If you check out this page: http://www.solstation.com/habitable.htm
There's a section there about stable orbits in binary systems, with a diagram from a paper by Trilling et al showing how the Spitzer telescope detected disks around multiple stars ( http://www.solstation.com/images/bi2sep.jpg ). I've read the paper, and it says that in most cases you'd have a disk around both stars if they're close together, a disk around only one star if they're widely separated, and no disk around either star if they're in between.
Problem is, that doesn't tally at all with existing orbital dynamics. Theoretically at least, you should be able to have stable orbits around each star in an intermediate system (assuming the stars don't get too close to eachother in their orbits, that is), or each star in a widely separated system. (see e.g. Long-Term Stability of Planets in Binary Systems (Holman, Matthew J.; Wiegert, Paul A.) The Astronomical Journal, Volume 117, Issue 1, pp. 621-628).
But the results from Spitzer seems to be implying that this doesn't happen, and I can't see a reason why you cannot have planets in stable orbits around each star. Right now I'm running a simulation in Gravity Simulator of two stars orbiting eachother with a semimajor axis of 30 AU and a eccentricity of 0.5, with asteroid belts containing 100 objects around each star between 0.5 and 10 AU. Initially a bunch of asteroids get stripped off from each star during closest approach (which I expected), but after about 12 orbits (1200 years or so) they seem to have stabilised, with each star having a smaller remaining bunch of asteroids in the closer and more stable orbits around them beyond the other star's reach. So while that's a pretty short timescale, unless something changes drastically it doesn't look like both stars are in any danger of losing their disks entirely (right now, it looks like the more massive star has 50 objects in orbit around it, and the less massive star has 40. So I guess if you turn that into percentage of material remaining, one is down to 50% and the other is down to 40%, but they don't seem to be losing much more material because they've already lost most the asteroids in the least stable orbits during the earlier perihelia).
So are we missing something here? The Spitzer results seem to be somewhat interpretative - they're not actually detecting the disks directly as far as I can tell, they're just detecting an IR excess that implies there are disks there. So could the excess IR signature be indicative of something else? Or could there be systems with fully-formed planets around them and no detectable IR excess because there's not enough excess dust remaining in the system? (as a secondary point, if we were looking at the Sun from outside the system from say 10 pc away, how much dust would we be able to detect around it? Would any excess IR show up at all?)
My suspicion is that planets are possible around all these orbital configurations, so long as the orbital dynamics allows them to be there - there's no reason that I can see why they shouldn't be able to exist within specific orbital ranges otherwise. If we're seeing excess IR from dust then that's a bonus confirmation (well, at least it shows there's dust and planet-forming material there, implying that there could be planets too), but are we really justified in definitively saying "if there's no excess IR, there's no planets".
Anyone got any thoughts on this?