If "some" show he publishes alone, that implies "some" show he collaborates. Why are you focusing on the one thing and not the other?
As for why a person publishes solo:
Maybe they're shy, suffer from impostor syndrome, etc., and this makes it inordinately difficult for them to suggest the idea to others even if they really want to. Maybe they try to collaborate, but then the collaborators contribute nothing for months by which time they've completely solved the problem and so publish solo. Maybe this is a highly specialized topic, and there are only a handful (or less) of potential collaborators; it'd be quite easy for them to all be busy with other matters (such as some other ongoing collaboration, committee work, etc.). Maybe they didn't need anyone else for that topic: how to proceed with the problem came quickly and clearly to them, and there was never any need to bring in another perspective. Et cetera, et cetera.
The short of the matter is: the correct thing for you to do here is to ask. As long as your topic seems reasonably within their skill set there's really nothing to lose. (Asking a PDE's guy to work on a group theory paper probably isn't smart if you don't have a very specifically good reason) Long as that's the case, worst case scenario is you lose a couple of minutes typing up an e-mail. And even then the person probably is happy to know someone finds his work and skills interesting, and may seek you out for potential future collaborations.