1996PASP..108..580I

The long-period (∼600 yr) binary system 36 Ophiuchi AB consists of two chrompherically active K dwarfs. If we constrain the period using the mass-luminosity relation and the observed parallax, the remaining orbit parameters can be estimated from the 170 years of visual-binary observations. In this paper, we further constrain the orbit by precise measurements of the differential radial velocities (with arbitrary zero points) of both 36 Oph A and B and the difference in radial velocity between 36 Oph B and A. Our best orbit gives a good fit to the visual-binary observations, the difference in velocity, and the mean radial acceleration of A; but the observed acceleration of B is a factor of 164 larger than the value predicted by the orbit. This factor is so large that no reasonable variation in the adopted sum of masses, mass ratio, parallax, or orbit parameters will remove the B acceleration discrepancy. Ths maximum companion mass allowed by the residuals from the visual-binary orbit is of order 8 Jupiter masses for assumed periods between 30 and 100 years so the 36 Oph B acceleration discrepancy would ordinarily make it a candidate for a substellar companion. However, the very high eccentricity (∼0.9) of the binary-star orbit means its closest approach is of order 6 a.u. making it unlikely that any substellar companions would form or survive with the semi-major axis exceeding ∼1.5 a.u. or period exceeding ∼2 years. Thus, 36 Oph B is an important counter-example which serves as a warning that for chromospherically active stars, at least, it is possible to have apparent radial accelerations in the absence of substellar companions.