The Distance and Interstellar Sight Line to Gx 339 – 4

The distance to the black hole binary GX 339–4 remains a topic of debate. We examine high-resolution optical spectra of the Na D lines resolving the velocity structure along the line of sight. We find this to be complex, with at least nine components, mostly blue-shifted, spanning a velocity range of nearly 200 km s −1. The presence of components with a large blue-shift rules out a nearby location and requires that the binary be located at or beyond the tangent point, implying a lower limit to the distance of ∼ 6 kpc. The presence of a significant red-shifted component at +30 km s −1 is even more intriguing as GX 339–4 also has a slightly positive systemic velocity, suggesting that the source, and this cloud, could be on the far side of the Galaxy, where the radial velocities due to Galactic rotation become positive again. If this is the case, we require a distance of ∼ 15 kpc. This is less secure than the 6 kpc lower limit however. We discuss the implications of these possible distances for the outburst and quiescent luminosities, and the nature of the companion star, and argue that a large distance is consistent with these characteristics. In particular, it would explain the non-detection of the companion star during the faintest states. The distance of X-ray binaries is challenging to determine precisely. They are usually too distant for current parallax measurements, and there are no really reliable standard candles, other than type-I X-ray bursts showing radius expansion, and these only occur in a subset of neutron star systems (see Kuulkers et al. 2003 and references therein). For black holes, such as GX 339– 4, the problem is much less tractable. Fortunately the majority of such systems undergo periods of deep quiescence when it is possible to detect the companion star, and determine the binary parameters. Hence both the radius and temperature (i.e. surface brightness) can be estimated and the distance derived from the apparent brightness (see e.g. Gelino, Harrison, & Orosz 2001). For GX 339–4 this is not possible, as even in a very low luminosity state, no spectral features of the companion star were seen in VLT/FORS observations (Shahbaz, Fender, & Charles 2001). This is frustrating for distance and parameter estimation, and very puzzling in its own right. One explanation for the non-detection of the companion is that the distance has been significantly underestimated. …