2001ApJ...546.1068W

We have searched for a methane signature in the infrared spectrum of τ Bootis, produced by the planetary companion. The observations comprise 598 low-noise (signal-to-noise ratio ∼100), high-resolution (λ/δλ=4x10⁴) spectra near 3044 cm^–1, which we analyze by cross-correlating with a modeled planetary spectrum based on the work of Burrows & Sharp and Sudarsky, Burrows, & Pinto. The 3 σ random noise level of our analysis is ∼5x10^–5 stellar continuum flux units, and the confusion noise limit–measuring the resemblance of a cross-correlation feature to the spectrum of methane–is ∼2.5x10^–4. We find a significant cross-correlation amplitude of ∼3.3x10^–4 continuum units at a velocity near that of the star. This is likely due to methane from a low-mass companion in a long-period orbit. Fischer, Butler, & Marcy report a long-term velocity drift indicative of such a companion. But the system is known to be a visual binary with an eccentric orbit and is rapidly approaching periastron. Whether the visual companion can account for our observations and the Fischer et al. velocity drift depends on knowing the orbit more precisely. The stability of planetary orbits in this system also depends crucially on the properties of the binary orbit. A second cross-correlation feature, weaker and much more diffuse, has intensity amplitude ∼2x10^–4 continuum units and occurs at a velocity amplitude of 71(±10) km.s^–1, in agreement with the orbit claimed for the planet by Cameron et al. Like the first feature, it has passed several tests designed to reject systematic errors. We discuss the possibility that this second feature is due to the planet.