betaFeedback

WorksMathematics

Relative Du Bois complexes over higher dimensional bases

Abstract

We extend the notion of the right relative Du Bois complex to a smooth base of arbitrary dimension. We further show that this complex satisfies base change in the sense of Kovács–Taji, for flat families with strict higher Du Bois fibers, extending the result by Ning over a smooth curve.

Introduction

For a smooth morphism f:X→Sf:X\to S of smooth complex varieties, the relative de Rham complex ΩX/S∙\Omega^\bullet_{X/S} is characterized by the exact sequence:

0⟶f∗ΩS1⟶ΩX1⟶ΩX/S1⟶0.0\longrightarrow f^*\Omega^1_S\longrightarrow\Omega^1_X\longrightarrow\Omega^1_{X/S}\longrightarrow0.

Exterior powers of this sequence give rise to a natural filtration of ΩXp\Omega^p_X with graded pieces f∗ΩSq⊗ΩX/Sp−qf^*\Omega^q_S\otimes\Omega^{p-q}_{X/S}. A key property is that ΩX/Sp\Omega^p_{X/S} satisfies base change, that is, for ff smooth,

ΩX/Sp⊗OXLOXs≃qisΩXsp.\Omega^p_{X/S}\mathbin{\mathop{\otimes}^{\mathrm L}_{\mathcal O_X}}\mathcal O_{X_s}\simeq_{\mathrm{qis}}\Omega^p_{X_s}.

For XX singular, Du Bois [1] replaced the de Rham complex ΩXq\Omega^q_X with the filtered complex Ω‾Xq\underline{\Omega}^q_X in the derived category of XX. It agrees with ΩXq\Omega^q_X when XX is smooth, and, from a cohomological point of view, this complex behaves very much like the usual de Rham complex. It is therefore natural to ask whether there is a way to produce an object analogous to the relative de Rham complex of a smooth family, equipped with a similar base change property.

Kovács constructed a (left) relative Du Bois complex, which we will denote ΩX/Sp,−\Omega^{p,-}_{X/S}, over a smooth curve in [6] and over a smooth base of arbitrary dimension in [8, 7]. It shows many applications to families of singular varieties; see [6, 7, 10, 11]. One important question of Kovács–Taji [12] asks whether the relative Du Bois complex satisfies base change, that is, does

ΩX/Sp,−⊗OXLOXs≃qisΩXsp\Omega^{p,-}_{X/S}\mathbin{\mathop{\otimes}^{\mathrm L}_{\mathcal O_X}}\mathcal O_{X_s}\simeq_{\mathrm{qis}}\Omega^p_{X_s}

hold? It turns out that, in general, base change for the left relative Du Bois complex fails in every degree [4, 13]. Ning proposed an alternative construction, dubbed the right relative Du Bois complex ΩX/Cp,+\Omega^{p,+}_{X/C} for f:X→Cf:X\to C, where CC is a smooth curve. It satisfies base change [13 Theorem 1.1] for flat families with strict higher Du Bois singularities (cf. Definition 2.2). As shown in loc. cit., base change properties have geometric applications such as deformation invariance of higher Du Bois singularities.

The purpose of this paper is to extend the notion of the right relative Du Bois complex ΩX/Sp,+\Omega^{p,+}_{X/S} to a smooth base of arbitrary dimension while preserving base change properties, see Definition 4.1 for the precise definition. We construct the following.

Theorem 1.1. (== Theorem 3.12) Let f:X→Sf:X\to S be a morphism of complex algebraic varieties with dim⁡X=n\dim X=n, dim⁡S=k\dim S=k and SS smooth. Then there exist objects ΩX/Sp,+,ΩX/Sp,−\Omega^{p,+}_{X/S},\Omega^{p,-}_{X/S} in Dcohb(X)D^b_{\mathrm{coh}}(X) for all p∈Zp\in\mathbb Z with the following properties:

(1) ΩX/Sp,+≃qis0\Omega^{p,+}_{X/S}\simeq_{\mathrm{qis}}0 for p<0p<0, ΩX/Sp,−≃qis0\Omega^{p,-}_{X/S}\simeq_{\mathrm{qis}}0 for p>n−kp>n-k.

:::`

(2) There exist hyperfiltrations Fq,+ΩXp\mathrm{F}^{q,+}\Omega_X^p, Fq,−ΩXp\mathrm{F}^{q,-}\Omega_X^p of ΩXp\Omega_X^p with 0≤q≤k+10 \le q \le k+1, such that F0,+ΩXp≃qisF0,−ΩXp≃qisΩXp\mathrm{F}^{0,+}\Omega_X^p \simeq_{\mathrm{qis}} \mathrm{F}^{0,-}\Omega_X^p \simeq_{\mathrm{qis}} \Omega_X^p, Fk+1,+ΩXp≃qisFk+1,−ΩXp≃qis0\mathrm{F}^{k+1,+}\Omega_X^p \simeq_{\mathrm{qis}} \mathrm{F}^{k+1,-}\Omega_X^p \simeq_{\mathrm{qis}} 0, whose associated graded complexes, with 0≤q≤k0 \le q \le k are, Gr⁡q,+ΩXp≃qisΩX/Sp−q,+⊗f∗ΩSq\operatorname{Gr}^{q,+}\Omega_X^p \simeq_{\mathrm{qis}} \Omega_{X/S}^{p-q,+} \otimes f^*\Omega_S^q, Gr⁡q,−ΩXp≃qisΩX/Sp−q,−⊗f∗ΩSq\operatorname{Gr}^{q,-}\Omega_X^p \simeq_{\mathrm{qis}} \Omega_{X/S}^{p-q,-} \otimes f^*\Omega_S^q.

(3) Fq,+ΩXp\mathrm{F}^{q,+}\Omega_X^p and Fq,−ΩXp\mathrm{F}^{q,-}\Omega_X^p are functorial.

(4) If ff is smooth, then ΩX/Sp,+≃qisΩX/Sp,−≃qisΩX/Sp\Omega_{X/S}^{p,+} \simeq_{\mathrm{qis}} \Omega_{X/S}^{p,-} \simeq_{\mathrm{qis}} \Omega_{X/S}^p.

(5) Let {νi:Ui↪X}\{\nu_i: U_i \hookrightarrow X\} be a finite open cover of XX. Then ΩX/Sp,±≃qisR(ν∙)∗ΩU∙/Sp,±\Omega_{X/S}^{p,\pm} \simeq_{\mathrm{qis}} R(\nu_\bullet)_*\Omega_{U_\bullet/S}^{p,\pm}.

Note that the statement for the left relative Du Bois complex ΩX/Sp,−\Omega_{X/S}^{p,-} follows from [7 Theorem 3.4]. Our construction for the right relative complex agrees with the one constructed in [13] over a smooth curve, see Proposition 3.14.

We also show base change properties for this generalized right relative Du Bois complex for fibers that are strict higher Du Bois (cf. Definition 2.2).

Theorem 1.2. (= Theorem 4.7) Let f:X→Sf : X \to S be a flat morphism of algebraic varieties with dim⁡X=n\dim X = n, dim⁡S=k\dim S = k and SS smooth. Let s∈Ss \in S be a closed point and suppose the fiber XsX_s has strict-mm-Du Bois singularities. Then XsX_s satisfies right-mm-base change (cf. Definition 4.1) for the relative Du Bois complex, that is, for all 0≤p≤m0 \le p \le m,

ΩX/Sp,+⊗LOXOXs≃qisΩXsp.\Omega_{X/S}^{p,+} \mathbin{\mathop{\otimes}^{\mathbf{L}}}_{\mathcal{O}_X}\mathcal{O}_{X_s} \simeq_{\mathrm{qis}} \Omega_{X_s}^p.

We remark that a recent work of Dirks in [2] constructs an alternative relative Du Bois complex using the theory of mixed Hodge modules, over smooth bases of arbitrary dimensions. It behaves similarly to the right relative Du Bois complex, and Dirks asks [2 §5, Questions 3, 4] whether the existing constructions agree. This result will be discussed in an upcoming work [14].

Paper Outline. Section 2 reviews preliminary material essential to the construction. In Section 3, we construct the right relative Du Bois complex ΩX/Sp,+\Omega_{X/S}^{p,+} directly for a singular total space, refining the methods of [7]. We verify that both relative Du Bois complexes ΩX/Sp,±\Omega_{X/S}^{p,\pm} agree with the existing definitions in the literature [6, 13] over a smooth curve. Section 4 shows the base change property of ΩX/Sp,+\Omega_{X/S}^{p,+} for flat families with strict higher Du Bois fibers.

Acknowledgments. The authors would like to thank Sándor Kovács and Brad Dirks for helpful discussions.

AI Disclosure. No generative AI is used for the mathematical contents of this paper.

Preliminaries

The Du Bois Complex

The Du Bois complex was originally constructed in [1] using simplicial hyperresolutions. In this paper, by hyperresolution we will mean cubical hyperresolutions in the sense of [3]. We refer the reader there for detail, see also [5 §10.6].

Theorem 2.1 ([1, 3]). Let XX be a reduced scheme of finite type over C\mathbb{C}. Then there exists an object Ω‾X∙\underline{\Omega}_X^\bullet in Dfilt,cohb(X)D_{\mathrm{filt,coh}}^b(X), with associated graded Ω‾Xp=Gr⁡filtpΩ‾X∙[p]\underline{\Omega}_X^p = \operatorname{Gr}_{\mathrm{filt}}^p \underline{\Omega}_X^\bullet[p] satisfying the following properties.

(1) If π∙:X∙→X\pi_\bullet: X_\bullet\to X is any hyperresolution of XX, then

Ω‾X∙≃qisRπ∙∗Ω‾X∙∙.\underline{\Omega}_X^\bullet\simeq_{\mathrm{qis}} R\pi_{\bullet*}\underline{\Omega}_{X_\bullet}^\bullet.

(2) The construction is functorial, i.e. if ϕ:Y→X\phi: Y \to X is a morphism of reduced schemes of finite type over C\mathbb{C}, then there exists a natural map ϕ∗:Ω‾X∙→Rϕ∗Ω‾Y∙\phi^* : \underline{\Omega}_X^\bullet\to R\phi_*\underline{\Omega}_Y^\bullet.

Furthermore, if ϕ\phi is proper, then ϕ∗\phi^* is a morphism in Dfilt,cohb(X)D_{\mathrm{filt,coh}}^b(X).

(3) Let ΩX∙\Omega_X^\bullet be the usual De Rham complex of Kähler differentials considered with the ‘filtration bête’. Then there exists a natural map of filtered complexes

ΩX∙⟶Ω‾X∙\Omega_X^\bullet\longrightarrow\underline{\Omega}_X^\bullet

and if XX is smooth, this map is a quasi-isomorphism.

(4) Let {νi:Ui↪X}\{\nu_i: U_i \hookrightarrow X\} be a finite open cover of XX, with ν:U∙→X\nu: U_\bullet\to X the associated map. Then

Ω‾X∙≃qisRν∗Ω‾U∙∙.\underline{\Omega}_X^\bullet\simeq_{\mathrm{qis}} R\nu_*\underline{\Omega}_{U_\bullet}^\bullet.

The definitions of higher Du Bois singularities have several distinct flavors outside of the local complete intersection setting. As observed in [13], the notion of strict-mm-Du Bois singularities appears to be the most suitable for base change type questions, so we only include this variant here. We refer the reader to [SVV23, Kov26, NN25] for detailed discussions on other notions.

Definition 2.2. Let XX be a reduced scheme of finite type over C\mathbb{C} and let m∈Nm \in\mathbb{N}. We say XX has strict-mm-Du Bois singularities if the natural map

Ω‾Xp→≃qisΩXp\underline{\Omega}_X^p \xrightarrow{\simeq_{\mathrm{qis}}} \Omega_X^p

is a quasi-isomorphism for each 0≤p≤m0 \leq p \leq m.

Hyperfiltrations

Let A\mathcal{A} be an abelian category and D(A)D(\mathcal{A}) its derived category. Let Γ:A→Ab\Gamma: \mathcal{A} \to\mathsf{Ab} be a left exact additive functor from A\mathcal{A} to the category of abelian groups.

Definition 2.3. [7 1.2.1] Let K∈Ob⁡(Db(A))K \in\operatorname{Ob}(D^b(\mathcal{A})) be a bounded complex. A hyperfiltration F\mathbb{F} of KK consists of a set of objects FqK∈Ob⁡(Db(A))\mathbb{F}^qK \in\operatorname{Ob}(D^b(\mathcal{A})) for q=l,…,k+1q=l,\ldots,k+1, where l,k∈Zl,k \in\mathbb{Z} and morphisms

ϕq∈Hom⁡Db(A)(Fq+1K,FqK)\phi_q \in\operatorname{Hom}_{D^b(\mathcal{A})}(\mathbb{F}^{q+1}K,\mathbb{F}^qK)

for q=l,…,kq=l,\ldots,k, with Fl≃qisK\mathbb{F}^l \simeq_{\mathrm{qis}} K and Fk+1≃qis0\mathbb{F}^{k+1} \simeq_{\mathrm{qis}} 0. The pp-th associated graded complex of a hyperfiltration F\mathbb{F} is

Gp:=Gr⁡FpK=Cone⁡(ϕp).\mathbb{G}^p := \operatorname{Gr}_{\mathbb{F}}^p K = \operatorname{Cone}(\phi_p).

This is the class of the mapping cone of a morphism in C(A)C(\mathcal{A}) whose class in D(A)D(\mathcal{A}) is ϕp\phi_p, so that Fp+1⟶Fp⟶Gp→+1\mathbb{F}^{p+1} \longrightarrow\mathbb{F}^p \longrightarrow\mathbb{G}^p \xrightarrow{+1} forms a distinguished triangle.

The right relative Du Bois complex

We extend the construction of the right relative Du Bois complex Ω‾X/Bp,+\underline{\Omega}^{p,+}_{X/B} from [13], where the base is a smooth curve, to a smooth base SS of arbitrary dimension. This is the ascending analogue of the construction of [7], in which the pieces Ω‾X/Sp\underline{\Omega}^p_{X/S} are built upward from p=0p=0 rather than downward from p=dim⁡X−dim⁡Sp=\dim X-\dim S.

Our approach further simplifies the constructions in loc. cit. by building Ω‾X/Sp\underline{\Omega}^p_{X/S} directly without assuming XX is smooth. The central observation is that the only data required from a hyperresolution is a wedge map on the Du Bois complex of XX. The following lemma provides the naturality that we need.

Lemma 3.1. Let f:X→Sf : X \to S be a morphism of algebraic varieties, with SS smooth. Then, for each p,q≥0p,q \geq0, there exists a functorial wedge map

∧p,q:Ω‾Xp⊗f−1OSf−1Ω‾Sq⟶Ω‾Xp+q.\wedge_{p,q}: \underline{\Omega}_X^p \otimes_{f^{-1}\mathcal{O}_S} f^{-1}\underline{\Omega}_S^q \longrightarrow\underline{\Omega}_X^{p+q}.

Furthermore, the following diagram commutes

Ω‾Xp⊗f∗Ω‾Sq⊗f∗Ω‾Sr→Ω‾Xp⊗f∗Ω‾Sq+r∧p,q↓↓∧p,q+rΩ‾Xp+q⊗f∗Ω‾Sr→∧p+q,rΩ‾Xp+q+r\begin{CD} \underline{\Omega}_X^p \otimes f^*\underline{\Omega}_S^q \otimes f^*\underline{\Omega}_S^r @>>> \underline{\Omega}_X^p \otimes f^*\underline{\Omega}_S^{q+r} \\ @V{\wedge_{p,q}}VV @VV{\wedge_{p,q+r}}V \\ \underline{\Omega}_X^{p+q} \otimes f^*\underline{\Omega}_S^r @>{\wedge_{p+q,r}}>> \underline{\Omega}_X^{p+q+r} \end{CD}

Proof. This is a simple extension of the observations in [6 §1], see also [13 Construction 3.1].

Fix a cubical hyperresolution π∙:X∙→X\pi_{\bullet}: X_{\bullet} \to X, so that we have a specific representative of ΩXp\Omega_X^p, namely

R(π∙)∗ΩX∙p≃qisΩXp.R(\pi_{\bullet})_*\Omega_{X_{\bullet}}^p \simeq_{\mathrm{qis}} \Omega_X^p.

Using the projection formula to identify forms pulled back from SS, we may define

∧p,q:R(π∙)∗ΩX∙p⊗f∗ΩSq→R(π∙)∗ΩX∙p+q.\wedge_{p,q}: R(\pi_{\bullet})_*\Omega_{X_{\bullet}}^p \otimes f^*\Omega_S^q \to R(\pi_{\bullet})_*\Omega_{X_{\bullet}}^{p+q}.

determined by

∧p,q((πα)∗ηα⊗ξ)=(πα)∗(ηα∧(πα)∗ξ).\wedge_{p,q}\bigl((\pi_\alpha)_*\eta_\alpha\otimes\xi\bigr)=(\pi_\alpha)_*\bigl(\eta_\alpha\wedge(\pi_\alpha)^*\xi\bigr).

where πα:Xα→X\pi_\alpha:X_\alpha\to X is a component of the hyperresolution chosen, and ηα,ξ\eta_\alpha,\xi are the appropriate differential forms. It is then clear that the diagram in the statement of the lemma commutes with the chosen representative.

We now only need to show that ∧p,q\wedge_{p,q} is independent of the hyperresolution chosen. Suppose π~∙:X~∙→X\widetilde{\pi}_{\bullet}:\widetilde{X}_{\bullet}\to X is another cubical hyperresolution and α:X∙→X~∙\alpha:X_{\bullet}\to\widetilde{X}_{\bullet} is a morphism between them. Then we have a commutative diagram

R(π∙)∗ΩX∙p⊗f∗ΩSq→∧p,qR(π∙)∗ΩX∙p+q≃qis↓↓≃qisR(π~∙)∗ΩX~∙p⊗f∗ΩSq→∧~p,qR(π~∙)∗ΩX~∙p+q\begin{CD} R(\pi_{\bullet})_*\Omega_{X_{\bullet}}^p\otimes f^*\Omega_S^q @>{\wedge_{p,q}}>> R(\pi_{\bullet})_*\Omega_{X_{\bullet}}^{p+q}\\ @V{\simeq_{\mathrm{qis}}}VV @VV{\simeq_{\mathrm{qis}}}V\\ R(\widetilde{\pi}_{\bullet})_*\Omega_{\widetilde{X}_{\bullet}}^p\otimes f^*\Omega_S^q @>{\widetilde{\wedge}_{p,q}}>> R(\widetilde{\pi}_{\bullet})_*\Omega_{\widetilde{X}_{\bullet}}^{p+q} \end{CD}

where the vertical arrows are quasi-isomorphisms induced by α\alpha, see [1 Lemme 2.1.4]. Therefore, the equivalence class of ∧p,q\wedge_{p,q} in D(X)D(X) is independent of the chosen hyperresolution, as desired. □

The following construction is an analogue to the induction presented in [7], built upward to match the construction presented in [13].

Construction 3.2. Let f:X→Sf:X\to S be a morphism of algebraic varieties, with SS smooth of dimension kk. With induction on i≥0i\geq0, for each fixed ii, our goal is to define auxiliary objects Fip,+\mathcal{F}_i^{p,+} with maps

μip,q:Fip,+⊗f∗ΩSq→Fip+q,+,\mu_i^{p,q}:\mathcal{F}_i^{p,+}\otimes f^*\Omega_S^q\to\mathcal{F}_i^{p+q,+},

for all p,q≥0p,q\geq0. We further require at each step that μip,q\mu_i^{p,q} is compatible with wedge maps from the base, meaning that the diagram

Fip,+⊗f∗ΩSp−r⊗f∗ΩSq→Fip,+⊗f∗ΩSp+q−rμip,r↓↓μir,p+q−rFip,+⊗f∗ΩSq→μip,qFip+q,+\begin{CD} \mathcal{F}_i^{p,+}\otimes f^*\Omega_S^{p-r}\otimes f^*\Omega_S^q @>>> \mathcal{F}_i^{p,+}\otimes f^*\Omega_S^{p+q-r}\\ @V{\mu_i^{p,r}}VV @VV{\mu_i^{r,p+q-r}}V\\ \mathcal{F}_i^{p,+}\otimes f^*\Omega_S^q @>{\mu_i^{p,q}}>> \mathcal{F}_i^{p+q,+} \end{CD}

commutes for all $p,q,r\geq 0.

We begin with the i=0i=0 case. Let p≥0p\geq0 and define the first term to be F0p,+:=ΩXp\mathcal{F}_0^{p,+}:=\Omega_X^p, with

μ0p,q:=∧p,q:ΩXp⊗f∗ΩSq→ΩXp+q.\mu_0^{p,q}:=\wedge_{p,q}:\Omega_X^p\otimes f^*\Omega_S^q\to\Omega_X^{p+q}.

Define F1p,+:=Cone⁡(μ00,p)\mathcal{F}_1^{p,+}:=\operatorname{Cone}(\mu_0^{0,p}), that is, it sits in the following distinguished triangle

F00,+⊗f∗ΩSp→μ00,pF0p,+→F1p,+→+1.(1)\mathcal{F}_0^{0,+}\otimes f^*\Omega_S^p \xrightarrow{\mu_0^{0,p}} \mathcal{F}_0^{p,+}\to\mathcal{F}_1^{p,+}\xrightarrow{+1}. \tag*{(1)}

For any q≥0q\geq0, we obtain a commutative diagram:

ΩX0⊗f∗ΩSp⊗f∗ΩSq→ΩX0⊗f∗ΩSp+q↓↓ΩXp⊗f∗ΩSq→μ0p,qΩXp+q↓↓F~1p,+⊗f∗ΩSq→μ1p,qF~1p+q,++1↓+1↓\begin{CD} \Omega_X^0 \otimes f^*\Omega_S^p \otimes f^*\Omega_S^q @>>> \Omega_X^0 \otimes f^*\Omega_S^{p+q} \\ @VVV @VVV \\ \Omega_X^p \otimes f^*\Omega_S^q @>{\mu_0^{p,q}}>> \Omega_X^{p+q} \\ @VVV @VVV \\ \widetilde{\mathscr{F}}_1^{p,+} \otimes f^*\Omega_S^q @>{\mu_1^{p,q}}>> \widetilde{\mathscr{F}}_1^{p+q,+} \\ @V{+1}VV @V{+1}VV \end{CD}

The two columns are distinguished triangles of the form 3.3 above, and the compatibility is given by Lemma 3.1. We see that it induces a map μ1p,q:F~1p,+⊗f∗ΩSq→F~1p+q,+\mu_1^{p,q}:\widetilde{\mathscr{F}}_1^{p,+}\otimes f^*\Omega_S^q\to\widetilde{\mathscr{F}}_1^{p+q,+} on the bottom. This gives the first step of the induction.

Suppose the inductive hypothesis holds for some i≥0i\ge0. In particular, there are maps

μip,q:F~ip,+⊗f∗ΩSq→F~ip+q,+\mu_i^{p,q}:\widetilde{\mathscr{F}}_i^{p,+}\otimes f^*\Omega_S^q\to\widetilde{\mathscr{F}}_i^{p+q,+}

for all p,q≥0p,q\ge0 that commutes with the wedge maps on the base. Define F~i+1p,+:=Cone⁡(μii,p−i)\widetilde{\mathscr{F}}_{i+1}^{p,+}:=\operatorname{Cone}(\mu_i^{i,p-i}), so that

F~ip,+⊗f∗ΩSp−i→μip−iF~ip,+⟶F~i+1p,+→+1(2)\widetilde{\mathscr{F}}_i^{p,+}\otimes f^*\Omega_S^{p-i}\xrightarrow{\mu_i^{p-i}} \widetilde{\mathscr{F}}_i^{p,+}\longrightarrow\widetilde{\mathscr{F}}_{i+1}^{p,+}\xrightarrow{+1} \tag*{(2)}

is a distinguished triangle. Note that for p<ip<i this gives F~ip,+≃qisF~i+1p,+\widetilde{\mathscr{F}}_i^{p,+}\simeq_{\mathrm{qis}}\widetilde{\mathscr{F}}_{i+1}^{p,+} since by convention ΩSp−i=0\Omega_S^{p-i}=0 for p<ip<i. Similar to before, this gives the diagram with exact columns of the form (2)

F~ip,+⊗f∗ΩSp−i⊗f∗ΩSq→F~ip,+⊗f∗ΩSp+q−i↓↓F~ip,+⊗f∗ΩSq→μip,qF~ip+q,+↓↓F~i+1p,+⊗f∗ΩSq→μi+1p,qF~i+1p+q,++1↓+1↓(3)\begin{CD} \widetilde{\mathscr{F}}_i^{p,+}\otimes f^*\Omega_S^{p-i}\otimes f^*\Omega_S^q @>>> \widetilde{\mathscr{F}}_i^{p,+}\otimes f^*\Omega_S^{p+q-i} \\ @VVV @VVV \\ \widetilde{\mathscr{F}}_i^{p,+}\otimes f^*\Omega_S^q @>{\mu_i^{p,q}}>> \widetilde{\mathscr{F}}_i^{p+q,+} \\ @VVV @VVV \\ \widetilde{\mathscr{F}}_{i+1}^{p,+}\otimes f^*\Omega_S^q @>{\mu_{i+1}^{p,q}}>> \widetilde{\mathscr{F}}_{i+1}^{p+q,+} \\ @V{+1}VV @V{+1}VV \tag*{(3)} \end{CD}

The top square commutes by the inductive hypothesis, therefore we obtain the induced bottom map to be μi+1p,q:F~i+1p,+⊗f∗ΩSq→F~i+1p+q,+\mu_{i+1}^{p,q}:\widetilde{\mathscr{F}}_{i+1}^{p,+}\otimes f^*\Omega_S^q\to\widetilde{\mathscr{F}}_{i+1}^{p+q,+}. To check the required compatibility, consider the following diagram:

Ω~ii,+⊗f∗ΩSp−i−1⊗f∗ΩSq⟶Ω~ii,+⊗f∗ΩS1⊗f∗ΩSp+q−i−1↓↓Ω~i+1i+1,+⊗f∗ΩSp−i−1⊗f∗ΩSq⟶Ω~i+1i+1,+⊗f∗ΩSp+q−i−1↓↓Ω~i+1i+1,+⊗f∗ΩSp−i−1⊗f∗ΩSq⇢Ω~i+1i+1,+⊗f∗ΩSp+q−i−1↓↓Ω~ii,+⊗f∗ΩSp−i⊗f∗ΩSq⟶Ω~ii,+⊗f∗ΩSp+q−i↓↓Ω~ip,+⊗f∗ΩSq⟶Ω~ip+q,+↓↓Ω~i+1p,+⊗f∗ΩSq⇢Ω~i+1p+q,+\begin{aligned} \widetilde{\Omega}^{i,+}_{i}\otimes f^{*}\Omega_{S}^{p-i-1}\otimes f^{*}\Omega_{S}^{q} &\longrightarrow\widetilde{\Omega}^{i,+}_{i}\otimes f^{*}\Omega_{S}^{1}\otimes f^{*}\Omega_{S}^{p+q-i-1}\\ \downarrow&& \downarrow\\ \widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p-i-1}\otimes f^{*}\Omega_{S}^{q} &\longrightarrow\widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p+q-i-1}\\ \downarrow&& \downarrow\\ \widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p-i-1}\otimes f^{*}\Omega_{S}^{q} &\dashrightarrow\widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p+q-i-1}\\ \downarrow&& \downarrow\\ \widetilde{\Omega}^{i,+}_{i}\otimes f^{*}\Omega_{S}^{p-i}\otimes f^{*}\Omega_{S}^{q} &\longrightarrow\widetilde{\Omega}^{i,+}_{i}\otimes f^{*}\Omega_{S}^{p+q-i}\\ \downarrow&& \downarrow\\ \widetilde{\Omega}^{p,+}_{i}\otimes f^{*}\Omega_{S}^{q} &\longrightarrow\widetilde{\Omega}^{p+q,+}_{i}\\ \downarrow&& \downarrow\\ \widetilde{\Omega}^{p,+}_{i+1}\otimes f^{*}\Omega_{S}^{q} &\dashrightarrow\widetilde{\Omega}^{p+q,+}_{i+1} \end{aligned}

The vertical faces commute as they are all diagrams of the form (3). The top square commutes by associativity of the wedge product, and the middle square commutes by the inductive hypothesis. Therefore, the bottom square is commutative on any chosen representative. For clarity, the bottom square reads:

Ω~i+1i+1,+⊗f∗ΩSp−i−1⊗f∗ΩSq⟶Ω~i+1i+1,+⊗f∗ΩSp+q−i−1↓μi+1i+1,p−i−1↓μi+1i+1,p+q−i−1Ω~i+1p,+⊗f∗ΩSq→μi+1p,qΩ~i+1p+q,+\begin{aligned} \widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p-i-1}\otimes f^{*}\Omega_{S}^{q} &\longrightarrow\widetilde{\Omega}^{i+1,+}_{i+1}\otimes f^{*}\Omega_{S}^{p+q-i-1}\\ \downarrow^{\mu^{i+1,p-i-1}_{i+1}} && \downarrow^{\mu^{i+1,p+q-i-1}_{i+1}}\\ \widetilde{\Omega}^{p,+}_{i+1}\otimes f^{*}\Omega_{S}^{q} &\xrightarrow{\mu^{p,q}_{i+1}} \widetilde{\Omega}^{p+q,+}_{i+1} \end{aligned}

This is precisely the compatibility required for the inductive step. Therefore, we conclude the construction of Ω~ip,+\widetilde{\Omega}^{p,+}_{i} and the necessary accompanying maps μip,q\mu^{p,q}_{i} for all i,p,q≥0i,p,q\geq0. We summarize this information as:

Definition 3.6. Following Construction 3.2, for p,i≥0p,i\geq0, we define

Ω~i+1p,+:=Cone⁡(Ω~ii,+⊗f∗ΩSp−i→μii,p−iΩ~ip,+).\widetilde{\Omega}^{p,+}_{i+1}:=\operatorname{Cone}\left(\widetilde{\Omega}^{i,+}_{i}\otimes f^{*}\Omega_{S}^{p-i}\xrightarrow{\mu^{i,p-i}_{i}}\widetilde{\Omega}^{p,+}_{i}\right).

By convention, we put Ω~ip,+=0\widetilde{\Omega}^{p,+}_{i}=0 if p<0p<0. Note that Ω~0p,+=ΩXp\widetilde{\Omega}^{p,+}_{0}=\Omega^{p}_{X}. Furthermore,

0≃qisΩ~p+1p,+≃qisΩ~ip,+0\simeq_{\mathrm{qis}}\widetilde{\Omega}^{p,+}_{p+1}\simeq_{\mathrm{qis}}\widetilde{\Omega}^{p,+}_{i}

for all i>pi>p.

This allows us to define the right relative Du Bois complex.

Definition 3.7. Let f:X→Sf:X\to S be a morphism of varieties with SS smooth. We define ΩX/Sp,+∈Ob⁡(D(X))\Omega^{p,+}_{X/S}\in\operatorname{Ob}(D(X)), called the right relative Du Bois complex of f:X→Sf:X\to S, as follows. Put

ΩX/Sp,+={0p<0,Ω~pp,+p≥0.\Omega^{p,+}_{X/S}= \begin{cases} 0 & p<0,\\ \widetilde{\Omega}^{p,+}_{p} & p\geq0. \end{cases}

In particular, the triangle in (2) reads

ΩX/Si,+⊗f∗ΩSp−i⟶Ω~ip,+⟶Ω~i+1p,+→+1.\Omega^{i,+}_{X/S}\otimes f^{*}\Omega_{S}^{p-i}\longrightarrow\widetilde{\Omega}^{p,+}_{i}\longrightarrow\widetilde{\Omega}^{p,+}_{i+1}\xrightarrow{+1}.

We use the objects defined in Definition 3.6 to construct a hyperfiltration of ΩXp\Omega^{p}_{X}.

Construction 3.8. As in Construction 3.2, let f:X→Sf:X\to S be a morphism of varieties with SS smooth of dimension kk. Note that by the iterative definition of Fip,+\mathcal{F}^{p,+}_i, there is a natural map Fip,+→Fi+1p,+\mathcal{F}^{p,+}_i\to\mathcal{F}^{p,+}_{i+1} given by the second map of the triangle (2). These yield a chain

ΩXp≃qisF0p,+→F1p,+→⋯→Fpp,+≃qisΩX/Sp,+→Fp+1p,+≃qis0.\Omega_X^p\simeq_{\mathrm{qis}}\mathcal{F}^{p,+}_0\to\mathcal{F}^{p,+}_1\to\cdots\to\mathcal{F}^{p,+}_p\simeq_{\mathrm{qis}}\Omega_{X/S}^{p,+}\to\mathcal{F}^{p,+}_{p+1}\simeq_{\mathrm{qis}}0.

For 0≤q≤p+10\le q\le p+1, define

FSq,+ΩXp:=Cone⁡(ΩXp→Fp−q+1p,+)[−1].\mathbb{F}^{q,+}_S\Omega_X^p:=\operatorname{Cone}\left(\Omega_X^p\to\mathcal{F}^{p,+}_{p-q+1}\right)[-1].

If no confusion will arise, we will suppress the notation of the base SS and denote it by Fq,+\mathbb{F}^{q,+}. Note that

F0,+ΩXp≃qisΩXpandFp+1,+ΩXp=0.\mathbb{F}^{0,+}\Omega_X^p\simeq_{\mathrm{qis}}\Omega_X^p\qquad\text{and}\qquad\mathbb{F}^{p+1,+}\Omega_X^p=0.

By convention, we also define Fq,+ΩXp=0\mathbb{F}^{q,+}\Omega_X^p=0 for q>p+1q>p+1. Now, consider the following diagram:

Fq+1,+ΩXp→ΩXp→Fp−qp,+→+1↓∥↓↓Fq,+ΩXp→ΩXp→Fp−q+1p,+→+1↓↓↓Grq,+ΩXp→0→(Fp−qp,+⊗f∗ΩSq)[1]→+1\begin{CD} \mathbb{F}^{q+1,+}\Omega_X^p @>>> \Omega_X^p @>>> \mathcal{F}^{p,+}_{p-q} @>{+1}>> \\ @VVV @V\Vert VV @VVV \\ \mathbb{F}^{q,+}\Omega_X^p @>>> \Omega_X^p @>>> \mathcal{F}^{p,+}_{p-q+1} @>{+1}>> \\ @VVV @VVV @VVV \\ \mathop{\mathrm{Gr}}\nolimits^{q,+}\Omega_X^p @>>> 0 @>>> \left(\mathcal{F}^{p,+}_{p-q}\otimes f^*\Omega_S^q\right)[1] @>{+1}>> \end{CD}

The third column is the shifted triangle (2) at i=p−qi=p-q, and the top two rows are triangles by construction. By commutativity of the top right square, we may extend to obtain Fq+1,+ΩXp→Fq,+ΩXp\mathbb{F}^{q+1,+}\Omega_X^p\to\mathbb{F}^{q,+}\Omega_X^p, and thereafter define the associated graded complex Grq,+ΩXp\mathop{\mathrm{Gr}}\nolimits^{q,+}\Omega_X^p to be its cone. The derived nine-lemma gives exactness of the third row, which implies that

Grq,+ΩXp≃qisΩX/Sp−q,+⊗f∗ΩSq(4)\mathop{\mathrm{Gr}}\nolimits^{q,+}\Omega_X^p\simeq_{\mathrm{qis}}\Omega_{X/S}^{p-q,+}\otimes f^*\Omega_S^q \tag*{(4)}

We will take a quick detour to establish the functorial properties of Fq,+ΩXp\mathbb{F}^{q,+}\Omega_X^p.

Lemma 3.10. Let ϕ:Y→X\phi:Y\to X be an SS-morphism, and g=f∘ϕg=f\circ\phi. Then for each p,q≥0p,q\ge0, there are natural maps of hyperfiltrations

FSq,+ΩXp→Rϕ∗FSq,+ΩYp.\mathbb{F}^{q,+}_S\Omega_X^p\to R\phi_*\mathbb{F}^{q,+}_S\Omega_Y^p.

That is, the maps are compatible with the hyperfiltration and descend to the associated graded. In particular, there are natural maps

ΩX/Sp,+→Rϕ∗ΩY/Sp,+.\Omega_{X/S}^{p,+}\to R\phi_*\Omega_{Y/S}^{p,+}.

Proof. We will temporarily use the notation Fi,Xp,+\mathcal{F}^{p,+}_{i,X} and Fi,Yp,+\mathcal{F}^{p,+}_{i,Y} for the auxiliary object in Definition 3.6 associated to XX and YY, respectively. Using the triangle (2), consider the following diagram

Fi,Xp,+⊗f∗ΩSp−i→μi,Xi,p−iFi,Xp,+→Fi+1,Xp,+→+1↓↓↓Rϕ∗Fi,Yp,+⊗f∗ΩSp−i→Rϕ∗μi,Yi,p−iRϕ∗Fi,Yp,+→Rϕ∗Fi+1,Yp,+.→+1\begin{CD} \mathcal{F}^{p,+}_{i,X}\otimes f^*\Omega_S^{p-i} @>{\mu^{i,p-i}_{i,X}}>> \mathcal{F}^{p,+}_{i,X} @>>> \mathcal{F}^{p,+}_{i+1,X} @>{+1}>> \\ @VVV @VVV @VVV \\ R\phi_*\mathcal{F}^{p,+}_{i,Y}\otimes f^*\Omega_S^{p-i} @>{R\phi_*\mu^{i,p-i}_{i,Y}}>> R\phi_*\mathcal{F}^{p,+}_{i,Y} @>>> R\phi_*\mathcal{F}^{p,+}_{i+1,Y}. @>{+1}>> \end{CD}

Note the use of the projection formula to identify the bottom left object. For i=0i=0, the vertical arrows are given by the map of Du Bois complexes, as F0,Xp,+≃qisΩXp\mathcal{F}^{p,+}_{0,X}\simeq_{\mathrm{qis}}\Omega_X^p and similarly for YY, see Theorem 2.1(2). Furthermore, the left square commutes by the functoriality of the wedge map μp,q=∧p,q\mu^{p,q}=\wedge_{p,q} in Lemma 3.1. Therefore, the same inductive argument on i≥0i\ge0 as in Construction 3.2 shows that we may extend iteratively to obtain a natural map

Fi+1,Xp,+→Rϕ∗Fi+1,Yp,+.\mathcal{F}^{p,+}_{i+1,X}\to R\phi_*\mathcal{F}^{p,+}_{i+1,Y}.

Moreover, it is compatible with the map on Du Bois complexes ΩXp→Rϕ∗ΩYp\Omega_X^p \to R\phi_*\Omega_Y^p. This allows us to extend the map to the hyperfiltrations, as in the following diagram:

Fq,+ΩXp→ΩXp→Fp−q+1,Xp,+→+1↓↓↓Rϕ∗Fq,+ΩYp→Rϕ∗ΩYp→Rϕ∗Fp−q+1,Yp,+→+1\begin{CD} F^{q,+}\Omega_X^p @>>> \Omega_X^p @>>> \mathscr{F}_{p-q+1,X}^{p,+} @>{+1}>> \\ @VVV @VVV @VVV \\ R\phi_*F^{q,+}\Omega_Y^p @>>> R\phi_*\Omega_Y^p @>>> R\phi_*\mathscr{F}_{p-q+1,Y}^{p,+} @>{+1}>> \end{CD}

This gives the first statement of the lemma. Applying the triangle for the associated graded in Construction 3.8 gives the compatibility. Lastly, the map ΩX/Sp,+→Rϕ∗ΩY/Sp,+\Omega_{X/S}^{p,+} \to R\phi_*\Omega_{Y/S}^{p,+} is simply given by Fp,Xp,+→Rϕ∗Fp,Yp,+\mathscr{F}_{p,X}^{p,+} \to R\phi_*\mathscr{F}_{p,Y}^{p,+}. This completes the proof. □

Lemma 3.11. Let π∙:X∙→X\pi_\bullet:X_\bullet\to X be a cubical hyperresolution of XX. Then for each p≥0p\ge0,

R(π∙)∗ΩX∙/Sp,+≃qisΩX/Sp,+.R(\pi_\bullet)_*\Omega_{X_\bullet/S}^{p,+}\simeq_{\mathrm{qis}}\Omega_{X/S}^{p,+}.

Proof. By (4) and Lemma 3.10 we obtain a morphism of distinguished triangles:

Fq+1,+ΩXp→Fq,+ΩXp→ΩX/Sp−q,+⊗f∗ΩSq→+1↓↓↓(Rπ∙)∗Fq+1,+ΩX∙p→(Rπ∙)∗Fq,+ΩX∙p→(Rπ∙)∗(ΩX∙/Sp−q,+)⊗f∗ΩSq→+1\begin{CD} F^{q+1,+}\Omega_X^p @>>> F^{q,+}\Omega_X^p @>>> \Omega_{X/S}^{p-q,+}\otimes f^*\Omega_S^q @>{+1}>> \\ @VVV @VVV @VVV \\ (R\pi_\bullet)_*F^{q+1,+}\Omega_{X_\bullet}^p @>>> (R\pi_\bullet)_*F^{q,+}\Omega_{X_\bullet}^p @>>> (R\pi_\bullet)_*(\Omega_{X_\bullet/S}^{p-q,+})\otimes f^*\Omega_S^q @>{+1}>> \end{CD}

Note the use of the projection formula for the last term. The proof follows from the same inductive argument as in Lemma 3.10 and the five-lemma, reducing down to the statement about the Du Bois complex, ΩXp≃qisR(π∙)∗ΩX∙p\Omega_X^p\simeq_{\mathrm{qis}}R(\pi_\bullet)_*\Omega_{X_\bullet}^p (cf. [1 2.3] [3 V.3.3]). See also [7 Lemma 3.3] for an analogous statement. □

We collect the results in this section in the following theorem. For easy comparison, we also include the left relative Du Bois complex by [7].

Theorem 3.12. Let f:X→Sf:X\to S be a morphism of algebraic varieties with dim⁡X=n\dim X=n, dim⁡S=k\dim S=k and SS smooth. Then there exist objects ΩX/Sp,+,ΩX/Sp,−\Omega_{X/S}^{p,+},\Omega_{X/S}^{p,-} in Dcohb(X)D^b_{\mathrm{coh}}(X) for all p∈Zp\in\mathbb{Z} with the following properties:

(1) ΩX/Sp,+≃qis0\Omega_{X/S}^{p,+}\simeq_{\mathrm{qis}}0 for p<0p<0, ΩX/Sp,−≃qis0\Omega_{X/S}^{p,-}\simeq_{\mathrm{qis}}0 for p>n−kp>n-k.

(2) There exist hyperfiltrations Fq,+ΩXp,Fq,−ΩXpF^{q,+}\Omega_X^p,F^{q,-}\Omega_X^p of ΩXp\Omega_X^p with 0≤q≤k+10\le q\le k+1, such that

F0,+ΩXp≃qisF0,−ΩXp≃qisΩXp,Fk+1,+ΩXp≃qisFk+1,−ΩXp≃0,F^{0,+}\Omega_X^p\simeq_{\mathrm{qis}}F^{0,-}\Omega_X^p\simeq_{\mathrm{qis}}\Omega_X^p,\qquad F^{k+1,+}\Omega_X^p\simeq_{\mathrm{qis}}F^{k+1,-}\Omega_X^p\simeq0,

whose associated graded complexes are, with 0≤q≤k0\le q\le k,

Gq,+ΩXp≃qisΩX/Sp−q,+⊗f∗ΩSq,Gq,−ΩXp≃qisΩX/Sp−q,−⊗f∗ΩSq.G^{q,+}\Omega_X^p\simeq_{\mathrm{qis}}\Omega_{X/S}^{p-q,+}\otimes f^*\Omega_S^q,\qquad G^{q,-}\Omega_X^p\simeq_{\mathrm{qis}}\Omega_{X/S}^{p-q,-}\otimes f^*\Omega_S^q.

(3) Fq,+ΩXpF^{q,+}\Omega_X^p and Fq,−ΩXpF^{q,-}\Omega_X^p are functorial, i.e. if ϕ:Y→X\phi:Y\to X is an SS-morphism, then there are natural maps Fq,±ΩXp→Rϕ∗Fq,±ΩYpF^{q,\pm}\Omega_X^p\to R\phi_*F^{q,\pm}\Omega_Y^p in D(X)D(X) forming the commutative diagram of distinguished triangles:

Fq+1,±ΩXp→Fq,±ΩXp→Gq,±ΩXp→+1↓↓↓Rϕ∗Fq+1,±ΩYp→Rϕ∗Fq,±ΩYp→Rϕ∗Gq,±ΩYp→+1\begin{CD} F^{q+1,\pm}\Omega_X^p @>>> F^{q,\pm}\Omega_X^p @>>> G^{q,\pm}\Omega_X^p @>{+1}>> \\ @VVV @VVV @VVV \\ R\phi_*F^{q+1,\pm}\Omega_Y^p @>>> R\phi_*F^{q,\pm}\Omega_Y^p @>>> R\phi_*G^{q,\pm}\Omega_Y^p @>{+1}>> \end{CD}

In particular, there are natural maps in D(X)D(X) forming the commutative diagram:

ΩXp→ΩX/Sp,±↓↓Rϕ∗ΩYp→Rϕ∗ΩY/Sp,±\begin{CD} \Omega_X^p @>>> \Omega_{X/S}^{p,\pm} \\ @VVV @VVV \\ R\phi_*\Omega_Y^p @>>> R\phi_*\Omega_{Y/S}^{p,\pm} \end{CD}

(4) If ff is smooth, then ΩX/Sp,+≃qisΩX/Sp,−≃qisΩX/Sp\Omega_{X/S}^{p,+}\simeq_{\mathrm{qis}}\Omega_{X/S}^{p,-}\simeq_{\mathrm{qis}}\Omega_{X/S}^p.

(5) Let {νi:Ui↪X}\{\nu_i: U_i \hookrightarrow X\} be a finite open cover of XX. Then

ΩX/Sp,±≃qisR(ν∙)∗ΩU∙/Sp,±.\Omega^{p,\pm}_{X/S} \simeq_{\mathrm{qis}} R(\nu_\bullet)_* \Omega^{p,\pm}_{U_\bullet/S}.

Proof. The statements for ΩX/Sp,−\Omega^{p,-}_{X/S} follow from [7 Theorem 3.4]. For the right complex ΩX/Sp,+\Omega^{p,+}_{X/S}, its existence and (1) follow from Construction 3.2 and Definition 3.7. (2) is Construction 3.8 and (3.9). (3) is Lemma 3.10. The rest are standard consequences of the analogous properties of the absolute Du Bois complex and the constructions of the hyperfiltrations, see [6 Theorem 1.3], [7 Theorem 3.4], [13 Theorem 3.2]. □

Corollary 3.13. Let f:X→Sf:X \to S be a morphism of algebraic varieties such that SS is smooth. Then for any p≥0p \ge0 and any locally free sheaf E\mathcal{E} on XX there exists a spectral sequence

E1q,i=Hq+i(X,ΩX/Sp−q,±⊗f∗ΩSq⊗E)⇒Hq+i(X,ΩXp⊗E).E_1^{q,i}=\mathbb{H}^{q+i}(X,\Omega^{p-q,\pm}_{X/S}\otimes f^*\Omega^q_S\otimes\mathcal{E}) \Rightarrow\mathbb{H}^{q+i}(X,\Omega^p_X\otimes\mathcal{E}).

Proof. The statement for ΩX/Sp,−\Omega^{p,-}_{X/S} is [7 Corollary 3.5]. The analogous statement for ΩX/Sp,+\Omega^{p,+}_{X/S} follows immediately from Theorem 3.12(2) and [7 Theorem 1.2.2]. □

The next proposition shows that the relative Du Bois complexes ΩX/Sp,±\Omega^{p,\pm}_{X/S} agree with the existing definition in the literature when the base is a smooth curve (cf. [6, 13]), justifying the notation. We include the comparison for the left relative Du Bois complex of Kovács here for completeness, as it is not included in [7].

Proposition 3.14. Let f:X→Sf:X \to S be a morphism of algebraic varieties, where SS is a smooth curve. Then ΩX/Sp,+,ΩX/Sp,−\Omega^{p,+}_{X/S},\Omega^{p,-}_{X/S} agree with the objects defined in [13 Theorem 3.2] for all p∈Np \in\mathbb{N}.

Proof. To avoid confusion, we will always use the notations ΩX/Sp,+,ΩX/Sp,−\Omega^{p,+}_{X/S},\Omega^{p,-}_{X/S} to mean the objects constructed in Theorem 3.12, and use Ω~X/Sp,+,Ω~X/Sp,−\widetilde{\Omega}^{p,+}_{X/S},\widetilde{\Omega}^{p,-}_{X/S} to temporarily denote those constructed in [13 Theorem 3.2]. The statement trivially holds if p<0p<0 for ΩX/Sp,+\Omega^{p,+}_{X/S} and if p>dim⁡X−dim⁡f(X)p>\dim X-\dim f(X) for ΩX/Sp,−\Omega^{p,-}_{X/S}. So let pp be in the appropriate range where Fip,+,Fip,−\mathcal{F}^{p,+}_i,\mathcal{F}^{p,-}_i are iteratively defined as cones.

Since dim⁡S=1\dim S=1, ΩSq=0\Omega^q_S=0 for q>1q>1. We see that for all 0≤i≤p−20\le i\le p-2,

Fi+1p,+=Cone⁡(Fip,+⊗f∗ΩS1→Fip,+)≃qisFip,+.\mathcal{F}^{p,+}_{i+1}=\operatorname{Cone}(\mathcal{F}^{p,+}_i\otimes f^*\Omega^1_S\to\mathcal{F}^{p,+}_{i})\simeq_{\mathrm{qis}}\mathcal{F}^{p,+}_i.

In other words, for all 0≤i≤p−10\le i\le p-1

Fip,+≃qisF0p,+≃qisΩXp.\mathcal{F}^{p,+}_i\simeq_{\mathrm{qis}}\mathcal{F}^{p,+}_0\simeq_{\mathrm{qis}}\Omega^p_X.

Therefore the triangle for ΩX/Sp,+=Fpp,+\Omega^{p,+}_{X/S}=\mathcal{F}^{p,+}_{p} reads

ΩX/Sp−1,+⊗f∗ΩS1→ΩXp→ΩX/Sp,+→+1.\Omega^{p-1,+}_{X/S}\otimes f^*\Omega^1_S\to\Omega^p_X\to\Omega^{p,+}_{X/S}\xrightarrow{+1}.

Furthermore, for a fixed representative KpK^p of ΩXp\Omega^p_X, the first map is given by the natural factorization of the wedge map Kp−1⊗f∗ΩS1→KpK^{p-1}\otimes f^*\Omega^1_S\to K^p. Therefore starting with the base case ΩX/S−1,+≃qisΩ~X/S−1,+=0\Omega^{-1,+}_{X/S}\simeq_{\mathrm{qis}}\widetilde{\Omega}^{-1,+}_{X/S}=0, we see that this agrees precisely with the defining triangle of Ω~X/Sp,+\widetilde{\Omega}^{p,+}_{X/S} in [13 Theorem 3.2(2)], and consequently

ΩX/Sp,+≃qisΩ~X/Sp,+\Omega^{p,+}_{X/S}\simeq_{\mathrm{qis}}\widetilde{\Omega}^{p,+}_{X/S}

for all p≥0p\ge0. A similar argument works for ΩX/Sp,−\Omega^{p,-}_{X/S}, with the difference being the inductive cone construction. The distinguished triangle in [8 2.4] (see also [7 §2]) reads, after index shift

Fi+1p,−=Cone⁡(Fip,−⊗f∗ωS→Fii,−⊗f∗ΩSi−p+1)[−1]≃qisFip,−⊗f∗ωS.\mathcal{F}^{p,-}_{i+1}=\operatorname{Cone}\left(\mathcal{F}^{p,-}_i\otimes f^*\omega_S\to\mathcal{F}^{i,-}_{i}\otimes f^*\Omega^{i-p+1}_S\right)[-1]\simeq_{\mathrm{qis}}\mathcal{F}^{p,-}_i\otimes f^*\omega_S.

The last quasi-isomorphism holds for all i<p−1i<p-1 since ΩSi−p+1=0\Omega^{i-p+1}_S=0 for i−p+1<0i-p+1<0. In other words, for all 0≤i≤p−10\le i\le p-1,

Fip,−≃qisF0p,−⊗f∗ωS⊗i≃qisΩXn−p⊗f∗ωS⊗i.\mathcal{F}^{p,-}_i\simeq_{\mathrm{qis}}\mathcal{F}^{p,-}_0\otimes f^*\omega_S^{\otimes i}\simeq_{\mathrm{qis}}\Omega^{n-p}_X\otimes f^*\omega_S^{\otimes i}.

Now, the triangle for i=p−1i=p-1 reads

Fpp,−→Fp−1p,−⊗f∗ωS→Fp−1p,−→+1.\mathcal{F}^{p,-}_p\to\mathcal{F}^{p,-}_{p-1}\otimes f^*\omega_S\to\mathcal{F}^{p,-}_{p-1}\xrightarrow{+1}.

Using the definition ΩX/Sn−p−1,−=Fp−p⊗f∗ωS−p\Omega^{n-p-1,-}_{X/S}=\mathcal{F}^{p}_{p-}\otimes f^*\omega_S^{-p}, this becomes

ΩX/Sn−p−1,−⊗f∗ωSp⟶ΩX/Sn−p⊗f∗ωSp−1⟶ΩX/Sn−p,+⊗f∗ωSp−1⟶,\Omega^{n-p-1,-}_{X/S}\otimes f^*\omega_S^p\longrightarrow\Omega^{n-p}_{X/S}\otimes f^*\omega_S^{p-1}\longrightarrow\Omega^{n-p,+}_{X/S}\otimes f^*\omega_S^{p-1}\longrightarrow,

which gives the defining triangle for ΩX/Sp,−\Omega^{p,-}_{X/S}, cf. [6 Theorem 1.3.2]. Note that after twisting with f∗ωSpf^*\omega_S^p, the second map in the triangle factorizes the wedge map KXn−p⊗f∗ωS⟶KXn−p+1K^{n-p}_{X}\otimes f^*\omega_S\longrightarrow K^{n-p+1}_{X}. Therefore, the same argument as the right complex concludes that

ΩX/Sp,+≃qisΩX/Sp,−,\Omega^{p,+}_{X/S}\simeq_{\mathrm{qis}}\Omega^{p,-}_{X/S},

once we identify the base case ΩX/Sk,−≃qisΩX/Sk,−=0\Omega^{k,-}_{X/S}\simeq_{\mathrm{qis}}\Omega^{k,-}_{X/S}=0 for k>dim⁡X−dim⁡f(X)k>\dim X-\dim f(X), as desired. □

Higher dimensional base change

The notions in [13 Definition 4.1] extend naturally to our setting over a higher dimensional base.

Definition 4.1. Let f:X→Sf:X\to S be a morphism of algebraic varieties with dim⁡X=n\dim X=n, dim⁡S=k\dim S=k and SS smooth. Let s∈Ss\in S, m∈Zm\in\mathbb{Z}. We say the fiber XsX_s satisfies

  • left-mm-base change for the relative Du Bois complex if there is a natural quasi-isomorphism

ΩX/Sp,+⊗LOXOXs≃qisΩXsp,\Omega^{p,+}_{X/S}\mathbin{\mathop{\otimes}^{\mathbf{L}}}_{\mathcal{O}_X}\mathcal{O}_{X_s}\simeq_{\mathrm{qis}}\Omega^p_{X_s},

for all p≤mp\leq m.

  • right-mm-base change for the relative Du Bois complex if there is a natural quasi-isomorphism

ΩX/Sp,+⊗LOXOXs≃qisΩXsp,\Omega^{p,+}_{X/S}\mathbin{\mathop{\otimes}^{\mathbf{L}}}_{\mathcal{O}_X}\mathcal{O}_{X_s}\simeq_{\mathrm{qis}}\Omega^p_{X_s},

for all p≤mp\leq m.

  • mm-base change for the relative Du Bois complex if it satisfies both left- and right-mm-base change.

  • base change (resp. left-base change, right-base change) for the relative Du Bois complex if it satisfies mm-base change (resp. left-mm-base change, right-mm-base change) for all m∈Zm\in\mathbb{Z}.

We will frequently omit “for the relative Du Bois complex” if it is clear from the context.

The following lemma identifies the relative complex over a point.

Lemma 4.2. Let f:X→Sf:X\to S be a morphism of algebraic varieties, with SS smooth of dimension 00. Then for each p∈Np\in\mathbb{N},

ΩX/Sp,+≃qisΩX/Sp,−≃qisΩXp.\Omega^{p,+}_{X/S}\simeq_{\mathrm{qis}}\Omega^{p,-}_{X/S}\simeq_{\mathrm{qis}}\Omega^p_X.

Proof. Recall as in Construction 3.2 that we iteratively defined

Fi+1p,+=Cone⁡(Fip,+⊗f∗ΩSp−i→Fip,+).\mathcal{F}^{p,+}_{i+1}=\operatorname{Cone}(\mathcal{F}^{p,+}_i\otimes f^*\Omega_S^{p-i}\to\mathcal{F}^{p,+}_i).

Since ΩSp−i=0\Omega_S^{p-i}=0 for all i≠pi\ne p, we have

ΩXp≃qisF0p,+≃qis⋯≃qisFpp,+≃qisΩX/Sp,+.\Omega^p_X\simeq_{\mathrm{qis}}\mathcal{F}^{p,+}_0\simeq_{\mathrm{qis}}\cdots\simeq_{\mathrm{qis}}\mathcal{F}^{p,+}_p\simeq_{\mathrm{qis}}\Omega^{p,+}_{X/S}.

For the left complex ΩX/Sp,−\Omega^{p,-}_{X/S}, it can be defined similarly with a representative of ΩXp\Omega^p_X as in Theorem 3.12 and then the proof would follow. However, since [7] defined it first assuming XX is smooth, we will check this case instead and use [7 Lemma 3.3] to conclude that it holds for singular XX as well. When XX is smooth with dim⁡X=n\dim X=n, we have

F0p,−=ΩXn−p⊗f∗ωS−1,\mathcal{F}^{p,-}_0=\Omega^{n-p}_X\otimes f^*\omega_S^{-1},

Iteratively, [7 Proposition 2.3] defines

Fi+1p,−=Cone⁡(Fip,−⊗f∗ωS→Fip,−⊗f∗ΩS−p+i)[−1],\mathcal{F}^{p,-}_{i+1}=\operatorname{Cone}(\mathcal{F}^{p,-}_i\otimes f^*\omega_S\to\mathcal{F}^{p,-}_i\otimes f^*\Omega_S^{-p+i})[-1],

which then simplifies to

ΩXp≃qisF0n−p,−≃qis⋯≃qisFn−pn−p,−≃qisΩX/Sp,−.\Omega^p_X\simeq_{\mathrm{qis}}\mathcal{F}^{n-p,-}_0\simeq_{\mathrm{qis}}\cdots\simeq_{\mathrm{qis}}\mathcal{F}^{n-p,-}_{n-p}\simeq_{\mathrm{qis}}\Omega^{p,-}_{X/S}.

This concludes the proof for both complexes. □

Lemma 4.3. Let f:X→Sf : X \to S, g:S→Tg : S \to T be morphisms of algebraic varieties, with S,TS,T smooth. Then there exists a functorial map

ΩX/Tp,+⟶ΩX/Sp,+\Omega_{X/T}^{p,+} \longrightarrow\Omega_{X/S}^{p,+}

and a compatible map of hyperfiltrations FT∙,+→FS∙,+F_T^{\bullet,+} \to F_S^{\bullet,+}. That is, for each p,q≥0p,q \ge0, the diagram

FTq+1,+⊗ΩXp→FTq,+⊗ΩXp↓↓FSq+1,+⊗ΩXp→FSq,+⊗ΩXp\begin{CD} F_T^{q+1,+}\otimes\Omega_X^p @>>> F_T^{q,+}\otimes\Omega_X^p \\ @VVV @VVV \\ F_S^{q+1,+}\otimes\Omega_X^p @>>> F_S^{q,+}\otimes\Omega_X^p \end{CD}

commutes. In particular, they are also compatible with the maps from ΩXp\Omega_X^p.

Proof. Recall in Construction 3.2, we iteratively defined intermediate objects Fi,Sp,+,Fi,Tp,+\mathcal{F}_{i,S}^{p,+},\mathcal{F}_{i,T}^{p,+} for p,i≥0p,i \ge0 associated to the morphisms f:X→Sf:X\to S, g∘f:X→Tg\circ f:X\to T, with base case

F0,Sp,+≃qisF0,Tp,+≃qisΩXp.\mathcal{F}_{0,S}^{p,+}\simeq_{\mathrm{qis}}\mathcal{F}_{0,T}^{p,+}\simeq_{\mathrm{qis}}\Omega_X^p.

Here the additional subscripts distinguish the bases SS and TT. The defining triangle for i=0i=0 reads

ΩX0⊗(gf)∗ΩTp→ΩXp→F1,Tp,+→+1↓∥↓ΩX0⊗f∗ΩSp→ΩXp→F1,Sp,+→+1\begin{CD} \Omega_X^0\otimes(gf)^*\Omega_T^p @>>> \Omega_X^p @>>> \mathcal{F}_{1,T}^{p,+} @>{+1}>> \\ @VVV @| @VVV \\ \Omega_X^0\otimes f^*\Omega_S^p @>>> \Omega_X^p @>>> \mathcal{F}_{1,S}^{p,+} @>{+1}>> \end{CD}

where the first map is induced by the natural inclusion of base forms g∗ΩTp→ΩSpg^*\Omega_T^p\to\Omega_S^p. At each step i>0i>0, we may obtain comparison maps from the diagram of (3.5),

Fi,Ti,+⊗(gf)∗ΩTp−i→Fi,Tp,+→Fi+1,Tp,+→+1↓↓↓Fi,Si,+⊗f∗ΩSp−i→Fi,Sp,+→Fi+1,Sp,+→+1\begin{CD} \mathcal{F}_{i,T}^{i,+}\otimes(gf)^*\Omega_T^{p-i} @>>> \mathcal{F}_{i,T}^{p,+} @>>> \mathcal{F}_{i+1,T}^{p,+} @>{+1}>> \\ @VVV @VVV @VVV \\ \mathcal{F}_{i,S}^{i,+}\otimes f^*\Omega_S^{p-i} @>>> \mathcal{F}_{i,S}^{p,+} @>>> \mathcal{F}_{i+1,S}^{p,+} @>{+1}>> \end{CD}

using instead g∗ΩTp−i→ΩSp−ig^*\Omega_T^{p-i}\to\Omega_S^{p-i}. For i+1=pi+1=p, this gives the desired ΩX/Tp,+→ΩX/Sp,+\Omega_{X/T}^{p,+}\to\Omega_{X/S}^{p,+}. Now, we populate the maps for the hyperfiltrations analogously using Construction 3.8. We have a diagram

FTq+1,+⊗ΩXp→FTq,+⊗ΩXp→ΩX/Tp−q,+⊗(gf)∗ΩTq→+1↓↓↓FSq+1,+⊗ΩXp→FSq,+⊗ΩXp→ΩX/Sp−q,+⊗f∗ΩSq→+1\begin{CD} F_T^{q+1,+}\otimes\Omega_X^p @>>> F_T^{q,+}\otimes\Omega_X^p @>>> \Omega_{X/T}^{p-q,+}\otimes(gf)^*\Omega_T^q @>{+1}>> \\ @VVV @VVV @VVV \\ F_S^{q+1,+}\otimes\Omega_X^p @>>> F_S^{q,+}\otimes\Omega_X^p @>>> \Omega_{X/S}^{p-q,+}\otimes f^*\Omega_S^q @>{+1}>> \end{CD}

identifying the associated graded as in Theorem 3.12(2). When q=0q=0, the two hyperfiltration objects are naturally quasi-isomorphic

FT0,+⊗ΩXp≃qisFS0,+⊗ΩXp≃qisΩXp.F_T^{0,+}\otimes\Omega_X^p\simeq_{\mathrm{qis}}F_S^{0,+}\otimes\Omega_X^p\simeq_{\mathrm{qis}}\Omega_X^p.

Hence we may iteratively extend to FTq+1,+→FSq+1,+F_T^{q+1,+}\to F_S^{q+1,+} using the functorial map ΩX/Tp−q,+→ΩX/Sp−q,+\Omega_{X/T}^{p-q,+}\to\Omega_{X/S}^{p-q,+} defined previously, combined with the inclusion of base forms. This shows all the claimed properties. □

The following is a relative version of the triangle for the relative Du Bois complex over a curve.

Proposition 4.4. Let f:X→Sf : X \to S, g:S→Tg : S \to T be morphisms of algebraic varieties, with S,TS,T smooth and gg smooth of relative dimension 1. Then for each p∈Np\in\mathbb{N}, there is a triangle

ΩX/Sp−1,+⊗f∗ΩS/T1⟶ΩX/Tp,+⟶ΩX/Sp,+→+1.\Omega_{X/S}^{p-1,+}\otimes f^*\Omega_{S/T}^1 \longrightarrow\Omega_{X/T}^{p,+}\longrightarrow\Omega_{X/S}^{p,+}\xrightarrow{+1}.

Proof. This is an exercise of patience working with the hyperfiltrations associated with ΩX/Sp,+\Omega_{X/S}^{p,+} and ΩX/Tp,+\Omega_{X/T}^{p,+}. We proceed first by induction on pp, noting

ΩX/S−1,+≃qis0andΩX/T0≃qisΩX/T0,+≃qisΩX/S0,+.\Omega_{X/S}^{-1,+}\simeq_{\mathrm{qis}}0\quad\text{and}\quad\Omega_{X/T}^{0}\simeq_{\mathrm{qis}}\Omega_{X/T}^{0,+}\simeq_{\mathrm{qis}}\Omega_{X/S}^{0,+}.

Suppose the triangle in the statement holds up to some fixed p≥0p \ge0. Using Lemma 4.3, define

Cp,q=Cone⁡(FTq,+⊗ΩXp→FSq,+⊗ΩXp).C^{p,q} = \operatorname{Cone}\left(\mathbb{F}^{q,+}_{T} \otimes\Omega^{p}_{X} \to\mathbb{F}^{q,+}_{S} \otimes\Omega^{p}_{X}\right).

Consider the following diagram associated to the hyperfiltrations in Theorem 3.12,

FTq+1,+⊗ΩXp⟶FTq,+⊗ΩXp⟶ΩX/Tp−q,+⊗(gf)∗ΩTq→+1↓↓↓αp,qFSq+1,+⊗ΩXp⟶FSq,+⊗ΩXp⟶ΩX/Sp−q,+⊗f∗ΩSq→+1↓↓↓Cp,q+1⟶Cp,q⟶Dp,q→+1↓+1↓+1↓+1(5)\begin{aligned} \mathbb{F}^{q+1,+}_{T} \otimes\Omega^{p}_{X} &\longrightarrow\mathbb{F}^{q,+}_{T} \otimes\Omega^{p}_{X} \longrightarrow\Omega^{p-q,+}_{X/T} \otimes(gf)^{*}\Omega^{q}_{T} \xrightarrow{+1} \\ \downarrow\quad&\qquad\qquad\downarrow\qquad\qquad\qquad\qquad\downarrow\alpha^{p,q} \\ \mathbb{F}^{q+1,+}_{S} \otimes\Omega^{p}_{X} &\longrightarrow\mathbb{F}^{q,+}_{S} \otimes\Omega^{p}_{X} \longrightarrow\Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{q}_{S} \xrightarrow{+1} \\ \downarrow\quad&\qquad\qquad\downarrow\qquad\qquad\qquad\qquad\downarrow\\ C^{p,q+1} &\longrightarrow C^{p,q} \longrightarrow D^{p,q} \xrightarrow{+1} \\ \downarrow^{+1}\quad&\qquad\downarrow^{+1}\qquad\qquad\downarrow^{+1} \tag*{(5)} \end{aligned}

where using the derived nine-lemma there is an object Dp,qD^{p,q} making the rows and columns exact triangles.

Claim 4.6. For q≥1q \ge1, we have

Cp,q≃qisΩX/Sp−q,+⊗f∗ΩS/T1⊗(gf)∗ΩTq−1.C^{p,q} \simeq_{\mathrm{qis}} \Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{1}_{S/T} \otimes(gf)^{*}\Omega^{q-1}_{T}.

Assuming Claim 4.6, the q=0q = 0 diagram of (5) reads

FT1,+⊗ΩXp⟶ΩXp⟶ΩX/Tp,+→+1↓∥↓FS1,+⊗ΩXp⟶ΩXp⟶ΩX/Sp,+→+1↓↓↓Cp,1⟶0⟶Cp,1[1].→+1\begin{aligned} \mathbb{F}^{1,+}_{T} \otimes\Omega^{p}_{X} &\longrightarrow\Omega^{p}_{X} \longrightarrow\Omega^{p,+}_{X/T} \xrightarrow{+1} \\ \downarrow\quad&\qquad\Vert\qquad\qquad\downarrow\\ \mathbb{F}^{1,+}_{S} \otimes\Omega^{p}_{X} &\longrightarrow\Omega^{p}_{X} \longrightarrow\Omega^{p,+}_{X/S} \xrightarrow{+1} \\ \downarrow\quad&\qquad\downarrow\qquad\qquad\downarrow\\ C^{p,1} &\longrightarrow0 \longrightarrow C^{p,1}[1]. \xrightarrow{+1} \end{aligned}

We can simply read off the shifted third column to obtain the desired triangle

Cp,1≃qisΩX/Sp−1,+⊗f∗ΩS/T1→ΩX/Tp,+→ΩX/Sp,+→+1C^{p,1} \simeq_{\mathrm{qis}} \Omega^{p-1,+}_{X/S} \otimes f^{*}\Omega^{1}_{S/T} \to\Omega^{p,+}_{X/T} \to\Omega^{p,+}_{X/S} \xrightarrow{+1}

This completes the proof of Proposition 4.4. □

Proof of Claim 4.6. We work with the same inductive assumptions as in the proof of Proposition 4.4, i.e. assuming the proposition holds up to some p≥0p \ge0. Now, the claim is local on XX by Theorem 3.12(5). Note that gg is smooth of relative dimension 11, so after shrinking the base if necessary, we make use of the local splitting

ΩSq≅g∗ΩTq⊕(ΩS/T1⊗g∗ΩTq−1).\Omega^{q}_{S} \cong g^{*}\Omega^{q}_{T} \oplus\left(\Omega^{1}_{S/T} \otimes g^{*}\Omega^{q-1}_{T}\right).

Thus for q≥1q \ge1, we may apply the inductive hypothesis to the third column of (5), and obtain

Dp,q≃qis(ΩX/Sp−q−1,+⊗f∗ΩS/T1⊗(gf)∗ΩTq)[1]⊕(ΩX/Sp−q,+⊗f∗ΩS/T1⊗(gf)∗ΩTq−1)[0].D^{p,q} \simeq_{\mathrm{qis}} \left(\Omega^{p-q-1,+}_{X/S} \otimes f^{*}\Omega^{1}_{S/T} \otimes(gf)^{*}\Omega^{q}_{T}\right)[1] \oplus\left(\Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{1}_{S/T} \otimes(gf)^{*}\Omega^{q-1}_{T}\right)[0].

Note that as in Theorem 3.12, FSq,+⊗ΩXp=0\mathbb{F}^{q,+}_{S} \otimes\Omega^{p}_{X} = 0 for q>dim⁡Sq > \dim S and FTq,+⊗ΩXp=0\mathbb{F}^{q,+}_{T} \otimes\Omega^{p}_{X} = 0 for q>dim⁡T=dim⁡S−1q > \dim T = \dim S - 1. Therefore when q=dim⁡Sq = \dim S, (5) consists of all zeroes except in the lower right square, and yields

Cp,q≃qisDp,q≃qisΩX/Sp−q,+⊗f∗ΩSq≃qisΩX/Sp−q,+⊗f∗ΩS/T1⊗(gf)∗ΩTq−1.C^{p,q} \simeq_{\mathrm{qis}} D^{p,q} \simeq_{\mathrm{qis}} \Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{q}_{S} \simeq_{\mathrm{qis}} \Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{1}_{S/T} \otimes(gf)^{*}\Omega^{q-1}_{T}.

Here we used the splitting of ΩSq\Omega^{q}_{S} and that ΩTq=0\Omega^{q}_{T} = 0 for q=dim⁡Sq = \dim S.

Now we proceed by descending induction on qq from q=dim⁡Sq = \dim S, with the base case just shown. Note that as in Lemma 4.3, the map αp,q\alpha^{p,q} in (5) factors as

ΩX/Tp−q,+⊗(gf)∗ΩTq→βp,qΩX/Sp−q,+⊗(gf)∗ΩTq→ΩX/Sp−q,+⊗f∗ΩSq\Omega^{p-q,+}_{X/T} \otimes(gf)^{*}\Omega^{q}_{T} \xrightarrow{\beta^{p,q}} \Omega^{p-q,+}_{X/S} \otimes(gf)^{*}\Omega^{q}_{T} \to\Omega^{p-q,+}_{X/S} \otimes f^{*}\Omega^{q}_{S}

with the last map induced by a split injection (gf)∗ΩTq↪f∗ΩSq(gf)^*\Omega_T^q \hookrightarrow f^*\Omega_S^q. Using both inductive hypotheses on pp and qq for the source and the target, the composition realizes a quasi-isomorphism

ΩX/Sp−q−1,+⊗f∗ΩS/T1⊗(gf)∗ΩTq[1]≃qisCone⁡(βp,q)⟶Dp,q≃qisCone⁡(αp,q)⟶Cp,q+1[1].\Omega_{X/S}^{p-q-1,+} \otimes f^*\Omega_{S/T}^1 \otimes(gf)^*\Omega_T^q[1] \simeq_{\mathrm{qis}} \operatorname{Cone}(\beta^{p,q}) \longrightarrow D^{p,q} \simeq_{\mathrm{qis}} \operatorname{Cone}(\alpha^{p,q}) \longrightarrow C^{p,q+1}[1].

Furthermore, the first map is split injective induced by the splitting (gf)∗ΩTq↪f∗ΩSq(gf)^*\Omega_T^q \hookrightarrow f^*\Omega_S^q. More precisely, it comes from this diagram extending left inverses:

ΩX/Tp−q,+⊗(gf)∗ΩTq=ΩX/Tp−q,+⊗(gf)∗ΩTqβp,q↓↓αp,qΩX/Sp−q,+⊗(gf)∗ΩTq→ΩX/Sp−q,+⊗f∗ΩSq↓↓Cone⁡(βp,q)→Cone⁡(αp,q)↓+1↓+1\begin{CD} \Omega_{X/T}^{p-q,+}\otimes(gf)^*\Omega_T^q @= \Omega_{X/T}^{p-q,+}\otimes(gf)^*\Omega_T^q \\ @V{\beta^{p,q}}VV @VV{\alpha^{p,q}}V \\ \Omega_{X/S}^{p-q,+}\otimes(gf)^*\Omega_T^q @>>> \Omega_{X/S}^{p-q,+}\otimes f^*\Omega_S^q \\ @VVV @VVV \\ \operatorname{Cone}(\beta^{p,q}) @>>> \operatorname{Cone}(\alpha^{p,q}) \\ @VV{+1}V @VV{+1}V \end{CD}

Combining the above, we see that the triangle coming from (5)

Cp,q⟶Dp,q⟶Cp,q+1[1]→+1C^{p,q} \longrightarrow D^{p,q} \longrightarrow C^{p,q+1}[1] \xrightarrow{+1}

is split. In particular, again applying the inductive hypothesis on qq identifying Cp,q+1C^{p,q+1} shows that

Cp,q≃qisΩX/Sp−q,+⊗f∗ΩS/Tq⊗(gf)∗ΩTq−1,C^{p,q} \simeq_{\mathrm{qis}} \Omega_{X/S}^{p-q,+} \otimes f^*\Omega_{S/T}^q \otimes(gf)^*\Omega_T^{q-1},

thus proving the claim. □

Theorem 4.7. Let f:X→Sf : X \to S be a flat morphism of algebraic varieties with dim⁡X=n\dim X = n, dim⁡S=k\dim S = k and SS smooth. Let s∈Ss \in S be a closed point and suppose the fiber XsX_s has strict-mm-Du Bois singularities. Then XsX_s satisfies right-mm-base change for the relative Du Bois complex, that is, for all 0≤p≤m0 \le p \le m,

ΩX/Sp,+⊗OXLOXs≃qisΩXsp.\Omega_{X/S}^{p,+} \otimes_{\mathcal{O}_X}^{\mathbf{L}} \mathcal{O}_{X_s} \simeq_{\mathrm{qis}} \Omega_{X_s}^p.

Proof. The case dim⁡S=0\dim S = 0 is Lemma 4.2 and the case dim⁡S=1\dim S = 1 follows from Proposition 3.14 and [13 Theorem 1.1(2)]. We will prove this by induction on the dimension of the base, so suppose dim⁡S=k≥2\dim S = k \ge2 and the theorem has been proved for all bases of dim⁡≤k−1\dim\le k - 1.

The statement is local on the base, so after possibly shrinking SS, choose étale coordinates t1,…,tkt_1,\ldots,t_k and project from tkt_k to obtain g:S→T=At1,…,tk−1k−1g : S \to T = \mathbb{A}^{k-1}_{t_1,\ldots,t_{k-1}}. We see that gg is smooth of relative dimension 11 and g∘fg \circ f is flat.

Fix s∈Ss \in S in the theorem and let t=g(s)t = g(s), so that we have a family Xt→StX_t \to S_t over a smooth curve StS_t. Over s∈Sts \in S_t, we have an inclusion of an effective Cartier divisor Xs→XtX_s \to X_t by flatness of ff. By assumption, XsX_s is strict-mm-Du Bois. In particular note that XtX_t is reduced in a neighborhood of XsX_s. Therefore, after further shrinking of XX to a neighborhood containing XsX_s, [13 Theorem 1.3] implies that XtX_t is strict-mm-Du Bois. We have thus arrived at the scenario of the diagram

Xs→Xt→X↓ft↓f↓{s}→St→S↓↓g↓{t}→T\begin{CD} X_s @>>> X_t @>>> X \\ @VVV @V{f_t}VV @V{f}VV \\ \{s\} @>>> S_t @>>> S \\ @VVV @VVV @V{g}VV \\ \{t\} @>>> T \end{CD}

where each square is Cartesian. Now, Proposition 4.4 pulled back to XtX_t gives a triangle

ΩX/Sp−1,+⊗OXLOXt⊗ft∗ΩSt1⟶ΩX/Tp,+⊗OXLOXt⟶ΩX/Sp,+⊗OXLOXt→+1,\Omega_{X/S}^{p-1,+} \otimes_{\mathcal{O}_X}^{\mathbf{L}} \mathcal{O}_{X_t} \otimes f_t^*\Omega_{S_t}^1 \longrightarrow\Omega_{X/T}^{p,+} \otimes_{\mathcal{O}_X}^{\mathbf{L}} \mathcal{O}_{X_t} \longrightarrow\Omega_{X/S}^{p,+} \otimes_{\mathcal{O}_X}^{\mathbf{L}} \mathcal{O}_{X_t} \xrightarrow{+1},

where we used that f∗ΩS/T1⊗OSOXt≅ft∗ΩSt/κ(t)1f^*\Omega^1_{S/T}\otimes_{\mathcal{O}_{S}} \mathcal{O}_{X_t}\cong f_t^*\Omega^1_{S_t/\kappa(t)}. By functoriality, we have a diagram

ΩX/Sp−1,+⊗OXLOXt⊗ft∗ΩSt1⟶ΩX/Tp,+⊗OXLOXt⟶ΩX/Tp,+⊗OXLOXt→+1ΩXt/Stp−1,+⊗ft∗ΩSt1⟶ΩXtp⟶ΩXt/Stp,+→+1.\begin{aligned} \Omega^{p-1,+}_{X/S}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_t}\otimes f_t^*\Omega^1_{S_t} &\longrightarrow\Omega^{p,+}_{X/T}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_t} \longrightarrow\Omega^{p,+}_{X/T}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_t} \xrightarrow{+1} \\ \Omega^{p-1,+}_{X_t/S_t}\otimes f_t^*\Omega^1_{S_t} &\longrightarrow\Omega^p_{X_t} \longrightarrow\Omega^{p,+}_{X_t/S_t} \xrightarrow{+1}. \end{aligned}

Applying the inductive hypothesis to g∘fg\circ f gives base change of the middle vertical map

ΩX/Tp,+⊗OXLOXt≃qisΩXtp,+.\Omega^{p,+}_{X/T}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_t}\simeq_{\mathrm{qis}}\Omega^{p,+}_{X_t}.

for 0≤p≤m0\leq p\leq m. On the other hand, we may perform induction on p≥0p\geq0, starting from

ΩX/S−1,+≃qisΩXt/St−1,+≃qis0,\Omega^{-1,+}_{X/S}\simeq_{\mathrm{qis}}\Omega^{-1,+}_{X_t/S_t}\simeq_{\mathrm{qis}}0,

and see that the left vertical map is also a quasi-isomorphism. Therefore by the five-lemma we have

ΩX/Sp,+⊗OXLOXt≃qisΩXt/Stp,+.\Omega^{p,+}_{X/S}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_t}\simeq_{\mathrm{qis}}\Omega^{p,+}_{X_t/S_t}.

Finally, apply curve base change [13 Theorem 1.1(2)] to the family Xt→StX_t\to S_t to obtain

ΩX/Sp,+⊗OXLOXs≃qisΩXt/Stp,+⊗OXtLOXs≃qisΩXsp\Omega^{p,+}_{X/S}\otimes_{\mathcal{O}_X}^{\mathbf{L}}\mathcal{O}_{X_s}\simeq_{\mathrm{qis}}\Omega^{p,+}_{X_t/S_t}\otimes_{\mathcal{O}_{X_t}}^{\mathbf{L}}\mathcal{O}_{X_s}\simeq_{\mathrm{qis}}\Omega^p_{X_s}

for 0≤p≤m0\leq p\leq m, as desired. □

References

University of Washington, Department of Mathematics, Seattle, WA 98195, USA

Email address: [email protected]

School of Mathematics, Georgia Institute of Technology, GA 30332, USA

Email address: [email protected]

References

  1. [1]Philippe Du Bois. Complexe de de Rham filtré d’une variété singulière. Bull. Soc. Math. France, 109(1):41–81, 1981.DOI
  2. [2]Bradley Dirks. A hodge module construction of relative du bois complexes, 2026. arXiv:2609.17148.arxiv.org/abs/2609.17148
  3. [3]F. Guillén, V. Navarro Aznar, P. Pascual Gainza, and F. Puerta. Hyperrésolutions cubiques et descente co-homologique, volume 1335 of Lecture Notes in Mathematics. Springer-Verlag, Berlin, 1988. Papers from the Seminar on Hodge-Deligne Theory held in Barcelona, 1982.
  4. [4]Caleb Ji and Sándor Kovács. General base change for relative Du Bois complexes, 2025. arXiv:2508.02848.arxiv.org/abs/2508.02848
  5. [5]János Kollár. Singularities of the minimal model program, volume 200 of Cambridge Tracts in Mathematics. Cambridge University Press, Cambridge, 2013. With a collaboration of Sándor Kovács.DOI
  6. [6]Sándor J. Kovács. Smooth families over rational and elliptic curves. J. Algebraic Geom., 5(2):369–385, 1996.
  7. [7]Sándor J. Kovács. Relative de Rham complex for non-smooth morphisms. In Birational algebraic geometry (Baltimore, MD, 1996), volume 207 of Contemp. Math., pages 89–100. Amer. Math. Soc., Providence, RI, 1997.DOI
  8. [8]Sándor J. Kovács. Spectral sequences associated to morphisms of locally free sheaves. In Recent progress in arithmetic and algebraic geometry, volume 386 of Contemp. Math., pages 57–85. Amer. Math. Soc., Providence, RI, 2005.DOI
  9. [10]Sándor J. Kovács and Behrouz Taji. Hodge sheaves underlying flat projective families. Math. Z., 303(3):Paper No. 75, 34, 2023.arxiv.org/abs/2103.03515
  10. [11]Sándor J. Kovács and Behrouz Taji. Arakelov inequalities in higher dimensions. J. Reine Angew. Math., 806:115–145, 2024.arxiv.org/abs/2205.00761
  11. [12]Sándor J. Kovács and Behrouz Taji. The relative Du Bois complex—on a question of S. Zucker. In Higher dimensional algebraic geometry—a volume in honor of V. V. Shokurov, volume 489 of London Math. Soc. Lecture Note Ser., pages 151–162. Cambridge Univ. Press, Cambridge, 2025.
  12. [13]Haoming Ning. Deformations, local freeness, and base change for higher Du Bois singularities, 2026. arXiv:2608.07849.
  13. [14]Haoming Ning. Duality and comparisons for relative Du Bois complexes, 2026. Preprint.

Paper details

vlnc:W_tm9stmmc
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents