EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 624 (April 2019) A&A, 624 (2019) A1 Full HTML
Free Access
Issue
A&A
Volume 624, April 2019
Article Number A1
Number of page(s) 26
Section Stellar atmospheres
DOI https://doi.org/10.1051/0004-6361/201834466
Published online 01 April 2019

© ESO 2019

1 Introduction

So far, only a small number of planetary nebulae (PNe) have been identified to host a binary with a giant or subgiant component dominating the optical wavelength range and showing peculiar surface element abundances that indicate late stage stellar evolution nuclear synthesis. These stars exhibit signatures of slow neutron-capture nucleosynthesis (s-process) in their spectra and in some cases an enrichment in carbon (C).

For the object of this work, the central star (CS) of the PN Hen 2−39 (PN G283.8−04.2, Wray 16−64; Henize 1967; Acker et al. 1992; Wray 1966), Miszalski et al. (2013a) determined an overabundance for the s-process element barium (Ba) of [Ba∕Fe]1 = 1.5 ± 0.25 and an enrichment of [C∕H] = 0.42 ± 0.02 in a spectral analysis based on mid-resolution spectra obtained with the Southern African Telescope (SALT; Buckley et al. 2006) with the Robert Stobie Spectrograph (RSS; Burgh et al. 2003; Kobulnicky et al. 2003). These findings confirm the membership of the K-type nucleus of Hen 2−39 in the small group of Ba central stars of planetary nebulae (CSPNe) along with LoTr 5, WeBo 1, and Abell 70 (Thevenin & Jasniewicz 1997; Bond et al. 2003; Miszalski et al. 2012; Tyndall et al. 2013; Aller et al. 2018).

Ba CSPNe are prime examples of progenitors of Ba stars that were described by Bidelman & Keenan (1951). Because of their evolutionary status, namely still being on the main sequence or a red giant, these stars did not yet experience AGB nucleosynthesis and, thus, cannot have synthesized heavy elements. McClure et al. (1980) discovered the binary nature of Ba stars and proposed that mass transfer was key to explain these sources. Boffin & Jorissen (1988) performed detailed simulations of wind mass transfer to explain the pollution of the Ba star from an evolved companion with the products of asymptotic giant branch (AGB) nucleosynthesis that are dredged up to the stellar surface (Herwig 2005; Werner & Herwig 2006). More recently, other mechanisms were proposed in which the material is transferred to the still unevolved companion (Boffin 2015) via Roche-lobe overflow (RLOF; e.g. Han et al. 1995) or wind-RLOF (e.g. Nagae et al. 2004; Mohamed & Podsiadlowski 2007; Abate et al. 2013).

This scenario is strongly supported by the fact that, so far, all Ba stars are found in binaries with white dwarf (WD) companions (McClure et al. 1980; McClure 1983; Jorissen & Mayor 1988; McClure & Woodsworth 1990; Jorissen et al. 1998), which is also definitely clear for the Ba CSPNe since the Ba star is not hot enough to ionize the ambient ejected material that is visible as the surrounding PN. Although the (pre-)WD companion must be there without any doubt, it can be difficult to detect against the bright companion even in the UV. Recently, more and more WD companions of Ba stars (Gray et al. 2011) and pre-WD companions of Ba CSPNe have been detected (e.g. Abell 70; Miszalski et al. 2012), which doubtlessly confirms the formation scenario. The still poorly understood mechanism of mass transfer in these systems is subject of ongoing research. The challenge is to determine their orbital parameters, such as eccentricity and period and to reproduce these with theoretical binary evolution models (Saladino et al. 2018; De Marco 2009).

These stars are expected to show orbital periods of several hundred days, which are typical values for Ba stars (Jorissen et al. 1998). However, there is the CSPN binary in the Necklace Nebula (PN G054.6−03.4; Corradi et al. 2011)standing out toward shorter periods. Miszalski et al. (2013b) found a period of 1.16 d for the post-CE system from the analysis of the C-dwarf secondary. On the other side of the period range, current analyses also indicate that there are systems with values up to several years and with larger eccentricities (Jones et al. 2017).

Ba CSPNe are ideal to study AGB nucleosynthesis. They provide a snapshot of an evolutionary stage with ideal conditions for analyzing not only the polluted cool (sub)giant star but also the ejected material of the nebula around the polluting post-AGB star (e.g. Madonna et al. 2017, 2018). The short duration of the PN phase (≈ 104 yr) guarantees that the mass transfer happened recently and that the companion has not yet had time to adjust. Also, in some cases, the polluted star is still unevolved and did not experience the first dredge-up (DU) that would affectthe surface element composition including the nucleosynthesis outcomes from the polluting post-AGB star.

By comparing the results of our comprehensive spectral analysis to theoretical AGB nucleosynthesis models (Karakas & Lugaro 2016; Karakas et al. 2018), new insights into Ba stars and PNe are gained. It is worth mentioning the s-process mechanisms including atomic reaction rates, the source of neutrons and neutron exposure, internal stellar structures, and mixing processes occurring in a thermal-pulsing AGB star. This allows us to constrain the progenitor mass of the post-AGB star and the number of thermal pulses (TPs) on the AGB. Including binary evolution models (Saladino et al. 2018; De Marco 2009), these objects offer the opportunity to study the CE process and (wind-) RLOF, which are still far from being understood (Miszalski et al. 2013b; Jones & Boffin 2017), and in addition the fraction of mass transferred (Boffin & Jorissen 1988) and, following from this, the dilution factor in the Ba star itself and, thus, the mixing processes at work in (sub)giant stars (Husti et al. 2009). In particular, Ba CSPNe such as Hen 2−39 offer the possibility to detect technetium (Tc), which is the lightest element with no stable isotopes, in their atmospheres. This element was first detected by Merrill (1952) in the atmospheres of red giants, which proved that it is synthesized in evolved stars, since the half-life of 99Tc of 210 000 yr2 is much shorter than the previous giant evolutionary phase. It is thus only observed in AGB stars currently undergoing thermal pulses (TPs; Van Eck & Jorissen 1999; Lebzelter & Hron 2003) and, hence, the determination of the Tc surface abundance of the Ba CSPN indicates the mass-transfer link between the binary components in the PN and establishes a definite indicator for the existence of the third dredge-up (TDU). Assuming a typical post-AGB age of some 103 − 104 yr (Miller Bertolami 2016) for the primary component and taking into account that the dynamical process of mass transfer is short compared to this number (Iben & Livio 1993; Chen et al. 2017), a large fraction of the transferred Tc should still be present in the stellar atmosphere.

We describe the observations, stellar atmosphere models, and analysis techniques in Sects. 2 and 3, respectively. The spectral analysis follows in Sects. 4 and 5. The results are discussed in Sect. 6. We summarize and conclude in Sect. 7.

2 Observations

The spectral analysis of Hen 2−39 is based on spectra in the optical wavelength range obtained with the Ultraviolet and Visual Echelle Spectrograph (UVES; Dekker et al. 2000)at the Very Large Telescope (VLT) at the Paranal Observatory of the European Southern Observatory under ESO program 093.D−0332(A). The data products created from this data were retrieved from the ESO Science Archive Facility. The observation log, including the signal-to-noise ratio (S/N), of the spectra used in this paper is shown in Table A.1. All spectra were taken with a resolving power of R = 42 000−44 000.

We used the spectral analysis code ISpec (Blanco-Cuaresma et al. 2014) to determine the radial velocities for each single observation via cross-correlation with a model template spectrum created using the fundamental parameters Teff, log g, C abundance, and metallicity determined by Miszalski et al. (2013a). The heliocentric corrected radial velocities for the 18 observations are given in Table 1. To improve the S/N, all observations were shifted to the rest-frame velocity and subsequently co-added. To simulate the resolution of the instrument, all synthetic spectra shown in this work were convolved with Gaussians (full width half maximum (FWHM) = 0.12Å).

3 Model atmospheres, atomic data, and analysis techniques

We used the stellar synthesis code SPECTRUM3 (Gray & Corbally 1994, version 2.76) to calculate synthetic spectra for the analysis of the observed high-resolution spectra with the ATLAS9 model atmosphere grids4 (Kurucz 1991; Castelli & Kurucz 2003) as input. These one-dimensional models are based upon the solar abundances from Grevesse & Sauval (1998) and are calculated under the presumption of plane-parallel geometry and local thermodynamic equilibrium (LTE), which is valid for stars in this temperature and gravity regime (Hubeny et al. 2003). For the wavelength values and oscillator strengths of the lines selected in our analysis, we used the values provided within the distribution of SPECTRUM. Data for Tc I were retrieved from the Atomic Spectra Database5 of the National Institute of Standards and Technology (NIST). For Tc II, we used the data provided by Palmeri et al. (2007). We calculated an extensive grid of synthetic spectra spanning from Teff = 3500 to 6000 K (ΔTeff = 250 K between 4000 and 5000 K and ΔTeff = 500 K otherwise) and from log g (cm-1 s-2) = 0.0 to 4.0 (Δlog g = 0.5) with a metallicity of [M∕H] = −0.3 around the literature values of Teff = 4250 ± 150 K and log g = 2.0 ± 0.5 (Miszalski et al. 2013a). For the determination of abundances, we relied on the model with Teff = 4250 K and log g = 2.5 and varied the abundance of one single element over a range of at least 2.5 dex with a step of 0.5 dex. The exceptions to this are C, for which we varied the abundance in steps of 0.05 dex over a range of 0.25 dex, and N with a range of 1.5 dex and steps of 0.3 dex. Since the spectrum is crowded with absorption lines that are broadened by rotation, we could not measure equivalent widths to determine the fundamental parameters. We performed the analysis of the different parameters by selecting wavelength regions that show a strong influence of these particular species. The final values were then derived using a χ2 -method applied to the synthetic spectra grid for the selected regions.

Table 1

Observation time and heliocentric radial velocities for the 18 observations of the Ba CSPN of Hen 2−39.

4 Stellar parameters

4.1 Rotation

To determine vrot sin i = 38 ± 5 km s-1, we used a fit of a synthetic spectrum calculated with the literature values given by Miszalski et al. (2013a) and convolved with rotational profiles for values from vrot sin i = 0 to 50 km s−1 to two regions spanning from 6440 to 6515Å and from 7030 to 7070Å dominated by strong C2 molecular absorption bands (Fig. 1).

4.2 Effective temperature and surface gravity

The many rotationally broadened lines in the observation also hamper the determination of equivalent widths. Thus, we used a set of diagnostic lines of Fe I, Fe II, Ti I, Ti II, Sc I, Sc II, and Mg I and performed a χ2 -fit for selected wavelength regions. The set is composed of the lines that Tabernero et al. (2018) used for a similar spectral analysis. Furthermore, we included some Fe I and Fe II lines that are used by the Brussels Automatic Code for Characterizing High accuracy Spectra (BACCHUS6; Masseron et al. 2016). Finally, we added lines for Fe I, Fe II, Ti I, Ti II, Sc I, and Sc II for which we obtained the largest theoretical equivalent widths in the calculation of the synthetic spectra and that did not saturate. The resulting collection of wavelength regions used for the analysis and the list of diagnostic lines is shown in Table A.2. They cover a wide range of different excitation potentials and oscillator strengths. After a first determination of Teff and log g, we found a degeneracy in log g due to the fact that the strength of the computed lines in the regime around the literature values varies very little as a function of log g for a fixed value of Teff. A spectroscopic determination of log g is hampered by uncertain values for the distance and brightness (Sect. 6.3). Thus, we adopt a value of log g = 2.5 ± 0.5 that is typical for Ba giants of that type (e.g. de Castro et al. 2016). This approach seems to be reasonable because a change of Δlog g = 0.5 only marginally affects the derived abundances compared to the significant statistical errors. We derive Teff = (4350 ± 150) K. Figures A.1 and A.2 illustrate the spectroscopic determination of these parameters by showing the difference due to a variation of Teff and log g.

thumbnail Fig. 1

Synthetic spectra (red) convolved with a rotational profile of vrot = 38 km s-1 around strong C2 and CN molecular absorption bands compared with observations (gray).

5 Element abundances

We used model atmospheres with Teff = 4250 K and log g = 2.5 from the grid and performed line-profile fits for the following elements to determine their abundances. The results are given in Table 2. In our analysis, we assumed the atomic data to be correct and did not propagate uncertainties on atomic data. The continuum placement uncertainty is also assumed to be small as thecontinuum placement is shifted during the fitting procedure. To estimate the impact of a varied Teff and log g on the determined abundances, we redid part of our analysis with models with varied Teff between 4000 and 4500 K at log g varied between 2.0 and 3.0. We also varied the microturbulence velocity by ± 2.0 km s−1, which was kept fixed at 2.0 km s−1 in the initialanalysis. By far, the impact of a change in temperature is the largest. Compared to this error, the variation in log g and microturbulence velocity become negligible. Furthermore, we investigated the influence of the metallicity of the model atmosphere grid that was chosen for the analysis on the determined abundances and repeated part of the analysis with different input model metallicities between − 0.5 and 0. Raising (lowering) the metallicity by a certain amount results in a Fe abundance that is lower (higher) by roughly 1.3-fold that amount. Consistency is reached for the grid with [M∕H] = −0.3 that gives a Fe abundance of [Fe∕H] = 0.3 ± 1.0. In our analysis, we find a C enrichment but cannot determine the O abundance and, thus, use the solar value. The resulting C/O ratio is larger than one. To test whether it is justified to use an O-rich model atmosphere grid (model 1) with solar abundances, we employed the ATLAS9 code and calculated a test model with a C abundance increased to the value found in our analysis (model 2). Furthermore, we calculated a second test model with the increased C abundance and an O abundance reduced by one dex (model 3). We redid the abundance analysis for C and Ba and found the same abundance for C from model 1 and 2. For the model with increased C and decreased O abundance, we get a C abundance that is 0.03 dex higher. Compared to model 1, the Ba abundance for model 2 is higher by 0.04 dex and by 0.2 dex for model 3. The difference between the O-rich model 1 and the C-rich model 2 is shown in Fig. A.3. The effect of the model atmosphere on the abundances is in a range that justifies relying on the available model atmosphere grid. Thus, we did not compute a C-rich grid. However, this adds another uncertainty to the abundances. We obtain large errors arising from the crowdedness of the observed spectrum that we estimate by detailed line profile fits and evaluation based on the χ-by-eye method. In many cases, this is the main contributor to the total error. It ranges between 0.05 dex for C and about 0.4 dex for the light metals up to about 1 dex for the iron group and trans-iron elements. To take the uncertainties in Teff and log g into account, we did this procedure for the corners of the grid stated above (Teff = 4000 K, log g = 3.0 and 4500 K, 2.0). The abundance errors arising from this effect range between 0.03 dex for C and about 0.5 dex for the other metals. The total errors given in Table 2 are the maximum differences for the abundances that are possible within the error limits of this grid.

Table 2

Element abundances determined for Hen 2−39 in log ɛ = 12 + log (nX∕nH), [X ∕H] = log (nX∕nH) − log (nX,⊙∕nH,⊙), and [X ∕Fe] = log (nX∕nFe) − log (nX,⊙∕nFe,⊙) with the number fraction nX for element X.

thumbnail Fig. 2

Observation (gray) of Hen 2−39 compared to model spectra with Teff = 4250 K and log g = 2.5 for a selectedregion of strong C2 absorption for [C∕H] = 0.46, 0.36, 0.26 (green, red, and blue, respectively).

Carbon

We analyzed the C abundance using spectrum synthesis calculations for the region of strong C2 absorption from 4650 to 4737 Å (Fig. 2). We confirm the C enhancement and our result of [C ∕H] = 0.36 ± 0.08 agrees within 1σ with the value derived by Miszalski et al. (2013a) from their mid-resolution spectra.

Nitrogen

Using the C abundances, we derived the N abundance from synthetic calculations for the wavelength regions 7030−7070 Å and 7900−8100 Å affected by strong CN absorption bands. Figure 3 shows the best result. We find N to be enriched to the same level as C with [N∕H] = 0.3 ± 0.8. We could not identify any line of oxygen in the observed spectrum and, thus, were unable to fix an abundance value for O. In our analysis, we adopt the solar value. To get an idea of the 12C/13C ratio, we analyzed the CN absorption band in the region 8100−8200 Å and included the line list for 13C14N from Sneden et al. (2014; Fig. 4). From the inspection of the observation, we cannot claim to find a enhancement in 13C resulting in 12C/13C lower than the solar value of 90, although this cannot be ruled out.

thumbnail Fig. 3

Like Fig. 2, for strong CN absorption for [N∕H] = 1.1, 0.3, −0.5 (green, red, and blue, respectively).
thumbnail Fig. 4

Like Fig. 2, for strong CN absorption for 12C∕13C = 90, 10, 5 (red, green, and blue, respectively). 13CN absorption lines are indicated at the bottom in cyan.

Light metals: sodium to potassium

For the following elements, we used the most prominent absorption features in the synthetic spectra, which show the largest impact of a change of the abundance (Table A.3). The Na doublet Na II λλ 5890.8, 5896.5Å (Fig. A.4) is used to find a Na abundance of [Na∕Fe] = 0.1 ± 0.7, which resembles well with the sample of Ba stars analyzed by de Castro et al. (2016, Fig. 6). From a fit to the regions affected by absorption due to Al we derive [Al ∕Fe] = −0.3 ± 1.0. Ba stars typically show a slight enhancement in Al, which is also present in the sample of de Castro et al. (2016). For S we derive [S ∕Fe] = 1.2 ± 1.2 and for K we found [K∕Fe] = 0.1 ± 1.0.

Iron-peak elements: calcium to copper

Selected absorption features due to neutral Ca to Cu are shown in Figs. A.5 and A.6. We performed a fit for a set of Fe I and Fe II absorption lines and found this star to be metal-poor with [Fe ∕H] = −0.3 ± 1.0. For all iron-peak elements prior to Fe, we found solar values or slight underabundances, namely [Ca ∕Fe] = −0.1 ± 1.0, [Ti ∕Fe] = −0.6 ± 1.4, [V ∕Fe] = −0.6 ± 1.0, [Cr ∕Fe] = −0.2 ± 1.3, and [Mn ∕Fe] = 0.0 ± 1.0. For Sc, we determined an upper limit of [Sc∕Fe] = −0.4. These low values for the α elements Caand Ti do not correspond with the trend in the sample of de Castro et al. (2016) who found an enrichment of these elements compared to Fe with decreasing metallicity. The values close to solar for the other elements, however, are in good agreement with the sample showing a clustering of the abundances around [X ∕Fe] = 0.0. For the elements of this group subsequent to Fe, we determined overabundances compared to Fe of [Co ∕Fe] = 0.5 ± 1.0, [Ni ∕Fe] = 0.6 ± 1.3, and [Cu ∕Fe] = 1.2 ± 1.5.

Trans-iron elements: zinc to osmium

Selected wavelength regions that are among those showing the largest impact of a change in abundance for the elements of this group are shown in Figs. A.7–A.9. Since we found only few significant absorption features for the majority of these elements, the statistical errors of the determined values are large (often > 1 dex) and in many cases only upper abundance limits could be established. We obtained [Zn ∕Fe] < 1.6 and [Rb ∕Fe] = 1.4 ± 1.3. For the elements around the first peak of the s-process, we found [Sr∕Fe] = 1.0 ± 1.5, [Y ∕Fe] = 0.4 ± 1.5, and [Zr ∕Fe] = 0.2 ± 1.5. Despite the large uncertainties, we find a good agreement with the Sr abundances of the sample of Karinkuzhi et al. (2018, Fig. 6). The Sr abundances for most of the Ba stars of Merle et al. (2016) are slightly lower but some reach up to 1 dex as well. Also the Y abundance of Hen 2−39 agrees well with the sample of Merle et al. (2016), whereas the majority of Ba giants of Karinkuzhi et al. (2018) and de Castro et al. (2016) crowd around higher values of Y. The [Zr ∕Fe] abundances of Karinkuzhi et al. (2018) are significantly higher (all >1 dex), whereas some of the Ba stars of Merle et al. (2016) agree with low values of [Zr ∕Fe]. Also the sample of de Castro et al. (2016) clusters around [Zr∕Fe] ≈ 1 but also shows stars with comparatively low values like that for Hen 2−39. For the elements subsequent to this peak, we determined [Nb∕Fe] < 1.0, [Mo ∕Fe] = 1.4 ± 1.3, and [Ru ∕Fe] < 2.1.

One key element of this analysis is the radioactive Tc. The strongest absorption features that appear in the synthetic spectra are Tc I λλ 4031.6, 4095.7, 4238.2, 4262.3, 4297.1Å. Unfortunately, these lines could not be clearly identified in the observed spectrum but it could be used to establish an upper abundance limit of log ɛTc < 2.57.

For the determination of the Ba abundance, we used Ba II λλ 4554.0, 4931.1, 5853.7, 6141.7, 6496.9Å. The first two are very strong and sensitive to small variations of the abundance (Fig. A.8). This helped to constrain [Ba ∕Fe] = 1.8 ± 0.5, which agrees within the error limits with the previous value of Miszalski et al. (2013a). The values determined by Karinkuzhi et al. (2018) for this element range from 0.81 ≤ [Ba∕Fe] ≤ 2.67. Also the sample of Merle et al. (2016) shows a scatter between almost solar and 2.5. Our strong enrichment found for Hen 2-39, thus, is not exceptional (Fig. 6).

La II absorption lines yield [La∕Fe] = 1.5 ± 1.6. For the other elements of the second peak of the s-process we found [Ce ∕Fe] < 2.3, [Pr ∕Fe] < 2.7, and [Nd ∕Fe] = 0.8 ± 1.5. The upper limit for Ce lies above the value range of 1.02 ≤ [Ce∕Fe] ≤ 1.76 determined by Karinkuzhi et al. (2018) and also above that of Merle et al. (2016) ranging from solar to 1.5 dex. The Ba stars of de Castro et al. (2016) cluster around an enrichment of 1.0 dex and none of these stars reach values above 2.5 dex. Our upper limit for Pr lies above the value range of 1.18≤ [Pr∕Fe] ≤ 2.55 of Merle et al. (2016). However, our value for the Nd abundance is below their range of values. The star with the lowest Nd abundance shows [Nd ∕Fe] = 1.18, whereas the sample of de Castro et al. (2016) clustering around [Nd∕Fe] ≈ 1.0 perfectly agrees with our value within the error limits.

For the further rare-earth elements Sm, Eu, Gd, Tb, Dy, and Er, we could only establish upper abundance limits (Table 2). These elements are not analyzed by Merle et al. (2016) and de Castro et al. (2016). Karinkuzhi et al. (2018) found ranges of 1.02≤ [Sm∕Fe] ≤ 2.17, 0.96≤ [Eu∕Fe] ≤ 1.43, and 1.60≤ [Dy∕Fe] ≤ 2.51. Our upper limits of [Sm∕Fe] < 1.1, and [Eu∕Fe] < 1.0 lie within these ranges; [Dy∕Fe] < 3.8 is significantly higher. Furthermore, we could determine the abundances of Hf, W, and Os to be [Hf ∕Fe] < 1.4, [W ∕Fe] = 1.1 ± 1.5, and [Os ∕Fe] < 1.8.

6 Discussion

6.1 Element abundances

We compared our results with the yields from nucleosynthesis calculations of Karakas & Lugaro (2016) for a metallicity of Z = 0.007, in line with the low metallicity of Z = 0.006 that we determined from Z = 10[Fe∕H]Z⊙ with [Fe∕H] = −0.3 and Z⊙ = 0.0134 (Asplund et al. 2009). From these models and those of Karakas et al. (2018), it becomes obvious that AGB nucleosynthesis does not affect the abundances of the iron peak elements and, thus, it seems reasonable to assume the same low metallicity for both components of the binary.

The fact that we cannot see a 13C enhancement agrees very well with the theoretical calculations predicting even an enhancement of the initial solar 12C/13C-ratio for models with initial masses 1.5 M⊙ ≤ Mini ≤ 4.0M⊙ where our estimated initial mass (Sect. 6.2) lies within.

The finding that the iron-peak elements prior to Fe show underabundances and those subsequent to Fe are enhanced leads to the speculation that this pattern may be caused by neutron capture on the former elements as seed species and the formation of elements heavier than Fe. Figure 5 also shows an enrichment due to AGB nucleosynthesis for the elements subsequent to Fe.

The observed N enhancement of [N∕Fe] = 0.7 ± 0.8 (Fig. 5) is in line with the enhancement found for the Ba stars of Karinkuzhi et al. (2018, Fig. 6). A high [N∕C] ratio as found for this object is discussed in the literature (e.g. Smiljanic et al. 2006; Merle et al. 2016). These authors argue that CN processing in Ba stars could result in higher N abundances. According to Smiljanic et al. (2006), an increased [N ∕C] ratio can becaused by mixing events such as the first DU or by a more complex mixing process due to rotation for intermediate mass stars. This would be an indicator for hydrogen burning via the CNO-cycle in the stellar core. With the assumed mass for the primary star (Sect. 6.2), this should be the dominating fusion process in this star. The fast rotation of the Ba-CSPN is most likely due to transfer of angular momentum from the primary and therefore does not imply that this star was rotating exceptionally fast initially so as to affect its [N ∕C] abundance ratio.

For Tc, we could not identify the presence of any line without doubt and, thus, cannot constrain the abundance further than log ɛTc < 2.5. Therefore, we cannot claim this star to have Tc in its atmosphere, which would directly lead to the necessity of prior mass transfer. The models of Karakas & Lugaro (2016) predicted a final surface abundance between log ɛTc = 1.11 and 1.24 for the models with initial masses between 2.1 and 2.5 M⊙, which lies well below the upper limit for Hen 2−39. Another diagnostic element reflecting recent s-process nucleosynthesis is Nb. According to Neyskens et al. (2015), this mono-isotopic species is synthesized by the decay of the radioactive 93Zr produced by s-process nucleosynthesis. Compared to 99Tc, this species has a longer half-life time of 1.53 Myr. Following our estimate made for Tc (Sect. 1), we do not expect an significant enrichment in Nb, since the primary’s post-AGB age should be much shorter than the 93Zr half-life and, thus, a large fraction of this species should still be present. Thus, the Nb/Zr would not represent the 93Zr/Zr ratio at the end of the AGB and cannot be employed as proof for prior mass transfer. Furthermore, the Zr abundance can be determined only within a very large error range and for Nb, we find an upper limit only.

The detection of Tc is not hampered by the resolution of the spectrograph. The limiting factor is the S/N. We estimate the needed S/N that would be necessary to clearly distinguish between a model without Tc and one with log ɛTc = 1.2. From Fig. 7, it becomes clear, that the current S/N is not sufficient to determine a Tc abundance of that level. Currently, the single spectra have a S/N of 3 at that wavelength region. This is increased by co-adding all the spectra, but still a S/N increased by a factor of 3 would be necessary. According to the UVES exposure time calculator (ETC), the needed S/N would require about a six fold longer exposure. For the future Extremely Large Telescope (ELT) the estimate is more promising. By using the E-ELT Spectroscopic ETC, we find that the required S/N is reached with an exposure of about half that of a single observation used in this analysis. For stars with a lower rotational velocity, the detection would become easier (Fig. 7). Unfortunately, all Ba CSPNe that are known up to now seem to rotate fast (shortest period of 4.7 d for Abell 70 and WeBo 1 and longest period of 5.9 d for LoTr 5; Bond et al. 2003; Miszalski et al. 2012; Aller et al. 2018), most likely because of the transfer of angular momentum by accretion of matter from the companion.

6.2 Mass transfer

By comparing our determined enrichment in s-process elements to the yields from evolutionary models for different initial masses of Karakas et al. (2018), we try to confirm that this can be the result of realistic mass transfer. For a primary that is currently in the stage of a CSPN, the secondary should have a mass that is lower by about 5% to be currently in the evolutionary stage of a red giant (assuming a mass dependent relation for the main-sequence lifetime t ~ M−2.5). According to Joss et al. (1987), these stars should have a radiative core of about 0.3 M⊙ and a convective envelope of a mass Menv = Mini − 0.3M⊙ within which the accreted mass becomes diluted. We want to determine a realistic mass range for the primary by comparing the total mass for thedifferent elements that is ejected during the AGB evolution with the mass that the secondary would have neededto accrete to become that enriched. The mass of element X that needs to be accreted is given by Mneed = Menv(mfX,final −mfX,initial) with a final mass fraction mfX,final according to our analysis results and an initial mass fraction mfX,initial according to the low metallicity. In Fig. 8, we show the needed mass compared to the total ejected mass for C, N, and the elements heavier than Fe that show a significant production due to AGB nucleosynthesis. Since most of the abundances could not be constrained within small error limits, we focus on the C and Ba abundances. For all models of Fig. 8 and some additional models, we calculated the percentage of the total ejecta that would need to be accreted to produce the observed enrichment (Table 3). It is obvious that only the models for an initial mass between 1.5 and 4.0 M⊙ can explain the enrichment due to a realistic mass transfer. For the models with the lowest initial masses as well as for those with the highest masses, the ratio of the yield of Ba to that of C is smaller, i.e., these models produce a smaller amount of Ba compared to C. The fraction of accreted mass of C and Ba should be equal. Thus, a 1.75–3.00 M⊙ progenitor seems to be most consistent with the abundance determinations. For this analysis, we used the models with the largest 13C-pocket that are available from Karakas & Lugaro (2016). The larger the pocket size, the lower the yield of C and the higher that of Ba, i.e., the ratio of the yield of Ba to that of C is larger. Even for the models with the largest 13C-pocket, the percentage of the total ejecta that would need to be accreted is higher considering the Ba abundance compared to the C abundance. This ratio becomes worse for smaller pocket sizes. In addition to the choice of the 13C-pocket, yields of the evolutionary models are affected by uncertainties due to mass loss, convective mixing, reaction rates, and neutron poisons (Karakas & Lattanzio 2014). These effects are not evaluated by Karakas & Lugaro (2016) and we take the tabulated yields without considering an error range. However, the progentitor mass estimate is affected by large uncertainties on the abundances and on the model yields and, thus, should be treated with caution.

This result leads to the conclusion that even such a high enrichment can reasonably be explained with realistic mass-transfer mechanisms such as wind-RLOF (Chen et al. 2017). In this scenario only a small fraction of mass becomes unbound from the binary and the percentage of accreted mass ranges between 20 and 40%. Simulations indicate that a binary with a wide separation, where mass transfer would act via the Bondi-Hoyle mechanism, can be ruled out since the percentage of accreted mass decreases to only 2–3% (Theuns et al. 1996). CE evolution would imply a short orbital period (≤ a few days), which is not the case for this binary and, thus, this scenario is also ruled out.

thumbnail Fig. 5

Atmospheric element abundances of Hen 2−39 compared to the final yields of a selection of evolutionary models from Karakas & Lugaro (2016) with a metallicity of Z = 0.007. The initial masses are indicated in the upper panel. Arrows indicate upper limits.
thumbnail Fig. 6

Atmospheric element abundances of Hen 2−39 (blue) compared to the values for the three Ba stars HD515959, HD88035, and HD121447 analyzed by Karinkuzhi et al. (2018, purple) and the ranges that Merle et al. (2016, red) and de Castro et al. (2016, green) found for their sets of Ba-stars. Arrows indicate upper limits.

6.3 Spectroscopic distance

For Hen 2−39, several different distances are published. The nebular analysis of Miszalski et al. (2013a) yields 5.7 kpc. Frew et al. (2016) found 7.6−1.3+1.5$7.6^{&#x002B;1.5}_{-1.3}$ kpc based on the companion spectral type. The CSPN of Hen 2−39 is in the Gaia data release (DR2, ID: 5256396485463285504, Gaia Collaboration 2018).

The parallax of 0.0564 ± 0.0340 mas is affected by a large error corresponding to relative errors of 60.3%. The object is also contained in the catalog of distances of Bailer-Jones et al. (2018) derived from the DR2 data. They found 9.071−1.962+2.939$9.071^{&#x002B;2.939}_{-1.962}$ kpc. With our result for the companion mass, we would like to get a clue for the distance for the binary in Hen 2–39. Using the assumed giant log g = 2.5, the derived mass of about 2.2 M⊙ for the Ba star, and the determined Teff, we get Mbol= Mbol,⊙+2.5(logg−logg⊙)−2.5log(MM⊙)−10log(TeffTeff⊙)=0.262. \begin{align*} M_{\mathrm{bol}} =&\, M_{\mathrm{bol,}\odot} &#x002B; 2.5 \left( \log g - \log g_{\odot} \right) \\ &- 2.5 \log \left( \frac{M}{M_{\odot}}\right) - 10 \log \left( \frac{{T_{\mathrm{eff}}}}{{T_{\mathrm{eff}}}_{\odot}}\right) = 0.262. \end{align*}

The solar values are taken from the Sun Facts Sheet from the NASA Goddard Space Flight Center8 and we use Mbol,⊙ = 4.74 mag for the Sun derived from the standardized absolute bolometric magnitude scale defined by the international astronomical union (Mamajek et al. 2015). The error of Mbol is dominated by the uncertainty of log g. Assuming an uncertainty of Δlog g = 0.5 and ΔM = 0.5M⊙, we get ΔMbol = 1.316. Nevertheless, we can derive the absolute visual magnitude MV = Mbol −BC(V) = 0.852 ± 1.525. The bolometric correction of BC(V) = −0.590 ± 0.209 is calculated using the approach of Alonso et al. (1999) including the preliminary values for Teff and the Fe abundance.

We can now estimate the distance via the distance modulus but the known V magnitude shows a much larger uncertainty than the more recent infrared magnitudes (Table 4). By comparing the calculated flux normalized on the K magnitude of Cutri et al. (2003) with the B and V magnitudes of Tylenda et al. (1991) and those for the I, J, and K bands from Epchtein et al. (1999), it becomes obvious that our model agrees very well with the brightness values in all filters (Fig. 9). Thus, we decided to rely on the precise I magnitude for the distance estimation. Using the color relation for (V − I) from Alonso et al. (1999) for the given Teff, we find (V−I)=1.543−0.110+0.124$(V-I) = 1.543^{&#x002B;0.124}_{-0.110}$ mag. This leads to an absolute MI brightness of MI=−0.691−1.649+1.635$M_I = -0.691^{&#x002B;1.635}_{-1.649}$ mag.

Miszalski et al. (2013a) determined an extinction of EB-V = 0.37 mag. With the Galactic extinction law with RV = 3.1 and the relation from Cardelli et al. (1989), we derive the total absorption for the I band of AI = 0.5 AV = 0.5 × 3.1 EB-V = 0.565 mag. Now, we find d=10−(MI−I+AI−5)/5=9.15−4.90+10.65$d = 10^{-\left( M_I - I &#x002B; A_I - 5\right)/5} = 9.15^{&#x002B;10.65}_{-4.90}$ kpc. The large error is again an effect of the assumed uncertainty in log g but, nevertheless, the value agrees with all other distance values within these limits. Furthermore, this value is very close to that derived from the Gaia parallax measurement. With a precise distance measurement, we could get a second handle on the mass of the Ba star. The mass could then be derived by comparing its properties to evolutionary tracks for different masses and compared with our value derived using the abundance yields of evolutionary models.

With its Galactic latitude of − 4.239° (Gaia Collaboration 2018), the star is located 0.676−0.362+0.789$0.676^{&#x002B;0.789}_{-0.362}$ kpc below the Galactic plane, which means that it is just below the edge of the Galactic thin disk (Rix & Bovy 2013) and, thus, should belong to the thick disk. This assignation is in agreement with the observed low metallicity since simulations for the Galactic metallicity distribution predict negative metallicity gradients for low scale heights and may change the sign at about a scale height of 1.5 kpc. Ivezić et al. (2012) also found thick disk stars to be more metal poor (median [Fe∕H] = −0.6) compared to thin disk stars (median [Fe∕H] = −0.2), where Hen 2−39 lies just in the middle. Furthermore, we can use Mbol to estimate the luminosity L/L⊙=10(Mbol,⊙−Mbol)/2.5=61.83−43.43+145.95$L / L_{\odot} = 10^{\left( M_{\mathrm{bol,}\odot} - {M_{\mathrm{bol}}} \right) / 2.5} = 61.83^{&#x002B;145.95}_{-43.43}$ and the radius RR⊙=LL⊙Teff⊙4Teff4=13.84−6.78+13.38.\begin{align*} \frac{R}{R_{\odot}} = \sqrt[]{\frac{L}{L_{\odot}} \frac{T_{\mathrm{eff \odot}} ^4}{T_{\mathrm{eff}}^4}} = 13.84^{&#x002B;13.38}_{-6.78}. \vspace*{-1.5pt}\end{align*}

Using the rotational velocity from Sect. 4.1 and the radius the star should have a rotational period of 18.4 d for a high inclination of i = 90°. Miszalski et al. (2013a) detected a photometrically variability of the star with a period of 5.46 d. Assuming this value for the rotation, we can find an inclination of i=17.22° −8.53+18.30$i = 17.22^{\circ}\,^{&#x002B;18.30}_{-8.53}$. The parameters are summarized in Table 5. We speculate that the rotational axis of the giant is perpendicular to the binary orbital plane. The low inclination is then in good agreement with the ring-like appearance of the nebula (Miszalski et al. 2013a), which indicates a nearly pole-on view and therefore a binary orbital plane almost in the plane of the sky (Hillwig et al. 2016).

thumbnail Fig. 7

Left panel: comparison between models with log ɛTc = 0.0, 1.2, 2.5 (red, blue, green) for models broadened with a rotational profile (vrot = 36 km s−1, solid lines with 0.35 offset) and without rotation (dashed). Right panel: comparison betweena rotationally broadened model without Tc (black) and models with log ɛTc = 1.2 with different levels of artificial noise (S/N of the single observation: red, S/N of co-added spectrum: blue, required S/N: green). For comparison, the observation is shown with an offset of 0.35 (single spectrum: gray, co-added spectra: black).
Table 3

Percentage of the total ejecta that would need to be accreted to produce the observed enrichment of Hen 2−39 for evolutionary models of Karakas & Lugaro (2016) with Z = 0.007 for different initial masses.

thumbnail Fig. 8

Total amount of mass ejected during AGB evolution of different elements for evolutionary models from Karakas & Lugaro (2016) for Z = 0.007 with initial masses of 1.0–5.0 M⊙ (blue, initial mass indicated in the panel) compared to the need of accreted mass to reproduce the determined abundance values of the secondary via mass transfer (red).
Table 4

Brightnesses in different filters for the Ba CSPN of Hen 2–39.

thumbnail Fig. 9

Synthetic spectrum of our best model of Hen 2−39 normalized to the H magnitude of Cutri et al. (2003) and convolved with a Gaussian with FWHM = 5Å for clarity, including interstellar reddening with EB-V = 0.37 (red) compared with the observed magnitudes (Table 4).
Table 5

Properties of the CSPN of Hen 2−39.

7 Summary and conclusions

We presented and discussed the spectral analysis of UVES spectra of the Ba CSPN of Hen 2−39. Within the error limits, we confirm the result of Miszalski et al. (2013a) that the observed nucleus of Hen 2−39 has a cool atmosphere of Teff = (4350 ± 150) K. Furthermore, we confirm the C and Ba enrichment but can significantly improve the abundances of [C ∕H] = 0.36 ± 0.08 and [Ba ∕Fe] = 1.8 ± 0.50 due to the high resolution of the analyzed spectra. We determined abundances or upper abundance limits for 26 trans-iron elements for the first time. For Tc, the lightest element with no stable isotope, we find an upper abundance limit of log ɛTc < 2.5. This does not confirm the presence of Tc in the atmosphere of the star proving prior mass transfer. The limiting factor is not the resolution of the spectrum. For a clear detection of Tc, an exposure time about six times longer than that of all the spectra combined would be required to obtain the necessary S/N ratio. We can find a low metallicity of [Fe ∕H] = −0.3 ± 1.0 for the Ba giant. The determined abundance pattern requires mass transfer from a companion with an extremely high enrichment of AGB nucleosynthesis products. The comparison with nucleosynthesis models of Karakas & Lugaro (2016) indicates an initial mass of 1.75–3.00 M⊙ for the primary.The percentage of ejected mass that needs to be accreted indicates that the preferred mass transfer mechanism is wind-RLOF. A wide binary involving Bondi-Hoyle accretion can be ruled out as can a CE evolution. For this star, the distance is rather uncertain. Thus, it cannot be used for a spectroscopic determination of the mass by interpolation from evolutionary tracks. A precise spectroscopic determination of the distance is hampered by the fact that log g cannot be constrained within narrow error limits from the analysis of the spectra. It is highly desirable to get a more precise distance measurement. This would also help to get a second value for the mass of the Ba CSPN to compare with that derived from the comparison with AGB models. With this second measurement, it would be possible to refine the primary mass estimate and place stronger constraints on the mass transfer. Our result for the height above the Galactic plane places this system among the thick disk population, in good agreement with the subsolar metallicity derived by our analysis.

Acknowledgements

We thank the anonymous referee for their constructive review of the manuscript. We thank Brent Miszalski and Thomas Rauch for their helpful comments and suggestions. L.L. is supported by the German Research Foundation (DFG, grant WE 1312/49-1) and by the Studentship Programme of the European Southern Observatory. D.J. gratefully acknowledges the Spanish Ministry of Economy and Competitiveness (MINECO) under the grant AYA2017-83383-P. This research has made use of NASA’s Astrophysics Data System and the SIMBAD database, operated at CDS, Strasbourg, France. This work has made use of data from the European Space Agency (ESA) mission Gaia (https://www.cosmos.esa.int/gaia), processed by the Gaia Data Processing and Analysis Consortium (DPAC, https://www.cosmos.esa.int/web/gaia/dpac/consortium). Funding for the DPAC has been provided by national institutions, in particular the institutions participating in the Gaia Multilateral Agreement.

Appendix A Additional figures and tables

Table A.1

Observation log for the UVES observations.

Table A.2

Diagnostic lines used for the determination of stellar parameters.

thumbnail Fig. A.1

Observation (gray) of Hen 2−39 compared to model spectra with Teff = 4000, 4250, 4500 K (blue, red, green) for selected regions that were used for the determination of Teff. CN, C2, and CH absorption lines are indicated at the bottom in green, purple, and cyan, respectively. All absorption lines that appear with an equivalent width ≥ 20 mÅ in the calculated spectrum are indicated.
thumbnail Fig. A.2

Like Fig. A.1, for log g = 2.0, 2.5, 3.0 (blue, red, green).
Table A.3

Diagnostic lines used for our determination of element abundances.

thumbnail Fig. A.3

Difference between the newly computed C-rich test model (model 2) and the O-rich model from the available grid (model 1) for Teff = 4250 K and log g = 2.5.
thumbnail Fig. A.4

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Na I for [Na ∕Fe] = 0.35, 0.05, −0.25 (green, red, and blue, respectively), S I for [S∕Fe] = 2.17, 1.17, 0.17, Al I for [Al∕Fe] = 0.68, −0.32, −1.32, and K I for [K∕Fe] = 1.07, 0.07, −0.93. Interstellar absorption lines are indicated with blue marks.
thumbnail Fig. A.5

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Ca I for [Ca∕Fe] = 0.41, −0.09, −0.59 (green, red, and blue, respectively), Sc I for [Sc∕Fe] = 0.08, −0.42, −0.92, Cr I for [Cr∕Fe] = 0.32, −0.18, −0.68, Ti I for [Ti∕Fe] = −0.13, −0.63, −1.13, and V I for [V∕Fe] = −0.05, −0.55, −1.05. Artifacts arising from the overcorrection of nebula lines are indicated.
thumbnail Fig. A.6

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Mn I for [Mn ∕Fe] = 0.52, 0.02, −0.48 (green, red, and blue, respectively), Co I for [Co∕Fe] = 1.03, 0.53, 0.03, Cu I for [Cu∕Fe] = 2.18, 1.18, 0.18, Fe I for [Fe∕H] = 0.15, −0.35, −0.85, and Ni I for [Ni∕Fe] = 1.09, 0.59, 0.09. Interstellar absorption lines are indicated with blue marks.
thumbnail Fig. A.7

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Zn I for [Zn∕Fe] = 2.51, 1.51, 0.51 (green, red, and blue, respectively), Sr I for [Sr∕Fe] = 1.99, 0.99, −0.01, Rb I for [Rb∕Fe] = 2.41, 1.41, 0.41, Y I and Y II for [Y∕Fe] = 1.37, 0.37, −0.63, Zr I and Zr II for [Zr∕Fe] = 1.19, 0.19 − 0.81, and Nb I for [Nb∕Fe] = 1.97, 0.97, −0.03. Artifacts arising from the overcorrection of nebula lines are indicated.
thumbnail Fig. A.8

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Mo I for [Mo ∕Fe] = 2.35, 1.35, 0.35 (green, red, and blue, respectively), Tc I for log ɛTc = 3.5, 2.5, 1.5, Ru I for [Ru∕Fe] = 3.05, 2.05, 1.05, Ba II [Ba ∕Fe]= 2.29, 1.79, 1.27, La II for [La∕Fe] = 2.44, 1.44, 0.44, Ce I and Ce II for [Ce∕Fe] = 3.29, 2.29, 1.29, Pr II for [Pr∕Fe] = 3.68, 2.68, 1.68, and Nd I for [Nd∕Fe] = 1.76, 0.76, −0.24. Artifacts arising from the overcorrection of nebula lines are indicated.
thumbnail Fig. A.9

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Sm I and Sm II for [Sm ∕Fe] = 2.08, 1.08, 0.08 (green, red, and blue, respectively), Eu II for [Eu∕Fe] = 1.95, 0.95, −0.05, Gd I and Gd II [Gd∕Fe] = 2.80, 1.80, 0.80, Tb II for [Tb∕Fe] = 1.79, 0.79, −0.21, Dy II for [Dy∕Fe] = 4.75, 3.75, 2.75, Er I and Er II[Er∕Fe] = 2.86, 1.86, 0.86, Hf I for [Hf∕Fe] = 2.31, 1.31, 0.31, W I for [W∕Fe] = 2.05, 1.05, 0.05, and Os I for [Os∕Fe] = 2.76, 1.76, 0.76. Artifacts arising from the overcorrection of nebula lines are indicated.

References

  1. Abate, C., Pols, O. R., Izzard, R. G., Mohamed, S. S., & de Mink, S. E. 2013, A&A, 552, A26 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  2. Acker, A., Marcout, J., Ochsenbein, F., et al. 1992, The Strasbourg-ESO Catalogue of Galactic Planetary Nebulae. Parts I, II (Garching: European Southern Observatory) [Google Scholar]
  3. Aller, A., Lillo-Box, J., Vučković, M., et al. 2018, MNRAS, 476, 1140 [NASA ADS] [CrossRef] [Google Scholar]
  4. Alonso, A., Arribas, S., & Martínez-Roger, C. 1999, A&AS, 140, 261 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  5. Asplund, M., Grevesse, N., Sauval, A. J., & Scott, P. 2009, ARA&A, 47, 481 [NASA ADS] [CrossRef] [Google Scholar]
  6. Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R. 2018, AJ, 156, 58 [NASA ADS] [CrossRef] [Google Scholar]
  7. Bidelman, W. P., & Keenan, P. C. 1951, ApJ, 114, 473 [NASA ADS] [CrossRef] [Google Scholar]
  8. Blanco-Cuaresma, S., Soubiran, C., Heiter, U., & Jofré, P. 2014, A&A, 569, A111 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  9. Boffin, H. M. J. 2015, Ecology of Blue Straggler Stars (Berlin, Heidelberg: Springer-Verlag) [Google Scholar]
  10. Boffin, H. M. J., & Jorissen, A. 1988, A&A, 205, 155 [NASA ADS] [Google Scholar]
  11. Bond, H. E., Pollacco, D. L., & Webbink, R. F. 2003, AJ, 125, 260 [NASA ADS] [CrossRef] [Google Scholar]
  12. Buckley, D. A. H., Swart, G. P., & Meiring, J. G. 2006, Proc. SPIE, 6267, 62670Z [Google Scholar]
  13. Burgh, E. B., Nordsieck, K. H., Kobulnicky, H. A., et al. 2003, in Instrument Design and Performance for Optical/Infrared Ground-based Telescopes, eds. M. Iye, & A. F. M. Moorwood, Proc. SPIE, 4841, 1463 [NASA ADS] [CrossRef] [Google Scholar]
  14. Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245 [NASA ADS] [CrossRef] [Google Scholar]
  15. Castelli, F., & Kurucz, R. L. 2003, in Modelling of Stellar Atmospheres, eds. N. Piskunov, W. W. Weiss, & D. F. Gray, IAU Symp., 210, A20 [Google Scholar]
  16. Chen, Z., Frank, A., Blackman, E. G., Nordhaus, J., & Carroll-Nellenback, J. 2017, MNRAS, 468, 4465 [NASA ADS] [CrossRef] [Google Scholar]
  17. Corradi, R. L. M., Sabin, L., Miszalski, B., et al. 2011, MNRAS, 410, 1349 [NASA ADS] [CrossRef] [Google Scholar]
  18. Cutri, R. M., Skrutskie, M. F., van Dyk, S., et al. 2003, VizieR Online Data Catalog: II/246 [Google Scholar]
  19. de Castro, D. B., Pereira, C. B., Roig, F., et al. 2016, MNRAS, 459, 4299 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  20. De Marco O. 2009, PASP, 121, 316 [NASA ADS] [CrossRef] [Google Scholar]
  21. Dekker, H., D’Odorico, S., Kaufer, A., Delabre, B., & Kotzlowski, H. 2000, in Optical and IR Telescope Instrumentation and Detectors, eds. M. Iye, & A. F. Moorwood, Proc. SPIE, 4008, 534 [Google Scholar]
  22. Epchtein, N., Deul, E., Derriere, S., et al. 1999, VizieR Online Data Catalog: II/240 [Google Scholar]
  23. Frew, D. J., Parker, Q. A., & Bojičić, I. S. 2016, MNRAS, 455, 1459 [NASA ADS] [CrossRef] [Google Scholar]
  24. Gaia Collaboration 2018, VizieR Online Data Catalog: I/345 [Google Scholar]
  25. Gray, R. O., & Corbally, C. J. 1994, AJ, 107, 742 [NASA ADS] [CrossRef] [Google Scholar]
  26. Gray, R. O., McGahee, C. E., Griffin, R. E. M., & Corbally, C. J. 2011, AJ, 141, 160 [NASA ADS] [CrossRef] [Google Scholar]
  27. Grevesse, N., & Sauval, A. J. 1998, Space Sci. Rev., 85, 161 [NASA ADS] [CrossRef] [Google Scholar]
  28. Han, Z., Eggleton, P. P., Podsiadlowski, P., & Tout, C. A. 1995, MNRAS, 277, 1443 [NASA ADS] [CrossRef] [Google Scholar]
  29. Henize, K. G. 1967, ApJS, 14, 125 [NASA ADS] [CrossRef] [Google Scholar]
  30. Herwig, F. 2005, ARA&A, 43, 435 [NASA ADS] [CrossRef] [Google Scholar]
  31. Hillwig, T. C., Jones, D., De Marco, O., et al. 2016, ApJ, 832, 125 [NASA ADS] [CrossRef] [Google Scholar]
  32. Hubeny, I., Mihalas, D., & Werner, K., eds. 2003, in Stellar Atmosphere Modeling, ASP Conf. Ser., 288, 17 [NASA ADS] [Google Scholar]
  33. Husti, L., Gallino, R., Bisterzo, S., Straniero, O., & Cristallo, S. 2009, PASA, 26, 176 [NASA ADS] [CrossRef] [Google Scholar]
  34. Iben, Jr. I., & Livio, M. 1993, PASP, 105, 1373 [Google Scholar]
  35. Ivezić, Ž., Beers, T. C., & Jurić, M. 2012, ARA&A, 50, 251 [NASA ADS] [CrossRef] [Google Scholar]
  36. Jones, D., & Boffin, H. M. J. 2017, Nat. Astron., 1, 0117 [CrossRef] [Google Scholar]
  37. Jones, D., Van Winckel, H., Aller, A., Exter, K., & De Marco, O. 2017, A&A, 600, L9 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  38. Jorissen, A., & Mayor, M. 1988, A&A, 198, 187 [NASA ADS] [Google Scholar]
  39. Jorissen, A., Van Eck, S., Mayor, M., & Udry, S. 1998, A&A, 332, 877 [NASA ADS] [Google Scholar]
  40. Joss, P. C., Rappaport, S., & Lewis, W. 1987, ApJ, 319, 180 [NASA ADS] [CrossRef] [Google Scholar]
  41. Karakas, A. I., & Lattanzio, J. C. 2014, PASA, 31, e030 [NASA ADS] [CrossRef] [Google Scholar]
  42. Karakas, A. I., & Lugaro, M. 2016, ApJ, 825, 26 [NASA ADS] [CrossRef] [Google Scholar]
  43. Karakas, A. I., Lugaro, M., Carlos, M., et al. 2018, MNRAS, 477, 421 [NASA ADS] [CrossRef] [Google Scholar]
  44. Karinkuzhi, D., Goswami, A., Sridhar, N., Masseron, T., & Purandardas, M. 2018, MNRAS, 476, 3086 [NASA ADS] [CrossRef] [Google Scholar]
  45. Kobulnicky, H. A., Nordsieck, K. H., Burgh, E. B., et al. 2003, in Instrument Design and Performance for Optical/Infrared Ground-based, eds. M. Iye & A. F. M. Moorwood Telescopes, Proc. SPIE, 4841, 1634 [Google Scholar]
  46. Kurucz, R. L. 1991, in Stellar Atmospheres: Beyond Classical Models, eds. L. Crivellari, I. Hubeny, & D. G. Hummer (Berlin: Springer Science & Business Media), NATO ASIC Proc., 341, 441 [Google Scholar]
  47. Lebzelter, T., & Hron, J. 2003, A&A, 411, 533 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  48. Madonna, S., García-Rojas, J., Sterling, N. C., et al. 2017, MNRAS, 471, 1341 [NASA ADS] [CrossRef] [Google Scholar]
  49. Madonna, S., Bautista, M. A., Dinerstein, H., et al. 2018, ApJL, 861, L8 [NASA ADS] [CrossRef] [Google Scholar]
  50. Mamajek, E. E., Torres, G., Prsa, A., et al. 2015, ArXiv e-prints [arXiv:1510.06262] [Google Scholar]
  51. Masseron, T., Merle, T., & Hawkins, K. 2016, Astrophysics Source Code Library [record ascl:1605.004] [Google Scholar]
  52. McClure, R. D. 1983, ApJ, 268, 264 [NASA ADS] [CrossRef] [Google Scholar]
  53. McClure, R. D., Fletcher, J. M., & Nemec, J. M. 1980, ApJ, 238, L35 [NASA ADS] [CrossRef] [Google Scholar]
  54. McClure, R. D., & Woodsworth, A. W. 1990, ApJ, 352, 709 [NASA ADS] [CrossRef] [Google Scholar]
  55. Merle, T., Jorissen, A., Van Eck, S., Masseron, T., & Van Winckel, H. 2016, A&A, 586, A151 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  56. Merrill, P. W. 1952, ApJ, 116, 21 [NASA ADS] [CrossRef] [Google Scholar]
  57. Miller Bertolami, M. M. 2016, A&A, 588, A25 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  58. Miszalski, B., Boffin, H. M. J., Frew, D. J., et al. 2012, MNRAS, 419, 39 [NASA ADS] [CrossRef] [Google Scholar]
  59. Miszalski, B., Boffin, H. M. J., Jones, D., et al. 2013a, MNRAS, 436, 3068 [NASA ADS] [CrossRef] [Google Scholar]
  60. Miszalski, B., Boffin, H. M. J., & Corradi, R. L. M. 2013b, MNRAS, 428, L39 [NASA ADS] [CrossRef] [Google Scholar]
  61. Mohamed, S., & Podsiadlowski, P. 2007, in 15th European Workshop on White Dwarfs, eds. R. Napiwotzki & M. R. Burleigh, ASP Conf. Ser., 372, 397 [NASA ADS] [Google Scholar]
  62. Nagae, T., Oka, K., Matsuda, T., et al. 2004, A&A, 419, 335 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  63. Neyskens, P., van Eck, S., Jorissen, A., et al. 2015, Nature, 517, 174 [NASA ADS] [CrossRef] [PubMed] [Google Scholar]
  64. Palmeri, P., Quinet, P., Biémont, É., et al. 2007, MNRAS, 374, 63 [NASA ADS] [CrossRef] [Google Scholar]
  65. Rix, H.-W., & Bovy, J. 2013, A&ARv, 21, 61 [NASA ADS] [CrossRef] [Google Scholar]
  66. Saladino, M. I., Pols, O. R., van der Helm, E., Pelupessy, I., & Portegies Zwart, S. 2018, A&A, 618, A50 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  67. Smiljanic, R., Barbuy, B., de Medeiros, J. R., & Maeder, A. 2006, Rev. Mex. Astron. Astrofis. Conf. Ser., 26, 45 [NASA ADS] [Google Scholar]
  68. Sneden, C., Lucatello, S., Ram, R. S., Brooke, J. S. A., & Bernath, P. 2014, ApJS, 214, 26 [Google Scholar]
  69. Tabernero, H. M., Dorda, R., Negueruela, I., & González-Fernández, C. 2018, MNRAS, 476, 3106 [Google Scholar]
  70. Theuns, T., Boffin, H. M. J., & Jorissen, A. 1996, MNRAS, 280, 1264 [NASA ADS] [CrossRef] [Google Scholar]
  71. Thevenin, F., & Jasniewicz, G. 1997, A&A, 320, 913 [NASA ADS] [Google Scholar]
  72. Tylenda, R., Acker, A., Stenholm, B., Gleizes, F., & Raytchev, B. 1991, A&AS, 89, 77 [NASA ADS] [Google Scholar]
  73. Tyndall, A. A., Jones, D., Boffin, H. M. J., et al. 2013, MNRAS, 436, 2082 [NASA ADS] [CrossRef] [Google Scholar]
  74. Van Eck, S., & Jorissen, A. 1999, A&A, 345, 127 [NASA ADS] [Google Scholar]
  75. Werner, K., & Herwig, F. 2006, PASP, 118, 183 [NASA ADS] [CrossRef] [Google Scholar]
  76. Wray, J. D. 1966, Ph.D. Thesis, Northwestern University, IL, USA [Google Scholar]
1

[A∕B] = log(nA∕nB) − log(nA,⊙∕nB,⊙) with the number fractions n for element A and B.

2

Los Alamos National Laboratory Periodic Table http:// periodic.lanl.gov

7

log ɛ = 12 + log(nX∕nH).

All Tables

Table 1

Observation time and heliocentric radial velocities for the 18 observations of the Ba CSPN of Hen 2−39.

In the text
Table 2

Element abundances determined for Hen 2−39 in log ɛ = 12 + log (nX∕nH), [X ∕H] = log (nX∕nH) − log (nX,⊙∕nH,⊙), and [X ∕Fe] = log (nX∕nFe) − log (nX,⊙∕nFe,⊙) with the number fraction nX for element X.

In the text
Table 3

Percentage of the total ejecta that would need to be accreted to produce the observed enrichment of Hen 2−39 for evolutionary models of Karakas & Lugaro (2016) with Z = 0.007 for different initial masses.

In the text
Table 4

Brightnesses in different filters for the Ba CSPN of Hen 2–39.

In the text
Table 5

Properties of the CSPN of Hen 2−39.

In the text
Table A.1

Observation log for the UVES observations.

In the text
Table A.2

Diagnostic lines used for the determination of stellar parameters.

In the text
Table A.3

Diagnostic lines used for our determination of element abundances.

In the text

All Figures

thumbnail Fig. 1

Synthetic spectra (red) convolved with a rotational profile of vrot = 38 km s-1 around strong C2 and CN molecular absorption bands compared with observations (gray).
In the text
thumbnail Fig. 2

Observation (gray) of Hen 2−39 compared to model spectra with Teff = 4250 K and log g = 2.5 for a selectedregion of strong C2 absorption for [C∕H] = 0.46, 0.36, 0.26 (green, red, and blue, respectively).
In the text
thumbnail Fig. 3

Like Fig. 2, for strong CN absorption for [N∕H] = 1.1, 0.3, −0.5 (green, red, and blue, respectively).
In the text
thumbnail Fig. 4

Like Fig. 2, for strong CN absorption for 12C∕13C = 90, 10, 5 (red, green, and blue, respectively). 13CN absorption lines are indicated at the bottom in cyan.
In the text
thumbnail Fig. 5

Atmospheric element abundances of Hen 2−39 compared to the final yields of a selection of evolutionary models from Karakas & Lugaro (2016) with a metallicity of Z = 0.007. The initial masses are indicated in the upper panel. Arrows indicate upper limits.
In the text
thumbnail Fig. 6

Atmospheric element abundances of Hen 2−39 (blue) compared to the values for the three Ba stars HD515959, HD88035, and HD121447 analyzed by Karinkuzhi et al. (2018, purple) and the ranges that Merle et al. (2016, red) and de Castro et al. (2016, green) found for their sets of Ba-stars. Arrows indicate upper limits.
In the text
thumbnail Fig. 7

Left panel: comparison between models with log ɛTc = 0.0, 1.2, 2.5 (red, blue, green) for models broadened with a rotational profile (vrot = 36 km s−1, solid lines with 0.35 offset) and without rotation (dashed). Right panel: comparison betweena rotationally broadened model without Tc (black) and models with log ɛTc = 1.2 with different levels of artificial noise (S/N of the single observation: red, S/N of co-added spectrum: blue, required S/N: green). For comparison, the observation is shown with an offset of 0.35 (single spectrum: gray, co-added spectra: black).
In the text
thumbnail Fig. 8

Total amount of mass ejected during AGB evolution of different elements for evolutionary models from Karakas & Lugaro (2016) for Z = 0.007 with initial masses of 1.0–5.0 M⊙ (blue, initial mass indicated in the panel) compared to the need of accreted mass to reproduce the determined abundance values of the secondary via mass transfer (red).
In the text
thumbnail Fig. 9

Synthetic spectrum of our best model of Hen 2−39 normalized to the H magnitude of Cutri et al. (2003) and convolved with a Gaussian with FWHM = 5Å for clarity, including interstellar reddening with EB-V = 0.37 (red) compared with the observed magnitudes (Table 4).
In the text
thumbnail Fig. A.1

Observation (gray) of Hen 2−39 compared to model spectra with Teff = 4000, 4250, 4500 K (blue, red, green) for selected regions that were used for the determination of Teff. CN, C2, and CH absorption lines are indicated at the bottom in green, purple, and cyan, respectively. All absorption lines that appear with an equivalent width ≥ 20 mÅ in the calculated spectrum are indicated.
In the text
thumbnail Fig. A.2

Like Fig. A.1, for log g = 2.0, 2.5, 3.0 (blue, red, green).
In the text
thumbnail Fig. A.3

Difference between the newly computed C-rich test model (model 2) and the O-rich model from the available grid (model 1) for Teff = 4250 K and log g = 2.5.
In the text
thumbnail Fig. A.4

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Na I for [Na ∕Fe] = 0.35, 0.05, −0.25 (green, red, and blue, respectively), S I for [S∕Fe] = 2.17, 1.17, 0.17, Al I for [Al∕Fe] = 0.68, −0.32, −1.32, and K I for [K∕Fe] = 1.07, 0.07, −0.93. Interstellar absorption lines are indicated with blue marks.
In the text
thumbnail Fig. A.5

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Ca I for [Ca∕Fe] = 0.41, −0.09, −0.59 (green, red, and blue, respectively), Sc I for [Sc∕Fe] = 0.08, −0.42, −0.92, Cr I for [Cr∕Fe] = 0.32, −0.18, −0.68, Ti I for [Ti∕Fe] = −0.13, −0.63, −1.13, and V I for [V∕Fe] = −0.05, −0.55, −1.05. Artifacts arising from the overcorrection of nebula lines are indicated.
In the text
thumbnail Fig. A.6

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Mn I for [Mn ∕Fe] = 0.52, 0.02, −0.48 (green, red, and blue, respectively), Co I for [Co∕Fe] = 1.03, 0.53, 0.03, Cu I for [Cu∕Fe] = 2.18, 1.18, 0.18, Fe I for [Fe∕H] = 0.15, −0.35, −0.85, and Ni I for [Ni∕Fe] = 1.09, 0.59, 0.09. Interstellar absorption lines are indicated with blue marks.
In the text
thumbnail Fig. A.7

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Zn I for [Zn∕Fe] = 2.51, 1.51, 0.51 (green, red, and blue, respectively), Sr I for [Sr∕Fe] = 1.99, 0.99, −0.01, Rb I for [Rb∕Fe] = 2.41, 1.41, 0.41, Y I and Y II for [Y∕Fe] = 1.37, 0.37, −0.63, Zr I and Zr II for [Zr∕Fe] = 1.19, 0.19 − 0.81, and Nb I for [Nb∕Fe] = 1.97, 0.97, −0.03. Artifacts arising from the overcorrection of nebula lines are indicated.
In the text
thumbnail Fig. A.8

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Mo I for [Mo ∕Fe] = 2.35, 1.35, 0.35 (green, red, and blue, respectively), Tc I for log ɛTc = 3.5, 2.5, 1.5, Ru I for [Ru∕Fe] = 3.05, 2.05, 1.05, Ba II [Ba ∕Fe]= 2.29, 1.79, 1.27, La II for [La∕Fe] = 2.44, 1.44, 0.44, Ce I and Ce II for [Ce∕Fe] = 3.29, 2.29, 1.29, Pr II for [Pr∕Fe] = 3.68, 2.68, 1.68, and Nd I for [Nd∕Fe] = 1.76, 0.76, −0.24. Artifacts arising from the overcorrection of nebula lines are indicated.
In the text
thumbnail Fig. A.9

Observation (gray) of Hen 2−39 compared to model spectra for selected regions around absorption lines of Sm I and Sm II for [Sm ∕Fe] = 2.08, 1.08, 0.08 (green, red, and blue, respectively), Eu II for [Eu∕Fe] = 1.95, 0.95, −0.05, Gd I and Gd II [Gd∕Fe] = 2.80, 1.80, 0.80, Tb II for [Tb∕Fe] = 1.79, 0.79, −0.21, Dy II for [Dy∕Fe] = 4.75, 3.75, 2.75, Er I and Er II[Er∕Fe] = 2.86, 1.86, 0.86, Hf I for [Hf∕Fe] = 2.31, 1.31, 0.31, W I for [W∕Fe] = 2.05, 1.05, 0.05, and Os I for [Os∕Fe] = 2.76, 1.76, 0.76. Artifacts arising from the overcorrection of nebula lines are indicated.
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.