Hot Jupiters, which I blogged about last week, can be so close to their parent stars that they can exert detectable influences on them. When they are very close, they can slightly elongate the star through tides, and that elongation can make the star’s brightness appear to change as the bulge goes around the star — the effect is tiny but Kepler can detect it.
One method of interaction is through magnetic fields. Stars have strong magnetic fields, as do planets. As a hot Jupiter moves through its host star’s field, it should “tickle” the magnetic field lines. These vibrations should travel down the field lines into the chromosphere of the star, where they should dissipate as heat. Hydrogen excited by this sort of heating emits red light, which is what gives the Sun’s chromosphere its red color (and, thus, its name).
Chromospheres also emit lots of blue light in the blue, a product of ionized calcium. Such emission can tell us how much magnetic activity a star experiences, which in turn tells us about their age and rotation. Or, maybe, their hot Jupiters: if the signal is strongest when the planet is in front of the star and gone when the planet is behind the star, we may be seeing a “hot spot” in the chromosphere due to the planet.
There have been multiple claims and theoretical investigations in the literature about this sort of planet-star interaction’s detectability, not just in the calcium lines but also in X-rays. Some of these claims have always had me scratching my head, though. First of all, the theoretical strength of these signals is very small, but the observed signals, usually near the limits of plausible detectability, tend to be much larger than that. Secondly, the signals seem to come and go. This could be due to “temporal variation” in the signal, but the amount of overall variation in the star’s emission is the same when the signal is “on” as when it is “off”: it seems that when the planet is not interacting, the star has lots of internal variability, but when the planet is interacting the signal is in lock step with the planet, and the star’s intrinsic variability goes away. And why should the signal be sporadic, anyway?
Brendan Miller (a PSU alum, now a postdoc at Michigan) has been working on this problem to settle the issue. He has found the planets most likely to interact with their star — big planets orbiting close to stars with strong magnetic fields — and has pointed Chandra and Magellan at them in an attempt to find the signals there. Brenden hope to find lots of signals so that we can measure the planets’ magnetic fields, but if Brendan can’t find them in these best-case scenarios, this would make detections around ostensibly poorer targets questionable.
The results of the first, best target are in and they are… disappointing, but maybe ambiguous. I worked with Brendan a bit on the paper, which you can find here. The target, WASP-18, should be a slam-dunk case of planet star interaction, but Brendan sees nothing. In fact, Brendan sees no X-rays from the star at all! This likely means that even in the best case scenario, planets don’t interact enough even for Chandra to pick up. The way out is that, despite prior appearances, WASP-18 is an unusually low-activity star, so there’s not much stellar magnetic field to work with in the first place. Maybe there is no interaction because there’s almost no stellar field for the planet to interact with in the first place?
So, are prior detections of the effect in other stars bogus or is WASP-18 just an unexpectedly bad target? The search continues, and other targets are forthcoming. Stay tuned…