9
$\begingroup$

Let $-CP^{2}$ denote the complex projective surface $CP^{2}$ with the reverse orientation. I have seen some results about the existence of symplectic structures on the connected sums $\#_{l}CP^{2}\#_{k}(-CP^{2})$ for some positive integers l,k.

My question is whether there is a complete result which can decribe the existence of symplectic structures on the connected sums $\#_{l}CP^{2}\#_{k}(-CP^{2})$ for all positive integers l,k.

$\endgroup$

1 Answer 1

12
$\begingroup$

For $l=1$, these are blowups of the projective plane, which are all Kaehler and hence symplectic. For $l>1$, these do not have symplectic structures. For if $l$ is even, then they don't even have almost complex structures (cf. this MO thread), which symplectic manifolds certainly do. If $l>1$ is odd, then their Seiberg-Witten invariants would vanish, since your manifolds decompose as a connected sum into pieces with positive $b_2^+$. But a famous theorem of Taubes says that there is a non-vanishing Seiberg-Witten invariant.

Finally, if $l=0$, then $b_2^+ = 0$, so your manifold cannot be symplectic, since the cohomology class of the symplectic form has positive square.

$\endgroup$
8
  • $\begingroup$ "The cohomology class of the symplectic form has positive square" : I wonder what is your definition of a symplectic strucuture. For me, a symplectic form remains symplectic if one changes the orientation. In particular, the answer should be positive for $k=1$ and all $l$. $\endgroup$
    – abx
    Dec 12, 2013 at 14:44
  • $\begingroup$ @abx But a symplectic form canonically defines an orientation of the underlying manifold. If you had already chosen a preferred orientation of your manifold, it makes sense to look for symplectic forms inducing the same orientation. There are no symplectic forms on $CP^{2}$ which would induce an orientation different from the standard one. $\endgroup$ Dec 12, 2013 at 15:48
  • 1
    $\begingroup$ I agree, but then the question should be slightly modified (… a symplectic structure compatible with the orientation). $\endgroup$
    – abx
    Dec 12, 2013 at 15:53
  • 1
    $\begingroup$ Thank a lot for a clear answer! Moreover, I'd like to know how about considering all the smooth manifolds homeomorphic to $\#_{l}CP^{2}\#_{k}(-CP^{2})$ for $l$ is odd. $\endgroup$
    – user44052
    Dec 13, 2013 at 8:13
  • 1
    $\begingroup$ That's a much more complicated answer; there are some that have symplectic structures, some that don't. By the argument I outlined above, you should assume that your manifold is irreducible, ie not a connected sum of manifolds with $b_2^+ >0$. There is a very large literature on the subject, starting with [Szabó, Zoltán, Simply-connected irreducible 4-manifolds with no symplectic structures. Invent. Math. 132 (1998), no. 3, 457–466.] I would also look at [R. Fintushel and R. Stern, Knots, links, and 4-manifolds, Invent. Math. 134 (1998), 363–400.] $\endgroup$ Dec 13, 2013 at 14:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.