The interaction of X-rays with interstellar dust grains leads not only to mere absorption but also to small angle scattering. Therefore, X-ray sources behind sufficiently large dust columns are surrounded by haloes of faint and diffuse radiation. This effect and its use as a powerful diagnostic tool was first described theoretically by Overbeck (1965), the first detection of an X-ray halo almost 20 years later was reported by Rolf (1983). Mauche & Gorenstein (1986) showed that the shape and strength of the haloes as derived from their measurements with the Einstein Observatory were consistent with common grain models, e.g., the one established by Mathis et al. (1977). The up to now most complete investigation on scattering haloes was done by Predehl & Schmitt (1994) based on observations with ROSATfrom both the all-sky survey as well as a number of pointed observations. They found a strong correlation between the strength of haloes (or the optical depth in scattering, respectively), the X-ray absorption and the visual extinction for X-ray sources with known optical counterparts. These relations provide a precise differentiation between interstellar and local matter: both extinction and X-ray absorption are produced by the total column density between source and observer, the scattering, however, only by dust on large scales, i.e., the ISM.
Scattered radiation has to travel along a slightly longer path than the direct, unscattered light. Any intensity variation of the source therefore occurs somewhat delayed in the halo. Trümper and Schönfelder (1973) proposed a method for using this behaviour to measure the distance to variable X-ray sources. Although it was presented almost 30 years ago, this method could never be applied so far. The main reason is that the amplitude and the timescale of the intensity variation together with the distance of the source have to fit the angular resolution of the observing instrument. Using instruments with moderate angular resolution, the presence of scattering haloes has only a damping effect on the measured intensity variation. These observational aspects are discussed by, e.g., Molnar & Mauche (1986) and Kitamoto et al. (1989).
The attractiveness of the `halo-method' is that it yields a geometrical rather than an physical distance from other methods like the 21 cm absorption, visual extinction, or X-ray absorption. A good example for a wrong distance estimate is for Sco X-1 which was assumed to be at a distance between 25 and 1000 pc until Bradshaw et al. (1997) obtained a distance of 2.8 kpc from VLBI measurements thus placing it far outside the galactic dust and gas layer.
In this paper we present, to our knowledge for the first time, a successful application of this distance determination method. Using data taken with the Chandra X-ray Observatory, we could measure the approximate distance to Cyg X-3. This object is an eclipsing X-ray binary with an orbital period of 4.8 hours (Brinkman et al. 1972).
© European Southern Observatory (ESO) 2000
Online publication: May 3, 2000