ETHOS 1: a high-latitude planetary nebula with jets forged by a post-common-envelope binary central star* Free

Abstract

We report on the discovery of ETHOS 1 (PN G068.1+11.0), the first spectroscopically confirmed planetary nebula (PN) from a survey of the SuperCOSMOS Science Archive for high-latitude PNe. ETHOS 1 stands out as one of the few PNe to have both polar outflows (jets) travelling at 120 ± 10 km s⁻¹ and a close binary central star. The light curve observed with the Mercator Telescope reveals an orbital period of 0.535 d and an extremely large amplitude (0.816 mag) due to irradiation of the companion by a very hot pre-white dwarf. ETHOS 1 further strengthens the long-suspected link between binary central stars of PNe (CSPN) and jets. The Isaac Newton Telescope/Intermediate Dispersion Spectrograph and Very Large Telescope (VLT) FORS spectroscopy of the CSPN reveals weak N iii, C iii and C iv emission lines seen in other close binary CSPN and suggests that many CSPN with these weak emission lines are misclassified close binaries. We present VLT FORS imaging and Manchester Echelle Spectrometer long-slit observations from which a kinematic model of the nebula is built. An unusual combination of bipolar outflows and a spherical nebula conspires to produce an X-shaped appearance. The kinematic age of the jets (1750 ± 250 yr kpc⁻¹) is found to be more than that of the inner nebula (900 ± 100 yr kpc⁻¹), consistent with previous studies of similar PNe. Emission-line ratios of the jets are found to be consistent with that of reverse-shock models for fast low-ionization emitting regions (FLIERs) in PNe. Further large-scale surveys for close binary CSPN will be required to securely establish whether FLIERs are launched by close binaries.

1 INTRODUCTION

In the protracted debate on the shaping mechanisms of planetary nebulae (PNe) (Balick & Frank 2002), appropriate solutions are sought to explain both the dominant nebula morphology (e.g. spherical, elliptical, bipolar; Balick 1987) and the accompanying collimated outflows or ‘jets’. Single star evolution readily explains spherical PNe and potentially elliptical PNe if there is interaction with the interstellar medium (Villaver, García-Segura & Manchado 2003; Wareing, Zijlstra & O’Brien 2007). Highly axisymmetric or bipolar nebulae (Corradi & Schwarz 1995) have been modelled with some success, including the generalized interacting stellar wind (GISW) model (Kwok, Purton & Fitzgerald 1978; Kwok 2000) and a combination of stellar rotation and magnetic fields (e.g. García-Segura 1997). There are, however, fundamental limitations in both these models. The GISW model depends upon an assumed density contrast to produce bipolar PNe (e.g. a dusty torus) and the latter cannot operate without an additional supply of angular momentum (e.g. a binary companion; Soker 2006; Nordhaus, Blackman & Frank 2007). While there are multiple advantages to a binary explanation over existing theories, we are only beginning to understand the prevalence and impact of the binarity in PNe (De Marco 2009; Miszalski et al. 2009a,b, Miszalski et al. 2010).

Decisive observational evidence is required to advance the shaping debate which is imbalanced towards theoretical models. PNe with dusty discs seem to support the GISW model, though there appears to be a strong dependence on a binary companion for their formation (Chesneau 2010). Similarly, magnetic fields are difficult to observe directly in PNe (Jordan, Werner & O’Toole 2005; Sabin, Zijlstra & Greaves 2007), yet they appear to be more influential as part of a binary-driven dynamos (Nordhaus & Blackman 2006; Nordhaus et al. 2007). Evidence for close binary central stars of PNe (CSPN) that passed through a common-envelope (CE) phase is now firmly established and gaining considerable momentum with at least 40 now known (De Marco, Hillwig & Smith 2008; Miszalski et al. 2009a, 2010). These make up at least 17 ± 5 per cent of all CSPN (Bond 2000; Miszalski et al. 2009a).

With a larger sample of 30 post-CE nebulae, Miszalski et al. (2009b) found that at least 30 per cent, perhaps 60–70 per cent, of post-CE binaries had bipolar nebulae, suggesting that the CE phase preferentially shapes bipolar nebulae. This is a substantial improvement over the study of Bond & Livio (1990) whose sample of around a dozen close binaries showed no clear morphological trends. High-resolution kinematics of nebulae around post-CE CSPN have shown a tendency for matched orbital and nebula inclinations that strengthen the connection between the binary and its nebula (e.g. Mitchell et al. 2007; Jones et al. 2010). Further progress in this aspect will require a statistically significant sample of post-CE nebulae to be identified and studied in detail to reveal trends bestowed upon the nebula during the CE phase (Miszalski et al. 2010). Miszalski et al. (2009b) made the first steps in this direction by finding a propensity for low-ionization structures (Gonçalves, Corradi & Mampaso 2001) amongst post-CE nebulae, either in the orbital plane as a ring (e.g. Sab 41, Miszalski et al. 2009b; the Necklace, Corradi et al. 2011) or in the polar direction as jets (e.g. A 63, Mitchell et al. 2007).

Jets occur in a wide variety of astrophysical objects including PNe in which they are the least understood (Livio 1997). The connection between jets and binarity has long been suspected (Soker & Livio 1994), but never proven, given the paucity of known PNe with jets and binary CSPN. Jets in PNe typically have low outflow velocities (∼100 km s⁻¹), but higher velocities of 200–300 km s⁻¹ are not uncommon, for example, M 1-16 (Schwarz 1992; Corradi & Schwarz 1993), NGC 6337 (Corradi et al. 2000) and NGC 2392 (Gieseking, Becker & Solf 1985), and can even reach 630 km s⁻¹ in the case of MyCn 18 (O’Connor et al. 2000). There is no strong evidence for highly collimated outflows (e.g. Herbig–Haro like outflows) in genuine PNe and Frew & Parker (2010) suggest that the few objects with the most-collimated outflows are instead related to B[e] stars. The most-striking example is Hen 2-90 (Sahai & Nyman 2000) which was later reclassified as a B[e] star (Kraus et al. 2005; Frew & Parker 2010).

Instead, jets in PNe manifest as point-symmetric ‘corkscrew’-like outflows (López, Meaburn & Palmer 1993; Vázquez et al. 2008) or as pairs of opposing knots often called fast low-ionization emitting regions (FLIERs; see e.g. Balick, Preston & Icke 1987; Balick et al. 1993; Dopita 1997) with notable examples being Hb 4 (López, Steffen & Meaburn 1997; Harrington & Borkowski 2000; Miszalski et al. 2009b), NGC 6337 (Corradi et al. 2000; García-Díaz et al. 2009), A 63 (Mitchell et al. 2007), IC 4634 (Guerrero et al. 2008), IC 4673 (Kovacevic et al. 2010), Sab 41 (Miszalski et al. 2009b) and the Necklace (Corradi et al. 2011). Further examples are listed by Gonçalves et al. (2001) and Harrington & Borkowski (2000).

Point-symmetric jets have received the most attention in the literature with models consisting of an episodic precessing jet powered by an accretion disc in a wide binary with P_orb∼ a few years (e.g. Raga, Cantó & Biro 1993; Cliffe et al. 1995; Haro-Corzo et al. 2009; Raga et al. 2009). Precession helps smear out or widen the jet and may also explain the often ‘bent’ position of jets with respect to the presumed major-axis (e.g. Hb 4 and Sab 41). Some jets appear in two pairs (e.g. NGC 6337) and these may have arose from multiple ejections (e.g. Nordhaus & Blackman 2006). A similar scenario involves a short-lived‘wobbling’ accretion disc (Livio & Pringle 1996, 1997; Icke 2003) and models with magnetic fields have also successfully been applied to the problem (García-Segura 1997). Dynamical lifetimes of jets suggest that they are ejected well before the main nebula (Mitchell et al. 2007; Corradi et al. 2011), that is, before the CE is ejected, if a close binary is present, and this is consistent with the short-lived nature of the modelled accretion discs.

The most-constructive way to establish the link between jets and binaries is to look for binary CSPN in a large sample of PNe with jets. This is one of the major aims behind our ongoing programme to look for close binary CSPN via the photometric monitoring technique (Miszalski et al. 2010). We are also targeting bipolar PNe and through non-detections we will also be able to discern whether larger orbital periods produce point-symmetric jets rather than FLIERs. The programme has been very successful so far with at least five close binary CSPN found using the 1.2-m Flemish Mercator Telescope. These include the Necklace and its SN 1987A like ring (Corradi et al. 2011), the Type Ia supernova candidate double-degenerate nucleus of Hen 2-428 (Santander-García et al. 2010) and the subject of this paper, ETHOS 1.

This paper is structured as follows. Section 2 introduces the new survey responsible for the discovery of ETHOS 1 and the spectroscopic confirmation of its PN status. Spectroscopic and photometric evidence for binarity is presented in Section 3. The nebula is examined in Section 4 with narrow-band images and high-resolution spectroscopy. Section 4.2 examines diagnostic emission-line ratios of the jets and we conclude in Section 5.

2 DISCOVERY AND PLANETARY NEBULA CONFIRMATION

2.1 The Extremely Turquoise Halo Object Survey

Beyond the confines of Galactic Plane Hα surveys (Drew et al. 2005; Parker et al. 2005), there are very few PNe currently known as less than 9 per cent of ∼2800 Galactic PNe are located at latitudes |b| ≳ 10°. Most of these were found from visual inspection of photographic broad-band sky survey plates (e.g. Kohoutek 1963; Abell 1966; Longmore 1977; Weinberger 1977; Longmore & Tritton 1980; Hartl & Tritton 1985). With the digitization of the same material, it is now possible to extend this work to lower surface brightnesses and to systematically search larger areas more efficiently (Jacoby et al. 2010). The MASH-II project (Miszalski et al. 2008) demonstrated that visual searches for PNe are largely insensitive to compact or star-like PNe. Several compact PNe were found from catalogue queries of the SuperCOSMOS Hα Survey (Parker et al. 2005) including the MASH-II PN MPA 1327−6031 that lies just near the outer lobe of the bipolar MASH PN PHR 1327−6032 (Parker et al. 2006). The number of distant (compact) halo PNe is therefore essentially unknown. Further discoveries are essential to determine whether halo PNe can only be formed via close binary evolution (Bond 2000) as there is currently only a small amount of supporting evidence (e.g. Peña & Ruiz 1998; Tovmassian et al. 2010).

The Extremely Turquoise Halo Object Survey (ETHOS, Miszalski et al., in preparation) is a new survey designed to find high-latitude compact PNe that would ordinarily be missed by visual searches. The ETHOS is essentially an extension to broad-band colours of the MASH-II catalogue search applied to the whole sky (excluding the Hα surveyed regions). Currently, four candidate PNe have been spectroscopically confirmed out of a few dozen candidates identified solely from queries of the SuperCOSMOS Science Archive (SSA, Hambly et al. 2004), a convenient sql interface to the entire data base of catalogue data generated during the digitization of SuperCOSMOS Sky Survey data (Hambly, Irwin & MacGillivray 2001b). Candidates are then visualized as for the MASH-II, but this time replacing the primary colour-composite Hα (red), short-red (green) and B_J (blue) image with an I_N (red), R_F (green) and B_J (blue) image (Reid et al. 1991; Hambly et al. 2001a). The IRB image is particularly effective in distinguishing PNe apart from other objects, since they often appear with a strong turquoise hue. Fig. 1 best shows the turquoise false-colour in the discovery image of ETHOS 1, the first PN candidate identified during the design phase of the ETHOS (Miszalski 2009). The object was selected based on the colours of the inner nebula, but the presence of faint extensions on either side made it stand out amongst the other candidates. Further details of the survey and initial results will be presented elsewhere (Miszalski et al., in preparation).

2.2 Spectroscopic confirmation of ETHOS 1

ETHOS 1 was observed during the Isaac Newton Telescope (INT) service time with the Intermediate Dispersion Spectrograph (IDS) on 2009 March 10. The R300V grating was used at a central wavelength of 5400 Å to provide a wavelength coverage from 3300 to 8000 Å at a dispersion of 1.87 Å pixel⁻¹. The spectrograph slit was placed along the jets at a position angle (PA) of 149° and its width of 1.0 arcsec gave a resolution of 5 Å (full width at half-maximum, henceforth FWHM) at 5400 Å. One 30-min exposure was taken at an airmass of 1.51 followed by a 30-s exposure of the spectrophotometric standard star HD 192281 at an airmass of 1.72 with an 8-arcsec slit. Standard reductions were performed with iraf and four different extractions corresponding to different nebula regions were made using the iraf task APALL. The stellar continuum of the CSPN was traced to guide the extractions. Aperture sizes used were 6 pixels for the CSPN (2.4 arcsec), 54 pixels for the inner nebula (21.6 arcsec, which includes the CSPN), 19 pixels for the SE jet (7.6 arcsec) and 23 pixels for the NW jet (9.2 arcsec). The jet extraction widths were measured from the jet tips inwards towards the CSPN. The extractions were then flux calibrated in the standard fashion with the aforementioned standard.

Fig. 2 shows the IDS spectra that confirm the PN nature of ETHOS 1 (PN G068.1+11.0) whose basic properties are given in Table 1. We measured c, the logarithmic extinction at Hβ, to be 0.18 using the Howarth (1983) reddening law and T_e = 17 700 K (Section 4.2). The high ratio of He ii λ4686/Hβ≳ 1.1 in the main nebula is typical of very high excitation PNe at high Galactic latitudes in which all the helium is doubly ionized giving a mass-bounded, optically thin nebula (Kaler 1981; Kaler & Jacoby 1989). Similar emission-line intensities are also found in the post-CE PNe K 1-2 (Exter, Pollacco & Bell 2003) and the Necklace (Corradi et al. 2011), but in our case we do not find any low-ionization species in the main nebula, perhaps suggesting a slightly higher CSPN temperature. The only low-ionization species are found in the tips of the jets which are caused by shocks (Section 4.2). Besides the [Ne v] and [Ar v] emission found close to the hot CSPN, the CSPN spectrum also shows C iii λ4650, C iv λ5801, λ5812 and λ7726 as the signature of an irradiated close companion (Section 3.1).

3 THE BINARY CENTRAL STAR

3.1 An irradiated spectroscopic binary

ETHOS 1 was also observed with the FORS2 instrument (Appenzeller et al. 1998) under the Very Large Telescope (VLT) visitor mode program 083.D-0654(A) on 2009 July 14. The objective of this programme was to measure radial velocity curves of binaries discovered by Miszalski et al. (2009a); however, the inclement weather forced the execution of a backup programme in which ETHOS 1 was included. The blue optimized E2V detector from the FORS1 was used in combination with the 1200g grism to give a continuous wavelength coverage from 4058 to 5556 Å. The CCD was readout with 2 × 2 binning to give a dispersion of 0.72 Å pixel⁻¹, a spatial resolution of 0.252 arcsec pixel⁻¹ and a resolution measured from arc lines of 1.57 Å (FWHM) at 4800 Å.

A 40-min spectrum of ETHOS 1 was obtained at an airmass of 2.04 and the seeing varied between 1.0–1.3 arcsec under thin-thick cloud (relative flux rms up to 0.07) limiting the signal-to-noise ratio (S/N) in the continuum to ∼40 near He ii λ4686. The 0.7-arcsec slit was placed at PA = 144° to cover the jets, but, because of the gap between FORS chips, not all of the NW jet was observed. The spectroscopic data were processed using the ESO FORS pipeline and the iraf task APALL was used to make three extractions using the CSPN continuum as a reference. These were 8 pixels for the CSPN (2.0 arcsec), 84 pixels for the inner nebula (21.2 arcsec, which includes the CSPN) and 70 pixels for the SE jet (17.6 arcsec). The jet extraction widths were measured from the jet tips inwards towards the CSPN. Each spectrum was flux calibrated using a 300-s exposure of the CSPN of NGC 7293 (Oke 1990) obtained at the end of the same night at an airmass of 1.27 and a slit width of 0.7 arcsec. The final spectra are depicted in Fig. 3.

Features identified in the IDS and FORS CSPN spectra are typical of an irradiated close binary CSPN (Pollacco & Bell 1993, 1994). The hot primary is revealed by He ii λ4686, λ5412 in absorption and the primary is irradiating a main-sequence companion to produce N iii (λ4634.14, λ4640.64), C iii (λ4647.42, λ4650.25) and C iv (λ4658.30, λ5801.33, 5811.98, λ7726.20) emission lines. These lines originate from the irradiated zone on the side of the companion facing the primary and therefore serve as a proxy for the motion of the secondary. However, it should be noted that these lines sample the centre of light and not the centre of mass of the secondary (e.g. Exter et al. 2003).

To measure any shifts present, we used the [O iii] and Hβ emission lines of our FORS spectrum of the inner nebula as a reference. The systemic heliocentric velocity of the inner nebula was set to be V_sys=−20 km s⁻¹ from higher resolution spectroscopy (Section 4.1). The N iii λ4640.64, C iii λ4647.42, λ4650.25, and C iv λ4658.30 emission lines gave an average V_sec=−122.0 ± 4.6 km s⁻¹. Estimating V_pri is much more difficult, given the prominent nebular He ii emission, but the He ii λ4686 absorption line is present (albeit at low S/N, Fig. 3) and redshifted compared to the nebula velocity. The exact line profile centre is difficult to establish, given the nebular emission and the chance that the line core may be partially filled in by irradiated emission. The large uncertainty means we can only estimate V_pri∼ 105 ± 80 km s⁻¹. Our single epoch spectrum therefore confirms the presence of a spectroscopic binary with both components well separated from each other and the nebula. Given the large uncertainty in V_pri and the lack of a well-sampled radial velocity curve, it is premature to estimate the masses of each component at this stage.

An important corollary is that the presence of the N iii, C iii and C iv emission lines alone would ordinarily be a sufficient ground for a weak emission-line star or wels classification (Tylenda, Acker & Stenholm 1993; Marcolino & Araújo 2003). The wels are a heterogeneous class of CSPN that are poorly defined (Fogel, De Marco & Jacoby 2003) and do not fit well-established classification schemes (Méndez 1991; Crowther, De Marco & Barlow 1998; Acker & Neiner 2003). According to Tylenda et al. (1993), wels are identified by (i) C iv λ5801, 5812 weaker and narrower than in [WC] CSPN; (ii) very weak or undetectable C iii λ5696; and (iii) the emission complex at λ4650. However, this is exactly what is observed in many close binary CSPN with the best examples being Abell 46 (Pollacco & Bell 1994), ETHOS 1 (Fig. 3) and the Necklace (Corradi et al. 2011). Many wels are therefore likely to be misclassified close binaries (Miszalski 2009). Further examples will be presented elsewhere and we expect further high-resolution spectroscopic observations of the so-called wels will identify further close binaries.

3.2 Light curve

Time-series photometry of ETHOS 1 was obtained with the MEROPE camera (Davignon et al. 2004) on the 1.2-m semirobotic Flemish Mercator Telescope (Raskin et al. 2004) from 2009 August 24 to 2009 September 4. Fig. 4 shows the light curve phased with our ephemeris: HJD (min I) = 245 5076.0312 ± 0.0007 + (0.535 12 ± 0.000 19) E. The IDS and FORS spectra taken earlier during the same year covered phases 0.40–0.44 and 0.80–0.85, respectively. As for the Necklace, the observed periodic variability is clear evidence of binarity. A sine fit to the light curve gives a very high amplitude of 0.816 mag that indicates a very hot primary irradiating a cool main-sequence companion (see e.g. fig. 1 of De Marco et al. 2008). Such a large amplitude places ETHOS 1 amongst a small group of post-CE binaries showing extreme irradiation effects with only Sab 41 having a larger amplitude (I = 0.849 mag, Miszalski et al. 2009a). The next largest amplitudes are found in the Necklace (I = 0.747 mag, Corradi et al. 2011) and K 1-2 (V = 0.680 mag, Exter et al. 2003). There are no other post-CE binaries without visible nebulae that come close to these extreme amplitudes simply because they have evolved significantly farther along the white dwarf cooling track (Aungwerojwit et al. 2007; Shimansky et al. 2009).

4 THE NEBULA

4.1 Morphology and kinematics

Narrow-band images of 300 s duration each was obtained during the VLT visitor mode program 085.D-0629(A) on 2010 June 19 with the FORS Hα+ 83, O iii+50 and O ii+44 filters. The central wavelengths and FWHM of each filter are 656.3/6.1 nm, 500.1/5.7 nm and 377.6/6.5 nm, respectively. Note that the Hα+ 83 filter includes the [N ii]λ6548, λ6583 emission lines. Fig. 5 presents a montage of the images which show three components to ETHOS 1: (i) the main nebula 19.4 arcsec across measured at 10 per cent of the peak Hα intensity; (ii) the jets measuring 62.6 arcsec tip to tip also at 10 per cent of the peak Hα intensity; and (iii) a faint bipolar outflow best visible in [O iii] (Fig. 5c).

The absence of an inner ring of low-ionization structures (LIS) sets ETHOS 1 apart from the other nebulae surrounding high-amplitude close binaries including K 1-2 (Exter et al. 2003), Sab 41 (Miszalski et al. 2009b) and the Necklace (Corradi et al. 2011). As the ultraviolet (UV) radiation field is comparably high in the other objects, this may indicate that the ring never existed rather than recent ablation of an LIS ring. There are in fact no low-ionization species detected in ETHOS 1 except for in the shocked jet tips described in Section 4.2. A faint detection in the [O ii] image is due to the nebula continuum, since there is no [N ii], [S ii] or [O ii] emission recorded by our INT spectrum of the main nebula (Fig. 2).

Spatially resolved long-slit echelle spectra of ETHOS 1 were made using the second Manchester Echelle Spectrometer on the 2.1-m San Pedro Mártir Telescope (Meaburn et al. 2003). The primary spectral mode was used with narrow-band filters to isolate the [O iii]λ5007, Hα and [N ii]λ6548, 6583 emission lines. Observations were performed on 2010 September 23 with a 2 arcmin × 2 arcsec slit and the Thomson 2k × 2k CCD which gave a spatial scale of 0.38 arcsec pixel⁻¹ and a resolving power of R∼ 30 000 (2.82 km s⁻¹ pixel⁻¹). A series of 1800-s exposures were taken with the slit oriented approximately along the minor (PA = 60°) and major (PA = 147°) axes.

Fig. 6 presents the emission-line position–velocity (P–V) diagrams observed. Alongside these are the same P–V diagrams extracted from a shape (Steffen & López 2006) best-fitting spatio-kinematic model of ETHOS 1 built following Jones et al. (2010). Fig. 7 shows the model that was inspired by the faint Hb 12 like bipolar extensions (Vaytet et al. 2009) that are visible in Fig. 5(c). The basic components are a spherical outflow and nested thin-walled bipolar outflows. The inner bipolar outflow helps reproduce the X-shape and the outer (possibly related) Hb 12 like bipolar outflow protrudes from the central sphere. Our approach was not tailored to each emission line observed, but instead aimed to create a good overall approximation to the features seen in Figs 5 and 6. The jets were not incorporated in our model, but they appear to be well aligned to the major-axis unlike other PNe (e.g. Hb4). The faint bipolar extensions (Fig. 5c) were included in our model, but their low surface brightness precluded their detection on the 2.1-m telescope at this resolution. We note that no attempt has been made to replicate the relative surface brightnesses of each component in the model and deeper observations should be able to detect these fainter features.

Along the major-axis, the P–V diagrams of the inner nebula display a figure-of-eight profile best seen in Hα that is typical of a bipolar outflow with a ‘pinched’ waist (e.g. Icke, Preston & Balick 1989). This bipolar profile is also apparent in [O iii], but is far more irregular and lacks emission from the central region. The southern component splits into two velocity components that are consistent with the front and back walls of a hollow bipolar shell. A velocity ellipse is also seen in the [O iii] profile surrounding the figure-of-eight profile. This is consistent with a spherical shell of material and the minor-axis P–V diagrams support this interpretation. Along the minor-axis, the outer ellipse represents the spherical shell, while the inner knot at negative velocity represents the pinched bipolar shell (e.g. Vaytet et al. 2009). There is a small amount of asymmetry in the minor-axis which is partly due to imperfect alignment of the slit along the axis and partly intrinsic.

No emission is detected from the central region in [N ii] consistent with our other observations. There is, however, an elongated excess of Hα emission in the inner nebula aligned along the major-axis that is not matched in [O iii] (e.g. Fig. 5a). While this could be explained by ionization stratification (e.g. the same feature may be visible in O v or O vi), it may also be related to similar but ‘tilted’ inner nebulae in PN G126.6+01.3 (Mampaso et al. 2006), M 2-19 (Miszalski et al. 2009b) and A 41 (Miszalski et al. 2009b; Jones et al. 2010). We suggest these nebulae could be the remnants of a precessing jet (Haro-Corzo et al. 2009; Raga et al. 2009) or nozzle (Balick 2000). The uncanny resemblance between PN G126.6+01.3 and the models of Haro-Corzo et al. (2009) supports this hypothesis.

The symmetry axis of our best-fitting shape model has an inclination of 60°± 5° to the line of sight. This inclination is consistent with the non-eclipsing light curve, the high degree of mirror symmetry evident in the main nebula about the minor-axis and the large visibility fraction of the jets. A heliocentric radial velocity of −20 ± 5 km s⁻¹ was determined from our model for the nebula. The expansion of the model nebula follows a Hubble-type flow, where all velocities are proportional to the radial distance from the nebular centre, equivalent to an expansion velocity of 55 km s⁻¹ at the edge of the spherical component. This is above average for a PN but not unexpected for a bipolar outflow (Corradi & Schwarz 1995) with He ii λ6560 present (Richer et al. 2008). Fig. 3 shows He ii λ4859 and we confirm the detection of He ii λ6560 in our MES spectroscopy.

Although the distance to ETHOS 1 is unknown, the spatio-kinematic model was used to determine the kinematical age per kpc of the inner nebula as 900 ± 100 yr kpc⁻¹. Similarly, assuming the jets have the same inclination as the internal bipolar structure, we find their kinematical age to be 1750 ± 250 yr kpc⁻¹ at their tips, almost double that of the central region!Table 2 compares the kinematic ages of ETHOS 1 against the Necklace and A63. The velocities of the jet tips reach −55 ± 5 km s⁻¹ (SE jet) and 65 ± 5 km s⁻¹ (NW jet). Their deprojected velocities at our assumed inclination are −110 ± 10 and 130 ± 10 km s⁻¹, respectively.

4.2 Emission-line diagnostics

Table 3 gives our measured emission-line intensities for our extracted IDS and FORS spectra. Dereddened intensities were calculated using the Howarth (1983) reddening law and c = 0.18. As Exter et al. (2003) point out an abundance analysis for such high-excitation nebulae requires photoionization modelling to properly measure the chemical abundances. Since our spectral information is rather limited and lacks reliable UV and far-blue coverage, we postpone this work for now. Abundances of the jets are likely heavily distorted by their shocked nature, but we can investigate their emission-line ratios for diagnostic purposes.

The nebular package in iraf (Shaw & Dufour 1995) was used to determine electron temperatures and densities in the main nebula and SE jet. Assuming n_e = 1000 cm⁻³, the main nebula has a very hot T_e = 17 700 ± 500 K. This is comparable to K 1-2 with T_e = 17 500 ± 1000 K, but is larger than T_e = 14 800⁺⁵³⁰₋₄₆₀ K measured in the Necklace (Corradi et al. 2011). The [Ar iv]λ4711/4740 density diagnostic from our VLT spectrum lies near the lower density limit so we find n_e = 850 ± 1000 cm⁻³. We also find the jets to be cooler T_e = 12 900 ± 1000 K and to have a similarly low density by using the [S ii]λ6731/6717 diagnostic to provide an upper limit of n_e≲ 1000 cm⁻³.

The emission-line ratios from separate extractions of the jet tips in Table 3 can be compared to model predictions (Dopita 1997; Raga et al. 2008). These works aim to reproduce FLIERs as the result of shocks in the strongly photoionized medium provided by the CSPN, giving the appearance of a ‘reverse shock’. In ETHOS 1, the presence of [O iii] in the jet and its relatively low T_e = 12 900 K indicates strong radiative cooling is occurring in the post-shock region. We find the line ratios agree well with log U₂∼−4.4 from Dopita (1997) and the measured values of FLIERs in NGC 6543 and 7009 (Raga et al. 2008). The Raga et al. (2008) models systematically underestimate the [O iii]λ5007/Hα ratio, although this is probably due to the relatively low stellar temperatures adopted in their models. The spatial distribution of the key emission lines Hα, [N ii]λ6583 and [O iii]λ5007 are also consistent with the Raga et al. (2008) models. Fig. 8 presents spatial profiles of these lines where the CSPN continuum was subtracted from extractions adjacent to the lines. The results qualitatively agree with both IC 4634 and the models presented by Raga et al. (2008) with [N ii] extending ∼2 arcsec farther out than the other emission lines. This effect is also evident in Fig. 5(a).

5 CONCLUSIONS

We have introduced ETHOS 1 (PN G068.1+11.1), the first spectroscopically confirmed PN discovered from the ETHOS (Miszalski et al., in preparation). An irradiated close binary central star was discovered with an orbital period of 0.535 d and an amplitude of 0.816 mag. The extreme amplitude is second only to Sab 41 (Miszalski et al. 2009a) and is consistent with the presence of a very hot central star that produces the highly ionized nebula (T_e = 17 700 K). The Necklace (Corradi et al. 2011) and K 1-2 (Exter et al. 2003) also share similarly large amplitudes and high-ionization nebulae, although the absence of low-ionization filaments in ETHOS 1 may suggest a slightly different evolutionary history. VLT FORS spectroscopy of the CSPN confirms the presence of a close binary with a large velocity separation between primary and secondary components and the nebula. The presence of N iii, C iii and C iv emission lines continues the trend seen in other irradiated close binary CSPN. These weak emission lines are typical of many CSPN classified as wels in the literature and we expect many of these will turn out to be close binaries. Further observations are required to constrain the orbital inclination, masses and radii of the binary CSPN.

A spectacular pair of jets travelling at 120 ± 10 km s⁻¹ accompanies the inner nebula of ETHOS 1, adding further evidence towards the long-suspected relationship between binary CSPN and jets. Their tips present emission-line ratios that are consistent with shocked models of FLIERs in PNe. The fact that the jets are detached suggests a limited period of jet activity consistent with a transient accretion disc before the CE is ejected. The kinematic age of the jets (1750 ± 250 yr kpc⁻¹) was found to be older than the inner nebula (900 ± 100 yr kpc⁻¹), supporting this hypothesis. ETHOS 1 therefore continues to follow this trend previously identified in A 63 (Mitchell et al. 2007) and the Necklace (Corradi et al. 2011). Both ETHOS 1 and the Necklace have younger kinematic ages than A 63, consistent with the apparent ongoing cooling in the younger jets via the [O iii] emission. A close binary engine powering jets is out of place in the literature with most models of jets incorporating orbital periods of several years (Cliffe et al. 1995; Haro-Corzo et al. 2009; Raga et al. 2009). These models may be adequate if the accretion disc is established before the companion begins its in-spiral phase as our kinematic ages suggest. Nevertheless, new models with shorter orbital periods may be necessary to model the jets in objects like ETHOS 1, which are relatively more collimated than extant model jets.

VLT FORS imaging and MES high-resolution spectroscopy of the inner nebula of ETHOS 1 was conducted. The X-shaped inner nebula was reconstructed using a shape kinematic model consisting of bipolar outflows and a spherical nebula. The faint bipolar extensions particularly visible in the [O iii] FORS image could be an earlier ejection occurring at about the same time as the jets or alternatively may represent a bubble produced by the jets clearing out a cavity around the PN. Similar bipolar outflows are seen in A 65 (Walsh & Walton 1996), PPA 1759−2834 (Miszalski et al. 2009b) and Fg 1 (Boffin & Miszalski, in preparation), suggesting they may have a binary-related origin. It may also be possible that ETHOS 1 represents a slightly more evolved state of Hb 12 (Vaytet et al. 2009) where the spherical nebula is ejected on top of a pre-existing Hb 12 system.

We thank the DUOPM observers supervised by Jean-Eudes Arlot and Agnès Acker at the OHP T120 during 2009 July 20 and 21 for observations that gave the first indication of the I-band variability in ETHOS 1. The work of RLMC and MS-G has been supported by the Spanish Ministry of Science and Innovation (MICINN) under grant AYA2007-66804. LS is in grateful receipt of a UNAM postdoctoral fellowship and is partially supported by PAPIIT-UNAM grant IN109509 (Mexico). We thank the staff at the San Pedro Martír Observatory who assisted with the observations. HMJB wishes to thank Yuri Beletsky and Leo Rivas for their efficient support at the VLT. We also thank the referee Orsola De Marco for her careful reading of this manuscript and useful comments. This research has made use of data obtained from the SSA, prepared and hosted by the Wide Field Astronomy Unit, Institute for Astronomy, University of Edinburgh, which is funded by the UK Science and Technology Facilities Council.

REFERENCES

Author notes