6. A phase shift of the C III/IV 4650 line complex
As noted in the previous section, the periastron dates of the emission line solutions show a clear bimodal distribution. The lines C IV , 4786, 7730 yield whereas C III/IV and C III give . The difference of three days is significant and as demonstrated in Fig. 6, the phase shift is obvious even to inspection by eye when the two data sets are plotted on top of each other.
We do not understand this phenomenon. For the C III line we find an orbit solution that is in perfect agreement with the other three lines and in particular, with a periastron date of 2,450,121.9, if we modify the fit procedure by including the nearby He II and use a fixed line width. This result indicates that for the C III the phase shift may not be real. The periastron date of the He II belongs to the group of early periastron dates as well. However, an inspection of its radial velocity curve reveals that for this line the early periastron date is only due to the periastron angle of the solution which differs considerably from the other solutions. Hence, the He II does not belong to the lines with a phase shift.
On the other hand, the shift of the strong C III /IV appears to be real. We have tried several different approaches with multiple-Gauss fits and all combinations of fixed and variable line widths. We never obtain a significant deviation from the elements given in Table 1 and in particular, the data always show the phase shift when plotted against other transitions. Although the radial velocity curve of the He II is not very well defined, it appears that also He II shows a significant phase shift of about 4 days.
As a hypothesis we propose that the line profiles are disturbed by additional emission from the bow shock region. Hydrodynamical calculations of the collision of the WR wind with that of the O star predict a shock zone at the wind-wind interface. Because in the case of Velorum the momentum of the WR wind is much larger than that of the O star, the shock is wrapped around the O star with the flow directed fairly straight away from the WR, and with velocities of the order of the WR wind speed (see Fig. 2 of Walder 1995; Stevens et al. 1996). The proposed additional emission cannot be observed as an extra feature but it widens the line profiles in such a way that we measure an overall shift to the blue, when the stars are in conjunction and the WR is behind the O star, and to the red during the other conjunction when the WR star is in front. The line centers are not shifted during quadrature when we observe the highest radial velocities. Thus, the velocity amplitude is not altered.
According to our hypothesis we expect a systematic variation of the line width. Indeed, we observe that the line widths vary systematically with orbital phase. However, instead of two maxima of the line widths we only observe one. At least, the observations agree partially with our model, in that the maximum occurs during one of the conjunctions. We conclude that our model is too simple and that reality is more complex. Clearly, a more elaborate investigation of this potentially interesting phenomenon is needed. Nevertheless, since in UV resonance lines there is clear evidence for colliding winds (St-Louis et al. 1993), it is well possible that the phase shift of at least the C III/IV is real and due to additional emission not associated with the WR star. Additional evidence for emission possibly arising in a wind-wind collision zone was found by St-Louis (1996). She reported phase-related variability in the profile of C III . The same type of profile variability is modeled successfully for WR 79 by Lührs (1997) with a simple model of the shock zone.
Similar phenomena are also observed in other WR binaries. Bertrand (1995) observed a line profile skewness in the case of WR 133. Moffat & Seggewiss (1977) found a phase shift close to a quarter period for the line C III in Mus. These observations are thought to be associated with effects of colliding winds.
© European Southern Observatory (ESO) 1997
Online publication: March 24, 1998