Astron. Astrophys. 338, 479-490 (1998)

1. Introduction

In this paper we describe the observations and analysis that have been performed in a search for photometrically detectable signals from the presence of extrasolar planets around the eclipsing binary CM Draconis. This is the first long-term observational application of the transit method for the detection of extrasolar planets. The transit method is based on observing small drops in the brightness of a stellar system, resulting from the transit of a planet across the disk of its central star. Such transits would cause characteristic changes in the central star's brightness and, to a lesser extend, color. The depth of a transit is proportional to the surface area of the planet, and the duration of a transit is indicative of the planet's velocity. If the central star's mass is known, the distance and period of the planet can then be derived. Once repeated transits of the same planet are observed, the period can be obtained with great precision. The transit method was first proposed by Struve (1952); later developments are described in Rosenblatt (1971), Borucki & Summers (1984), Deeg (1997). Previous observational tests have been prevented by the required photometric precision (which is about 1 part in 10⁵ in the case of an Earth-sized planet transiting a sun-like star), and by the generally low probability that a planetary plane is aligned correctly to produce transits. This probability of orbital alignment is about 1% for planetary systems similar to our solar system. An observationally appealing application is available with close binary systems, where the probability is high that the planetary orbital plane is coplanar with the binary orbital plane, and thus in the line sight. This makes the observational detection of planetary transits feasible in systems with an inclination very close to (Schneider & Chevreton 1990 , Schneider & Doyle 1995). Furthermore, repeated planetary transits across the binary's components will result in unique sequences of transit-lightcurves, whose exact shape depends on the phase of the binary system at the time of the planetary transit (Fig. 1). Jenkins et al. (1996) demonstrated, that these unique sequences can be used for the detection of planetary signatures with amplitudes below the noise of the observed lightcurves, if signal detection techniques based on cross correlations with model lightcurves are applied.

Fig. 1. Model lightcurves from planetary transits across CM Dra. The upper graph gives the brightness of the CM Dra system, normalized to an off-transit magnitude of zero. A model planet with 2 RE causes transits with a maximum brightness loss of 3.8 mmag (0.35 %). Mutual binary eclipses are removed here. The lower graphs show the elongation (in Solar radii) of the two CM Dra components (CM Dra A: black band, B: white band) and the model-planet (thin line) from the common barycenter. These graphs are of the same style as the well-known diagrams of the positions of Jupiter's moons, and the thickness of the bands is scaled to the sizes of CM Dra's components. From the left to the right, the two leftmost panels show transits caused by a planet with an orbital period of 9 days. The leftmost panel shows the normal case, with two short transits separated by several hours (0.1 CM Dra phase unit corresponds to 3 hours). The second panel shows the rarer case, where the planet transits when CM Dra is close to a mutual eclipse (here shown a secondary one at phase 0.5), and long transits occur. The complicated shape of this transit results from the planet covering zones of different surface brightness on CM Dra (considering limb-darkening and the planet's ingress/egress). The two rightmost panels show transits from a planet with a 36 day period. Such a planet has a transversal velocity that is slower than CM Dra's components, and multiple transits with complicated shapes are more likely to occur.

The near ideal characteristics of the eclipsing binary system CM Dra for an observational test on the presence of planets has been suggested by Schneider and Doyle (1995). The CM Dra system is the eclipsing binary system with the lowest mass known, with components of spectral class dM4.5/dM4.5 (see Lacy 1977, for all system elements). The total surface area of the systems' components is about 12% of the sun's, and the transits of a planet with 3.2 R_E, corresponding to 2.5% of the volume of Jupiter, would cause a brightness drop of about 0.01 mag, which is within easy reach of current differential photometric techniques. The low temperature of CM Dra also implies that planets in the thermal regime of solar system terrestrial planets would circle the central binary with orbital periods on the order of weeks. This allows for a high detection probability of planetary transits by observational campaigns with coverages lasting more than one planetary period. Planets with orbital periods of 10 - 30 days around CM Dra are especially interesting, since they would would lie within the habitable zone, which is the region around a star where planetary surface temperatures can support liquid water, and therefore the development of organic life (see Kasting et al. 1993; Doyle 1996). CM Dra is relatively close (17.6 pc) and has a near edge-on inclination of . With this inclination, coplanar planets within a distance of CM Dra of 0.35 AU will cause a transit event. This maximum distance corresponds to a circular orbit with a period of about 125 days. There is also a low probability of observing orbits from planets inclined out of CM Dra's binary orbital plane, if the ascending or descending nodes of the planetary orbits are precessing across the line of sight (Schneider 1994). The observations of CM Dra presented in this paper are therefore the first attempt to obtain observational evidence of the existence of sub-Jupiter sized planets around main-sequence binary stars, and to evaluate the probability that such detections are possible. For these observations we used differential CCD-photometry and employed 1m class telescopes. To obtain sufficient observational coverage, the 'TEP' (Transits of Extrasolar Planets) network was formed with the participation of several observatories in 1994. Preliminary accounts of TEP network observations are given by Doyle et al. (1996), Martín et al. (1997) and Deeg et al. (1997). A list of the observatories that have been participating is given in Table 1.

Table 1. TEP Network Telescopes and their Location

© European Southern Observatory (ESO) 1998

Online publication: September 14, 1998
helpdesk.link@springer.de