Two hot pre-white dwarfs inside the red-giant-branch planetary nebula Pa 13 -- Double core evolution or common envelope-induced rejuvenation?

Abstract

Close binary central stars of PNe offer a unique window for investigating the conditions immediately following the ejection of a common envelope (CE). Double eclipsing and double-lined double systems are particularly valuable as they provide minimally model-dependent constraints on fundamental binary parameters. We report that the nucleus of Pa13 (P=0.3988d) belongs to this rare class of systems and present a comprehensive analysis of its double-degenerate binary. We performed a two-component NLTE spectral analysis based on phase-resolved X-Shooter spectroscopy, multi-band light-curve modeling, SED fitting, as well as a kinematic analysis. Both stars are found to be hot pre-white dwarfs, with Star1 being cooler but larger (Teff=50kK, R=0.40Rsol) than Star2 (Teff=75kK, R=0.16Rsol). The weakness of spectral lines of Star2 made both the atmospheric and RV analyses challenging, and we uncovered a strong sensitivity of the assumed surface ratio to its derived RV curve. Yet, the RV curve and Kiel mass of Star1 (M1=0.41+/-0.02Msol) could be determined precisely, allowing for a dynamical mass determination of Star2 (M2=0.39+/-0.04Msol). We uncovered that Pa13 exhibits a small but significant orbital eccentricity (e=0.02+/-0.01), making it only the second post-CE binary nucleus with a measured eccentricity. We conclude that Pa13 provides hitherto the strongest evidence that PNe can be observed around post-RGB stars. Immediately after the CE-ejection, Star1 likely still filled its Roche lobe, suggesting that Pa13 is a more evolved, detached descendant of over-contact double-degenerate systems such as Hen2-428. Since the mass ratio of Pa13 is close to unity the system may have formed through double-core CE evolution. Alternatively, there must exist an efficient CE-induced rejuvenation mechanism capable of reheating the cool white dwarf in the binary, as already indicated by Hen2-428. (abbreviated)

Figures (7)

• Figure 1: Two examples X-Shooter spectra (gray) taken close to maximum RV separation, respectively. Light gray regions indicate the location of a weak residual nebular line, which have been excluded from the fit. The red and blue lines represent the contribution of Star 1 and Star 2, respectively, and the dashed, purple line shows the combined fit combined fit. The upper panel show the predicted line profiles assuming a surface ratio of 0.17, while the lower panel the line profiles assuming a surface ratio of 0.23. Applying a consistent approach to all spectra, the resulting derived RV semi-amplitudes of Star 2 are $K_2=143$ km/s and $K_2=78$ km/s for an assumed surface ratio of 0.17 and 0.23, respectively.
• Figure 2: Kiel diagram, showing the position of the two CSs of Pa 13 compared to H-shell burning post-AGB tracks (black lines) from MillerBertolami2016 and post-RGB tracks (dashed, gray lines) from Hall+2013.
• Figure 3: The top three panels show our best fit eccentric PHOEBE model (black) compared to the observed ZTF g- (green) and r-band (red) and SARA i-band (brown) light curves. The gray line indicates our best model that does not account for eccentricity in the fit. The bottom panel shows the RVs of Star 1 (red crosses) and Star 2 (blue crosses) as measured assuming a surface ratio of 0.17. The formal fitting error bars for the assumed surface ratio (0.17) are smaller than the symbol sizes. The red and the pink lines indicated the best fit to the RV curve of Star 1 from the eccentric and circular model, respectively. For Star 2 the predicted RV curves from our best fit eccentric/circular model are indicated by the dashed blue/cyan lines. The gray dashed-dotted line indicates the system velocity.
• Figure 4: Two-component spectral energy distribution fit. Top panel: Filter-averaged fluxes converted from observed magnitudes are shown in different colors ( Gaia: blue, Skymapper: red, UHS: pink). The light red and blue lines correspond to the flux contribution of Star 1 and Star 2, respectively, and the gray line is the combined best fit to the observation. The model fluxes are degraded to a spectral resolution of $6\,\AA$. To reduce the steep SED slope, the flux is multiplied by the wavelength cubed. Bottom panel: Uncertainty-scaled difference between synthetic and observed magnitudes.
• Figure 5: Toomre diagram showing the location of Pa 13. The solid and dashed lines indicate the one- and two-sigma contours, respectively, of the U, V, and W velocity distributions of main-sequence stars from Kordopatis2011. The gray, green, and blue lines indicate the velocity distribution of thin disk, thick disk, and Galactic halo stars, respectively.