SPITZER 24 μm IMAGES OF PLANETARY NEBULAE

Low- and intermediate-mass stars expel their envelopes to form planetary nebulae (PNe), while their exposed cores evolve into white dwarfs (WDs). The morphology of a PN can be very complex. First of all, a PN may show a multiple-shell morphology due to interactions between fast and slow winds (Kwok 1983) and between the winds and the ambient interstellar medium (Dgani & Soker 1998; Villaver et al. 2003). A bipolar structure may be developed in a PN owing to a fast stellar rotation or close binary interaction (Soker & Livio 1989). Moreover, the effective temperature of the central star of a PN (CSPN) increases as the nebula expands, and the varying stellar UV flux results in varying ionization and excitation structure in the PN (Schönberner 1983). Therefore, images of a PN taken at different wavelengths may exhibit very different morphologies, and comparisons among these images help us understand the PN's physical conditions and structure.

Infrared (IR) imaging of PNe from the ground is easy at wavelengths ≲2 μm, but is hampered by the bright sky background at longer wavelengths. While high-resolution images at wavelengths up to ∼20 μm can be obtained from the ground for the bright, young, compact PNe (e.g., Volk et al. 2007), it is difficult to observe extended PNe with lower surface brightnesses. The Spitzer Space Telescope (Werner et al. 2004) makes it possible for the first time to obtain high-sensitivity images of a large number of older and more extended PNe at near- and mid-IR wavelengths. Its Infrared Array Camera (IRAC; Fazio et al. 2004) takes images at 3.6, 4.5, 5.8, and 8.0 μm, while the Multiband Imaging Photometer for Spitzer (MIPS; Rieke et al. 2004) takes images at 24, 70, and 160 μm. IRAC images are available for over 60 PNe and they have detected emission from ionized gas, warm dust, polycyclic aromatic hydrocarbons (PAHs), and H₂ molecular gas (e.g., Hora et al. 2004, 2008; Hora 2009). MIPS images, before our work, were available for only about a dozen mature PNe, excluding proto-PNe that were usually unresolved or saturated in MIPS 24 μm images. Analyses of MIPS observations of NGC 650 (Ueta 2006), NGC 2346 (Su et al. 2004), and NGC 7293 (Su et al. 2007) suggest that diffuse emission in the 24 μm band is dominated by the [O iv] 25.9 μm line, and the 70 and 160 μm bands by dust continuum emission.

We have conducted a Spitzer MIPS 24 μm survey of 72 hot WDs or pre-WDs, selected from McCook & Sion (1999) and Napiwotzki (1999), to search for dust disks at radii of a few tens of AU (PID 40953). Thirty-six of these hot stars surveyed are still surrounded by visible PNe; therefore, our MIPS survey of hot WDs has provided a threefold increase of 24 μm images of spatially resolved PNe in the entire Spitzer archive. Diffuse emission in the 24 μm band is detected from 28 PNe, among which ∼50% show shell morphologies and ∼35% show emission only in the central region near the CSPN. In this paper, we describe the observations in Section 2, present the results in Section 3, and discuss the origins of the 24 μm emission in Section 4 and implications of this emission in Section 5.

The MIPS onboard the Spitzer Space Telescope was used to image the 36 PNe in the 24 μm band as part of Program 40953. Each object was imaged in the photometry mode using the small offset scale with 10 s exposure time for three cycles. The Basic Calibrated Data processed by the Spitzer Science Center's pipeline software were used to produce mosaics of the PNe. Prior to building each mosaic, bad pixels and latent images of bright point sources in each frame were flagged, and background brightness offsets between individual frames were determined and corrected using the method outlined by Regan & Gruendl (1995). The MOPEX software package was used to combine data and produce the mosaics. The field of view of each mosaic covers ∼80 × 75, of which the central ∼35 × 30 portion has a total exposure time of ∼420 s and the outer regions have exposure times 1/2 or 1/4 that of the center. We note that latent images of the brighter PNe sometimes persist in the outer portions of the final mosaics. The point-spread function (PSF) of MIPS at 24 μm has a FWHM of ∼6''.

Among the 36 PNe that have been imaged in the MIPS 24 μm band, 28 are detected and exhibit a variety of morphologies. To understand the origin of this emission, we compare each PN's 24 μm morphology to its optical morphology. Hα images, showing all dense ionized gas, are best for such comparison; however, not all PNe have Hα images¹⁰ available, and we have to resort to the Digitized Sky Survey 2 red (DSS2r) images. For PNe with hot CSPNs, the nebular excitation and ionization is high and the strongest emission line in the DSS red passband is the Hα line; thus DSS2r images are excellent substitutes for Hα images. In one exception, HaTr 7, the DSS 2 blue (DSS2b) image is used because it shows the nebula better than the DSS2r image (Hartl & Tritton 1985). For the rest of the paper, we will loosely use these DSS2r and DSS2b images as "Hα" images, assuming that they are similar.

Table 1 lists all 36 PNe observed: Column 1, PN name; Column 2, Galactic coordinates; Column 3, CSPN name; Column 4, the stellar effective temperature (T_eff); Column 5, the angular size (diameter or major × minor axes) of the bright main nebula measured from the Hα images; Column 6, distance; Column 7, the linear size of the main nebula in Hα; Column 8, the He ii λ4686/Hβ flux ratio; and Column 9, the peak surface brightness at 24 μm. The eight PNe that were not detected in the 24 μm images are listed at the end of Table 1.

To visualize the spatial distribution of the 24 μm and Hα emission, we have extracted surface brightness profiles along cuts through the nebular center (i.e., the CSPN) at position angles (PAs) selected to avoid contaminating stars and to sample the well-defined shell rims of the main nebula. The width of each cut, typically a few arcsec, has been adjusted to achieve a reasonable signal-to-noise ratio (S/N) in the resulting profile. For each PN, similar widths are used to extract the 24 μm and Hα surface brightness profiles, and the profiles are individually zeroed at the background level and normalized to the peak of nebular emission. Note that the Hα surface brightness profiles are not normalized in the cases of Abell 61, HaTr 7, and Sh 2-216, because of stellar contamination in the first object and extremely low S/N in the images of the latter two PNe.

Using the images and surface brightness profiles, we find three kinds of correspondence between the 24 μm and Hα morphologies of the 28 PNe: (1) the 24 μm emission is more extended than the Hα emission, (2) the 24 μm emission has a similar spatial extent as the Hα emission, and (3) the 24 μm emission peaks near the CSPN within the Hα shell. We have used these different correspondences between 24 μm and Hα emission to divide the PNe into three groups accordingly, and arranged them in Table 1 and Figures 1–3 in the order of these groups to facilitate easy comparisons among objects. Some PNe have properties intermediate between Groups 1 and 2, and they are assigned to groups based on larger likelihoods. More detailed descriptions of these groups are given below.

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

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.1. Group 1

3.2. Group 2

3.3. Group 3

3.4. Nondetections

What are the origins of the diffuse emission detected in the Spitzer MIPS 24 μm band?  The [Ne v] 24.3 μm and [O iv] 25.9 μm nebular lines and dust continuum are the three main candidates contributing to the MIPS 24 μm fluxes, as indicated by Infrared Space Observatory (ISO) SWS spectra of the young PN NGC 7027 and the bipolar PN NGC 2346. In NGC 7027, [O iv] and [Ne v] lines and dust continuum are all significant, although the [Ne v] line is only 1/2 as bright as the [O iv] line (Bernard Salas et al. 2001). In NGC 2346, the [O iv] line contributes 27%, and dust continuum contributes the rest of the emission in the MIPS 24 μm band (Su et al. 2004). The excitation potentials of Ne v and O iv are 97.1 and 54.9 eV, respectively. It is not surprising that NGC 7027 and NGC 2346 have such different [Ne v]/[O iv] ratios, as the former CSPN has an effective temperature of ∼160,000 K (Beintema et al. 1996) and the latter CSPN has a Zanstra temperature of only ∼100,000 K (Méndez 1978). It is conceivable that the [Ne v]/ [O iv] ratio increases with stellar temperature, and for CSPNs with effective temperatures ⩽ 150,000 K, the [O iv] 25.9 μm line is by far the strongest emission line in the MIPS 24 μm band.

The ubiquitous presence of [O iv] line emission in our PNe is indicated by the He ii λ4686 emission that is detected in all available spectroscopic observations of these PNe (see Column 8 of Table 1; compiled by Tylenda et al. 1994). The He ii line is emitted through recombinations of He iii, of which the excitation potential is 54.4 eV. This excitation potential is close to that of O iv, thus wherever the He ii emission is present there ought to be [O iv] emission, as long as the density is not excessively higher than the critical density of [O iv], ∼9 × 10³ cm⁻³.

Note that the He ii/Hβ ratios in Column 8 of Table 1 appear to be higher for PNe in Groups 1 and 2 than for those in Group 3, but this is an artificial effect of how the observations were conducted. Optical spectrophotometric observations of PNe are usually made in regions of high surface brightness. For PNe in Groups 1 and 2, the optically bright regions are also bright in the MIPS 24 μm band; however, for PNe in Group 3, the bright 24 μm emission peaks near the center of a nebula, offset from the brightest optical emission in the shell where spectrophotometric observations are made. The He ii/Hβ ratios at the central diffuse 24 μm emission regions should be much higher than those in the nebular shells for Group 3 PNe. The radial variations of the He ii/Hβ ratio, as well as the [Ne v]/[O iv] ratio, reflect the well-known excitation and ionization stratifications in PNe.

The relative importance of nebular line emission and dust continuum in the MIPS 24 μm band depends on the ionization structure and dust content of a PN. It has been illustrated by Spitzer Infrared Spectrograph (IRS) observations of PNe in the Large and Small Magellanic Clouds that bipolar PNe are generally dustier than round/elliptical PNe and that the largest PNe exhibit the least dust continuum emission (Stanghellini et al. 2007). Group 1 PNe have 24 μm emission more extended than the Hα emission. As [O iv] tracks He ii and the He ii/Hα ratio decreases with radial distance from the CSPN, due to ionization and excitation stratification, the extended emission in the MIPS 24 μm band must be dominated by dust continuum. For the main shells of PNe in Groups 1 and 2, it is likely that both the nebular [O iv] line emission and dust continuum contribute to the MIPS 24 μm flux, but the exact ratios cannot be determined from the current data; spectroscopic observations are needed.

To determine the ratio of line emission to dust continuum for Group 3 PNe, we take a detailed look at Spitzer observations of diffuse emission from the Helix Nebula (Su et al. 2007). The IRAC images and MIPS 70 and 160 μm images of the Helix Nebula show the prominent helical ring structure with a hollow center, while the MIPS 24 μm image shows bright diffuse emission in the central region but only faint emission in the helical ring. Spitzer IRS observations were made at the central star and its immediate surroundings without additional observations of an external background region. These observations were adequate for deriving a spectrum of the central star but not a spectrum of the nebular emission because the zodiacal background was bright toward the Helix Nebula. MIPS 24 μm observations indicated that the nebular surface brightness was only about 1/7 that of the zodiacal emission. Indeed, the spectrum of a diffuse-emission region near the Helix central star, shown in Figure 4, is dominated by the zodiacal dust continuum emission, and it is not clear whether a small fraction of the dust continuum originates in the Helix Nebula. Nevertheless, the nebular spectrum in Figure 4 shows strong [O iv] 25.9 μm line emission; the [Ne v] 24.3 μm line is much weaker than that of the region encompassing the central star (see Figure 2 of Su et al. 2007).

Zoom In Zoom Out Reset image size

Recently, we obtained Spitzer IRS observations of the CSPN of Sh 2-188 in the mapping mode (PID 50629). While these observations will be presented in another paper, we have used these data to extract a spectrum of the diffuse emission near the central cavity. Figure 5 shows the slit positions marked on a MIPS 24 μm image of Sh 2-188. The spectra extracted from the nebular aperture "Neb" and from the background aperture "Bkg" show significant amounts of dust continuum emission and PAH emission. The origins of this dust continuum and PAH emission are not clear and could be nebular, interstellar, or even zodiacal. The background-subtracted nebular spectrum shows essentially all fine-structure line emission, with the [O iv] line dominating in the MIPS 24 μm band. Therefore, at least in the case of Group 3 PN Sh 2-188 the central diffuse emission in the MIPS 24 μm band is by far dominated by the [O iv] line emission. It is likely that the central diffuse 24 μm emission from all Group 3 PNe is dominated by the [O iv] line. The contribution of [Ne v] line is smaller and depends on the effective temperature of the CSPN and the radial distance to the CSPN.

Zoom In Zoom Out Reset image size

Finally, we note a complication in NGC 2438. While NGC 2438 is placed in Group 3 with centrally peaked 24 μm emission, diffuse 24 μm emission at lower levels is detected from its nebular shell as well as its halos. As shown in Figure 6, an archival IRAC 8 μm image of NGC 2438 has detected two halos around its main nebular shell, similar to the Hα+[N ii] image reported by Corradi et al. (2003); our MIPS 24 μm image of NGC 2438 has detected the main nebular shell and the inner halo clearly, and perhaps some faint emission from the outer halo as well. The IRAC 8 μm band contains strong PAH emission features at 7.7 and 8.6 μm, forbidden lines of [Ar ii], [Ar iii], [Ar v], and [Ne vi], and weak H i recombination lines, such as Pfα. Since these halos are detected in all IRAC bands, including the 4.5 μm band that does not include PAH features, it is likely that line emission dominates the near-IR IRAC bands. The 24 μm image of NGC 2438 resembles its [O iii] λ5007 image (Corradi et al. 2003), suggesting that [O iv] 25.9 μm line emission may make significant contributions to the MIPS 24 μm flux. The relative importance of nebular line and dust continuum emission at 24 μm is unknown. NGC 2438 is projected near the open cluster M46, but their different radial velocities suggest that they are unrelated (Kiss et al. 2008).

Zoom In Zoom Out Reset image size

We have defined three groups of PNe according to the spatial extent and surface brightness profile of the 24 μm emission relative to those of the Hα emission. To compare physical properties among these three groups of PNe, we have used the information in Table 1 and plotted the distribution of stellar T_eff and nebular sizes in Figure 7. A rough trend of size increase is seen from Group 1 through Groups 2 and 3 to the nondetections, suggesting an evolutionary sequence. The stellar T_eff, on the other hand, does not show any obvious trend. This is not surprising because stellar evolutionary tracks bend at the high-temperature extremes and different initial stellar masses lead to different extreme stellar temperatures; thus stellar temperature distributions are not sensitive indicators of evolution.

Zoom In Zoom Out Reset image size

We have measured the peak 24 μm surface brightness in the shell rim of each PN in Groups 1 and 2 and in the central diffuse-emission region of each PN in Group 3. These surface brightness measurements are listed in Column 9 of Table 1. Two Group 2 PNe, Lo 4 and NGC 246, have peculiar bright spots within their nebular shells; the surface brightnesses of these bright spots are about twice as high as the peak surface brightness in the shell rim. For the PNe that have distances available, we plot their 24 μm surface brightness versus their linear optical sizes in Figure 8. It is evident that the 24 μm surface brightness decreases with nebular size. This correlation is expected from nebular evolution, as nebular expansion leads to lower density and lower surface brightness. Column 9 of Table 1 shows that Group 1 PNe on average have the highest surface brightnesses, and Group 3 PNe have the lowest surface brightnesses. The PNe that do not show detectable diffuse 24 μm emission are still fainter than Group 3 PNe. The trend that the 24 μm surface brightness decreases from Group 1 through Groups 2–3 to the nondetections provides further support of an evolutionary sequence among these PNe.

Zoom In Zoom Out Reset image size

We have obtained Spitzer MIPS 24 μm observations of 36 Galactic PNe, and detected diffuse emission in 28 of them. Comparing the 24 μm surface brightness profiles with those in the Hα line, we find the 24 μm emission can be more extended, similar, or present only in the shell center, and divided these PNe into Groups 1–3, respectively. The optical shell sizes and 24 μm surface brightnesses of the PNe observed suggest that Groups 1–3 and nondetections represent an age or evolutionary sequence, with Group 1 being the youngest.

The diffuse emission in the MIPS 24 μm band is attributed to [O iv] 25.9 μm and [Ne v] 24.3 μm lines and dust continuum emission, but the relative proportions of these three components depend on the nebular excitation and dust content. The extended 24 μm emission in Group 1 PNe is most likely dominated by dust continuum, as high-ionization fine-structure line emission is expected to decline radially outward. Spectroscopic observations are needed to determine the ratio of nebular line emission to dust continuum for the main bodies of PNe in Groups 1 and 2. Using a Spitzer IRS spectrum of the central diffuse emission region of PN Sh 2-188, we illustrate that the [O iv] line is by far the dominant contributor to the MIPS 24 μm emission. It is likely that [O iv] line emission, and some [Ne v] line emission for PNe ionized by the hottest central stars, dominates the central diffuse 24 μm emission among all Group 3 PNe.

This research was supported by NASA grants JPL 1319342 and JPL 1343946. M.A.G. acknowledges support by grants AYA2005-01495 of the Spanish Ministerio de Educación y Ciencia (MEC) and AYA2008-01934 of the Spanish Ministerio de Ciencia e Innovación (MICINN). We thank the anonymous referee for helpful comments.