Thursday, February 28, 2008

More on Flares

In my last post discussing my research for this semester, I explained that, in order to determine whether or not massive, close in Jupiter like planets could induce super flares on their parent start, I have two available paths:

1) Determine whether superflare stars have planets.
2) Determine if stars with planets have superflares.

The trouble with the first is that looking for planets isn't exactly an easy task. The main method requires good spectroscopy over the length of the entire orbit. For these hypothetical massive, close in planets, it shouldn't be too hard, but that's still telescope time we'd have to apply for (or ask someone else to tack on to their observing run).

Another problem is that so few stars have been observed to have these flares, we're still dealing with small number statistics which makes convincing correlations difficult.

The second approach also has problems. We avoid the problem of not having many candidates since there's nearly 250 known planetary systems, but we run into another: We'd have to be watching all these stars and hope we catch it during a flare. Currently, we have no idea how often such stars might flare, making it an even more daunting task.

So the second way isn't much better unless we can find some way to make a substitute argument about whether or not the star will flare.

It should be pretty obvious that stars with more active atmospheres will be more prone to flares, whether large or small. Thus, if we can find a quick way to measure the photospheric activity, and that activity gets larger when you have close in planets, we can support the hypothesis.

As luck would have it, there's a nifty trick to determine the photospheric activity. It's called the S value for the star and has to do with the relative strengths of two main calcium emission lines (the H & K lines) in the star's spectra. The larger this value is, the more active the star's outer layers. The more active the outer layers, the more likely the star will be to flare (big or small), even if we don't observe the flare directly.

So the obvious next step would be to take all the known extra solar planetary systems, and plot up the S value against the semi-major axis of the known planets. If the planets are indeed having a strong enough effect to induce superflares on the star, we should see the S value increase as the planets semi-major axis gets smaller.

Although S values have been found for well over 3,000 stars, they haven't been found for many of the known planetary systems. After dropping all the systems that didn't have stars with spectral classes within 1/2 a spectral class of the sun, we were left with 15 stars that had S values. Not great numbers, but hopefully enough to see if something exciting was going on.

So after spending 3 hours looking for all these values, and plotting it up, what did it look like?

Nothing.

No noticeable correlation what-so-ever.

So what does it mean?

Well, we can't say just yet.

Before we can really be certain, we need to cross check this against the extreme cases, both looking at average S values for solar type stars that don't undergo massive flares as well as looking to make sure there's actually a correlation between the semi-major axis and the S value for the RS CVn stars. If there is, then we know the results of this test actually mean something. If not, then this entire angle on it wasn't any use and we'll have to find some other way to approach the problem.

In the meantime, we're also looking at either asking a planet hunter to glance at these superflare systems to search for planets, or doing it ourselves. Other options would involve going through images from some of the massive sky surveys to see if any of the stars with known planets have ever flared while those images were being taken, and also to see if these systems we're looking at have flared again.

5 comments:

Randy said...

OK, I would think you would have covered this already amongst yourselves, but it's not clear from your post or my own background. Is the mechanism for the connection between these high-S-value stars and their heightened flare activity known? If not, or if the S-value is a marker rather than a cause, it would seem to me that you might well have two causes for photospheric activity, one related to the S-value, and another one completely unrelated to S-value involving planets. If the S-value causes the activity and flares in some way, all the superflares in particular, then it would seem to me you'd made this hypothesis rather unlikely.

Jon Voisey said...

The S-value is directly correlated to the activity of the stellar atmosphere. In really active stars (K and M class) it can be very high. In solar type stars, it's pretty low.

The mechanism is that the additional turbulence in the atmosphere of the stars excites the Calcium II in the outer layers, allowing us to see the emission lines.

The idea here is that if planets actually can get close enough and actually interact enough to cause these flares, they'd have to do so by tangling up the magnetic field. If they tangle up the magnetic field, that puts more energy into the atmosphere, causing the CaII lines to strengthen.

So the S value doesn't cause flares, it's just a sign that flares are likely. If flares (and therefore the S values) are more likely when there are close in planets, then we've supported the hypothesis. But you're right. From what we have so far, it looks like the hypothesis is on shaky ground. But there's still several things to ask.

Since it looks like that's not the case that S is inversely proportional to the semi major axis, from the graph I showed, we have to check to make sure for the RS CVn systems this approach is even valid. Another thing that we're going to have to do (especially for the RS CVn systems) is try to find a way to estimate magnetic field strength from the companion instead of just the semi-major axis since stars magnetic fields vary greatly.

K. Rice said...

I stumbled onto this blog and really enjoyed reading the 10 most recent posts or so. You've got a great style and I find myself intrigued by your science (though I am more of a biology person myself). Best of luck with your research. :)

K. Rice said...

I stumbled onto this blog and really enjoyed reading the 10 most recent posts or so. You've got a great style and I find myself intrigued by your science (though I am more of a biology person myself). Best of luck with your research. :)

Jon Voisey said...

The S-value is directly correlated to the activity of the stellar atmosphere. In really active stars (K and M class) it can be very high. In solar type stars, it's pretty low.

The mechanism is that the additional turbulence in the atmosphere of the stars excites the Calcium II in the outer layers, allowing us to see the emission lines.

The idea here is that if planets actually can get close enough and actually interact enough to cause these flares, they'd have to do so by tangling up the magnetic field. If they tangle up the magnetic field, that puts more energy into the atmosphere, causing the CaII lines to strengthen.

So the S value doesn't cause flares, it's just a sign that flares are likely. If flares (and therefore the S values) are more likely when there are close in planets, then we've supported the hypothesis. But you're right. From what we have so far, it looks like the hypothesis is on shaky ground. But there's still several things to ask.

Since it looks like that's not the case that S is inversely proportional to the semi major axis, from the graph I showed, we have to check to make sure for the RS CVn systems this approach is even valid. Another thing that we're going to have to do (especially for the RS CVn systems) is try to find a way to estimate magnetic field strength from the companion instead of just the semi-major axis since stars magnetic fields vary greatly.