Nebular versus Stellar Wind Abundances in NGC 6543

Over the last several years, sensitive X‐ray observations of planetary nebulae (PNe) by Chandra and XMM‐Newton have led to results that both astonish and puzzle astronomers. One such result was presented by Chu et al. (2001) on the Chandra‐observed X‐ray spectrum of the Cat's Eye nebula, NGC 6543. Chu et al. (2001) found that the X‐ray spectrum of NGC 6543 could be adequately fitted using stellar wind abundances but not with abundances that had been reported for the nebular material in this object. Specifically, they found that because He and N are expected to be significantly more enhanced in the stellar wind than in the nebular material (Aller & Czyzk 1983; Pwa, Pottasch, & Mo 1984; Manchado & Pottasch 1989; de Koter et al. 1996), only models derived from stellar wind abundances would match the observed spectrum. Chu et al. (2001) emphasized, however, that the low temperature they derived for the X‐ray–emitting material (1.7 × 10⁶ K) is puzzling since the expected postshock temperature for the fast (1750 km s ^-1) stellar wind of NGC 6543 is on the order of 10⁸ K. They also note that the location of the X‐ray–emitting gas requires that a significant fraction of the X‐ray–emitting material be nebular. In this paper, we reanalyze the spectrum of NGC 6543 using updated abundances found for the nebular material in an attempt to resolve this apparent anomaly.

Chandra observed NGC 6543 for 46.0 ks (obs. ID 630) with the Advanced CCD Imaging Spectrometer (ACIS; Garmire et al. 1988) as the focal plane instrument. The telescope bore sight was positioned near the center of the spectroscopy CCD array (ACIS‐S); NGC 6543 was imaged on the central back‐illuminated CCD (S3), which provides moderate spectral resolution (E/ΔE) of ∼4.3 at 0.5 keV and ∼9 at 1.0 keV. Analysis of these data has already appeared in Chu et al. (2001). Data from the observation have been reprocessed by the Chandra X‐ray Center (CXC) subsequent to the publication by Chu et al. (2001); the analysis presented here was performed on the reprocessed files. Each spectrum was extracted using the Chandra Interactive Analysis of Observations (CIAO) software within a region judged to contain all the X‐ray flux from the nebula (see Fig. 1). We note that the current extraction (CIAO ver. 2.2.1) reflects upgrades to the calibration system that were made after Chu's paper on this system was published. The extracted events are aspect corrected, bias subtracted, energy calibrated, and limited to grade 02346 events (ASCA system). The background count rate as determined from a large, off‐source annulus region (30 and 50 pixel radii) was negligible in comparison to the source count rate.

Zoom In Zoom Out Reset image size

Spectra were extracted from the entire nebula and from three regions corresponding to the central elliptical shell, the northern extension, and the southern extension (see Fig. 1). The extractions for the entire nebula and the central region excluded counts from the central point source. The divisions were made following the regions listed by Chu et al. (2001).

The extracted spectra were first modeled with the abundances used by Chu et al. (2001): they used abundances reported by Aller & Czyzak (1983), Pwa, Pottasch, & Mo (1984), and Manchado & Pottasch (1989) for nebular material and abundances reported by de Koter et al. (1996) for the stellar wind. For purposes of comparison, we used both the model used by Chu et al. (2001) (VRAYMOND with adopted absorption cross sections from Balucinska‐Church & McCammon 1992, hereafter RBM) as well as a VMEKAL model with absorption cross sections adopted from Morrison & McCammon (1983, hereafter VMM). Both RBM and VMM are appropriate for optically thin thermal plasma in ionization equilibrium. When abundance values for elements that are fitted by these models are not available, we set the abundance to solar. Abundances were held fixed, and only temperature, intervening column density, and normalization were allowed as free parameters. We are unable to reproduce the results of Chu et al. (2001), finding that stellar wind abundances they used modeled the observed X‐ray spectra as well as their nebular abundances (Figs. 2 and 3). The abundances and best‐fit parameters are listed in Tables 1 and 2.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We then fitted the observed spectra with nebular abundances determined from optical observations by Hyung et al. (2001). Hyung et al.'s abundances, given relative to hydrogen by number, were converted to abundances relative to solar following Grevesse & Sauval (1998); the converted abundances are listed in Table 1. We again performed two sets of fits: RBM and VMM. Fits using Hyung et al.'s abundances produced good results, giving a lower χ² value than both the nebular and the stellar wind abundances used by Chu et al. (2001). The temperature did not vary significantly over any of the models, but column density decreased by 10% for fits using VMM (Fig. 4).

Zoom In Zoom Out Reset image size

Since the Hyung et al. (2001) abundances were based on observations taken in only one wavelength and sampled only at two bright regions of the nebula, they may not be representative of the entire nebula. We therefore used abundances for the nebula determined by Bernard‐Salas et al. (2003) from multiple observations in the infrared, optical, and ultraviolet. These provide a better sample of the nebular region, and the multiwavelength determination is more robust. However, we note that Bernard‐Salas et al. is not yet published, and these numbers are subject to small changes (J. Bernard‐Salas & S. R. Pottasch 2003, private communication). Again, the abundances provided by Bernard‐Salas and our best‐fit parameters are listed in Tables 1 and 2. The model fits are shown in Figure 5. The similarity between the Hyung and Bernard‐Salas nebular abundances and the improvement in fits to the data using these abundances give us confidence in our result.

Zoom In Zoom Out Reset image size

We fitted the spectra extracted from the central shell and the northern and southern extensions using the same abundances as those adopted for the entire nebula. The main result from the regional fits is that while temperature remains constant over the emitting region, absorption appears to vary significantly. Specifically, the intervening absorption is greater for the southern extension and the central shell than for the northern extension (Table 3). This result is in agreement with the predictions of Miranda & Solf (1992) and with the work of Chu et al. (2001).

Chu et al. (2001) pioneered Chandra X‐ray spectral work on the planetary nebula NGC 6543 and found that models using abundances reported for the stellar wind fitted the observed spectrum better than those using abundances reported for the nebula. They concluded that the X‐ray emission from this PN arose primarily from stellar wind material but emphasized that this result was troubling when considered with the X‐ray temperature derived from their model and with the location of the X‐ray–emitting gas. In an attempt to resolve this issue, we re‐modeled the X‐ray spectrum of NGC 6543 using abundances reported by Hyung et al. (2001) and J. Bernard‐Salas & S. R. Pottasch (2003, private communication). The resulting models fitted our spectra better than either the nebular or the stellar wind models proposed by Chu et al. (2001), suggesting that the X‐ray emission from NGC 6543 arises primarily from nebular gas. We point out that our result does not require that diffuse X‐ray emission from this PN originate solely from nebular material. Given the quality of the spectrum, a definite separation of nebular and stellar wind abundances cannot be achieved, and in reality, the X‐ray–emitting material may be some mix of stellar wind and nebular material. Our finding that observed X‐ray spectral properties from NGC 6543 allow for emission predominantly from nebular material resolves the low‐temperature anomaly identified by Chu et al. (2001).