Complexes, Residues and obstructions
for
log-symplectic manifolds
Abstract.
We consider compact Kählerian manifolds of even dimension 4 or more, endowed with a log-symplectic structure , a generically nondegenerate closed 2-form with simple poles on a divisor with local normal crossings. A simple linear inequality involving the iterated Poincaré residues of at components of the double locus of ensures that the pair has unobstructed deformations and that deforms locally trivially.
Key words and phrases:
Poisson structure, log-symplectic manifold, deformation theory, log complex, mixed Hodge theory2010 Mathematics Subject Classification:
14J40, 32G07, 32J27, 53D17Data availaibility statement
There is no data set associated with this paper.
Introduction
A log-symplectic manifold is a pair consisting of a complex manifold , usually compact and Kählerian, together with a log-symplectic structure. A log-symplectic structure can be defined either as a generically nongegenerate meromorphic closed 2-form with normal-crossing (anticanonical) polar divisor , or equivalently as a generically nondegenerate holomorphic tangential 2-vector such that with normal-crossing degeneracy divisor . The two structures are related via . See [4] or [11] or [3] or [12] for basic facts on Poisson and log-symplectic manifolds and [5] (especially the appendix), [6], [1], [8] or [10], and references therein, for deformations.
Understanding log-symplectic manifolds unavoidably involves understanding their deformations. In the very special case of symplectic manifolds, where , the classical theorem of Bogomolov [2] shows that the pair has unobstructed deformations. In [14] we obtained a generalization of this result which holds when satisfied a certain ’very general position’ condition with respect to (the original statement is corrected in the subsequent erratum/corrigendum). Namely, we showed in this case that has ’strongly unobstructed’ deformations, in the sense that it has unobstructed deformations and deforms locally trivially.
Further results on unobstructed deformations (in the sense of Hitchin’s generalized geometry [7]) and Torelli theorems in the case where has global normal crossings were obtained by Matviichuk, Pym and Schedler [9], based on their notion of holonomicity.
Our purpose here is to prove a more precise strong unobstructedness result compared to [14], nailing down the generality required: we will show in Theorem 6 that strong unobstructedness can fail only when the log-symplectic structure , more precisely its (iterated Poincaré) residues at codimension-2 strata of the polar divisor (which are essentially the (locally constant) coefficients of with respect to a suitable basis of the log forms adapted to ) satisfy certain special linear relations with integer coefficients. Explicitly, at a triple point of with branches labelled 1,2,3 and associated residues , the condition is
Essentially, if this never happens over the entire triple locus then has strongly unobstructed deformations.
The strategy of the proof as in [14] is to study the inclusion of complexes
albeit from a more global viewpoint. In fact as in [14] it turns out to be more convenient to transport the situation over to the De Rham side where it becomes an inclusion
where the latter ’log-plus’ complex is a certain complex of meromorphic forms with poles on . We study a filtration, introduced in [14], interpolating between the two complexes, especially its first two graded pieces. As we show, the first piece is automatically exact, while 0-acyclicity for the second piece leads to the above cocycle condition. See §3 for details.
We begin the paper with a couple of auxiliary, independent sections. In §1 we construct a ’principal parts complex’ associated to an invertible sheaf on a smooth variety, extending the principal parts sheaf together with the universal derivation . We show this complex is always exact. In §2 we show that, for any normal-crossing divisor on any smooth variety, the log complex - unlike itself- can be pulled back to a complex of vector bundles on the normalization of . These complexes play a role in our analysis of the aforementioned inclusion map.
I am grateful to Brent Pym for helpful communications, in particular for communicating Example 8.
1. Principal parts complex
In this section denotes an arbitrary -dimensional smooth complex variety and denotes an invertible sheaf on .
1.1. Principal parts
The Grothendieck principal parts sheaf (see EGA) is a rank- bundle on defined as
where is the diagonal and are the projections. We have a short exact sequence
whose corresponding extension class in coincides with . The sheaf
which likewise has extension class , is called the normalized principal parts sheaf. The map admits a splitting that is a derivation, i.e.
In fact, the universal derivation on . Moreover is generated over by the image of . Likewise, is generated by elements of the form where is a local generator of .
1.2. Complex
It is well known that which in particular yields a derivation . In fact, This map extends to a complex that we denote by or just and call the ( st) principal parts complex of :
(1) |
The differential is given, in terms of local -generators of L, by
and extending using additivity and the derivation property. There are also similar shorter complexes
Note the exact sequences
These sequences splits locally and also split globally whenever is a flat line bundle. In such cases, we get a short exact sequence
(2) |
The principal parts complex may be tensored with , for any , yielding the -th principal parts complex:
(3) |
The differential is defined by setting
where are local generators for , and extending by additivity and the derivation property. Thus, .
An important property of principal parts complexes is the following:
Proposition 1.
For any local system , the complexes are null-homotopic and exact for all .
Proof.
The assertion being local, we may assume is trivial and so the -th term of is just and the differential is Then a homotopy is given by Thus, is null-homotopic, hence exact. ∎
1.3. Log version
The above constructions have an obvious extension to the log situation. Thus let be a divisor with normal crossings on . We define as the image of under the inclusion , and likewise for . Then as above we get complexes
(4) |
1.4. Foliated version
Let be an integrable subbundle of rank . Then gives rise to a foliated De Rham complex , we well as a foliated principal parts sheaf . Putting these together, we obtain the foliated principal parts complexes (where ):
(5) |
Note that the proof of Proposition 1 made no use of of the acyclicity of the De Rham complex. Hence the same proof applies verbatim to yield
Proposition 2.
For any local system , the complexes are null-homotopic and exact for all .
2. Calculus on normal crossing divisors
In this section denotes a smooth variety or complex manifold and denotes a locally normal-crossing divisor on . Our aim is to show that the log complex on , unlike its De Rham analogue, can be pulled back to the normalization of .
2.1. Branch normal
Let be the normalization of the -fold locus of . A point on consists of a point on together with a choice of distinct local branches of at it. There is a canonical induced normal-crossing divisor on : at a point where is an equation for and are the chosen branches, the equation of is . Note the exact sequence
(6) |
where is the normal bundle to which fits in an exact sequence
Locally, coincides with where is a ’branch equation’: to be precise, if denotes the kernel of the natural surjection , then is an invertible -module locally generated by and . Note that
2.2. Pulling back log complexes
Interestingly, even though the differential on the pullback De Rham complex does not extend to , the analogous assertion for the log complex does hold: the differential on extends to what might be called the restricted log complex:
This is due to the identity (where denotes a branch equation)
Note that the residue map yields an exact sequence
(7) |
Note that the rsidue map commutes with exterior derivative. Therefore this sequence induces a short exact sequence of complexes
(8) |
Furthermore, a twisted form of the restricted log complex, called the normal log complex, also exists:
(9) |
this is thanks to the identity, where is any log form,
Now recall the exact sequence coming from the residue map
In fact, it is easy to check that this exact sequence has extension class hence identifies with so that the normal log complex (9) may be identified with the principal parts complex :
Lemma 3.
The normal log complex is isomorphic to , hence is exact.
Similarly, a pull back log complex exists for all . A similar twisted log complex also exists the determinant of the normal bundle :
(10) |
This comes from (where are the branch equations at a given point of ):
2.3. Iterated residue
We have a short exact sequence of vector bundles on :
(11) |
where is the local system of branches of along and the right map is multiple residue. Taking exterior powers, we get various exact Eagon-Northcott complexes. In particular, we get surjections, called iterated Poincaé residue:
(12) |
(13) |
is a rank-1 local system on which may be called the ’normal orientation sheaf’. The maps for together yield a surjection
(14) |
3. Comparing log and log plus complexes
In this section denotes a log-symplectic smooth variety with log-symplectic form and corresponding Poisson vector , and denotes the degeneracy divisor of or polar divisor of . Our aim is to prove Theorem 6 which shows that deformations of coincide with locally trivial deformations of and are unobstructed.
3.1. Setting up
We will use to denote and similarly for the log versions. This to match with the Lichnerowicz-Poisson complex and . Thus, interior multiplication by induces and isomorphism . Equivalently, itself is a form im inducing a nondegenerate pairing on . In terms of local coordinates, at a point of multiplicity on , we have a basis for of the form
and then
We have an inclusion of complexes
where, for compact Kähler, the first complex controls ’locally trivial’ deformations of , i.e. deformations of inducing a locally trivial deformation of , and the second complex controls all deformations of . It is known (see e.g. [14]) that locally trivial deformations of are always unobstructed and have an essentially Hodge-theoretic (hence topological) character, so one is interested in conditions to ensure that the above inclusion induces an isomorphism on deformation spaces; as is well known, the latter would follow if one can show that the cokernel of this inclusion has vanishing .
Our approach to this question starts with the above ’multiplication by ’ isomorphism
This isomorphism extends to an isomorphism to with a certain subcomplex of , the meromorphic forms regular off , that we call the log plus complex and denote by .
Our goal then becomes that of comparing the log and log-plus complexes. To this end we introduce a filtration on , essentially the filtration induced by the exact sequence
and its isomorphic copy
where is the normalization of and is the associated normal bundle (’branch normal bundle’). We will show that the first graded piece is always an exact complex. The second graded piece is much more subtle. We will show that it is locally exact in degree 0 unless the log-symplectic form , i.e. the matrix above satisfies some special relations with integer coefficients.
The computations of this section are all local in character, though the applications are global.
3.2. Residues and duality
Let be the normalization of the -fold locus of , the induced normal-crossing divisor on . Thus a point of consists of a point of together with a choice of an unordered set of branches of through and is the union of the branches of not in . We consider first the codimension-1 situation. As above, we have a residue exact sequence
(15) |
(the right-hand map given by residue is locally evaluation on where is a local equation for the branch of through the given point of ). Note that if comes from a closed form on near then is a constant.
Dualizing (15), we get
(16) |
where the left-hand map, the ’co-residue’, is locally multiplication by where is a branch equation). Set
Then is canonical as section of , independent of the choice of local equation . By contrast, as section of is canonical only up to a tangential field to , and generates modulo .
Now and admit mutually inverse isomorphisms
The composite
has a rank-1 kernel that is the kernel of the Poisson vector on induced by , aka the conormal to the symplectic foliation on . Now set
Then is locally the form in denoted by in [14]. Again is canonically defined, independent of choices and corresponds to the first column of the matrix for a local coordinate system compatible with the normal-crossing divisor . By contrast, , which depends on the choice of local equation , is canonical up to a log form in and generates modulo the latter.
In , is locally of the form (symplectic), so there . Note that by skew-symmetry we have
Thus, locally . In terms of the matrix above, . Note that which corresponds to the Hamiltonian vector field , is a closed form. Consequently, defines a foliation on . Let be the associated foliated De Rham complex :
endowed with the foliated differential.
Note that the residue exact sequence (15) induces the Poincaré residue sequence
Again the Poincaré residue of a closed form is closed. Now the exact sequence
yields
(17) |
and this sequence induces the filtration on the log-plus complex .
3.3. First graded piece
Now consider first the first graded which is supported in codimension 1. (the shift is so that starts in degree 0). Then is a (finite) direct image of a complex of modules:
Using Lemma 3, we can easily show:
Proposition 4.
is isomorphic to , hence is null-homotopic and exact, hence is exact.
3.4. Second graded piece
Next we study , which is supported on . We consider a connected, nonempty open subset , for example an entire component, over which the ’normal orientation sheaf’ , i.e. the local -system of branches of along , is trivial (we can take if, e.g. has global normal crossings). Such a subset of is said to be a normally split subset of and a normal splitting of is an ordering of the branches is specified. Obviously is covered by such subsets . Likewise, for a subset .
3.4.1. Iterated residue
Over a normally split subset , we have a diagram
(18) |
where is the map induced by . The composite map is just the alternating form induced by , and has the form where is the hyperbolic plane . In terms of a local frame for containing , is the coefficient of in . Note must be constant because is closed. In fact we have
where denote the (Poincaré) residues along the branches of over . Set
This is essentially what is called the biresidue by Matviichuk et al., see [9]. Thus, when , we have a basis for the log forms
multiplicity of , , and then
where
If may be not be normally orientable (e.g. an entire component of ) then is defined only up to sign; if we say that is non-residual, otherwise it is residual.
3.4.2. Non-residual case
Here we consider the case .
Note that in that case we may express along in the form
where the gammas are closed log forms in the coordinates on , i.e. . Moreover, because is divisible by . Also, unless are both holomorphic (pole-free), there is another component of such that (in particular, ). Hence if no such exists, we may by suitably modifying coordinates, assume locally that . A similar argument, or induction, applies to . This means we are essentially in the P-normal case considered in [13]. This we conclude:
Lemma 5.
Unless is P-normal, there exists a nonempty residual open subset of .
3.4.3. Residual case: identifying
Next we analyze a residual normally oriented open subset . As above, we get a composite map of , whose image we denote by . It has a local basis corresponding to the basis of . In term of the -matrix, we have
As are closed, is integrable. Let denote the quotient . Then we have an isomorphism
(19) |
given explicitly by
(because , residues with respect to the two branches of ). Now set , an invertible sheaf on . Then is the direct image of a complex on :
(20) |
where a local generator of has the form and the differential has the form
3.4.4. Zeroth differential
Using the identification (19), the zeroth differential has the form
(21) |
The form has zero residues with respect to , hence yields a form in . Changing the local equations changes by adding a holomorphic (pole-free) form on .
For nonzero (21) can be rewritten
(22) |
When does this operator have a nontrivial kernel? First, if is constant then on which is im[possible if meets . Next, locally at a point , clearly holomorphic and nonzero in the kernel exists locally since is closed and holomorphic so for a holomorphic function and we can take . Moreover nonzero solutions to differ by a multiplicative constant. The condition that the local solutions patch is clearly that be an integer for any loop in . Now is defined only modulo a holomorphic form on while is generated modulo by small loops normal to components of , So the relevant condition is just integrality over such loops .
At a simple point of , the condition that exist locally as a holomorphic function with no pole on is clearly that for as above, oriented positively, the integer is nonnegative, so that has no pole on . In other words, that the sum of the first 2 columns of the matrix, normalized so that , should be a nonnegative integer vector. Finally by Hartogs, if is holomorphic off the singular locus of , it extends holomorphically to .
3.4.5. Special components
Now let be a component of and assume and are both normally split so that the branches of along may be labelled 12 while those along may be labelled 123. Thus branches of over are labelled 12, 23, 31 and the preceding discussion shows that the zeroth differential has nontrivial kernel along only if the iterated residues of along these branches, denoted , assuming , satisfy
(23) |
We call such a component special; then is said to be special if every (normally split) component of is special.
What about the normally split hypothesis? Suppose first is contained in a connected open set which is not normally split. Then as is locally constant in it follows that , i.e. is not residual. Now suppose is contained in open connected and not normally split. Then monodromy acts on the branches of along cyclically and consequently the above are all equal. Then (23) holds automatically with so is special.
3.4.6. Conclusion
What we have so far proven is the following: if is a normally oriented residual open subset of of then the stalk of the zeroth cohomology vanishes somewhere on unless either
(i) , or
(ii) is special.
Note that if the stalk of vanishes somewhere in , then because is coherent and torsion-free, it follows that , hence a similar vanishing holds for the entire component of containing . Now recall that, minding the index shift, if then the cokernel of the inclusion has vanishing (and ). On the other hand, it is well known (see e.g. [14]) that controls deformations of or where deforms locally trivially, and those deformations are unobstructed thanks to Hodge theory.
Summarizing this discussion, we conclude:
Theorem 6.
Let be a log-symplectic manifold with polar divisor . With notations as above, let
Then the inclusions
induce isomorphisms on and injections on , hence isomorphisms on and injections on , unless either
(i) has a non-residual component; or
(ii) has a special component.
As noted above, any component of that is disjoint from , i.e. contains no triple points of , is automatically non-residual.
Corollary 7.
Notations as above, if is compact and Kählerian and conditions (i), (ii) both fail, then the pair has unobstructed deformations and the polar divisor of deforms locally trivially.
In the case where has global normal crossings, i.e. is a union of smooth divisors, this result also follows from results in [9], which also states a partial converse: when is not a quasi-isomorphism, has obstructed deformations and admits deformations where either smooths or deforms locally trivially.
Example 8.
(Due to M. Matviichuk, B. Pym, T. Schedler, see [9], communicated by B. Pym) Consider the matrix
(24) |
and the corresponding log-symplectic form on , and corresponding Poisson structure , both of which extend to with Pfaffian divisor , hyperplane at infinity. Then admits the 1st order Poisson deformation with bivector , which in fact extends to a Poisson deformation of over the affine line , and the Pfaffian divisor deforms as , hence non locally-trivially. Correspondingly, the log-plus form is closed ( and not exact). That corresponds to the integral column relation
where the and are the columns of the matrix and the identity, respectively, showing that and are residual triples of type II and (12), i.e. is a special component of .
Remark 9.
As we saw above, the presence of monodromy on the branches of is related to non-residual or special components. This suggests that log-symplectic manifolds with irreducible polar divisor may often be obstructed. However we don’t have specific examples.
References
- [1] I. Mǎrcut and B. Osorno Torres, Deformations of log symplectic structures, J. London Math. Soc. (2014).
- [2] F. A. Bogomolov, Hamiltonian Kählerian manifolds, Dokl. Akad. Nauk. SSSR 243 (1978), no. 5, 1101–1104.
- [3] N. Ciccoli, From Poisson to quantum geometry, Notes taken by P. Witkowski, avaliable on http://toknotes.mimuw.edu.pl/sem4/files/Ciccoli_fpqg.pdf.
- [4] J.-P. Dufour and N. T. Zung, Poisson structures and their normal forms, Prog. Math., vol. 242, Birkhauser, Basel- Boston- Berlin, 2005.
- [5] V. Ginzburg and D. Kaledin, Poisson deformations of symplectic quotient singularities, Adv. Math. 186 (2004), 1–57, arxiv.org/0212279v5.
- [6] R. Goto, Rozanski- Witten invariants of log-symplectic manifolds, Integrable systems, topology, and Physics, Tokyo 2000, Contemp. Math., vol. 309, 2002.
- [7] N. Hitchin, Generalized Calabi-Yau manifolds, Quarterly J. Math. 54 (2003), no. 3, 281–308.
- [8] R. Lima and J. V. Pereira, A characterization of diagonal Poisson structures, Bull. Lond. Math. Soc. 46 (2014), 1203–1217.
- [9] M. Matviichuk, B. Pym, and T. Schedler, A local torelli for log symplectic manifolds, Arxiv.math (2020), no. 2010.08692.
- [10] Y. Namikawa, Flops and poisson deformations of symplectic varieties, Publ. RIMS, Kyoto Univ. 44 (2008), 259–314.
- [11] A. Polishchuk, Algebraic geometry of Poisson brackets, J. Math. Sci. 84 (1997), 1413–1444.
- [12] B. Pym, Constructions and classifications of projective Poisson varieties, Arxiv:1701.08852v1 (2017).
- [13] Z. Ran, Deformations of holomorphic pseudo-symplectic Poisson manifolds, Adv. Math. 304 (2017), 1156–1175, arxiv.org/1308.2442.
- [14] by same author, A Bogomolov unobstructedness theorem for log-symplectic manifolds in general position., J. Inst. Math. Jussieu (2018), Erratum/corrigendum (2021):; arxiv.org/1705.08366.