Fig. 1 shows the whole SWS spectrum of NGC7538 IRS1 in the 2.4-45 µm region. The spectrum presents several absorption features attributed to silicates and to frozen species such as H2 O, CO, CH3 OH while other features still remain unidentified.
The absorption bands at about 4.27 µm (2340 cm-1) and 15.1 µm (660 cm-1) are attributed to the stretching and bending mode of solid carbon dioxide (CO2) respectively. We have estimated from both bands an abundance of solid CO2 of about 3.4 molecules/cm2 (using the integrated absorbance values A=7.6 cm/molecule and 1.28 cm/molecule for the stretching and bending mode respectively; Yamada & Person 1964) and from the 3 µm band an abundance of solid H2 O of about 1.5 molecules/cm2 (using A=2 cm/molecule; Hagen et al. 1981) which give a relative abundance of carbon dioxide to water ice of about 22%. This is similar to the values observed towards other sources (de Graauw et al. 1996a; Gürtler et al. 1996).
Because of the low resolution of our spectrum, the only ice band that we can compare with laboratory spectra is the bending mode of CO2. In fact in this spectral region resolution is about 2 cm-1. This band shows a double structure with peaks centered at 656 cm-1 and 662 cm-1.
Preliminary comparisons of laboratory spectra of icy mixtures containing CO2 with the ISO spectra (de Graauw et al. 1996a) have shown that it is not possible to find a single spectrum which fits the CO2 observed bands and it has been suggested that the observed features are due to the sum of at least two independent components. In particular a combination of a polar mixture (dominated by H2 O) with a nonpolar mixture (dominated by CO2) has been used. However these fits do not attempt to reproduce narrow substructure in the profiles of the bending mode. In fact it is not possible to find different laboratory spectra which fit each subpeak separately and most laboratory spectra of CO2 in several mixtures do not show substructure in the bending profile. Among astrophysically relevant mixtures, only the bending mode of pure CO2 and O2:CO2 mixtures show a substructure. However in the case of pure CO2 particle shape effects modify the band profile and good fits to the observed bands are only obtained for a very restrictive set of parameters (de Graauw et al. 1996a), similarly in O2:CO2 mixtures subpeaks are very sensitive to matrix conditions and concentrations. In particular Ehrenfreund et al. (1997) have studied several ice mixtures containing CO2. As an example in CO:O2:CO2 mixtures the bending mode shows two peaks which shift from 658.7 and 661 cm-1 in a 100:50:4 mixture at 10 K to 658.2 and 662.2 cm-1 in a 100:50:32 mixture and in some cases also a third peak at 662.7 appears. In H2 O:CO2 mixtures (Sandford & Allamandola 1990) the peak position of the CO2 bending mode shifts from 653.3 cm-1 in a 50:1 mixture to 654.7 in a 1:1 mixture. When CO2 is produced after ion irradiation of ice mixtures (Palumbo et al. 1998) the peak position of the bending mode varies from 651 cm-1 in a H2 O:CH3 OH=2:1 mixture to 661 cm-1 in a CO:N2 =1:1 mixture. The CO2 bending mode observed towards obscured objects shows subpeaks at about 656 cm-1 and 662 cm-1 (de Graauw et al. 1996a; this work). In the case of GL 2136 also a third peak at about 649 cm-1 appears (de Graauw et al. 1996a). It is interesting to note that the relative intensity of these peaks changes from source to source suggesting that different components (polar, nonpolar, other?) give a different relative contribution along different lines of sight probably depending on the different thermal and energetic processing of the dust.
We have compared the band due to the bending mode in NGC7538 IRS1 with several laboratory spectra of ion irradiated mixtures 1. CO2 is in fact easily produced by ion irradiation of ice mixtures containing C-bearing and O-bearing species (Strazzulla et al. 1997; Palumbo et al. 1998). Again, in most cases, the bending mode of CO2 produced doesn't show subpeaks. This is not the case when CO2 is produced after ion irradiation of CO:O2 mixture (see Fig. 3 in Palumbo et al. 1998). Fig. 2 shows the profile of the CO2 bending mode produced after ion irradiation of CO at 10 K and warmed up to 20 K and 40 K. It is interesting to note that the band profile changes significantly at 40 K. At this temperature most of the original CO has sublimated; most CO2 produced leave the target with CO while some is left over trapped in a suboxides matrix (Palumbo et al. 1998). The profile of the bending mode band in this case shows a substructure which however is different from that of pure CO2.
Fig. 3 shows a comparison of the observed feature (crosses) with the laboratory spectra of CO2 produced after ion irradiation of CO at 10 K and warmed-up to 40 K (dashed line) and of CO2 produced in a H2 O:CH3 OH mixture (100 K; dotted line). The solid line is the sum of the two independent components. Also in this case it is then possible to separate a nonpolar and a polar component which contribute comparably to the optical depth of the observed feature. Furthermore the substructure in the profile seems to be well reproduced.
© European Southern Observatory (ESO) 1998
Online publication: June 2, 1998