Monthly Archives: November 2014

How Planets Make Stars More Magnetically Active

My thesis was on planet detection and stellar magnetic activity, so I’ve been following the progress of detections of star-planet magnetic interactions pretty closely over the years.  It started for me when Evgenya Shkolnik published her work on the HD 179949 system, where she detected a modulation of the activity of the star in phase with the giant planet orbiting it.

This was quite a surprise to me.  Tiny little planets don’t seem like they should be able to have much of an effect on the stars they orbit, but in principle one can see how it would work for very close-in, giant planets like the hot Jupiters:  the planets have magnetic fields, which collide with the stars’ fields as they go around.  If the planet is close enough to be encountering very strong stellar field lines, then it will “pluck” these field lines like the strings of a guitar.  The resulting waves will travel down along the field lines to the chromosphere of the star, where it will be dissipated as heat, and that hot gas will cool by emitting radiation via calcium ions in the H & K Fraunhoffer lines (a subject of of my thesis).   

Image from Science Magazine

Image from Science Magazine

Evgenya would go on to detect the effect in the υ Andromedae system as well, and I was one of four Penn Staters on Stephen Kane’s team that tried this for the planet host star HD 63454 (we didn’t find anything.)  But, at least with HD 179949, it seems to be a frustratingly intermittent effect: sometimes it’s there, sometimes its not, which makes it hard to confirm or refute the claim that this is going on in any system.  

And we really want to know!  The effect would teach us about the magnetic fields of these planets, which is not something we know very much about.

Another approach is to try to determine whether stars that host hot Jupiters have excess emission on average.  The effect that Evgenya detected was very small, but detectable because it was modulated at a specific frequency.  A more general claimed effect is that hot Jupiter hosts consistently give off more X-ray and/or chromospheric emission due to interactions with the planet, far in excess of what would be generated by their rotational dynamos.

This is a much trickier claim to evaluate, because of selection effects.  Active stars aren’t great radial velocity targets, so we rarely find small or distant planets around them (which is to say that we generally ONLY find hot Jupiters, because we are not sensitive to anything else).  Also, more massive stars tend to be more active (because, having shorter lifetimes than less massive stars, they tend to be younger, and activity decreases with age).  More massive stars are brighter, and so more commonly the targets of ground-based transit searches, and those searches are only sensitive to hot Jupiters.  So you can’t just look at the activity levels of hot Jupiter hosts and say “hey, look, they’re more active than stars that host other kinds of planets!”  You would expect that even in the absence of any star-planet magnetic interaction.  You have to much more careful than that, and different authors have come to different conclusions when trying to untangle these effects.

For a good review of the topic, I recommend Katja Poppenhager’s recently posted review article here (in a small world-ism, Katja was the REU adviser for my student, Katherina Feng, when she was at CfA).

Brendan

Brendan Miller, now of Macalester University

Back in 2011, Penn State grad Brendan Miller contacted me (I didn’t know him before that) about a Chandra proposal he was submitting with Elena Gallo at the University of Michigan.  Brendan’s background was in X rays from radio-loud quasars (I always found that term technically redundant), but he was interested in testing this star-planet interaction issue.  Was I interested in collaborating?  I said yes, and we studied WASP-18, which should have been a “slam dunk” case of star-planet interaction… and found nothing.

We continued working on the project, and Brendan got more Chandra time to follow up many more systems.  The trick he used was to sort the systems not just by whether they had a hot Jupiter or not, but by how much star-planet interaction we expect, based on scalings of the planets’ masses and the stars’ magnetic fields.  This way, if the higher X ray activity of hot Jupiter hosts was really caused by the planets, one would expect it to track roughly with the degree of interaction.  If it’s just a selection effect, then there should be no correlation.

And indeed, that’s what Brendan finds.  Dividing our sample of hot-Jupiter-bearing stars into those with strong and weak potential interactions, he finds no correlation, or any difference at all between the two.

Screen Shot 2014-11-13 at 8.14.09 PMLooking at a broader sample of 198 FGK planet hosts, a correlation did jump out, and with high significance (99%), thanks to a small number of extreme systems (not a broad trend).  But in these extreme systems, the stars themselves look to be inherently, not anomalously, active.  That is, they are fast rotators, with lots of calcium emission all of the time (not modulated by the planet’s orbit), and so are expected to have correspondingly high X ray emission.

So what’s going on?  It looks like it’s a second-hand star-planet interaction due to gravity, not magnetic fields.  That is, there is growing evidence that close-in planets can actually spin up their parent stars’ outer envelopes, which drives the magnetic dynamo up and increases X-ray and calcium emission.  Katja Poppenhager and Scott Wolk found this to be the case earlier this year by comparing stars in binary systems where one component has a hot Jupiter.  They find that the hot-Jupiter-hosting star typically looks “younger” (that is, spins faster and has more magnetic activity than you would expect given the state of its sibling star).

The poster child for this for me has been the τ Boötis system.  The host star is rotating very rapidly — at the same period as its hot Jupiter companion, actually.  This had always puzzled us — was it a weird coincidence?  Surely that tiny little planet couldn’t spin up the entire star!  But a quick check of the angular momentum and energy budget, especially when you consider that you only have to torque the outer envelope, shows it’s not impossible.  Apparently, it’s common!

So star-planet interactions look like they ARE important for magnetic activity measurements, but not for the reasons we originally thought.

Brendan’s paper describing his comprehensive statistical assessment of star-planet interaction is on the arXiv here.  Go check it out!