In this section, we use the near-infrared imaging data to discuss the morphological properties of the infrared nebula. First, conclusions are drawn from the visual appearance by interpreting the different features as astrophysical phenomena. This indicates an outflow and the principal, already well-known orientation of the disk. Going further, we use colour indices and the wavelength-dependent position of the object to confirm the details of the geometry. The second part of the section deals with the spectra and the derivation of the dust composition therefrom.
3.1. Disk and outflow
In Figs. 1 through 3, the NIR images of the IR nebula can be seen. They all show some common features: From the bright knot at the reference position, the nebula stretches to the east and, much weaker, to the west. The general features of the Cha IRN are already well known with bright elongated structures to the northeast (feature A), the northwest (B) and the southwest (D). These bright lobes are generally believed to be the brightened rims of parabolic cavities, cleared by outflow activity (e.g. Gledhill et al. 1996). These cavities point to the east and to the west. The southeastern rim is then missing, probably due to foreground extinction. The parabolic cavities exhibit a position angle of , which would in turn point towards a molecular disk at P.A. . However, the polarization studies by Ageorges et al. (1996) and Gledhill et al. (1996) show an aligned polarization vector pattern at P.A. towards the central bright knot. This indicates a tilted inner part of the disk, an effect which according to Gledhill et al. (1996) might be caused by tidal forces from a young binary system inside.
In Fig. 1 through 3, two more features can be seen: A tail-like structure pointing from the bright knot to the west and turning north (Feature C) and another tail, pointing east and turning south (E). Features (C) and (E) are best visible at the shorter NIR wavelengths, which is obvious from the J-band image shown in Fig. 3. Fig. 4 shows the contours of the high-resolution speckle image, superimposed on the grey-scale image from Fig. 1. These contours reveal some inner structure of the bright knot. Apart from the bright peak itself, the highest contours trace a structure which seems to be the inner origin of feature E. Also, the upper rim of the contour pattern is obviously the inner part of the bright rim marked as feature (A). The beginning of the missing lower rim can marginally be seen at the southern edge of the pattern. This indicates that foreground extinction already starts shortly south of the main peak. West of the bright knot, almost no contour lines are visible. Only a very weak structure at position (, ) is visible. This feature shows two elongations which could be interpreted as the origins of features (B) and (C). If this is indeed the case, then features (C) and (E) meet at the approximate position (, ) with P.A. . Altogether (C) and (E) form an S-shaped feature and could be interpreted as the signature of a bipolar outflow. The curvature indicates the presence of a second stellar component in the system. Such a binary can provide the bending of the two lobes, e.g. by causing a tilt in the inner disk through tidal forces. The disk would then start to precess and force the outflow to follow the motion of its polar axis. This scenario is also consistent with the presence of a tilted inner disk mentioned above. A second explanation can be thought of by assuming the companion star itself to be the source of the outflow. In this case the curvature could be caused by the orbital motion (provided that the orbital velocity is of the same order of magnitude as the outflow velocity) like in the model for T Tau by van Langevelde et al. (1994). In both cases, the main polar axis is defined by the large-scale molecular disk at P.A. , which is in agreement with the bipolar cavities and probably with the orbital plane of the binary system. As a counterpoint to this line of reasoning it has to be mentioned that no apparent molecular outflow from Cha IRN has been found up to now, despite a CO search with the SEST by Ageorges (1997, priv. comm.). From Figs. 1, 2, and 3, we computed the colour indices of the Cha IRN. The J, H, and K' images have been aligned using the star at (, ) (see the northern edge of Fig. 1) as a reference point. The reddest parts of the images are the bright knot itself and the two rims of the western outflow cavity. The knot has indices of J - mag and H - mag, whereas the western rims show indices around J - mag and H - mag. In fact, the western rims, denoted as features (B) and (D), are hardly visible in Figs. 2 and 3. East of the bright knot, colour indices range up to only J - mag and H - mag. This leads to the conclusion that the disk is inclined such that the western lobes are viewed through the outer parts of the disk plane.
Using the distance of 190 pc towards the object and the fact that all the previous investigations indicate a disk inclination of , we can now derive the linear size of the disk: Since the western lobes appear strongly reddened out to a distance of 680 AU (and thus appear "covered" by the disk), we compute its radius to be at least 2000 AU. This value is larger than the 1000 AU obtained by Ageorges et al. (1996) and the 1400 AU assumed by Cohen & Schwartz (1984) and Gledhill et al. (1996). However, the latter two sizes were based on a distance estimate of 140 pc.
Another fact can be utilized to confirm the spatial orientation of the disk: If we calculate the centroid position of the bright knot, we see that it is different in the three bands. The position in H is shifted by (, ) with respect to J, in K by ( =-0.5", ). This positional shift between wavelengths is a classical indicator for reflection at cavity walls emerging from an inclined disk (see, e.g., Stecklum et al. (1997) for GGD 27 or Close et al. (1997) for HL Tau). The small shifts in declination point to a very small bending angle out of the north south plane. From these measurements, the position angle of the disk plane would be between and . This indicates that the large-scale disk (not the tilted, inner one) is responsible for the effect of the position shift between wavelengths. The direction of the shifts also confirms that we are facing the eastern surface of the disk, the inclination therefore obscuring and thus reddening the western cavity walls. However, the errors in these measurements are of the order , so this can only be seen as a principal indicator. The fact that we do not see the small-scale inner disk with this method is consistent with Gledhill et al. (1996), who assume this part to be unresolvable by their observations. This means that the inner disk should not show any direct effect in our images.
3.2. IR spectra
We will now discuss the infrared spectra taken with IRAS, ISO, and IRSPEC. The combined spectra of the Cha IRN are shown in Fig. 5. Several features can be identified in them, the most obvious being the H2 O ice band near 3.1 µm. The optical depth of 3.3 at the centre of this feature makes it one of the most prominent ones known in protostellar spectra, similar to the one in AFGL 2136 (Kastner & Weintraub 1996) and W3-IRS5 (Smith et al. 1989). The long wavelength wing of the feature is slightly different in the IRSPEC and ISOPHOT spectra. The IRSPEC data indicate a more pronounced wing.
Characteristics of the features are summarized in Table 2. The line centre, optical depth and full width half maxima have been determined by fitting a Gaussian profile to each absorption line and using the parameters from the fit. For continuum subtraction, quadratic baselines were fitted to the immediate neighbourhood of the features. In case of the 2.9 µm NH3 feature, the fit of the 3.1 µm H2 O was subtracted as background before fitting the line. All features were identified in the ISO data, the results of IRSPEC being too noisy to deliver useful results. The IRSPEC data do, however, confirm the general course of the spectrum. The IRAS data also show no features, apart from a slightly noise contaminated confirmation of the 8.9 µm feature.
In Table 2 we also give the derived column densities N for each molecular species. The latter are calculated using the relation , where A is the integrated absorbance, the measured line width (FWHM), and the optical depth of the line. The ISO data show a narrow secondary dip within the H2 O absorption feature. Decomposing the two features and measuring their parameters yields the values given in Table 2.
These data indicate that ammonia is approximately 2.5 times more abundant than water, a fact which seems rather unlikely. Due to the complicated line subtraction process and the fact that the feature shows up in essentially only one single ISO data point, this result should be viewed with a healthy amount of suspicion. More so, because the 8.9 µm feature, if attributed to ammonia as well, would yield a column density of only cm-2, which makes only 8% of the water abundance. To a limited extent, the inconsistency between the two column densities for ammonia could be explained by the fact that radiation with wavelength 8.9 µm is probing deeper into the edge-on disk, while at 3 µm we observe mostly photons leaking out from the poles (Pendleton et al. 1990). If this is true, then ammonia must be more abundant (compared to water) in the ambient cloud material, than in the disk itself. The origin of the 8.9 µm feature is, however, uncertain altogether. It might also be due to methanol or a forbidden line of Ar III gas.
The feature at 6.8 µm is clearly visible. However, it is still unclear, to which molecular species it can be attributed. Possible explanations include methanol, and a blend of several different absorption components (Henning 1996). We do not derive abundances for methanol here, since we do not detect absorption features at 3.4 µm, 3.5 µm, 9.7 µm, or 14.2 µm. The other two identified features are at 4.2 µm and 4.7 µm, indicating the presence of CO2 and CO. The derived abundance ratios relative to water are consistent with the ratios for which the absorbance strengths were determined. The latter are, however, anyhow only weakly dependent on the mixture (Gerakines et al. 1995). To summarize the spectral data, we detected a grain mixture of
with the value for NH3 being rather uncertain.
Despite the disk is seen nearly edge-on, we do not find any silicate absorption feature at 10 µm. Therefore, the silicate abundance in the disk seems to be unusually low. Silicate absorption in the spectra of young stellar objects usually appears in form of a wide, structureless feature peaking at 9.7 µm (Jäger et al. 1994). In comets and Vega-type stars, silicate features are found that have an additional peak at 11.2 µm, which indicates the presence of crystalline silicates. As this shows that the actual position of the silicate feature might appear shifted by a considerable amount, one might now be tempted to speculate that we did detect some silicate absorption in form of the 8.9 µm feature. However, we do not ascribe this feature to an exotic species of silicates and do not take it as an indication of any weak silicate absorption at all.
© European Southern Observatory (ESO) 1998
Online publication: March 30, 1998