Forum Springer Astron. Astrophys.
Forum Whats New Search Orders

Astron. Astrophys. 329, 1035-1044 (1998)

Previous Section Next Section Title Page Table of Contents

4. Analysis of the astronomical "11.3 µm" feature

The analyses of the LRS spectra presented by Baron et al. (1987) and Papoular (1988) are reconsidered here in the light of the previous discussion.

Fig. 9a represents 6 spectra, each of which is the average of from 24 to 142 spectra extracted from Class 4n of the LRS Catalog, and having closely similar profiles. In each case, a power-law continuum (passing through the spectral points at 9.7 and 14 µm) was extracted from the LRS spectrum to give the excess emission associated with the band carrier, which is shown in the figure. We are not interested, here, in the peak between 8 and 10 [FORMULA] m, which is probably emitted by another dust component. For the central features, however, the SiC model is relevant, since the laboratory features cover exactly the same wavelength range. However, in order to compare LRS spectra with laboratory spectra taken in solid matrices, we have first to convert the latter to vacuum. The procedure is the same as in Sect. 3.4 (Eq. 14-16), except that,here, the target matrix has [FORMULA] =1. Fig. 8 shows the result for both our powders.

[FIGURE] Fig. 8. Predicted emissivity spectra of our powders in vacuum. a  Continuous line: SiC212 (laser pyrolysis); b  dashes: SiC4h (mechanical synthesis).

[FIGURE] Fig. 9. a Shortwave spectra of SiC excess emission of the 535 C-rich stars in the LRS Catalog (Papoular, 1988); b  linear combinations of the 2 spectra in Fig. 6: 1) 0 [FORMULA] (a)+(b); 2) 0.39 [FORMULA] (a)+0.78 [FORMULA] (b); 3) 0.77 [FORMULA] (a)+0.58 [FORMULA] (b).

The feature peaks at 11.9 µm for SiC212, and 11.15 µm for SiC4h. Since the LRS spectra (Fig. 9a) exhibit peaks near 11.3 and 11.8, it is tempting to try and fit them with linear combinations of the two spectra in Fig. 8. This is done in Fig. 9b, for spectra 1,3 and 6 of Fig. 9a. We emphasize that no arbitrary frequency shift was introduced in this procedure, and that the only free parameter in each fit was the ratio of the multiplying coefficients of the two curves in Fig. 8. Such small discrepancies as may be seen between Fig. 9a and b are only to be expected since the grain growth and evolution processes in space are not necessarily exactly the same as in the laboratory. A more rigorous procedure would require the multiplication of the emissivities by a Planck function corresponding to the "average" dust temperature for each set of spectra. However, we checked that this has not much of an impact on the profiles or on the peak wavelengths, for temperatures between 200 and 1250K, characteristic of circumstellar dust.

As they are, these fits indicate that the diversity of the observed feature profiles can be understood in terms of different physical conditions in the dust birthplace, giving rise to various grain shape distributions.It is not possible to explain this variety by invoking polytypes. On the other hand, one can state that the cosmic SiC is mostly crystalline.

Goebel et al. (1995), using another classification scheme, also displayed a series of spectra obtained by averaging the LRS spectra in each of their 5 classes of 11.3 µm features. The features differ in width and peak wavelength; the latter ranges between [FORMULA] 11.3 and [FORMULA] 11.7 µm and the overall picture is quite reminescent of our Fig. 9a. The authors interpret this variation as being due to a varying mixture of SiC and a-C:H dust in the CS shells. Papoular (1988) suggested a similar interpretation, based on the apparent correlation of the 11.7 µm peak with the peak near 8.6 µm, both of which he tentatively attributed to another, carbonaceous, dust component. However, a-C:H, for instance, also carries features at 3.4, 6.2 and 7-9 µm, which are not yet known to occur in the spectra of C stars. The recently launched ISO sattelite may help confirm (or otherwise), their existence in the future. Meanwhile, the interpretation given above seems preferable.

Roche et al.(1991) and others more recently, such as Beintema et al. (1996), have detected a feature at about 12.7 µm in various hot nebulae. Baron et al.(1987) also pointed to a feature at 12.8 µm, exhibited by 79 LRS spectra of Class 1n, on the wing of a stronger feature peaking at 11.3 µm and covering the interval 10-14 µm. It should be interesting to explore the possibility that these features be linked to the 12.6-µm feature discussed at the end of Sect. 3, in connection with fiber matts. A linear string of small, connected spheres of crystalline SiC (a more plausible structure in space) should also exhibit this feature. Unfortunately, these wavelengths are at the edges of the useful pass-bands of both spectrographs on-board the IRAS satellite, so that the feature is not easily documented in the LRS. Again, the ISO satellite should give more accurate profiles and positions.

Previous Section Next Section Title Page Table of Contents

© European Southern Observatory (ESO) 1998

Online publication: December 16, 1997