Selecting Planet Candidates
Once the data has been reduced and de-trended, and Stage 1 candidates have been selected (any objects that display periodic transits) and the light curves have been phase-folded, pull up a DSS image of the object on DS9 for comparison to APT images. Eliminate any candidates that meet the following criteria (links go to examples of light curves that meet criteria):
* A secondary transit is visible;
* Fewer than 3 transit events;
* Transit depth is greater than 150 mmag;
* Phase folded light curve contains significant sampling gaps;
* A period less than 1 day;
* Light curve contains high ellipsoidal variations;
* 3 or more stars of comparable brightness within the APT aperture; or
* A period cannot be determined;
* The I-band magnitude of the star is < 8.0.
If you are not sure the candidate has a secondary transit, time bin the curve (myplot #### 15. This will bin the light curve in 15 minute intervals) and check again. For a circular orbit, the secondary transit should occur at Phase 0.5.
When the light curve is phase folded, a simple fit is used to estimate the basic parameters of the light curve, including the transit duration, depth, and duration of the flat bottom. The will output derived physical parameters using the equations of Seager & Mallen-Ornelas (2003). These estimates include the star density, the impact parameter, the star mass and radius, the orbit radius, the orbital inclination, and the planet radius.
If there are no other bright stars within the APT aperture, the given parameters can be used to estimate the star's spectral type. Using the estimate of the star density, determine the spectral type using the graph below. If there are brighter stars, the light curve fitting algorithm can be used to determine the amplitude depth change due to the blended stars. The depth of the transit in the simple fit (discussed above) can be adjusted and those parameters used.
Do the derived parameters, namely the planet radius and star density, coincide with a planet transit? If not, flag the candidate. Using the estimated transit parameters and a constant planet radius of 1 R_Jup, find the Tingley & Sackett (2004) diagnostic np, where Z = 1, d is the transit depth, D is the observed transit duration, P is the period, and Rp is the planet radius:
Flag any candidates that have a value of np > 1.0 as lower priority.
From the remaining candidates, you must collect catalog information on the target and any fainter stars that are within 4 magnitudes of the brightest star that fall within 30 arcsec of the center of the APT aperture. This data includes the BVRI and JHK magnitudes (from the NOMAD catalog), the proper motion (with errors), a given spectral type (if any), and any special or relevant data (is it an X-ray source, a known binary, etc).
Using the BVRI and JHK magnitudes, find the colour index of each and compare them to the Intrinsic Colour catalogs (to the left under "Links") to estimate the spectral type of the star. These colour relations include (B - V), (J - H), and (H - K).
Next, use the light curve fitting algorithm to fit the light curve to a range of stellar spectral types. Does the fitted estimate agree with the colour index estimate of the star? Using the results of this fit, eliminate any candidates that pass the following criteria:
* Planet radius estimate > 2 R_Jup;
* S/N of ellipsoidal variations > 8.0;
We should now have 4 estimates of the star radius;
1) the estimate using the colour index,
2) the estimate using the Seager & Mallen-Ornelas equations,
3) the estimate using the derived star density, and
4) the estimate using the light curve fit.
If the estimates of the star's spectral type all agree, the estimated planet radius is < 2 R_Jup, and the diagnostic gives a value np < 1, consider the candidate high priority. If the estimates of the star radius do not all agree, but some give a reasonable planet radius estimate and a disgnostic np ~ 1, consider the candidate medium priority. If none of the values agree and the planet radius is >> 2 R_Jup, discard the candidate.