HAT-P-32 Ab

image_normal.jpeg

Dr. Ming Zhao, research associate at the Penn State Center for Exoplanets and Habitable Worlds.

Regular readers will recall that Penn State Research Associate Ming Zhao has been working on precise photometry of hot Jupiters, especially in the near infrared.

Ming works at Palomar with the WIRC instrument, and with data from the Spitzer Space Telescope to get data further into the midinfrared that we can do from the ground.

Ming has three secrets for precise photometry from the ground:

  1. Guide carefully.  Not a trivial task with a 200-inch telescope, but Ming worked carefully with Jennifer Milburn to get a control loop going between the guiding and the camera to get the image of the star stable to a pixel or so.
  2. Diffuse the image: usually with defocusing, but soon with holographic diffusers.
  3. Correct for detector nonlinearities

I’ve written before about the first two.  The third one is especially tricky for infrared detectors.  It turns out that they aren’t linear (no surprise there), but that their nonlinearities are themselves sort of nonlinear!  In particular they are flux-dependent.

The technical term for this is reciprocity failure, and it’s apparently a well known feature of infrared detectors among those that work with them closely.  For people attempting sub-millimag photometry, though, it’s a huge pain.

The gist is that a purely linear detector gives you an output that is proportional to the number of photons that hit your detector, and that constant of proportionality is called the “gain.”  This is the number of photons (well, photoelectrons, technically) required to get one “count” out of the detector.

The problem is that as you accrue a lot of electrons, things start to go nonlinear.  In the extreme limit of a “full well” on a CCD, more photons produce no new counts at all (in that pixel) for any of a few reasons, and so the gain becomes infinite (you’ve “saturated” your detector).  But even before then, subtle nonlinearities can set in that make the gain slightly higher for pixels with lots of counts.

No problem, you say, just measure the nonlinearity and divide it out.  The problem is reciprocity failure, which makes the gain dependent not just on the number of electrons already in the pixel, but the rate at which photons are generating new electrons.  That is, a high flux of photons makes the detector more nonlinear.  So a 5 second exposure of a bright source and a 50 second exposure of a 10-times-fainter source collect the same number of photons but produce different numbers of counts.  Evil.

Because of this problem, you can’t do super precise photometry without help.  Stars are sharp dance around, and so the flux of photons onto a pixel varies dramatically from pixel to pixel and even from second to second on a given pixel, so you can’t even count on there being a well defined flux rate for an image.

What Ming did was use the fact that WIRC exposures are short, so the flux in a given exposure is approximately constant.  This let him approximate the flux as simply counts / time, and generate a nonlinearity curve for a given short exposure time.

Screen Shot 2014-10-03 at 1.49.39 PM

This was the final piece of the puzzle needed to characterize the detector, and a procedure that needs to be done on any Hawaii-2 detector (not the RG’s, though).  As Murphy would have it, shortly after doing all that work, Ming was informed that the detector had given up the ghost, meaning we’d need an RG to replace it, or else characterize a new detector all over again.

But with the old detector Ming got great images of HAT-P-32, which is transited by HAT-P-32 b, a hot Jupiter thought to maybe be in a non-circular orbit.  With his new precise secondary eclipse measurements, he shows that the orbit is probably actually circular, after all.

Of course, nothing’s ever that easy.  As Murphy would have it, HAT-32 had to go and be a binary (so the planet is actually HAT-32 Ab), requiring careful subtraction of the “third light” of the system from this M dwarf companion.  Robo-AO and Keck II AO helped Ming’s team determine the nature of the secondary, and many models and isochrones later, Ming can conclude that we now have a good radius for the planet (1.79 ± 0.03 Jupiter radii).

Oh!  And combined with Spitzer IRAC data Ming also gets an SED for the planet, and finds its temperature profile is consistent with a thermal inversion. Which is really cool to know!  (Sometimes it’s easy to forget amid all this detail that we’re measuring the conditions of the atmosphere on another planet 1,000 light years from home!)

The paper has lots more detail, too.  It’s classic Ming Zhao: a thorough piece of work, with lots of modeling, sanity checks, and careful analysis.

You can find it on astro-ph at arXiv:1410.0968.

Leave a Reply

Your email address will not be published. Required fields are marked *