On holomorphic -actions
Abstract.
In this paper we study holomorphic actions of the complex multiplicative group on complex manifolds around a singular (fixed) point. We prove linearization results for the germ of action and also for the whole action under some conditions on the manifold. This can be seen as a follow-up to previous works of M. Suzuki and other authors.
Key words and phrases:
Stein manifold, holomorphic flow, quasihomogeneous singularity, foliation.2000 Mathematics Subject Classification:
Primary 37F75, 57R30; Secondary 32M25, 32S65.1. Introduction
A fundamental contribution to the study of holomorphic actions on Stein surfaces, was made by M. Suzuki who introduced the use of techniques of Theory of Foliations, Potential Theory and the Theory of Analytic Spaces (cf. [9] and [10]). Given a holomorphic action of the complex multiplicative group on a complex analytic space of dimension two, we denote by the holomorphic foliation on induced by . In [9] it is proved that if is Stein then the admits a meromorphic first integral. In a very sharp study ([10]) the same author proves that any -action on the affine space is analytically linearizable, i.e., conjugate to an action where , and .
As a sort of extension of the above result, in [1] the authors consider the classification of a Stein complex analytic space of dimension two, with a normal singularity , endowed with a -action having a dicritical singularity at , meaning that is a fixed point such that every non-singular orbit of , close enough to accumulates at and only at . These are called quasi-homogeneous singularities in the framework of Analytic and Algebraic Geometry ([8]).
In [3] the authors study a -action with isolated singularities on a nonsingular complex analytic Stein space , the action having some nondicritical singularity and without dicritical singularities. In few words it is proved analytic linearization at the level of leaf spaces.
1.1. Suzuki’s theory
For our current purposes Suzuki’s main result probably is:
Theorem 1 ([10]).
Any analytic -action on is analytically linearizable, i.e., analytically equivalent to an operation of the form , for some .
The classification of holomorphic -actions (i.e., holomorphic flows) with proper orbits on is also obtained by Suzuki ([10], Theorem 4), and not all of them are linearizable.
1.2. Quasi-homogeneous singularities: Orlik-Wagreich results
A 2-dimensional complex analytic variety , with a distinguished point , is called a quasi-homogeneous complex surface singularity, if it admits a holomorphic action of the complex multiplicative group such that every non-singular orbit in a neighborhood of , accumulates only at (see for instance [8], page 67 and [2, 7]). Such an action is called a good action. The study of algebraic quasi-homogeneous singularities is a main topic in the theory of singularities with major contributions from Saito ([6]), giving an algebraic description in the local context, and from Orlik and Wagreich ([4], [5]). The later studied algebraic affine varieties of dimension two, with an isolated singularity at the origin, under the hypothesis that the variety is invariant by an algebraic action of the form where , i.e. all are positive integers. Algebraic surfaces embedded in endowed with such an action are completely classified by them.
1.3. Main results
In this paper we study linearization of -actions and apply our results to some of the previous works above. We denote by a holomorphic action of a complex Lie group on a complex manifold having as fixed point (not necessarily isolated). Given an element we shall denote by the (biholomorphic) map . Let be a holomorphic action, with a fixed point at and let be a local chart. Using the natural identification , we can use the derivative of the at to define the following holomorphic action:
Although we have used to define , any other chart defines an equivalent action. So we will refer to as the action by linear transformations defined by the derivative of at and may write .
In the first part of this work, we study the case , and we obtain the following semilocal linearization result below:
Theorem A (Semi-local linearization in the affine space).
Let be a holomorphic action of on having as a fixed point and let be the action by linear transformations defined by the derivative of at . Then there is an entire map such that and
for each . Moreover, there is a -invariant open set in such that is a biholomorphism.
Next we study actions on complex manifolds. Let be a complex manifold of dimension and a holomorphic action of the group on . Denote by the one-dimensional holomorphic foliation with isolated singularities on induced by . The one-dimensional orbits of are leaves of and the singular set of is contained in the set of fixed points of . Following the foliation theory terminology, an isolated singularity of a one-dimension holomorphic foliation in dimension two will be called dicritical if, for some neighborhood of , there are infinitely many leaves of the restriction accumulating only at . The closure of any such a local leaf is an invariant local analytic curve in , called a separatrix of through . Thus, a dicritical singularity exhibits infinitely many separatrices. A singularity of a good -action on is clearly dicritical. With a notion of dicricity similar to the one above we obtain the following linearization result:
Theorem B (Local linearization on a manifold).
Let be a holomorphic action on a complex manifold , with an isolated fixed point at such that is of the dicritical type. Let be the action by linear transformations defined by the derivative of of at . There are a -invariant open set in , a -invariant open set in and a biholomorphism such that
for each .
In the last section we apply our techniques to the classification of -actions on Stein manifolds, in particular we obtain variants of Suzuki’s results above (cf. Theorems C and D).
Acknowledgement. This work was partially supported by Faperj and CNPq. On behalf of all authors, the corresponding author states that there is no conflict of interest.
2. Local linearization of -actions
In this section we prove Theorem A. We shall write to denote the action . We begin with the following lemma, a sort of local version of Bochner-Cartan theorem for our situation:
Lemma 1.
Let be a holomorphic action with a fixed point at and let be the action by linear transformations defined by the derivative of at . There is a holomorphic map such that:
-
.
-
e .
Proof.
Consider the action (holomorphic flow) on
We define the holomorphic map
We write in the form
where . If , then
(2.1) | |||||
Since , we have that
(2.2) | |||||
It follows from (2.1) and (2.2) that
(2.3) |
A similar computation shows that . This, together with (2.3) implies that
for all . Now, fix an arbitrary . The map
is holomorphic. Since , we have that is identically zero. Since was taken arbitrarily, the proof is finished. ∎
Lemma 2.
Let be a holomorphic action with a fixed point at and let be the action by linear transformations defined by the derivative of at . Let be the holomorphic map defined in Lemma 1. If and are open sets in satisfying
-
(1)
is injective in and
-
(2)
is connected, for each ,
then is injective in the saturation of by (the orbits of) .
Proof.
The saturation of by is the union of all orbits of intersecting . For each , it follows from Lemma 1 that
(2.4) |
in some open set . Using and the Identity Principle, we conclude that (2.4) happens in . We define the map
where satisfies . We need to show that is well defined. It is obvious if . Take . Suppose that there exist such that and . In this case, there is a such that and this implies that
that is, . Since , it follows from (2.4) that . Therefore,
This implies that is well defined and that is holomorphic, since is open. We define the map
where satisfies . By item () of Lemma 1 we have that . On the other hand and , that is, is the inverse of . ∎
Now, we can prove Theorem A:
Proof of Theorem A.
Consider the flow action
Let be the complete holomorphic vector field whose flow is . If we identify with the map , then the action
is the flow of the linear vector field . Since the isotropy of at each point contains , a Jordan canonical form for has no nilpotent part. Therefore, there is a linear isomorphism and an action such that
where and . Let be an open ball centered at the origin of .
Claim 1.
For each , is connected.
Proof.
In fact, take and . There exists such that . Let be the straight line segment joining and the origin. It is enough to show that . So take . Each real coordinate of has modulus less than the modulus of respective coordinate of . So the same is valid for their complex coordinates. Since the -th complex coordinate of is
(2.5) |
we conclude that each complex coordinate of has modulus less than the modulus of respective coordinate of . Therefore, and Claim 1 is proved. ∎
3. Actions of on complex manifolds
Now we search for results similar to Theorem A for complex manifolds. With this aim the following definition is given:
Definition 1.
Let and be holomorphic actions on complex manifolds and . Let and connected open sets in and respectively and let be a biholomorphism such that . We write
if, for each , there exists an open set such that
Occasionally, we will use the notation without mentioning the map .
Remark 1.
It is easy to see that is symmetric. More precisely, if by , then by .
Remark 2.
Suppose that by . Let be an connected open set such that . Since is a fixed point, there exists an open set (the neutral element) in , such that when . So by Identity Principle, we have that, for each ,
In addiction, it also follows from Identity Principle that if is connected, then . So using the symmetry of , we conclude that
Definition 2.
Let be a holomorphic action with a fixed point at and let be the action by linear transformations defined by the derivative of at . If , we say that is locally linearizable at .
A very preliminary result towards Theorem B is the following:
Lemma 3.
Every holomorphic action with a fixed point at is locally linearizable at .
Proof.
Consider the action
Let be a local chart at and let be the action by linear transformations defined by the derivative of at , constructed from . Since is a fixed point, there exists an open set such that for all . We define the map
Fixed , we can repeat the argument in Lemma 1 to obtain
(3.1) |
Since , we can assume that is injective, reducing if necessary. Take a connected open set in and an open set in such that whenever , . Take and define the holomorphic map
By (3.1), whenever . Therefore , that is,
(3.2) |
The set
is a subgroup of , so by (3.2), . This implies that and the proof is complete. ∎
Definition 3.
Let
-
•
and be holomorphic actions;
-
•
and be open sets in and respectively;
-
•
be an open set in such that ;
-
•
be a biholomorphism such that .
We write
if, for each , some of the following alternatives occurs:
-
(1)
and in ,
-
(2)
and in .
Occasionally, we will use the notation without mentioning the application and the open sets.
Lemma 4.
implies that . Moreover, the relation is symmetric: if by , then by , where .
Proof.
The proof is an easy verification.∎
Lemma 5.
Let and be holomorphic actions such that by . Suppose that there exists a connected open set such that, for each , some of the following alternatives occurs:
-
(1)
,
-
(2)
.
Then, by .
Proof.
This is an immediate consequence of the Identity Principle.∎
Lemma 6.
If and are holomorphic actions such that by then there is a biholomorphism that extends and satisfies
Proof.
We define
where satisfies . We need to prove that is well defined. If , this follows from Lemma 4. Take and suppose that there exist such that e . In this case, there exists a such that . We have that , that is, Since , so . Therefore
This implies that is well defined and that is holomorphic, since is open. Given , there exist and such that . If , then
By Lema 4, by . Thus we can construct analogously the inverse of , and the proof is complete. ∎
Remark 3.
Given a holomorphic action , consider the action
Let be a local chart at and let be the action by linear transformations defined by the derivative of at , constructed from . Let be the complete holomorphic vector field defined by . If we identify with the map , we have that is the flow of the linear vector field . Since the isotropy of at each point contains , a Jordan canonical form for has no nilpotent part. Therefore, there is a linear isomorphism such that
where . We say that is a fixed point of the dicritical type if all are nonzero with a same signal. This does not contradict the notion of dicriticity introduced above, about having all orbits near being contained in separatrices of the corresponding foliation .
Proof of Theorem B.
Take , and as in Remark 3. We know that (up to linear conjugation)
(3.3) |
where , . Let . The -th coordinate of is
(3.4) |
Since is of the dicritical type, there are no ’s with different signals. Suppose that neither is negative (the other case is identical). If is a open ball centered at origin, then
-
•
if ,
-
•
if and
-
•
if .
By Lemma 3, there is a biholomorphism such that by . Thus, by . We can suppose that is an open ball. If we take an open ball centered at origin, then follows from Lemma 5 that by . So the desired result follows from Lemma 6. ∎
4. Actions of on Stein manifolds
Our above results lead to natural variants of results proven for holomorphic actions on Stein manifolds. For instance, joining the results in [7] and [3] we can state the following global linearization theorem:
Theorem 2 ([2], Theorem 1.1).
Let be a complete holomorphic vector field with isolated singularities on a Stein manifold of dimension . Assume that has isolated singularities and some dicritical singularity with first jet of the form , where , . If is finite and then is holomorphically conjugate to . In particular, is biholomorphic to .
Applying our linearization results to this theorem we obtain:
Theorem C.
Let be a -action with isolated singularities on a Stein manifold of dimension . Assume that has an isolated dicritical singularity at . If then is analytically linearizable, indeed, there is a biholomorphic map which conjugates to the linear flow of , where , . In particular, is biholomorphic to .
Proof.
The -action induces a holomorphic flow on by . The flow is -periodic and induces a complete holomorphic vector field on by . By hypothesis is a dicritical singularity for the action and by Theorem B there is an analytic linearization for and therefore for in an invariant neighborhood of in . This implies that has a first jet at of the form , where , for some . Applying then Theorem 2 above we conclude the proof. ∎
Next we obtain a sort of extension of Suzuki’s linearization theorem in [10] to -actions on Stein surfaces:
Theorem D.
Let be a -action with isolated singularities on a Stein surface with and . The we have the following possibilities:
-
(i)
has a nondicritical singularity and the action admits a holomorphic first integral ( or ) of the form where , j=1,2, are irreducible curves. The foliation is the pull-back of the linear foliation given on by the vector field and corresponding to the action , . The map is bijective as a map between leaf spaces.
-
(ii)
has a dicritical singularity, is globally linearizable and is biholomorphic to .
Proof.
Because is Stein there is a meromorphic first integral for on (cf. [9]). We can also assume that this first integral is primitive and is onto an open Riemann surface in the sense of [10] as explained in [2]. If has no fixed point on then defines a holomorphic fibration what is not possible thanks to the hypothesis (recall that the fibers are which is topologically a cylinder). Therefore must have some fixed point, i.e., some singularity for . If the singularity is nondicritical then we apply [2]. If the singularity is dicritical then we apply Theorem C above (essentially, [7]).
∎
References
- [1] C. Camacho, H. Movasati, B. Scárdua: The Moduli of Quasi-Homogeneous Stein Surface Singularities. J Geom Anal 19, 244-260 (2009). https://doi.org/10.1007/s12220-008-9062-7.
- [2] C. Camacho; B. Scárdua: Dicritical holomorphic flows on Stein manifolds; Archiv der Mathematik, 89 (2007), 339-349.
- [3] C. Camacho; B. Scardua: Nondicritical - actions on two dimensional Stein manifolds. Manuscripta Mathematica, Volume 129, Number 1 / May, 2009, 91–98.
- [4] P. Orlik; Wagreich: Isolated singularities of algebraic surfaces with -action; Annals of Math. 93, 205-228 (1971).
- [5] P. Orlik: Seifert manifolds. Springer Lecture Notes in Mathematics 291. 1-155 (1970)
- [6] K. Saito: Quasihomogene isolierte Singularitäten von Hyperflächen. Invent. Math. 14 (1971), 123–142.
- [7] B. Scardua: On the classification of holomorphic flows and Stein surfaces; Complex Variables and Elliptic Equations Vol. 52, No. 1, January 2007, 79-83.
- [8] J. Seade: On the topology of isolated singularities in analytic spaces. Progress in Mathematics, 241. Birkhäuser Verlag, Basel, 2006.
- [9] M. Suzuki: Sur les opérations holomorphes de et de sur un espace de Stein, Séminaire Norguet, Springer Lect. Notes, 670 (1977), 80-88.
- [10] M. Suzuki: Sur les opérations holomorphes du groupe additif complexe sur l’espace de deux variables complexes; Ann. Sci. Éc. Norm. Sup. 4e série, t.10, 1977, p. 517 à 546.