Astron. Astrophys. 324, 656-660 (1997)

3. Discussion

We first focus attention on LV 5 (OW 158-323) since the observed spectrum of this object is of much higher quality than the others. The line profile, illustrated in Fig. 2a, shows three strong components close to the systemic velocity, at , 22 and , together with a weaker blue-shifted component at , which possesses a wing extending out to . Because of uncertainties in subtraction of the stong and highly variable background (see Fig. 1), it is possible that the peak is spurious and hence we ignore it in our model fits. It is obviously not entirely satisfactory to exclude the 10 to region from our fit, but in order to reliably isolate this portion of the proplyd spectrum from the background nebular emission, much higher spatial resolution is necessary than is available from ground-based observations.

Our model reproduces well the redmost peak, together with the blue peak and wing. However, the peak, although present in our model (due to the Mach disk) is not strong enough. We note that in order to obtain a peaked rather than broad Mach disk component, a rather special combination of parameters is required. In particular, it is necessary that and that . This leads to a value for the inclination of the disk normal with respect to the direction of  Ori C of for LV 5, which is at the upper limit of that allowed by our simple dynamical model of the proplyds (see Paper I).

Apart from shortcomings in our dynamical model, one factor that could effect the proplyd line profiles and which is not considered in this paper is dust scattering in the photoevaporated wind. The dust optical depth to the base of the wind in our models is of order 0.2-1.0 at the wavelength of the [3 ] 5007Å line and this could have a substantial effect on the wind component of our model profile (c.f. Henney 1994; Henney & Axon 1995). Since the wind is both divergent and accelerating, any scattered component will be red-shifted with respect to the direct emission from the wind. Preliminary results of detailed modelling of this process (to be presented in a subsequent paper) show that the principal effect on the line profiles is to broaden the component due to the photoevaporated wind by 10- .

From the position-velocity array of LV 5 (Fig. 1) it is apparent that the blue-shifted emission peak is more extended spatially than either the far blue wing or the red-shifted peak. We find that our model spectrum also shows this feature, which arises because the blue peak is due to the swept-back tail of the proplyd, which is its most spatially extended part.

Turning now to the other proplyds, illustrated in Figs. 3a-d, the observations are somewhat poorer (mainly because of problems with background subtraction) and all have unreliable portions somewhere in the range - (indicated by light gray shading on the figures). The quality of the model fits is variable, ranging from very good in the case of OW 163-317 to very poor in the case of LV 1 (OW 168-326E). In other cases, such as LV 2 (OW 167-317), the fit to the red portion of the spectrum is good but the model predicts an extended blue wing, which is not observed.

The deduced values of the parameter for our fitted models (Table 2) are all reasonable, lying in the range expected for proplyds at a distance from  Ori C of 0.01-0.05 parsecs. They are, however, a little higher than the values (10-30) found by fitting the morphology of the proplyds in Paper I. In addition, the angles , and of the fits listed in Table 2 are, in general, different from the fits to the same objects found in Paper I. However, we feel that the line profiles are a much more effective discriminant of the model parameters than the morphologies, and indeed the fits found here still show reasonable agreement with the observed morphologies.

In summary, our proplyd model is successful in accounting for most of the features in the [3 ] 5007Å line profiles of the five emission line knots presented here. It is hoped that the remaining discrepancies between theory and observation will be removed by improvements in the model (for instance, a fuller treatment of the effects of gravity on the disk wind and a more detailed calculation of the flow in the tail) and by a more realistic treatment of the radiative transfer, including dust absorption and scattering.

© European Southern Observatory (ESO) 1997

Online publication: May 26, 1998

helpdesk.link@springer.de