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

2. Predicted and observed line profiles

2.1. Observations

Longslit spectrographic observations were carried out with the Manchester Echelle Spectrometer on the 2.5 m Isaac Newton Telescope between 1991 and 1993. Full details of the observational setup are reported in Massey & Meaburn (1995). In this paper, we present the extracted spectra of five emission line knots (proplyds), which are listed in Table 1, together with the observational parameters for each object.

Table 1. Parameters of the longslit spectroscopic observations.

In each case, an attempt was made to isolate the emission due to the proplyd itself from that due to the nebular background. This was more successful in some cases than in others. For LV 5 (OW 158-323), excellent seeing resulted in relatively straightforward background subtraction and we are confident that in this case all the features in the spectrum outside of the narrow range - belong to the emission knot. The raw position-velocity array for this object is shown in Fig. 1, from which the quality of the data can be appreciated. Similarly, for the other objects, portions of the spectra in the heliocentric velocity range - are very uncertain.

Fig. 1a and b. Positive a and negative b greyscale representations and c logarithmic contours of the same position/velocity array of [3 ] 5007Å profiles through LV 5 (arrowed). The continuous horizontal band is the spectrum of a field star. The bright emission from the background gas (the irregular vertical band) can be seen in a and c to contaminate the spectrum of LV 5 in the range of to . Note, however, that the velocity spike attributable to LV 5 is continuous through this range. The deep presentation in b displays the fainter emission associated with LV 5 at large negative radial velocities with respect to the systemic radial velocity. The log10 contours in c are evenly spaced between 2.4 and 4.5.

The background subtraction was carried out as follows. For each emission knot, a spatial segment along the slit was isolated that completely covered the knot emission. This knot segment was divided into 3 sub-segments. Two background segments were then extracted from the spectra to each side of the knot segment and spatially contiguous with it. For each wavelength, the background from these two segments was linearly interpolated across the knot segment and subtracted from each sub-segment separately. The background-subtracted sub-segments were then added together to give the background-subtracted knot spectra shown in Figs. 2 and  3. Although this subtraction technique can be sensitive to non-linearity in the CCD detector, we find that the peak count rate over 4 binned pixels is in the bright background profiles near the systemic radial velocity. This is significantly less than the saturation value of , so we are confident that the detector response is highly linear.

Fig. 2. [3 ] 5007Å line profile of LV 5 (OW 158-323). a  Observed profile. b  Comparison with model profile (heavy solid line). The different components of the model are indicated by the light solid line (photoevaporated wind), dotted line (Mach disk) and dashed line (swept-back tail). Inset box shows magnification of the blue wing. The region of the spectrum that may be uncertain due to background subtraction is indicated by the light gray rectangle.

Fig. 3a and b. Comparison of observed [3 ] 5007Å line profiles of other proplyds (gray histogram) with the models (heavy solid line). Model components as in Fig. 2b. The region of the spectrum that may be uncertain due to background subtraction is indicated by the light gray rectangle. a  LV 1 (OW 168-326E). b  LV 2 (OW 167-317). c  LV 3 (OW 163-317). d  OW 171-334 .

2.2. Model Profiles

Simulated spectra were generated from the proplyd models of Paper I, using the technique outlined in Henney (1996). The models consist of a circumstellar disk around a low mass star, from the illuminated face of which flows a transonic wind due to photoionization by the O7 star  Ori C. This disk wind in turn interacts with the stellar wind from  Ori C, forming a blunt Mach disk where the ram pressure of the two winds balance, together with a long tail where the disk wind material is swept back by the stronger wind from the O star.

The predicted line profiles are affected by five parameters. Two of these are the parameters that describe the proplyd model: , which is the ratio of the momentum flux of newly ionized material entering the base of the disk wind to the momentum flux in the wind of  Ori C, and , which is the angle between the disk normal and the direction to  Ori C. In addition, there is the orientation of the proplyd with respect to the observer (characterized by an inclination angle and an azimuthal angle ) and the displacement of the slit center from the position of the center of the proplyd.

For each object, the slit parameters from Table 1 are used and the model parameters, orientation and slit position are varied in an attempt to find the best fit to the line profile. A sound speed of is used (appropriate for ionized gas at K) and a systemic heliocentric velocity of is assumed for each proplyd (corresponding to the velocity of the Orion molecular cloud). The fits would not be substantially improved by allowing peculiar motions of up to . Since the absolute positions on the sky of the spectrograph slits are rather uncertain, we have allowed displacements of the slit by up to one half of the seeing width from the fiducial positions indicated in Fig. 1 of Massey & Meaburn (1995).

The resultant parameters of the best-fit models are shown in Table 2 and the fits themselves are shown superimposed on the observed spectra in Figs. 2 and 3. The model line profiles have three components. The brightest is due to the photoevaporated wind itself and is always close to the systemic velocity. There is also a faint, usually broad, component due to the Mach disk (where the disk wind interacts with  Ori C's wind) and a brighter, more peaked component that is due to the swept-back tail. These three components are shown in the figures by light solid, dotted and dashed lines respectively.

Table 2. Parameters of model fits to the line profiles.

The bulk of the material in the photoevaporated wind moves at only slightly supersonic velocities, so that the Doppler shift of the peak of the wind emission depends only on the angle between the normal to the disk and the line of sight and is very approximately given by . Note that, since this depends on all three of the angular parameters, it would be impossible to find a unique fit to the wind component of the line profile alone, even in the absence of background subtraction problems near the systemic velocity. The width and shape of the wind component also depend on the angular parameters but in a rather complicated way, especially for instances where the base of the wind is partially occulted by the opaque circumstellar disk.

However, if the tail component of the line profile can also be identified, then a reasonably unique fit can be determined. This is because the ratio of the intensities of the wind component to the tail component depends almost exclusively on the parameter , whereas the Doppler shift of the peak in the tail component depends mainly on , but only very weakly on and not at all on the other angular parameters. The relative intensity of the Mach disk component is rather insensitive to all the model parameters except that, for slits that are more or less perpendicular to the proplyd axis, it does depend on the position of the slit. Hence, we can fit the models to the observations in a systematic way.

© European Southern Observatory (ESO) 1997

Online publication: May 26, 1998

helpdesk.link@springer.de