A major discovery of the International Ultraviolet Explorer satellite, which operated from 1978 until it was "switched off" in 1996, was that the accretion disks of cataclysmic variables are the seat of substantial mass outflows, or winds. This phenomenon is observed at times of high mass accretion rate and is characterised by strong, broad ultraviolet spectral lines. These are formed by scattering in the wind of disk continuum photons at wavelengths corresponding to atomic transitions in the wind atoms. The lines are broad because the wind is being accelerated. Thus different parts of the wind are travelling at different velocities with respect to the disk and the wavelength of the continuum light that can be scattered is modified by the Doppler shift. In some systems, we see a P-Cygni-type line profile, the classic signature of mass outflow. In others the lines have broad emission or absorption components, depending on the angle at which we view the accretion disk.
While the existence of these winds is well established, they remain one of the least understood phenomena associated with accretion disks in cataclysmic variables. The physical cause of the outflow has yet to be properly identified, though radiation pressure from the accretion disk undoubtedly plays a role in accelerating the wind to the high velocities seen (several thousand km/s). Likewise there is considerable uncertainty over the wind geometry and kinematics. However, it is clear that the outflows modify our view of accretion disks in these systems and may influence the binary evolution, so it is important that they are much better characterised than at present. Because winds are also known to be associated with accretion disks in other astrophysical settings (e.g. in young stellar objects of the FU Ori class), this problem assumes a more general significance.
At MSSL we have been working with the group at Imperial College London and others in a programme to study winds in cataclysmic variables using the Hubble Space Telescope. The HST is much more sensitive than IUE was, and we can observe the wind-formed ultraviolet lines with much better spectral resolution, and follow how they change as the binary system rotates. We have specifically concentrated on observing systems where the accretion disk is eclipsed by the companion star in the binary. By relating changes in the line profiles during eclipse to the (known) position of the companion star in its orbit, information can be obtained on the spatial and velocity distribution of the wind.
The results so far have been very interesting. We have been able to show that the wind has a component of rotation as well as outflow, arguing strongly that the outflow is seated in the inner accretion disk, rather than the central white dwarf star. The study also reveals narrow absorption lines superimposed on the broad emission. One possibility being explored is that these absorption lines arise in an extended, relatively static atmosphere that interfaces between the disk and the wind. This could be very important in understanding how the wind material gets ejected from the disk, and thus in understanding how and why the wind forms.
Another unexpected result to come out of the HST study concerns the overall orbital light curves of these CV (see Mason et al. (compressed postscript file). As noted above, the stars being studied have accretion disks which are viewed very nearly edge-on. In each of the two systems observed so far with HST the ultraviolet light exhibits a substantial "dip" about a quarter of an orbital cycle before the time when the accretion disk is eclipsed by the companion star. The spectral signature of this dip suggests that it is caused by the occultation of the hot inner disk by cooler material further out. This clearly demonstrates that the disk has substantial vertical structure in places, and underlines the fact that accretion disks in these interacting binaries are not simple two-dimensional entities. Their three dimensional nature may well be an important determining factor leading to the formation of the wind.
Other Galactic Astronomy pages:
MSSL home pages:
This page written by Keith Mason (kom@mssl.ucl.ac.uk).
Last modified 8th Jan 1998