EDP Sciences logo
Subscriber Authentication Point
Search
Menu
Home All issues Volume 667 (November 2022) A&A, 667 (2022) A84 Full HTML
Open Access
Issue
A&A
Volume 667, November 2022
Article Number A84
Number of page(s) 12
Section Stellar structure and evolution
DOI https://doi.org/10.1051/0004-6361/202142918
Published online 09 November 2022

© V. Karambelkar et al. 2022

Licence Creative CommonsOpen Access article, published by EDP Sciences, under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

This article is published in open access under the Subscribe-to-Open model. Subscribe to A&A to support open access publication.

1. Introduction

R Coronae Borealis (RCB) and dustless Hydrogen-deficient Carbon (dLHdC)1 stars are supergiants characterized by a peculiar chemical composition – an acute deficiency of hydrogen and an overabundance of carbon (Clayton 1996; Feast 1996; Lambert & Rao 1994). Together, they form the class of Hydrogen-deficient Carbon (HdC) stars. Their chemical compositions suggest that they are remnants of He-core and CO-core white dwarf (WD) mergers (Webbink 1984; Iben & Tutukov 1984; Saio & Jeffery 2002), making them low mass counterparts of type Ia supernovae in the double-degenerate (DD) scenario (Fryer et al. 2008). In addition to peculiar chemical compositions, RCB stars experience rapid photometric declines (up to 8 mag in V band) that are attributed to dust formation episodes (Clayton 2012). dLHdC stars do not show any such declines or any significant infrared (IR) excess, suggesting that they do not undergo any dust formation (Warner 1967; Goswami et al. 2010; Tisserand 2012). Why RCB stars produce dust while dLHdC stars do not, despite having no other known chemical differences is still a mystery.

It has been relatively easier to identify RCB stars than dLHdC stars owing to their spectacular photometric variations. There are 128 RCB stars known in the Milky Way, while only five Galactic dLHdC stars had been known for the last four decades (Clayton 2012; Tisserand et al. 2020; Karambelkar et al. 2021). However, the sample of Galactic dLHdC stars has more than quintupled in the last year with the discovery of 27 new dLHdC stars (Tisserand et al. 2022). Additionally, several new RCB stars have been discovered and observed spectroscopically at near-infrared (NIR) wavelengths (Karambelkar et al. 2021).

NIR spectroscopic observations were key to identifying He-core and CO-core WD mergers as the progenitors of dLHdC and RCB stars. The K-band spectra of these stars show anomalously strong 12C18O first overtone bands in addition to the 12C16O first overtone bands (Clayton et al. 2005). The values of 16O/18O in dLHdC and RCB stars cover a remarkably wide range (from 0.3 to 50), thus in all cases 1–3 orders of magnitude smaller than the solar value (∼500). The large amount of 18O is thought to be produced by partial helium burning in a thin shell around the core of the WD merger remnant, which is convectively dredged up to the surface (Clayton et al. 2007; Crawford et al. 2020).

It has been suggested that in addition to dust production, 16O/18O ratios could be a second difference between RCB and dLHdC stars. Clayton et al. (2007) and García-Hernández et al. (2009) noted that dLHdC stars have 16O/18O < 1 and most RCB stars have 16O/18O > 1. However, this analysis is based on a small sample of only two dLHdC and five RCB stars2. Both dLHdC stars in the sample have 16O/18O < 1, while three of the five RCB stars have 16O/18O > 1. Karambelkar et al. (2021) measured the ratios for six additional RCB stars using medium resolution spectra and found that they had 16O/18O > 1. However, additional measurements of these ratios for dLHdC stars were not possible in the past because of the five dLHdC stars known at the time, only three had temperatures cold enough to show the CO bands (Clayton et al. 2007; García-Hernández et al. 2009, 2010).

In this paper, we present the largest sample of 16O/18O values in dLHdC and RCB stars measured to date. We use medium resolution NIR spectra to measure 16O/18O for seven dLHdC and 33 RCB stars. We also present JHK-band spectra of 21 (20 new, one previously known) dLHdC stars for the first time. Our results conclusively show that most dLHdC stars have lower 16O/18O than most RCB stars. In Sect. 2, we describe our spectroscopic observations, which include data collected over the last 15 years. In Sect. 3, we describe the NIR spectral features of dLHdC stars. We outline our methods for measuring 16O/18O in Sect. 4.1. We discuss the implications of the different oxygen isotope ratios for formation scenarios of RCB and dLHdC stars in Sect. 5, and conclude with a summary of our results in Sect. 6.

2. Data

Our data comprise medium resolution NIR spectra of 24 dLHdC stars and 49 RCB stars. All but two of the stars are resident in the Milky Way; the others are in the Magellanic Clouds. The complete log of spectroscopic observations together with the S/N for each spectrum is presented in Table 1.

Table 1.

Log of spectroscopic observations.

The spectra of 18 of the 20 newly discovered dLHdC stars were taken with the Triplespec spectrograph (Herter et al. 2008, R ≈ 3000) on the 200-inch Hale telescope at Palomar observatory on several nights between June and October 2021. Two newly discovered dLHdC star (A166 and A183) were observed with the SpeX spectrograph (R ≈ 2500) on the 3 m NASA Infrared Telescope Facility (IRTF, Rayner et al. 2003) on UT 20210623. We include a previously unpublished spectrum of the previously known dLHdC star HD 137613 taken with the UKIRT Imager Spectrometer (UIST, Ramsay Howat et al. 2004, R ≈ 3100) on the 3.8 m United Kingdom Infrared Telescope (UKIRT) on UT 20050309. We also include spectra of three other previously known dLHdC stars obtained with the Gemini Near-infrared Spectrograph (GNIRS, Elias et al. 2006) on the 8.1 m Gemini-South telescope in the long-slit mode with R ≈ 5900 in 2005, reported by Clayton et al. (2007).

The spectra of the RCB stars were collected over the last 15 years at three telescopes with four different instruments. We include spectra of six RCB stars obtained with Gemini South/GNIRS in September 2005 (Clayton et al. 2007) and 4 more RCB stars observed with the Flamingos 2 spectrograph (Eikenberry et al. 2004) on the Gemini-South telescope in 2015 and 2016. We observed the RCB star Z Umi with GNIRS on the Gemini-North telescope in 2011. In 2013, we observed nine RCB stars with IRTF/SpeX at R ≈ 2500. We observed 12 more RCB stars in 2014 and 26 RCB stars in 2015 with IRTF/Spex at R ≈ 5000.

The Triplespec and IRTF spectra were reduced using the IDL package spextool (Cushing et al. 2004), and were flux calibrated and corrected for telluric absorption with standard star observations using xtellcor (Vacca et al. 2003). The UKIRT spectrum of HD 137613 was reduced in a standard manner using the Figaro software package for flatfielding, removal of the effects of cosmic ray hits, spectral and spatial rectification of the spectral images, wavelength calibration using an argon lamp, and ratioing by the spectrum of a standard star. The Gemini spectra were reduced similarly, but used a combination of IRAF and Figaro as described in Clayton et al. (2007).

3. NIR spectra of dLHdC stars

We present JHK-band NIR spectra of the 20 new dLHdC stars in Fig. 1. We also include in the figure spectra of the previously known dLHdC star HD 137613 (green) and the RCB star NSV11154 (red, taken from Karambelkar et al. 2021).

thumbnail Fig. 1.

Near-infrared spectra of the newly discovered dLHdC stars. Near-infrared spectra of the RCB star NSV11154 and the previously known dLHdC star HD137613 are also shown for comparison. The rest-frame positions of the CN and 12C16O absorption bands are indicated by dashed vertical lines.

The spectra of the dLHdC stars closely resemble NIR spectra of RCB stars taken at maximum light. The continuum shapes resemble those of F-G type stars. Numerous absorption lines are present. We identify strong absorption features attributed to C I (most prominently at 1.06883, 1.0686, 1.0688 and 1.06942 μm in the J-band and 1.73433, 1.74533 and 1.75104 μm in the H-band) and a blend of numerous Fe I, K I and Si I lines (see Rayner et al. 2003 for the wavelengths). H lines are absent in the NIR spectra of the dLHdC stars3. Only five of the twenty new dLHdC stars show 12C16O and 12C18O molecular features (presumably because the others are too hot for CO to exist in detectable amounts). Of the five, all but A166 also show the 12C14N bands at 1.0875, 1.0929, 1.0966, 1.0999 μm. Two additional stars B565 and A811 show CN but no CO bands.

3.1. Helium 1.0833 μm triplet

Our NIR spectra cover the He I 1.0833 μm triplet. This feature can serve as a tracer of high velocity winds around RCB and dLHdC stars. The levels of this transition are 20 eV above the ground state, and cannot be populated by photospheric radiation of dLHdC and RCB stars. Instead, they can be collisionally excited in high velocity winds around these stars. The He I feature has been observed in several RCB stars either as blueshifted absorption or as a P-Cygni profile with velocities as high as 500 km s−1 (Clayton et al. 2003, 2011; Karambelkar et al. 2021). The strength and velocity of the winds around RCB stars are greatest when the stars have just emerged from dust enshrouded minima, and decrease with time thereafter. This suggests that the winds around RCB stars are dust-driven; that is, the gas is dragged to high velocities by dust grains that are accelerated by radiation pressure. As dLHdC stars do not show dust excesses in their spectral energy distributions, we do not expect to see dust-driven winds around them. Geballe et al. (2009) observed the four previously known dLHdC stars and found no evidence for strong RCB-like He I absorption lines in their spectra.

Figure 2 shows a zoom-in of the NIR spectra of the new dLHdC stars around the He I 1.0833 μm triplet. We do not detect RCB-like (width > 200 km s−1) helium absorption in the NIR spectra for any of the newly discovered dLHdC stars, except possibly A166. We cannot determine whether a lower velocity He I component is present in this star from our medium resolution spectra, as we cannot resolve possible contribution of He I to the Si I (1.0831 μm) line. A166 is the only dLHdC star that shows signs of an extended absorption component in addition to Si I absorption. If this is indeed the He I line, it would imply a wind velocity of ≈400 km s−1. Higher resolution observations are necessary to identify the sources of this absorption.

thumbnail Fig. 2.

Zoom-in of the spectral region around the He I (10 833 Å) triplet. The velocity is measured with respect to 1.0833 μm. The same spectral region of He I profile of an RCB star is shown in red – RCB-ToI-223 (a.k.a. WISE J182010.96−193453.4), at a phase when it has just emerged from a dust decline. We cannot resolve the contribution from the Si I line to any He I features from our low resolution spectra. However, it is evident that we do not detect any strong (> 200 km s−1), RCB-like winds in any dLHdC star, suggesting that there are no dust-driven winds around them. This is consistent with no hot dust excesses seen in the SEDs of any of these stars. A166 is the only dLHdC star that shows a possible high velocity He I absorption profile (plotted in blue).

Although we cannot completely resolve any possible low velocity He I component, our observations rule out the presence of strong, RCB-like dust-driven mass loss in dLHdC stars. This is expected, as none of these stars (except A166) shows a significant IR excess (Tisserand et al. 2022). We note, however, that this does not rule out the possibility that dLHdC stars were forming dust at some point in their history (⪆10 years ago). Observations of XX Cam – an RCB star that has not entered a dust-enshrouded decline for the last six decades – do not show any significant He I features, similar to dLHdC stars (Geballe et al. 2009). The dust-driven He I wind is thus only a tracer of recent (few years to decades) dust-formation.

4. 16O/18O: Analysis and results

4.1. Analysis

In this section we constrain the 16O/18O ratios of the five newly discovered dLHdC stars that show 12C16O and 12C18O bands, the previously known dLHdC stars HD 137613 and HD 182040 and 33 RCB stars that show the CO bands. Figure 3 shows the first four overtone CO bandheads of the seven dLHdC stars and six representative RCB stars. Even from visual inspection of the spectra, it is evident that in general the 12C18O absorptions are stronger in dLHdC stars than RCB stars. Here, we describe our procedure to derive constraints on the 16O/18O ratios from these and other spectra.

thumbnail Fig. 3.

Zoom-in of the 12C16O and 12C18O absorption bands of all seven dLHdC stars (left) and six representative RCB stars (right) from our sample. All dLHdC stars except A166 show strong 12C18O absorption, comparable in strength to the 12C16O absorption. The RCB stars show a range of 12C18O absorption strengths. Shown are two RCB stars where the 12C18O absorption is strongest, two where it is of intermediate strength and two where the absorption is weak. Of the 33 RCB stars analyzed here, only five have strong dLHdC-like 12C18O absorption, while the remaining majority has intermediate or weak 12C18O absorption. We also plot examples of synthetic spectra that fit the observed spectra as red dashed lines. Note that the RCB star EROS2-CG-RCB4 shows 13C16O absorption bands (right panel).

We follow the procedure described in Karambelkar et al. (2021). We first generated synthetic spectra using a grid of hydrogen-deficient spherically symmetric MARCS (Model Atmospheres in Radiative and Convective Scheme) atmospheric models with input compositions characteristic of RCB and dLHdC stars (log ϵ(H) = 7.5, log ϵ(He) = 11.5, log and C/He = 0.01, Gustafsson et al. 1975, 2008; Bell et al. 1976; Plez 2008). The models assumed a solar metallicity to derive compositions of other elements, a total mass of 1 M⊙, ξ = 5 km s−1 and surface gravity log g = 1.0. We generated synthetic spectra using the package TURBOSPECTRUM (Alvarez & Plez 1998) and line lists from Goorvitch (1994), Plez (priv. commun. and described in Hedrosa et al. 2013) and Yurchenko et al. (2018) for CO, CN and C2 molecules respectively. We varied the effective temperatures from 4000 to 7500 K in intervals of 250 K. The C, N and O abundances are relevant to measure the oxygen isotope ratios. However, we cannot measure these abundances from our low to medium resolution spectra directly. We assumed a C/He ratio of 0.01 (Asplund et al. 2000), corresponding to log ϵ(C) = 9.5. We also fixed the total oxygen abundance to log ϵ(O) = 8.8 (consistent with the range of observed O abundances for RCB stars Asplund et al. 2000), and only varied the oxygen isotope ratios to vary the relative contribution of 16O and 18O. We chose 16O/18O values of 0.01, 0.05, 0.1, 0.2, 0.5, 1, 2, 5, 10, 20, 50, 500 and infinity (no 18O). Finally, to partially mitigate the effects of fixing the abundances, we varied the nitrogen abundances, choosing values of log ϵ(N) = 7.0, 7.5, 8.0, 8.5 and 9.4 (Asplund et al. 2000) Additionally, for RCB stars, the warm dust shell can contribute significantly to the K-band flux (up to 80%, Tisserand et al. 2013), veiling the absorption bands. To account for this, we introduced an additional parameter f dust = F shell , K F total , K $ f_{\mathrm{dust}} = \frac{F_{\mathrm{shell,K}}}{F_{\mathrm{total,K}}} $, and varied it between 0 and 0.8 in steps of 0.1 for RCB stars. We do not expect this to be a significant effect for dLHdC stars as they do not have infrared dust excesses. We fit the synthetic spectra to continuum-normalised NIR spectra and visually examine each of the fits to determine the range of isotope ratios that is consistent with the observed spectra. We are only able to place wide range estimates on the oxygen isotope ratios owing to the several free parameters and assumptions in our fitting. Future higher resolution spectra will determine the CNO abundances precisely and provide precise oxygen isotope ratios for the stars presented here. The derived values of 16O/18O are listed in Table 2, and Figs. 4, 5 and 6 show some examples of the model fits.

thumbnail Fig. 4.

Model fits to the CO absorption bandheads of the dLHdC star A166 and RCB stars ASAS-RCB-17, WXCrA and ASAS-RCB-7. The spectra (with errorbars) are plotted in black, and the best-fit models for a wide allowed range of model parameters (Table 2) are plotted as solid colored lines. The measured 16O/18O for these stars are > 20, 5 − 50, 0.05 − 1 and > 500 respectively. The dotted and dashed lines show models with oxygen isotope ratios outside these derived ranges (16O/18O = 10 for A166, 1 and 500 for ASAS-RCB-17, and 5 for WXCrA). It is challenging to distinguish between models with different N abundances given the medium resolution and S/N of the spectra. We focus on the regions around the 12C18O and 12C16O bandheads (dashed green and gray lines) and report the range of values consistent with the observed spectra.

thumbnail Fig. 5.

Zoom-in of the spectra of the RCB star WXCrA (solid black line). We plot the best fit-model synthetic spectrum to this star (gray, dashed line), and show the contribution of 12C14N (solid red), 12C16O blue and 12C18O green. The top, middle and bottom panels show different spectral regions.

thumbnail Fig. 6.

Zoom-in of the spectra of the RCB star ASAS-RCB-7 (solid black line). We plot the best fit-model synthetic spectrum to this star (gray, dashed line), and show the contribution of 12C14N (solid red), 12C16O blue and 12C18O green. Unlike WXCrA (Fig. 5), this spectrum does not show very strong 12C18O bandheads. As our oxygen ratio measurements are based on these bandheads, we estimate 16O/18O≥500 for this source. It is possible that the low spectroscopic resolution and S/N of this spectrum cannot resolve the blended weak bandhead. Higher resolution spectra are required to measure the ratio accurately. The other targets with 16O/18O≥500 could also be affected by a similar issue.

Table 2.

Range of model parameters that best fit the observed spectra.

As noted in García-Hernández et al. (2009), it is challenging to measure the 16O/18O ratios accurately from medium resolution spectra. The 12C16O and 12C18O absorption bands can be contaminated by absorption from other molecules such as 12C14N and C2. The absorption lines of CN and C2 are densely packed in the K band. It is also possible that the 12C16O absorption bands are saturated, resulting in a smaller measured 16O/18O ratio than the true value. There is also a degeneracy between the effects of the effective temperatures and nitrogen abundances on the depths of the CO absorption bands. For these reasons, we are not able to tightly constrain 16O/18O, but instead provide a range of values combining the effects of the temperatures, nitrogen abundances and fdust. Wherever possible, we use estimates of temperature ranges from the literature to tighten the constraints. Despite the wide range of ratios for each star, our measurements strongly suggest that most dLHdC stars have significantly lower 16O/18O than RCB stars.

4.2. Results

4.2.1. dLHdC stars

With the exception of A166, the depths of the 12C18O absorption bands in the dLHdC stars are comparable to those of the 12C16O bands (see Fig. 3). We use temperature estimates from the V − I color calibration of the HdC spectral classification system found in Crawford et al. (in prep.). From comparisons to synthetic spectra we derive 16O/18O ratios in the ranges 0.05–0.2 for B42, C38 and B566, 0.1–0.5 for A223, HD 182040 and HD 137613. Our derived values for HD 137613 and HD 182040 are consistent with those reported by Clayton et al. (2007) from medium resolution and García-Hernández et al. (2009) from high resolution spectra.

For B42, none of the spectral models fit the CO features at 2.349 μm and 2.352 μ. This could be because the 2.352 μm 12C16O band is saturated. Fits to the 2.378 and 2.383 μm bands suggest 16O/18O = 0.05–0.2, but the only models that fit these bands have a low nitrogen abundance of log ϵ(N) = 7.0. The optical spectrum of B42 shows strong CN bands (see Tisserand et al. 2022), suggesting that the nitrogen abundance may not be so low. It is possible that the 2.383 μm 12C16O band in B42 is also saturated, and the value we report is lower than the true 16O/18O. If this is the case, our reported uncertainty on the 16O/18O of B42 is also likely underestimated, as it does not include the systematic uncertainty associated with the saturated bands and the nitrogen abundances. Higher resolution observations are necessary to measure the true nitrogen abundance and the true 16O/18O value of B42.

Unlike the other dLHdC stars, A166 shows significantly weaker 12C18O absorptions than the 12C16O absorption; we constrain 16O/18O > 20 for this star. Tisserand et al. (2022) note that A166 is an outlier among the newly discovered dLHdC stars. In the HR diagram, it is located near two cool RCB stars and is distant from most dLHdC stars. It also has IR excesses in the WISE W3 and W4 bands. This, together with its high, RCB-like value of 16O/18O and possible broad He I absorption suggests that it is an RCB star. However, the lack of IR excesses in the NIR and WISE W1 and W2 bands indicates that it is in a phase of low dust production, similar to XX Cam (see Sect. 4.6 of Tisserand et al. 2022).

4.2.2. RCB stars

Of the 33 RCB stars that show the CO overtone absorption bands, estimates of effective temperatures are available for 19 (Crawford et al., in prep.). We are able to derive the tightest constraints on 16O/18O for them. For most of the remaining 12, we can only derive lower limits on 16O/18O.

Most of the RCB stars in our sample have much weaker 12C18O absorption bands relative to 12C16O, than dLHdC stars. A total of 28 of the 33 RCB stars have 16O/18O > 1. Eight of them do not show any 12C18O bandheads in our medium resolution spectra, and are consistent with the 16O/18O ⪆ 500 models. Figure 6 shows a zoom-in of the spectrum of one such star. However, it is possible that these measurements are not accurate for the following reasons. First, it is possible that the resolution and S/N of these spectra is not sufficient to resolve a weak 12C18O bandhead from the 12C14N bands (that are possibly enhanced by the low effective temperatures) in this region of these cold stars. Second, it is possible that the strong CO bandheads in the spectra of these cold RCB stars are saturated. As our medium resolution spectra cannot resolve the weaker, unsaturated CO lines, we cannot accurately measure their oxygen isotope ratios. Third, it is also possible that the spectra of these cold stars suffer strong dust dilution, making it difficult to detect the diluted 12C18O bandheads at medium resolution. We caution that a combination of these resolution, saturation and dilution effects could be preventing the detection of 12C18O bands in our spectra of these cold stars. Higher resolution spectra are required to obtain reliable measurements of oxygen isotope ratios in these stars, which we highlight in Table 2. However, it is unlikely that they have very strong dLHdC-like 12C18O bands. Thus, the medium resolution spectra suggest that a large majority of RCB stars have higher 16O/18O than dLHdC stars.

Five RCB stars – HD 175893, ASAS-RCB-11, IRAS 18135.5-2419, WX CrA and EROS2-CG-RCB-6 – have prominent 12C18O absorption features and are consistent with 16O/18O < 1. We derive 16O/18O values of 0.01–0.2, 0.01–0.2, 0.5–5, 0.05–1 and 0.2–5 for these stars, respectively. These values are similar to those of dLHdC stars. We note that our derived 16O/18O values for HD 175893, S Aps, SV Sge, ES Aql, U Aqr, Z Umi and WX Cra agree with previous medium resolution measurements from Clayton et al. (2007). Our derived values also agree with high resolution measurements for all of these except S Aps, SV Sge and ES Aql, as the 12C16O bandheads are saturated in these three (García-Hernández et al. 2010). Finally, the RCB star EROS2-CG-RCB-4 shows the 13C16O absorption bandhead at 2.345, 2.374, 2.404 and 2.434 μm (see Fig. 3).

To summarize, we find that most dLHdC stars have 16O/18O < 1, lower than most RCB stars. However, there is an overlap – a small fraction of dLHdC stars have 16O/18O > 1 while a small fraction of RCB stars have 16O/18O < 1. We illustrate this in Fig. 7. Finally, we emphasize that high resolution spectra are required to confirm the oxygen isotope ratio measurements, especially in the case of stars that have low effective temperatures. High-resolution spectra will help identify the stars for which the bandheads are saturated, and also provide accurate O and N abundance measurements which are crucial for precise 16O/18O measurements. These observations should help confirm our conclusion that most dLHdC stars have smaller 16O/18O than most RCB stars.

thumbnail Fig. 7.

Cumulative distribution plot showing the fraction of dLHdC and RCB stars that have 16O/18O below a given value. Of the dLHdC stars, 85% (6 out of 7) have 16O/18O < 0.5 (black dashed line) while only ≈15% of RCB stars (5 out of 33) are consistent with such low 16O/18O values. Most dLHdC stars have a lower 16O/18O than most RCB stars. A166 is the only dLHdC star that has 16O/18O > 1. This star is an outlier amongst the newly discovered dLHdC stars, and is likely an RCB star in a low phase of dust production (see Sec. 4.1).

5. Discussion

Our NIR spectroscopic observations have revealed that in addition to dust-formation, dLHdC and RCB stars can in most cases be distinguished based on their values of 16O/18O. dLHdC stars in general have a lower 16O/18O than RCB stars. It is not surprising that the oxygen isotope ratios are an important factor in the study of dLHdC and RCB stars. Anomalously low 16O/18O in dLHdC and RCB stars was key to identifying the merger of a He-core and a CO-core white dwarf as their formation channel.

18O is synthesized by the partial helium burning reaction 14N(α,γ) F18(β+ν)18O. This reaction is efficient at temperatures of ≈108 K (Clayton et al. 2007; Jeffery et al. 2011). At higher temperatures, 18O is burnt to 22Ne. These conditions can be achieved in a thin helium burning shell around the merger remnant of a He-core and a CO-core white dwarf (Clayton et al. 2007). The 18O is convectively dredged up to the surface of the star within the first few hundred years after the white dwarf merger (Crawford et al. 2020; Munson et al. 2021; Lauer et al. 2019). The photospheric value of 16O/18O in dLHdC and RCB stars is thus set within the first hundred years and remains constant for the rest of their lifetimes (≈104 − 5 years). Here, we explore the properties of the merging white dwarfs that set the values of 16O/18O in the remnant supergiants. We also discuss the implications of our observations on the dLHdC–RCB connection.

5.1. The origin of different 16O/18O for dLHdC and RCB stars

Studies have just begun to examine the quantities that affect 16O/18O in white dwarf merger remnants. The effects on 16O/18O of the helium burning shell temperature, convective extent of the supergiant envelope and the amount of hydrogen in the shell have been explored (Crawford et al. 2020; Munson et al. 2021). Their values in turn depend on the properties of the merging He-core and CO-core white dwarfs, such as their masses, mass ratios and compositions. The different values of 16O/18O in dLHdC and RCB stars thus suggest a correlation between the progenitor white dwarfs’ properties and dLHdC/RCB formation, which we discuss here. Note however that there are several additional factors that could contribute to the oxygen isotope ratios and remain to be modeled. A large unmodeled source of uncertainty is the effect of the initial composition of the progenitor white dwarfs and post-merger remnant on 16O/18O. The correlations that we discuss below are based on current understanding and are hence preliminary. Nevertheless, they demonstrate that different properties of merging white dwarfs can result in different values of 16O/18O, possibly providing a natural explanation for dLHdC vs. RCB formation. Additional modeling will help establish robust relations between progenitor properties and dLHdC–RCB stars.

Crawford et al. (2020) explored the effect of the temperature of the helium burning shell (THe) on 16O/18O. In particular, 16O/18O drops below unity for shell temperatures in the range THe ≈ 2.5 − 3.5 × 108 K. Thus, dLHdC stars could be associated with WD merger remnants that have THe in this range. For shell temperatures outside this range, 16O/18O increases to RCB-like values, suggesting that RCB stars either have THe < 2.5 × 108 K or THe > 3.5 × 108 K.

From simulations of white dwarf mergers, Staff et al. (2012) found that THe is inversely correlated with the mass ratio (q = MHe/MCO, where MHe and MCO are the masses of the He white dwarf and CO white dwarf, respectively). They derive values of THe = 3, 2.5 and 1.5 × 108 K for q = 0.5, 0.6 and 0.7 respectively. The helium shell temperatures of dLHdC stars are consistent with mass ratios 0.5 < q < 0.6. RCB stars are consistent either with q > 0.6 or q < 0.5 depending on whether they have THe < 2.5 × 108 K or THe > 3.5 × 108 K, respectively. However the Staff et al. (2012) analysis does not include an important factor – additional energy contributions from nucleosynthesis in the shell. This additional energy will increase the shell temperature, and thus the above estimates of THe are likely lower limits, and the q ranges are unrealistic. No studies of temperature increase due to nucleosynthesis exist in literature.

To first order, we can estimate the temperature increase timescale assuming triple-α burning is the dominant nucleosynthetic energy source (Hansen et al. 2004). Assuming the shell parameters from Staff et al. (2012) and a 100% energy to temperature conversion, we find that the temperature of a q = 0.7 merger with initial THe, init = 1.5 × 108 K doubles within the first few years after merger. This time reduces to a few days for q = 0.6 (THe, init = 2 × 108 K) and a few minutes for q = 0.5 (THe, init = 3 × 108 K). The shell-burning temperatures for q < 0.6 mergers can thus exceed 4 × 108 K. RCB stars would then be consistent with low mass-ratio (q < 0.6) mergers and dLHdC stars would be consistent with higher mass-ratio mergers (q > 0.7)4. More detailed studies of nucleosynthetic energy production are required to determine the exact ranges of q that might form RCB and dLHdC stars

THe also depends on the total mass of the white dwarf merger (Staff et al. 2018); however, the exact dependence has not been studied extensively. The Staff et al. (2012) estimates mentioned above assumed Mtot ≈ 0.9 M⊙. Without including nucleosynthetic energy, Staff et al. (2018) found that a Mtot = 0.7 M⊙, q = 0.5 merger has THe < 2 × 108 K – lower than the estimate of 3 × 108 K for the 0.9 M⊙, q = 0.5 merger. This suggests that lower mass mergers have lower THe than higher mass mergers. Adding nucleosynthetic energy can increase THe for the 0.7 M⊙ merger to the dLHdC range (2.5 − 3.5 × 108 K) and the 0.9 M⊙ merger outside this range. dLHdC stars would then be consistent with arising from lower mass mergers than RCB stars. This would be consistent with the observation in Tisserand et al. (2022) that the population of dLHdC stars has lower luminosities than RCB stars, which they interpret as a consequence of dLHdC stars originating from lower mass mergers than RCB stars.

In addition to THe, 16O/18O can also depend on the extent of the convective envelope and the mass of hydrogen in the helium burning shell (Zhang et al. 2014; Munson et al. 2021). Munson et al. explored the effect of convective overshoot factor (f) on 16O/18O. They found that for f < 0.07, 16O/18O ≈ 1, but increases rapidly for f > 0.07. Large values of f correspond to cases where the convective envelope extends all the way to the CO core and dredges up additional 16O from the core into the envelope. The high values of 16O/18O in RCB stars could be partly explained due to such a deep convective dredge up. That will also reduce C/O in RCB stars compared to dLHdC stars. The initial conditions of a merging system that would give rise to a deep convective zone remain to be explored. Munson et al. (2021) also showed that 16O/18O increases with increasing mass of hydrogen in the helium-burning shell. The presence of hydrogen leads to a decrease of 18O and increase in 16O due to proton capture reactions. The source of this hydrogen is a thin hydrogen envelope around the progenitor He white dwarf (Zhang et al. 2014). The mass of the hydrogen envelope is inversely correlated with the mass of the He white dwarf (Staff et al. 2012; Driebe et al. 1998). However, a significant fraction of this hydrogen is expected to be burned during the merger. How much of the envelope hydrogen actually reaches the helium shell is not known.

In conclusion, the 16O/18O values in white dwarf merger remnants are affected by the properties of the merging white dwarfs such as their total mass, mass ratios and individual compositions. In this context, the different values of 16O/18O in dLHdC and RCB stars indicate that they are formed from distinct populations of progenitor white dwarf binary mergers.

5.2. The evolutionary link between dLHdC and RCB stars

García-Hernández et al. (2010) speculated the possibility of an evolutionary link between dLHdC and RCB stars. This was based on their surface abundances, together with the assumptions that dLHdC → RCB (→ Extreme Helium Stars) is the common sequence and that their surface abundances change with evolution. They note that depending on the masses and compositions of the merging white dwarfs, the merger remnant will have different initial temperatures on the supergiant track. They associate dLHdC stars with the cooler parts of the supergiant track and RCB stars with hotter temperatures. In their model, depending on merger conditions the merger may first form a cold dLHdC star and eventually evolve to an RCB star, or directly form an RCB star.

The new observations, if corroborated by high-resolution spectroscopy, would show that the picture of a dLHdC-RCB evolutionary link could be incorrect. First, several newly discovered dLHdC stars have higher photospheric temperatures than many RCB stars (Tisserand et al. 2022). Second, the spectra presented and analyzed here demonstrate that dLHdC stars in general may have different values of 16O/18O than RCB stars (Sect. 4.1). In addition, recent theoretical modeling by Crawford et al. (2020) suggest that the surface abundances of dLHdC and RCB stars remain constant throughout their lifetimes; that is, it would not be possible for one class to evolve into another.

Instead, we propose that whether a white dwarf merger forms a dLHdC or an RCB depends solely on the properties of the merging white dwarfs. The merger forms a dLHdC star (no significant dust formation, low 16O/18O) or an RCB star (dust formation, high 16O/18O) based on the mass ratios, masses and compositions of the white dwarfs. In this picture, it is still a mystery why RCB stars form large amounts of dust while dLHdC stars do not. It seems reasonable to expect that only those white dwarf mergers that are more massive than a certain threshold or have particular ranges of chemical compositions form remnants that can undergo dust formation. Future theoretical models and studies that more precisely identify the differences between dLHdC and RCB star progenitors could provide an answer to this question.

5.3. The overlap between dLHdC and RCB stars

An intriguing observation is the five RCB stars (classified based on their observed IR excesses and brightness declines) that show low, dLHdC-like 16O/18O. These “overlap” stars may be a consequence of there being several different factors that contribute to 16O/18O. For example, they could arise from typical RCB-like mass-ratio mergers but with a smaller total mass, smaller convective envelope or a smaller hydrogen content than typical. Another interesting possibility is that some dLHdC stars undergo a short-lived, dust forming phase. During this phase, these stars would be observed as dust-forming RCB-like stars with low 16O/18O – similar to the five RCB stars mentioned above. This scenario does not contradict the observation that dLHdC stars do not show significant excesses in the mid-IR WISE bands. Tisserand et al. 2022 show that the emission from a circumstellar dust shell ejected from the surface of a star shifts from the mid-IR to longer wavelengths within a few decades. Thus, even if some dLHdC stars experienced dust formation ∼10 years before their WISE observations, they would not show any excesses in the WISE bands. The dust around them would have cooled to < 50 K (see Montiel et al. 2018) and would be observable only at very long wavelengths. Only one dLHdC star has been observed at far-infrared (FIR) wavelengths to date – HD 173409 which did not show any FIR excess (Montiel et al. 2018). Far-infrared and sub-mm observations of the newly identified dLHdC stars will help confirm or rule out the presence of a cold dust shell around them.

Finally, we note that four newly discovered dLHdC stars (A166, C526, F75 and F152) show definite signs of dust formation from their mid-IR colors and light curves (see Sect. 4.5 of Tisserand et al. 2022). Of these, A166 is an outlier within the dLHdC population (see Sect. 4.2.1), has RCB-like 16O/18O and is likely an RCB star in a low dust-production phase. The three other stars are not outliers, but are too hot to sustain 12C16O and 12C18O in their photospheres so their 16O/18O cannot be constrained. Sub-mm searches for the colder circumstellar medium of these stars can determine whether they have dLHdC-like or RCB-like 16O/18O values. Additional studies of such “overlap” stars are necessary to understand the dust-formation connection between dLHdC and RCB stars.

6. Summary and way forward

18O has been key to understanding the origins of the enigmatic dLHdC and RCB stars. The anomalously large 18O abundance in them can be explained by invoking a He-core and a CO-core white dwarf merger model for their formation. Although there were indications that RCB and dLHdC stars have different values of 16O/18O, the limited sample of dLHdC stars prevented a quantitative comparison. In this paper, we have utilized the revolutionary discovery of 27 new Galactic dLHdC stars to revisit this question.

We analyzed NIR spectra of 24 dLHdC stars together with unpublished spectra of 49 RCB stars. Owing to the several free parameters and assumptions in our fitting methodology, we obtained wide range estimates of the 16O/18O ratios for 7 dLHdC and 33 RCB stars whose spectra contain the 12C16O and/or 12C18O bands. We find that six of the seven dLHdC stars have 16O/18O < 0.5, while 28 of the 33 RCB stars have 16O/18O > 1. Thus, the wide range estimates on the 16O/18O ratios (although more uncertain in the coolest RCBs) obtained from the medium resolution spectra analyzed here suggest that most dLHdC stars have lower 16O/18O than most RCB stars. If corroborated by future high resolution spectroscopic observations, this will be the first established chemical difference between the two classes of HdC stars. It remains to be seen if and how the lower 16O/18O can be related to the lack of dust formation in dLHdC stars.

The different oxygen isotope ratios suggest that there is no evolutionary link between the class of dLHdC and RCB stars. Instead, this observation is consistent with the picture that dLHdC and RCB stars are formed from merging white dwarfs with distinct masses, mass ratios and compositions. Further theoretical studies are required to accurately determine the properties of white dwarfs that merge to form dLHdC versus RCB stars. A small number of RCB stars have uncharacteristically low, dLHdC-like 16O/18O values. This could be a consequence of multiple white dwarf properties that can affect the value of 16O/18O in the merger product, or can be explained by a short-lived dust formation phase in dLHdC stars. Theoretical models will test the former scenario, while FIR and sub-mm observations will confirm or rule out the latter. Further investigations of these “overlap” stars will shed light on the dLHdC-RCB dust formation mystery.

Future higher resolution NIR spectroscopy of the newly discovered dLHdC stars with CO bands will allow accurate determinations of their 16O/18O ratios and will help validate the results presented here. High resolution optical spectroscopy should allow the determination of accurate fluorine abundances in them. As fluorine is also a signature of a white dwarf merger (Pandey et al. 2007), these observations will help determine the properties of progenitors of the hot dLHdC stars that do not show CO overtone bands. High resolution spectroscopy also will potentially identify additional chemical differences between dLHdC and RCB stars (e.g., between their H and N abundances as indicated in Tisserand et al. 2022). and shed further light on their progenitor white dwarf populations.

1

Historically, dLHdC stars were referred to as HdC stars. Here, we follow the updated nomenclature of Tisserand et al. (2022).

2

Note that the star HD 175893 was originally classified as a dLHdC star but is now known to be an RCB star based on an IR excess found by Tisserand (2012).

3

See here for a list of NIR H lines.

4

Note that for mergers where THe > 4.5 × 108 K, other elemental abundances predicted by models do not agree with the observed RCB abundances (Crawford et al. 2020). This suggests that there is a lower limit on the mass-ratios of white dwarf binaries that can produce RCB/dLHdC stars.

Acknowledgments

We thank the anonymous referee for helpful comments that improved the quality of this paper. We thank Bradley Munson for useful comments and discussions. PT acknowledges financial support from “Programme National de Physique Stellaire” (PNPS) of CNRS/INSU, France. MMK acknowledges the Heising-Simons foundation for support via a Scialog fellowship of the Research Corporation. MMK acknowledges generous support from the David and Lucille Packard Foundation. SA acknowledges support from the GROWTH PIRE grant 1545949. GC and CC are grateful for support from National Science Foundation Award 1814967. This research is based in part on observations for programs GS-2005B-Q-20, GN-2011A-Q-112, GS-2015B-FT-1 and GS-2016B-FT-6 obtained at the international Gemini Observatory, a program of NSF’s NOIRLab, which is managed by the Association of Universities for Research in Astronomy (AURA) under a cooperative agreement with the National Science Foundation. on behalf of the Gemini Observatory partnership: the National Science Foundation (United States), National Research Council (Canada), Agencia Nacional de Investigación y Desarrollo (Chile), Ministerio de Ciencia, Tecnología e Innovación (Argentina), Ministério da Ciência, Tecnologia, Inovações e Comunicações (Brazil), and Korea Astronomy and Space Science Institute (Republic of Korea).

References

  1. Alvarez, R., & Plez, B. 1998, A&A, 330, 1109 [NASA ADS] [Google Scholar]
  2. Asplund, M., Gustafsson, B., Lambert, D. L., & Rao, N. K. 2000, A&A, 353, 287 [NASA ADS] [Google Scholar]
  3. Bell, R. A., Eriksson, K., Gustafsson, B., & Nordlund, A. 1976, A&AS, 23, 37 [NASA ADS] [Google Scholar]
  4. Clayton, G. C. 1996, PASP, 108, 225 [NASA ADS] [CrossRef] [Google Scholar]
  5. Clayton, G. C. 2012, J. Am. Assoc. Variable Star Obs., 40, 539 [NASA ADS] [Google Scholar]
  6. Clayton, G. C., Geballe, T. R., & Bianchi, L. 2003, ApJ, 595, 412 [NASA ADS] [CrossRef] [Google Scholar]
  7. Clayton, G. C., Herwig, F., Geballe, T. R., et al. 2005, ApJ, 623, L141 [NASA ADS] [CrossRef] [Google Scholar]
  8. Clayton, G. C., Geballe, T. R., Herwig, F., Fryer, C., & Asplund, M. 2007, ApJ, 662, 1220 [NASA ADS] [CrossRef] [Google Scholar]
  9. Clayton, G. C., Sugerman, B. E. K., Stanford, S. A., et al. 2011, ApJ, 743, 44 [NASA ADS] [CrossRef] [Google Scholar]
  10. Crawford, C. L., Clayton, G. C., Munson, B., Chatzopoulos, E., & Frank, J. 2020, MNRAS, 498, 2912 [NASA ADS] [CrossRef] [Google Scholar]
  11. Cushing, M. C., Vacca, W. D., & Rayner, J. T. 2004, PASP, 116, 362 [NASA ADS] [CrossRef] [Google Scholar]
  12. Driebe, T., Schoenberner, D., Bloecker, T., & Herwig, F. 1998, A&A, 339, 123 [Google Scholar]
  13. Eikenberry, S. S., Elston, R., Raines, S. N., et al. 2004, SPIE Conf. Ser., 5492, 1196 [NASA ADS] [Google Scholar]
  14. Elias, J. H., Rodgers, B., Joyce, R. R., et al. 2006, in SPIE Conf. Ser., 6269, 626914 [NASA ADS] [Google Scholar]
  15. Feast, M. W. 1996, ASP Conf. Ser., 96, 3 [NASA ADS] [Google Scholar]
  16. Fryer, C. L., & Diehl, S. 2008, ASP Conf. Ser., 391, 335 [NASA ADS] [Google Scholar]
  17. García-Hernández, D. A., Hinkle, K. H., Lambert, D. L., & Eriksson, K. 2009, ApJ, 696, 1733 [CrossRef] [Google Scholar]
  18. García-Hernández, D. A., Lambert, D. L., Kameswara Rao, N., Hinkle, K. H., & Eriksson, K. 2010, ApJ, 714, 144 [CrossRef] [Google Scholar]
  19. Geballe, T. R., Rao, N. K., & Clayton, G. C. 2009, ApJ, 698, 735 [NASA ADS] [CrossRef] [Google Scholar]
  20. Goorvitch, D. 1994, ApJS, 95, 535 [NASA ADS] [CrossRef] [Google Scholar]
  21. Goswami, A., Karinkuzhi, D., & Shantikumar, N. S. 2010, ApJ, 723, L238 [NASA ADS] [CrossRef] [Google Scholar]
  22. Gustafsson, B., Bell, R. A., Eriksson, K., & Nordlund, A. 1975, A&A, 42, 407 [NASA ADS] [Google Scholar]
  23. Gustafsson, B., Edvardsson, B., Eriksson, K., et al. 2008, A&A, 486, 951 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  24. Hansen, C. J., Kawaler, S. D., & Trimble, V. 2004, Stellar Energy Sources (New York: Springer), 271 [Google Scholar]
  25. Hedrosa, R. P., Abia, C., Busso, M., et al. 2013, ApJ, 768, L11 [NASA ADS] [CrossRef] [Google Scholar]
  26. Herter, T. L., Henderson, C. P., Wilson, J. C., et al. 2008, Proc. SPIE, 7014, 70140X [NASA ADS] [CrossRef] [Google Scholar]
  27. Iben, I., J., & Tutukov, A. V. 1984, ApJS, 54, 335 [NASA ADS] [CrossRef] [Google Scholar]
  28. Jeffery, C. S., Karakas, A. I., & Saio, H. 2011, MNRAS, 414, 3599 [NASA ADS] [CrossRef] [Google Scholar]
  29. Karambelkar, V. R., Kasliwal, M. M., Tisserand, P., et al. 2021, ApJ, 910, 132 [NASA ADS] [CrossRef] [Google Scholar]
  30. Lambert, D. L., & Rao, N. K. 1994, J. Astrophys. Astron., 15, 47 [NASA ADS] [CrossRef] [Google Scholar]
  31. Lauer, A., Chatzopoulos, E., Clayton, G. C., Frank, J., & Marcello, D. C. 2019, MNRAS, 488, 438 [NASA ADS] [CrossRef] [Google Scholar]
  32. Montiel, E. J., Clayton, G. C., Sugerman, B. E. K., et al. 2018, AJ, 156, 148 [NASA ADS] [CrossRef] [Google Scholar]
  33. Munson, B., Chatzopoulos, E., Frank, J., et al. 2021, ApJ, 911, 103 [NASA ADS] [CrossRef] [Google Scholar]
  34. Pandey, G., Lambert, D. L., & Kameswara Rao, N. 2007, ArXiv e-prints [arXiv:0712.3971] [Google Scholar]
  35. Plez, B. 2008, Phys. Scripta, Vol. T, 133, 014003 [NASA ADS] [CrossRef] [Google Scholar]
  36. Ramsay Howat, S. K., Todd, S., Leggett, S., et al. 2004, SPIE Conf. Ser., 5492, 1160 [NASA ADS] [Google Scholar]
  37. Rayner, J. T., Toomey, D. W., Onaka, P. M., et al. 2003, PASP, 115, 362 [NASA ADS] [CrossRef] [Google Scholar]
  38. Saio, H., & Jeffery, C. S. 2002, MNRAS, 333, 121 [CrossRef] [Google Scholar]
  39. Staff, J. E., Menon, A., Herwig, F., et al. 2012, ApJ, 757, 76 [Google Scholar]
  40. Staff, J. E., Wiggins, B., Marcello, D., et al. 2018, ApJ, 862, 74 [NASA ADS] [CrossRef] [Google Scholar]
  41. Tisserand, P. 2012, A&A, 539, A51 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  42. Tisserand, P., Clayton, G. C., Welch, D. L., et al. 2013, A&A, 551, A77 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  43. Tisserand, P., Clayton, G. C., Bessell, M. S., et al. 2020, A&A, 635, A14 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  44. Tisserand, P., Crawford, C. L., Clayton, G. C., et al. 2022, A&A, 667, A83 [NASA ADS] [CrossRef] [EDP Sciences] [Google Scholar]
  45. Vacca, W. D., Cushing, M. C., & Rayner, J. T. 2003, PASP, 115, 389 [NASA ADS] [CrossRef] [Google Scholar]
  46. Warner, B. 1967, MNRAS, 137, 119 [NASA ADS] [CrossRef] [Google Scholar]
  47. Webbink, R. F. 1984, ApJ, 277, 355 [NASA ADS] [CrossRef] [Google Scholar]
  48. Yurchenko, S. N., Szabó, I., Pyatenko, E., & Tennyson, J. 2018, MNRAS, 480, 3397 [NASA ADS] [CrossRef] [Google Scholar]
  49. Zhang, J., Wang, X., Mazzali, P. A., et al. 2014, ApJ, 797, 5 [NASA ADS] [CrossRef] [Google Scholar]

All Tables

Table 1.

Log of spectroscopic observations.

In the text
Table 2.

Range of model parameters that best fit the observed spectra.

In the text

All Figures

thumbnail Fig. 1.

Near-infrared spectra of the newly discovered dLHdC stars. Near-infrared spectra of the RCB star NSV11154 and the previously known dLHdC star HD137613 are also shown for comparison. The rest-frame positions of the CN and 12C16O absorption bands are indicated by dashed vertical lines.
In the text
thumbnail Fig. 2.

Zoom-in of the spectral region around the He I (10 833 Å) triplet. The velocity is measured with respect to 1.0833 μm. The same spectral region of He I profile of an RCB star is shown in red – RCB-ToI-223 (a.k.a. WISE J182010.96−193453.4), at a phase when it has just emerged from a dust decline. We cannot resolve the contribution from the Si I line to any He I features from our low resolution spectra. However, it is evident that we do not detect any strong (> 200 km s−1), RCB-like winds in any dLHdC star, suggesting that there are no dust-driven winds around them. This is consistent with no hot dust excesses seen in the SEDs of any of these stars. A166 is the only dLHdC star that shows a possible high velocity He I absorption profile (plotted in blue).
In the text
thumbnail Fig. 3.

Zoom-in of the 12C16O and 12C18O absorption bands of all seven dLHdC stars (left) and six representative RCB stars (right) from our sample. All dLHdC stars except A166 show strong 12C18O absorption, comparable in strength to the 12C16O absorption. The RCB stars show a range of 12C18O absorption strengths. Shown are two RCB stars where the 12C18O absorption is strongest, two where it is of intermediate strength and two where the absorption is weak. Of the 33 RCB stars analyzed here, only five have strong dLHdC-like 12C18O absorption, while the remaining majority has intermediate or weak 12C18O absorption. We also plot examples of synthetic spectra that fit the observed spectra as red dashed lines. Note that the RCB star EROS2-CG-RCB4 shows 13C16O absorption bands (right panel).
In the text
thumbnail Fig. 4.

Model fits to the CO absorption bandheads of the dLHdC star A166 and RCB stars ASAS-RCB-17, WXCrA and ASAS-RCB-7. The spectra (with errorbars) are plotted in black, and the best-fit models for a wide allowed range of model parameters (Table 2) are plotted as solid colored lines. The measured 16O/18O for these stars are > 20, 5 − 50, 0.05 − 1 and > 500 respectively. The dotted and dashed lines show models with oxygen isotope ratios outside these derived ranges (16O/18O = 10 for A166, 1 and 500 for ASAS-RCB-17, and 5 for WXCrA). It is challenging to distinguish between models with different N abundances given the medium resolution and S/N of the spectra. We focus on the regions around the 12C18O and 12C16O bandheads (dashed green and gray lines) and report the range of values consistent with the observed spectra.
In the text
thumbnail Fig. 5.

Zoom-in of the spectra of the RCB star WXCrA (solid black line). We plot the best fit-model synthetic spectrum to this star (gray, dashed line), and show the contribution of 12C14N (solid red), 12C16O blue and 12C18O green. The top, middle and bottom panels show different spectral regions.
In the text
thumbnail Fig. 6.

Zoom-in of the spectra of the RCB star ASAS-RCB-7 (solid black line). We plot the best fit-model synthetic spectrum to this star (gray, dashed line), and show the contribution of 12C14N (solid red), 12C16O blue and 12C18O green. Unlike WXCrA (Fig. 5), this spectrum does not show very strong 12C18O bandheads. As our oxygen ratio measurements are based on these bandheads, we estimate 16O/18O≥500 for this source. It is possible that the low spectroscopic resolution and S/N of this spectrum cannot resolve the blended weak bandhead. Higher resolution spectra are required to measure the ratio accurately. The other targets with 16O/18O≥500 could also be affected by a similar issue.
In the text
thumbnail Fig. 7.

Cumulative distribution plot showing the fraction of dLHdC and RCB stars that have 16O/18O below a given value. Of the dLHdC stars, 85% (6 out of 7) have 16O/18O < 0.5 (black dashed line) while only ≈15% of RCB stars (5 out of 33) are consistent with such low 16O/18O values. Most dLHdC stars have a lower 16O/18O than most RCB stars. A166 is the only dLHdC star that has 16O/18O > 1. This star is an outlier amongst the newly discovered dLHdC stars, and is likely an RCB star in a low phase of dust production (see Sec. 4.1).
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.