To achieve the required photometric precision to find terrestrial-size planets, the photometer and the data analysis system must be designed to detect the very small changes part in 104) in stellar flux that are characteristic of transits by terrestrial planets. The Kepler mission approach is best described as "differential relative photometry." In this approach:
• Target stars are always measured relative to the ensemble of similar stars on the same part of the same CCD and read out by the same amplifier;
• Only the time change of the ratio of the target star to the ensemble is of interest. Only decreases from a trend line based on a few times the transit duration are relevant (long-term stability of the trend is not required);
• Target star and ensemble stars are read out every five seconds to avoid drift and saturation; and
• Correction for systematic errors is critical.
Photometry is not done on the spacecraft. Instead, all of the pixels associated with each star image and the collateral, bias, and smear pixels are sent to the ground for analysis. This choice allows many different approaches to be used to reduce systematic errors.
The spacecraft will be placed in an Earth-trailing heliocentric orbit by a Delta II 2925-10L launch vehicle. The heliocentric orbit provides a benign thermal environment to maintain photometric precision. It also allows continuous viewing of a single FOV
for the entire mission without the Sun, Earth or Moon obtruding. Only a single FOV is monitored during the entire mission to avoid missing transits and to maintain a high duty cycle.
A pattern of at least three transits, showing that the orbital period repeats to a precision of at least 10 ppm and showing at least a 7a detection, is required to validate any discovery. A detection threshold of 7a is required to avoid false positives due to random noise. To obtain a higher recognition rate, the mission is designed to provide a lifetime of four years to allow four transits in the HZ of a solar-like star to be observed. Note that transit signatures with a mean detection statistic of 8a will be recognized 84% of the time, whereas those with a mean of 7a will be recognized only 50% of the time.
Classical signal-detection algorithms that whiten the stellar noise, fold the data to superimpose multiple transits, and apply matched filters are employed to search for the transit patterns down to the statistical noise limit (6). From measurements of the period, change in brightness and known stellar type, the planetary size, and the semi-major axis, the characteristic temperature of the planet can be determined. The latter gives some indication of whether liquid water could be present on the surface; i.e., whether the planet is in the habitable zone.
Because of the limitations on the telemetry stream, only data from those pixels illuminated by the pre-selected target stars are saved for transmission to Earth. Data for each pixel are co-added onboard to produce one brightness measurement per pixel per 15-minute integration. Data for a subset of target stars can be measured at a cadence of once per minute. This option will be exercised to obtain detailed emersion-immersion profiles, for detecting changes in transit timing due to the presence of multiple planets, and for conducting observations for astroseismology.
Was this article helpful?