Wind Models

As discussed in previous chapters, wind models seem to have a few difficulties in explaining the Balmer line profiles. While IPC line profiles cannot, surely, result from these models, can Pa Beta and/or Br Gamma Type I profiles with the observed characteristics be produced by them?

The top panel of Figure 6.3 shows an example of the results obtained for a typical wind model, characterized by the parameters displayed in the lower panel of the same figure.

The H Alpha and the H Beta lines display normal P Cygni profiles. This is in good agreement with the results obtained for these lines, in a wind scenario, by other authors (eg. Hartmann et al. 1990). Pa Beta and Br Gamma display the same kind of characteristics as the Balmer lines. The lines have normal P Cygni profiles, with a redshifted line peak. These characteristics are typical of the results obtained for wind models and are not at all observed in the data set discussed in the previous chapter.

Increasing the mass loss rate from about M yr^-1 to about M yr^-1, results in stronger and broader line profiles, as the lines become optically thicker. Eventually, the blueshifted absorption disappears and the line becomes a Type I profile (see panels a) and b) in Figure 6.4). Together with the increase in width there is an increase of the line peak, which however remains redshifted relative to the rest velocity. A similar effect results from increasing the temperature of the wind from 7000 K to 10000 K (see panels c) and d) in Figure 6.4). Increasing the wind temperature still further, to about 15000 K, results in line profiles similar to those produced for 10000 K but with slightly lower peak intensities.

For these typical wind models the task of producing Type I profiles with the line peaking at the observed intensities seems not possible. The Type I profiles resulting from the conditions discussed above always display peak intensities much higher than observed, usually by one order of magnitude. Furthermore, the model line peaks are redshifted whereas the observations show blueshifted line peaks.

The spherically symmetric characteristic of the wind models computed here imply that higher wind densities produce stronger and broader line profiles. While the latter is a desirable feature in order to obtain Type I line profiles one would like to avoid the former. A possible way of doing that is relaxing the assumption of spherical symmetry which might lead to situations for which the line profiles are broad enough but the smaller filling factors do not let the line become too strong. The results from Mitskevitch, Natta & Grinin (1993), discussed in Section 5.4.1, show that type of behaviour, despite not giving completely satisfactory results.

The results presented in this section correspond to models of a region ten stellar radii in size. Calculations were carried out where the size of the region to be modeled was increased to twenty stellar radii. The results are very similar to those obtained for a region of size ten stellar radii. It is therefore sufficient to use the latter for the size of the region modeled. A smaller size for the modeled region was not used since that would alter the line profiles significantly. That would result from the fact that source functions and optical depths for Pa Beta and Br Gamma lines at radii smaller than ten stellar radii are important. Beyond ten stellar radii both source functions and optical depths become suficiently small for not being a determinant factor for the final line profile. Source functions for the H Alpha line, for wind models, show a similar behaviour [Hartmann et al. 1990].

CW Tau is the only star in the sample discussed in this work with a blueshifted absorption, and only at Pa Beta. One would expect a wind model to be able to explain the observed Pa Beta line. Line profiles were computed for a wind model similar to that used by Johns & Basri (1995b) for SU Aur. The characteristics of the wind and the resulting profiles are displayed in Figure 6.5. There is a good agreement between the model result for Pa Beta and the observations, however, the model Br Gamma line displays a normal P Cygni profile while the observations show a IPC profile. It should be kept in mind that the Pa Beta and Br Gamma observations were made one day apart and due to line variability one cannot exclude that the Br Gamma line profile changed for normal P Cygni to IPC. A variability study of these two lines in CW Tau would help solving this problem. In order to obtain a good match between model and observations for the Pa Beta line the turbulent velocity at the base of the wind does not need to be as large as in Johns & Basri (it goes up to 120 km/s in their model), however, the extent of the high turbulence region must be somewhat larger than the one used by those authors.

Apart from this one star, and only at Pa Beta, the wind models explored in this work, are (not surprisingly) unable to reproduce the observations, even of the Type I profiles which are observed to be skewed to the blue.