A Search for Variability in the Spectral Line Shapes of τ Bootis: Does This Star Really Have a Planet?

Abstract

An analysis is made of the spectral line shapes of τ Bootis using high-resolution (0.026 A) and high signal-to-noise (S/N ≈ 400) data in an effort to confirm the planet hypothesis for this star. Changes in the line shape are quantified using spectral line bisectors and line residuals. We detect no variations in either of these quantities above the level of the noise in the data. One spectral line, Fe I λ6213, does show a hint of sinusoidal variations in the bisector velocity span when phased to the radial velocity period of 3.3 days, but this is not seen in the bisectors for two other lines nor in the line residuals. Comparisons of the data to the bisector and residual variations expected for nonradial pulsations indicate that we can exclude those sectoral nonradial modes having m > 2 and all sectoral modes with k > 1, where k is the ratio of the horizontal to vertical velocities for the pulsations. The lack of line shape variability and the 469 m s-1 radial velocity amplitude is still consistent with nonradial sectoral modes m = 1 and possibly with m = 2, but with k ≈ 1, which is at least 3 orders of magnitude less than the predicted value given the 3.3 day period of τ Bootis. Such low values of k can probably be excluded given the lack of photometric variations for this star. Although the measurements presented here do not prove beyond a doubt that τ Boo has a planetary companion, they do add significantly to the increasing body of evidence in favor of this hypothesis.