betaFeedback

WorksMathematics

Holomorphic Tischler fibration theorem for Stein spaces

Abstract

We study a holomorphic analogue of Tischler approximation for closed differential forms normalized by a prescribed holomorphic vector field. The main setting consists of connected normal Stein spaces satisfying a degree-one reflexive Poincaré property, with finitely generated first integral homology and a holomorphic C∗\mathbb C^*-action admitting a closed normalized reflexive one-form. Under these hypotheses, the given form can be approximated on compact subsets of the regular locus by closed normalized reflexive forms with discrete period groups. When the first Betti number is at least two, the approximating forms may be chosen with rank-two lattice periods. We relate these approximations to holomorphic locally trivial fibrations over quotients of C\mathbb C, with connected Stein analytic fibers that may be singular. We also establish a finite-equivariant version for suitable Stein quotients and discuss the associated normalization obstructions, including those arising from local singularity topology. Further results concern closed holomorphic time forms, transverse foliations, and the relation between the period groups and the dynamics of complete holomorphic flows. Examples illustrate both the applicability and the limitations of the hypotheses.

Introduction

The theory of foliations connects several areas of mathematics, including differential topology, algebraic topology, and dynamical systems. A classical example of this interaction is Tischler’s fibration theorem [39], which relates three objects: (i) nowhere-vanishing closed real one-forms on smooth compact manifolds; (ii) fibrations over the circle; and (iii) smooth codimension-one foliations invariant under transverse real flows (see [35, 21]). This viewpoint is also useful in the study of certain Anosov flows [35, 36]. The holomorphic setting has been considered in [36].

In this work we investigate the noncompact holomorphic setting, principally under the assumption that the ambient space is Stein. Our first objective is to approximate closed holomorphic one-forms on noncompact Stein spaces by closed holomorphic one-forms with discrete periods. We therefore study closed holomorphic time forms normalized by a fixed complete vector field, especially the infinitesimal generator of a holomorphic C∗\mathbb C^*-action. Our main results establish normalized approximation on normal Stein spaces under a local reflexive Poincaré hypothesis and identify the holomorphic fibrations associated with discrete periods. We also consider finite quotients, surface-specific existence and uniformization, and obstructions arising in singular and nonperiodic settings. The transverse-flow and suspension viewpoint builds on earlier work on invariant holomorphic foliations [36, 21].

Basic notions

Let MM be a connected complex manifold, and let ΩM1\Omega_M^1 denote its sheaf of holomorphic one-forms. For every open subset U⊆MU\subseteq M, we have ΩM1(U)=H0(U,T∗M),{\Omega_M^1(U)=H^0(U,T^*M),} where T∗MT^*M is the holomorphic cotangent bundle. Given a nowhere-vanishing holomorphic vector field XX on MM we put:

TX={ω∈H0(M,ΩM1):dω=0, ω(X)=1},FX={β∈H0(M,ΩM1):dβ=0, β(X)=0}.\begin{equation} \begin{aligned} \mathcal T_X&=\{\omega\in H^0(M,\Omega_M^1):d\omega=0,\ \omega(X)=1\},\\ \mathcal F_X&=\{\beta\in H^0(M,\Omega_M^1):d\beta=0,\ \beta(X)=0\}. \end{aligned} \tag*{(1.1)} \end{equation}

We shall refer to TX\mathcal T_X as the space of closed time forms of XX and to FX\mathcal F_X as the space of closed tangent forms of XX. Whenever ω0∈TX≠∅\omega_0 \in \mathcal T_X\ne\varnothing we have the structure of an affine space TX=ω0+FX\mathcal T_X=\omega_0+\mathcal F_X. Every ω∈TX\omega\in\mathcal T_X defines a nonsingular holomorphic codimension-one foliation Fω=ker⁡ω\mathcal F_\omega=\ker\omega, invariant under the local flow of XX ([21], [36]). Our approximation problem asks whether a closed time form ω\omega can be approximated on compact sets by ωj∈TX\omega_j\in\mathcal T_X whose period images in C\mathbb C are discrete. Throughout this work we shall adopt the following terminology. Given a holomorphic action Φ:C∗×M→M\Phi:\mathbb C^*\times M\to M, it defines a complete holomorphic vector field, called infinitesimal generator, by

X(p)=ddt∣t=0Φ(et,p).X(p)=\left.\frac{d}{dt}\right|_{t=0}\Phi(e^t,p).

Its flow is given by φt(p)=Φ(et,p)\varphi_t(p)=\Phi(e^t,p), and φ2πi=id⁡M\varphi_{2\pi i}=\operatorname{id}_M. The zeros of the infinitesimal generator are the fixed points of the action.

For a class ξ∈H1(M,C)\xi\in H^1(M,\mathbb C), its period homomorphism is

Per⁡ξ:H1(M,Z)⟶C,[γ]⟼⟨ξ,[γ]⟩.\operatorname{Per}_{\xi}:H_1(M,\mathbb Z)\longrightarrow\mathbb C, \qquad [\gamma]\longmapsto\langle\xi,[\gamma]\rangle.

The discrete-period locus is defined as

D(M)={ξ∈H1(M,C):Per⁡ξ(H1(M,Z)) is discrete in C}.\begin{equation} \mathcal D(M)=\bigl\{\xi\in H^1(M,\mathbb C): \operatorname{Per}_{\xi}(H_1(M,\mathbb Z))\text{ is discrete in }\mathbb C\bigr\}. \tag*{(1.2)} \end{equation}

Since every discrete additive subgroup of C\mathbb C is free abelian of rank at most two, every period group associated with a class in D(M)\mathcal D(M) has rank at most two.

We turn to normal complex analytic spaces, using standard terminology as in [25].

Definition 1.1 (Normalized reflexive holomorphic one-form). Let SS be a normal complex analytic space, and let SregS_{\mathrm{reg}} denote its regular locus. Write

j:Sreg↪Sj:S_{\mathrm{reg}}\hookrightarrow S

for the natural inclusion.

We denote by ΩSreg1\Omega^1_{S_{\mathrm{reg}}} the sheaf of holomorphic differential one-forms on the complex manifold SregS_{\mathrm{reg}}, and define the sheaf of reflexive one-forms in SS as

ΩS[1]:=j∗ΩSreg1.\Omega_S^{[1]}:=j_*\Omega^1_{S_{\mathrm{reg}}}.

Let XX be a holomorphic vector field on SS. A reflexive holomorphic one-form

ω∈H0(S,ΩS[1])\omega\in H^0(S,\Omega_S^{[1]})

is said to be normalized with respect to XX if

ω(X)=1on Sreg.\omega(X)=1 \quad\text{on } S_{\mathrm{reg}}.

It is called a closed normalized reflexive holomorphic one-form if, in addition,

dω=0on Sreg.d\omega=0 \quad\text{on } S_{\mathrm{reg}}.

Recall that, since SS is normal, there is a canonical identification

ΩS[1]≃(ΩS1)∗∗,\Omega_S^{[1]} \simeq (\Omega_S^1)^{**},

where ΩS1\Omega_S^1 denotes the sheaf of Kähler holomorphic differentials and (⋅)∗∗(\cdot)^{**} denotes the double dual as an OS\mathcal O_S-module.

Thus, a reflexive holomorphic one-form is a holomorphic one-form on SregS_{\mathrm{reg}}, regarded as a section of its reflexive extension to SS. Such a form need not extend as a Kähler differential across the singular locus. When SS is smooth, one has ΩS[1]=ΩS1\Omega_S^{[1]}=\Omega_S^1.

Definition 1.2 (Degree-one reflexive Poincaré property). Let SS be a normal complex analytic space, and write j:Sreg↪Sj:S_{\rm reg}\hookrightarrow S for the inclusion of its regular locus. Set

ZS1:=ker⁡(d:ΩS[1]⟶j∗ΩSreg2).\mathcal Z_S^1:=\ker\bigl(d:\Omega_S^{[1]}\longrightarrow j_*\Omega^2_{S_{\rm reg}}\bigr).

The space SS has the degree-one reflexive Poincaré property if the sequence of sheaves of abelian groups

0⟶C‾S⟶OS→dZS1⟶0\begin{equation} 0\longrightarrow\underline{\mathbb C}_S\longrightarrow\mathcal O_S \xrightarrow{d}\mathcal Z_S^1\longrightarrow0 \tag*{(1.3)} \end{equation}

is exact. Equivalently, every closed reflexive holomorphic one-form has a local holomorphic primitive, including at singular points.

The degree-one reflexive Poincaré property is not automatic for a normal Stein space. It holds for several natural classes of complex spaces. It is automatic for smooth complex manifolds by the classical holomorphic Poincaré lemma. It also holds for normal spaces with isolated complete-intersection singularities of complex dimension at least three, as well as their products with complex manifolds (Proposition 2.5).

Another class is provided by finite quotients of smooth complex manifolds. Indeed, closed reflexive one-forms pull back locally to closed holomorphic one-forms on the covering manifold; their local primitives can be averaged over the finite group and descended to the quotient. Thus the property holds, for instance, for the quadratic cone

{xy=z2}⊂C3\{xy=z^2\}\subset\mathbb C^3

and, by Proposition 2.5, for the three-dimensional ordinary double point

{xy−zw=0}⊂C4.\{xy-zw=0\}\subset\mathbb C^4.

Tischler approximation on normal Stein spaces

Our first main result extends normalized Tischler approximation to normal Stein spaces.

Theorem A (Tischler approximation on normal Stein spaces: analytic form). Let SS be a connected normal Stein space equipped with a holomorphic C∗\mathbb C^*-action with infinitesimal generator XX. Assume that:

  1. The first homology H1(S,Z)H_1(S,\mathbb Z) is finitely generated.

  2. SS has the degree-one reflexive Poincaré property.

Given a closed reflexive one-form ω\omega normalized by XX, then ω\omega is a compact-open limit on SregS_{\rm reg} of closed, normalized reflexive forms with discrete period groups. If b1(S)≥2b_1(S)\geq 2, the approximating forms can be chosen so that their period groups are lattices of rank two in C\mathbb C.

In complex dimension two, the existence of a closed normalized holomorphic one-form is automatic for fixed-point-free holomorphic C∗\mathbb C^*-actions on normal Stein surfaces, as shown in Proposition 6.1.

We can also state it in the geometric form below:

Theorem B (Tischler approximation on normal Stein spaces: geometric form). Let SS be a connected normal Stein space with finitely generated H1(S,Z)H_1(S,\mathbb Z), satisfying the degree-one reflexive Poincaré property. Let XX be the infinitesimal generator of a holomorphic C∗\mathbb C^*-action on SS admitting a closed reflexive normalized holomorphic one-form ω\omega.

Then the holomorphic foliation defined by ω\omega on SregS_{\mathrm{reg}} is approximated, in the compact-open topology, by fibrations Fj:S⟶C/ΛjF_j:S\longrightarrow\mathbb C/\Lambda_j where Λj⊂C,j∈N\Lambda_j\subset\mathbb C, j \in \mathbb N is a discrete subgroup.

A few remarks about the geometric version is:

∙\bullet We have Fj∗(dtj)⟶ωF_j^*(dt_j)\longrightarrow\omega uniformly on compact subsets of SregS_{\mathrm{reg}}. Here dtjdt_j denotes the invariant holomorphic one-form on C/Λj\mathbb C/\Lambda_j induced by dzdz on C\mathbb C.

∙\bullet Each base C/Λj\mathbb C/\Lambda_j is biholomorphic either to C∗\mathbb C^* or to an elliptic curve.

∙\bullet If b1(S)≥2b_1(S)\geq2, the subgroups Λj\Lambda_j may all be chosen to be lattices of rank two, so that every base C/Λj\mathbb C/\Lambda_j is an elliptic curve.

Theorem C (Finite-quotient Tischler approximation). Let S=M/GS=M/G be a finite quotient of a connected Stein manifold MM with finitely generated first homology, and assume that MM carries a holomorphic C∗\mathbb C^*-action whose infinitesimal generator is nowhere zero and which commutes with GG. Consequently SS inherits a fixed-point-free holomorphic C∗\mathbb C^*-action. Every GG-invariant normalized closed holomorphic form on MM admits GG-invariant normalized approximants with discrete periods; the descended reflexive forms on SS have discrete periods on SregS_{\rm reg}.

Manifold case

When SS is smooth, the reflexive Poincaré property follows from the classical holomorphic Poincaré lemma (see [40]). The general approximation theorem specializes to the following smooth statement.

Corollary D (Tischler approximation: manifold case). Let MM be a connected Stein manifold with finitely generated first homology group H1(M,Z)H_1(M,\mathbb Z). Let XX be the nowhere-vanishing infinitesimal generator of a holomorphic C∗\mathbb C^*-action on MM. Then the following three cases occur:

  1. If b1(M)=0b_1(M)=0, then TX=∅\mathcal T_X=\varnothing.

  2. If b1(M)=1b_1(M)=1 and TX≠∅\mathcal T_X\ne\varnothing, every ω∈TX\omega\in\mathcal T_X has a nonzero discrete cyclic period group; its integral defines a holomorphic locally trivial fibration with connected fibers over C∗\mathbb C^*, whose fibers are the leaves of ker⁡ω\ker\omega.

  3. If b1(M)≥2b_1(M)\ge2 and TX≠∅\mathcal T_X\ne\varnothing, every ω∈TX\omega\in\mathcal T_X is a compact-open limit of closed normalized holomorphic forms with rank-two lattice period groups. Their leaves are connected fibers of holomorphic locally trivial fibrations over elliptic curves.

In each case where TX≠∅\mathcal T_X\ne\varnothing, the cohomology classes of its members constitute the affine hyperplane {ξ∈H1(M,C):⟨ξ,[γ]⟩=2πi},\bigl\{\xi\in H^1(M,\mathbb C):\langle\xi,[\gamma]\rangle=2\pi i\bigr\}, where [γ][\gamma] is the common orbit-circle class.

On Stein surfaces, fixed-point-free multiplicative actions also yield an automatic existence result.

References

  1. [21]A. Behague and B. Scárdua, Foliations invariant under Lie group transverse actions, Monatsh. Math. 153 (2008), 295--308. https://doi.org/10.1007/s00605-008-0523-7 doi:10.1007/s00605-008-0523-7 .
  2. [22]S. Bochner, Compact groups of differentiable transformations, Ann. of Math. (2) 46 (1945), 372--381.
  3. [23]C. Camacho, H. Movasati, and B. Scárdua, The moduli of quasi-homogeneous Stein surface singularities, J. Geom. Anal. 19 (2009), 244--260. https://doi.org/10.1007/s12220-008-9062-7 doi:10.1007/s12220-008-9062-7 .
  4. [24]F. Forstnerič, Holomorphic submersions from Stein manifolds, Ann. Inst. Fourier (Grenoble) 54 (2004), no. 6, 1913--1942. https://doi.org/10.5802/aif.2071 doi:10.5802/aif.2071 .
  5. [25]H. Grauert and R. Remmert, Coherent Analytic Sheaves, Grundlehren der mathematischen Wissenschaften, vol. 265, Springer-Verlag, Berlin, 1984.
  6. [26]H. A. Hamm, Lokale topologische Eigenschaften komplexer Räume, Math. Ann. 191 (1971), 235--252. https://doi.org/10.1007/BF01578709 doi:10.1007/BF01578709 .
  7. [27]A. Iannuzzi and S. Trapani, A classification of taut, Stein surfaces with a proper $\R$-action, Math. Ann. 352 (2012), 965--986. https://doi.org/10.1007/s00208-011-0664-1 doi:10.1007/s00208-011-0664-1 . See in particular Proposition 4.1 and Remark 4.3.
  8. [28]C. Jörder, On the Poincaré Lemma for reflexive differential forms, arXiv:1401.7495 (2014). https://arxiv.org/abs/1401.7495 arXiv:1401.7495 .
  9. [29]I. Majcen, Closed holomorphic 1-forms without zeros on Stein manifolds, Math. Z. 257 (2007), 925--937. https://doi.org/10.1007/s00209-007-0153-0 doi:10.1007/s00209-007-0153-0 .
  10. [30]J. Milnor, Singular Points of Complex Hypersurfaces, Annals of Mathematics Studies, vol. 61, Princeton University Press, Princeton, 1968.
  11. [31]A. Némethi and L. I. Nicolaescu, Seiberg--Witten invariants and surface singularities II: Singularities with good $\C^*$-action, J. London Math. Soc. 69 (2004), 593--607.
  12. [32]W. D. Neumann, A calculus for plumbing applied to the topology of complex surface singularities and degenerating complex curves, Trans. Amer. Math. Soc. 268 (1981), 299--344.
  13. [33]P. Orlik and P. Wagreich, Isolated singularities of algebraic surfaces with $\C^*$ action, Ann. of Math. (2) 93 (1971), 205--228.
  14. [34]H. Pinkham, Normal surface singularities with $\C^*$ action, Math. Ann. 227 (1977), 183--193.
  15. [35]J. F. Plante, Anosov flows, Amer. J. Math. 94 (1972), 729--754.
  16. [36]B. Scárdua, Holomorphic Anosov systems, foliations and fibring complex manifolds, Dynamical Systems 18 (2003), 365--389.
  17. [37]G. C. Shephard and J. A. Todd, Finite unitary reflection groups, Canadian J. Math. 6 (1954), 274--304; see also the classical Chevalley--Shephard--Todd smooth-quotient criterion and invariant-differential descent for complex reflection groups.
  18. [38]L. Solomon, Invariants of finite reflection groups, Nagoya Math. J. 22 (1963), 57--64. https://doi.org/10.1017/S0027763000011028 doi:10.1017/S0027763000011028 .
  19. [39]D. Tischler, On fibering certain foliated manifolds over $S^1$, Topology 9 (1970), 153--154.
  20. [40]R. O. Wells, Jr., Differential Analysis on Complex Manifolds, 3rd ed., Graduate Texts in Mathematics, vol. 65, Springer, New York, 2008.

Paper details

vlnc:W_dd4ybu9u
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

ChatGPT (OpenAI)

View statement
Looking for related papers…
Contents