4. Results and discussion
The molecular abundances relative to the total gas number density are listed in Table 3 for various shock strengths and positions in the wind. As we do not know precisely the position where the wind is chemically frozen due to the acceleration induced by grains, we have considered gas layers very close to the photosphere. Therefore, abundances at 2.2 may not be the exact values frozen in the outflow but are indicative of trends on destruction/formation of species in the inner wind. The dominant molecules are, apart from molecular hydrogen, CO, H2O, N2 and SiO. This was known already from TE calculations applied to O-rich AGB stars and confirms the "parent" character of these molecules. However, and as for the case of carbon stars (see WC98), caution should be exerted in comparing observations with TE calculations. For some "parent" species, the non-equilibrium chemistry does not alter significantly the initial TE abundances. However, other species abundant in the TE photosphere according to Table 3 (e.g. OH, O, SiS and HS) are quickly destroyed in the outflow by the non-equilibrium chemistry generated by shocks.
Table 3. Calculated fractional abundances (relative to the total gas number density) versus shock strength and radius. Data from millimeter observations are listed with the following references: 1 - Bujarrabal et al. (1994), 2 - Lindqvist et al. (1988), 3 - Omont et al. (1993), 4 - Justtanont et al. (1998): in other O-rich miras. No observations for IK Tau , 5 - Menten & Alcolea (1995).
This chemistry is also responsible for the formation of several carbon-bearing species close to the star, in particular CO2, HCN, and CS. The theoretical values derived for HCN and CS are in excellent agreement with the abundances derived from millimeter lines in the outer envelope. Chemically, these species are quite stable and in carbon stars, they travel the entire envelope unaltered until they reach the photo-dissociation regions of the outer wind. The same should occur in O-rich winds as these molecules do not participate to the formation of dust grains (e.g., silicate, corundum) in the inner envelope.
The chemical processes reponsible for the formation of HCN and CS are linked in that both CS and HCN are produced from reactions involving cyanogen, CN. While HCN is formed by the reaction
the formation of CS results from the reaction
CN acts as an intermediate in the formation of the two molecules and is quickly destroyed by atomic hydrogen. Furthermore, the destruction of HCN by atomic hydrogen possesses a high activation barrier while Reaction (2) is highly dependent on temperature and then quite fast in the gas excursions.
The chemical routes to the formation of CO2 are
where hydroxyl OH is formed from the collisional destruction of water, H2O, with atomic hydrogen, and
Reaction (4) represents the dominant formation pathway at small radii while Reaction (5) becomes important at larger radii, explaining the jump in the CO2 abundance at 2 . The rate for Reaction (4) has been determined experimentally but that of Reaction (5) was calculated from thermodynamical data. In view of the uncertainty of this rate, we also consider Reaction (5) to proceed with a typical three-body reaction rate of k cm6 s-1 to test the variation of CO2 abundance with radius. The abundance was lower than that quoted in Table 3 but of the order of . Therefore, CO2 is a direct result of shock chemistry involving the destruction of CO by OH radicals.
Other species appear to be absent from the inner regions of the wind but are observed in the outer envelopes of Miras stars. This is the case for SO which is produced in the photo-dissociation regions by ion-molecule reactions according to WM97. However, WM97 inject at large radii SiS when the molecule does not appear to be "parent" and has a very low abundance in the inner wind (see Table 3). This could explain the discrepancy found by WC97 between their theoretical value and the observed value for SiS which is much lower, implying that SiS is produced in the outer envelope of IK Tau. As for methane and ammonia, they are also absent from the inner wind of Mira stars, a result already derived for carbon stars (WC98). This points to the necessity of invoking formation processes different from pure gas-phase mechanisms and these species have to be produced in the intermediate regions of the wind from gas-grain interaction as first suggested by Nejad & Millar (1988). Whether they enter the outer envelope of IK Tau with the large abundances used by WM97 needs to be confirmed by theoretical models or observations. The same conclusion is drawn for hydrogen sulphide, H2S, which has a very low abundance at small stellar radii. Omont et al. (1993) have proposed a formation route for H2S from gas-grain chemistry in the intermediate wind of Miras involving atomic sulphur and atomic hydrogen. We notice from Table 3 that both S and H enter the intermediate part of the wind with quite large abundances. This result for sulphur could be tested observationally with the search of the [S I] fine structure lines at 25 and 56 microns.
We conclude that the inner envelopes of O-rich Miras are regions of efficient formation of molecules and dust. As for carbon stars, the non-equilibrium chemistry is triggered by the propagation of periodic shocks induced by stellar pulsation. Some of the carbon molecules so far observed in these stars, in particular HCN, CS and CO2 are formed in the inner wind and act as "parent" species throughout the envelope. This prediction awaits confirmation from new ISO data and sub-millimeter observations on O-rich Mira stars.
© European Southern Observatory (ESO) 1999
Online publication: December 4, 1998