In this program our goal is to measure the total light received from the star+planet system,
as accurately as possible. The host star of the planet is assumed to be spherical, so that the light we receive directly
from the star is constant with time. Superimposed on top of this constant level is a small fluctuation (about one part
in a thousand) that is due to the planet. This is because, as the planet orbits the star, first it exposes the
bright side (that which faces the host star) and then the night side (that which faces away from the star). Thus, we
need to perform a very accurate measurement, as the variation is only a tenth of one percent of the total light
we receive from that star.
In order to do an accurate measurement, it is very useful if you have some reference to which you can compare the object you are measuring. The availability of such a source is what dictated our choice of instrument and wavelength for this observation, because observing at 24 microns means that we can use the Zodiacal Light as a reference source. Our own solar system contains a cloud of dust, which absorbs light from the sun and then shines in the infrared. By measuring the star relative to this Zodiacal reference source, we can perform a more accurate measurement than would be possible at other wavelengths. One subtlety associated with this procedure is that we have to account for the fact that the level of the Zodiacal light changes slowly over time because the telescope is not stationary, but rather orbiting the sun, so that the line of sight passing through the dust cloud to the object in question is slowly changing.
An important point to note about this method is that it is, in principle, possible to make this kind of measurement for every star that harbours a close-in giant planet. Unlike previous measurements of temperatures on extrasolar planets, which rely on the fortuitous geometrical alignment necessary to observe an eclipse, the flux variations that result from day-night temperature differences can be seen for a wide range of orbit orientations with respect to the observer. Only when the orbit is almost face-on, so that the day side and night sides are equally projected onto the plane of the sky, will there be no signal. This is an important point because it means we can observe the very closest and very brightest stars with such planets, thereby achieving the most accurate measurements (the brighter a star is, the stronger the signal relative to the noise and hence the more accurate the measurement).Brad Hansen