rham degeneration and crystalline torsion-freeness one degree above each form ensure unobstructed deformations over the Witt vectors and formal smoothness of forgetting the forms. This refines Brantner–Taelman’s Calabi–Yau theorem and gives a symplectic criterion involving only degree three. We also give criteria for smooth projective lifts and symplectic examples with degree-three crystalline torsions.

Key words and phrases:
crystalline cohomology, deformation theory, differential forms, symplectic varieties, period maps, liftings
2020 Mathematics Subject Classification:
Primary 14D15; Secondary 14F30, 14J32, 14J42

1. Introduction

The Bogomolov–Tian–Todorov theorem asserts that a compact Kähler manifold X with trivial canonical bundle has unobstructed deformations, even though the obstruction space H2⁢(X,TX) need not vanish. Contraction with a volume form makes obstructions visible in de Rham cohomology, where Hodge theory forces them to vanish. In positive characteristic this argument meets an essential obstruction: a closed form determines a de Rham class on the special fibre, but this class need not lift across a mixed-characteristic extension. Non-liftable varieties with trivial canonical bundle show that additional hypotheses are necessary [15, 9].

Brantner–Taelman resolve this difficulty for Calabi–Yau varieties whose Hodge–de Rham spectral sequence degenerates and whose crystalline cohomology is torsion-free [8, Theorem A]. Their animated crystalline period map and divided-power bootstrapping provide the framework for this paper. We prove a criterion whose hypotheses are tailored to the forms that detect obstructions. For a two-form the relevant cohomological degree is three, regardless of the dimension; for a volume form it is the middle degree; and several forms may detect obstructions jointly even when none does so individually. All conditions are imposed on the special fibre, and none presupposes a lift, even for p=2.

We also construct a smooth projective symplectic fourfold in characteristic three whose Hodge–de Rham spectral sequence degenerates at E1 and whose mixed-characteristic deformations are obstructed, despite admitting a projective symplectic lift over the Witt vectors (Theorem E).

1.1. The main criterion

Let k be a perfect field of characteristic p>0, and put W=W⁢(k). Throughout the paper, X/k is smooth, proper, and geometrically connected, of dimension d≥1. We write DefX/W for the functor of isomorphism classes of deformations of X over local Artinian W-algebras with residue field k, and similarly for other objects over k.

For an integer n, write (HDn) for the numerical equality

dimkHdRn⁢(X/k)=∑a+b=ndimkHb⁢(X,ΩXa),

which holds for every n if the Hodge–de Rham spectral sequence of X degenerates at E1. Fix differential forms 𝜶=(α1,…,αs) with αj∈H0⁢(X,ΩXrj) and 1≤rj≤d. Contraction with these forms defines the joint contraction map on obstructions

H2⁢(X,TX)⟶⨁j=1sH2⁢(X,ΩXrj−1),v⟼(ιv⁢αj)j. (1.1)

When the forms are closed, a deformation of (X,𝜶) is a deformation of X together with closed relative lifts of all the forms.

Theorem A.

Assume that (1.1) is injective and that, for every j, (HDrj+1) holds and Hcrysrj+1⁢(X/W) is torsion-free.

  1. (i)

    The deformation functors of X and of (X,𝜶) are unobstructed over W.

  2. (ii)

    The morphism Def(X,𝜶)/W→DefX/W forgetting the forms is formally smooth on discrete Artinian W-algebras.

Formal smoothness is meant in the relative sense: given a surjection A′↠A, a deformation with forms over A, and a marked lift to A′ of its underlying variety, the forms lift to this given variety. In particular, every Artinian deformation of X carries lifts of the forms 𝜶. Theorem A holds without geometric connectedness, which we assume throughout only for the sake of the volume-form and symplectic applications.

1.2. Volume and symplectic forms

A nowhere-vanishing top form induces an isomorphism TX≃ΩXd−1, and integral crystalline duality shows that Hcrysd⁢(X/W) is torsion-free if and only if Hcrysd+1⁢(X/W) is. Applying Theorem A gives the following degree-wise criteria for unobstructedness.

Corollary B.

Suppose that ωX≃𝒪X, that (HDd+1) holds, and that Hcrysd⁢(X/W) is torsion-free. Fix a nowhere-vanishing top form η. Then the deformation functors of X and of (X,η) are unobstructed over W, and forgetting the volume form is formally smooth on discrete Artinian bases. 111When preparing this paper, we noticed that Takamatsu also obtain Corollary B (see [28, Proposition 2.11]) with different methods.

Corollary C.

Let σ∈H0⁢(X,ΩX2) have invertible contraction map ισ:TX⟶∼ΩX1. Assume that (HD3) holds and that Hcrys3⁢(X/W) is torsion-free. Then σ is closed, the deformation functors of X and of (X,σ) are unobstructed over W, and forgetting the form is formally smooth on discrete Artinian bases.

Theorem A also applies beyond these two settings: for instance, Example 6.11 treats a product of plane quartics, whose canonical bundle is ample and whose obstruction space H2⁢(X,TX) is nonzero.

1.3. Projective lifts and algebraization

To obtain a projective W-scheme from compatible formal lifts, one must also extend an ample line bundle. For a line bundle L on X, put c1H⁢(L)=d⁢log⁡(L)∈H1⁢(X,ΩX1) and define

λL:H1⁢(X,TX)⟶H2⁢(X,𝒪X),ξ⟼ξ⁢⌟⁢c1H⁢(L). (1.2)
Theorem D.

Assume the hypotheses of Theorem A, and let L be an ample line bundle on X for which (1.2) is surjective. Then the deformation functors of (X,L) and (X,𝛂,L) are unobstructed over W. In particular, there is a smooth projective W-scheme lifting X, together with a relatively ample lift of L and closed relative lifts of all the αj.

The proof in Section 7.2 varies a geometric lift by H1⁢(X,TX) to cancel the obstruction in H2⁢(X,𝒪X), and then lifts the prescribed forms to the resulting variety. When H2⁢(X,𝒪X)=0, every compatible formal lift algebraizes with the chosen L (Proposition 7.1). Section 7 also treats prescribed formal lifts, using a crystalline Chern-class criterion independent of Theorem A; in particular, every smooth proper formal lift algebraizes projectively when the canonical or anticanonical bundle is ample.

Surjectivity of λL is not necessary for the existence of a projective lift. In forthcoming work, we give more easily verified criteria for the unobstructedness of polarized deformations of primitive symplectic varieties.

1.4. An obstructed symplectic fourfold

The crystalline torsion-freeness assumption in Corollary C cannot be omitted, even when the Hodge–de Rham spectral sequence degenerates in every degree and a projective symplectic W-lift exists.

Theorem E.

Let k=𝔽̅3 and W=W⁢(k). There exists a smooth projective symplectic fourfold (Y,σY) over k whose Hodge–de Rham spectral sequence degenerates at E1 and for which

Hcrys3⁢(Y/W)tors≃(W/3)⊕6.

Both DefY/W and Def(Y,σY)/W are obstructed, although (Y,σY) admits a smooth projective symplectic lift over W.

Since σY∧σY trivializes ωY, this gives a characteristic-three symplectic answer to the existence question raised in [8, Remark 1.1]: a variety with trivial canonical bundle and E1-degenerate Hodge–de Rham spectral sequence whose mixed-characteristic deformations are obstructed. Thus the crystalline torsion-freeness assumption in their Theorem A cannot be omitted without replacement. The same question is addressed by Zock using a bielliptic surface in characteristic two [34].

Proposition 8.5 constructs the fourfold as a free cyclic quotient of a product of ordinary elliptic curves and realizes three linearly independent geometric obstruction directions by a coherent Bockstein. In contrast, Proposition 8.6 gives an unobstructed smooth projective symplectic sixfold with nonzero degree-three crystalline torsion. Neither example is irreducible symplectic.

1.5. Outline of the proof

For a closed r-form α, let ℱα be the space-valued formal moduli problem of deformations of X carrying a closed lift of α. The period problem 𝒫αr parametrizes lifts of the point determined by α to marked maps A⁢[−r]→R⁢Γcrys⁢(X/W)⊗W𝐋A; its definition (3.7) requires no chosen integral period or rank-one direct summand.

Crystalline invariance gives a natural period map ℱα→𝒫αr on animated divided-power bases. The Kaledin–Verbitsky complex computes the tangent object of ℱα (Lemma 4.2). Using divided-power envelopes of the diagonal, we identify the derivative with contraction by α on vector fields and the filtered inclusion on forms. When

ια:TX⟶∼ΩXr−1, (1.3)

which happens precisely for volume and symplectic forms (Remark 6.3), the derivative is simply R⁢Γ⁢(X,Fr−1⁢ΩX∙)⁢[r]→R⁢Γ⁢(X,ΩX∙)⁢[r]. The calculation on shifted square-zero tests also identifies the map on complete obstruction spaces.

The equality (HDr+1) makes the relevant filtered maps strict, so injectivity of the joint contraction (1.1) detects the complete obstruction of a tuple. On the target, universal coefficients show that degree-r period classes lift whenever Hcrysr+1⁢(X/W) is torsion-free. Brantner–Taelman’s bootstrapping theorem [8, Proposition 7.8] then extends the conclusion from divided-power elementary extensions to all Artinian extensions. At p=2 this step requires an initial W2-point: Achinger–Suh’s description of the conjugate differential [1, Theorem 1.3 and Corollary 4.4] supplies a geometric W2-lift, and a filtered Bockstein argument lifts the forms to it using only the hypotheses of the theorem. Throughout, we rely on the foundations for formal moduli problems and animated crystalline cohomology developed in [8, 5, 23].

1.6. Torsion and the scope of the criterion

The torsion hypotheses ensure that every period lifts across Artinian extensions, and they are necessary for this: by the explicit hull of Proposition 3.9, the classical truncation 𝒫αrcl of the deformation problem of periods is unobstructed if and only if the adjacent crystalline cohomology group is torsion-free. This necessity concerns only the deformation problem of periods; it does not imply necessity for geometric unobstructedness. In the symplectic case, every actual obstruction lies in a canonical selective subspace, whose vanishing gives a refined criterion that tolerates some degree-three torsion (Section 8).

1.7. Relation with earlier work

Bogomolov proved unobstructedness for compact Kähler holomorphic symplectic manifolds [6], and Tian and Todorov did so for compact Kähler manifolds with trivial canonical bundle [30, 31]. Iacono–Manetti gave an algebraic proof for smooth projective varieties over an algebraically closed field of characteristic zero [16, Theorem 6.2 and Corollary 6.5]. Using the Cartan homotopy ℒv=d⁢ιv+ιv⁢d and Hodge–de Rham degeneration, their Theorem 6.4 places obstruction classes in the kernel of contraction against the cohomology of the canonical bundle; a volume form makes this test injective, and they show moreover that the governing differential graded Lie algebra is homotopy abelian. Theorem A keeps only the injectivity of contraction on H2⁢(X,TX) from this mechanism, and it concludes unobstructedness and relative lifting of the forms rather than homotopy abelianity.

Kaledin–Verbitsky classify deformations of a smooth complex algebraic manifold with a closed nondegenerate two-form by their periods, which lie in the formal completion of ℍ2⁢(X,F1⁢ΩX∙) at the class of the form, under an admissibility condition on every infinitesimal deformation [19, Definitions 2.1 and 2.4, Theorem 2.6]. The same truncated complex appears in our symplectic tangent calculation as R⁢Γ⁢(X,F1⁢ΩX∙)⁢[2]. Properness allows us to impose all conditions on the special fibre: (HD3) for strictness and degree-three torsion-freeness for lifting periods. We obtain smoothness; a classification by periods would also require control of the whole tangent fibre, including infinitesimal automorphisms.

Brantner–Taelman’s period map attaches to a deformation Y/A the arrow R⁢Γcrys⁢(X/W)⊗W𝐋A→R⁢Γ⁢(Y,𝒪Y), whose fibre is the first step of the Hodge filtration [8, Definition 6.13]. Their Theorem 7.16 sharpens [8, Theorem A]: it assumes a splitting R⁢Γcrys⁢(X/W)≃M⊕N lifting the quotient R⁢Γ⁢(X,𝒪X), torsion-freeness of π0⁢MapModW⁡(N,M), and explicit conditions detecting obstructions, and thereby isolates the integral linear problem more precisely than torsion-freeness in all degrees [8, Remark 7.17]. At p=2 it requires a W2-lift, which their Theorem 7.18 supplies under the hypotheses of their Theorem A. Our deformation problem of periods (3.7) tracks a specified map A⁢[−r]→(R⁢Γcrys⁢(X/W)⊗W𝐋A), so its defining complex is R⁢Γcrys⁢(X/W)⁢[r] and the integrality condition lies in the adjacent degree r+1. We compute the derivative of our period map, which differs from the first-filtration derivative of [8, Theorem 6.18], in Propositions 4.4 and 4.6, and we construct the initial lift of the tuple at p=2 in Section 5.4.

Organization

Section 2 recalls formal deformation theory; Sections 3 and 4 construct the period map and compute its derivative, and Section 5 proves Theorem A. Sections 6 and 7 give applications and algebraization results, and Section 8 treats the selective refinement and torsion examples.

Conventions

We write Wn=Wn⁢(k)=W/pn⁢W and K=W⁢[1/p]. We abbreviate ΩXa=ΩX/ka, put TX=ℋ⁢o⁢m𝒪X⁢(ΩX1,𝒪X) and ωX=ΩXd, and write hb⁢(X,ℰ)=dimkHb⁢(X,ℰ) for a coherent sheaf ℰ; the argument X is omitted when clear. Deformations in this paper are marked, that is, equipped with an identification of the special fibre with X, and they are unpolarized unless a line bundle is specified. A deformation functor is unobstructed if it takes every surjection of such algebras to a surjection of sets.

We write DefX/W for the set-valued classical deformation functor and DefX for the space-valued geometric deformation problem, retaining automorphisms and higher homotopies. The deformation problems ℱα of pairs and 𝒫αr of periods are also space-valued. The superscript cl denotes classical truncation, so DefXcl=DefX/W; the superscript closed used below refers instead to closed differential forms.

We use sans-serif letters for auxiliary ∞-categories, calligraphic letters for sheaves, complexes, and algebraic lifts, and Fraktur letters for formal schemes. We write H for cohomology and ℍ for sheaf hypercohomology; crystalline and filtered de Rham complexes are denoted by R⁢Γcrys⁢(X/W), R⁢ΓdR⁢(Y/A), and Fa⁢R⁢ΓdR⁢(Y/A).

Acknowledgement

The author thanks Lie Fu, Zhiyuan Li and Teppei Takamatsu for helpful discussions and comments.

AI declaration

This paper grew out of several research notes, written with the help of large language models (LLMs), aimed at understanding the formal deformation theory of Calabi–Yau varieties developed in [8]. We asked ChatGPT Pro to survey the classical algebraic proofs of the Bogomolov–Tian–Todorov theorem, which drew our attention to the deformation theory of symplectic varieties in characteristic zero due to Kaledin and Verbitsky [19]. It also supplied elementary deformation-theoretic computations suggesting that the argument should extend to positive and mixed characteristic under the natural assumptions of Hodge–de Rham degeneration and torsion-free crystalline cohomology. We then tried to construct animated period maps following the ideas of Kaledin and Verbitsky, but for arbitrary closed forms rather than only symplectic ones. Finally, we arrived at the divided-power Cartan homotopy formula of Proposition 4.4 with the help of ChatGPT 6 Astra and ChatGPT 6.1 Sol. ChatGPT was also used to investigate the case where Hcrys3⁢(X/W) is not torsion-free, and it produced the symplectic fourfold and sixfold of Section 8.3.

We also pushed ChatGPT 6 Astra to find useful refinements in the presence of torsion via prismatic deformation theory, but it did not produce interesting results beyond the classical framework.

The author drafted the paper. Claude Code was used to help with proofreading, readability, organization, and notation.

2. Preliminaries on formal deformation theory

We collect the conventions and results from [8, §§2–4, 6.1, 7.4] that we use to construct the deformation problems of varieties, pairs, and periods. Throughout, k is a perfect field of characteristic p>0 and W=W⁢(k). Our tangent object 𝕋ℱ is the tangent fibre of [8, Remark 3.11], with the same shift convention.

2.1. Categorical conventions

Write 𝒮 for the ∞-category of spaces. For an ∞-category 𝖢, we denote its mapping spaces by Map𝖢⁡(−,−) and its core, consisting of objects and equivalences, by 𝖢≃. All limits and colimits are taken in the relevant ∞-category, so a point of E×BD includes a path between the images of its two components in B. In particular, a point of the fibre fibb⁡(E→B)=E×B{b} includes an identification with b; in a deformation problem, this identification is the marking. Natural transformations include the compatible homotopies in their naturality diagrams [21, §§1.2, 4.2].

For an animated algebra A, write ModA for its stable ∞-category of modules and Perf⁡(A) for the perfect modules, namely the retracts of finite cell modules. For discrete A, the homotopy category of ModA is the unbounded derived category of A-modules. We use cohomological shifts Hj⁢(M⁢[s])=Hj+s⁢(M) and hence πi⁢M=H−i⁢(M). In particular, a discrete module V⁢[n] is concentrated in cohomological degree −n. The underlying infinite loop space is denoted Ω∞⁢M, so

MapModA⁡(A⁢[−r],M)≃Ω∞⁢(M⁢[r]),πi⁢MapModA⁡(A⁢[−r],M)=Hr−i⁢(M)(i≥0). (2.1)

In a stable ∞-category, fib⁡(M→N)≃cofib⁡(M→N)⁢[−1]. Tensor products in stable module ∞-categories are derived, and the symbol ⊗𝐋 emphasizes this convention. Unadorned tensor products of discrete modules denote ordinary tensor products. Tensoring with a fixed module preserves finite limits and colimits, and Ω∞ preserves limits [22, §1.1]. The stable ∞-category of decreasing filtered modules is ModAfil=Fun⁡(ℤop,ModA), with grFa⁡D=cofib⁡(Fa+1⁢D→Fa⁢D) and limits and colimits computed at each filtration step.

2.2. Animated algebras and cotangent complexes

The ∞-category AnCAlgW of animated commutative W-algebras is obtained from simplicial commutative W-algebras by inverting weak equivalences, or from polynomial W-algebras by animation [8, §2.2]. In positive characteristic this ∞-category differs from that of connective 𝔼∞-W-algebras, but its module ∞-categories are still stable.

Following [8, Definition 3.1], let ArtWan be the ∞-category of augmented animated W-algebras A→k such that π0⁢A is local Artinian with residue field k and ⨁i≥0πi⁢A has finite length over π0⁢A. Write ArtW for the ordinary category of its discrete objects; for A∈ArtW, let 𝔪A denote its maximal ideal. For a finite-dimensional k-vector space V and n≥0, the split square-zero algebra k⊕V⁢[n] has augmentation ideal V⁢[n] and W-structure through k, which is denoted by sqz⁡(V⁢[n]) in [8]. The case n=0, V=k gives the discrete ring of dual numbers k⁢[ϵ]/(ϵ2).

For R→B in animated algebras, the cotangent complex LB/R∈ModB represents derived derivations: for a connective B-module M,

DerR⁡(B,M):=Map(AnCAlgR)/B⁡(B,B⊕M)≃MapModB⁡(LB/R,M). (2.2)

Here the maps on the left are sections of B⊕M→B. Cotangent complexes satisfy transitivity and commute with derived base change [8, Corollary 2.18 and Propositions 2.19–2.20]. For a smooth map of discrete algebras R→B, LB/R≃ΩB/R1 in degree zero. The same derivation formula holds for sheaves of animated algebras [8, Proposition 2.31].

2.3. Formal moduli problems and tangent objects

Definition 2.1.

A functor ℱ:ArtWan→𝒮 is a formal moduli problem if ℱ⁢(k) is contractible and the natural map

ℱ⁢(B×AD)⟶ℱ⁢(B)×ℱ⁢(A)ℱ⁢(D)

is an equivalence whenever B→A←D is surjective on π0 on both sides. We denote the resulting ∞-category by FMPW. Its classical truncation is ℱcl⁢(A)=π0⁢ℱ⁢(A) for A∈ArtW.

This is [8, Definition 3.6]; the pullback condition is called infinitesimal cohesiveness, and it also holds when only one of the two maps is surjective on π0. By [8, Lemma 3.7], ℱcl is a deformation functor: it sends classical pullback squares with one map surjective to surjections of sets, and to bijections when the common base is k and the other ring is k⁢[ϵ]/(ϵ2). The condition ℱ⁢(k)≃∗ fixes the special object and its marking, but it does not imply that ℱ⁢(A) is nonempty for every A.

The tangent object 𝕋ℱ∈Modk is characterized by natural equivalences of pointed spaces

ℱ⁢(k⊕V⁢[n])≃Ω∞⁢(𝕋ℱ⊗kV⁢[n]),π0⁢ℱ⁢(k⊕V⁢[n])=Hn⁢(𝕋ℱ⊗kV), (2.3)

for finite-dimensional V and n≥0 [8, Remark 3.11]. The basepoints come from the split maps k→k⊕V⁢[n], and the equivalences are compatible with the deloopings

k⊕V⁢[n]≃k×k⊕V⁢[n+1]k,ℱ⁢(k⊕V⁢[n])≃Ω⁢ℱ⁢(k⊕V⁢[n+1]).

More generally,

πi⁢ℱ⁢(k⊕V⁢[n])=Hn−i⁢(𝕋ℱ⊗kV)(i≥0). (2.4)

Thus H−1⁢(𝕋ℱ) records infinitesimal automorphisms, H0⁢(𝕋ℱ) is the ordinary tangent space, and H1⁢(𝕋ℱ) is the complete obstruction space described below.

A natural transformation ℱ→𝒢 induces a map 𝕋ℱ→𝕋𝒢. It is an equivalence of formal moduli problems if and only if it induces equivalences on all shifted square-zero tests [8, Proposition 3.9]. The criterion applies to a given transformation: an equivalence of bare tangent complexes does not construct one. The additional nonlinear operations are encoded by the partition Lie formalism [8, Remark 3.11]; see also [7]. Our arguments use the functorial tangent and obstruction constructions rather than this formalism.

2.4. Elementary extensions and complete obstructions

An elementary extension A′↠A in ArtW is one whose kernel I satisfies 𝔪A′⁢I=0, so that I is a finite-dimensional k-vector space. By [8, Lemma 3.3], it is classified by a map κ:A→k⊕I⁢[1] in a homotopy pullback

A′Akk⊕I⁢[1].κ (2.5)

The lower map is the zero section. Equivalently, the extension class is represented by an A-linear map LA/W→I⁢[1], where A acts on I through k. The W-structure on κ retains mixed-characteristic extension classes: Lk/W≃k⁢[1], and the nonsplit extension W2⁢(k)↠k gives a nonzero element of π0⁢MapModk⁡(Lk/W,k⁢[1])≃k.

For ℱ∈FMPW and x∈ℱ⁢(A), applying ℱ to (2.5) identifies the marked lift space with

fibx⁡(ℱ⁢(A′)→ℱ⁢(A))≃Pathℱ⁢(k⊕I⁢[1])⁡(ℱ⁢(κ)⁢(x),0). (2.6)

The obstruction class of x is

obℱ⁡(x;A′/A)∈π0⁢ℱ⁢(k⊕I⁢[1])=H1⁢(𝕋ℱ)⊗kI. (2.7)

A lift of x on A exists exactly when obℱ⁡(x;A′/A) vanishes. Once a marked lift is chosen, the lift space is identified with Ω∞⁢(𝕋ℱ⊗kI); its components form a torsor under H0⁢(𝕋ℱ)⊗kI, and its relative automorphisms are H−1⁢(𝕋ℱ)⊗kI [8, §3.3].

These constructions are natural in ℱ. In particular, a transformation f:ℱ→𝒢 carries obℱ⁡(x;A′/A) to ob𝒢⁡(f⁢(x);A′/A) by the induced map on H1, and gives a compatible map of marked lift spaces. Every surjection in ArtW factors into elementary extensions [8, Corollary 3.4]. Hence ℱcl is unobstructed exactly when all its obstruction classes vanish [8, Proposition 3.14], and H1⁢(𝕋ℱ)=0 is a sufficient condition for this.

2.5. Divided-power bases and tangent tests

Give W its canonical divided powers on p⁢W. Animated divided-power rings are generated from divided-power polynomial rings by sifted colimits [8, Definition 4.4]. Let PDArtWan be the ∞-category of animated divided-power W-algebras whose divided-power reduction is k and whose underlying augmented algebra belongs to ArtWan [8, Definition 6.1]. A divided-power formal moduli problem satisfies Definition 2.1 with pullbacks taken in this ∞-category [8, Definition 6.2]. The forgetful functor und preserves limits, so ℱ|pd:=ℱ∘und is a divided-power formal moduli problem for every ℱ∈FMPW [8, Corollary 4.7 and Lemma 6.3].

For a connective perfect k-module V, the divided-power square-zero algebra k⊕pdV has underlying animated algebra k⊕V [8, Definition 4.32 and §6.3]. Consequently ℱ|pd has the same tangent object as ℱ, with the same shifted tests and degree conventions. Restriction to divided-power bases should not be confused with deforming sheaves of divided-power rings; the two problems are related by the induced divided-power construction of [8, Construction 6.8].

A discrete divided-power extension A′↠A with kernel I is elementary when [8, Definition 7.7]

𝔪A′⁢I=0,γm⁢(I)=0(m≥2). (2.8)

By [8, Proposition 7.15], its classifying square (2.5) exists in PDArtWan with target k⊕pdI⁢[1]. Thus the same complete obstruction and lift-space descriptions apply to the restricted problems ℱ|pd, using their divided-power square-zero tests. At p=2, the extension W2⁢(k)↠k is not elementary in this sense, because γ2⁢(2)=2(mod4). This is why the bootstrapping theorem Theorem 5.5 has a separate initial-lift hypothesis.

2.6. Geometric deformations

For a k-scheme X, fix its small Zariski site and let 𝖣=Shv⁡(X,AnCAlg) be the ∞-category of Zariski sheaves of animated algebras on X. Let c⁢(A) be the sheafification of the constant presheaf with value A, and write 𝖣A=𝖣c⁢(A)⁣/≃Shv⁡(X,AnCAlgA). Consider the geometric deformation problem DefX constructed as in [8, Construction 3.22]. The marked-fibre description of Remark 3.23 therein will serve as our working definition.

Lemma 2.2.

For A∈ArtWan with augmentation a:A→k, the geometric deformation problem DefX admits a natural equivalence

DefX⁡(A)≃fib𝒪X⁡(𝖣A≃⟶k⊗A𝐋(−)𝖣k≃). (2.9)

Thus its points are pairs (ℬ,β), where ℬ∈𝖣A and β:k⊗A𝐋ℬ⟶∼𝒪X is a specified equivalence in 𝖣k, or equivalently pushout squares

c⁢(A)c⁢(k)ℬ𝒪X.c⁢(a)

A path from (ℬ,β) to (ℬ′,β′) consists of an equivalence u:ℬ⟶∼ℬ′ in 𝖣A together with a specified homotopy β′∘(k⊗A𝐋u)≃β; all higher compatibility homotopies are retained. For an augmented map f:A→A′, base change sends this pair to (A′⊗A𝐋ℬ,βf), where

βf:k⊗A′𝐋(A′⊗A𝐋ℬ)≃k⊗A𝐋ℬ⟶𝛽𝒪X.
Proof.

In [8, Construction 3.22], consider

𝖢=AnCAlg×𝖣Fun⁡(Δ1,𝖣),q:𝖢⟶AnCAlg,

where the two maps to 𝖣 are c and source evaluation, and q is the first projection. Its objects are structure maps c⁢(A)→ℬ. For f:A→A′, the pushout universal property identifies q-cocartesian transport with f!⁢ℬ=c⁢(A′)∐c⁢(A)ℬ=A′⊗A𝐋ℬ. An arrow (A,ℬ)→(A′,ℬ′) over f is therefore cocartesian exactly when the comparison f!⁢ℬ→ℬ′ is an equivalence, or equivalently its square in 𝖣 is a pushout. Thus the ∞-subcategory 𝖢cocart of all objects and cocartesian arrows is the left fibration classified by A↦𝖣A≃; this is the dual of [21, Corollary 2.4.2.5].

The fibre of 𝖢/(k,𝒪X)cocart→AnCAlg/k over a:A→k consists of cocartesian arrows (A,ℬ)→(k,𝒪X), hence of equivalences β:k⊗A𝐋ℬ⟶∼𝒪X. This gives (2.9) and the pushout description, with all pushouts computed in sheaves. Paths in this homotopy fibre are equivalences u together with the stated marking homotopy and all higher compatibilities. For fixed ℬ, the space of markings is Map𝖣k≃⁡(k⊗A𝐋ℬ,𝒪X), which need not be contractible. Associativity of sheaf pushouts gives the displayed marking βf and naturality in A. ∎

Proposition 2.3 ([8, Proposition 3.25]).

The geometric deformation problem DefX is a formal moduli problem.

For a quasi-compact and quasi-separated scheme X, the sheaf-algebra construction agrees with deformation in derived schemes, and its objects are flat over A [8, Corollaries 3.32 and 3.35]. The flatness criterion states that a connective A-module is flat if and only if its derived special fibre is discrete [8, Lemma 3.34]. For smooth proper X and discrete A, the resulting space is the usual groupoid of smooth proper marked deformations. Its classical truncation is therefore DefXcl=DefX/W.

Let Mod⁡(X,𝒪X) denote the stable ∞-category of sheaves of 𝒪X-modules. We record the tangent calculation of [8, Corollary 3.26 and Remark 3.27], including the passage from square-zero tests to the tangent object.

Proposition 2.4.

For a smooth proper k-scheme X, there is a natural equivalence

𝕋DefX≃R⁢Γ⁢(X,R⁢ℋ⁢o⁢m𝒪X⁢(LX/k,𝒪X))⁢[1]. (2.10)

Here R⁢ℋ⁢o⁢m is the derived internal sheaf Hom. Since LX/k≃ΩX1, this specializes to

𝕋DefX≃R⁢Γ⁢(X,TX)⁢[1]. (2.11)

Consequently, the infinitesimal automorphism, tangent, and complete obstruction spaces are respectively H0⁢(X,TX), H1⁢(X,TX), and H2⁢(X,TX); all three are finite-dimensional.

Proof.

Put E=R⁢Γ⁢(X,R⁢ℋ⁢o⁢m𝒪X⁢(LX/k,𝒪X)). For a finite-dimensional k-vector space V and n≥0, [8, Corollary 3.26] and the internal-Hom adjunction give

DefX⁡(k⊕V⁢[n]) ≃MapMod⁡(X,𝒪X)⁡(LX/k,𝒪X⊗kV⁢[n+1])
≃Ω∞⁢(E⁢[1]⊗kV⁢[n]).

Here LX/k≃ΩX1 is finite locally free, and the coefficient complex is a finite direct sum of shifts of k, so its tensor product commutes with derived global sections. Coefficient naturality and compatibility with square-zero deloopings identify 𝕋DefX with E⁢[1] by (2.3), as in [8, Remark 3.21]. Since R⁢ℋ⁢o⁢m𝒪X⁢(ΩX1,𝒪X)≃TX, this proves both formulas; taking cohomology and using properness gives the remaining assertions. ∎

2.7. Descent

Let U∙→X be the Čech nerve of a finite affine Zariski cover. Descent for sheaves of animated algebras and their equivalences gives

DefX⁡(A)≃Tot[q]∈Δ⁡DefUq⁡(A). (2.12)

Here Tot is the limit of the cosimplicial diagram in 𝒮, and the local functors use the same marked sheaf-algebra construction. The limit includes overlap equivalences and all their compatibility homotopies. Mapping spaces and homotopy fibres preserve this descent, so the same procedure applies to the closed-form problems in Section 3.

Applying the tangent construction to all shifted square-zero tests linearizes this limit. On a smooth affine U, the local tangent object is Γ⁢(U,TU)⁢[1], and its descent limit is R⁢Γ⁢(X,TX)⁢[1]. More generally, limits in a stable ∞-category preserve fibre sequences and hence cofibre sequences, so local tangent triangles and natural transformations globalize by taking the same descent limit. We use this procedure for the pair tangent object and the derivative of the period map in Section 4.

3. Deformations of closed forms and crystalline periods

Fix a smooth proper k-scheme X of dimension d, an integer r≥1, and a closed form α∈H0⁢(X,ΩXr). We construct the deformation problem ℱα of marked pairs carrying closed lifts of α, the period problem 𝒫αr and its classical hull, and the crystalline period map ℱα→𝒫αr on divided-power bases. The constructions require only smooth properness and closedness; further hypotheses are stated in the lifting results.

3.1. Relative de Rham complexes and closed forms

For A∈ArtWan and Y∈DefX⁡(A), let dRY/Ader denote the Hodge-filtered relative derived de Rham complex, with derived p-completion understood. We use Illusie’s construction [17, VIII, §2.1] in the polynomial-resolution form of [5, Definition 2.1 and Remark 2.2]: locally, we realize the de Rham complexes of a simplicial polynomial resolution, filtered by form degree Fa⁢Ω∙=Ω≥a. We write

R⁢ΓdR⁢(Y/A) =R⁢Γ⁢(Y,dRY/Ader),
Fa⁢R⁢ΓdR⁢(Y/A) =R⁢Γ⁢(Y,Fa⁢dRY/Ader).

Derived p-completion is taken at each filtration step [5, Definition 8.1] and is automatic here because p is nilpotent in π0⁢A. Hodge completion is a different operation [5, Remark 2.6]; it actually has no effect by the finite filtration proved below.

For the animated divided-power formulation and its compatibility with derived de Rham cohomology for animated rings, see [23, Definition 4.31 and Lemma 4.32]. On a divided-power base A, the induced structure on 𝒪Y gives the same filtered complex [8, Corollary 5.12].

Lemma 3.1.

The filtration on R⁢ΓdR⁢(Y/A) is finite, with F0⁢R⁢ΓdR⁢(Y/A)=R⁢ΓdR⁢(Y/A) and Fa⁢R⁢ΓdR⁢(Y/A)=0 for a>d. Its graded pieces, for a≥0, are

grFa⁡R⁢ΓdR⁢(Y/A)≃R⁢Γ⁢(Y,Λa⁢LY/A)⁢[−a]. (3.1)

Here Λa denotes the derived exterior power of the cotangent complex. Every Fa⁢R⁢ΓdR⁢(Y/A) is a perfect A-module, and for every A→B in ArtWan, with YB=Y×Spec⁡ASpec⁡B the derived base change, there is a natural equivalence

B⊗A𝐋Fa⁢R⁢ΓdR⁢(Y/A)≃Fa⁢R⁢ΓdR⁢(YB/B). (3.2)

For discrete A, the filtered complex is R⁢Γ⁢(Y,ΩY/A∙) with its Hodge filtration Fa⁢R⁢ΓdR⁢(Y/A)=R⁢Γ⁢(Y,ΩY/A≥a).

Proof.

The associated graded formula for the Hodge filtration is part of Illusie’s construction; see [2, §1.2] and, for the animated formulation, [8, Propositions 5.10 and 5.15]. The deformation Y/A is smooth and proper in the derived sense, and LY/A is finite locally free of rank at most d. The polynomial-resolution construction commutes with derived base change, including each Hodge step; see [5, Proposition 2.7] and [23, Corollary 4.38]. Together with the smooth comparison modulo powers of p [5, Corollary 3.10], this identifies, on each affine open, the reduction of Fa⁢dRY/Ader with Fa⁢ΩX∙=0 for every a>d. Tensoring with k is conservative on all A-modules: the finite Postnikov tower of A and composition series of its homotopy groups build A from finitely many shifts and extensions of k as an A-module. Thus M⊗A𝐋k=0 implies M≃M⊗A𝐋A=0. Applying this to each Fa⁢dRY/Ader, a>d, proves finiteness of the filtration before taking global sections.

The exterior powers of LY/A are perfect, and their derived global sections are perfect and commute with derived base change by proper smooth pushforward. Applying this to (3.1) and using the finite filtration proves perfectness and (3.2). The identification for discrete A is again [5, Corollary 3.10]; its applicability uses the nilpotence of p on A. ∎

Definition 3.2.

The space of closed relative r-forms of degree zero on Y/A is

𝒜r,closed⁢(Y/A) =MapModA⁡(A⁢[−r],Fr⁢R⁢ΓdR⁢(Y/A)) (3.3)
≃Ω∞⁢(Fr⁢R⁢ΓdR⁢(Y/A)⁢[r]).

The degree-zero argument is suppressed in the notation. This space is functorial in Y/A, including changes of animated base. Projection to grFr⁡R⁢ΓdR⁢(Y/A) gives its underlying-form map

𝒜r,closed⁢(Y/A)⟶Ω∞⁢R⁢Γ⁢(Y,Λr⁢LY/A).

On animated bases, a point retains the homotopies expressing closedness, rather than only an underlying differential form.

Lemma 3.3.

If A is discrete, then 𝒜r,closed⁢(Y/A) is discrete and its points are precisely

ker⁡(H0⁢(Y,ΩY/Ar)⟶𝑑H0⁢(Y,ΩY/Ar+1)).
Proof.

The complex Fr⁢R⁢ΓdR⁢(Y/A) has no cohomology below degree r, and its degree-r cohomology is the displayed kernel; apply (2.1). ∎

3.2. The deformation problem of the pair

For Y∈DefX⁡(A), its marking gives a reduction map 𝒜r,closed⁢(Y/A)→𝒜r,closed⁢(X/k). Define ℱα⁢(A) as the total space over DefX⁡(A) whose fibre at Y is

fibY⁡(ℱα⁢(A)→DefX⁡(A))=fibα⁡(𝒜r,closed⁢(Y/A)⟶𝒜r,closed⁢(X/k)). (3.4)

Thus a point consists of a marked deformation Y, a closed form αA, and a path identifying its reduction with α. Marked equivalences of Y transport the forms and their reduction paths. The total space is formed by the Grothendieck construction [21, §2.2], which defines ℱα and its forgetful map to DefX functorially.

Proposition 3.4.

The functor ℱα is a formal moduli problem satisfying Zariski homotopy descent. On discrete bases it is the groupoid of marked pairs (Y,αA), and

ℱαcl=Def(X,α)/W. (3.5)

Its tangent object is perfect, and forgetting the form gives a natural equivalence

fib⁡(𝕋ℱα→𝕋DefX)≃Fr⁢R⁢ΓdR⁢(X/k)⁢[r]. (3.6)

In particular, ℱαcl is a deformation functor with finite-dimensional tangent space.

Proof.

The value at k is contractible. Consider a pullback B=B1×B0B2 as in Definition 2.1. By the formal moduli property of DefX, compatible geometric deformations glue to Y/B. The coefficient fibre sequence

B⟶B1⊕B2⟶B0

remains a fibre sequence after tensoring with Fr⁢R⁢ΓdR⁢(Y/B). By (3.2), the tensor product Fr⁢R⁢ΓdR⁢(Y/B)⊗B𝐋Bi is identified with the corresponding filtered de Rham complex of YBi/Bi. Applying the shift [r], then Ω∞, and taking the fibre at α proves the formal moduli axiom for (3.4), and hence for the total spaces. Sheaf descent for the filtered complexes, combined with (2.12), gives the same assertion on a Zariski cover.

On discrete bases, Lemma 3.3 gives (3.5), and [8, Lemma 3.7] supplies the classical deformation-functor axioms.

For finite-dimensional V and n≥0, put A=k⊕V⁢[n] and let XA be the constant deformation. Its constant form α gives a basepoint, and (3.2) identifies its marked form fibre with

fibXA⁡(ℱα⁢(A)→DefX⁡(A))≃Ω∞⁢(Fr⁢R⁢ΓdR⁢(X/k)⁢[r]⊗kV⁢[n]).

This is natural on all shifted tests, so it gives (3.6). The fibre is perfect by Lemma 3.1, and 𝕋DefX≃R⁢Γ⁢(X,TX)⁢[1] is perfect as X is smooth proper, so 𝕋ℱα is perfect. ∎

Remark 3.5.

Once the geometric deformation has been glued, gluing its forms uses exactness in the stable module ∞-category, as in the proof above; this argument does not require individual relative Hodge cohomology groups to be flat. If (1.3) holds at the special fibre, derived Nakayama shows that every infinitesimal lift also satisfies it.

3.3. The deformation problem of periods

Recall that crystalline cohomology is perfect and its derived reduction is de Rham cohomology:

R⁢Γcrys⁢(X/W)∈Perf⁡(W),R⁢Γcrys⁢(X/W)⊗W𝐋k≃R⁢ΓdR⁢(X/k).

The form α defines a point cα∈Ω∞⁢R⁢ΓdR⁢(X/k)⁢[r] by the composite k⁢[−r]⟶𝛼Fr⁢R⁢ΓdR⁢(X/k)→R⁢ΓdR⁢(X/k). For any A∈ArtWan, set

ℰr⁢(A) =MapModA⁡(A⁢[−r],R⁢Γcrys⁢(X/W)⊗W𝐋A)
≃Ω∞⁢((R⁢Γcrys⁢(X/W)⊗W𝐋A)⁢[r]),

and define the deformation problem of periods by

𝒫αr⁢(A)=ℰr⁢(A)×ℰr⁢(k){cα}. (3.7)

A point consists of a period A⁢[−r]→(R⁢Γcrys⁢(X/W)⊗W𝐋A) and a path identifying its reduction with cα. This is defined on all animated Artinian bases.

Proposition 3.6.

The functor 𝒫αr is a formal moduli problem, with

𝕋𝒫αr≃R⁢ΓdR⁢(X/k)⁢[r]. (3.8)

For finite-dimensional V and n≥0, the split lift of cα gives a natural equivalence

𝒫αr⁢(k⊕V⁢[n])≃Ω∞⁢(R⁢ΓdR⁢(X/k)⁢[r]⊗kV⁢[n]).

Its ordinary tangent and complete obstruction spaces are HdRr⁢(X/k) and HdRr+1⁢(X/k), respectively.

Proof.

The underlying W-modules of an infinitesimal pullback of animated algebras form a homotopy pullback in ModW: surjectivity on π0 makes the homotopy pullback connective. Tensoring with R⁢Γcrys⁢(X/W), shifting by r, applying Ω∞, and taking the fibre at cα preserve this limit. Since 𝒫αr⁢(k)≃{cα} is contractible, the functor 𝒫αr is thus a formal moduli problem.

On a split test, the augmentation (R⁢Γcrys⁢(X/W)⊗W𝐋(k⊕V⁢[n]))→R⁢ΓdR⁢(X/k) is split with fibre R⁢ΓdR⁢(X/k)⊗kV⁢[n]. Translation by the split lift of cα gives the stated equivalence, and hence (3.8). ∎

The reduction path remains part of a point of 𝒫αr⁢(A) even for discrete A, and this period space can have higher homotopies. It therefore records more than a set-theoretic fibre of degree-r cohomology classes.

3.4. Hulls for the deformation problem of periods

Let C∈Perf⁡(W) and c∈Ω∞⁢((C⊗W𝐋k)⁢[r]). We set

Fc⁢(A)=π0⁢fibc⁡(Ω∞⁢((C⊗W𝐋A)⁢[r])⟶Ω∞⁢((C⊗W𝐋k)⁢[r])) (3.9)

for any A∈ArtW.

For any C∈Perf⁡(W), the same argument as in Proposition 3.6 shows that the space-valued fibre defines a formal moduli problem; hence its classical truncation Fc is a deformation functor. When C=R⁢Γcrys⁢(X/W) and c=cα, this recovers Fc=(𝒫αr)cl from (3.7). Unobstructedness of Fc means that Fc⁢(A′)→Fc⁢(A) is surjective for every surjection A′↠A in ArtW.

Lemma 3.7.

For a perfect W-complex C and a discrete W-module A, there is a natural exact sequence

0⟶Hr⁢(C)⊗WA⟶Hr⁢(C⊗W𝐋A)⟶Tor1W⁡(Hr+1⁢(C),A)⟶0. (3.10)

If Hr+1⁢(C) is torsion-free, then Hr⁢(C⊗W𝐋A)=Hr⁢(C)⊗WA naturally, and every surjection A′↠A of discrete W-modules induces a surjection Hr⁢(C⊗W𝐋A′)↠Hr⁢(C⊗W𝐋A).

Proof.

Since the DVR W has global dimension one, the hyper-Tor spectral sequence E2−a,b=ToraW⁡(Hb⁢(C),A)⇒Hb−a⁢(C⊗W𝐋A) has only the columns a=0,1; its degree-r edge sequence is (3.10), also given by [29, Tag 0H80]. If Hr+1⁢(C) is torsion-free, it is finite free, so the Tor term vanishes. Right exactness of tensor product then gives the asserted surjectivity, without any freeness assumption on Hr⁢(C). ∎

Lemma 3.8.

Let A′↠A in ArtW. A point ξ∈Fc⁢(A) lifts to Fc⁢(A′) if and only if its underlying class in Hr⁢(C⊗W𝐋A) lifts to Hr⁢(C⊗W𝐋A′).

Proof.

Necessity follows by forgetting the marking. Conversely, a class lift gives a period e′ over A′ and a path η:e′|A≃e for a representative (e,γ:e|k≃c) of ξ. The path γ∘(η|k) marks e′ at c, and η identifies its image with ξ. ∎

For a complete Noetherian local W-algebra R with residue field k, write hR⁢(A)=HomW,augcont⁡(R,A). A hull is a formally smooth natural transformation hR→Fc inducing an isomorphism on tangent spaces [29, Tag 06IY].

A minimal representative of C is a bounded complex P of finite free W-modules, quasi-isomorphic to C, such that di⁢(Pi)⊆p⁢Pi+1 for every i; equivalently, P⊗Wk has zero differential. Such a representative exists by [29, Tag 0BCC]: the complex constructed there has rank dimkHi⁢(C⊗W𝐋k) in degree i, so it is bounded because C is perfect. Concretely, starting with a bounded finite free representative, one successively cancels a contractible summand [W⟶1W] whenever a differential has a unit entry; see [29, Tag 00MT]. Each cancellation lowers the total rank, so the process terminates with all differential entries in p⁢W.

Proposition 3.9.

Choose a minimal bounded finite free representative P≃C and Smith bases in which the nonzero entries of dr are pe1,…,pes, with ej≥1. Then

Hr+1⁢(C)tors≃⨁j=1sW/pej,N:=rankW⁡Pr=dimkHr⁢(C⊗W𝐋k).

Transport c to this model and choose a lift c˜∈Pr of its vector in Pr⊗Wk. Number the first s coordinates by the nonzero Smith entries. A hull of Fc is

Rc=W[[x1,…,xN]]/(pej(xj+c˜j):1≤j≤s). (3.11)

The hull map sends x∈𝔪AN satisfying these equations to the cycle c˜+x with its constant reduction path to c.

Proof.

Minimality gives zero differentials on P⊗Wk, and hence the formula for N. We first explain why the Smith form of dr splits off two-term summands of the entire complex. For a nonzero diagonal entry pej, the identity dr⁢dr−1=0 forces the corresponding row of dr−1 to vanish, while dr+1⁢dr=0 forces the corresponding column of dr+1 to vanish. Here cancellation of pej is valid because the terms are free over the domain W. Thus each block [W⟶pejW] in degrees r,r+1 is a direct summand of P and contributes W/pej to Hr+1⁢(C). In the complementary complex the differential in degree r is zero, so its degree-(r+1) cohomology is a submodule of a free module and is torsion-free. This proves the asserted description of Hr+1⁢(C)tors.

For a discrete A, the marked classes have the description

Fc⁢(A)≃{z∈Pr⊗WA:dr⁢z=0,z̅=c}dr−1⁢(Pr−1⊗W𝔪A). (3.12)

Indeed, a reduction path is represented in degree r−1 over k. Lifting it to Pr−1⊗WA and changing the cycle by its boundary makes the marking constant. Since P⊗Wk has zero differential, higher homotopies do not change the vector representing this path. Changes between constant markings therefore have coefficients in 𝔪A, giving precisely the displayed denominator. Writing z=c˜+x, the cycle equations are pej⁢(xj+c˜j)=0. Thus hRc represents the numerator of (3.12).

For formal smoothness, let A′↠A, fix such a cycle z over A, and prescribe a class in Fc⁢(A′) reducing to its image. Choose a representative z′ of this class. Its reduction differs from z by dr−1⁢b for some b∈Pr−1⊗W𝔪A. Lift b to b′∈Pr−1⊗W𝔪A′. Then z′−dr−1⁢b′ represents the same marked class and lifts z exactly. On k⁢[ϵ]/(ϵ2) every differential of P vanishes, so the hull map induces the tangent isomorphism kN≃Hr⁢(C⊗W𝐋k). ∎

Torsion in Hr⁢(C) arises from incoming summands [W⟶pfW] in degrees r−1,r. Each such summand contributes a free hull parameter whose values in Fc⁢(A) are identified modulo pf⁢𝔪A, and these boundary identifications can make the hull map noninjective.

Corollary 3.10.

The deformation functor Fc is unobstructed if and only if Hr+1⁢(C) is torsion-free.

Proof.

A deformation functor admitting a hull is unobstructed exactly when that hull is formally smooth over W, equivalently a power-series ring over W [29, Tag 0DZK]. If s=0, (3.11) is such a ring. If s>0, choose 1≤j≤s. The element xj+c˜j is nonzero, since its image in Rc/p⁢Rc=k⁢[[x1,…,xN]] is nonzero, but it is killed by pej, which is impossible in a power-series ring over W. Thus unobstructedness is equivalent to s=0, which is precisely the asserted torsion-freeness. ∎

Corollary 3.11.

The classical truncation 𝒫αrcl of the deformation problem of periods is unobstructed on all discrete Artinian W-algebras if and only if Hcrysr+1⁢(X/W) is torsion-free.

Proof.

Apply Corollary 3.10 with C=R⁢Γcrys⁢(X/W) and c=cα. ∎

Remark 3.12.

Under the torsion-free hypothesis of Corollary 3.11, every actual obstruction to lifting a period across a discrete elementary extension vanishes. The complete obstruction space HdRr+1⁢(X/k) of Proposition 3.6 can nevertheless be nonzero, since classical unobstructedness does not assert that 𝕋𝒫αr is concentrated in degree zero. The necessity of torsion-freeness for the deformation problem of periods does not imply its necessity for unobstructedness of the geometric deformation functor.

3.5. Crystalline comparison and the period map

Give W its canonical divided powers on p⁢W. For divided-power bases, the crystalline property identifies the total relative complex with the fixed coefficient complex (R⁢Γcrys⁢(X/W)⊗W𝐋A).

Lemma 3.13.

For A∈PDArtWan and Y∈DefX⁡(A), there are natural equivalences of unfiltered module objects

R⁢ΓdR⁢(Y/A)≃R⁢Γ⁢(X,dR𝒪X/Apd)≃R⁢Γcrys⁢(X/W)⊗W𝐋A. (3.13)

They are compatible with marked equivalences of deformations, divided-power base change, and Zariski descent.

Proof.

Since the induced divided-power algebra of 𝒪Y has divided-power reduction 𝒪X, [8, Corollary 5.12 and Theorem 5.22] give the first equivalence. For the second, form the pushout A⊗Wpd𝒪X in sheaves of animated divided-power algebras. Its divided-power reduction is 𝒪X [8, Proposition 4.11]. Filtered base change and unfiltered crystalline invariance [8, Proposition 5.11 and Theorem 5.22] therefore identify dR𝒪X/Apd with A⊗W𝐋dR𝒪X/Wpd. Global-sections base change [8, Lemma 6.12] and crystalline comparison [8, Proposition 5.26] give R⁢Γcrys⁢(X/W)⊗W𝐋A. All these constructions are functorial, which gives the stated compatibilities. ∎

Remark 3.14.

The equivalences (3.13) concern only the total complex; its Hodge filtration still varies with Y, as emphasized in [8, Warning 5.23]. The period map uses the moving filtered piece Fr⁢R⁢ΓdR⁢(Y/A) and its map to (R⁢Γcrys⁢(X/W)⊗W𝐋A).

Proposition 3.15.

Crystalline comparison defines a natural transformation of divided-power formal moduli problems

perα:ℱα|pd⟶𝒫αr|pd, (3.14)

with the following properties:

  1. (i)

    It sends a marked pair (Y,αA) over A∈PDArtWan to

    A⁢[−r]⟶αAFr⁢R⁢ΓdR⁢(Y/A)⟶R⁢Γcrys⁢(X/W)⊗W𝐋A, (3.15)

    with its induced reduction path to cα.

  2. (ii)

    It commutes with base change and Zariski homotopy descent; on discrete bases, its underlying cohomology class is the crystalline class of αA.

  3. (iii)

    For a finite-dimensional k-vector space V and n≥0, its restriction to the marked form fibre of the constant deformation over k⊕pdV⁢[n] is

    Ω∞⁢(Fr⁢R⁢ΓdR⁢(X/k)⁢[r]⊗kV⁢[n])⟶Ω∞⁢(R⁢ΓdR⁢(X/k)⁢[r]⊗kV⁢[n]), (3.16)

    induced by the Hodge inclusion.

Proof.
  1. (i)

    The marking of αA and the naturality of

    R⁢ΓdR⁢(Y/A)⟶∼R⁢Γcrys⁢(X/W)⊗W𝐋A

    identify the reduction of (3.15) with cα, so this composite is a point of 𝒫αr⁢(A). The filtered inclusion and crystalline comparison are natural transformations of sheaf-valued functors. Their composite therefore defines (3.14).

  2. (ii)

    Naturality gives compatibility with divided-power base change, and passage to the Zariski descent limit gives compatibility with Zariski homotopy descent. The smooth comparison identifies the underlying cohomology class with the usual crystalline class.

  3. (iii)

    Take the constant geometric deformation and translate the form and period fibres by the constant lift of α. Filtered base change identifies the resulting map with (3.16), induced by the Hodge inclusion.

∎

Formula (3.16) describes only the variation of the form on a constant deformation; Section 4 computes the full derivative, including the contribution from geometric deformations of underlying varieties.

Remark 3.16.

In general, (3.14) does not extend to a morphism ℱα→𝒫αr on all animated Artinian bases. Indeed, take X=E an elliptic curve, r=1, and α a nonzero invariant differential. For a discrete Artinian k-algebra A, translations of the constant pair by Eˆ⁢(A)=ker⁡(E⁢(A)→E⁢(k)) would map to the period automorphism group π1⁢(𝒫αr⁢(A),yA)≃HdR0⁢(E/k)⊗k𝔪A≃(𝔪A,+), where yA is its image under the proposed extension. Naturality would therefore give a homomorphism f:Eˆ→𝐆ˆa to the additive formal group. Compatibility on dual numbers with trivial higher divided powers and Theorem 4.6 force d⁢f0 to be an isomorphism, given up to sign by contraction with α. Thus f is invertible as a power series, whereas

f∘[p]Eˆ=[p]𝐆ˆa∘f=0

would force [p]Eˆ=0, contradicting finiteness of [p]:E→E.

The same obstruction occurs for the symplectic surface X=E1×E2, with r=2 and α=pr1∗⁡α1∧pr2∗⁡α2, where αi is a nonzero invariant differential on the elliptic curve Ei and pri is the projection to Ei. Translations in the first factor would give a homomorphism from Eˆ1 to the additive formal group associated with HdR1⁢(X/k), whose differential is, up to sign,

v⟼α1⁢(v)⁢[pr2∗⁡α2].

The class [pr2∗⁡α2] is nonzero by de Rham Künneth and the Hodge inclusion for E2. Composing with a k-linear functional nonzero on this class yields the same contradiction.

4. The derivative of the period map

Keep the notation of Section 3. We compute the derivative of perα from the tangent extension of the pair and the crystalline comparison homotopy.

4.1. The tangent extension and the Kaledin–Verbitsky complex

Work in the stable ∞-category Mod⁡(X,k) of sheaves of k-modules on X. Put

a:TX⟶Fr⁢ΩX∙⁢[r],a⁢(v)=ℒv⁢α.

Since α is closed, a is a k-linear map of sheaf complexes. Here TX is concentrated in degree zero. Write cofib⁡(a) for its cofibre. It is the middle row of the commutative diagram of complexes

−101Fr⁢ΩX∙⁢[r]0ΩXrΩXr+1⋯cofib⁡(a)TXΩXrΩXr+1⋯TX⁢[1]TX00⋯(−1)r⁢d(−1)r⁢dℒ(−)⁢α(−1)r⁢d(−1)r⁢d

whose columns are indexed by cohomological degree; here d denotes the unshifted de Rham differential. The vertical maps are the canonical maps of the cofibre sequence of a, and each column is split exact. As in [19, §6], we also use the complex

𝒦α∙=[TX⟶ℒ(−)⁢αΩXr⟶𝑑ΩXr+1⟶𝑑⋯],

with TX in degree zero, ΩXr in degree one, and the displayed positive differentials. We call it the Kaledin–Verbitsky complex of the closed form α, or KV complex for short.

Remark 4.1.

We use the cohomological shift convention of Section 2, so the differential of M⁢[s] is (−1)s⁢dM. The resulting chain isomorphism

I:𝒦α∙⁢[1]⟶∼cofib⁡(a)

is −1 on TX in degree −1 and (−1)(r+1)⁢n on ΩXr+n in degree n≥0. In particular, it is the identity on ΩXr in degree zero, as required by (3.16). Under I, the projection cofib⁡(a)→TX⁢[1] is the negative of the termwise projection from 𝒦α∙⁢[1].

Lemma 4.2.

There is a natural cofibre sequence in Perf⁡(k):

R⁢Γ⁢(X,Fr⁢ΩX∙)⁢[r]⟶𝕋ℱα⟶R⁢Γ⁢(X,TX)⁢[1]⟶−ℒ(−)⁢αR⁢Γ⁢(X,Fr⁢ΩX∙)⁢[r+1], (4.1)

and an identification

𝕋ℱα≃R⁢Γ⁢(X,𝒦α∙)⁢[1]. (4.2)
Proof.

By Proposition 3.4 and (2.11), the fibre and quotient of the tangent extension are R⁢Γ⁢(X,Fr⁢ΩX∙)⁢[r] and R⁢Γ⁢(X,TX)⁢[1]. To identify the connecting map, trivialize the geometric deformation on smooth affine charts. Over the dual numbers k⁢[ϵ]/(ϵ2), an infinitesimal automorphism 1+ϵ⁢v changes the constant form by

(1+ϵ⁢v)∗⁢α−α=ϵ⁢ℒv⁢α.

Thus the infinitesimal action is a. The deformation problem of a local pair is the homotopy quotient of the form fibre by changes of this trivialization. Linearizing this quotient gives the cofibre of the action map a. Rotating the cofibre sequence of a gives the connecting arrow −a⁢[1] in (4.1).

This action calculation is natural in the coefficient module. For a connective perfect module V, choose a simplicial vector space V∙ representing it. On a polynomial chart P, filtered base change identifies the relative complex in degree j with ΩP/k∙⊗k(k⊕Vj). The universal derivation gives the same coefficient-linear Lie derivative in every simplicial degree. Realizing these module maps computes the action with coefficients in V, including all shifted tests. Taking the stable descent limit of Section 2.7 proves (4.1). Changing the signs of the terms, with the convention made explicit in Remark 4.1, identifies cofib⁡(a) with 𝒦α∙⁢[1] and gives (4.2). ∎

4.2. The crystalline comparison homotopy

By Proposition 3.15, the derivative on the form fibre is induced by the filtered inclusion

j:Fr⁢ΩX∙⁢[r]⟶ΩX∙⁢[r].

We use j for the inclusion of the indicated Hodge step into the total de Rham complex. For every M in the stable ∞-category Mod⁡(X,k), the universal property of the cofibre gives a pullback square of spaces

MapMod⁡(X,k)⁡(cofib⁡(a),M)MapMod⁡(X,k)⁡(Fr⁢ΩX∙⁢[r],M){0}MapMod⁡(X,k)⁡(TX,M).a∗ (4.3)

Here a∗ is precomposition with a, and the bottom arrow selects the zero morphism. Extending j to the pair tangent object therefore requires a choice of nullhomotopy of j⁢a. For the period map, this homotopy comes from Lemma 3.13. Cartan’s formula gives

ℒv⁢α=(d⁢ιv+ιv⁢d)⁢α=d⁢(ιv⁢α).

The following calculation identifies contraction with the homotopy induced by that crystalline–de Rham comparison, including its naturality on shifted divided-power tests.

Lemma 4.3.

Let A=k⊕V, where V is a discrete k-vector space with trivial higher divided powers, and put

B=A⁢[x1,…,xe],Q=B⁢⟨t1,…,te⟩.

The vertex maps xi↦xi and xi↦xi+ti induce p0,p1:ΩB/A∙⇉ΩQ/A∙,pd. They admit a homotopy H satisfying

d⁢H+H⁢d=p1−p0,c⁢H=0.

Here c sets ti=d⁢ti=0. For every A-derivation δ:B→V⁢B, evaluation at ti=δ⁢(xi) sends H to the contraction ιδ, a homotopy from the identity to (1+δ)∗.

Proof.

Write R=A⁢⟨t1,…,te⟩ and put

Ci=[A⁢⟨ti⟩⟶diA⁢⟨ti⟩⁢d⁢ti],C=C1⊗A⋯⊗ACe=ΩR/A∙,pd,

with Ci in degrees zero and one. Since Q=B⊗AR, the decomposition of differentials into the d⁢xi and d⁢ti gives an identification of complexes

ΩQ/A∙,pd=ΩB/A∙⊗AC.

Here the differential on a tensor ω⊗η, with deg⁡(ω)=a, is dB⁢ω⊗η+(−1)a⁢ω⊗dC⁢η. Thus [29, Tag 07LC] applies in the t-direction with coefficient ring A and coefficient module ΩB/Aa in each x-degree a. We now make its contraction and the signs of the total differential explicit.

Let εi:Ci→A set ti=d⁢ti=0, let ui:A→Ci include the constants, and put Ei=ui⁢εi. The identity di⁢γm⁢(ti)=γm−1⁢(ti)⁢d⁢ti for m≥1 shows that the A-linear degree-minus-one map

hi⁢(γm⁢(ti)⁢d⁢ti)=γm+1⁢(ti),hi⁢(γm⁢(ti))=0

satisfies di⁢hi+hi⁢di=1−Ei and εi⁢hi=0. On C use the tensor homotopy

ht=∑i=1eE1⊗⋯⊗Ei−1⊗hi⊗1⊗(e−i),

with the Koszul sign convention for tensor products of graded maps. The identities for the hi telescope to dC⁢ht+ht⁢dC=1−E1⊗⋯⊗Ee. Moreover, (ε1⊗⋯⊗εe)⁢ht=0. Extend this to the total complex by

h⁢(ω⊗η)=(−1)a⁢ω⊗ht⁢(η),ω∈ΩB/Aa.

This sign cancels the two terms involving dB⁢ω in d⁢h+h⁢d. Since p0 includes the constants in C and c augments C, we obtain d⁢h+h⁢d=1−p0⁢c and c⁢h=0. Finally c⁢p1=1, so H=h⁢p1 satisfies d⁢H+H⁢d=p1−p0 and c⁢H=0.

Put J=V⁢B and g=1+δ, and equip Q with the PD ideal JQ generated by J and the ti. Evaluation ti↦δ⁢(xi), d⁢ti↦d⁢δ⁢(xi) is a map of PD de Rham complexes: it is induced by a PD map (Q,JQ)→(B,J), since the higher divided powers on J vanish. Give ΩB/A∙ weight zero and assign weights 1,1,m to ti,d⁢ti,γm⁢(ti), respectively. Both dC and ht preserve weight. After evaluation, every term of weight at least two vanishes: J2=0, the coefficients of d⁢δ⁢(xi) lie in J, and γm⁢(δ⁢(xi))=0 for m>1. The weight-zero part is killed by h, as are the weight-one terms containing a ti but no d⁢ti. For a weight-one term containing d⁢ti in the jth differential position, moving it past the a−j remaining differentials and applying h gives the sign (−1)a−j⁢(−1)a−1=(−1)j−1 and replaces d⁢ti by ti. Evaluation then gives ∑i∂fj/∂xi⁢δ⁢(xi)=δ⁢(fj). Thus the evaluated homotopy is contraction:

Hδ⁢(f0⁢d⁢f1∧⋯∧d⁢fa)=∑j=1a(−1)j−1⁢f0⁢δ⁢(fj)⁢d⁢f1∧⋯⁢d⁢fĵ⁢⋯∧d⁢fa. (4.4)

Evaluation sends p1 to g∗ and p0 to the identity, so

g∗−1=d⁢Hδ+Hδ⁢d=ℒδ,Hδ=ιδ. (4.5)

∎

Locally on X, let A=k⊕pdV for a finite-dimensional k-vector space V, and write αA for the constant lift of α. For an automorphism g of the constant deformation over A reducing to the identity, the reduction-compatible naturality of (3.13) gives a path from the identity to g∗ on the total de Rham complex. Evaluating this path at αA and translating by −αA gives a path from 0 to (g∗−1)⁢αA. Linearizing these paths in such automorphisms gives a path

0⟹j⁢ainMapMod⁡(X,k)⁡(TX,ΩX∙⁢[r]).

The following proposition identifies this homotopy with contraction.

Proposition 4.4.

The nullhomotopy of j⁢a:TX→ΩX∙⁢[r] in Mod⁡(X,k) constructed above is given, before shifting, by contraction

ια:TX⟶ΩXr−1,v⟼ιv⁢α, (4.6)

with d⁢(ιv⁢α)=ℒv⁢α. In the shifted target, this nullhomotopy is represented by the graded map h:TX→ΩX∙⁢[r] of degree −1, given by

h⁢(v)=(−1)r⁢ιv⁢α,(−1)r⁢d⁢h=j⁢a.

Here the sign comes from the differential (−1)r⁢d on ΩX∙⁢[r]. This identification is compatible with restriction and natural on k⊕pdV for every connective perfect k-module V.

Proof.

Work first on a polynomial chart, with A,B,Q,J,JQ as in Lemma 4.3. Filtered base change and the induced divided-power comparison [8, Proposition 5.11 and Corollary 5.12] identify the polynomial and PD-polynomial complexes in that lemma with

MB=dRB/Apd,MQ=dRQ/Apd.

Put P=B/J=Q/JQ=k⁢[x1,…,xe] and MP=dRP/Apd. The natural unfiltered equivalences of [8, Theorem 5.22] give

qB:MB⟶∼MP,qQ:MQ⟶∼MP,

with qQ⁢p0=qQ⁢p1=qB and qB⁢c=qQ, including their natural coherences. These are the reduction equivalences used in Lemma 3.13; [8, Proposition 5.26] identifies their globalization with Berthelot’s crystalline cohomology. The classical computation by PD diagonal envelopes is given in [29, Tag 07LG and Tag 07LH].

Here c:MQ→MB is an equivalence, either by the explicit contraction of Lemma 4.3 or by the displayed reduction equivalences. Consequently

c∗:MapModA⁡(MB,MQ)⟶∼MapModA⁡(MB,MB)

has contractible homotopy fibre over 1MB. The vertices p0,p1 belong to this fibre with their constant augmentation paths, and c⁢H=0 says that H is a path between them in this fibre. Postcomposition with qB identifies it with the contractible fibre

fibqB⁡(MapModA⁡(MB,MQ)⟶(qQ)∗MapModA⁡(MB,MP)).

The comparison path is the path in this latter fibre obtained from the natural reduction maps: both vertices reduce to 1P, and their paths to qB are those supplied by the same augmentation diagram. Thus its specified augmentation normalization agrees with that of H, and contractibility identifies the two paths together with their coherences. For an A-derivation δ:B→J, put g=1+δ and let eδ:Q→B fix B and send ti to δ⁢(xi). It satisfies qB⁢eδ=qQ with these same reduction coherences. It therefore takes this identification to the comparison path from 1 to (1+δ)∗ on MB, whose homotopy is ιδ by (4.4). Here the substitution ti↦δ⁢(xi) and d⁢ti↦d⁢δ⁢(xi) composes the universal homotopy H with the map of de Rham complexes defined by eδ, giving Hδ. For a closed r-form β on Spec⁡P, let βA be its pullback to Spec⁡B. Applying this comparison homotopy to βA means precomposing its path with βA:A⁢[−r]→MB. Its chain representative is Hδ⁢(βA)=ιδ⁢βA; closedness gives d⁢(Hδ⁢(βA))=(g∗−1)⁢(βA).

This uniqueness also holds for parameter diagrams. Fix the polynomial algebra P, and let I index a diagram of discrete coefficient spaces V and derivations δ:P→P⊗kV, with coefficient maps carrying each derivation to the next. All the preceding arrows, their reductions, and their evaluations form diagrams of k-modules, and the formula for H is natural in these coefficient maps. Since c and qB are objectwise equivalences, they are equivalences in Fun⁡(I,Modk). The same mapping-fibre argument in this ∞-category, with the diagram MB as source, identifies H with the comparison path as a path of natural transformations. In particular, it supplies all compatibilities in the coefficient and derivation parameters.

For a connective perfect k-module V, choose a simplicial vector space V∙ representing it, with each Vj finite-dimensional. The full derived derivation space on P is

MapModP⁡(ΩP/k1,P⊗kV).

Since ΩP/k1 is finite free, this space is computed by the simplicial spaces of degreewise derivations into P⊗kVj. Apply the preceding argument to these parameter diagrams, with Aj=sqzkpd⁡(Vj), Bj=Aj⁢[x1,…,xe], and Qj=Bj⁢⟨t1,…,te⟩. The functors sqzkpd and dRpd preserve geometric realizations [8, Definition 4.32, Corollary 4.7, and Definition 5.6]. Realizing the diagrams of divided-power arrows and their comparison maps therefore gives the corresponding diagram over A=sqzkpd⁡(V), together with its augmentation normalization. In every simplicial degree,

Ω(k⊕Vj)⁢[x]/(k⊕Vj)∙=Ωk⁢[x]/k∙⊗k(k⊕Vj).

The evaluated homotopy ιδ is linear in the coefficients Vj, so its realization is contraction with coefficients in V. The diagramwise identification above makes it the comparison homotopy for every such V, including every shifted test.

Finally, cover X by smooth affine opens admitting étale coordinates. The evaluated contraction is intrinsic and commutes with localization and étale pullback, so these formulas agree on coordinate refinements. Since the crystalline comparison and its normalized identification are natural on the same refinements, they define the same morphism of sheaf tangent complexes, which descends to X by Section 2.7. Evaluation at the closed form α gives ℒv⁢α=d⁢(ιv⁢α) and (4.6). ∎

Remark 4.5.

The animated divided-power comparison supplies the higher coherences, and the polynomial calculation identifies its derivative. The passage from divided-power elementary extensions to arbitrary Artinian extensions uses the bootstrapping argument of Section 5.

4.3. The tangent morphism

Combining the Hodge inclusion with Proposition 4.4 gives the full derivative.

Theorem 4.6.

With the sign convention of Remark 4.1, the tangent morphism of the period map perα:ℱα|pd→𝒫αr|pd has the following descriptions.

  1. (i)

    Under (4.2) and (3.8), it is R⁢Γ⁢(X,fα)⁢[1], where

    fα:𝒦α∙⟶ΩX∙⁢[r−1] (4.7)

    is defined in nonnegative cohomological degrees as

    TXΩXrΩXr+1⋯ΩXr−1ΩXrΩXr+1⋯ℒ(−)⁢α(−1)r−1⁢ιαdd(−1)r−1⁢id(−1)r−1⁢d(−1)r−1⁢d(−1)r−1⁢d

    Here d is the ordinary differential on de Rham complex, and the columns are indexed by cohomological degree. The vertical map in degree q≥1 is (−1)(r−1)⁢(q−1)⁢id; all components in negative degrees are zero.

  2. (ii)

    If contraction ια:TX→ΩXr−1 is an isomorphism, as in (1.3), this morphism identifies with the filtered inclusion

    𝕋ℱα≃R⁢Γ⁢(X,Fr−1⁢ΩX∙)⁢[r]⟶𝑗R⁢Γ⁢(X,ΩX∙)⁢[r]≃𝕋𝒫αr. (4.8)

    For every finite-dimensional k-vector space V and n≥0, the period map on k⊕pdV⁢[n] is then

    Ω∞⁢(R⁢Γ⁢(X,Fr−1⁢ΩX∙)⁢[r]⊗kV⁢[n])⟶Ω∞⁢(R⁢ΓdR⁢(X/k)⁢[r]⊗kV⁢[n]).
Proof.

By Lemma 4.2, the source tangent object is R⁢Γ⁢(X,cofib⁡(a)). The derivative restricts to j on the form fibre by (3.16), and Proposition 4.4 identifies its comparison nullhomotopy with h. The cofibre property (4.3) therefore gives a chain map

Φ:cofib⁡(a)⟶ΩX∙⁢[r]

with component h=(−1)r⁢ια:TX→ΩXr−1 in degree −1 and the identity on ΩXr+n in every degree n≥0. Transporting along I and shifting by −1 gives fα=(Φ∘I)⁢[−1], with the vertical maps shown in the diagram. Naturality in connective perfect coefficients and descent, proved in Proposition 4.4, identify R⁢Γ⁢(X,Φ) with the tangent morphism of the period map on every shifted divided-power test.

If contraction is an isomorphism, the same components define a chain isomorphism ψ:cofib⁡(a)⟶∼Fr−1⁢ΩX∙⁢[r] with Φ=j⁢ψ. Taking derived global sections gives (4.8). Under this identification, the tangent triangle (4.1) is the filtered triangle

R⁢Γ⁢(X,Fr⁢ΩX∙)⁢[r]⟶R⁢Γ⁢(X,Fr−1⁢ΩX∙)⁢[r]⟶R⁢Γ⁢(X,ΩXr−1)⁢[1].

Here the geometric quotient R⁢Γ⁢(X,TX)⁢[1] is identified with the last term by (−1)r⁢ια. Finally, tensoring the tangent morphism with V⁢[n] and applying Ω∞ gives the formula on divided-power square-zero tests. ∎

4.4. Tangent and obstruction spaces

With our conventions, degrees −1, 0, and 1 of a tangent object describe infinitesimal automorphisms, tangent vectors, and complete obstructions, respectively.

Corollary 4.7.

The period map perα:ℱα|pd→𝒫αr|pd induces the maps

ℍ1⁢(X,𝒦α∙) ⟶HdRr⁢(X/k) on tangent spaces,
ℍ2⁢(X,𝒦α∙) ⟶HdRr+1⁢(X/k) on complete obstruction spaces.

Both are induced by fα. If (1.3) holds, the infinitesimal automorphism, tangent, and complete obstruction spaces of the pair are, respectively,

H−1⁢(𝕋ℱα) =ℍr−1⁢(X,Fr−1⁢ΩX∙),
H0⁢(𝕋ℱα) =ℍr⁢(X,Fr−1⁢ΩX∙), (4.9)
H1⁢(𝕋ℱα) =ℍr+1⁢(X,Fr−1⁢ΩX∙),

and their maps to the corresponding spaces of 𝒫αr are induced by the filtered inclusion Fr−1⁢ΩX∙→ΩX∙.

Proof.

Restriction to divided-power bases preserves the tangent objects, because k⊕pdV⁢[n] has underlying animated algebra k⊕V⁢[n]. Taking H0 and H1 of the morphism in Theorem 4.6 therefore gives the two displayed maps. Under (1.3), taking H−1, H0, and H1 in (4.8) gives (4.9) and the maps induced by the filtered inclusion. ∎

For a divided-power elementary extension A′↠A with kernel I satisfying (2.8), the classifying square of Section 2 places each pair obstruction in ℱα⁢(k⊕pdI⁢[1]). The case n=1 therefore identifies the map on actual obstruction classes with that of Corollary 4.7 tensored with I; its crystalline boundary description is Proposition 5.1. Complete obstruction spaces may remain nonzero when every actual class vanishes (Remark 3.12).

5. Proof of the main criterion

The proof of Theorem A combines injectivity on complete obstruction spaces, unobstructedness of periods, and Brantner–Taelman bootstrapping. At p=2, this requires the initial tuple lift constructed in Section 5.4.

5.1. Naturality of complete obstructions

Let A′↠A be a divided-power elementary extension with kernel I, and let x=(Y,αA)∈ℱα⁢(A). By Section 2, both restrictions in (3.14) send its classifying square to homotopy pullbacks, and naturality gives a homotopy-commutative diagram

ℱα⁢(A)𝒫αr⁢(A)ℱα⁢(k⊕pdI⁢[1])𝒫αr⁢(k⊕pdI⁢[1]).perαobob (5.1)

Its lower map is the derivative on the shifted square-zero test with coefficient module I⁢[1]. On connected components, it is the complete obstruction map of Corollary 4.7 tensored with I.

On the target, the coefficient extension supplies the distinguished triangle

R⁢ΓdR⁢(X/k)⊗kI⟶R⁢Γcrys⁢(X/W)⊗W𝐋A′⟶R⁢Γcrys⁢(X/W)⊗W𝐋A⟶+1. (5.2)

After shifting by r and applying the mapping-space construction, its connecting map is the obstruction map for the deformation problem of periods. We denote the induced boundary map on cohomology by

∂A′/A: Hr⁢(R⁢Γcrys⁢(X/W)⊗W𝐋A)
⟶HdRr+1⁢(X/k)⊗kI.
Proposition 5.1.

Under (1.3), the complete pair obstruction satisfies

j⁢(obℱα⁡(x;A′/A))=∂A′/A[αA]crysin ⁢HdRr+1⁢(X/k)⊗kI. (5.3)

Here [αA]crys∈Hr⁢(R⁢Γcrys⁢(X/W)⊗W𝐋A) is the underlying crystalline period class, and the pair obstruction belongs to ℍr+1⁢(X,Fr−1⁢ΩX∙)⊗kI.

Proof.

Take connected components of (5.1), use Corollary 4.7 for the lower map, and use (5.2) for the target obstruction. Completeness of the source obstruction follows from its lift-space description (2.6), since ℱα is a formal moduli problem. ∎

Remark 5.2.

Identity (5.3) is induced by a morphism of lift-space homotopy pullbacks, so it is compatible with automorphisms and changes of representatives.

Remark 5.3.

Under (1.3), classical Čech–de Rham descent also gives a complete obstruction in ℍr+1⁢(X,Fr−1⁢ΩX∙)⊗kI; we neither use this construction nor identify its class with obℱα⁡(x;A′/A).

5.2. Degreewise Hodge conditions

The following lemma translates the numerical equality (HDn) into the filtered injectivity and surjectivity used below.

Lemma 5.4.

Let (D,F) be a bounded complex over a field with a finite filtration and finite-dimensional cohomology in every graded piece. Suppose

dimHn⁢(D)=∑adimHn⁢(grFa⁡D).

For every a, the maps Hn⁢(Fa⁢D)→Hn⁢(D) are injective, and both

Hn⁢(Fa⁢D)⟶Hn⁢(grFa⁡D),Hn−1⁢(Fa⁢D)⟶Hn−1⁢(grFa⁡D)

are surjective.

Proof.

Choose a bounded finite-dimensional, degreewise split filtered model P whose graded complexes have zero differential. Such a model is obtained by representing the successive filtration extensions by cofibres of chain maps and cancelling the contractible pairs in their diagonal graded blocks; these cancellations preserve the filtration. Thus

dimPn=∑adimHn⁢(grFa⁡D).

For a finite-dimensional complex,

dimPn−dimHn⁢(P)=rank⁡dn−1+rank⁡dn.

The hypothesis therefore forces dn−1=dn=0. In degree n, the asserted maps are inclusions and quotients of the split filtered vector space Pn. In degree n−1, every representative of a class in grFa⁡Pn−1 lifts to an element of Fa⁢Pn−1, which is a cycle since dn−1=0. This proves the remaining surjectivity. ∎

5.3. From divided-power bases to all Artinian bases

We use the following bootstrapping theorem of Brantner–Taelman [8, Proposition 7.8].

Theorem 5.5.

Assume that k is infinite. Let G:ArtW→Set be a deformation functor with finite-dimensional tangent space. Suppose G⁢(A′)→G⁢(A) is surjective for every divided-power elementary extension. If p=2, suppose also that G⁢(W2⁢(k))≠∅. Then G is unobstructed.

Its mixed-characteristic truncated divided-power test rings detect equations in a hull, whereas testing only divided-power k-algebras would not suffice.

5.4. The initial lift in characteristic two

To supply the initial tuple point required by Theorem 5.5 at p=2, we first lift forms on a fixed W2⁢(k)-lift by a filtered Bockstein argument, and then construct such a lift using the conjugate spectral sequence. The first step works for every p.

Lemma 5.6.

Let Y/W2⁢(k) be a smooth proper lift of X. If (HDr+1) holds and Hcrysr+1⁢(X/W) is torsion-free, then every α∈H0⁢(X,ΩXr) is closed and has a closed relative lift on Y.

Proof.

Put D2=R⁢Γ⁢(Y,ΩY/W2∙) and D1=R⁢Γ⁢(X,ΩX∙), with their stupid filtrations. Smooth proper filtered base change gives Fr⁢D2⊗W2𝐋k≃Fr⁢D1. Tensoring the coefficient sequence 0→k⟶𝑝W2→k→0 over W2 with Fr⁢D2 therefore gives a distinguished triangle

Fr⁢D1⟶Fr⁢D2⟶Fr⁢D1⟶+1.

The filtered inclusions give a morphism from this triangle to the corresponding triangle for D2, and hence a natural square of Bocksteins

Hr⁢(Fr⁢D1)Hr+1⁢(Fr⁢D1)Hr⁢(D1)Hr+1⁢(D1).βFjβ

Crystalline–de Rham comparison and derived crystalline base change identify Di≃R⁢Γcrys⁢(X/W)⊗W𝐋Wi⁢(k) for i=1,2, where W1⁢(k)=k, compatibly with reduction [29, Tag 07MW, Tag 07MU]. These identifications are unfiltered: under them, the total coefficient triangle is R⁢Γcrys⁢(X/W)⊗W𝐋(k→W2→k), whereas the filtered triangle uses the filtration of the specified lift Y and the filtered base change above. By universal coefficients and torsion-freeness of Hcrysr+1⁢(X/W), the map Hr⁢(D2)→Hr⁢(D1) is surjective, so β=0. By Lemma 5.4 in degree r+1, the map j is injective. The same numerical equality kills d1r,0, so α is closed and belongs to Hr⁢(Fr⁢D1). Naturality yields j⁢βF⁢(α)=β⁢[α]=0, hence βF⁢(α)=0. The filtered coefficient sequence lifts α to Hr⁢(Fr⁢D2), which is the kernel of d:H0⁢(Y,ΩY/W2r)→H0⁢(Y,ΩY/W2r+1). ∎

Proposition 5.7.

Suppose (1.1) is injective and (HDrj+1) holds for every j. Then X admits a smooth proper lift to W2⁢(k).

Proof.

Put X′=X×k,Fkk and write αj′ for the base change of αj. The conjugate spectral sequence is

E2a,b=Ha⁢(X′,ΩX′b)⟹HdRa+b⁢(X/k).

Since k is perfect, the sum of the dimensions of the terms of the conjugate spectral sequence in total degree rj+1 equals the corresponding Hodge dimension sum, which by (HDrj+1) already equals the dimension of the abutment. Passing to the next page decreases a total-degree dimension sum by the ranks of the incoming and outgoing differentials. Consequently, all differentials touching this diagonal vanish; in particular, d20,rj=0.

Let o=ob⁡(X′/W2⁢(k))∈H2⁢(X′,TX′). By [1, Theorem 1.3 and Corollary 4.4],

0=d20,rj⁢(αj′)=−H2⁢(ιαj′)⁢(o).

To identify the operation, note that [1, §2.2 and Corollary 4.3] describe the underlying extension-class map as exterior comultiplication followed by 1⊗(−o), whose evaluation on a decomposable form is the usual alternating sum of contractions. No factor rj or rj! appears, so the identity also holds at p=2.

Flat base change by the field automorphism Fk preserves the injectivity of (1.1). Hence o=0, and the complete obstruction theory of smooth schemes gives a W2⁢(k)-lift of X′. Base change by the inverse Witt Frobenius automorphism of W2⁢(k) yields a lift of X, which is smooth and proper because these properties persist across the nilpotent thickening. ∎

Corollary 5.8.

Under the hypotheses of Proposition 5.7, assume also torsion-freeness of Hcrysrj+1⁢(X/W) for every j. Then the tuple (X,𝛂) has a lift to W2⁢(k) with all the forms closed.

Proof.

Choose one geometric lift from Proposition 5.7 and apply Lemma 5.6 to each rj on this same lift. The lifts of the individual forms are chosen independently, and together they lift all the prescribed forms simultaneously. ∎

The fixed-lift comparison used here is valid for the canonical divided powers on W2⁢(k): the smooth-lift comparison of [29, Tag 07MW] requires p nilpotent in the base, without requiring its divided powers to be nilpotent. At p=2, the identity γ2⁢(2)=2(mod4) means that W2⁢(k)→k is not a divided-power elementary extension. This does not affect the comparison just used, and the constructed pair lift supplies the separate initial point required by Theorem 5.5.

5.5. Lifting forms on a prescribed deformation

Lemma 5.9.

Let f:ℱ→𝒢 be a morphism of space-valued formal moduli problems over W. If ℱcl is unobstructed and H0⁢(𝕋ℱ)→H0⁢(𝕋𝒢) is surjective, then for every surjection of discrete Artinian bases A′↠A the map

ℱ⁢(A′)⟶ℱ⁢(A)×𝒢⁢(A)h𝒢⁢(A′)

is surjective on connected components. In particular, f is formally smooth on discrete bases, and 𝒢cl is unobstructed; no hull assumption is needed.

Proof.

This is the space-valued analogue of the classical criterion [29, Tag 0E3R]. It suffices to consider an elementary extension A′↠A with kernel I. Fix (x,y′,γ) in the displayed homotopy fibre product, where γ identifies f⁢(x) with y′|A. By (2.6), the components of the marked lift spaces over x and f⁢(x) are functorial torsors under H0⁢(𝕋ℱ)⊗kI and H0⁢(𝕋𝒢)⊗kI, respectively. The first is nonempty by unobstructedness, and tangent surjectivity therefore supplies a lift mapping to y′ compatibly with γ. Factorization into elementary extensions proves formal smoothness. Taking A→k shows that every 𝒢⁢(A)-point has a preimage in ℱ⁢(A); lifting such a preimage proves unobstructedness of 𝒢cl. ∎

5.6. Joint detection of obstructions

Proof of Theorem A.

First, every αj is closed: the Hodge differential d1rj,0 enters total degree rj+1, so (HDrj+1) forces it to vanish.

The tuple deformation problem is the finite homotopy fibre product of the individual closed-form problems over DefX. It is a formal moduli problem by closure under finite limits, and its tangent cohomology is finite-dimensional by properness and the bounded coherent terms below. Put

Vj=R⁢Γ⁢(X,Frj⁢ΩX∙)⁢[rj],Gj=R⁢Γ⁢(X,Frj−1⁢ΩX∙)⁢[rj],Qj=R⁢Γ⁢(X,ΩXrj−1)⁢[1],

and let V,G,Q be their respective direct sums. If 𝕋𝜶 denotes the tuple tangent object, Lemmas 4.2 and 4.6 give a morphism of triangles

V𝕋𝜶R⁢Γ⁢(X,TX)⁢[1]V⁢[1]VGQV⁢[1].⊕(−1)rj⁢ιαj

Indeed, the action of a vector field on each closed form is d⁢ιv⁢αj, so its action morphism factors through the connecting morphism of the corresponding filtered triangle. Composing the middle vertical arrow with the sum of the filtered inclusions gives the derivative of the product period map.

By Lemma 5.4, applied in degree rj+1, both connecting maps H0⁢(Q)→H1⁢(V) and H1⁢(Q)→H2⁢(V) vanish. The triangles therefore yield a diagram of short exact sequences

0H1⁢(V)H1⁢(𝕋𝜶)H2⁢(X,TX)00H1⁢(V)H1⁢(G)⨁jH2⁢(X,ΩXrj−1)0.

The right map is joint contraction followed by the diagonal signs (−1)rj from Proposition 4.4; it is injective by assumption, so the middle map is injective. Moreover, since H1⁢(Gj)→HdRrj+1⁢(X/k) is injective by Lemma 5.4, the tuple period map is injective on complete obstruction spaces.

By universal coefficients and the torsion-freeness of the indicated crystalline groups, the product of the deformation problems of periods is classically unobstructed. Naturality and completeness of obstruction classes then imply that the tuple functor is unobstructed across divided-power elementary extensions. At p=2, Corollary 5.8 supplies the initial tuple lift. Brantner–Taelman bootstrapping then gives unobstructedness across all Artinian surjections when k is infinite.

For finite k, extend scalars to k̅ and W⁢(k̅). Proper base change and flatness preserve all the hypotheses, and when p=2 the same construction supplies the initial point over W2⁢(k̅); hence the result holds over the infinite field k̅. For an elementary extension over k with kernel I, the complete obstruction lies in H1⁢(𝕋𝜶)⊗kI, its formation commutes with scalar extension, and its image over k̅ vanishes. Since extension of scalars is injective on this finite-dimensional vector space, the original obstruction vanishes, and factorization into elementary extensions gives unobstructedness over k.

Finally, the boundary H1⁢(X,TX)→H1⁢(V) factors through the zero map H0⁢(Q)→H1⁢(V), so H0⁢(𝕋𝜶)→H1⁢(X,TX) is surjective. Applied to the morphism from the tuple problem to DefX, Lemma 5.9 gives formal smoothness on discrete bases and unobstructedness of X. ∎

6. Volume forms, symplectic forms, and examples

We deduce Corollaries B and C from Theorem A and then give examples of the joint criterion, including varieties with a nontrivial canonical bundle and a nonzero geometric obstruction space. This section and the next depend on Sections 2, 3, 4 and 5 only through the statement of Theorem A.

6.1. Varieties with trivial canonical bundle

Let X be a smooth proper variety of dimension d over k.

Lemma 6.1.

Integral crystalline Poincaré duality for X gives perfect torsion pairings (with Frobenius twists omitted):

Hcrysi⁢(X/W)tors×Hcrys2⁢d+1−i⁢(X/W)tors⟶K/W. (6.1)

In particular, Hcrysd⁢(X/W) is torsion-free if and only if Hcrysd+1⁢(X/W) is torsion-free.

Proof.

Integral crystalline duality gives

R⁢Γcrys⁢(X/W)≃R⁢HomW⁡(R⁢Γcrys⁢(X/W),W)⁢[−2⁢d].

Apply Berthelot’s crystalline Poincaré duality [3, VII, §2.1] with coefficient crystal 𝒪 over each Wm⁢(k). The finite-level trace pairings are perfect, and passing to the compatible perfect complexes over W gives the displayed equivalence; equivalently, the trace-induced map is an equivalence modulo p, and hence over W by derived Nakayama. No freeness of individual crystalline cohomology groups is assumed here. Since W has global dimension one, the universal-coefficient sequence for the derived dual is

0⟶ ExtW1⁡(Hcrys2⁢d+1−i⁢(X/W),W)⟶Hcrysi⁢(X/W)
⟶HomW⁡(Hcrys2⁢d−i⁢(X/W),W)⟶0.

The last term is free, and for a finite torsion module T, ExtW1⁡(T,W)=HomW⁡(T,K/W). Thus the torsion in the middle is exactly the dual of the torsion in degree 2⁢d+1−i. This proves the claim; the extra +1 in this degree is essential. ∎

Proof of Corollary B.

Contraction with the chosen volume form η identifies TX with ΩXd−1, so it is injective on H2. By Lemma 6.1, torsion-freeness of Hcrysd⁢(X/W) implies torsion-freeness of Hcrysd+1⁢(X/W). Apply Theorem A with r=d. Top-degree forms are closed, and a lift of η remains nowhere vanishing by Nakayama’s lemma. ∎

6.2. Symplectic varieties

Lemma 6.2.

If ισ:TX→ΩX1 is an isomorphism for a two-form σ, then d=2⁢n and ωX≃𝒪X. Under (HD3), d⁢σ=0.

Proof.

The corresponding nondegenerate alternating pairing has even rank and admits local symplectic bases. Changes between these bases have determinant one; equivalently, the integral Pfaffian polynomial transforms as a top exterior form. The resulting nowhere-vanishing section Pf⁡(σ) trivializes ωX. This argument also works at p=2 and when p divides n!, since the relation σ∧n=n!⁢Pf⁡(σ) is not used to define Pf⁡(σ).

Finally, the map d:H0⁢(X,ΩX2)→H0⁢(X,ΩX3) is the d12,0 differential of the Hodge–de Rham spectral sequence. Since it enters total degree three, (HD3) forces it to vanish. ∎

Proof of Corollary C.

The contraction isomorphism is injective on H2, so Theorem A applies with r=2. The assumed numerical equality makes σ closed, and degree-three crystalline torsion-freeness is the only integral hypothesis required, even at p=2. Nondegeneracy persists on every lift by Nakayama’s lemma. ∎

Remark 6.3.

Let α be a closed r-form with 1≤r≤d. Invertibility of contraction in (1.3) forces d=(dr−1), hence r=2 or r=d when d≥2, since (dr−1)>d for 2≤r−1≤d−2. For r=d, the identification ΩXd−1≃TX⊗ωX makes this the nowhere-vanishing volume-form condition. For r=2, it makes α symplectic, so d is even and ωX is trivial by Lemma 6.2. In dimension one only the volume case occurs; in dimension two the cases coincide, and r≥3 always forces d=r. The joint injectivity on H2 in Theorem A imposes no such bundle-rank restriction, permitting the examples with nontrivial canonical bundle in Section 6.3.

Example 6.4.

Let A/k be an abelian variety of dimension g. A basis α1,…,αg of invariant one-forms identifies TA with 𝒪A⊕g by joint contraction. The Hodge–de Rham spectral sequence degenerates [25, Proposition 5.1], and crystalline cohomology is the exterior algebra on the free module Hcrys1⁢(A/W) [18, II, §7.1]. Thus Theorem A applies in every characteristic. It recovers unobstructedness of the unpolarized deformation functor, whose tangent space has dimension g2, and also gives relative lifting of the chosen invariant basis. With these one-forms, the criterion uses crystalline degree two in every characteristic, independently of g.

Example 6.5.

For a K3 surface S/k, crystalline cohomology is free, the Hodge–de Rham spectral sequence degenerates, and H0⁢(S,TS)=H2⁢(S,TS)=0, while h1⁢(S,TS)=20; see [11, Proposition 1.1]. A generator of ωS is a nondegenerate two-form. Hence Corollary C recovers unobstructedness over W, including in characteristic two, and relative lifting of this form. Here the usual geometric obstruction space already vanishes; the products below give examples in which it does not. If L is ample and c1H⁢(L)≠0, Serre duality makes λL surjective, so Theorem D also gives unobstructed deformations of the polarized surface with its two-form and a smooth projective W-lift. Projectivity of a specified formal lift requires an additional condition; see Theorem 7.4.

Moduli spaces of stable sheaves give higher-dimensional examples. For the following proposition and examples, assume that k is algebraically closed. If S is a K3 or abelian surface, write MH⁢(S,v) for the moduli space of Gieseker H-stable sheaves with Mukai vector v=(r,c1,s), where r is the rank and s=χ⁢(E)−r in the K3 case, while s=χ⁢(E) in the abelian case. We put v2=c12−2⁢r⁢s and call H v-general if every H-semistable sheaf of vector v is stable. On an abelian surface the determinant is allowed to vary within its numerical class.

Proposition 6.6.

Let p≥5, let S/k be a K3 surface, and let v=(r,c1⁢(L),s) be primitive with r>0 and v2>0. Suppose that H is v-general and that (S,H,L) admits a smooth projective lift over W. Then M=MH⁢(S,v) is a smooth projective symplectic variety of dimension v2+2, and

Hcrys3⁢(M/W)=0,Hb⁢(M,ΩMa)=0(a+b=3).

Consequently, Corollary C applies to M and its Mukai two-form, with no bound on dimM.

Proof.

The geometric assertions and the relative moduli construction are given in [12, Proposition 4.4 and its proof]. The chosen lift of (S,H,L) gives a smooth projective moduli space ℳ/W with special fibre M; generality of the polarization persists on the geometric generic fibre. After passage to characteristic zero, this fibre is of K3[n] deformation type, where 2⁢n=v2+2. The Betti-number formula and integral torsion-freeness for Hilbert schemes of K3 surfaces [32, Theorems 1.1 and 3.1] therefore give

He´⁢t3⁢(ℳK̅,ℤp)=0,He´⁢t4⁢(ℳK̅,ℤp)⁢ is torsion-free.

The adjacent degree-four assertion ensures that He´⁢t3⁢(ℳK̅,ℤ/p)=0. Fontaine–Messing comparison in degree 3≤p−2, in the form recalled in [24, Theorem 0.3], now gives HdR3⁢(M/k)=0. The crystalline universal-coefficient sequence

0⟶Hcrys3⁢(M/W)⊗Wk⟶HdR3⁢(M/k)⟶Hcrys4⁢(M/W)⁢[p]⟶0

and Nakayama’s lemma give the vanishing of crystalline cohomology. Finally, the lift modulo p2 and the truncated Deligne–Illusie decomposition [10, Theorem 2.1] imply

∑a+b=3hb⁢(M,ΩMa)=dimkHdR3⁢(M/k)=0,

since 3<p. Thus (HD3) holds, and Corollary C applies. ∎

The unramified lifting hypothesis in Proposition 6.6 is part of the criterion: the existence of a lift over a ramified extension of W alone does not justify this comparison argument. The following examples reach the smaller primes by direct cohomological calculations.

Example 6.7 (K3 Hilbert squares in odd characteristic).

Let S/k be a K3 surface and p>2. Its Hilbert square is the stable-sheaf moduli space

S[2]=MH⁢(S,(1,0,−1)),Z⟼ℐZ,

for any polarization H, and carries the Mukai two-form [12, Proposition 4.4]. Let B be the blowup of S×S along the diagonal. There is a finite degree-two morphism q:B→S[2]. The blowup formula adds the cohomology of S shifted by (1,1) to that of S×S; it holds integrally for crystalline cohomology and for Hodge cohomology [13, IV, Theorem 1.1.9, Corollary 1.1.11, and Theorem 1.2.1]. Together with Künneth and Example 6.5, it gives

Hcrys3⁢(B/W)=0,Hb⁢(B,ΩBa)=0(a+b=3).

Pullback followed by pushforward is multiplication by two [13, II, Proposition 4.2.9 and its proof], so pullback is injective in these theories. The same vanishing therefore holds on S[2]. Its degree-three de Rham cohomology vanishes as well, so (HD3) holds. Thus Corollary C applies, including in characteristic three, without a lifting assumption on the polarized surface.

Example 6.8 (Sheaves on abelian surfaces at small primes).

Let A/k be an abelian surface and Aˆ=PicA/k0 its dual abelian variety. First suppose that p>2, that v=(r,c1,s) is primitive with r>0 and v2=4, and that H is v-general. Then M=MH⁢(A,v) is a smooth projective sixfold with a Mukai two-form. Its Albanese fibre F is a K3 surface, and base change of the Albanese map by [2] on A×Aˆ gives a finite étale cover

F×A×Aˆ⟶Mof degree ⁢28;

see [12, Proposition 6.9(i) and diagram (50) in its proof]. To see that F is a K3 surface, note that the smooth proper lift constructed there has a K3 geometric generic fibre; hence F is connected and χ⁢(𝒪F)=2. Its symplectic form trivializes ωF, so h1⁢(𝒪F)=0. By Examples 6.4 and 6.5 and Künneth, the covering variety has torsion-free crystalline cohomology and E1-degeneration. Since 28 is invertible in W, normalized trace splits pullback on crystalline cohomology and on the filtered de Rham complex. Both properties descend to M, proving the hypotheses of Corollary C, also when p=3. For instance, v=(1,0,−2) gives M≃A[2]×Aˆ.

There is also a four-dimensional stable-sheaf example in characteristic two. In every characteristic,

MH⁢(A,(1,0,−1))≃A×Aˆ,(x,L)⟼ℐx⊗L.

Indeed, a rank-one torsion-free sheaf with these invariants has double dual L∈Aˆ⁢(k) and quotient of length one, and is stable for every polarization. The sum of nonzero invariant two-forms from the two abelian surfaces is nondegenerate, including at p=2. The cohomological assertions follow from Example 6.4, so Corollary C applies here as well. These examples verify the theorem’s hypotheses at the stated primes; they do not assert them for arbitrary sheaf moduli spaces in characteristics two and three.

6.3. Joint contraction on products

For forms 𝜶Y on Y, let cYq denote the joint contraction map from Hq⁢(Y,TY) to the sum of the Hq⁢(Y,ΩYrj−1).

Lemma 6.9.

Let Y1,…,Ym be smooth proper geometrically connected varieties, each equipped with a finite tuple of forms. Suppose that cYiq is injective for 0≤q≤2 and every i. Then the pullbacks of all these forms to X=∏iYi jointly detect H2⁢(X,TX).

If every Yi has E1-degeneration and torsion-free crystalline cohomology in all degrees, then X has both properties as well. In this case Theorem A applies to the pulled-back tuple in every characteristic.

Proof.

Write Zi=∏j≠iYj. The tangent bundle splits by factors, and coherent Künneth gives

H2⁢(X,TX)=⨁i⨁q=02Hq⁢(Yi,TYi)⊗kH2−q⁢(Zi,𝒪Zi).

Forms pulled back from Yi contract to zero on the other tangent factors, and Künneth identifies their joint contraction on the displayed summands with cYiq⊗id. These maps are injective; for one-forms, the separately indexed targets for the forms keep the tangent factors distinct.

The crystalline Künneth theorem and freeness give the integral assertion. Coherent and de Rham Künneth give E1-degeneration on the product by comparing dimensions in each total degree. ∎

Products X=∏iCi of smooth proper geometrically connected curves of genus at least two have factorwise deformation theory over Artinian k-algebras [4, Theorem 3.3], although H2⁢(X,TX)≠0 with at least two factors [4, Warning 3.5]. The same splitting holds over W by induction on elementary extensions: H2⁢(Ci,TCi)=0 lifts the factors, while H0⁢(Ci,TCi)=0 and Künneth identify H1⁢(X,TX) with ⨁iH1⁢(Ci,TCi) and split the lift torsors accordingly. Relative canonical sections are free and commute with base change, so the chosen one-forms lift to every prescribed deformation and their pullbacks are closed. These conclusions require no nonhyperellipticity assumption; the following proposition determines when joint detection holds.

Proposition 6.10.

Let C1,…,Cm be smooth proper geometrically connected curves of genera gi≥2. Assume that each Ci has genus two or is geometrically nonhyperelliptic. On X=∏iCi, take the pullbacks of bases of H0⁢(Ci,ωCi). These forms satisfy the hypotheses of Theorem A in every characteristic. Moreover, ωX is ample and

h1⁢(X,TX) =∑i(3⁢gi−3), (6.2)
h2⁢(X,TX) =∑i(3⁢gi−3)⁢∑j≠igj. (6.3)
Proof.

The canonical multiplication map

H0⁢(Ci,ωCi)⊗H0⁢(Ci,ωCi)⟶H0⁢(Ci,ωCi⊗2)

is surjective by Max Noether’s theorem, including genus two in every characteristic [33, Theorem 3.2.2]. By Serre duality, surjectivity is precisely the injectivity of

H1⁢(Ci,TCi)⟶Homk⁡(H0⁢(Ci,ωCi),H1⁢(Ci,𝒪Ci)).

In degrees zero and two, the source Hq⁢(Ci,TCi) vanishes.

Smooth proper curves have E1-degeneration [25, Corollary 5.2]. Their crystalline cohomology is torsion-free: the universal-coefficient sequence in degree zero, together with Hcrys0=W and HdR0=k, first gives freeness of Hcrys1, and Lemma 6.1 then gives freeness of Hcrys2. Apply Lemma 6.9. The dimension formulas follow from Künneth and h1⁢(Ci,TCi)=3⁢gi−3. The canonical bundle ωX is ample, being the external tensor product of the ample bundles ωCi. ∎

Example 6.11.

For two smooth plane quartics C,D, the six pulled-back one-forms detect the obstruction space of X=C×D, where

h1⁢(X,TX)=12,h2⁢(X,TX)=36,h2⁢(X,𝒪X)=9.

Since ωC≃𝒪C⁢(1), canonical multiplication is simply multiplication of linear forms into quadrics on ℙ2. No single one-form can detect the 36-dimensional space through contraction into the 9-dimensional H2⁢(X,𝒪X). Thus Theorem A recovers classical unobstructedness by joint detection on a variety with ample canonical bundle.

For comparison, let C be hyperelliptic of genus g≥3 and let D satisfy Proposition 6.10, with genus h. Canonical multiplication on C has cokernel of dimension g−2 [33, §3.3]. Serre duality and Künneth therefore give a kernel of dimension (g−2)⁢h>0 for joint contraction by the pulled-back canonical bases on C×D. The variety and its tuple of forms remain unobstructed: the detection criterion is sufficient but not necessary.

Example 6.12.

Let S be a K3 surface, and let C,D satisfy Proposition 6.10. On X=S×C×D, take the two-form from S and bases of one-forms from C and D. Contraction with the K3 form is an isomorphism of tangent and cotangent bundles, so Lemma 6.9 applies, and both the tuple and the underlying variety are unobstructed over W in every characteristic.

The K3 form detects the summands from TS, and the curve forms those from TC and TD; neither subcollection detects all of H2⁢(X,TX). This illustrates joint detection by forms of different degrees on a variety whose nontrivial canonical bundle is pulled back from the ample canonical bundle of C×D.

7. Polarizations and projective lifts

7.1. Lifting a polarization and algebraization

Proposition 7.1.

Assume the hypotheses of Theorem A. Suppose also that X is projective and H2⁢(X,𝒪X)=0, and let L be an ample line bundle on X. Then deformations of (X,𝛂,L) are unobstructed, and forgetting L is formally smooth on discrete Artinian bases. Moreover, every compatible formal lift of (X,𝛂) over W is the completion of a smooth projective W-scheme carrying closed relative lifts of the forms and a lift of L.

Proof.

Let A′↠A be an elementary extension with kernel I, and fix a lift Y′/A′ of a deformation Y/A. The exact sequence of units for this thickening gives an obstruction to lifting a line bundle on Y in

H2⁢(Y,I⁢𝒪Y′)=I⊗kH2⁢(X,𝒪X)=0;

see [29, Tag 0C6R]. Since ampleness persists across a nilpotent thickening, factorization into elementary extensions proves the relative formal smoothness, and the unobstructedness assertion then follows from Theorem A.

Along a compatible system of Wn-lifts, lift L recursively with restriction identifications. This gives an invertible sheaf, ample on the special fibre, on the resulting proper adic formal scheme. Grothendieck’s algebraization theorem [14, Theorem 5.4.5] supplies a projective scheme 𝒳/W with this completion and line bundle. It is smooth near the special fibre, and its closed nonsmooth locus is proper over W, so it would meet the special fibre if nonempty; hence 𝒳/W is smooth.

Coherent formal GAGA algebraizes each compatible section of Ω𝒳/Wrj. Their exterior derivatives vanish modulo every pn and hence vanish, since H0⁢(𝒳,Ω𝒳/Wrj+1) is a finite, p-adically separated W-module. ∎

In particular, a projective variety of dimension at least three with trivial canonical bundle and H2⁢(X,𝒪X)=0 has a smooth projective W-lift whenever it satisfies the hypotheses of Corollary B.

7.2. Choosing a lift that preserves a polarization

We replace the vanishing assumption in Proposition 7.1 by surjectivity of λL, allowing the geometric lift to vary; prescribed formal lifts are treated separately below.

Proof of Theorem D.

Let A′↠A be an elementary extension with kernel I, and let (Y,LA) deform (X,L) over A. Choose a marked smooth proper lift Y′/A′ using Theorem A. The exact sequence of units gives a complete obstruction

oY′⁢(LA)∈H2⁢(X,𝒪X)⊗kI

to extending LA to Y′; see [29, Tag 0C6R]. Changing the marked geometric lift by ξ∈H1⁢(X,TX)⊗kI changes this obstruction by

oYξ′⁢(LA)−oY′⁢(LA)=(λL⊗kI)⁢(ξ), (7.1)

for compatible choices of signs.

To verify this also in mixed characteristic, choose affine trivializations of LA, with lifted geometric and line-bundle transition maps gi⁢j′ and hi⁢j′. The obstruction is represented by hi⁢j′⁢gi⁢j′⁢(hj⁢k′)⁢(hi⁢k′)−1∈1+I⁢𝒪Y′. Replacing gi⁢j′ by (1+θi⁢j)⁢gi⁢j′ for a Čech cocycle representing ξ changes this discrepancy, under 1+I⁢𝒪Y′≃I⊗k𝒪X, by

θi⁢j⁢(hj⁢k)hj⁢k=θi⁢j⁢⌟⁢d⁢log⁡(hj⁢k).

Here all coefficients multiplying I reduce to X, since 𝔪A′⁢I=0. This proves (7.1); compare also [26, §1.12 and Corollary 1.14].

Since λL is surjective, we may choose ξ to cancel oY′⁢(LA), and then (Y,LA) lifts. If prescribed closed forms are also part of the data, the relative formal smoothness in Theorem A lifts them to this modified geometric lift. Factoring an arbitrary Artinian surjection into elementary extensions proves both unobstructedness assertions.

Recursively choose actual lifts and restriction identifications over Wn⁢(k). The resulting proper formal scheme carries a compatible invertible sheaf restricting to L. The algebraization and formal GAGA argument in Proposition 7.1 supplies the asserted smooth projective scheme, ample line bundle, and closed forms. ∎

Remark 7.2.

The proof chooses the geometric lift at each step, and it does not algebraize every prescribed formal lift. Indeed, the tangent image of Def(X,L)/W→DefX/W is ker⁡λL, so this morphism is not formally smooth when λL≠0.

7.3. A prescribed formal lift

Let 𝔛 be a smooth proper adic formal scheme over Spf⁡W with special fibre X, and put Xm=𝔛×Spf⁡WSpec⁡(W/pm). Write

HdR2⁢(𝔛/W)=ℍ2⁢(𝔛,Ωˆ𝔛/W∙),F1⁢HdR2=im⁡(ℍ2⁢(𝔛,F1⁢Ωˆ∙)⟶HdR2⁢(𝔛/W)).

Here the differential forms are continuous, and the filtration is the stupid filtration of their de Rham complex. Crystalline comparison gives an isomorphism

comp𝔛:Hcrys2⁢(X/W)⁢[1/p]⟶∼HdR2⁢(𝔛/W)⁢[1/p].

The filtration step F1 depends on the chosen lift.

Proposition 7.3.

For L∈Pic⁡(X), the following are equivalent:

  1. (i)

    Some L⊗pN extends to an invertible sheaf on 𝔛.

  2. (ii)

    comp𝔛⁡(c1crys⁢(L)) belongs to F1⁢HdR2⁢(𝔛/W)⁢[1/p].

More precisely, let pt annihilate the torsion of the finite W-module H2⁢(𝔛,𝒪𝔛). If (ii) holds, then L⊗pt extends for p>2, and L⊗2t+1 extends for p=2.

Proof.

This is the formal p-adic Lefschetz criterion; we give the obstruction argument, including the issue in characteristic two. The compatibility of the logarithmic boundary with crystalline first Chern classes is established in [20, Lemma 3.2.7]; the same reference, Theorem 4.2.1, proves the assertion for flat proper formal schemes of ramification index e<p. Here e=1.

For p>2, the exact sequence

1⟶1+p⁢𝒪𝔛⟶𝒪𝔛×⟶𝒪X×⟶1

and the convergent logarithm 1+p⁢𝒪𝔛≃p⁢𝒪𝔛 give an exact obstruction sequence

Pic⁡(𝔛)⟶Pic⁡(X)⟶𝛿H2⁢(𝔛,p⁢𝒪𝔛). (7.2)

After inverting p, the image of δ⁢(L) in H2⁢(𝔛,𝒪𝔛)⁢[1/p] equals the de Rham projection of comp𝔛⁡(c1crys⁢(L)). This compatibility can be seen in the Čech construction: logarithms of the triple-overlap discrepancies of lifted transition functions are the degree-zero components of the crystalline Chern cocycle. The exact sequence of complexes F1⁢Ωˆ∙→Ωˆ∙→𝒪 shows that the kernel of this projection is precisely F1; no Hodge–de Rham degeneration is required.

Consequently, (ii) says that δ⁢(L) vanishes after inverting p. Since 𝔛/W is flat, multiplication by p identifies 𝒪𝔛 with p⁢𝒪𝔛. Thus the torsion of the target in (7.2) is annihilated by pt, so δ⁢(L⊗pt)=0, and exactness proves the claimed extension.

For p=2, the obstruction to extending L from X to X2 lies in H2⁢(X,𝒪X) and is killed by 2. Hence L⊗2 has an extension L2 on X2. Now use

1⟶1+4𝒪𝔛⟶𝒪𝔛×⟶𝒪X2×⟶1,log:1+4𝒪𝔛⟶∼4𝒪𝔛.

Functoriality of crystalline Chern classes and their comparison maps identifies the rational image of this boundary with the projection of 2⁢comp𝔛⁡(c1crys⁢(L)); this is the comparison for the intermediate thickening in the proof of [20, Theorem 4.2.1]. Under (ii), the boundary of L2 therefore vanishes after inverting 2. Since the torsion of the target is killed by 2t, the sheaf L2⊗2t extends to 𝔛, and its restriction to X is L⊗2t+1.

Conversely, the de Rham Chern class of an invertible sheaf on 𝔛 lies in F1 and restricts to its crystalline Chern class. Dividing by pN proves (ii) from (i). Finally, the finiteness of formal coherent cohomology used above is [14, Theorem 3.4.2]. ∎

Theorem 7.4.

Suppose the special fibre X is projective. Then 𝔛 is the completion of a smooth projective W-scheme if and only if there is an ample L∈Pic⁡(X) such that

comp𝔛⁡(c1crys⁢(L))∈F1⁢HdR2⁢(𝔛/W)⁢[1/p]. (7.3)

In that case, every compatible closed relative form on 𝔛 algebraizes. In particular, every such 𝔛 with H2⁢(𝔛,𝒪𝔛)⁢[1/p]=0 is projectively algebraizable.

Proof.

Under (7.3), Proposition 7.3 supplies an invertible sheaf whose special-fibre restriction is a positive tensor power of L, hence ample. Apply [14, Theorem 5.4.5]. Smoothness and the algebraization of closed forms follow exactly as in the proof of Proposition 7.1 (no assumption on H2⁢(X,𝒪X) is required). Conversely, restrict a relatively ample line bundle on a projective algebraization and use the necessity in Proposition 7.3. If the rational second coherent cohomology vanishes, de Rham projection is zero, so its kernel F1 contains every degree-two crystalline class after comparison. ∎

Corollary 7.5.

Let X/k be smooth and proper. If either ωX or ωX−1 is ample, every smooth proper formal lift 𝔛/Spf⁡W of X has a smooth projective algebraization. Compatible closed relative forms on 𝔛 algebraize as well.

Proof.

The relative canonical bundle ω𝔛/W=detΩˆ𝔛/W1 or its inverse restricts to an ample line bundle on X. Grothendieck’s algebraization theorem [14, Theorem 5.4.5] gives a projective algebraization; smoothness and the algebraization of closed forms follow as in Proposition 7.1. ∎

For example, every prescribed smooth proper formal lift of a product of curves in Proposition 6.10 is projectively algebraizable. With two or more factors, H2⁢(X,𝒪X) is nonzero, so this does not follow from Proposition 7.1; instead, the canonical line bundle lifts to every deformation, whether or not λωX is surjective.

8. Crystalline torsion: a refinement and two examples

The torsion-freeness hypothesis makes the deformation problem of periods classically unobstructed. For unobstructedness of geometric deformations this is stronger than necessary, because the derivative sees only a filtered part of de Rham cohomology. We describe the intersection that can contain actual obstructions, and in Section 8.3 we construct two symplectic varieties in characteristic three with torsion in Hcrys3 but opposite deformation behaviour. Throughout this section, (X,σ) is a smooth proper symplectic pair over k, and

D=R⁢Γcrys⁢(X/W)⊗W𝐋k≃R⁢Γ⁢(X,ΩX∙),T=Hcrys3⁢(X/W)tors.

We assume that σ is closed; this follows automatically from (HD3) whenever that condition is imposed below.

8.1. The support of coefficient boundaries

Reduction and universal coefficients give a canonical injection

T/p⁢T↪Hcrys3⁢(X/W)/p⁢Hcrys3⁢(X/W)↪H3⁢(D). (8.1)

Indeed, Hcrys3⁢(X/W)/T is free, so T∩p⁢Hcrys3⁢(X/W)=p⁢T. Let S⊂H3⁢(D) denote the image of (8.1). Thus S≃T/p⁢T, and S is defined without choosing a decomposition of the crystalline complex.

Lemma 8.1.

For every elementary extension A′↠A in ArtW, with kernel I, the coefficient boundary satisfies

im⁡(∂A′/A:H2⁢(R⁢Γcrys⁢(X/W)⊗W𝐋A)⟶H3⁢(D)⊗kI)⊂S⊗kI.
Proof.

Represent R⁢Γcrys⁢(X/W) by a bounded complex of finite free W-modules. Successive Smith decompositions split it into free one-term summands and two-term summands [W⟶paW]. Only a two-term summand in degrees two and three can contribute to the displayed boundary. Its degree-two coefficient cohomology is A⁢[pa], and its connecting map is

A⁢[pa]⟶I,x⟼pa⁢x˜,

where x˜∈A′ lifts x. The expression belongs to I and is independent of the lift because p⁢I=0. Its target is the reduction of the upper term of that torsion summand, so its image lies in S⊗kI. All other summands have zero boundary in this degree. Summing over the summands proves the assertion, and the support S itself is the canonical subspace defined above. ∎

8.2. The selective obstruction space

Write

Oσ=H3⁢(F1⁢D),ji:Hi⁢(F1⁢D)⟶Hi⁢(D),ei:Hi⁢(D)⟶Hi⁢(X,𝒪X).

Here Fa⁢D=R⁢Γ⁢(X,Fa⁢ΩX∙) uses the Hodge filtration under the displayed de Rham identification, and ei is induced by projection to form degree zero. By Theorem 4.6, Oσ is the complete obstruction space of the pair problem. Set Osel=j3−1⁢(S).

Proposition 8.2.

Every pair obstruction over a divided-power elementary extension with kernel I belongs to Osel⊗kI. There is a canonical exact sequence

0⟶coker⁡(e2)⟶Osel⟶j3ker⁡(e3|S)⟶0. (8.2)

Under (HD3), j3 identifies this space with S∩F1⁢H3⁢(D), where F1⁢H3⁢(D)=im⁡j3.

Proof.

By Proposition 5.1, the image of the complete pair obstruction is the coefficient boundary of its period. Lemma 8.1 therefore places it in S⊗kI, proving the first assertion. The triangle F1⁢D→D→R⁢Γ⁢(X,𝒪X) identifies ker⁡j3 with coker⁡(e2) and im⁡j3 with ker⁡e3. Taking the inverse image of S gives (8.2). Under (HD3), Lemma 5.4 makes j3 injective, so its restriction identifies Osel with the stated intersection. ∎

Corollary 8.3.

Assume (HD3) and

S∩F1⁢HdR3⁢(X/k)=0. (8.3)

Then Def(X,σ)/W and DefX/W are unobstructed over W, and forgetting the form is formally smooth on discrete Artinian bases. This holds for every prime p.

Proof.

By Proposition 8.2 and completeness, every obstruction across a divided-power elementary extension vanishes. At p=2, we supply the initial pair separately. Since contraction by σ is an isomorphism, Proposition 5.7 gives a smooth proper geometric lift Y/W2⁢(k). The filtered coefficient sequence for its de Rham complex gives a Bockstein

H2⁢(F2⁢D)⟶H3⁢(F2⁢D).

By crystalline comparison and filtered naturality, the image of the Bockstein of σ in H3⁢(D) is the total coefficient boundary for W2⁢(k)→k. Lemma 8.1 places this image in S, and the filtered construction places it in F2⁢H3⁢(D), so it vanishes by (8.3). Since H3⁢(F2⁢D)→H3⁢(D) is injective by Lemma 5.4, the filtered Bockstein vanishes. Hence σ lifts to a closed relative two-form on Y. Its contraction remains invertible by Nakayama’s lemma.

For infinite k, Theorem 5.5 now proves unobstructedness of the pair functor. For finite k, apply this argument over k̅; the complete obstruction spaces commute with scalar extension, so their actual classes vanish already over k. Finally, Lemma 5.4 gives the surjection H2⁢(F1⁢D)↠H1⁢(X,ΩX1)≃H1⁢(X,TX), so Lemma 5.9 proves both the relative assertion and unobstructedness of the underlying variety. ∎

Condition (8.3) concerns the position of the torsion reduction relative to the Hodge filtration rather than its length or exponent, and taking T=0 recovers the symplectic criterion of Corollary C. On the other hand, a nonzero Osel is only a space of possible obstruction values, and its nonvanishing does not prove that an actual obstruction is realized. When this space is nonzero, the argument asserts neither the continuation of every finite pair nor the existence of a compatible ramified lifting branch; these questions require additional control of the period and its coefficient boundaries.

8.3. Two projective examples with crystalline torsion

We exhibit both possible behaviours of the classical deformation functor when Hcrys3 has torsion. The obstructed example admits a projective symplectic W-lift, so the existence of one lift must be distinguished from the unobstructedness of the deformation functor. Throughout these examples, k=𝔽̅3 and W=W⁢(k). For a smooth proper k-scheme V, write

bi⁢(V)=dimKHcrysi⁢(V/W)⁢[1/3],ti⁢(V)=dimkHcrysi⁢(V/W)⁢[3].

Universal coefficients give

dimkHdRi⁢(V/k)=bi⁢(V)+ti⁢(V)+ti+1⁢(V).

Write Φ for crystalline Frobenius, with its semilinearity over Witt-vector Frobenius understood, and C3 for the cyclic group of order three, whose generator acts by the indicated automorphism in each construction.

We use the elliptic scheme

ℰ/ℤ3:y2+xy+y=x3,P=(0,0).

Its discriminant is −26. The tangent line y=0 meets the cubic with multiplicity three at P, so P has order three. Moreover, the points 0,P,2⁢P=(0,−1) remain distinct modulo three. We also write ℰ for its base change to W, and E for its special fibre. The three distinct torsion points show that E is ordinary. Put CE=E/⟨P⟩.

Ordinarity makes Cartier bijective on an invariant volume form, so Frobenius duality gives a normalized splitting of Frobenius on an ordinary abelian variety. Since its canonical bundle is trivial, the space of splitting maps is the one-dimensional space of sections of ω1−3, and normalization fixes the scalar. The splitting is therefore invariant under every automorphism and descends through any finite étale quotient, including one of degree three. Petrov’s decomposition theorem then gives full Hodge–de Rham degeneration on these quotients [27, Theorem 1.1].

These quotients are also Bloch–Kato ordinary: for every r≥1, the sheaf BZ/kr=im⁡(d:ΩZr−1→ΩZr) of exact r-forms has vanishing cohomology in all degrees [8, Definition 8.8]. Indeed, exact forms commute with pullback on the étale sites, so for either abelian cover A→Z the Cartan–Leray spectral sequence is

Ha⁢(C3,Hb⁢(A,BA/kr))⟹Ha+b⁢(Z,BZ/kr).

Every term on its second page vanishes because A is ordinary. This proves ordinarity of the quotient also when the covering degree is three.

All quotients below arise from free actions on abelian schemes over W. They are smooth and proper, and projectivity follows by tensoring an ample bundle over its finite group orbit and descending the resulting linearized ample bundle. For both crystalline cohomology and relative Hodge cohomology, pullback followed by trace is multiplication by three, and the corresponding groups on the abelian cover are W-free. Hence three kills all torsion in the crystalline and relative Hodge groups of either quotient.

Lemma 8.4.

Let X be a smooth proper scheme over a perfect field of characteristic p. If crystalline Frobenius is bijective on T=Hcrys3⁢(X/W)tors, then

S∩F1⁢HdR3⁢(X/k)=0.
Proof.

The reduction of crystalline Frobenius is induced on the de Rham complex by f↦fp in form degree zero and by zero in positive form degrees, with the corresponding Frobenius twist on the coefficient field understood. This is functorial derived base change; it does not replace derived reduction by reduction of an individual crystalline cohomology group. In particular, it kills the image of H3⁢(F1⁢D). On the canonical torsion reduction S≃T/p⁢T, it is the reduction of Frobenius on T and is therefore bijective, because the universal-coefficient injection Hcrys3⁢(X/W)/p↪H3⁢(D) is Frobenius-equivariant, even when Hcrys4⁢(X/W) has torsion. A vector in the stated intersection is thus killed by this map and lies in a subspace on which the map is injective, so it is zero. ∎

8.4. An obstructed fourfold admitting a projective lift

Let g act on ℰ4 by translation by P on the first factor, by

M=(0−11−1)

on the second and third factors, and by the identity on the fourth. Since M3=1, this defines an order-three action. Each nonidentity element translates the first factor by a nonzero section, so the action is free over W. Put

𝒴=ℰ4/⟨g⟩,Y=𝒴k.

For an invariant differential ω on ℰ, let ωi denote its pullback from the ith factor. The form

σ˜Y=ω1∧ω4+ω2∧ω3

is invariant, closed and nondegenerate; invariance on the middle factors follows from detM=1. It therefore descends to a relative symplectic form on 𝒴, and we write σY for its reduction.

Proposition 8.5.

The smooth projective symplectic fourfold (Y,σY) satisfies

Hcrys2⁢(Y/W)tors≃(W/3)⊕2,Hcrys3⁢(Y/W)tors≃(W/3)⊕6.

Both DefY/W and Def(Y,σY)/W are obstructed over W. More precisely, three linearly independent geometric obstruction directions are realized on an elementary extension with one-dimensional kernel. Nevertheless, the pair has the projective symplectic lift 𝒴/W just constructed.

Proof.

We first compute the cohomology, without using the crystalline obstruction comparison. Put

Z=(E1×E2×E3)/⟨g⟩.

Then Y=Z×E4, and f:Z→CE is an abelian-surface fibration trivialized by the connected étale torsor E→CE. Modulo three, the matrix M is a nontrivial unipotent block of size two, and its associated vector bundle 𝒱 is a nonsplit extension

0⟶𝒪CE⟶𝒱⟶𝒪CE⟶0.

Indeed, the extension cocycle is the image in H1⁢(CE,𝒪CE) of the nonzero torsor class in He´⁢t1⁢(CE,𝔽3). The Artin–Schreier sequence makes this map injective, because F−1:k→k is surjective. Thus h0⁢(𝒱)=h1⁢(𝒱)=1. Descent gives

R0⁢f∗⁢𝒪Z=𝒪CE,R1⁢f∗⁢𝒪Z=𝒱,R2⁢f∗⁢𝒪Z=𝒪CE.

Since CE is a curve, there are no Leray differentials. Hence (hi⁢(𝒪Z))i=03=(1,2,2,1), and product Künneth gives

(hi⁢(𝒪Y))i=04=(1,3,4,3,1). (8.4)

Write Jd=k⁢[t]/(td), with t=g−1. The representation on invariant one-forms of the cover is J1⊕2⊕J2. Taking global sections after descent gives

h0⁢(ΩY1)=3,h0⁢(ΩY2)=4,

since ∧2(J1⊕2⊕J2)=J1⊕2⊕J2⊕2. To compute the remaining Hodge number, we use

ΩY1≃𝒪Y⊕2⊕prZ∗⁡f∗⁢𝒱.

In characteristic three, J2⊗J2=J3⊕J1. The regular block J3 corresponds to the direct image of 𝒪E under E→CE, so h0⁢(𝒱⊗𝒱)=2. The projection formula and Leray now give h1⁢(Z,f∗⁢𝒱)=1+2=3, and Künneth with E4 gives h1⁢(Y,prZ∗⁡f∗⁢𝒱)=3+1=4. Consequently

h1⁢(ΩY1)=2⋅3+4=10. (8.5)

In the rational degree-one cohomology of the cover, the trivial character has multiplicity four and each nontrivial character of C3 has multiplicity two. Rational descent and exterior powers therefore give b1⁢(Y)=4 and b2⁢(Y)=(42)+2⋅2=10. Hodge degeneration gives de Rham dimensions 6 and 4+10+4=18 in degrees one and two. Since t1⁢(Y)=0 by universal coefficients in degree zero, the dimension formula gives t2⁢(Y)=2 and t3⁢(Y)=6. Since the trace argument above shows that the torsion is killed by three, this proves the asserted modules.

To prove actual obstructedness, put Qi=Hi⁢(𝒴,Ω𝒴/W1). These modules also have torsion killed by three. After extending the generic field to contain a primitive cube root of unity, let χ be a nontrivial character of C3. On the generic fibre of the cover, both invariant one-forms and H1⁢(𝒪) have characters 1⊕2⊕χ⊕χ−1. Therefore

rank⁡Q0=2,rank⁡Q1=4+1+1=6.

Since Q0 is torsion-free, coherent universal coefficients, (8.5), and h0⁢(ΩY1)=3 give

dimQ1⁢[3]=3−2=1,dimQ2⁢[3]=10−6−1=3.

It follows that (Q2)tors=(W/3)⊕3. Contraction with σ˜Y identifies this module with H2⁢(𝒴,T𝒴/W)tors.

Let 𝒴2=𝒴×WW2. The coefficient Bockstein

δ:H1⁢(Y,TY)⟶H2⁢(Y,TY)

for 0→TY→T𝒴2/W2→TY→0 factors as the universal-coefficient surjection onto H2⁢(𝒴,T𝒴/W)⁢[3], followed by its reduction into H2⁢(Y,TY). Its image is therefore the canonical reduction of H2⁢(𝒴,T𝒴/W)⁢[3]. This reduction is injective since the torsion has exponent three, so rank⁡δ=3.

Choose a first-order deformation with tangent ξ and glue it to the fixed lift 𝒴2 over

A=W2×kk⁢[ϵ]/ϵ2.

The map

A′=W2⁢[ϵ]/ϵ2⟶A

is an elementary extension with kernel k⋅3⁢ϵ. A lift over A′ would extend ξ to a square-zero deformation of 𝒴2, hence lift its class to H1⁢(𝒴2,T𝒴2/W2). Its complete obstruction is exactly δ⁢(ξ). Moreover, a reduction isomorphism over A identifies the W2-restriction of any proposed lift with 𝒴2, so the necessity just proved applies to every lift in the absolute deformation functor, including lifts without a separately chosen W2-marking. Every vector in the three-dimensional image of δ is therefore realized as an actual geometric obstruction on this extension.

Finally, by contraction and the ordinary Cartan formula, the tangent object of the pair is identified with R⁢Γ⁢(Y,F1⁢ΩY∙)⁢[2]; this identification uses only changes of local coordinates and forms. Hodge degeneration and Lemma 5.4 make its tangent map

ℍ2⁢(Y,F1⁢ΩY∙)⟶H1⁢(Y,ΩY1)≃H1⁢(Y,TY)

surjective. Choose a pair tangent above ξ and glue it to the fixed W2 pair. If δ⁢(ξ)≠0, this pair cannot lift over A′ because its underlying variety cannot lift. This proves both obstructedness assertions using ordinary coherent deformation theory. ∎

Proposition 8.5 gives an affirmative answer in characteristic three to the existence question raised in [8, Remark 1.1]. Indeed, σY∧σY trivializes ωY, the Hodge–de Rham spectral sequence degenerates at E1 by the Frobenius-splitting argument above, and DefY/W is obstructed. Thus Hodge–de Rham degeneration alone does not imply the conclusion of [8, Theorem A], even for a smooth projective symplectic variety admitting a projective symplectic W-lift. For a related bielliptic example in characteristic two addressing the same question, see [34]. Here Calabi–Yau is understood in the sense of [8], namely triviality of the canonical bundle; the fourfold does not satisfy the additional vanishing Hi⁢(Y,𝒪Y)=0 for 0<i<4, as (8.4) shows.

8.5. An unobstructed sixfold with nonzero torsion

The second construction changes the integral representation on the cover while retaining its ordinary elliptic factors. Define

ℬ=ℰ3/⟨(P,P,P)⟩,Q=[(P,0,0)]∈ℬ⁢(W),𝒜=ℬ2.

Let ρ cyclically permute the three coordinates. It preserves the diagonal subgroup and hence acts on ℬ. Set

g⁢(b1,b2)=(ρ⁢b1+Q,ρ⁢b2).

The norm Q+ρ⁢Q+ρ2⁢Q is the class of (P,P,P), which is zero; thus g3=1. The sum map s:ℬ→ℰ descends because 3⁢P=0, and it satisfies s⁢(ρ⁢b+Q)=s⁢(b)+P. Both nonidentity powers of g therefore act freely. Put 𝒳=𝒜/⟨g⟩ and X=𝒳k. The abelian variety B=ℬk is ordinary because it is isogenous to E3.

Let αj,i, with j=1,2 and i=0,1,2, be the invariant differentials on the two copies of ℬ which pull back to the coordinate differentials on ℰ3. They exist because the isogeny is étale. The form

σ˜X=∑i=02α1,i∧α2,i

is invariant, closed and nondegenerate, and it descends to 𝒳. We write σX for its reduction.

Proposition 8.6.

The smooth projective symplectic sixfold (X,σX) satisfies

Hcrys2⁢(X/W)tors≃(W/3)⊕2,Hcrys3⁢(X/W)tors≃(W/3)⊕2.

Crystalline Frobenius is bijective on the latter module. Both DefX/W and Def(X,σX)/W are unobstructed over W, and forgetting the form is formally smooth on discrete Artinian bases.

Proof.

We give the Hodge and integral representation calculations needed for the conclusion. Let I=ker⁡(ℤ3⁢[C3]→ℤ3) be the augmentation lattice and put 𝐞=(1,1,1). On étale Tate modules the isogeny E3→B adjoins 𝐞/3 to ℤ3⊕3. Taking dual lattices gives, equivariantly,

He´⁢t1⁢(B,ℤ3)={(xi)∈ℤ3⊕3:∑ixi∈3⁢ℤ3}=I⊕ℤ3⁢𝐞.

The last decomposition follows by subtracting (∑ixi)/3 times 𝐞. The sum map s:B→E has the section x↦[(x,0,0)], so B≃E×B0 for B0=ker⁡s. Pullback from E is the primitive summand ℤ3⁢𝐞, and therefore He´⁢t1⁢(B0,ℤ3)=I as a C3-lattice.

For an ordinary abelian variety, its unit-root crystalline summand is He´⁢t1⁢(−,ℤ3)⊗ℤ3W; reduction of this summand maps isomorphically to H1⁢(𝒪). These identifications are compatible with endomorphisms [25, 18]. Since I/3⁢I=J2, we obtain

H1⁢(B,𝒪B)=J1⊕J2,H1⁢(B0,𝒪B0)=J2.

The map X→CE induced by s becomes a product with fibre B0×B after the étale base change E→CE. The descent action can include translations, but these act trivially on the coherent cohomology of abelian varieties, so the degree-one fibre representation is Nfib=J1⊕J2⊕2. For a representation N associated to this elliptic torsor, the descended bundle has h0=dimNC3 and h1=h0: the first equality follows by descent of global sections, and the second by Riemann–Roch for a degree-zero bundle. Using J2⊗J2=J3⊕J1 and ∧2J2=J1, one finds

dimNfibC3=3,dim(∧2Nfib)C3=6.

Coherent Leray therefore gives h1⁢(𝒪X)=4 and h2⁢(𝒪X)=9.

The representation on one-forms of A=B2 is k⁢[C3]⊕2. Its exterior square is k⁢[C3]⊕5. For the covering map π:A→X, regular representation descent gives

ΩX1≃(π∗⁢𝒪A)⊕2,ΩX2≃(π∗⁢𝒪A)⊕5.

Consequently

h0⁢(ΩX1)=2,h1⁢(ΩX1)=12,h0⁢(ΩX2)=5.

Rationally, Hcrys1⁢(A/W)⁢[1/3] is a sum of four regular representations. The invariant part in exterior degree two has dimension (42)+4⋅4=22; hence b1⁢(X)=4 and b2⁢(X)=22. Hodge degeneration gives de Rham dimensions 6 and 9+12+5=26 in degrees one and two. As above, t1⁢(X)=0, and the dimension formula gives t2⁢(X)=t3⁢(X)=2. Since the trace argument shows that the torsion is killed by three, this proves the asserted crystalline modules.

It remains to determine Frobenius on degree-three torsion. The ordinary connected–étale decomposition gives a C3-stable integral decomposition

Hcrys1⁢(A/W)=U⊕V,ΦU⁢ bijective,ΦV=3⁢ΨV⁢ with ⁢ΨV⁢ bijective.

Translations act trivially on cohomology, and the linear action preserves this canonical decomposition. The ordinary Hodge filtration identifies V/3⁢V with H0⁢(A,ΩA1). Lifting two regular generators of this last representation gives a map W⁢[C3]⊕2→V that is an isomorphism modulo three, so Nakayama and equality of ranks show that V≃W⁢[C3]⊕2.

Crystalline étale descent along A→X gives the Frobenius-compatible Cartan–Leray spectral sequence

E2a,b=Ha⁢(C3,Hcrysb⁢(A/W))⟹Hcrysa+b⁢(X/W).

One can equally construct it from the étale Čech nerve of the lifted torsor and crystalline–de Rham comparison. Frobenius compatibility comes from crystalline Frobenius on that nerve, so no Frobenius lift of 𝒜 is needed. The crystalline cohomology algebra of the abelian variety is the exterior algebra on U⊕V. A diagonal tensor product of a regular module with any W-free module is induced. Moreover, ∧2V≃W⁢[C3]⊕5: unordered pairs in each regular basis form free orbits, and the cross tensor product is a sum of three regular modules. Thus V, U⊗V, and ∧2V have no positive group cohomology.

It follows that the positive-column terms in total degree three are precisely

H1⁢(C3,∧2U),H2⁢(C3,U),H3⁢(C3,W).

They have finite length and bijective Frobenius. Every Frobenius-stable submodule and quotient of a finite-length Frobenius-bijective module remains Frobenius-bijective: injectivity on a stable submodule and equality of lengths imply surjectivity there, and the quotient follows. Therefore their E∞ subquotients remain Frobenius-bijective. The positive-column filtration step of Hcrys3⁢(X/W) has finite length, while its quotient E∞0,3 is a submodule of the free module Hcrys3⁢(A/W)C3, so this filtration step is exactly the torsion subgroup. Since bijectivity of Frobenius is preserved under extensions of finite-length modules, Frobenius is bijective on T.

Now Lemma 8.4 gives S∩F1⁢HdR3⁢(X/k)=0. Hodge degeneration and Corollary 8.3 prove unobstructedness of both classical deformation functors and formal smoothness of forgetting the form. ∎

8.6. Comparison with Serre–Tate theory

Brantner–Taelman’s ordinary theory gives a further interpretation of these examples, beyond their crystalline splitting criterion [8, Theorem 7.16]. Their Serre–Tate theorem applies to a smooth proper geometrically irreducible Bloch–Kato 2-ordinary variety Z of dimension d with trivial canonical bundle when He´⁢td⁢(Z,ℤp) is torsion-free [8, Theorem 8.28]. It describes the deformation problem by a formal group of multiplicative type, which can contain nonsmooth factors μpn; ordinarity alone therefore does not imply unobstructedness [8, Remark 1.5]. Over our algebraically closed field, their description of the classical deformation functor involves

MZ=Hd−1⁢(Z,W⁢ΩZ1)F=1,

with coefficient module He´⁢td⁢(Z,ℤp) [8, Remark 8.30]; when the latter is free of rank one, a choice of generator identifies MZ with the character module. Here F denotes de Rham–Witt Frobenius.

We first check the top-degree hypothesis for both examples: let Z be either Y or X, and put d=dimZ. The invariant volume forms on the abelian covers descend, so ωZ≃𝒪Z and hd⁢(Z,𝒪Z)=1. The determinant of the action on the degree-one 3-adic étale cohomology of each cover is one, and rational descent consequently gives

dimℚ3He´⁢td⁢(Z,ℚ3)=1.

Bloch–Kato 1-ordinarity gives

He´⁢td⁢(Z,𝔽3)⊗𝔽3k≃Hd⁢(Z,𝒪Z)

by [8, Proposition 8.11]. The universal-coefficient injection of He´⁢td⁢(Z,ℤ3)/3 into He´⁢td⁢(Z,𝔽3) then shows that He´⁢td⁢(Z,ℤ3) is free of rank one. Thus their Serre–Tate theorem applies to both examples; the fourfold’s obstructedness is compatible with the possible nonsmooth finite factors.

For the sixfold, we can also verify smoothness of the Serre–Tate group directly. Retain the notation A, U, and V from the proof of Proposition 8.6. On the ordinary abelian cover, the integral Hodge–Witt decomposition and the exterior algebra description of crystalline cohomology give

Hb⁢(A,W⁢ΩA1)≃⋀bU⊗WV,0≤b≤6,

compatibly with the C3-action and with F=⋀bΦU⊗ΨV [18, II, Remark 4.10(a), (4.10.1), and § 7.1, (7.1.1)]. In particular, F is bijective on each of these modules. Since V≃W⁢[C3]⊕2, every displayed module is induced and has vanishing positive group cohomology. Étale descent for de Rham–Witt cohomology therefore gives

H5⁢(X,W⁢ΩX1)≃(⋀5U⊗WV)C3.

This is a free W-module of rank 2⁢(65)=12, with bijective F. Taking F-fixed elements over algebraically closed k gives MX≃ℤ3⊕12 [18, II, Lemma 5.3]. By [8, Theorem 8.28 and Remark 8.30], the unpolarized deformation functor DefX/W is consequently a formal torus, noncanonically isomorphic to 𝐆ˆm,W12, and is unobstructed.

Thus unpolarized unobstructedness of the sixfold is also a consequence of Brantner–Taelman’s full ordinary theory. The construction supplies explicit crystalline torsion and illustrates the selective criterion; our argument also proves relative lifting of the specified symplectic form, through formal smoothness of Def(X,σX)/W→DefX/W.

References

  • [1] P. Achinger and J. Suh (2023) Some refinements of the Deligne–Illusie theorem. Algebra & Number Theory 17 (2), pp. 465–496. Note: arXiv:2003.09857v2 External Links: Document Cited by: §1.5, §5.4.
  • [2] A. Beilinson (2012) p-adic periods and derived de Rham cohomology. Journal of the American Mathematical Society 25 (3), pp. 715–738. External Links: Document, Link Cited by: §3.1.
  • [3] P. Berthelot (1974) Cohomologie cristalline des schémas de caractéristique p>0. Lecture Notes in Mathematics, Vol. 407, Springer-Verlag, Berlin. External Links: Document Cited by: §6.1.
  • [4] B. Bhatt, W. Ho, Z. Patakfalvi, and C. Schnell (2013) Moduli of products of stable varieties. Compositio Mathematica 149, pp. 2036–2070. External Links: Document Cited by: §6.3.
  • [5] B. Bhatt (2012) p-adic derived de Rham cohomology. Note: arXiv:1204.6560v1 External Links: Link Cited by: §1.5, §3.1, §3.1, §3.1.
  • [6] F. A. Bogomolov (1978) Hamiltonian Kählerian manifolds. Doklady Akademii Nauk SSSR 243 (5), pp. 1101–1104. Note: In Russian External Links: Link Cited by: §1.7.
  • [7] L. Brantner and A. Mathew (2019) Deformation theory and partition Lie algebras. Note: arXiv:1904.07352 External Links: Link Cited by: §2.3.
  • [8] L. Brantner and L. Taelman (2025) Deformations and lifts of Calabi–Yau varieties in characteristic p. Note: arXiv:2407.09256v3, 25 May 2025 External Links: Link Cited by: §1.4, §1.5, §1.7, §1, §1, §2.2, §2.2, §2.2, §2.3, §2.3, §2.3, §2.4, §2.4, §2.4, §2.5, §2.5, §2.5, §2.6, §2.6, §2.6, §2.6, §2.6, Proposition 2.3, §2, §3.1, §3.1, §3.2, §3.5, Remark 3.14, §4.2, §4.2, §5.3, §8.3, §8.4, §8.6, §8.6, §8.6.
  • [9] S. Cynk and D. van Straten (2009) Small resolutions and non-liftable Calabi–Yau threefolds. Manuscripta Mathematica 130 (2), pp. 233–249. External Links: Document Cited by: §1.
  • [10] P. Deligne and L. Illusie (1987) Relèvements modulo p2 et décomposition du complexe de de Rham. Inventiones Mathematicae 89 (2), pp. 247–270. External Links: Document Cited by: §6.2.
  • [11] P. Deligne (1981) Relèvement des surfaces K3 en caractéristique nulle. In Surfaces algébriques, Lecture Notes in Mathematics, Vol. 868, pp. 58–79. Note: Rédigé par Luc Illusie Cited by: Example 6.5.
  • [12] L. Fu and Z. Li (2021) Supersingular irreducible symplectic varieties. In Rationality of Varieties, Progress in Mathematics, Vol. 342, pp. 147–200. External Links: Document Cited by: §6.2, Example 6.7, Example 6.8.
  • [13] M. Gros (1985) Classes de Chern et classes de cycles en cohomologie de Hodge–Witt logarithmique. Mémoires de la Société Mathématique de France (N.S.) 21, pp. 1–87. External Links: Document Cited by: Example 6.7.
  • [14] A. Grothendieck (1961) Éléments de géométrie algébrique. III. étude cohomologique des faisceaux cohérents. I. Publications Mathématiques de l’IHÉS 11, pp. 5–167. Note: Avec la collaboration de Jean Dieudonné Cited by: §7.1, §7.3, §7.3, §7.3.
  • [15] M. Hirokado (1999) A non-liftable Calabi–Yau threefold in characteristic 3. Tohoku Mathematical Journal (2) 51 (4), pp. 479–487. External Links: Document Cited by: §1.
  • [16] D. Iacono and M. Manetti (2010) An algebraic proof of Bogomolov–Tian–Todorov theorem. In Deformation Spaces: Perspectives on Algebro-Geometric Moduli, H. Abbaspour, M. Marcolli, and T. Tradler (Eds.), Aspects of Mathematics, pp. 113–133. Note: Theorem numbers refer to arXiv:0902.0732v2 External Links: Document Cited by: §1.7.
  • [17] L. Illusie (1972) Complexe cotangent et déformations II. Lecture Notes in Mathematics, Vol. 283, Springer-Verlag, Berlin–New York. External Links: Document Cited by: §3.1.
  • [18] L. Illusie (1979) Complexe de de Rham–Witt et cohomologie cristalline. Annales Scientifiques de l’École Normale Supérieure (4) 12, pp. 501–661. External Links: Link Cited by: Example 6.4, §8.5, §8.6.
  • [19] D. Kaledin and M. Verbitsky (2002) Period map for non-compact holomorphically symplectic manifolds. Geometric and Functional Analysis 12 (6), pp. 1265–1295. External Links: Link Cited by: §1.7, §1, §4.1.
  • [20] H. Liu (2025) Lifting line bundles to schemes over p-adic base. Note: Preprint, 18 November 2025 External Links: Link Cited by: §7.3, §7.3.
  • [21] J. Lurie (2009) Higher topos theory. Annals of Mathematics Studies, Vol. 170, Princeton University Press. Note: Author version dated 9 April 2017 External Links: Link Cited by: §2.1, §2.6, §3.2.
  • [22] J. Lurie (2017) Higher algebra. Note: Version dated 18 September 2017 External Links: Link Cited by: §2.1.
  • [23] Z. Mao (2024) Revisiting derived crystalline cohomology. Bulletin de la Société Mathématique de France 152 (4), pp. 659–784. External Links: Document, Link Cited by: §1.5, §3.1, §3.1.
  • [24] Y. Min (2020) Integral p-adic Hodge theory of formal schemes in low ramification. Note: Version 2, 9 December 2020 External Links: 2004.04436, Link Cited by: §6.2.
  • [25] T. Oda (1969) The first de Rham cohomology group and Dieudonné modules. Annales Scientifiques de l’École Normale Supérieure (4) 2, pp. 63–135. External Links: Link Cited by: §6.3, Example 6.4, §8.5.
  • [26] A. Ogus (1979) Supersingular K3 crystals. Astérisque 64, pp. 3–86. External Links: Link Cited by: §7.2.
  • [27] A. Petrov (2025) Decomposition of de Rham complex for quasi-F-split varieties. Note: Version 1, 19 February 2025 External Links: 2502.13356, Link Cited by: §8.3.
  • [28] T. Takamatsu (2026) On the witt liftability of quasi-f-split calabi–yau varieties. External Links: 2610.38427, Link Cited by: footnote 1.
  • [29] The Stacks Project Authors (2026) Stacks Project. External Links: Link Cited by: §3.4, §3.4, §3.4, §3.4, §4.2, §4.2, §5.4, §5.4, §5.5, §7.1, §7.2.
  • [30] G. Tian (1987) Smoothness of the universal deformation space of compact Calabi–Yau manifolds and its Peterson–Weil metric. In Mathematical Aspects of String Theory, Advanced Series in Mathematical Physics, Vol. 1, pp. 629–646. External Links: Document Cited by: §1.7.
  • [31] A. N. Todorov (1989) The Weil–Petersson geometry of the moduli space of S⁢U⁢(n≥3) (Calabi–Yau) manifolds. I. Communications in Mathematical Physics 126 (2), pp. 325–346. External Links: Document Cited by: §1.7.
  • [32] B. Totaro (2020) The integral cohomology of the Hilbert scheme of points on a surface. Forum of Mathematics, Sigma 8. External Links: Document Cited by: §6.2.
  • [33] J. Voight and D. Zureick-Brown (2022) The canonical ring of a stacky curve. Memoirs of the American Mathematical Society 277 (1362). Note: Corrected author version, June 2023 External Links: Link Cited by: §6.3, Example 6.11.
  • [34] H. Zock A variety with trivial canonical bundle, degenerate Hodge-to-de Rham spectral sequence, and obstructed deformations. Note: Undated note External Links: Link Cited by: §1.4, §8.4.