First Planet Detection From Minerva

An artist's impression of the WASP-12 planetary system.
IMAGE COURTESY ESA, NASA AND FRÉDÉRIC PONT, GENEVA UNIVERSITY OBSERVATORY

The Project Minerva team has been hard at work calibrating and characterizing Telescope #1, which is in an Aqawan enclosure on California Blvd, on the Caltech campus in Pasadena. This test rig allows us to work out the kinks with the telescope and enclosure control systems without having to fly/drive back and forth from campus to an observatory. Having a test rig right next to the Minerva team's offices has paid huge dividends thus far. We have overcome a major telescope hardware/software malfunction, identified improper coatings on some of our optical elements, and brought our telescope throughput up to spec, all in far less time than it would have taken if we went straight to the mountaintop.

This past Sunday night, Caltech postdoc Jon Swift and fourth-year PhD candidate Mike Bottom were at the controls of the telescope measuring the amount of starlight that makes it through the telescope optics and onto the actual detector, known as the telescope throughput. To do this, they measured the flux received from several well-characterized standard stars over a range of air masses. After running their tests they found a system throughput of 72%, which is very close to spec (at long last!).

Jon Swift and me giving a tour of the Minerva test facility at Caltech.
Photo courtesy Mike Wong.

With clear, stable skies overhead, they decided to look at a far less stable star named WASP-12. The star has a hot Jupiter that was discovered by the Wide-Angle Search for Planets (WASP) team back in 2009, in a paper led by my collaborator Leslie Hebb (See Hebb et al. 2009 for the gory details). With the planet scheduled to transit WASP-12 in less than an hour, Jon and Mike slewed the telescope to the star's coordinates and set up on the field. The plan was to repeatedly image WASP-12 and the stars within ~10 arcminutes (~1/6 of a degree) on either side of it. 

When photons from the star strike pixels on our CCD array, electrons are freed via the photoelectric effect from the silicon substrate, and the number of free electrons floating in each pixel is proportional to the amount of flux received. By summing the total number of electrons in all of the pixels on which the star's image fell, Jon and Mike measured the star's flux in time. However, due to variations in the transparency in the Earth's atmosphere above the telescope, these absolute flux levels vary. Fortunately, the other stars in the field of view also vary with the atmospheric transparency, so the ratio of WASP-12 to the nearby comparison stars provides a relative flux measurement.

Once the planet blocked the star, the flux level dropped by an amount approximately equal to the radius of the planet squared, $R_P^2$, divided by the radius of the star, $R_\star^2$, or ${\rm depth} = (R_P/R_\star)^2$. Here's what Jon and Mike observed during the 4.8 hour observing sequence:

This is Project Minerva's first planet! Granted, it had already been discovered, but this light curve is ours and we're very, very proud of it. See also Jason Wright's writeup over at his blog.

Our primary science goal is to search for planets using precise Doppler-shift measurements. Our secondary science goal is to search for previously unknown transits of RV-detected planets, and until our spectrometer is ready, we'll be doing a lot of transit work. For our targets we'll know that our transit targets have planets, and we'll have predictions for when transits should occur if the geometry of the orbit is just right. 

This test shows that we can achieve the precision necessary to detect a transiting hot Jupiter. The great news is that Jon did the simplest possible data reduction and analysis procedure, so this is likely an underestimate of our attainable photometric precision. Also, the camera was a loaner; we'll soon purchase a camera better optimized for transit work. Says Jon:

You also might want to add the fun facts that these measurements were made 40 ft from E. California Blvd. where cars whiz by, street lights flicker, and we are losing anywhere from 20 to 50% of our light just from the atmosphere alone!

Our hope is to regularly achieve a photometric precision better than 1 part in a 1000 per minute, or about 1 millimag per minute (For more, check out Greg Laughlin's post on ground-based photometry.)