Blurry Eyes and Stellar Contamination in the Kepler Light curves

The light curves you see on Planet Hunters are not always the light of a single star. Kepler has very very precise but blurry vision. The CCD pixels on Kepler’s focal plane are very big, four arcseconds to be exact. The light measured at each observation from several of these pixels are added together to create the light curve you see on Planet Hunters. So what does this exactly mean? In some cases the Kepler stars are pretty isolated, but in others there are fainter background stars that appear nearby in the sky can get blended with the light from the Kepler target star. It turns out you can hide a lot within 4 arcseconds.

This stellar contamination can impact what we see in the final light curve. If the main Kepler star has a transiting planet, the contaminating star can dilute the transits. The transits will look shallower than they really are, and you’ll estimate a small planet radius. Sometimes the fainter contaminating star is an eclipsing binary. Combined with the light from the brighter Kepler target star, the stellar eclipses from the eclipsing binary are diluted. The secondary eclipse (when the fainter cooler star goes behind the larger brighter star and the smaller cooler star’s light is blocked out) can be diluted such that it’s not seen and the primary stellar eclipse (when the smaller cooler star transits in front of the larger brighter star and blocks out a portion of the brighter star’s light) get shallower, looking like a planet transit. Other times depending on the brightness of the eclipsing binary, it will look like the main Kepler target is the eclipsing binary when it’s not.

This is something the Kepler mission always had to deal with and there are some observational checks and data tests that can help determine whether the transit-like signal is likely coming from the actual Kepler target star. You can take follow-up observations like we did for PH1 b and PH2 b using telescopes with adaptive optics that minimize the blurring effects of the Earth’s atmosphere to zoom in around the Kepler target star to look for contaminating stars. Also you can look for shifts in the position of the brightest pixel during and before and after a transit which signals the transit signal isn’t coming from the primary Kepler target star.  Also you can look at the individual pixel by pixel light curves from Kepler  (Kepler reads out a subimage around each target star and a small number of those pixels get added together to make the Kepler light curve)and see if the transit signal or eclipsing binary signal is present in every pixel or if you see say an eclipsing binary signal in one pixel making the light curve and in pixels near by around a different star. Here’s an example from some of the Planet Hunters volunteers who examined to see if an eclipsing binary was contaminating a light curve.

Despite Kepler’s slightly blurry eyes, we can use a host of techniques to try and rule out false positives, identify where there is stellar contamination, and still find planets. So bear this in mind when you see the light curves, that although it’s likely most of the star’s light is from the Kepler target star, a tiny portion (in most cases) is contributed by neighboring stars.

Source contaminating the main Kepler target that confirmed circumbinary planet PH1 b orbits (the contaminating source itself happens to be a visual binary)

 

Advertisements

2 responses to “Blurry Eyes and Stellar Contamination in the Kepler Light curves”

  1. Art says :

    Is that true that NASA has troubles contacting Kepler spacecraft right now?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: