5. Approximate hyperkähler metrics [02HG]
Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.
Complete original source context · Original author HTML
5. Approximate hyperkähler metrics
In this section we patch together ALF gravitational instantons and the incomplete hyperkähler structure of the previous section to construct a closed definite triple which is approximately hyperkähler. In the next section we will use analysis to deform into a genuine hyperkähler structure for each sufficiently small.
5.1. The –manifold
Let be the hyperkähler triple defined in (4.6). By Lemma 4.9 for small enough we think of as defined on , a smooth manifold with boundary obtained by restricting the line bundle to the complement of (arbitrarily) small balls centred at the punctures and then taking the quotient by the involution . The boundary of has components, each of which has a collar neighbourhood diffeomorphic either to for , or for .
For each let be the smooth –manifold underlying a ALF space and for each let be the smooth manifold underlying an ALF space. We construct a smooth –manifold by cutting the ends of and and gluing the resulting manifolds with boundary to in a neighbourhood of or , respectively.
We will construct an approximate hyperkähler structure on in the next subsection. Here we pause for a moment to determine the Betti numbers of . While we will not use this result in an essential way in the rest of the paper, it is interesting to note how the balancing condition (4.1) appears naturally in the calculation of the Euler characteristic of .
Proposition 5.1.
The Betti numbers of the compact orientable –manifold are
Proof.
Decompose into the union of a piece , an ALF space for each and a ALF space for each . These pieces are identified along their common boundaries, which are homology spheres. Since all components have vanishing first Betti number, the reduced Mayer–Vietoris sequence yields . The Euler characteristic is also easily calculated:
by the balancing condition (4.1).
It remains to calculate the signature . Below we will construct a definite triple on which is close to define a hyperkähler structure. By changing basis of one can always deform this triple to a genuine –structure (without requiring any differential constraint). In particular, can be endowed with an almost complex structure with . Since , Hirzebruch’s Signature Theorem and the equality of characteristic classes yield . ∎
Remark.
In Remark 4.8 we noted that the case where and for all reduces to the usual Kummer construction. Hence we know that is diffeomorphic to the K3 surface in this special case. It seems likely one can prove that the diffeomorphism type of does not depend on the configuration of punctures satisfying the balancing condition (4.1). Since we are going to construct a hyperkähler metric on , the calculation of the Betti numbers will anyway imply that is always diffeomorphic to the K3 surface.
5.2. The definite triple
We are now going to define an approximately hyperkähler triple on the –manifold .
For each denote by the hyperkähler triple obtained from the Gibbons–Hawking ansatz (3.3) using the harmonic function
| (5.2) |
By abuse of notation we think of as defined both on the circle bundle as well as on its quotient by the involution that acts as the simultaneous standard involution on and the circle fibres.
Similarly, for each let be the hyperkähler triple obtained from the Gibbons–Hawking ansatz using the harmonic function
| (5.3) |
Here and are the constants and linear functions appearing in Lemma 4.7. We will assume that is small enough to guarantee that .
For all let be a complete ALF space. By Definition 3.6 there exists a compact set , and a diffeomorphism such that
for with and similar estimates on the derivatives.
For each let be a complete ALF space. By the classification of ALF spaces of cyclic type [35] the hyperkähler structure on is explicitly given via the Gibbons–Hawking ansatz starting from a harmonic function on with singularities. We can add to this function the smooth harmonic function . Over a ball in of radius much smaller than we can regard the resulting hyperkähler structure as a small perturbation of the ALF hyperkähler structure . By abuse of notation we denote this perturbed hyperkähler structure with the same symbol . The advantage of this modification is that now approaches with a smaller error: by Definition 3.6 there exists a compact set , and a diffeomorphism such that
with (by scaling) and similar estimates on the derivatives.
Remark.
The choice of perturbing the gravitational instanton of type by adding a linear function on can be regarded as an intermediate choice between resolving the singularity of by assuming all punctures have weight and the direct gluing of to . While not strictly necessary, the choice of perturbing by a linear function makes the exposition more uniform. In particular, the closed definite triple we will construct below fails to be hyperkähler by the same amount in a neighbourhood of and .
We will assume that is chosen so small as to make sure that , where was fixed in Lemma 4.7. By choosing larger if necessary we can assume that the harmonic functions and in (5.2) and (5.3) are as close to constant functions as we please for . Then the hyperkähler triples and define metrics and which are uniformly equivalent to in the regions . In the rest of the section all norms and covariant derivatives will be computed with respect to this metric.
A crucial observation is that we can take and to be exact.
Lemma 5.4.
- (i)
For all there exists a triple of –forms on such that
and .
- (ii)
For all there exists a triple of –invariant –forms on such that
and .
Proof.
The proof is identical in the two cases. Set in case (i) and in case (ii). In case (ii) we work with –invariant forms on the double cover .
By scaling we can assume that . It is enough to prove that every closed –form with can be written as with .
Since the restriction of to an exterior domain in is diffeomorphic to with an homology sphere, we can write for some –dependent –form and –form on with .
The condition implies . We then define . The Lemma follows. ∎
By a similar radial integration, Lemma 4.7.(i) implies that in the regions and , respectively, we can write
| (5.5a) | |||
| for triples and of –forms such that | |||
| (5.5b) | |||
for .
Now, let and be cut-off functions with the following properties:
| (5.6) |
We finally define a triple of closed –forms on by
| (5.7) |
Remark.
When there is no need to glue in an ALF space ( endowed with the Taub–NUT metric), since already extends smoothly over . It is however useful (we will use this in setting up the analysis for the deformation problem) to think of a rescaled Taub–NUT space localised around .
5.2.1. The error
We conclude this section by quantifying the failure of to define a hyperkähler structure. Since by construction for all , we only have to check that is a definite triple and estimate the difference between the associated intersection matrix and the identity.
In the regions , and when for all the triple defines a genuine hyperkähler structure. In the transition regions and we have, respectively,
Since and are hyperkähler triples, we conclude that is a definite triple for sufficiently small.