THE CHANDRA PLANETARY NEBULA SURVEY (ChanPlaNS). II. X-RAY EMISSION FROM COMPACT PLANETARY NEBULAE

The PNe that constitute the sample observed with Chandra during Cycle 14, listed in Table 1, were primarily selected so as to further probe the onset, evolutionary timescale, and properties of diffuse X-ray sources within PNe. Hence, we selected objects on the basis of their relatively small physical dimensions, as ascertained via optical imaging. More specifically, we chose the 24 Table 1 objects from the compilation of solar-neighborhood PNe in Frew (2008) such that the resulting merged Cycle 12+14+archival sample would constitute a representative volume-limited (D ≲ 1.5 kpc) sample of PNe with inner bubble or inner shell radii ≲0.4 pc (excluding very low surface brightness PNe, which typically have faint central stars). We estimate that our completeness in surveying compact (R_neb ≲ 0.4 pc) nebulae within ∼1.5 kpc is 90%, based on the solar neighborhood census of PNe recently compiled by Frew et al. (2014b). The chosen nebular radius roughly corresponds to a dynamical PN lifetime of ≲ 10⁴ yr—approximately twice the apparent hot bubble dissipation timescale inferred from Cycle 12 plus archival Chandra data shown in Paper I—assuming typical PN expansion velocities of ∼40 km s⁻¹ (Jacob et al. 2013). Basic PN and CSPN data for the sample objects can be found in Table 1.

All Cycle 14 PNe were observed with the back-illuminated (BI) CCD S3 of Chandra's Advanced CCD Imaging Spectrometer (ACIS). The use of CCDs as X-ray detectors provides determinations of incident photon energies as well as locations, which in turn allows for filtering of images by photon energy. Chandra/ACIS-S3 has an energy sensitivity of ∼0.3–8.0 keV, with a field of view of ∼8' × 8' and a pixel size of 0492. The BI CCD S3 has greater low-energy (<1 keV) sensitivity compared to the front-illuminated ACIS-I array, extending sensitivity down to ∼0.2 keV for high soft photon fluxes, albeit with uncertain calibration (Montez & Kastner 2013). Additionally, the use of S3 facilitates subpixel event repositioning (SER) in processing so as to better sample the core of the point spread function generated by Chandra's High-Resolution Mirror Assembly (Li et al. 2004). Observation IDs, dates, and exposure times for the Cycle 14 ChanPlaNS observations are listed in Table 2.

All data were reduced using CIAO²⁴ (version 4.5). Data reduction made use of the ChanPlaNS pipeline, which is detailed in Paper I. Briefly, this pipeline consists of processing the following steps: reprocessing of the primary and secondary data files and the application of SER using chandra_repro; detection of sources in the full field of view of each observation using wavdetect; calculation of statistics from a 35 radius region centered on the CSPN, to place constraints on possible X-ray point source emission; and generation of annotated two-panel images displaying a filtered Chandra X-ray image highlighting the predominately soft X-ray (0.3–2.0 keV) nature of the nebulae and the best available optical (Hα or R band) image obtained from the Hubble Legacy Archive,²⁵ the Wisconsin-Indiana-Yale-NOAO (WIYN) 0.9 m telescope (see Paper I), the SuperCOSMOS H-Alpha Survey (Parker et al. 2005; Frew et al. 2014a), or the Digitized Sky Survey (DSS²⁶), as indicated in the panels (Figure 1).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We have detected five of the Table 1 PNe in diffuse X-ray emission: Hb 5, NGC 1501, NGC 3918, NGC 6153, and NGC 6369 (Figure 2). The last four objects are detected for the first time; Hb 5 was detected previously by XMM-Newton (Montez et al. 2009, see below). These detections represent a ∼21% (5/24) detection rate of hot bubbles within the Cycle 14 sample. When these detections are combined with the previous ChanPlaNS observations, the ChanPlaNS detection rate of PNe hot bubbles now stands at ∼27% (16/59). Two of the newly detected objects, NGC 1501 and NGC 6369, host Wolf-Rayet-type (WR) CSPNe, which brings the ChanPlaNS diffuse detection rate of such objects to 100% (5/5). The detection rate of diffuse X-ray emission from [WR]-type CSPNe remains high when including PNe beyond ∼1.5 kpc (see Kastner et al. 2008). These diffuse X-ray detections follow the previously established trend that diffuse X-ray PNe generally display compact (R_neb ≲ 0.15 pc), elliptical optical morphologies (Figure 3; Paper I). A notable exception is Hb 5, which has a bipolar morphology.

The detections of X-ray point sources within three Cycle 14 PNe, HbDs1, NGC 6337, and Sp 1 (Table 1) yielded a ∼13% detection rate of CSPNe. These Cycle 14 detections bring the overall ChanPlaNS detection rate of CSPNe to ∼36% (21/59). Two of the three new CSPNe detections (within NGC 6337 and Sp 1) are known binaries, which brings the overall ChanPlaNS detection rate of all CSPN that are known to be binaries (at the time this paper was written) to 60% (Table 4).

The diffuse X-ray emitting PNe detected in the Cycle 14 sample are displayed in Figure 2 and are briefly described below. As is evident in Figures 3 and 4 and Table 4, these detections reinforce previous ChanPlaNS results indicating that PNe displaying diffuse X-rays are predominantly elliptical in morphology with R_neb ≲ 0.15 pc; furthermore, diffuse X-ray PNe generally lack detections of near-IR H₂ emission. In a forthcoming paper (R. Montez et al. 2014, in preparation), we present a global analysis of the ChanPlaNS diffuse X-ray sources. Here, we describe the properties of the diffuse X-ray PNe detected during Cycle 14 observations.

3.2.1. Classical Hot Bubbles

3.2.2. PNe with [WR]-type CSPNe

3.2.3. The Bipolar Planetary Nebula, Hb 5

The three Cycle 14 X-ray point source detections (HbDs1, NGC 6337, Sp 1) continue the trend, noted in Paper I, that soft (∼0.6–1.0 keV) X-ray sources likely indicate the presence of CSPN photospheric emission and/or wind shocks, while hard (≳ 1.0 keV) sources appear related to the presence of binary companions to CSPNe (Figure 4; R. Montez et al. 2014a, in preparation). Below are brief summaries of each of the Table 1 PNe with newly detected point-like X-ray emission.

HbDs1 is an irregularly elliptical nebula around the hot blue star LSS 1362 (T_eff = 116 kK; Herald & Bianchi 2011). The nebula has enhanced Hα emission along its northern and southern limbs. The magnetic field properties of the CSPN have been debated. A kG field was inferred by Jordan et al. (2005, 2012) from spectropolarimetric data taken with the ESO Very Large Telescope; however, more recent FORS2 observations showed no evidence of such a strong field (with upper limits of B < 600 G; Leone et al. 2011; Bagnulo et al. 2012). The detected X-ray point source has a median energy of 0.65 keV, which may indicate self-shocking within the stellar winds previously detected at the CSPN (Herald & Bianchi 2011; Guerrero & De Marco 2013), possibly combined with photospheric emission (see Montez & Kastner 2013).

NGC 6337 is seen as a ring nebula with radial filaments. This PN harbors a nearly face-on (i ≲ 15°) close binary (0.17 day period) central star system (Hillwig et al. 2010). It is likely that the nebula is viewed nearly pole-on as well, suggesting a bipolar structure with a narrow waist around the binary (García-Díaz et al. 2009). The close binary is most likely the product of a common envelope phase (De Marco et al. 2011), leaving behind a pre-WD and a cool main-sequence star (Hillwig et al. 2010). The X-ray point source associated with this binary has a median energy of 1.27 keV, the hardest emission detected from a point source in the ChanPlaNS sample thus far. This hard X-ray detection provides strong support for a model in which the X-rays originate from a cool, coronally active companion (see Section 4).

Sp 1 appears similar to NGC 6337; this PN is also a diffuse ring on the plane of sky and harbors a close binary (2.9 day period) at its center (Bond & Livio 1990; Jones et al. 2012; Bodman et al. 2012). Spatiokinematical modeling of the nebular shell shows that the nebula is a nearly pole-on hourglass shape (Jones et al. 2012). This is borne out by modeling of the CSPN by Bodman et al. (2012), which indicates a G0 or F8 main sequence companion with an orbital inclination of 10° < i < 20°, consistent with the generation of a near pole-on nebula. UV spectra of the CSPN show evidence for stellar winds (Patriarchi & Perinotto 1991), which could be at best partly responsible for the X-ray point source at the CSPN. However, the median energy for the X-rays from the Sp 1 CSPN is 0.93 keV, making it one of the harder X-ray sources in the ChanPlaNS sample, and suggesting that the companion dominates the detected emission.

Table 3 lists the point source X-ray characteristics (net counts, count rates, median photon energies, and photon energy ranges) of each PN in the Cycle 14 sample. To obtain these results, the 0.3–3.0 keV background counts were extracted from a source-free region near the central star. These background counts were scaled to the point source extraction radius of 35, and then subtracted from the counts of the same extraction region centered on the CSPN, to obtain the net source counts (or source count upper limits). Sources with less than six net counts are classified as undetected. Two sources lie below this background-subtracted limit, A65 and NGC 6026, but could be considered as tentative detections; there are ∼4 (source+background) counts found near their CSPNe, both of which have companion stars. There are a few sources above the net count limit that we consider to be doubtful as point source detections: Hb 5, NGC 3918, and NGC 6369. Contamination from the diffuse X-ray emission in NGC 3918 makes it difficult to determine the contribution from a possible point source at the CSPN in the circular source region. NGC 6369 is a luminous, albeit reddened, PN that lies near the limit of our volume-limited sample, at a distance of 1.55 kpc (Monteiro et al. 2004). Additionally, the diffuse X-ray emission of NGC 6369 is not concentrated at the CSPNe, which makes it unlikely to contaminate a point source detection. Finally, as noted in Section 3.2, Hb 5 appears to have no point-like emission within the CSPN extraction region.

Several PNe that were undetected appear to fit within morphology trends typical of previous ChanPlaNS point source and diffuse X-ray detections. IC 5148/50 is undetected, yet it has a morphology similar to Sp 1, suggesting that it may be a pole-on hourglass and, hence, could be the product of a binary system. In optical images, NGC 6894 has the multiple-shell structure of a hot bubble from which we would expect to see diffuse X-ray emission, though it is important to note that the central star is faint and on the WD cooling track. Similarly, NGC 7354 hosts multiple elliptical shells and ansae, similar to NGC 3918 and NGC 6153; however, no X-ray photons were detected from the innermost bubble of NGC 7354. These non-detections appear to place stringent limits on the timescales for X-ray emitting hot bubbles within PNe (Section 4.2)—although we note that NGC 7354 is quite heavily reddened and, hence, any soft X-rays from its interior might suffer considerable absorption.

The objects with no detectable X-ray emission have varying morphologies (round, elliptical, bipolar), with at least half exhibiting inherent asymmetries. IC 4637 is an example of a PNe that falls within the range of nebular properties for the diffuse X-ray detections (Paper I), but was not detected. Specifically, IC 4637 has a radius of 0.05 pc and elliptical morphology (Frew 2008). Additionally, IC 2149 is a young, compact bipolar nebula with an elliptical central shell and may be the product of a low-mass progenitor and/or binary evolution (Vázquez et al. 2002). Upper limit estimates for X-ray luminosity L_X for each of these sources (R. Montez et al. 2014, in preparation) will help place constraints on the sizes, plasma densities, and morphologies of X-ray emitting nebulae.

The CSPNe that are H-deficient and display spectroscopic evidence for fast stellar winds and high mass-loss rates (up to 10⁻⁶ M_☉ yr⁻¹; Crowther et al. 1998; DePew et al. 2011) are classified as [WR]-type because of their similarity, as a class, to luminous WR stars. The primary difference between the two classes is that the [WR]-type stars descend from intermediate mass (1–8 M_☉) stars, whereas "classical" WR stars are the descendants of massive (>25–30 M_☉) stars (Crowther et al. 2006). In total, there are five known [WR]-type CSPNe in the ChanPlaNS sample: NGC 1501 and NGC 6369 from Cycle 14 (Table 1), as well as NGC 40, NGC 2371, and BD +30°3639 from Cycle 12 and archival observations (Paper I). All five of these [WR]-type CSPN objects display diffuse X-ray emission. The only consistent morphological similarity among all of the [WR] type CSPNe is that their main bubbles appear elliptical (Frew 2008). In each case, the diffuse X-ray emission is enclosed by the central optical bubble, although the morphology of the X-ray emission varies considerably from object to object.

BD +30°3639: The unusually bright, diffuse emission fills the central bubble, perhaps as a consequence of its very young age (∼10³ yr; Frew 2008).

NGC 40: The emission appears within a partial ring that traces the inner edge of the bright optical rim of the nebula (Montez et al. 2005).

NGC 2371: The diffuse emission is contained within the bright nebular rim but does not seem to follow any significant optical feature, and is highly asymmetric.

NGC 1501: The X-ray emission traces the rim of the nebula, echoing the limb-brightened morphology of WR star wind-blown bubble S308 (Toalá et al. 2012).

NGC 6369: The diffuse emission is asymmetrical and strongest along the northern limb of the central bubble.

The X-ray results for most of the above PNe, with the possible exception of NGC 2371, reveal some potential similarities between PNe with [WR]-type CSPNe and the wind-blown bubbles of massive WR stars. The X-ray limb-brightening in massive WR star wind-blown bubbles, such as S308, is the result of the hot WR wind cooling as it comes in contact with the previously ejected red (or yellow) supergiant or luminous blue variable material (García-Segura et al. 1996a, 1996b; Freyer et al. 2003; Toalá & Arthur 2011; Toalá et al. 2012; Dwarkadas & Rosenberg 2013). A similar process occurs in PNe, wherein the X-ray emission is contained within the ejected AGB envelope and does not come in direct contact with the ISM. The question remains: is the diffuse X-ray emission from [WR] CSPNe limb-brightened or not? Given the photon-starved nature of our detections, we will require Poisson-based models for comparison with the observations to adequately answer this question.

The diffuse X-ray emission from hot bubbles in PNe is evidently confined to early phases in PN evolution. In Paper I, we pointed out that the central bubbles of PNe with diffuse emission tend to have elliptical, nested-shell morphologies and radii ≲ 0.15 pc (Figure 3 and Tables 1 and 4). This radius limit corresponds to dynamical ages ≲ 5 × 10³ yr, roughly a quarter of the (21 ± 5) × 10³ yr mean visibility time of PNe (Jacob et al. 2013). Our Cycle 14 detections and non-detections thereby reinforce the inference, made in Paper I, that diffuse X-ray emission is restricted to the most compact nebulae. We note that the correlation of diffuse X-rays with elliptical morphology is ambiguous (Table 4), but it remains the case that PNe with closed structures (whether bubbles or lobes) are more likely to be detected as diffuse X-ray sources.

Furthermore, from Figure 5 we see that with the exception of NGC 2371, all ChanPlaNS PNe with diffuse X-ray emission have nebular densities n_e ≳ 1000 cm⁻³, as determined via their Hα line luminosities and nebular radii, which are in turn determined from the Hα surface brightness versus radius relation (Frew 2008; Frew et al. 2014b). In contrast, the diffuse X-ray detection rate is essentially independent of distance and nebular excitation. Since PN density and radius are correlated, the remarkably sharp n_e boundary between diffuse X-ray detections and non-detections in Figure 5 further underscores the notion of a limit on the timescale for energetic (nebula-shaping) wind interactions in PNe. Moreover, Figure 5 suggests that, from the spectroscopy of density-sensitive lines (e.g., [O ii], [S ii]), it is possible to ascertain whether a PN might currently be in a wind-collision phase, when we would expect to see diffuse X-ray emission. Nevertheless, the duration of the wind-interaction phase is a complex problem that must account for (1) luminosity of the central star (e.g., Hb 5 high, NGC 5148 low), (2) expected hydrogen column density, and (3) spectral type of the central star (i.e., whether it is [WR] or not).

This density limit may explain the X-ray non-detection of nebulae like NGC 6894 and NGC 6804 (Paper I), which appear morphologically similar to known diffuse X-ray emitters. NGC 6894 has a nebular radius of 0.17 pc and NGC 6804 has a nebular radius of 0.19 pc, and both have a density log n_e ∼ 2.7. Thus, evidently, both NGC 6894 and NGC 6804 lie close to, but just below, the log n_e X-ray detection threshold. It is important to keep in mind that electron density is not a good indicator of PN evolutionary state. Fast evolving stars reach low luminosities at relatively high nebular densities, but remain undetected in diffuse X-rays (Steffen et al. 2008). Upper limit estimates and observations of additional nebulae similar to these—with n_e ≳ 1000 cm⁻³ and closed optical morphologies (multiple shells, well-defined inner bubbles)—will allow us to answer questions about diffuse emission such as the following. What is the lifetime of the hot bubble shaping phase? How does it relate to the mass of the CSPN progenitor? What is the relationship between diffuse X-ray emission and the final mass of the core, i.e., the WD remnant? We aim to revisit the question of diffuse X-ray dependence on age in R. Montez et al. (2014, in preparation; also see Schönberner et al. 2014).

Six additional known binary CSPNe were observed in Cycle 14, all of which are close binaries (p < 3 days). Two of these six, NGC 6337 and Sp 1, had detectable point source X-ray emission. Both PNe appear to have nearly pole-on hourglass structures. The median energy for both point sources is relatively hard (≳ 1.0 keV), which implies that the CSPN X-ray emission is due to late-type companions that display elevated levels of coronal activity. These binary, X-ray-luminous CSPNe are therefore similar to the CSPNe in DS 1, HFG 1, and LoTr 5, which are thought to be systems in which mass transfer from PN progenitor to late-type companion has spun up the companion (Jeffries & Stevens 1996; Soker & Kastner 2002; Montez et al. 2010); alternatively, in many or most of these cases of X-ray-emitting binary CSPNe, it is possible the coronal activity of the companion is due to tidal locking that has resulted in rapid, synchronous rotation (a possibility we further explore in Montez et al. 2014). Both NGC 6337 and Sp 1 represent the two most compact (and therefore youngest) PNe harboring X-ray point sources at known binary CSPNe in the ChanPlaNS sample (Figure 4). The periods of the close binaries within A65 (1.00 day; Bond & Livio 1990; Shimansky et al. 2009; T. Hillwig et al. 2014, in preparation), He 2-11 (0.61 days; Jones et al. 2014b), and NGC 6026 (0.53 days; Hillwig et al. 2010) lie between those of NGC 6337 (0.17 days; Hillwig et al. 2010) and Sp 1 (2.91 days; Bond & Livio 1990), but no point source X-ray emission was detected from the former three PNe. This leaves us with unanswered questions: What is the relationship between binary period/separation and X-ray point source emission? What are the characteristics of secondaries within binaries that do exhibit X-ray emission?

IC 5148/50 is a potentially significant non-detection, given its morphological similarity to Sp 1. Based on imagery alone, Sp 1 appears to have a spherical morphology, but spatiokinematical modeling reveals this is a projection effect and its intrinsic structure is that of a pole-on hourglass (Jones et al. 2012). Such bipolar structures are likely to be the products of binary interactions (see, e.g., Balick & Frank 2002, and references therein). The non-detection of IC 5148/50 may indicate that while it is morphologically similar to Sp 1, structurally it may be quite different. From Hα, [N ii], and [O iii] observations (Hua et al. 1998), it is clear that IC 5148/50 hosts not only internal elliptical structure but also a helical structure. Observations of H₂ (Kastner et al. 1996) and/or spatially resolved, high-resolution emission line spectroscopy (for use in spatiokinematical modeling) are necessary to determine whether or not IC 5148/50 has inherently bipolar versus elliptical structure. The CSPN of IC 5148/50 also has no infrared excess indicative of an unresolved cool companion (De Marco et al. 2013; Douchin et al. 2014).

It is possible that many other relatively X-ray-luminous CSPN have gone undetected in our sample of PNe. From Figure 5 we see that the fraction of PNe with detected point-like X-ray emission is strongly dependent on distance. All six sample PNe within ≲ 0.6 kpc have point-like emission, whereas beyond 1.3 kpc the detection fraction drops to ∼35%. This drop in detection fraction is likely the result of increased interstellar extinction obscuring the softer X-ray point sources, and furthermore implies that we lack the sensitivity necessary to detect even the (harder) binary CSPNe at distances much beyond ∼1 kpc. Further discussion concerning the implications of the distance dependence of the X-ray detection rate of CSPNe can be found in Montez et al. (2014).

Several elliptical PNe display diffuse X-ray emission morphologies wherein the elongated diffuse emission appears to include peaks on two opposite locations along the long axis of the elliptical rim. Examples are NGC 3242, NGC 6543, NGC 7009, NGC 7662 (Paper I), and NGC 3918 (Figure 1). One possible interpretation of such an X-ray morphology, with far-reaching implications, is that the diffuse emission emanates from gas heated by shocks as low-mass jets with velocities of a few 100 km s⁻¹ passed through material previously ejected in a common envelope interaction (Akashi et al. 2008). Such post-common envelope jets would most likely be launched by an accretion disk around a main sequence companion to the PN progenitor star during the progenitor's post-AGB phase. Indeed, it is possible that such jets are the "last gasps" of outflows launched during the final stages of common envelope evolution (Tocknell et al. 2014). If so, then the diffuse X-ray emission from elliptical PNe might hint at the operation of jets in the final removal of common envelopes formed by evolved, close binary systems. The X-ray emission morphologies of elliptical PNe, and their potential implications for common envelope models of PN formation, will be further explored in subsequence papers in the ChanPlaNS series (e.g., R. Montez et al. 2014, in preparation).