The UV scattering halo of the central source associated with η Carinae

We have made an extensive study of the UV spectrum of eta Carinae and find that we do not directly observe the star and its wind in the UV. Because of dust along our line of sight, the UV light that we observe arises from bound-bound scattering at large impact parameters. We obtain a reasonable fit to the UV spectrum by using only the flux that originates outside 0''033. This explains why we can still observe the primary star in the UV despite the large optical extinction: it is due to the presence of an intrinsic coronagraph in the eta Car system and to the extension of the UV-emitting region. It is not due to peculiar dust properties alone. We have computed the spectrum of the purported companion star and show that it could only be directly detected in the UV spectrum, preferentially in the FUSE spectral region (912 - 1175 A). However, we find no direct evidence for a companion star, with the properties indicated by X-ray studies and studies of the Weigelt blobs, in UV spectra. This might be due to reprocessing of the companion's light by the dense stellar wind of the primary. Broad Fe II and [Fe II] emission lines, which form in the stellar wind, are detected in spectra taken in the southeastern lobe, 0''2 from the central star. The wind spectrum shows some similarities to the spectra of the B and D Weigelt blobs but also shows some marked differences in that lines pumped by Ly alpha are not seen. The detection of the broad lines lends support to our interpretation of the UV spectrum and to our model for eta Car.