Standard theories of planet formation require a disk of gas and dust to form planets by coagulation of small particles due to gas drag and perturbations of the particle orbits (Weidenschilling 1980). The orbits of these particles are almost circular due to viscous dissipation; this fact suggests that newly formed planets should have also small eccentricities.
Very recently, researchers at McDonald and Lick Observatories have independently discovered a planet orbiting the star 16 Cygni B (Cochran at al. 1997; Butler & Marcy 1997; Kamper 1997), which is part of a triple star system about 21.4 pc from Earth. This multiple star consists of an inner binary with two G-dwarf stars separated by about 835 AU. A distant M-dwarf is orbiting the binary in a hierarchical arrangement. The companion to 16 Cygni B has a mass that could be as little as 1.6 times that of Jupiter ( = 9.6 ), so it is probably a true planet rather than a brown dwarf. It circles the star every 2.2 years in a highly eccentric orbit, e =0.67. Earlier discoveries of eccentric giant planets (hereafter EGPs) are the planet around 70 Vir with e =0.40 (Marcy & Butler 1996), and one orbiting HD 114762 with e =0.35 (Mazeh et al. 1996). Mechanisms for generating EGPs have been considered recently (Artymowicz 1993, 1997; Rasio & Ford 1996; Holman et al. 1997; Katz 1997; Mazeh et al. 1997a; Lin & Ida 1997; Mazeh et al. 1997b).
In this Letter, we consider the possibility of EGP formation around stars in open clusters together with probable parameters of observable systems. In 2 and 3, we present the results of detailed numerical calculations of the dynamical evolution of planetary systems in open star clusters. For simplicity, the planetary systems studied in this work consist of only one giant planet and its host star. The calculations were done with the N -body code NBODY5 (Aarseth 1985, 1994) appropriately modified for our present purpose. This code includes the effect of the Galactic tidal field (Aarseth 1985, 1994) and mass loss due to stellar evolution (Eggleton et al. 1989). The calculations also consider a realistic mass spectrum (Scalo 1986) in the range [0.08, 15.0] . Spherical symmetry and constant density are assumed for generating initial positions, with the rate of the total kinetic and potential energy fixed to 0.25. The initial velocities are random and isotropic. The initial mean radii are in the range [1, 1.5] pc. Multiple interactions are computed by using the Bulirsch-Stoer integrator; numerical errors (relative energy errors) are smaller than 10-4 per unit time. The evolution of the models is followed at least for 300 Myr. In 4 we discuss the feasible observational properties of stellar systems containing a giant planet in open clusters.
© European Southern Observatory (ESO) 1997
Online publication: April 8, 1998