next up previous
Up: Chapter 7: Carbon Stars Previous: An Evolutionary Sequence

Diamonds in the sky

As discussed earlier in this chapter and in chapter 6, various carbonaceous solids are expected to form around carbon-rich AGB stars. So far, the discussion has centred on silicon carbide, PAHs and various forms of hydrogenated amorphous carbon. Another form of carbon that could be present around these stars is diamond.

Diamonds in the interstellar medium were were first proposed by Saslaw & Gaustad (1969), who argued that, although graphite is the thermodynamically stable form of carbon in dust forming regions and in interstellar space, it was possible that diamond could form as a metastable product(see section 3.2.3). Following this theoretical work, interest in interstellar diamonds lay dormant until, in 1987, they were found, in meteorites, rather than in interstellar spectra (Lewis et al., 1987).

Diamonds in meteorites

Interstellar diamonds have been found in various carbonaceous chondrites. Their extra-solar origin has been established in various ways. Firstly, objects incorporated into meteorites must predate the meteorite, and chondrites date from the formation of the solar system, therefore the presence of diamonds within chondrites indicates that the diamonds predate the solar system. The presolar diamonds were actually discovered whilst attempting to find the source of an isotopically anomalous noble gas component in carbonaceous chondrites. There was isotopic evidence that Xenon was enriched by up to a factor of 2 in both the lightest and the heaviest isotopes compared to the dominant xenon component in meteorites and xenon in the Earth's atmosphere. This anomalous component is known as the Xe-HL component. This Xe-HL component is only found in diamond inclusions and is not present in any other presolar grain. Furthermore, it is found in all presolar diamond inclusions that have been studied. The two parts to the Xe-HL component (the heavy and the light isotopes) have proved to be completely inseparable in the laboratory. It is clear that the anomalous xenon is an indicator that the meteoritic diamond is extra-solar in origin. Physical details of these meteoritic presolar diamonds can be found in section 3.2.3.

Diamonds in the interstellar medium

When meteoritic diamonds were discovered, there had been no observational evidence for diamonds in interstellar space. Then, in 1992, Allamandola et al. (1992) found an absorption feature in the spectra of protostars embedded in dense molecular clouds at 3.47µm (2880 cm-1 ), which they attributed to sp3 bonded C-H, i.e. hydrogenated diamond-like particles. Thus, we have evidence of diamonds in space. Their observations suggested that the diamond-like grains seemed to be ubiquitous in dense clouds, while methyl (-CH3) and methylene (-CH2) rich material dominated the diffuse ISM. The diamond signature was only found in dense molecular clouds and not in the diffuse ISM, which was surprising because thorough mixing is expected to occur between the dense and diffuse media. Allamandola et al. (1992) suggested that the absence of -CH2 and -CH3 bands, together with appearance of the sp3 bonded -CH not found in the diffuse ISM, implies that C-rich materials in the diffuse medium do not become incorporated into, or do not survive incorporation into, dense molecular clouds. Neither destruction by shocks (which are weak in dense clouds) or UV photolysis (mild as a result of dust extinction), or H atom attack (which would make -CH2 and -CH3 groups rather than destroy them) can explain the apparent lack of the carbon-rich diffuse cloud components in dense clouds.

It is my proposal that we need to turn the argument around. I would argue that, without any explanation of how it got there, the sp3 bonded -CH is stable in the dense molecular clouds where it is protected from the severe environment outside. Only when it leaves the dense cloud is it converted into -CH2 and -CH3 by the harsher environment. If this is the case, we need to seek out a very effective mechanism by which -CH2 and -CH3 groups would be converted into sp3 bonded -CH inside the dense molecular clouds. However, if this paradigm is correct, the diamonds are formed in dense molecular clouds and are not of interest in a discussion of dust formation around carbon-rich AGB stars. Therefore, let us consider the meteoritic evidence for the origins of interstellar diamonds further.

Sources of meteoritic diamonds

The isotopic complexity of the noble gas component in meteoritic diamonds indicates that it comes from several sources. Other isotopically anomalous elements have also been found, including barium and strontium, which are slightly enriched in r-process isotopes. Nitrogen is also anomalous. The most studied aspect of meteoritic diamond isotope anomalies is the noble gas component. The noble gases are released by stepped heating of diamond residues extracted from meteorites. This has revealed three components, each including all five noble gases: 1) the roughly solar system component released between 200 and 900°C; 2) the HL component (anomalies in all noble gases) released at 1100-1600°C; and 3) a mixture of the first two released at even higher temperatures. Whether the carrier of the first and second components are distinct phases has been the subject of some discussion. It seems that the carrier of the solar system component is a more disordered carbon, with both sp3 and sp2 bond C-H, known as a-C:H (amorphous hydrogenated carbon). There are indications that this phase is lost through metamorphism, since the diamonds in the less primitive meteorites do not have this phase. It has been suggested that the carrier of the solar system component is merely the hydrogenated surface of the diamonds, which is supported by EELS (electron energy loss spectroscopy; Bernatowicz et al. 1990) data. However, the isotopic data implies that it formed at a different time or place to the more anomalous diamond. it is possible that in the diffuse interstellar medium, the diamonds acquire a coating of the a-C:H through hydrogenation of their surfaces, rendering the diamonds invisible by changing their spectral properties to those of methyl or methylene groups. Somehow this coating must disappear when the grains enter the dense molecular clouds in order to fit observations. Therefore a mechanism for the loss of the -CH2 and -CH3 coating needs to be established.

The origin of the meteoritic diamonds is still an enigma. The Xe-HL component must have been formed near a supernova, since the enrichment of the lightest and heaviest isotopes of xenon would proceed through the p-process and r-process respectively, both of which are associated with supernovae. This has led to various hypotheses for the formation of diamonds in space.

Theoretical models of diamond formation in astrophysical environments

Firstly, there is the chemical vapour deposition (CVD) method proposed by Saslaw & Gaustad (1969) and others (e.g. Anders & Zinner 1993 and references therein; Colangeli et al. 1994). This is favoured by the size distribution of meteoritic diamonds which is log-normal and is indicative of grain growth rather than fragmentation of larger grains, which tends to give a power law distribution. For this mechanism to be practicable, the diamonds must form in an environment with a carbon-to-oxygen ratio greater than unity. However, the isotope studies imply that a supernova is involved in the process. Carbon stars are not massive enough to become Type II supernovae and the supernova precursors are carbon poor. A possible mechanism, expounded by various authors (e.g. Lewis et al. 1987), is that diamond grains form and grow around C-rich AGB stars and then receive the noble gases by implantation in the vicinity of a supernova. I believe this mechanism is somewhat implausible because this scenario would imply that some diamonds would be formed around C-rich AGB stars and not necessarily be exposed to supernovae noble gases. These grains should also have been incorporated into the early solar system, a result which is not substantiated by the isotopic evidence.

Another suggestion for the formation of diamond is the collision of graphitic grains in supernova shocks (Tielens et al. 1987). As the efficiency of this process is estimated to be only ~5%, the diamonds should be accompanied by a twenty-fold excess of unconverted graphitic carbon with the same isotopic composition and the same noble gas components. However, there is no evidence of such unconverted graphitic carbon in meteorites. The lack of graphitic carbon cannot be blamed on preferential destruction in the early solar system, because deuterium-rich, and therefore interstellar, organic carbon survived in the same meteorites, despite greater fragility. Even if the graphitic carbon was converted to organic carbon through reaction with hydrogen in the ISM, this would also produce an unseen twenty-fold increase in the amount of organic carbon. Furthermore, work by Daulton et al. (1996) has given persuasive evidence for a low pressure mechanism based on high-resolution TEM studies. Their nanostructural comparison of meteoritic and synthetic diamond crystallites strongly favours a CVD-like process as opposed to one involving high pressure, shock induced metamorphism of pre-existing carbonaceous material.

Jørgensen (1988) tried to solve this problem by invoking a binary carbon star system. Matter flows to the more massive star after it becomes a white dwarf, permitting it to explode as a Type I supernova. Xe-HL ions in the high speed ejecta overtake the diamond dust shell produced during the red giant phase and implant themselves in the diamonds. Clayton (1989), on the other hand, proposed that Xe-HL is made in Type II supernova by neutrino-produced neutrons in the helium shell, and diamonds condense from the expanding shell about a year after the explosion, trapping the ambient Xe-HL. However, this theory only accounts for the heavy xenon component. The heavy and light xenon components have been inseparable in the laboratory, which led Manuel et al. (1972) to argue that the p- and r-process nuclei must have been mixed in the gas phase before Xe-HL was incorporated into the diamonds. There have been several major objections to the theories of both Jørgensen (1988) and Clayton (1989; e.g. Lewis et al. 1989), which are beyond the scope of the present work, but which are taken as negating these mechanisms for the formation of meteoritic diamonds.

Nuth & Allen (1992) invoked a supernova for the Xe-HL, but suggested that the diamonds were made from pre-existing carbonaceous dust rather than supernova or red-giant material. They proposed that small (<100Å) hydrocarbon grains in the vicinity of a supernova are ``annealed'' to diamond by absorption of several far ultraviolet photons, lose all their pre-existing gases and then trap heavy ions and neutral atoms from the supernova ejecta. However, this quantum heating may be a problem: such ``annealing'' of carbonaceous grains requires temperatures of over 1000K. Since the energy of the photon is distributed over the entire grain, such temperatures are only achievable by grains of <90 atoms (Anders & Zinner 1993). It is not clear whether this mechanism can yield diamonds larger than 90 atoms, which comprises the bulk of the size distribution of the diamonds in meteorites.

Another proposed mechanism for diamond formation uses photolysis of hydrocarbons (Buerki & Leutwyler 1991). Ethene (C2H4) and mixtures of ethene, molecular hydrogen (H2) and silane (SiH4) have been decomposed using a laser to obtain spherules of cubic and hexagonal diamonds along with PAHs, organic polymers, graphite and amorphous carbon. The diamond spherules formed had a mean size ranging from 63±24 to 1200±240Å. However, it is not yet known whether this process would work under astrophysically relevant conditions, i.e. lower photon fluxes, lower pressures and higher H/C ratios. If it was a viable formation mechanism, then the range of possible diamond formation sites would be greatly expanded.

At present it seems that the CVD method for producing diamonds is most favoured (Saslaw & Gaustad 1969; Wright 1992; Lewis et al. 1989; Daulton et al. 1996), although detailed mechanisms by which the diamonds obtain their isotopically anomalous noble gases are yet to be understood. This suggests that diamond production around C-rich AGB stars is not unfeasible, and is possibly expected.

Spectra of meteoritic diamonds

There have been various attempts to get representative spectra from meteoritic diamonds from several different meteorites (Murchison, Allende, Orgeuil; Lewis et al. 1989; Koike et al. 1995b; Mutschke et al. 1995; Hill et al. 1997; Andersen et al. 1998). A summary of the features seen in these spectra can be seen in Table 7.5. All of these spectra seem to be entirely different from one another. Only one of them shows the 3.47µm (2880 cm-1 ) band seen in the spectra of dense molecular clouds (Hill et al. 1997). In fact Mutschke et al. (1995) have discredited many of these spectra, claiming that many of the features are artifacts of the extractions and spectroscopic techniques.

The most recently published meteoritic diamond spectrum can be seen in Fig. 7.5, which shows the entire spectrum from 2.5 to 25µm, and Fig. 7.6 which show the 7.5-13.5µm region of the spectrum, relevant to our observations. This spectrum comes from Andersen et al. (1998), who use their newly measured optical properties of meteoritic diamond in a model of the stellar atmosphere of a carbon star. They suggest that diamonds form in the atmospheres of carbon stars and act as nucleation seeds for other dust grains. Hitherto, PAHs were expected to act as nucleation centres. However, Andersen et al. (1998) found that the timescales for PAH formation are too long compared with the dynamical timescales and the gas temperature is too high for PAH formation. The relatively modest opacity and higher condensation temperature of diamond may cause nucleation of diamond grains at relatively high atmospheric temperatures where the velocity field is still negligible.

Given that there are various models for the formation of diamonds around C-rich AGB stars, it would appear that we need to be looking for evidence of diamonds around such stars. The diamond spectrum from Andersen et al. (1998), shown in Figs. 7.5 & 7.6, shows various features which may be detectable in the infrared spectra of C-rich AGB stars. It is not yet clear what environmental ingredients are necessary to induce the IR emission in the diamond grains, however, even if they require ultraviolet radiation to emit in this region, we have already shown that C-rich AGB stars in binary systems could fulfil this requirement (see section 7.2). The meteoritic diamond spectrum shown in Fig. 7.5 covers a relatively large wavelength range (2.5-25µm), and the majority of the features are beyond the wavelength range of our observational spectra. However, as seen in Fig. 7.6, there is a prominent feature at ~8.5-9.5µm which may be detectable in the carbon stars in our sample. Unfortunately this is in the same position as the hypothesised a-C:H (see section 6.2.3), making it difficult to conclusively attribute the feature in the observed spectra to diamond grains. It would be interesting to study the high-resolution, wider wavelength range, ISO-SWS spectra of C-rich AGB stars with a view to matching all the features seen in Fig. 7.5 to features in carbon star spectra.

At present our observations have too limited a wavelength range to satisfactorily investigate the possibility of diamond grain formation around carbon stars.


  Table 7.5: Spectral features from interstellar diamonds
Lewis et al. (1987) Koike et al. (1995b) Mutschke et al. (1995) Hill et al. (1997) Andersen et al. (1998)
cm-1 µm cm-1 µm cm-1 µm cm-1 µm cm-1 µm
3402 2.94 3402 2.94 3700-3000 2.70-3.33 3420 2.92
3236 3.09
3125-2850 3.2-3.5 3000-2800 3.33-3.57 2961 3.38 2954 3.39
2919 3.43 2935 3.41 2924 3.42
2849 3.51 2875 3.48 2854 3.50
~2040 ~4.9
1774 5.64 1725-1590 5.8-6.3 1746 5.73 1728 5.79
1640 6.10 1632 6.13
1462 6.84
1456 6.87
1403 7.13 1400 7.14 1402 7.13
1385 7.22
1361 7.35
1282 7.80
1234 8.10
1173 8.53 1178 8.49 1175 8.51
1144-1138 8.74-8.78
1122 8.91 1125 8.89 1122 8.91
1103 9.03 1111-1109 9.00-9.02 1109 9.02
1090 9.17 1090 9.17 1090 9.17
1072 9.33
1054 9.49
1028 9.73
943 10.61
744 13.44
721 13.86
637 15.71
630 15.88 633 15.80
626 15.98
607 16.5
471 21.2


Figure 7.5: Infrared spectrum of meteoritic diamond from Andersen et al. (1998)

Figure 7.6: 7.5-13.5µm infrared spectrum of meteoritic diamond from Andersen et al. (1998)

Up: Chapter 7: Carbon Stars Previous: An Evolutionary Sequence

Wicked Witch's Webpage
Angela's Home Page
email me
Angela's Research page
Angela's Research Page