- Open Access
On the detections of C60 and derivatives in circumstellar environments
Earth, Planets and Space volume 65, Article number: 1 (2013)
C60 (buckminsterfullerene) was recently discovered in a variety of circumstellar environments by the Spitzer Space Telescope, suggesting that the envelopes around evolved stars are active sites for the synthesis of fullerenes. However, the physical state, excitation mechanism, and formation route of circumstellar C60 are not completely understood so far. These open issues are discussed in this paper. For that purpose we investigate the observed wavelengths and strengths of C60 bands and compare them with the experimental values. We also statistically study the environments and emission properties of the C60 sources. We would like to stress that improved flux measurements and more accurate Einstein coefficients are required to draw solid conclusions. Furthermore, we present possible detections of hydrogenated C60 and, and discuss their implications on fullerene chemistry in circumstellar environments.
The Ih-symmetrical buckminsterfullerene C60, arranged as 12 pentagons and 20 hexagons, is the smallest closed carbon cage molecule (fullerene) satisfying the isolated pentagon rule (IPR; followed by higher fullerenes C70, C74, C76, C78, C80, and so on), and thus is remarkably stable. Although its existence has been early predicted (Osawa, 1970; Bochvar and Galpern, 1973), C60 was first discovered in laboratory experiments simulating the circumstellar chemistry (Kroto et al., 1985). Krätschmer et al. (1990b) developed a method to efficiently produce C60 in the laboratory, making it possible to study in detail its electronic and vibrational properties. Because of its high stability and symmetry, C60 can be taken as a benchmark to study other fullerenes. C60 and its derivatives have long been suspected to be present in the universe. The search for this compound in nature started soon after its synthesis in the laboratory, and its detections in geological materials and meteorites have been reported by several groups (e.g. Buseck et al., 1992; Becker and Bada, 1994; Becker et al., 1994; Heymann et al., 1996) 1 Footnote 1.
The detection of C60 in deep space has proven a long lasting challenging task (Snow and Seab, 1989; Clayton et al., 1995; Kwok et al., 1999; Moutou et al., 1999; Herbig, 2000; Sassara et al., 2001). C60 has three broad peaks at 216, 264, and 339 nm in its electronic spectrum, and four infrared (IR) vibrational modes at 7.0, 8.5, 17.4, and 18.9 μm. Hydrogen-poor and carbon-rich circumstellar envelopes, such as R Coronae Borealis (RCB) stars (Goeres and Sedlmayr, 1992), are analogous to the laboratory conditions for the synthesis of fullerenes and thus were suggested to be ideal sites to search for C60; however, this was not supported by observations (García-Hernández et al., 2011b). So far, there is no successful report on the detection of electronic transitions of C60. The Infrared Spec-trograph (IRS; Houck et al., 2004) on the Spitzer Space Telescope (Spitzer; Werner et al., 2004) now provides an unprecedented opportunity to detect the IR vibrational transitions. The first convincing detection was recently made by Cami et al. (2010), who detected C60 and C70 in the Spitzer/IRS spectrum of the young planetary nebula (PN) Tc 1. Soon after, C60 was detected in a variety of evolved stars, including PNs in the Milky Way and the Magellanic Clouds, a protoplanetary nebula (PPN), post asymptotic giant branch (AGB) stars, modestly hydrogen-deficient RCB stars, and a peculiar binary object (García-Hernández et al., 2010, 2011a, b; Clayton et al., 2011; Gielen et al., 2011; Zhang and Kwok, 2011; Evans et al., 2012; Roberts et al., 2012). Fullerenes may also be widely present in the interstellar medium (ISM). has been proposed as the possible carrier of diffuse interstellar bands (Léger et al., 1988), and its electronic spectra in Ar and Ne matrices indeed show a reasonable match to two near-IR diffuse bands (Foing and Ehrenfreund, 1994). Photoabsorption by fullerenes and multi-layered fullerenes (buckyonions) has been suggested as the origin of the 217 nm extinction feature (de Heer and Ugarte, 1993; Iglesias-Groth, 2004; Li et al., 2008). Recently, the C60 IR bands have been detected in reflection nebulae (RNs) and the Orion nebula (Sellgren et al., 2010; Rubin et al., 2011). Furthermore, Roberts et al. (2012) detected C60 in pre-main-sequence objects including young stellar objects and a Herbig Ae/Be star. These detections suggest that C60 can be formed in a short time scale between AGB and PN stages, and can survive (or be resynthesized under favorable conditions) in the ISM.
However, based on a laboratory spectroscopic investigation, Duley and Hu (2012) presented that the 7.0, 8.5, 17.4, and 18.9 μm features in the spectra dominated by aromatic infrared bands (AIBs) actually arise from proto-fullerenes that are precursors of C60, and thus the C60 molecules were detected only in sources that do not show AIBs. García-Hernández et al. (2012b) did not detect C60 electronic transitions in the optical spectrum of a RCB star that exhibit AIBs and the (supposed) C60 IR features, and suggested that this supported the conclusion of Duley and Hu (2012). However, it is surprising that C60 electronic transitions are not detected in the high-quality optical spectrum of the strongest C60 source without AIB emission, Tc 1 (García-Hernández and Díaz-Luis, 2013).
The formation route of C60 is a subject of debate. Investigating this problem may provide significant insights into the circumstellar chemistry. The experiments of Kroto et al. (1985) suggest that small carbon clusters can self-assemble into C60 in a hydrogen-poor environment (bottom-up), and otherwise the formation of polycyclic aromatic hydrocarbons (PAHs) is favored. This scenario is supported by the fact that no AIBs are detected in the PN Tc 1 (Cami et al., 2010). Nevertheless, Jäger et al. (2009) found that fullerenes can be formed by gas-phase condensation reactions in the presence of hydrogen at very high temperature (>3500 K). Another formation mechanism of C60 is through shock- or UV-induced decomposition of hydrogenated amorphous carbon (HAC) grains (top-down), through which PAHs can be simultaneously formed (García-Hernández et al., 2010; Bernard-Salas et al., 2012; Micelotta et al., 2012). The supporting evidence comes from the facts that most of the C60 sources also exhibit AIBs, and C60 is not detected in extremely hydrogen-poor RCB stars (García-Hernández et al., 2011b). A similar mechanism was proposed by Berné and Tielens (2012), in which C60 is formed through UV-induced dehydrogenation and isomerization of graphenes. This model can explain the observation of the RN NGC 7023 by Sellgren et al. (2010) that the C60/AIB flux ratios decrease with increasing distance from the central star. Moreover, Boersma et al. (2012) found that C60 can coexist with the carriers of AIBs in the shielded regions of Orion.
A related question is what drives the excitation of C60 IR emission. Thermal excitation of solid fullerenes (Cami et al., 2010; García-Hernández et al., 2011a) and UV-excitation of gas-phase C60 molecules (Sellgren et al., 2010) have been discussed. The first systematical comparison of the two mechanisms was made by Bernard-Salas et al. (2012) who found that the IR spectra of three PNs with strong C60 features favor fluorescence as the excitation mechanism. However, Roberts et al. (2012) and García-Hernández et al. (2012a) argued that thermal excitation can better explain the observed C60 flux ratios.
Mapping the spatial distributions of C60 and AIBs is essential to understand the formation routes of fullerenes in astrophysics. However, this is very difficult for circumstellar envelopes given their compactness and the weakness of C60 emission. Another approach to investigate this problem is to probe hydrides of fullerenes (fulleranes). The presence of fulleranes in space has been predicted (e.g. Webster, 1992, 1993; Petrie and Bohme, 2000; Cataldo and Iglesias-Groth, 2009). When hydrogen atoms are attached to carbon atoms of C60, the conversion of delocalized π-into σ-character orbitals decreases the bond angle strain. Therefore, when exposed to atomic hydrogen, C60 can be quickly hydrogenated into C60H m (m = 2,4,...,60). On the other hand, heavily hydrogenated C60 is highly unstable because of large strain resulting from hydrogen-hydrogen repulsion. It follows that moderately hydrogenated C60 is likely to exist in circumstellar environments. Indeed, in laboratory conditions, it is easy to produce C60H36 that can further form C60H18 through thermal annealing (e.g. Iglesias-Groth et al, 2012) but the production of fulleranes with a larger or lower degree of hydrogenation is much more difficult. Roberts et al. (2012) noticed that two C60 sources also exhibit the very rare C-H emission from hydrogen-coated nanodiamonds, a species that can be formed together with fullerenes, so it is reasonable to believe that C-H emission from C60H m is also detectable in some C60 sources.
In this paper, we study the properties of all the reported C60 sources and compare the observations with existing experimental and theoretical results. In Section 2, we investigate the wavelengths and widths of the C60 IR bands, aiming at understanding the physical state of circumstellar C60. In Sections 3 and 4, we respectively discuss the excitation and formation problems of fullerenes based on the measured band strengths. In Section 5, we present a search for fulleranes in the carbon-rich PPN IRAS 01005+7910 utilizing the spectrum obtained by the Infrared Space Observatory (ISO; Kessler et al, 1996), and briefly discuss and in the PN Tc 1. The conclusions are given in Section 6.
2. Wavelengths of the IR-active Bands
C60 has 174 normal vibrational modes that are distributed among 46 frequencies owning to its icosahedral symmetry. Of these modes, only 4 IR-active vibrational bands of F1u symmetry and 10 Raman-active bands of Ag and Hg symmetries have been observed in laboratory experiments. Throughout this paper we only focus on the four IR C60 bands that are of astrophysical interest. Because of inter-molecular interactions, the wavelengths of IR bands from solid C60 clusters might differ from those from gas-phase molecules and depend on the environment temperatures. Therefore, the comparison between the observed and experimental wavelengths of the C60 IR bands enables us to investigate the physical state of circumstellar fullerenes.
Krätschmer et al. (1990a) and Frum et al. (1991) have made the laboratory measurements of the solid- and gas-phase spectra of C60, respectively. By comparing their results and the observed spectra, Evans et al. (2012) concluded that C60 in the peculiar binary XX Oph is in the solid phase. Temperature-dependent studies of C60 IR spectra have been presented by Chase et al. (1992), Nemes et al. (1994), and Iglesias-Groth et al. (2011) showing that the peak wavelengths shift to higher values with increasing temperature. In Fig. 1, we compare these experimental wavelengths and the spectra of three C60 sources, Tc 1, NGC 7023, and IRAS 01005+7910. There is no evidence showing that the peak wavelengths of C60 features shift between different sources 2 Footnote 2, suggesting that fullerenes in all the circumstellar envelopes are probably in a similar physical state. An inspection of Fig. 1 indicates that the peak wavelengths of the 17.4 μm feature seem to suggest that circumstellar C60 is in the solid phase at a low temperature (<500 K) rather than in the gas phase. The conclusion, however, is arguable in that the spectra of NGC 7023 and IRAS 01005+7910 exhibit the 16.4 μm features, and thus the 17.4 μm feature is probably blended with an AIB (Sellgren et al., 2010; Berné and Tielens, 2012). The observed peak-wavelengths of the other three C60 features lie well inside the experimental wavelength ranges of both solid- and gas-phase C60. Therefore, we cannot completely rule out a gas-phase origin although a solid-phase origin is favored.
Experimentally, the 17.4 and 18.9 μm features in the gas-phase IR spectrum have a full-width at half maximum (FWHM) of about 0.4μm at 1065 K (Frum et al., 1991). Chase et al. (1992) found that the band widths of C60 shift little in the solution and in the solid state, suggesting that C60 in the solid state is virtually a gas-phase-like, freely rotating molecule. The band widths are partially due to vibration-rotation coupling, scaling roughly with the square root of the temperature. Therefore, one can, in principle, estimate the temperature by comparing the observed band widths with the experimental values. A comparison between the experimental results of Frum et al. (1991) and Chase et al. (1992) suggests that the band widths of solid- phase C60 are 2–3 times narrower than those of gas-phase at the same temperature. The FWHMs of the observed 17.4 and 18.9 μm features are 0.26–0.36μm, implying a gas temperature of 400–900 K. The implicated temperature might be higher if C60 is in the solid phase. While interesting, these results should be viewed with some caution because they might be related to the experimental conditions, which are quite different from those of circumstellar envelopes. The 17.4 and 18.9 μm features might be blended with weak C70 emission (see, e.g., figure 1 of Cami et al., 2010), slightly affecting their FWHMs. Moreover, if the C60 bands arise from regions with large temperature fluctuations, the situation will be complicated in that the shifts of peak wavelengths can make an additional contribution to the band broadening.
3. On the Excitation of C60 Vibrational Spectra
It is essential to understand the excitation mechanism of fullerene emission as it enables us to accurately determine the abundance of C60 and establish how useful the C60 bands are to probe circumstellar environments. Two possible scenarios have been discussed in previous paper, but results remain conflicting. Cami et al. (2010) found that a thermal model can interpret the C60 emission in Tc 1. In that case, C60 is attached to dust grains which are in equilibrium with the UV radiation field, and one can calculate the thermal temperature from the Boltzmann excitation diagram of C60 band ratios. García-Hernández et al. (2011a) derived thermal temperatures of 200–500 K for a sample of C60– detected PNs. On the other hand, Sellgren et al. (2010) assumed that C60 is isolated free molecules in the gas phase and has a similar excitation mechanism with PAHs as discussed by Allamandola et al. (1989). In this scenario, C60 is excited to high electronic states by absorbing UV starlight, and then the electronic energy is quickly redistributed via internal conversion, followed by IR emission through cascading down the vibrational ladder to the ground vibrational state. Since the probabilities of UV-photon absorption and IR-photon emission respectively depend on the absorption cross-sections and Einstein A-values of the corresponding energy levels, a Monte Carlo technique3 can be used to simulate this process. Therefore, if the UV-excitation model is valid, the C60 IR bands can be used to probe the circumstellar radiation field. The model of Sellgren et al. (2010) can reasonably explain the C60 intensity ratios observed in NGC 7023, but predicts too large I7.0/I18.9 ratios in NGC 2023. A recent study of three PNs by Bernard-Salas et al. (2012) does not favor the thermal model because (1) the excitation classes of these PNs are very different while the C60 flux ratios are fairly similar, and (2) C60 is located in a region far from the central star and would be difficult to be heated to the deduced thermal temperatures. However, a different conclusion was subsequently drawn by García- Hernández et al. (2012a) who subtracted the contribution of C70 to the observed 7.0 μm flux and found that the UV- excitation model cannot explain the C60 flux ratios for a larger sample of PNs. Bernard-Salas et al. (2012) did not take into account the contribution of C70 because the band strengths cannot be accurately estimated without appropriate excitation model.
To investigate this problem, in Fig. 2 we compare the 17.4 μm / 18.9 μm and 7.0 μm / 18.9 μm band ratios for three well C60–detected sources at different evolutionary stages. The predictions of UV-excitation (Sellgren et al., 2010) and thermal models are also plotted in Fig. 2, where the A-values are derived from Choi et al. (2000) for both models. The thermal model is constructed using the same way as described by Bernard-Salas et al. (2012). Similar to the thermal model for > 500 K, the UV-excitation model predicts a nearly invariable 17.4 μm / 18.9 μm band ratio (also see Bernard-Salas et al., 2012). An inspection of Fig. 2 shows that the observations are situated at positions closer to the predictions of the UV-excitation model. In the scenario of UV-excitation, the 7.0 μm /18.9 μm band ratios increase with increasing UV radiation. This is consistent with the observational facts that among the three sources the central star of the PN Tc 1 has the highest effective temperature, and it is followed by the PPN IRAS 01005+7910 and the RN NGC 7023. Nevertheless, the observed 17.4 μm / 18.9 μm band ratios appear larger than the model predications. This is partially due to the contamination of an AIB to the 17.4 μm feature (except for Tc 1 which does not show AIBs). Another possible cause is that the A17.4 value of Choi et al. (2000) has been underestimated. This is also shown in figures 5 and 6 of Bernard-Salas et al. (2012).
In Fig. 3, we compare the A-values obtained from the band strengths given in previous literatures, as listed in Table 1. The theoretical band strengths are remarkably inconsistent with the experimental values. The theoretical calculations, based on semi-empirical or first-principle models, differ significantly in the resultant band strengths, especially for the 7.0 μm / 18.9 μm band ratios. The experimental band strengths are more concentrated in Fig. 3 than the theoretical values, yielding an average strength ratio of (0.35 ± 0.07) : (0.27 ± 0.09) : (0.34 ± 0.08) : 1.0 for the 7.0, 8.5, 17.4, and 18.9 μm bands. We can see that, indeed, the A17.4 value of Choi et al. (2000) is a factor of 1.3 lower than the average value. Through the comparison, we estimate that the A-values have a 20-30% error. It should be noted that all the experiments were performed in condensed phases and the results might be perturbed by inter- molecular effects. Experiments of gas-phase C60 will be invaluable to improve the A-values in the future. In addition, the UV-excitation model also relies on the photoabsorption cross section of C60 which contains some uncertainties (Yasumatsu et al., 1996; Iglesias-Groth et al, 2002; Yagi et al, 2009).
The thermal-model predictions using A-values obtained by different groups are compared in Fig. 4 along with the observed band ratios of all the C60 sources. It is clear that with higher A17.4 value the prediction of thermal models can better match the observed 17.4 μm / 18.9 μm flux ratios. Bernard-Salas et al. (2012) and García-Hernández et al. (2012a) gave quite different flux ratios.
The C60 8.5 μm band can be detected only in objects that do not exhibit strong AIBs. In Fig. 5, we compare the 8.5 μm/18.9 μmand7.0 μm/18.9 μm band ratios. Compared with those by Bernard-Salas et al. (2012), the band ratios obtained by Bernard-Salas et al. (2012). The main problem of the thermal model is that a high fullerene temperature (mostly > 300 K) is required, and hardly agrees with those implicated by dust thermal emission. On the other hand, Fig. 5 demonstrates that the UV-excitation model predicts an unacceptably low average photon energy (mostly <5 eV). This is the main argument of García-Hernández et al. (2012a) against the UV-excitation model. However, the UV-excitation model of Sellgren et al. (2010) applies only for free-flying C60 molecules. If the circumstellar fullerenes are in a cluster state, more intense radiation field will be expected to excite the C60 IR bands, and the model of Sellgren et al. (2010) only gives the lower limits of the average photon energy. If this was the case, the scattering band ratios in Figs. 4 and 5 would partially reflect the variation of cluster sizes in these sources.
Some alternative mechanisms, such as the release of chemical energy (Duley and Williams, 2011) and atom impacts (Papoular, 2012), have been introduced to explain the excitation of AIBs. It remains unknown whether these mechanisms play a role on the excitation of C60 IR bands.
4. On the Origin of C60
The synthesis of fullerenes in nature has been a subject of intense discussion for many years. It is unclear whether they form through assembly of small carbon- bearing molecules (bottom-up) or fragmentations of large compounds (top-down). The discovery of fullerenes in rocks (Buseck et al., 1992) suggests that fullerenes can be generated through a solid-state process. Many interesting scenarios, such as cataclysmic impact, extensive wildfires, chondritic impactor, vaporization of carbon by lightning strike, and pyrolysis of organic matter, have been proposed to explain the formation of fullerenes in geological environments (see Buseck, 2002, and references therein). To investigate C60 formation in circumstellar environments, it is instructive to compare the observations with experimental knowledge. In laboratory, the most effective two ways to produce fullerenes are through vaporization of graphite followed by growth of carbon clusters and through combustion of hydrocarbons. The other experimental routes, such as pyrolysis of hydrocarbons and chemical synthesis, are unlikely to happen under circumstellar environments. The method of graphite evaporation (bottom-up) requires a carbon-rich environment, otherwise the fullerene formation is strongly suppressed. However, C60 has been detected in O-rich binary post-AGB stars (Gielen et al, 2011), contrary to what is required by the bottom-up route 4 Footnote 4.
In Table 2 we compare the effective temperatures of the central stars and the other emission features of C60 objects. They all have effective temperatures of <35,000 K, suggesting that fullerenes cannot form or survive in strong UV-radiation fields. This situation is similar to that of the unidentified 21 μm feature that has been discovered in a small number of PPNs (Kwok et al, 1989). Most of the sources exhibit AIBs arising from C-H modes, consistent with the experimental results of hydrocarbon combustion. Not all of them show the 30 μm feature and the 15–20 μm plateau. It is intriguing that the plateau emission at 6–10 /im and 10–14 μm is revealed in all the sources, indicating that their carriers are likely to be related to fullerene formation. Kwok and Zhang (2011) presented that AIBs and plateau emission can be uniformly attributed to stretching and bending modes of mixed aromatic-aliphatic organic nanoparticles (MAONs), and during stellar evolution, aliphatic chains can be processed into aromatic rings. García-Hernândez et al (2010) and Micelotta et al (2012) proposed that fullerenes can be formed from HAC grains and “arophatic” clusters, both of which have a structure similar to that of MAONs. Thus, we infer that circumstellar fullerenes are produced by MAONs following the scheme suggested by Micelotta et al (2012). In this scenario, proto- fullerenes, as proposed by Duley and Hu (2012), are first formed through dehydrogenation of MAONs accompanied by introduction of pentagons following the IPR and a pathway of minimizing the number of dangling bonds, and later on are processed into vibrational-excited closed cages by reducing the number of dangling bonds to zero. The fullerenic cages can be transformed into the most stable isomers by means of a series of Stone-Wales rearrangements. Finally, the fullerenic cages are shrink to smaller cages like C60 by releasing excess energy. The dehydrogenation may be induced by either UV photons or shocks, but the latter seems to be favored at least for a few C60sources whose central stars have relatively low effective temperatures (<7000 K; Table 2). The percentage of carbon in circumstellar C60 is estimated to be less than 0.3% (Zhang and Kwok, 2011; García-Hernândez et al, 2012a). Since the carriers of AIBs presumably lock up 6% of the cosmic carbon (Cerrigone et al, 2009), we infer that only small amounts of MAONs can be transformed into C60 even under the most favorable conditions.
Because MAONs have a 3D structure, the formation process of fullerenic cages can take place in a layered structure. This provides a possible route for the formation of buckyonions, which have been suggested as the carrier of the 217 nm interstellar absorption feature.
The C70/C60 abundance ratio has important implications in understanding the formation route of fullerenes. The C70/C60 ratio of PNs ranges from 0.02 to 0.21 (García- Hernândez et al, 2012a), comparable to the values of geological fullerenes (0.08 in the Allende meteorite, Becker et al., 1994; 0.21–0.36 in terrestrial clays, Heymann et al., 1996). In laboratory, the C70/C60 ratio is highly variable. The hydrocarbon combustion process yields a C70/C60 ratio ranging from 0.26 to 0.57 (Howard et al., 1991), which tends to increase with increasing pressure. This ratio is much larger than that obtained through graphite evaporation methods (0.02–0.18, e.g. Ajie et al., 1990). One cannot simply argue that the low C70/C60 ratio in PNs seems not to favor the top–down scenario, because the circumstel- lar envelopes have a significantly lower pressure than laboratories and the fullerene formation in combustion experiments is not completely understood. Moreover, this range of C70/C60 ratio is very uncertain since C70 has been measured only in a very few sources. The abundance of generated higher fullerenes beyond C70 is relatively low, which cannot be theoretically explained. In addition, we cannot rule out the possibility that the C70/C60 ratio can be modified by chemical processes in circumstellar envelopes. Cir- cumstellar envelopes around evolved stars, therefore, provide an unique laboratory to investigate fullerene formation and processes under low pressure environments.
5. Derivatives of C60
5.1 A possible detection of fullerane in IRAS 01005+7910
The pioneer work of Webster (1992) suggested that the wavelength of the aliphatic C–H stretching vibration may shifts toward longer wavelengths from 3.4μm due to the absence of hydrogen atoms on neighboring carbon atoms. Each carbon atom of C60H m has three neighboring carbon atoms, results in three possibilities of the absence of hydrogen atoms. Therefore, C60H m is expected to exhibit three peaks in the wavelength range longer than 3.4 μm. The experimental spectra of C60H18 (Iglesias-Groth et al., 2012) do exhibit three peaks in the wavelength range from 3.4–3.6 μ m. Iglesias-Groth et al. (2012) proposed that because the C–H stretching band of C60H m is intense and can be easily distinguished from other features, it can serve as an indicator to search for these molecules in astrophysical environments. This band, however, is out of the wavelength range accessible to Spitzer/IRS.
Previously, we have detected C60 and AIBs in IRAS 01005+7910 (Zhang and Kwok, 2011). If C60 and hydrogen are located in the same region, fulleranes are very likely to be present in this object. As a PPN, IRAS 01005+7910 does not emit atomic lines, and thus provides a good platform to search for fulleranes. The ISO spectrum of IRAS 01005+7910 has been described by Hrivnak et al. (2000). Figure 6 displays the continuum-subtracted ISO spectrum of IRAS 01005+7910 in the C–H stretching region. The sp2 C–H stretch at 3.3 μm has been detected by Hrivnak et al. (2000). The weak peak at 3.42 μm might be due to the asymmetric stretch of -CH2- groups or just an artifact. The three peaks, presumably ascribed to fulleranes by Webster (1992), are clearly visible at 3.48, 3.51, and 3.58μm, and have fluxes comparable to the 3.3 μm feature. Through a multi-Gaussian fitting, we estimate the fluxes of the four features, corresponding to 0, 1, 2, and 3 non-H-bonded neighboring carbon atoms (referred as R0, R1, R2, and R3 hereafter), tobe 1.64 × 10−15, 2.69 × 10−15, 1.78 × 10−15, and 1.61 × 10−15 Wm−2, respectively. Note that the R3 feature might be blended with an unknown feature (Fig. 6).
Assuming that the C60 emission and the C–H stretching emission of C60H m origin from UV-excitation, the fractions of carbon locked up in fullerenes and fulleranes and ) can be estimated using the method described by Berné and Tielens (2012). Based on the fluxes of C60 bands given by Zhang and Kwok (2011), we find, suggesting that about 50 percent of fullerenes have been hydrogenated.
Under the assumption that all the features have the same oscillator strength, the relative strengths of R0, R1, R2, and R3 are equal to the relative probabilities of the four cases (0, 1, 2, 3 non-H-bonded neighboring carbon atoms), and thus the relative strengths of the four C60H m features can reflect the degree of hydrogenation (the m value). With increasing number of C atoms bonded with H, the strength of R3 decreases compared to R0. Figure 7 depicts the calculated relative strengths as functions of m values. The observed strengths indicate that the m value is very likely to lie within the range from 25–40. This is consistent with the experimental results in which C60H36 is the dominant product of hydrogenation reaction of C60. Furthermore, experiments show that C60H36 can be transferred to C60H18 through thermal annealing. We find that the relative strength of R3 is slightly higher than that expected for C60H36 (Fig. 7), suggesting that C60H18 might be also present. On the other hand, the calculations were purely from a mathematical consideration and did not take into account the chemical structure. Therefore, the fraction of R3 relative to R0 might be underestimated due to ignoring the hydrogen-hydrogen repulsion.
IRAS 01005+7910 does not exhibit the 21μm feature. Many candidate carriers of this feature have been proposed in previous studies (see Zhang et al., 2009, and references therein), among which Justtanont et al. (1996) and García- Lario et al. (1999) have attributed it to the mixture of fullerenes with various degree of hydrogenation. The absence of the 21 μm feature in IRAS 01005+7910 and the non-detection of C60 in 21 μm-detected PPNs do not support this identification.
During the post-AGB phase, the UV radiation field gradually increases with the evolution of the central star, and molecular H2 is more likely to be photodissociated into atomic H. Moreover, shocks created by fast stellar winds tend to dissociate H2 through collisions. This enhances the possibility of hydrogenation of C60. If C60 is highly hydrogenated, the bond-breaking may occur due to large angle strain, and thus destruct this compound. Besides, intense UV light can directly destroy fullerenes. This can account for the fact that fullerenes were never observed in evolved PNs.
5.2 and in Tc 1
C60 has an ionization potential of 7.6 eV, and thus is possibly present as the cation in circumstellar environments. Ionized C60 may contribute to the diffuse interstellar bands and initiate intriguing chemical reactions (e.g. Foing and Ehrenfreund, 1994; Moutou et al, 1999; Petrie and Bohme, 2000; Leidlmair et al., 2011). is generally thought to be scarce in circumstellar environments in that electron attachment to C60 is prohibited due to its high activation barrier (0.26 eV). However, there are alternative routes generating (Petrie and Bohme, 2000). Recently, Berné et al. (2013) detected IR emission bands in NGC 7023. have never been detected in space so far.
Based on an experiment in 5K neon matrices, Fulara et al. (1993) obtained the IR spectrum of and, which reveals the vibrational features at 7.11 and 7.51 μm from, and those at 7.22 and 8.32 μm from. Figure 8 shows the spectrum of Tc 1, where the wavelengths of and bands are indicated. It is clear that there is no detectable emission, and thus the content of circumstel- lar is safely negligible. However, the presence of cannot be ruled out. Although the strong forbidden line and C60 band hamper the detection of the 7.11 μm band, there is a blended feature on the red side of the H I (Pfα) line at 7.46μm, which may be partially contributed by the 7.51μm band. After subtracting the continuum and 6–10 μm plateau emission, we decompose this feature, as shown in the insert of Fig. 8. A broaden 7.51 μm band with a FWHM of 0.1 μm and two narrow atomic lines that are interpreted as H I and [Ne VI] lines can reasonably fit the observations. According to the fitting, the flux of the 7.51 μm band is estimated to be about 1.0 x 10−15 Wm−2.
The abundance of can bedetermined using the for-mulae given by Moutou et al. (1999). For the calculations (see equation 4 of Moutou et al., 1999), the C/H abundance ratio was taken from Pottasch et al. (2011), and the 7.11 μm / 7.51 μm cross-section ratio was assumed to be 2.2. Based on a blackbody fitting of the Spitzer spectrum of Tc 1, we estimate that the total IR emission from dust is (1.5–3.0)× 10−12 Wm−2. As a result, we obtained that about 0.12%–0.23% carbon is locked up in in Tc 1 if the detection is real. This is comparable to the upper limit of 0.26% in NGC 7023 (Moutou et al., 1999), and the value of 0.3%–0.9% in the ISM that was estimated by Foing and Ehrenfreund (1994) from two near-IR diffuse bands.
However, as noted recently by Berné et al. (2013), a new spectroscopic measurement performed by Kern et al. (2012) suggests that the 7.51 μm band might be due to rather than. Thus the above discussion is only tentative.
C60 and higher fullerenes have been detected in circum–stellar and geological environments. This strengthens the idea that a variety of carbon–based compounds, including fullerenes, can be efficiently produced by stars, be ejected into the ISM, reach the early Solar System, and be partially brought to Earth by comets and asteroids (Kwok, 2011). The study of C60 and its derivatives in circumstel–lar envelopes can help to understand the chemical evolution of galaxies. However, it is unclear how circumstellar C60 is formed and excited. In this paper, we investigate the relations between the emission properties of C60-detected sources, as well as between the observations and existing experimental results. Through a comparison of wavelengths and fluxes of C60 bands, we conclude that the UV-excitation of C60 in cluster state may account for the observations, although other excitation schemes, such as the release of chemical energy, remain possible. We would like to emphasize that the Einstein A-values and flux measurements are too uncertain to allow definite conclusions. We also propose that C60 is one of the products of the dehydrogenation of MAONs through a scenario presented by Micelotta et al. (2012). The fullerene formation thus reflects a transformation from sp3 to sp2 hybridization in MAONs. We have tentatively detected hydrogenated C60 in IRAS 01005+7910 and in Tc 1. The presence of C60 derivatives has the implications that fullerenes have been UV photochemically processed during the post–AGB evolution and the C70/C60 ratio can be significantly modified. A combination of further experiments and observations are required to obtain complete picture of fullerene formation and processes in circumstellar environments.
2Note that the 8.5μm feature in the spectra of NGC 7023 and IRAS 01005+7910 is essentially invisible because of severe blending with the AIB at 8.6μm, and the 7.0 μm feature in Tc1 is contaminated with a [Ar II] line at 7.0μm. This statement also applies to most of the C60 sources, and thus additional complexity can be introduced into the discussion of band strength ratios in the next section.
3The Monte Carlo model of Sellgren et al. (2010) applies to single-photon excitation. If the UV radiation field is sufficiently intense, it is more appropriate to use the statistical equilibrium method (A. Li, private communication).
4The detection of C60 in O-rich sources is rare. It is unclear whether C60 in these sources was produced through a different mechanism with those in C-rich objects.
Adams, G. B., J. B. Page, O. F. Sankey, and M. óKeeffe, Polymerized C60 studied by first–principles molecular dynamics, Phys. Rev. B, 50, 17471–17479, 1994.
Ajie, H. et al., Characterization of the soluble all-carbon molecules C60and C70,J. Phys. Chem., 94, 8630–8633, 1990.
Allamandola, L. J., A. G. G. M. Tielens, and J. R. Barker, Interstellar poly–cyclic aromatic hydrocarbons—The infrared emission bands, the excitation/emission mechanism, and the astrophysical implications, Astrophys. J. Suppl., 71, 733–775, 1989.
Becker, L. and J. L. Bada, Fullerenes in Allende meteorite, Nature, 372, 507–507,1994.
Becker, L., J. L. Bada, R. E. Winans, J. E. Hunt, T. E. Bunch, and B. M. French, Fullerenes in the 1.85-billion–year-old Sudbury impact structure, Science, 265, 642–645, 1994.
Bernard–Salas, J., J. Cami, E. Peeters, A. P. Jones, E. R. Micelotta, and M. A. T. Groenewegen, On the excitation and formation of circumstellar fullerenes, Astrophys. J., 757, 41, 2012.
Berné, O. and A. G. G. M. Tielens, Formation of buckminsterfullerene (C60) in interstellar space, PNAS, 109, 401–406, 2012.
Berné, O., G. Mulas, and C. Joblin, Interstellar C60+, Astron. Astrophys., 550, L4, 2013.
Bertsch, G. F, A. Smith, and K. Yabana, Optical properties of C60 vibrations, Phys. Rev. B, 52, 7876–7878, 1995.
Bochvar, D. A. and E. G. Galpern, Hypothetical systems: carbododecahe–dron, s–icosahedron, and carbo–s–icosahedron, Dokl. Acad. Nauk SSSR, 209, 610–612, 1973 (in Russian).
Boersma, C, R. H. Rubin, and L. J. Allamandola, Spatial analysis of the polycyclic aromatic hydrocarbon features southeast of the Orion Bar, Astrophys. J., 753, 168, 2012.
Buseck, P. R., Geological fullerenes: review and analysis, Earth Planet. Sci. Lett., 203, 781–792, 2002.
Buseck, P. R., S. J. Tsipursky, and R. Hettich, Fullerenes from the geological environment, Science, 257, 215–217, 1992.
Cami, J., J. Bernard–Salas, E. Peeters, and S. E. Malek, Detection of C60 and C60 in a young planetary nebula, Science, 329, 1180–1182, 2010.
Cataldo, F and S. Iglesias–Groth, On the action of UV photons on hydro–genated fulleranes C60H36 and C60D36, Mon. Not. R. Astron. Soc, 400, 291–298, 2009.
Chase, B., N. Herron, and E. Holler, Vibrational spectroscopy of fullerenes (C60 and C70) temperature dependant studies, J. Phys. Chem., 96,4262–4266, 1992.
Cerrigone, L., J. L. Hora, G. Umana, and C. Trigilio, Spitzer Detection of polycyclic aromatic hydrocarbons and silicate features in post–AGB stars and young planetary nebulae, Astrophys. J., 703, 585, 2009.
Choi, C. H., M. Kertesz, and L. Mihaly, Vibrational assignment of all 46 fundamentals of C60 and C606−: Scaled quantum mechanical results performed in redundant internal C60rdinates and compared to experiments, J. Phys. Chem. A, 104, 102–112, 2000.
Clayton, G. C, D. M. Kelly, J. H. Lacy, I. R. Little–Marenin, P. A. Feldman, and P. F Bernath, A mid–infrared search for C60 in R Coronae Borealis stars and IRC+10216, Astron. J., 109, 2096, 1995.
Clayton, G. C. et al., The dust properties of two hot R Coronae Borealis stars and a Wolf-Rayet central star of a planetary nebula: In search of a possible link, Astron. J., 142, 54, 2011.
de Heer, W. A. and D. Ugarte, Carbon onions produced by heat treatment of carbon soot and their relation to the 217.5 nm interstellar absorption feature, Chem. Phys. Lett., 207, 480–486, 1993.
de Vries, M. S. et al., A search for C60 in carbonaceous chondrites, Geochim. Cosmochim. Acta, 57, 933–938, 1993.
Duley, A. A. and A. Hu, Fullerenes and proto–fullerenes in interstellar carbon dust, Astrophys. J., 745, L11, 2012.
Duley, A. A. and D. A. Williams, Excitation of the aromatic infrared emission bands: Chemical energy in hydrogenated amorphous carbon particles?, Astrophys. J., 737, L44, 2011.
Ebbesen, T. W. et al., Origins of fullerenes in rocks, Science, 268, 1634–1635, 1995.
Esfarjani, K., Y. Hashi, J. Onoe, K. Takeuchi, and Y. Kawazoe, Vibrational modes and IR analysis of neutral photopolymerized C60 dimers, Phys. Rev. B, 57, 223–229, 1998.
Evans, A. et al., Solid–phase C60 in the peculiar binary XX Oph?, Mon. Not. R. Astron. Soc, 421, L92, 2012.
Foing, B. H. and P. Ehrenfreund, Detection of two interstellar absorption bands coincident with spectral features of C60+, Nature, 369, 296–298, 1994.
Frum, C. I., R. Engelman, H. G. Hedderich, P. F Bernath, L. D. Lamb, and D. R. Huffman, The infrared emission spectrum of gas-phase C+60 (buckmisterfullerene), Chem. Phys. Lett., 176, 504–508, 1991.
Fu, K.-J. et al., Giant vibrational resonances in A6C60 compounds, Phys. Rev. B, 46, 1937–1940, 1992.
Fulara, J., M. Jakobi, and J. P. Maier, Electronic and infrared spectra ofC+60 andC−60 in neon and argon matrices, Chem. Phys. Lett., 21, 227–234, 1993.
García-Hernández, D. A. and J. J. Díaz-Luis, Diffuse interstellar bands in fullerene planetary nebulae: The fullerenes—diffuse interstellar bands connection, Astron. Astrophys.,550, L6, 2013.
García-Hernández, D. A. et al., Formation of fullerenes in H–containing planetary nebulae, Astrophys. J., 724, L39, 2010.
García-Hernández, D. A. et al., The formation of fullerenes: Clues from new C60, C60, and (possible) planar C60 detections in Magellanic Cloud planetary nebulae, Astrophys. J., 737, L30, 2011a.
García–Hernández, D. A., R. N. Kameswara, and D. L. Lambert, Are C60 molecules detectable in circumstellar shells of R Coronae Borealis stars?, Astrophys. J., 729, 126, 2011b.
García–Hernández, D. A. et al., Infrared study of fullerene planetary nebulae, Astrophys. J., 760, 197, 2012a.
García-Hernández, D. A., R. N. Kameswara, and D. L. Lambert, High-resolution optical spectroscopy of DY Cen: Diffuse interstellar bands in a proto-fullerene circumstellar environment?, Astrophys. J., 759, L21, 2012b.
García-Lario, P., A. Manchado, A. Ulla, and M. Manteiga, Infrared space observatory observations of IRAS 16594–4656: A new Proto-planetary nebula with a strong 21 micron dust feature, Astrophys. J., 513, 941, 1999.
Giannozzi, P. and S. Baroni, Vibrational and dielectric properties of C60 from densityfunctional perturbation theory, J. Chem. Phys., 100, 8537, 1994.
Gielen, C, J. Cami, J. Bouwman, E. Peeters, and M. Min, Carbonaceous molecules in the oxygen–rich circumstellar environment of binary post–AGB stars. C60 fullerenes and polycyclic aromatic hydrocarbons, Astron. Astrophys., 536, 54, 2011.
Goeres, A. and E. Sedlmayr, The envelopes of R Coronae Borealis stars. I—A physical model of the decline events due to dust formation, Astron. Astrophys., 265, 216–236, 1992.
Hara, T., J. Onoe, and K. Takeuchi, In Situ High–Resolution FT–IR Study of the Glass Transition in a C60 Film, J. Phys. Chem.B, 101,9532–9536, 1997.
Herbig, G. H., The search for interstellar C60, Astrophys. J., 542,334–343, 2000.
Heymann, D., Astrophys. J., 489, 111, 1997.
Heymann, D. et al., Fullerenes of possible wildfire origin in Cretaceous–Tertiary boundary sediments, :Geol. Soc. Am. Spec. Pap., 307, 453–464, 1996.
Houck, J. R. et al., The Infrared Spectrograph (IRS) on the Spitzer Space Telescope, Astrophys. J. Suppl., 154, 18–24, 2004.
Howard, J. B., J. T. McKinnon, Y Makarovsky, A. L. Lafleur, and M. E. Johnson, Fullerenes C60 and C70 in flames, Nature, 352,139–141,1991.
Hrivnak, B. J., K. Volk, and S. Kwok, 2–45 micron infrared spectroscopy of carbon–rich proto–planetary nebulae, Astrophys. J., 535,275–292, 2000.
Iglesias–Groth, S., Fullerenes and buckyonions in the interstellar medium, Astrophys. J., 608, L37–L40, 2004.
Iglesias-Groth, S., A. Ruiz, J. Bretón, and J. M. Gómez Llorente, Photoabsorption spectra of icosahedral fullerenes: A semiempirical approach, J. Chem. Phys., 116, 10648, 2002.
Iglesias–Groth, S., F Cataldo, and A. Manchado, Infrared spectroscopy and integrated molar absorptivity of C60 and C70 fullerenes at extreme temperatures, Mon. Not. R. Astron. Soc, 413, 213–222, 2011.
Iglesias-Groth, S., D. A. García–Hernández, F Cataldo, and A. Manchado, Infrared spectroscopy of hydrogenated fullerenes (fulleranes) at extreme temperatures, Mon. Not. R. Astron. Soc, 423, 2868–2878, 2012.
Jäger, C, F Huisken, H. Mutschke, I. L. Jansa, and T. Henning, Formation of Polycyclic aromatic hydrocarbons and carbonaceous solids in gas-phase condensation experiments, Astrophys. J., 696,706–712, 2009.
Justtanont, K., M. J. Barlow, C. J. Skinner, P. F Roche, D. K. Aitken, and C. H. Smith, Mid–infrared spectroscopy of carbon–rich post–AGB objects and detection of the PAH molecule chrysene, Astron. Astrophys., 309, 612–628,1996.
Kern, B., D. Strelnikov, P. Weis, A. Böttcher, and M. M. Kappes, C+60 and C−60; in neon and argon matrices, in 67th International Symposium on Molecular Spectroscopy, 2012.
Kessler, M. F. et al., The Infrared Space Observatory (ISO) mission, Astron. Astrophys., 315, L27–L31, 1996.
Krätschmer, W., The infrared and ultraviolet absorption spectra of laboratory–produced carbon dust: Evidence for the presence of the C60 molecule, K. Fostiropoulos, and D. R. Huffman, Chem. Phys. Lett., 170, 167–170, 1990a.
Krätschmer, W, L. D. Lamb, K. Fostiropoulos, and D. R. Huffman, C60, a new form of carbon, Nature, 347, 354–358, 1990b.
Kroto, H. W., J. R. Heath, S. C. Obrien, R. F. Curl, and R. E. Smalley, C60: buckminsterfullerene, Nature, 318, 162–163, 1985.
Kwok, S., Amorphous organic solids as a component of interstellar dust, Earth Planets and Space, 63, 1021–1026, 2011.
Kwok, S. and Y. Zhang, Mixed aromatic–aliphatic organic nanoparticles as carriers of unidentified infrared emission features, Nature, 479, 80–83, 2011.
Kwok, S., K. Volk, and B. J. Hrivnak, A 21 micron emission feature in four proto–planetary nebulae, Astrophys. J., 345, L51–L54, 1989.
Kwok, S., S. Volk, and B. J. Hrivnak, Chemical evolution of carbonaceous materials in the last stages of stellar evolution, Astron. Astrophys., 350, L35–L38, 1999.
Léger, A., L. D. Hendecourt, L. Verstraete, and W. Schmidt, Remarkable candidates for the carrier of the diffuse interstellar bands—C60(+) and other polyhedral carbon ions, Astron. Astrophys.,203, L145–L148, 1988.
Leidlmair, C. et al., On the possible presence of weakly bound fullerene–H2 complexes in the interstellar medium, Astrophys. J., 738, L4, 2011.
Li, A., J.-H. Chen, M.-P. Li, Q.-J. Shi, and Y.–J. Wang, On buckyonions as an interstellar grain component, Mon. Not. R. Astron. Soc, 390, 39–42, 2008.
Martin, M. C, D. Koller, and L. Mihaly, In situ infrared transmission study of Rb–and K–doped fullerenes, Phys. Rev. B, 47, 14607–14610, 1993.
Micelotta, E. R., A. P. Jones, J. Cami, E. Peeters, J. Bernard–Salas, and G. Fanchini, The formation of cosmic fullerenes from arophatic clusters, Astrophys. J., 761, 35, 2012.
Moutou, C, K. Sellgren, L. Verstraete, and A. Léger, Upper limit on C60 and C60(+) features in the ISO–SWS spectrum of the reflection nebula NGC 7023, Astron. Astrophys., 347, 949–956, 1999.
Nemes, L. et al., Gas–phase infrared emission spectra of C60 and C70. Temperature–dependent studies, Chem. Phys. Lett., 218,295–303,1994. Onoe, J. and K. Takeuchi, In situ high–resolution infrared spectroscopy of a photopolymerized C60 film, Phys. Rev. B, 54, 6167–6171, 1996.
Onoe, J. and K. Takeuchi, In situ high–resolution infrared spectroscopy of a photopolymerized C60 film, Phys. Rev. B, 54, 6167–6171, 1996.
Osawa, E., Superaromaticity, Kagaku, 25, 854–863, 1970 (in Japanese).
Papoular, R., Excitation of unidentified infrared bands by H atom impacts, Mon. Not. R. Astron. Soc, 419, 2396–2406, 2012.
Petrie, S. and D. K. Bohme, Laboratory studies of ion/molecule feactions of fullerenes: Chemical derivatization of fullerenes within dense interstellar clouds and circumstellar shells, Astrophys. J., 540, 869–885, 2000.
Pottasch, S. R., R. Surendiranath, and J. Bernard–Salas, Abundances in planetary nebulae: NGC 1535, NGC 6629, He2-108, and Tc1, Astron. Astrophys., 531, A23, 2011.
Roberts, K. R. G., K. T. Smith, and P. J. Sarre, Detection of C60 in embedded young stellar objects, a Herbig Ae/Be star and an unusual post-asymptotic giant branch star, Mon. Not. R. Astron. Soc., 421, 3277–3285, 2012.
Rubin, R. H. et al., Spitzer reveals what is behind Orion73x0027;s Bar, Mon. Not. R. Astron. Soc., 410, 1320–1348, 2011.
Sassara, A., G. Zerza, M. Chergui, and S. Leach, Absorption wavelengths and bandwidths for interstellar searches of C60 in the 2400–4100 Å region, Astrophys. J. Suppl., 135, 263–273, 2001.
Sellgren, K., M. W. Werner, J. G. Ingalls, J. D. T. Smith, T. M. Carleton, and C. Joblin, C60 in reflection nebulae, Astrophys. J., 722, L54–L57, 2010.
Snow, T. P. and C. G. Seab, A search for interstellar and circumstellar C60, Astron. Astrophys., 213, 291–294, 1989.
Stanton, R. E. and M. D. Newton, Normal vibrational modes of buckmin–sterfullerene, J. Phys. Chem., 92, 2141–2145, 1988.
Webster, A., Unsaturated fulleranes and the minor features of the unidentified infrared emission near 3 microns, Mon. Not. R. Astron. Soc., 257, 463–470, 1992.
Webster, A., A theory of the diffuse interstellar bands, Mon. Not. R. Astron. Soc., 262, 831–838, 1993.
Werner, M. et al., The Spitzer Space Telescope mission, Astrophys. J. Suppl., 154, 1–9, 2004.
Winkler, R., T. Pichler, and H. Kuzmany, Vibrational analysis of IR reflection–transmission from single crystal C60, Phys. Rev. B, 96, 39–45, 1994.
Yagi, H. et al., Photoabsorption cross section of C60 thin films from the visible to vacuum ultraviolet, Carbon, 47, 1152–1157, 2009.
Yasumatsu, H., T. Kondow, H. Kitagawa., K. Tabayashi, and K. Shobatake, Absorption spectrum of C60 in the gas phase: Autoionization via core–excited Ryberg states, J. Chem. Phys., 104, 899, 1996.
Zhang, K., B.–W. Jiang, and A. Li, On the carriers of the 21μm emission feature in post–asymptotic giant branch stars, Mon. Not. R. Astron. Soc., 396, 1247–1256, 2009.
Zhang, Y. and S. Kwok, Detection of C60 in the protoplanetary nebula IRAS 01005+7910, Astrophys. J., 730, 126, 2011.
YZ would like to thank the SOC of the 5th meeting on Cosmic Dust for the invitation to give this talk at Center for Planetary Science, c/o Integrated Research Center of Kobe University. We also thank two anonymous referees for helpful comments. Financial support for this work was provided by the Research Grants Council of the Hong Kong under grants HKU7073/11P.
About this article
Cite this article
Zhang, Y., Kwok, S. On the detections of C60 and derivatives in circumstellar environments. Earth Planet Sp 65, 1 (2013). https://doi.org/10.5047/eps.2013.06.003
- AGB and post-AGB
- circumstellar matter