A radial velocity survey for post-common-envelope Wolf–Rayet central stars of planetary nebulae: first results and discovery of the close binary nucleus of NGC 5189

MNRAS 448, 1789–1806 (2015) doi:10.1093/mnras/stv074 A radial velocity survey for post-common-envelope Wolf–Rayet central stars of planetary nebulae: first results and discovery of the close binary  nucleus of NGC 5189 Rajeev Manick,1,2,3 † Brent Miszalski1,4 and Vanessa McBride1,2 1 South African Astronomical Observatory, PO Box 9, Observatory 7935, South Africa Cosmology and Gravity Centre, Department of Astronomy, University of Cape Town, Private Bag X3, Rondebosch 7701, South Africa 3 Instituut voor Sterrenkunde, KU Leuven, Celestijnenlaan 200D bus 2401, B-3001 Leuven, Belgium 4 Southern African Large Telescope Foundation, PO Box 9, Observatory 7935, South Africa 2 Astrophysics, ABSTRACT The formation of Wolf–Rayet central stars of planetary nebulae ([WR] CSPNe) whose spectroscopic appearance mimics massive WR stars remains poorly understood. Least understood is the nature and frequency of binary companions to [WR] CSPNe that may explain their H-deficiency. We have conducted a systematic radial velocity (RV) study of six [WR] CSPNe to search for post-common-envelope (post-CE) [WR] binaries. We used a cross-correlation method to construct the RV time series as successfully done for massive close binary WR stars. No significant RV variability was detected for the late-[WC] type nuclei of Hen 2-113, Hen 3-1333, PMR 2 and Hen 2-99. Significant, large-amplitude variability was found in the [WC4] nucleus of NGC 5315. In the [WO1] nucleus of NGC 5189, we discovered significant periodic variability that reveals a close binary with Porb = 4.04 ± 0.1 d. We measured a semi-amplitude of 62.3 ± 1.3 km s−1 that gives a companion mass of m2 ≥ 0.5 M or m2 = 0.84 M (assuming i = 45◦ ). The most plausible companion type is a massive white dwarf (WD) as found in Fleming 1. The spectacular nebular morphology of NGC 5189 fits the pattern of recently discovered post-CE PNe extremely well with its dominant low-ionization structures (e.g. as in NGC 6326) and collimated outflows (e.g. as in Fleming 1). The long 4.04 d orbital period is either anomalous (e.g. NGC 2346) or it may indicate that there is a sizeable population of [WR] binaries with massive WD companions in relatively wide orbits, perhaps influenced by interactions with the strong [WR] wind. Key words: stars: AGB and post-AGB – binaries: spectroscopic – stars: Wolf–Rayet – planetary nebulae: general – planetary nebulae: individual: NGC 5189. 1 I N T RO D U C T I O N Planetary nebulae (PNe) are circumstellar gas envelopes ejected at the end of the asymptotic giant branch (AGB) phase of lowintermediate mass stars of ∼1 M to 8 M . A majority of central stars of planetary nebulae (CSPNe) show hydrogen-rich atmospheres (e.g. Méndez et al. 1988; Méndez 1991; Napiwotzki & Schoenberner 1995), with their spectra containing only weak absorption lines, mainly of hydrogen (H) and helium (He). Perhaps the least understood are the less numerous emission-line CSPNe  Based on observations made with the South African Astronomical Observatory (SAAO) 1.9 m telescope and the Southern African Large Telescope (SALT) under programme 2013-2-RSA-005 (PI: Miszalski). † E-mail: with H-deficient atmospheres whose spectral appearance mimics the massive Wolf–Rayet (WR) stars. These CSPNe exhibit dense and strong stellar winds with mass-loss rates up to ∼10−7 M yr−1 (e.g. Koesterke 2001; Crowther 2008). Their classification is similar to massive WR stars where the dominant emission lines indicate either a [WC] (He, C and O; Crowther, De Marco & Barlow 1998) or [WN] (He and N; Smith, Shara & Moffat 1996) subtype, where the brackets around the spectral type distinguish them from their massive counterparts. In contrast to the roughly equal division between WN and WC subtypes in massive WR stars (van der Hucht 2001), the known examples of [WR] stars are heavily skewed towards early and late [WC] types with few intermediate [WC] types (Crowther 2008; see however Górny 2014), whereas only two bona fide [WN] CSPNe were established recently in IC 4663 (Miszalski et al. 2012) and A 48 (Todt et al. 2013; Frew et al. 2014). Comprehending this disparity may shed much needed light on how [WR] stars form.  C 2015 The Authors Published by Oxford University Press on behalf of the Royal Astronomical Society Accepted 2015 January 13. Received 2014 December 18; in original form 2014 October 10 1790 R. Manick, B. Miszalski and V. McBride These scenarios cannot reproduce the moderate level of ∼55 per cent He by mass in the possible intermediate [WN/WC] CSPN of PB 8 (Todt et al. 2010). Miller Bertolami et al. (2011) proposed a diffusion-induced nova scenario to explain the abundance pattern of PB 8; however, these models require low metallicities and core masses to reach the extremely He rich compositions of [WN] CSPNe (see the discussion in Todt et al. 2013). While it seems that low metallicities may facilitate [WR] formation (Zijlstra et al. 2006; Kniazev et al. 2008), the α-element nebular abundances of both IC 4663 and A 48 are notably solar, suggesting that the Miller Bertolami et al. (2011) scenario may not be applicable. 1 MNRAS 448, 1789–1806 (2015) Several different possibilities for binary interactions were proposed by De Marco & Soker (2002) and De Marco (2008); however, the paucity of observed binary systems makes it difficult to discern to what degree these or some other scenarios may play a role. In this work, we aim to address this by searching for companions in a sample of [WR] CSPNe. It is more practical with modest amounts of telescope time to search for close binaries where the orbital periods are a few days or less, signifying that a common-envelope (CE) interaction has produced the observed short orbital periods (Iben & Livio 1993; Ivanova et al. 2013). Only two previous post-CE examples have been discovered, namely SDSS J212531.92−010745.9 that consists of a PG 1159 star with an M dwarf companion (Nagel et al. 2006; Schuh, Beeck & Nagel 2009) and the [WC7] CSPN of PN G222.8−04.2 with an undetermined companion type (Hajduk, Zijlstra & Gesicki 2010). Their literature orbital periods are 0.29 d and 0.63 or 1.26 d, respectively. The majority of close binary CSPNe have been found via photometry monitoring including PN G222.8−04.2.2 There are however several limitations with this approach. A notable exception is the double degenerate nucleus of Fleming 1 (Boffin et al. 2012), discovered from radial velocity (RV) monitoring. The companion in Fleming 1 is a more massive white dwarf (WD) whose presence is not detectable in the optical light curve. At least 37 unique earlytype [WC], PG 1159 and O(He) CSPNe have been subjected to photometric monitoring campaigns to search for non-radial pulsations (Ciardullo & Bond 1996; González Pérez, Solheim & Kamben 2006); however, no close binaries have been discovered as part of these campaigns. If ∼15–20 per cent of PNe are expected to host close binaries (Bond 2000; Miszalski et al. 2009a), then (...truncated)