SALT HRS discovery of the binary nucleus of the Etched Hourglass Nebula MyCn 18
Brent Miszalski, Rajeev Manick, Joanna Mikołajewska, Hans Van Winckel, Krystian Iłkiewicz
PPublications of the Astronomical Society of Australia (PASA)doi: 10.1017/pas.2018.xxx.
SALT HRS discovery of the binary nucleus of theEtched Hourglass Nebula MyCn 18 ∗ Brent Miszalski , † , Rajeev Manick , Joanna Mikołajewska , Hans Van Winckel and KrystianIłkiewicz South African Astronomical Observatory, PO Box 9, Observatory, 7935, South Africa Southern African Large Telescope Foundation, PO Box 9, Observatory, 7935, South Africa Instituut voor Sterrenkunde, KU Leuven, Celestijnenlaan 200D bus 2401, B-3001 Leuven, Belgium Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences, Bartycka 18, PL-00716 Warsaw, Poland
Abstract
The shaping of various morphological features of planetary nebulae (PNe) is increasingly linked tothe role of binary central stars. Identifying a binary within a PN offers a powerful tool with which todirectly investigate the formation mechanisms behind these features. The Etched Hourglass Nebula,MyCn 18, is the archetype for several binary-linked morphological features, yet it has no identifiedbinary nucleus. It has the fastest jets seen in a PN of 630 km s − , a central star position offset from thenebula centre, and a bipolar nebula with a very narrow waist. Here we report on the Southern AfricanLarge Telescope (SALT) High Resolution Spectrograph (HRS) detection of radial velocity variability inthe nucleus of MyCn 18 with an orbital period of 18 . ± .
04 days and a semi-amplitude of 11 . ± . − . Adopting an orbital inclination of 38 ± . ± . M (cid:12) yields asecondary mass of 0 . ± . M (cid:12) corresponding to an M5V companion. The detached nature of thebinary rules out a classical nova (CN) as the origin of the jets or the offset central star as hypothesisedin the literature. Furthermore, scenarios that produce the offset central star during the AGB and thatform narrow waist bipolar nebulae result in orbital separations 80–800 times larger than observed inMyCn 18. The inner hourglass and jets may have formed from part of the common envelope ejecta thatremained bound to the binary system in a circumbinary disk, whereas the offset central star positionmay best be explained by proper motion. Detailed simulations of MyCn 18 are encouraged that arecompatible with the binary nucleus to further investigate its complex formation history. Keywords: planetary nebulae: individual: MyCn 18 – planetary nebulae: general – binaries: spectroscopic –techniques: radial velocites – stars: AGB and post-AGB
At least 1 in 5 planetary nebulae host a close binarycentral star with an orbital period of ∼ ∗ Based on observations made with the Southern African LargeTelescope (SALT) under programmes 2016-2-SCI-034 and 2017-1-MLT-010. † E-mail: [email protected] whether the small number of discoveries made so far(Van Winckel et al. 2014; Jones et al. 2017; Miszalski etal. 2018) signify the existence of a substantial populationof intermediate period binaries. We are systematicallysearching for this population (Miszalski et al. 2018) withthe High Resolution Spectrograph (HRS, Bramall etal. 2010, 2012; Crause et al. 2014) on the SouthernAfrican Large Telescope (SALT, Buckley et al. 2006;O’Donoghue et al. 2006). Further details of the scientificmotivation behind the survey can be found in Miszalskiet al. (2018).Included in our survey is the young planetary nebulaMyCn 18 (PN G307.5 − a r X i v : . [ a s t r o - ph . S R ] M a y Miszalski et al.
Figure 1.
Hubble Space Telescope colour-composite image ofMyCn 18 (Sahai et al. 1999) made from F658N (red), F656N(green) and F502N (blue) filters. Image credit: Raghvendra Sahaiand John Trauger (JPL), the WFPC2 science team, and NASA.
O’Connor et al. 2000; Clyne et al. 2014). MyCn 18 ex-hibits several features for which binary interactions arethe preferred formation mechanism, yet no evidence hasyet been found supporting the presence of a binary com-panion. These features include polar outflows or jets(Bryce et al. 1997; O’Connor et al. 2000), a central starposition offset up to 0.2” from the geometric centre andits bipolar morphology with a narrow waist (Sahai et al.1999; Clyne et al. 2014). MyCn 18 has often been com-pared against bipolar nebulae around low and high massevolved stars. It is particularly notable for its intriguingnested hourglass structure (Sahai et al. 1999; Clyne etal. 2014) that closely resembles the nebular remnantof SN 1987A (e.g. Burrows et al. 1995; Sugerman etal. 2015), the symbiotic nebula Hen 2-104 (Corradi &Schwarz 1993; Corradi et al. 2001; Santander-García etal. 2008; Clyne et al. 2015) and Hb 12 which may bea young PN or possibly a symbiotic nebula (Hsia et al.2006; Kwok & Hsia 2007; Vaytet et al. 2009; Clark etal. 2014). There are also strong similarities with bipo-lar and hourglass nebulae around several luminous bluevariables (e.g. García-Segura et al. 1999; Smith et al.2007; Gvaramadze & Menten 2012; Taylor et al. 2014).Based on these rich connections to several bipolar nebu-lae, resolving the unknown binary status of MyCn 18 isan important step towards further understanding howbipolar nebulae form.This paper is structured as follows. Section 2 presentsSALT HRS radial velocity monitoring observations of MyCn 18. We clarify the spectral class of the centralstar in Sect. 3.1, followed by a description of our radialvelocity measurements in Sect. 3.2. We demonstrate thesignificant periodic variability of these measurements inSect. 3.3 that prove the binary nature of MyCn 18. Sect.3.3 also derives the orbital parameters of the binarynucleus. We discuss our results in Sect. 4 in the contextof formation scenarios concerning MyCn 18 and otherpost-CE PNe. We conclude in Sect. 5.
Multiple observations of the nucleus of MyCn 18 weretaken with SALT HRS (Bramall et al. 2010, 2012; Crauseet al. 2014), a dual-beam, fibre-fed échelle spectrographenclosed in a vacuum tank within an insulated, temper-ature controlled enclosure in the spectrograph room ofSALT (Buckley et al. 2006; O’Donoghue et al. 2006).The medium resolution (MR) mode of HRS was usedto obtain spectra covering 3700–8900 Å with resolvingpowers R = λ/ ∆ λ of 43000 and 40000 for the blue andred arms, respectively. During an observation a secondfibre, separated at least 20 arcsec from the science fibre,simultaneously observes the sky spectrum. Both objectand sky spectra are interleaved on the separate blue (2kx 4k) and red (4k x 4k) CCDs. Regular bias, ThAr arclamp and quartz lamp flat field calibrations are takenas part of SALT operations.Table 1 presents a log of the 26 HRS MR observationsof MyCn 18 in addition to the radial velocity measure-ments described in Sect. 3.2. Basic processing of the datawas performed by p y s a lt (Crawford et al. 2010) be-fore a pipeline based on the m i da s packages e c h e l l e (Ballester 1992) and f e ro s (Stahl et al. 1999) devel-oped by A. Y. Kniazev (Kniazev et al. 2016) reduced thedata and distributed the data products. The blue spec-tra which do not feature prominent sky emission lineswere not sky subtracted and the red arm spectra weresky subtracted. The small angular size of MyCn 18 andthe fibre fed design of HRS precluded the subtractionof nearby nebula emission. The order-merged spectrawere converted to a logarithmic wavelength scale using i r a f before heliocentric radial velocity corrections wereadded using the v e l s e t task of the rv s ao package(Kurtz & Mink 1998). The spectral classification of the central star of MyCn 18is unclear in the literature. An Of(H) classification wasmade based on an unpublished spectrum (Méndez 1991)and more recently a rare H-deficient Of(C) classificationwas made by Lee et al. (2007) that requires a spectrum‘dominated by strong C emissions’ (Méndez 1991). We yCn 18yCn 18
Multiple observations of the nucleus of MyCn 18 weretaken with SALT HRS (Bramall et al. 2010, 2012; Crauseet al. 2014), a dual-beam, fibre-fed échelle spectrographenclosed in a vacuum tank within an insulated, temper-ature controlled enclosure in the spectrograph room ofSALT (Buckley et al. 2006; O’Donoghue et al. 2006).The medium resolution (MR) mode of HRS was usedto obtain spectra covering 3700–8900 Å with resolvingpowers R = λ/ ∆ λ of 43000 and 40000 for the blue andred arms, respectively. During an observation a secondfibre, separated at least 20 arcsec from the science fibre,simultaneously observes the sky spectrum. Both objectand sky spectra are interleaved on the separate blue (2kx 4k) and red (4k x 4k) CCDs. Regular bias, ThAr arclamp and quartz lamp flat field calibrations are takenas part of SALT operations.Table 1 presents a log of the 26 HRS MR observationsof MyCn 18 in addition to the radial velocity measure-ments described in Sect. 3.2. Basic processing of the datawas performed by p y s a lt (Crawford et al. 2010) be-fore a pipeline based on the m i da s packages e c h e l l e (Ballester 1992) and f e ro s (Stahl et al. 1999) devel-oped by A. Y. Kniazev (Kniazev et al. 2016) reduced thedata and distributed the data products. The blue spec-tra which do not feature prominent sky emission lineswere not sky subtracted and the red arm spectra weresky subtracted. The small angular size of MyCn 18 andthe fibre fed design of HRS precluded the subtractionof nearby nebula emission. The order-merged spectrawere converted to a logarithmic wavelength scale using i r a f before heliocentric radial velocity corrections wereadded using the v e l s e t task of the rv s ao package(Kurtz & Mink 1998). The spectral classification of the central star of MyCn 18is unclear in the literature. An Of(H) classification wasmade based on an unpublished spectrum (Méndez 1991)and more recently a rare H-deficient Of(C) classificationwas made by Lee et al. (2007) that requires a spectrum‘dominated by strong C emissions’ (Méndez 1991). We yCn 18yCn 18 Table 1
Observation log of SALT HRS spectra of MyCn 18 and radial velocity measurements. The Julian day represents themidpoint of each exposure and the radial velocity measurements were made from stellar N III λ λ Julian day Exposure time Orbital Phase RV (N III) RV (He I)(s) (km s − ) (km s − )2457847.41564 3000 0.26 − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± − ± Miszalski et al. created an average blue and red spectrum after shiftingall HRS spectra to the rest frame of N III λ i r a f and the rv s ao package (Kurtz & Mink 1998). Rel-evant portions are displayed in Fig. 2 which includesa Gaussian fit to He II λ ± l m f i t package (Newville et al. 2016). The data do notsupport an Of(C) classification as the only prominent Cemission features belong to C IV λ λ λ < λ − . ± . − ), whereas supporting an Of-WR classification thereis no absorption at H γ (possibly with a red wing ofemission) and blueshifted He II λ −
50 km s − ( − . ± . − ). We favour an Of(H)classification based on the FWHM of He II λ T eff = 50 ±
10 kK as He IInebular emission and He I absorption lines are all absentfrom the HRS spectra.
The radial velocity measurements in Table 1 were deter-mined from the stellar emission feature N III λ λ l m f i t package (Newville et al. 2016) was used to fit modelsconsisting of a Voigt function and a straight line to N IIIand a Gaussian function to He I. The fits to N III aredisplayed in Fig. 3. Uncertainties in each radial veloc-ity measurement are 1 σ uncertainties determined fromthe standard error in the fit centroid provided by thefitting routine. The average nebular radial velocity of − . ± .
10 km s − agrees well with −
71 km s − measured from spatio-kinematic modeling (Clyne et al.2014). The SALT HRS radial velocity measurements of thenucleus of MyCn 18 exhibit significant variability com-pared to the stationary nebular emission (Tab. 1). Periodanalysis of the radial velocities using a Lomb-Scargleperiodogram (Press et al. 1992) reveals the strongestpeak to have a period of 18.15 d, significant at the 5 σ level, while other peaks correspond to the 1 − f and 1+ f aliases of this peak (Fig. 4). The 18.15 d period was usedas the basis for a Keplerian orbit model that was builtusing a least-squares minimization method applied tothe phase-folded data. Figure 4 displays the Keplerianorbit fit overlaid on the radial velocity measurements intime and folded with the orbital period. The observedsinusoidal radial velocity shifts prove the binary natureof MyCn 18.Table 2 presents the orbital parameters of the Keple-rian orbit fit determined using Monte Carlo simulationsin which the eccentricity was fixed to be zero (for detailssee Miszalski et al. 2018). The systemic velocity of the or-bit ( V hel = − . ± .
24 km s − ; V LSR = − . ± . − ) agrees well with the nebular radial velocityof −
71 km s − (Clyne et al. 2014). This suggests theN III emission line from the Of(H)-type primary is notsignificantly disturbed by the stellar wind and tracesthe motion of the primary.The estimation of the secondary mass requires someassumptions concerning the orbital inclination i and theprimary mass M . Spatio-kinematic studies of post-CEPNe have demonstrated that the orbital inclination ofthe binary central star always matches the inclinationdetermined from analysis of the nebula (Hillwig et al.2016). The inclination of MyCn 18 is well determinedto be 38 deg by spatio-kinematic studies (O’Connor etal. 2000; Clyne et al. 2014), although no estimate ofthe uncertainty was provided by these studies. We havetherefore adopted a nominal error of 5 deg in the orbitalinclination.The primary mass M is more difficult to constrainwith the information available in the literature. Massesmay be estimated via comparison with post-AGB evolu-tionary tracks if the temperature and surface gravity orluminosity are known. Determining the surface gravity iscomplicated by the strong stellar wind which has likelycontaminated the absorption line profiles. We there-fore adopt an alternative distance-dependent approachthat estimates the luminosity of the central star follow-ing Sect. 9.4.5 of Frew (2008). We adopt a weightedmean distance of 3092 ±
507 pc, which incorporatesdistance estimates of 3342 ±
668 pc (Stanghellini &Haywood 2010) and 2750 ±
780 pc (Frew et al. 2016),along with an apparent magnitude m V = 14 . T eff = 50 ±
10 kK and a reddening of A V = 2 .
63 mag (Tsamis et al. 2003). We determine using yCn 18yCn 18
63 mag (Tsamis et al. 2003). We determine using yCn 18yCn 18 N III(a) (b) (c) (d) (e)
Wavelength (Å) I n t e n s i t y Figure 2.
Relevant portions of the average SALT HRS spectrum of the nucleus of MyCn 18. (a) He II λ λ λ λ γ λ Table 2
Orbital parameters of the binary nucleus of MyCn 18.
Orbital period P (d) 18 . ± . e K (km s − ) 11 . ± . γ (km s − ) − . ± . a sin i (AU) 0 . ± . f ( m ) ( M (cid:12) ) 0 . ± . T (d) 2457965 . ± . − ) 2.61Inclination i (degrees) 38 ± M ( M (cid:12) ) 0 . ± . M ( M (cid:12) ) 0 . ± . Miszalski et al.
Wavelength (Å) I n t e n s i t y Figure 3.
The observed N III λ −
71 kms − (Clyne et al. 2014). Each panel is labelled with the Julian day of each spectrum minus 2457000 days. yCn 18yCn 18
71 kms − (Clyne et al. 2014). Each panel is labelled with the Julian day of each spectrum minus 2457000 days. yCn 18yCn 18 A m p li t u d e Frequency (d −1 ) JD−2457000 (days) −90−85−80−75−70−65−60−55 R V ( k m s − ) R V ( k m s − ) Phase −16−80816 R e s ( k m s − ) Figure 4.
The binary nature of MyCn 18 revealed by SALT HRS. (Top panel)
The Lomb-Scargle periodogram of radial velocitymeasurements (top two segments). The strongest peak at a 5 σ significance level corresponds to the 18.15 d orbital period. The 1 − f and 1 + f aliases are also visible and the lower segment shows the window function. (Middle and Bottom panels) Radial velocitymeasurements displayed in time (middle) and folded with the orbital period (bottom). The solid lines represent the Keplerian orbit fitand the shaded region indicates the residuals are within 3 σ of the fit where σ = 2 .
61 km s − . Miszalski et al. equation 9.13 of Frew (2008) a bolometric correction of − . ± .
66 mag and an absolute bolometric magnitudeof − . ± .
77, where the error in the latter includes er-rors in the bolometric correction and absolute magnitudeadded in quadrature which are themselves dominated bythe uncertainties in the temperature and the distance, re-spectively. This results in log
L/L (cid:12) = 3 . ± .
31 which istypical for young central stars of PNe still on the horizon-tal part of post-AGB evolutionary tracks. Consideringthe uncertainties in this distance-dependent method, weadopt M = 0 . ± . M (cid:12) after making comparisonswith post-AGB evolutionary tracks (Miller Bertolamiet al. 2016). It is unlikely that the mass is above thisrange as the nebular chemical abundances of MyCn 18are not consistent with a massive AGB progenitor thathas a core mass above ∼ M (cid:12) (García-Hernández etal. 2016). This is dependent, however, on whether singlestar AGB models are applicable to post-CE PNe, whichis yet to be determined.The radial velocity semi-amplitude of K = 11 . ± . − , M = 0 . ± . M (cid:12) and i = 38 ± M = 0 . ± . M (cid:12) .According to the Teff-radius-mass relation for dwarfswith Z=0.014 (Bressan et al. 2012; Chen et al. 2014),the range of masses and temperatures give an M5Vcompanion with an uncertainty of one spectral class(Rajpurohit et al. 2013). The current orbital separationis 0.124 au or 26.6 R (cid:12) and the M5V companion hasa Roche lobe radius of ∼ R (cid:12) . Even if the radiuswere inflated by ∼ V >
21 mag (Coveyet al. 2007), explaining the lack of any secondary featuresin our spectra. A periodic photometric signal caused bythe irradiated atmosphere of the companion may bepresent with an amplitude of ∼ The system of collimated outflows or jets of MyCn 18 areremarkable as the fastest observed amongst PNe withdeprojected velocities of up to 630 km s − (Bryce et al.1997; O’Connor et al. 2000; Clyne et al. 2014). The jetswere ejected ∼ ∼ − M (cid:12) . O’Connor et al. (2000) suggested that since classicalnova (CN) ejecta have similar masses and kinematics, it may be that a CN was responsible for producing the jetsystem of MyCn 18. More recently, Soker & Kashi (2012)supported the assessment of O’Connor et al. (2000) andfurther commented that the kinetic energy of the jetswas too low to have been formed from an intermediateluminosity optical transient (ILOT) event. Clyne et al.(2014) further speculated on other nova-like formationevents, some of which include planetary bodies thatmay be destroyed. Apart from the knots demonstratingsimilar mass and velocities to CN ejecta, we emphasisethat these hypothetical CN or nova-like scenarios do nothave any other observational support. Despite this, thesescenarios appear to have been adopted as the de factoexplanation for jet formation in MyCn 18 in the litera-ture. As such, it is important to determine whether thesehypothetical scenarios are physically compatible withthe binary nucleus of MyCn 18. The observed orbitalparameters of the binary can be assumed unchangedsince the main nebula was ejected at the end of a CEinteraction phase. These parameters uniquely enable usto determine whether any CN event could have producedthe jet system since the main nebula was ejected.Classical novae occur in cataclysmic variables wherethe short orbital separation allows for mass transfer andaccumulation onto the white dwarf leading to a ther-monuclear explosion (Warner 1995). The time needed forthe WD to build up a critical layer of hydrogen dependsprimarily on the mass of the WD and the accretion rate.While models for novae on degenerate WDs in thermalequilibrium are well developed (e.g. Yaron et al. 2005),the same cannot be said for pre-WDs that are found inPNe. Supposing accretion and thermonuclear explosionscan occur on pre-WDs (which may also have strongwinds), which is currently unclear as no models haveexplored the topic, then in the following we examinewhether a CN could have launched the jets of MyCn 18.The strictest constraints come from the magnitudeof permissible mass transfer rates. In CNe, the pri-mary must have accreted a minimum amount of mass M env ∼ − –10 − M (cid:12) for a 0.6 M (cid:12) white dwarf (Yaronet al. 2005). Considering that a CN launched jet musthave accreted mass over the time span between the mainnebula and jet formation (only ∼ M env requires that theaverage accretion rate is an implausible ˙ M acc ∼ − –10 − M (cid:12) yr − that cannot be supplied from the M5Vcompanion. Even if the companion were an active M-dwarf, the accretion rate could only reach a maximumof ˙ M acc ∼ − M (cid:12) yr − during coronal mass ejections(Mullan 1996; Osten & Wolk 2015), falling far short ofthe required rate. The inability to achieve sufficientlyhigh accretion onto the primary from the companion Shara et al. (2010) explored very rare luminous red novae thatcan occur on low-mass WDs ( M WD (cid:46) . M (cid:12) ) accreting at lowrates ( ˙ M acc ∼ − –10 − M (cid:12) yr − ), however they must stillaccumulate M env ∼ − –10 − M (cid:12) to ignite a nova explosion. yCn 18yCn 18
21 mag (Coveyet al. 2007), explaining the lack of any secondary featuresin our spectra. A periodic photometric signal caused bythe irradiated atmosphere of the companion may bepresent with an amplitude of ∼ The system of collimated outflows or jets of MyCn 18 areremarkable as the fastest observed amongst PNe withdeprojected velocities of up to 630 km s − (Bryce et al.1997; O’Connor et al. 2000; Clyne et al. 2014). The jetswere ejected ∼ ∼ − M (cid:12) . O’Connor et al. (2000) suggested that since classicalnova (CN) ejecta have similar masses and kinematics, it may be that a CN was responsible for producing the jetsystem of MyCn 18. More recently, Soker & Kashi (2012)supported the assessment of O’Connor et al. (2000) andfurther commented that the kinetic energy of the jetswas too low to have been formed from an intermediateluminosity optical transient (ILOT) event. Clyne et al.(2014) further speculated on other nova-like formationevents, some of which include planetary bodies thatmay be destroyed. Apart from the knots demonstratingsimilar mass and velocities to CN ejecta, we emphasisethat these hypothetical CN or nova-like scenarios do nothave any other observational support. Despite this, thesescenarios appear to have been adopted as the de factoexplanation for jet formation in MyCn 18 in the litera-ture. As such, it is important to determine whether thesehypothetical scenarios are physically compatible withthe binary nucleus of MyCn 18. The observed orbitalparameters of the binary can be assumed unchangedsince the main nebula was ejected at the end of a CEinteraction phase. These parameters uniquely enable usto determine whether any CN event could have producedthe jet system since the main nebula was ejected.Classical novae occur in cataclysmic variables wherethe short orbital separation allows for mass transfer andaccumulation onto the white dwarf leading to a ther-monuclear explosion (Warner 1995). The time needed forthe WD to build up a critical layer of hydrogen dependsprimarily on the mass of the WD and the accretion rate.While models for novae on degenerate WDs in thermalequilibrium are well developed (e.g. Yaron et al. 2005),the same cannot be said for pre-WDs that are found inPNe. Supposing accretion and thermonuclear explosionscan occur on pre-WDs (which may also have strongwinds), which is currently unclear as no models haveexplored the topic, then in the following we examinewhether a CN could have launched the jets of MyCn 18.The strictest constraints come from the magnitudeof permissible mass transfer rates. In CNe, the pri-mary must have accreted a minimum amount of mass M env ∼ − –10 − M (cid:12) for a 0.6 M (cid:12) white dwarf (Yaronet al. 2005). Considering that a CN launched jet musthave accreted mass over the time span between the mainnebula and jet formation (only ∼ M env requires that theaverage accretion rate is an implausible ˙ M acc ∼ − –10 − M (cid:12) yr − that cannot be supplied from the M5Vcompanion. Even if the companion were an active M-dwarf, the accretion rate could only reach a maximumof ˙ M acc ∼ − M (cid:12) yr − during coronal mass ejections(Mullan 1996; Osten & Wolk 2015), falling far short ofthe required rate. The inability to achieve sufficientlyhigh accretion onto the primary from the companion Shara et al. (2010) explored very rare luminous red novae thatcan occur on low-mass WDs ( M WD (cid:46) . M (cid:12) ) accreting at lowrates ( ˙ M acc ∼ − –10 − M (cid:12) yr − ), however they must stillaccumulate M env ∼ − –10 − M (cid:12) to ignite a nova explosion. yCn 18yCn 18 More generally, it is unlikely that CNe are in a positionto produce jets in post-CE PNe. Jets are extremely rarein cataclysmic variables (Körding et al. 2008, 2011) andare unlikely to account for the large numbers of post-CEPNe with jets. Furthermore, CNe are rarely observed toappear in PNe (e.g. Wesson et al. 2008) and the mostpromising additional candidates have been reclassifiedas old nova shells (e.g. Miszalski et al. 2016; Shara et al.2017). In summary, while it is not completely unexpectedfor a CN to appear within a PN (e.g. Wesson et al. 2008),they are not expected to routinely form jets and are notcommon enough to play a significant role in producingjets of post-CE PNe.
The main hourglass nebula with its narrow waist wasejected about 2700 years ago (Clyne et al. 2014) likely atthe end of a CE interaction phase. Preferential depositionof material in the orbital plane during the CE phase likelyhad a defining influence in shaping the bipolar nebula(e.g. Sandquist et al. 1998; Ricker & Taam 2012; Passy etal. 2012). The inner hourglass nebula was ejected around950 years later and was closely followed by the jet system100 years later (Clyne et al. 2014). The close associationin time between the inner hourglass and the jets stronglysuggests they were formed from the same reservoir ofmatter. This matter would likely have been bound tothe binary (circumbinary) or either component of thebinary. As there could not have been any mass transferbetween the binary components (Sect. 3.3), the mostlikely source of this matter would be bound materialleft over from the CE phase (Passy et al. 2012; Ricker& Taam 2012), which simulations suggest can fallbackonto the binary or either component to form a disk ofa few ∼ M (cid:12) (Kashi & Soker 2011; Kuruwita et al.2016). The observed jet velocities (O’Connor et al. 2000;Clyne et al. 2014) are consistent with being produced bythis disk, whether the jets originate near the secondary ,the primary (Blackman & Lucchini 2014), or aroundboth stars. Another more speculative possibility is thata stellar or planetary tertiary companion provides thematter required (e.g. Clyne et al. 2014). Note that any accretion onto the primary would also behindered by the strong wind of the primary which can reach v ∞ ∼ − in Of-type central stars (Pauldrach et al.2004). An M5 dwarf has an escape velocity of ∼
582 km s − , assuming M = 0 . M (cid:12) and R = 0 . R (cid:12) (Bressan et al. 2012; Chen etal. 2014). We emphasise that such scenarios were developed prior toknowledge of the binary system we have discovered in MyCn 18.Any inclusion of a tertiary stellar or planetary companion information scenarios concerning MyCn 18 must reproduce the
The self-similarity of the inner hourglass with themain nebula (Clyne et al. 2014) suggests the accretedmatter may have formed a second CE (or mini-CE) thatwould likely lead to the formation of the inner hourglassand jets. Precisely how this process may have occurredin MyCn 18 remains to be determined with the aidof detailed simulations. The process may be similar tothat outlined by Soker (2017) in which the formationof a circumbinary disk plays a central role and the jetsfacilitate the removal of any remaining envelope.
The
Hubble Space Telescope imaging of MyCn 18 re-vealed the central star position to be offset from thegeometric centre of each nebula component (Fig. 1, Sa-hai et al. 1999; Clyne et al. 2014). The largest offsetof 0.2” occurs with respect to the inner hourglass andcorresponds to 618 ±
101 au at our adopted distance of3092 ±
507 pc. As with the high velocity jets, severalscenarios have been hypothesised to explain how thisoffset was produced in the time between the main nebulawas ejected and the jets were launched ( ∼ R (cid:12) , which is80–800 times smaller than the expected final separationrange of ∼ Perhaps the most promising explanation of the offsetcould be the proper motion. While Sahai et al. (1999)found this explanation problematic, we note that thedirection of the offset matches the direction of the propermotion (Fig. 5) which is µ ( α ) = − . ±− . − and µ ( δ ) = − . ± . − (Zachariaset al. 2017). This match alone justifies a more thor-ough investigation than available data permit and is observed binary system. As our observations do not allow usto determine whether a tertiary component is present, furtherdiscussion of the wide variety of such hypothetical scenarios iswell beyond the scope of this paper. It is unclear whether this can occur given the pre-WD natureof the primary which also exhibits strong winds. Miszalski et al.
Figure 5.
Hubble Space Telescope image of MyCn 18 taken withthe F N filter showing the direction of proper motion (greenarrow) of the central star (cyan star). The red line represents theminor axis. certainly beyond the scope of this paper. We suggestsuch a study would involve 3D smoothed particle hy-drodynamics (SPH) simulations that include a morecomprehensive prescription of the 3D geometry andproperties of the inner region of MyCn 18 as well ashigher quality proper motion data anticipated from the Gaia mission.
The 18.15 d orbital period of MyCn 18 lies betweenthe 16 d orbit of NGC 2346 and the 142 d orbit ofNGC 1360 (Miszalski et al. 2018). It is only the sixthbinary central star known with a measured orbital pe-riod above 10 days (Miszalski et al. 2018). Populationsynthesis models predict large numbers of multiple daypost-CE central stars (e.g. Nie et al. 2012), though it isunclear whether this population exists due to a historicaltendency towards photometric surveys to find binarycentral stars (Miszalski et al. 2018). We will address thisquestion empirically as our ongoing SALT HRS surveyprogresses, though we can say we have already discov-ered some new multiple day binaries. The multiple daysystems appear to be very rare in the similar populationof WD main-sequence binaries that has carefully studiedselection effects (Nebot Gómez-Morán et al. 2011). Ifthe CE phase operates similarly in central stars andWDMS binaries, then the initial SALT HRS discoveries Clyne et al. (2014) considered only two slit positions coveringthe inner region. Additional slit positions would be beneficial tounderstand the elusive nature of the inner rings identified by Sahaiet al. (1999). might suggest multiple day post-CE binaries are morecommon than previously thought. Individually, multipleday post-CE central stars may be used to improve ourunderstanding of CE population synthesis models (e.g.Davis et al. 2010) and simulations of the CE phase in-volving wider initial orbital separations (e.g. Iaconi etal. 2017).Whether longer orbital periods favour the formationof bipolar PNe like NGC 2346 and MyCn 18 remainsunclear and is currently biased by the small sample sizeof known binaries. The existence of canonical bipolarpost-CE PNe with much shorter orbital periods, e.g.M 2-19 ( P = 0 .
67 d, Miszalski et al. 2009) and Hen 2-428 ( P = 0 .
18 d, Santander-García et al. 2015), suggeststhere may be no simple correlation between orbital pe-riod and morphology. We also note that the often-citedformation scenario for bipolar PNe with narrow waistslike MyCn 18 produces final orbital separations of ∼ ) of 1.8 in MyCn 18. Sowicka et al.(2017) suggested that longer period systems tend tohave low measured ADFs based on NGC 5189 (Man-ick et al. 2015; García-Rojas et al. 2012) and IC 4776(Sowicka et al. 2017). This depends on the purported 9d orbital period of the central star of IC 4776 which isnot definitive. The orbital period could be much shorter(Sowicka et al. 2017) and more observations are requiredbefore IC 4776 can be meaningfully compared againstother post-CE PNe. We presented SALT HRS échelle observations of thenucleus of the PN MyCn 18, also known as the EtchedHourglass Nebula. Radial velocity measurements from26 spectra demonstrate a significant periodic variabilityof 18 . ± .
04 d. The data prove the presence of apost-CE binary nucleus in MyCn 18 which has longbeen suspected of being formed by a binary system.Several scenarios concerning the formation of MyCn 18in the literature are not compatible with the orbitalparameters of the binary nucleus. Our main conclusionsare as follows: yCn 18yCn 18
04 d. The data prove the presence of apost-CE binary nucleus in MyCn 18 which has longbeen suspected of being formed by a binary system.Several scenarios concerning the formation of MyCn 18in the literature are not compatible with the orbitalparameters of the binary nucleus. Our main conclusionsare as follows: yCn 18yCn 18 • The RV time series measured from the N III λ . ± .
04 d at a significance levelof 5 σ . The orbital period was used as the basis fora circular Keplerian orbit fit with a semi-amplitudeof 11 . ± . − and a systemic velocity of − . ± .
24 km s − . The latter is in good agree-ment with the nebula systemic velocity ( −
71 kms − , Clyne et al. 2014). Residuals of the Keplerianfit (2.61 km s − ) are relatively high because ofthe influence of stellar winds in the Of(H) primarywhose classification which we have clarified basedon a deep stacked spectrum. • Assuming a distance of 3092 ±
507 pc, we estimatethe luminosity of the primary to be log
L/L (cid:12) =3 . ± .
31. At T eff = 50 ±
10 kK, we estimatethe primary mass to be M = 0 . ± . M (cid:12) andadopt an orbital inclination of i = 38 ± M = 0 . ± . M (cid:12) corre-sponding to an M5 dwarf with an uncertainty ofone spectral class. • We rule out previous hypotheses that the jet systemof MyCn 18 (O’Connor et al. 2000) formed as theresult of a CN explosion. The orbital separation of0.124 au or 26.6 R (cid:12) is too large to allow for masstransfer via Roche-lobe overflow. Furthermore, anM5V secondary cannot provide the average accre-tion rate of ˙ M acc ∼ − –10 − M (cid:12) yr − required toaccrete the M env ∼ − –10 − M (cid:12) mass requiredto form a CN (Yaron et al. 2005) in the ∼ • An alternative formation scenario for MyCn 18 wasproposed whereby material from the CE ejectionthat formed the main nebula still bound to thebinary settles or falls back to envelop the binary(e.g. Kashi & Soker 2011; Kuruwita et al. 2016).This may have then lead to the formation of theinner hourglass and jets. The exact process in whichthis occurs remains uncertain and requires detailedsimulations, though it could resemble that describedby Soker (2017). There may be other mechanismsthat contribute to forming MyCn 18, but these mustbe compatible with the observed binary system. • We discussed the 0.2” offset the central star showswith respect to the centre of the inner hourglass.The accepted scenario for producing observable off-sets during the AGB (Soker et al. 1998) results inorbital separations 80–800 times larger than theobserved separation in MyCn 18. The orbital pa-rameters also rule out a CN or nova-like event toprovide a ‘kick’ to produce the observed offset (e.g.Sahai et al. 1999; Clyne et al. 2014). We favour the proper motion of MyCn 18 to explain the offset cen-tral star, especially since the offset is observed alongthe minor axis of MyCn 18 in the same direction asthe proper motion. Detailed 3D SPH simulationsof the motion of MyCn 18 that incorporates antici-pated
Gaia proper motion data are encouraged toexplore this possibility further. • The discovery of the binary nucleus of MyCn 18strengthens the long suspected link between bipo-lar nebulae and interacting binary stars (e.g. Soker& Rappaport 2000). Characterising binary starsin bipolar nebulae across a wide variety of stellarmasses and orbital separations may help shed lighton common physical mechanisms behind the mass-loss histories of bipolar nebulae (e.g. Sugerman etal. 2005). One such common mechanism may be CEevolution, given the strong resemblance betweenMyCn 18 (a post-CE PN) and the nebular rem-nant of SN 1987A (a post-CE merger, Morris &Podsiadlowski 2009).
This paper is based on spectroscopic observations made withthe Southern African Large Telescope (SALT) under jointSouth African-Polish programmes 2016-2-SCI-034 and 2017-1-MLT-010 (PI: B. Miszalski). Polish participation in SALTis funded by grant No. MNiSW DIR/WK/2016/07. We aregrateful to our SALT colleagues for maintaining the telescopefacilities and conducting the observations. B. M. acknowl-edges support from the National Research Foundation (NRF)of South Africa. This study has been supported in part by thePolish MNiSW grant 0136/DIA/2014/43, and NCN grantsDEC-2013/10/M/ST9/00086 and 2015/18/A/ST9/00746. H.V. W acknowledges support from the Research Council of theKU Leuven under grant number C14/17/082. We thank theanonymous referee for a helpful and constructive report. Thispaper also features observations made with the NASA/ESAHubble Space Telescope, and obtained from the HubbleLegacy Archive, which is a collaboration between the SpaceTelescope Science Institute (STScI/NASA), the Space Tele-scope European Coordinating Facility (ST-ECF/ESA) andthe Canadian Astronomy Data Centre (CADC/NRC/CSA).IRAF is distributed by the National Optical AstronomyObservatory, which is operated by the Association of Univer-sities for Research in Astronomy (AURA) under a coopera-tive agreement with the National Science Foundation. B. M.thanks S. Mohamed for discussions and A. Y. Kniazev formaking available his HRS pipeline data products.
REFERENCES
Afşar, M., & İbanoˇglu, C. 2008, MNRAS, 391, 802Ballester, P. 1992, European Southern Observatory Con-ference and Workshop Proceedings, 41, 1772
Miszalski et al.
Blackman, E. G., & Lucchini, S. 2014, MNRAS, 440,L16Bond, H. E. 2000, Asymmetrical Planetary Nebulae II:From Origins to Microstructures, 199, 115Bramall, D. G., Sharples, R., Tyas, L., et al. 2010, Proc.SPIE, 7735, 77354FBramall, D. G., Schmoll, J., Tyas, L. M. G., et al. 2012,Proc. SPIE, 8446, 84460ABressan, A., Marigo, P., Girardi, L., et al. 2012, MNRAS,427, 127Bryce, M., López, J. A., Holloway, A. J., & Meaburn, J.1997, ApJL, 487, L161Buckley, D. A. H., Swart, G. P., & Meiring, J. G. 2006,Proc. SPIE, 6267, 62670ZBurrows, C. J., Krist, J., Hester, J. J., et al. 1995, ApJ,452, 680Chen, Y., Girardi, L., Bressan, A., et al. 2014, MNRAS,444, 2525Clark, D. M., López, J. A., Edwards, M. L., & Winge,C. 2014, AJ, 148, 98Clyne, N., Redman, M. P., Lloyd, M., et al. 2014, A&A,569, A50Clyne, N., Akras, S., Steffen, W., et al. 2015, A&A, 582,A60Corradi, R. L. M., & Schwarz, H. E. 1993, A&A, 268,714Corradi, R. L. M., Livio, M., Balick, B., Munari, U., &Schwarz, H. E. 2001, ApJ, 553, 211Covey, K. R., Ivezić, Ž., Schlegel, D., et al. 2007, AJ,134, 2398Crause, L. A., Sharples, R. M., Bramall, D. G., et al.2014, Proc. SPIE, 9147, 91476TCrawford, S. M., Still, M., Schellart, P., et al. 2010, Proc.SPIE, 7737, 773725Davis, P. J., Kolb, U., & Willems, B. 2010, MNRAS,403, 179Dayal, A., Sahai, R., Watson, A. M., et al. 2000, AJ,119, 315De Marco, O., Hillwig, T. C., & Smith, A. J. 2008, AJ,136, 323De Marco, O. 2009, PASP, 121, 316Frew, D. J. 2008, Ph.D. Thesis, Macquarie UniversityFrew, D. J., Parker, Q. A., & Bojičić, I. S. 2016, MNRAS,455, 1459García-Díaz, M. T., Clark, D. M., López, J. A., Steffen,W., & Richer, M. G. 2009, ApJ, 699, 1633García-Hernández, D. A., Ventura, P., Delgado-Inglada,G., et al. 2016, MNRAS, 461, 542García-Rojas, J., Peña, M., Morisset, C., Mesa-Delgado,A., & Ruiz, M. T. 2012, A&A, 538, A54García-Segura, G., Langer, N., Różyczka, M., & Franco,J. 1999, ApJ, 517, 767Gleizes, F., Acker, A., & Stenholm, B. 1989, A&A, 222,237 Gvaramadze, V. V., & Menten, K. M. 2012, A&A, 541,A7Hillwig, T. C., Bond, H. E., Afşar, M., & De Marco, O.2010, AJ, 140, 319Hillwig, T. C., Jones, D., De Marco, O., et al. 2016, ApJ,832, 125Hsia, C. H., Ip, W. H., & Li, J. Z. 2006, AJ, 131, 3040Iaconi, R., Reichardt, T., Staff, J., et al. 2017, MNRAS,464, 4028Ivanova, N., Justham, S., Chen, X., et al. 2013, A&ARv,21, 59Jones, D., Van Winckel, H., Aller, A., Exter, K., & DeMarco, O. 2017, A&A, 600, L9Kashi, A., & Soker, N. 2011, MNRAS, 417, 1466Körding, E., Rupen, M., Knigge, C., et al. 2008, Science,320, 1318Körding, E. G., Knigge, C., Tzioumis, T., & Fender, R.2011, MNRAS, 418, L129Kniazev, A. Y., Gvaramadze, V. V., & Berdnikov, L. N.2016, MNRAS, 459, 3068Kurtz, M. J., & Mink, D. J. 1998, PASP, 110, 934Kuruwita, R. L., Staff, J., & De Marco, O. 2016, MNRAS,461, 486Kwok, S., & Hsia, C. H. 2007, ApJ, 660, 341Lee, T.-H., Stanghellini, L., Ferrario, L., & Wickramas-inghe, D. 2007, AJ, 133, 987Manick, R., Miszalski, B., & McBride, V. 2015, MNRAS,448, 1789Mayall, M. W., & Cannon, A. J. 1940, Harvard CollegeObservatory Bulletin, 913, 7Méndez, R. H., Herrero, A., & Manchado, A. 1990, A&A,229, 152Méndez, R. H. 1991, Evolution of Stars: the PhotosphericAbundance Connection, 145, 375Miller Bertolami, M. M. 2016, A&A, 588, A25Miszalski, B., Acker, A., Moffat, A. F. J., Parker, Q. A.,& Udalski, A. 2009, A&A, 496, 813Miszalski, B., Mikołajewska, J., Köppen, J., et al. 2011a,A&A, 528, A39Miszalski, B., Jones, D., Rodríguez-Gil, P., et al. 2011b,A&A, 531, A158Miszalski, B., Woudt, P. A., Littlefair, S. P., et al. 2016,MNRAS, 456, 633Miszalski, B., Manick, R., Mikołajewska, J., et al. 2018,MNRAS, 473, 2275Morris, T., & Podsiadlowski, P. 2009, MNRAS, 399, 515Mullan, D. J. 1996, Cool Stars, Stellar Systems, and theSun, 109, 461Nebot Gómez-Morán, A., Gänsicke, B. T., Schreiber,M. R., et al. 2011, A&A, 536, A43Newville, M., Stensitzki, T., Allen, D. B., et al. 2016,Astrophysics Source Code Library, ascl:1606.014Nie, J. D., Wood, P. R., & Nicholls, C. P. 2012, MNRAS,423, 2764 yCn 18yCn 18