EDP Sciences
Free Access
Issue
A&A
Volume 559, November 2013
Article Number L8
Number of page(s) 6
Section Letters
DOI https://doi.org/10.1051/0004-6361/201322466
Published online 14 November 2013

© ESO, 2013

1. Introduction

As the eighth most abundant element in the Universe, silicon plays an important role in understanding nucleosynthesis and Galactic chemical evolution (GCE). The main isotope 28Si is mainly produced by early-generation massive stars that become Type II supernovae. The other two stable isotopes 29Si and 30Si are mainly produced by O and Ne burning in massive stars or by slow neutron capture (the s-process) and by explosive burning in the final stages of stellar evolution, that is, the asymptotic giant branch (AGB) phase for low- and intermediate-mass stars and supernova explosions for high-mass stars (see, e.g., Woosley & Weaver 1995; Timmes & Clayton 1996; Alexander & Nittler 1999).

In the thermally pulsing AGB (TP-AGB) phase, thermonuclear runaways are periodically caused by He burning in a thin shell between the H-He discontinuity and the electron-degenerate C-O core. This energy goes directly into heating the local area and raises the pressure, which initiates an expansion and a series of convective and mixing events (Herwig 2005; Iben & Renzini 1983). During the so-called third dredge-up, the products of He burning and the s-process elements are brought to the surface, e.g., 12C, which can lead to the formation of S- (C/O ≈ 1) or C-type (C/O > 1) stars. In conjunction with dredge-ups, the Si-bearing molecules (e.g., SiC and SiO) formed in the stellar surface eventually condense onto dust grains or actually form dust grains. The silicon isotopic ratios will be preserved and go through the journey in the interstellar medium (ISM) until they are used again to form stars.

AGB stars can produce almost all grains of interstellar dusts, and their dust production is one order of magnitude higher than that of supernovae in the Milky Way (see, e.g., Dorschner & Henning 1995; Gehrz 1989). It is generally believed that oxygen-rich M-type stars produce mainly silicate grains and carbon-rich stars mainly carbonaceous grains (Gilman 1969). However, the actual situation may be more complicated and grain composition may change during the AGB phase (Lebzelter et al. 2006).

The measured 29Si/30Si ratios in the ISM are about 1.5 (Wolff 1980; Penzias 1981), very close to that of the solar system (Anders & Grevesse 1989; Asplund et al. 2009). However, near-infrared SiO observations of Tsuji et al. (1994) showed that some evolved stars have 29Si/30Si ratios slightly below 1.5. Our new observations of SiO isotopologues in the radio domain with the APEX and Herschel telescopes confirm the low 29Si/30Si ratios for oxygen-rich M-type stars.

Table 1

Spectral parameters of the observed SiO isotopologue transitions.

2. Observations

Observations of the SiO isotopologue lines toward VY CMa, o Ceti, W Hya, and R Leo were carried out with the 12-m APEX telescope in 2011 September and 2012 December on Llano de Chajnantor in Chile. The single-sideband heterodyne receivers APEX-1 and APEX-2 (Vassilev et al. 2008; Risacher et al. 2006) were used during the observations. The focus and pointing of the antenna were checked on Jupiter and Mars. The pointing and tracking accuracy were about 2′′ and 1′′, respectively. The extended bandwidth Fast Fourier Transform Spectrometer (XFFTS; Klein et al. 2012) backend was mounted and configured into a bandwidth of 2.5 GHz and ~0.1 km s-1 resolution. In addition, the Herschel/HIFI data of VY CMa, o Ceti, W Hya, χ Cyg, R Cas, and R Dor were retrieved from the Herschel science archive.

thumbnail Fig. 1

Left: APEX ground-vibrational 28SiO (black), 29SiO (red), and 30SiO (blue) J = 6−5 spectra toward VY CMa, o Ceti, and W Hya. The intensities of the 29SiO and 30SiO lines were multiplied by two for clarity. Right: Herschel/HIFI ground-vibrational 29SiO (red) and 30SiO (blue) J = 26−25 spectra at around 1.1 THz toward the same sources. The 29SiO emission of VY CMa is blended by the 13CO J = 10−9 line from the other sideband. The dashed lines indicate the VLSR of the sources.

Open with DEXTER

All spectra were converted to the main beam brightness temperature unit, TMB = (ηMB = Beff/Feff), using the forward efficiencies (Feff) and the beam-coupling efficiencies (Beff) from the APEX documentation1. The beam efficiencies of HIFI were taken from the Herschel/HIFI documentation webpage. We adopted ηMB of 0.75, 0.73, and 0.76 for the APEX-1, 2, and HIFI data, respectively. All data were reduced and analyzed by using the standard procedures in the GILDAS2 package. The SiO spectroscopic data were taken from the Cologne database for molecular spectroscopy (CDMS3) and are listed in Table 1.

3. Results and discussion

The APEX and Herschel SiO isotopologue spectra of the selected evolved stars (with both APEX and Herschel detections) are shown in Figs. 1 and 4, and the SiO intensity measurements are summarized in Table 2. The 29SiO and 30SiO emission is expected to be optically thin because the abundance of the main isotopologue 28SiO is at least ten times larger than 29SiO in the ISM (Penzias 1981). Additionally, the solar and terrestrial 29Si/30Si ratios are close to 1.5 (de Bièvre et al. 1984; Anders & Grevesse 1989). The 29SiO/30SiO J = 26−25 intensity ratios observed with the Herschel/HIFI instrument for o Ceti and W Hya (Fig. 1) are consistent with the low-J results obtained with the APEX telescope. Fitting two Gaussian profiles to the 29SiO line and the partially blended 13CO J = 10−9 line in the HIFI VY CMa spectra, we obtained a 29SiO/30SiO J = 26−25 intensity ratio of 1.4 ± 0.1, also consistent with the low-J APEX data. Because the upper-state energies Eup/k of J = 26−25 lines are about 700 K higher than those of J = 6−5, the constant 29SiO/30SiO intensity ratio of low- and high-J transitions indicates optically thin 29SiO and 30SiO emission with similar distributions and excitation conditions. In addition, we believe that the 29SiO and 30SiO emission obtained for our sample stars is unlikely to be dominated by masing effects due to the lack of any narrow spectral features. Therefore, the 29SiO/30SiO intensity ratio directly reflects the abundance ratio between 29Si and 30Si in the circumstellar envelopes of these stars, assuming any differences in chemical fractionation or photodissociation are minor. The derived 29Si/30Si ratios are listed in Table 3.

3.1. Silicon isotope ratios

Since 28Si is mainly produced via the α-process in massive stars, the 28Si in low-mass stars comes from their natal clouds. Additionally, stable isotopes 29Si and 30Si can be formed via slow neutron capture (the s-process) in both low- and high-mass stars. It has been shown by Timmes & Clayton (1996) that 28Si is the primary isotope in the GCE with a roughly constant silicon-to-iron ratio over time, independent of the initial metallicity. On the other hand, neutron-rich isotopes 29Si and 30Si show strong dependence on the composition and initial metallicity.

In Fig. 2, the 29Si/30Si ratios derived from the SiO integrated intensities are plotted against the mass-loss rates for different evolved stars, and they show a tendency to increase with increasing mass-loss rates. The two supergiants VY CMa and NML Cyg and the carbon star IRC+10216 have 29Si/30Si ratios close to the solar value of 1.5. The rest of the samples (see also Table 3) have 29Si/30Si ratios < 1.5, for example, the 29Si/30Si ≈ 1 for W Hya. There are two possibilities for the different 29Si/30Si ratios seen in our sample. One is that the silicon isotope ratios merely reflect the initial chemical composition of the environment where these stars were born and the different ratios are the results of different ages, which mainly depend on their masses and metallicities. The other possibility is that the stellar evolution can significantly change the silicon isotope ratios.

thumbnail Fig. 2

Comparison of 29Si/30Si in evolved stars. The dashed line indicates the terrestrial and solar 29Si/30Si abundance ratio of 1.51 (de Bièvre et al. 1984; Anders & Grevesse 1989; Asplund et al. 2009). The red line is a linear fit to the 29Si/30Si– relation.

Open with DEXTER

The first possibility implies that the 29Si/30Si ratio in the ISM has not significantly changed in the past 4.6 Gyr when the Sun was born. In comparison, VY CMa and IRC+10216 were born ~107 and 1−5 × 108 years ago, assuming masses of 25−32 and 3−5 M, respectively (see Portinari et al. 1998). We found stars that we believe to be significantly older than the Sun, such as W Hya (based on the mass-loss rate, initial mass, and current age), to have lower 29Si/30Si ratios. For instance, with an initial mass of 1−1.2 M, W Hya has an age of 5−10 Gyr. In either the lower or higher age limit, this suggests a significant change in the 29Si/30Si ratio between the pre- and post-solar period: the 29Si/30Si ratios in the ISM increase from about 1 to 1.5 between 5 to 10 Gyr ago and remain roughly constant after the Sun was born. Given the time it takes low-mass stars to evolve onto the AGBs, it is unlikely that many low-mass AGB stars existed in our Galaxy between 5 and 10 Gyr ago, even if they had been formed at the beginning of the Milky Way formation. It is therefore also unlikely that low-mass AGB stars were significant contributors to the GCE in the presolar era. The 29Si/30Si ratio in the presolar era may be due to supernovae and/or other massive evolved stars. We note that the stars in our sample only trace the 29Si/30Si ratio of their natal clouds if they do not modify this ratio via nucleosynthesis (see, e.g., Zinner et al. 2006).

The second possibility for different 29Si/30Si ratios is that the stars in the AGB phase can significantly modify these ratios. Some of the M-type stars will become C-type stars after several dredge-up episodes with higher mass-loss rates toward the end of the AGB phase (see, e.g., Herwig 2005). If the 29Si/30Si ratio can be modified by the s-process in the He-burning shell in evolved stars, it must be done efficiently because the AGB time scale is short (a few times 106 yr, see Marigo & Girardi 2007). However, the modeling results of Zinner et al. (2006) show that the 29Si/30Si ratios of low-mass stars do not significantly change during the AGB phase (see also the discussion of Decin et al. 2010).

3.2. 29Si/ 30Si ratio in presolar grains

thumbnail Fig. 3

Silicon three-isotope plot for presolar grains. The delta notation is defined as δiSi/28Si = [(iSi/28Si)/(iSi/28Si) − 1] × 1000. The δ29Si/δ30Si ratios do not translate directly into the 29Si/30Si ratios, which are plotted as black dashed lines. Left: the data of different subgroups (X, Y, and Z) of SiC grains were taken from Lin et al. (2002), Amari et al. (2001), and Hoppe et al. (1997), and the Orgueil silicate grains from Zinner & Jadhav (2013). The mainstream type (~93%) of grains are indicated as a solid red line with a slope of 1.37. The dashed red line indicates the possible extension from the mainstream grains to X2 grains. The filled triangle indicates the ISM value (Wolff 1980), and the black arrow indicates the direction of the Galactic chemical evolution. The orange crosses are the evolved star sample (about 3 M) from Tsuji et al. (1994). Right: similar plot as the left one, but on a smaller scale, and the different silicate grain data were taken from Nguyen et al. (2007), Mostefaoui & Hoppe (2004), and Nagashima et al. (2004).

Open with DEXTER

Assuming the 29Si/30Si ratio in the gas-phase SiO is the same as it condenses onto dust grains or forms silicates, this primitive 29Si/30Si ratio may be carried by those grains when they are incorporated into new stellar and planetary systems. The 29Si/30Si ratio in presolar SiC grains has been studied in some meteorites (e.g., the Murchison meteorite, see the review by Zinner 1998). They have been categorized into different types (e.g., X, Y, and Z) according to their silicon isotopic anomalies. Most of the SiC grains found in meteorites are the so-called mainstream grains (~93%, see Fig. 3), i.e., those with a slope of 1.34 on a silicon three-isotope plot (Hoppe et al. 1994). On the other hand, SiO is expected to condense onto the dust formation regions near O-rich stars, or via a possible heteromolecular nucleation of Mg, SiO, and H2O to form silicates (Goumans & Bromley 2012). In the studies of Si isotopes in primitive silicate grains, the Si isotopic compositions of the majority of presolar silicates are similar to the SiC mainstream grains (Nguyen et al. 2007; Mostefaoui & Hoppe 2004; Nagashima et al. 2004; Vollmer et al. 2008), indicating that the amount of Si isotopes locked in the SiC grains and the SiO group in silicates may be similar; an example are the Orgueil silicate grains shown in Fig. 3.

Most of the presolar grains have 29Si/30Si ratios around 1.5, but evidence of lower 29Si/30Si ratios are also found in presolar SiC grains, for instance, types X2 and Z in Fig. 3. The type Z grains may have originated from a nearby evolved star (see also Zinner et al. 2006). Additionally, the type X grains have been proposed to have a supernova origin (e.g., Amari et al. 1992; Hoppe et al. 1994), and have two or more subgroups (see, e.g., Hoppe et al. 1995; Lin et al. 2002) with possible different stellar origins. According to the study of Lin et al. (2002) on the Qingzhen enstatite chondrite, the subgroups X1 and X2 show somewhat similar N and O isotopes abundance ratios, but have different slopes on the Si three-isotope plot (0.7 vs. 1.3 for X1 and X2, respectively). Because the metallicity in the local ISM is increasing owing to the GCE, δ29Si and δ30Si will increase with time accordingly. It is possible that the lower 29Si/30Si ratios seen in X2 grains may have originated from a population of evolved stars (such as the evolved stars with lower 29Si/30Si ratios). On the other hand, X1 grains with higher 29Si/30Si ratios are likely to be attributable to Type II supernovae (see also Zinner & Jadhav 2013). Moreover, it is important to point out that the higher-mass (about 3 M) evolved star sample from Tsuji et al. (1994) can be well explained by the GCE (Fig. 3), considering the possible uncertainty in the 29Si/30Si ratio estimate for the present-day ISM.

4. Conclusions

We investigated the 29Si/30Si ratios of 15 evolved stars from the thermal SiO isotopologue emission obtained by the APEX and Herschel telescopes and from the literature. The inferred 29Si/30Si ratios tend to be lower among the older low-mass O-rich stars. Because the 29Si/30Si ratios are not significantly modified during the AGB phase and the contributions from the low-mass AGB stars are less important due to their long lifetimes, the lower 29Si/30Si ratios imply different enrichment of 29Si and 30Si in the Galaxy between 5 to 10 Gyr ago with a nearly constant value of 1.5 after that. Noting that presolar grains may also have 29Si/30Si ratios lower than 1.5 (i.e., Type X2 and Z), we suggest that these grains could have been produced by one or more AGB stars with masses high enough to evolve onto the AGB in time to contribute to presolar grains.


Acknowledgments

We thank the Swedish APEX staff for preparing observations and the referee for helpful comments. M.G.R. gratefully acknowledges support from the National Radio Astronomy Observatory (NRAO). The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. I.d.G. acknowledges the Spanish MINECO grant AYA2011-30228-C03-01 (co-funded with FEDER fund).

References

Online material

thumbnail Fig. 4

Upper panels: APEX ground-vibrational 29SiO (red) and 30SiO (blue) J = 7−6 spectra toward o Ceti and R Leo. Lower panels: Herschel/HIFI ground-vibrational 29SiO (red) and 30SiO (blue) J = 26−25 spectra at around 1.1 THz toward χ Cyg, R Cas, and R Dor. The dashed lines indicate the VLSR of the sources.

Open with DEXTER

Table 2

APEX and Herschel SiO integrated intensity measurements.

Table 3

Overview of envelope terminal velocities, mass-loss rates, and 29Si/30Si ratios toward the selected evolved stars.

All Tables

Table 1

Spectral parameters of the observed SiO isotopologue transitions.

Table 2

APEX and Herschel SiO integrated intensity measurements.

Table 3

Overview of envelope terminal velocities, mass-loss rates, and 29Si/30Si ratios toward the selected evolved stars.

All Figures

thumbnail Fig. 1

Left: APEX ground-vibrational 28SiO (black), 29SiO (red), and 30SiO (blue) J = 6−5 spectra toward VY CMa, o Ceti, and W Hya. The intensities of the 29SiO and 30SiO lines were multiplied by two for clarity. Right: Herschel/HIFI ground-vibrational 29SiO (red) and 30SiO (blue) J = 26−25 spectra at around 1.1 THz toward the same sources. The 29SiO emission of VY CMa is blended by the 13CO J = 10−9 line from the other sideband. The dashed lines indicate the VLSR of the sources.

Open with DEXTER
In the text
thumbnail Fig. 2

Comparison of 29Si/30Si in evolved stars. The dashed line indicates the terrestrial and solar 29Si/30Si abundance ratio of 1.51 (de Bièvre et al. 1984; Anders & Grevesse 1989; Asplund et al. 2009). The red line is a linear fit to the 29Si/30Si– relation.

Open with DEXTER
In the text
thumbnail Fig. 3

Silicon three-isotope plot for presolar grains. The delta notation is defined as δiSi/28Si = [(iSi/28Si)/(iSi/28Si) − 1] × 1000. The δ29Si/δ30Si ratios do not translate directly into the 29Si/30Si ratios, which are plotted as black dashed lines. Left: the data of different subgroups (X, Y, and Z) of SiC grains were taken from Lin et al. (2002), Amari et al. (2001), and Hoppe et al. (1997), and the Orgueil silicate grains from Zinner & Jadhav (2013). The mainstream type (~93%) of grains are indicated as a solid red line with a slope of 1.37. The dashed red line indicates the possible extension from the mainstream grains to X2 grains. The filled triangle indicates the ISM value (Wolff 1980), and the black arrow indicates the direction of the Galactic chemical evolution. The orange crosses are the evolved star sample (about 3 M) from Tsuji et al. (1994). Right: similar plot as the left one, but on a smaller scale, and the different silicate grain data were taken from Nguyen et al. (2007), Mostefaoui & Hoppe (2004), and Nagashima et al. (2004).

Open with DEXTER
In the text
thumbnail Fig. 4

Upper panels: APEX ground-vibrational 29SiO (red) and 30SiO (blue) J = 7−6 spectra toward o Ceti and R Leo. Lower panels: Herschel/HIFI ground-vibrational 29SiO (red) and 30SiO (blue) J = 26−25 spectra at around 1.1 THz toward χ Cyg, R Cas, and R Dor. The dashed lines indicate the VLSR of the sources.

Open with DEXTER
In the text

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.