Spatially Resolved STIS Spectra of WR + OB Binaries with Colliding Winds

We present spatially resolved spectra of the visual WR+OB massive binariesWR86, WR146, and WR147, obtained with the Space Telescope Imaging Spectrographon board the Hubble Space Telescope. The systems are classified as follows:WR86 = WC7 + B0 III, WR146 = WC6 + O8 I-IIf, WR147 = WN8 + O5-7 I-II(f). BothWR146 and WR147 are known to have strong non-thermal radio emission arising ina wind-wind collision shock zone between the WR and OB components. We find thatthe spectra of their O companions show Halpha profiles in emission, indicativeof large mass-loss rates, and consistent with the colliding-wind model. Ourspectra indicate that the B component in WR86 has a low mass-loss rate, whichpossibly explains the fact that WR86, despite being a long period WR+OB binary,was not found to be a strong non-thermal radio emitter. Because of the smallmass-loss rate of the B star component in WR86, the wind collision region mustbe closer to the B star and smaller in effective area, hence generating smalleramounts of non-thermal radio emission. Absolute magnitudes for all the starsare estimated based on the spectral types of the components (based on thetables by Schmidt-Kaler for OB stars, and van der Hucht for WR stars), andcompared with actual, observed magnitude differences. While the derivedluminosities for the WC7 and B0 III stars in WR86 are consistent with theobserved magnitude difference, we find a discrepancy of at least 1.5 magnitudesbetween the observed luminosities of the components in each of WR146 and WR147and the absolute magnitudes expected from their spectral types. In both cases,it looks as though either the WR components are about 2 magnitudes too brightfor their spectral types, or that the O components are about 2 magnitudes toofaint. We discuss possible explanations for this apparent discrepancy.Comment: Accepted for publication in the Astronomical Journa