Discussion

It is clear from the results of the fitting that it is difficult to discern a definite evolutionary trend in the type of dust found around oxygen-rich stars, since the same minerals give the best fits for all but group O1, for which no dust feature is required. It is also clear that the first detectable minerals to form around these stars are the magnesium-rich silicates. The non-detection of the refractory oxides or of silicates earlier in the condensation sequence (e.g. augite or diopside) may allow one to estimate an upper limit to the amounts of such dust grains around these stars.

There seems to be a distinct lack of evidence for Al₂O₃ grains. Contrary to the identification of this material in ground-based spectra by Hackwell (1972) and in IRAS LRS spectra by VdJW and OdJW, through the presence of the ~12µm emission feature, we find here that this feature may be fit by silicon dioxide (SiO₂). Condensation models (Sedlmayr 1989; Tielens 1990) suggest that Al₂O₃ should be the first dust type to condense around O-rich stars. According to some previous research (e.g VdJW; OdJW; and Tielens 1990) Al₂O₃ then forms a nucleation seed on which the silicates can form a mantle. Alternatively, Nuth & Hecht (1990) and Stencel et al. (1990) suggested that the condensate is a ``chaotic silicate'' in which, initially, the emission from Al-O bonds dominates the spectra, but is then overwhelmed by emission from the more abundant Si-O bonds. In both cases, the above authors agreed that the 12-13µm band seen in the spectra of oxygen-rich stars can be attributed to Al-O bonds and that it signifies the presence of some form of aluminium oxide. Moreover, Al₂O₃ grains found in meteorites (Nittler et al. 1994a,b; 1997; Huss et al. 1995; see section 3.2.3) have isotopic signatures which suggest they were formed around oxygen-rich AGB stars. However, the abundance of such AGB star Al₂O₃ grains is very low (<1ppm; c.f. 6ppm for presolar SiC and 400ppm for presolar diamond). Begemann et al. (1997) studied the laboratory spectra of various forms of aluminium oxide, both crystalline and amorphous, with a view to identifying the 12-13µm feature more clearly. They found that amorphous aluminium oxide could not account for the observations, and that, while one crystalline form of Al₂O₃ could account for the 12-13µm feature, a second feature seen at 21µm in laboratory spectra was not observed in astronomical spectra. They suggested that the 12-13µm feature may come from a form of silicate rather than aluminium oxide. This is confirmed by the results in this thesis which uses optical data of silicon dioxide to fit the 12-13µm feature rather than Al₂O₃, as discussed below and in chapter 5. In fact the spectra showing the 12-13µm feature could not be fit using any of the laboratory optical constants of Al₂O₃. The non-detection of aluminium oxide may allow one to constrain the amount of such material present around oxygen-rich stars using radiative transfer models. Furthermore, this may explain the low abundance of such grains in meteorites relative to other pre-solar grain types.

From the results obtained here it is not possible to be more specific about the dust species detected other than to say that they are various forms of amorphous magnesium silicates. On this point we concur with Day (1979) that the exact nature of the magnesium silicate cannot be determined from the 10µm spectra. In order to get a more unambiguous mineralogy for the dust, observational spectra with a broader wavelength range coverage, together with a better understanding of the nature of the laboratory samples, would be required. This is beyond the scope of the present work, but will hopefully be addressed by the study of ISO spectra (see e.g. Waters et al. 1996). While this sort of ambiguous result is not what had been hoped for, the use of the chi²-fitting technique has confirmed that the mineralogy cannot be determined from the 10µm spectra alone. It may also be possible to use these results to constrain the amount of crystalline silicates present around oxygen-rich stars. It is clear from the fitting results that there is little evidence of crystalline silicates in the 10µm region. In particular, we have not observed the 11.3µm crystalline olivine feature. Non-observance of this feature (and others) may constrain the relative quantities of crystalline and amorphous silicates around these stars, since the crystalline silicates have been observed at longer wavelengths in the ISO-SWS spectra (Waters et al. 1996).

One of the interesting results is the necessity to use SiO₂ to explain the narrow ~9.25µm peak seen in some of the spectra. As elaborated in chapter  5 and section 8.1.1, one of the molecules/minerals that is likely to be involved in the formation of silicates around O-rich stars is silicon dioxide (SiO₂; Pégourié & Papoular 1985). However, the spectral features/optical constants of SiO₂ have never been included in investigations of the features in the spectra of such stars. The forms of SiO₂ found on Earth are diverse and are discussed in more detail in chapter 5, and a summary is given here, in Table 8.12.

    Table 8.12: Silicon dioxide stability temperatures

SiO₂ stable/formation Crystal Structure Main Peak of 2nd peak

type temperature in K Spectral feature

alpha-quartz < 846 trigonal trapezoidal 9.20 12.5&12.8

beta-quartz 846-1143 hexagonal trapezoidal 9.20 12.5&12.8

tridymite 1143-1743 hexagonal holohedral 9.15 12.6

cristobalite 1743-1986 cubic 9.25 12.65

silica glass <~ 1300 amorphous 9.25 12.5

obsidian <~ 1300 amorphous 9.20 --

lechatelierite <~ 1300 amorphous 9.20 12.7

Given the temperatures of stability of each form of SiO₂, it is not unreasonable to assume that all these forms could be constituents of the circumstellar dust around oxygen-rich stars. The spectra of six samples of SiO₂, used in the chi²-fitting routine, are presented in section 5.4. Three of these samples are crystalline SiO₂: quartz, tridymite and cristobalite; and three samples are amorphous: silica glass, obsidian and amorphous SiO₂. All these spectra have a strong features peaking at about 9.2µm and all but obsidian show a minor peak at ~12.5µm.

The inclusion of silicon dioxide in models of circumstellar dust is reasonable. Those sources whose ``silicate'' features actually peak at 9.25µm are possibly best explained in this way. The 12.5µm feature seen in some O-rich circumstellar spectra (see Hackwell 1972; VdJW; OdJW) may also be attributable to SiO₂, rather than Al₂O₃ as has previously been asserted. The only problem with this hypothesis is that the 12.5µm feature and the 9.25µm feature seem to be mutually exclusive in the observed astronomical spectra whereas they appear together in the laboratory spectra (Fig. 5.13 and Table 8.12), with the exception of obsidian which exhibits only the 9.25µm feature. Therefore, obsidian can be used to explain those spectra which exhibit the 9.25µm only, but this does not resolve the problem of those spectra which exhibit the 12.5-13µm feature without a 9.25µm feature. The variations in the relative strengths of the two features from polytype to polytype may be the key. The attribution of the 12.5-13.0µm feature to SiO₂ would not be a problem if the 12.5µm were strong enough, as the 9.25µm feature would be masked by other features in that region. Indeed, the broad feature, most closely associated with the appearance of the of the 12.5-13.0µm feature, extends from 8.5-12µm. This is a wide enough range to obscure the 9.2µm feature, and allow the 12.5-13.0µm feature to be observed. The nature of the dust responsible for the broad feature is very ambiguous and may well be contributed to by some form of SiO₂. It may be possible to use the lack of 9.2µm feature in these cases to constrain the SiO₂ polytype. For instance, the ratio of the strengths of the features is much more in favour of the 12.5-13.0µm feature for the quartz polytypes than for the other polytypes. It should be possible to calculate how much of a given polytype is needed to produce the 12.5-13.0µm feature without the 9.2µm feature protruding above the broad feature. The formation of various types of SiO₂ will be dependent on the ambient conditions. A closer look at the optical properties of the different SiO₂ polytypes may be expedient if more precise identifications are to be made. It may be that different polytypes of silicon dioxide form in the differing conditions around different stars, or that all the SiO₂ polytypes are formed in different regions around a single star.

The stars in group O6 which exhibit a 9.25µm peak (DU Pup, GX Mon, V342 Sgr and V635 Aql) do not show the 12.5µm feature. Although this may be too small a sample to be statistically significant, this could imply that the 9.25µm feature in these stars is due to an amorphous silicon dioxide which does not exhibit the 12.5µm feature. This is substantiated by the chi²-fitting by obsidian, the only form of SiO₂ in our laboratory sample that does not exhibit the 12.5µm feature. If the 12.5µm feature is attributable to crystalline silicon dioxide, then this implies that the earlier groups contain crystalline SiO₂, while the later groups contains amorphous SiO₂ (which may represent a dust evolutionary process). The diminution of the 12.5µm feature in the supposed evolution of circumstellar dust spectra (see Little-Marenin & Little 1990; Stencel et al. 1990) could be explained in terms of the types of silicon dioxide grains being formed. Since the inclusion of SiO₂ into the fitting program has produced reasonable results and there is no theoretical reason for excluding silicon dioxide dust, it is obviously desirable to include the optical constants for various SiO₂ types in radiative transfer models of oxygen-rich stars. The strict temperature ranges of some of the different crystalline SiO₂ types and their (slightly) differing optical constants may be useful in constraining models of dust formation around such stars.

Up: Chapter 8: Oxygen-rich Stars

Angela's Home Page

Angela's Research Page