3 Diagnostic Mapping Results

An excitation diagnostic diagram consisting of [OIII]/H$\alpha$ versus [SII]/H$\alpha$ was first produced by Phillips & Cuesta (1999) to determine bow-shock regions in the bipolar outflows of the PN M2-9. Such an excitation diagnostic diagram was also used to distinguish between the shock- and photo-ionized regions in K4-47 (Gonçalves et al., 2004), which is a PN consisting of a high-ionization core and a pair of LISs. To discriminate photoionized nebulae from shock-excited PNe, Raga et al. (2008) constructed a set of diagnostic diagrams, including [OIII]/H$\alpha$ versus [SII]/H$\alpha$, based on axisymmetric simulations of fast, dense LISs moving through a low-density environment, and away from an ionizing source. These diagnostic diagrams have been employed to distinguish low-ionization knots from photo-ionizated nebulae in numerous studies (e.g., Gonçalves et al., 2009; Akras & Gonçalves, 2016; Ali & Dopita, 2017). More recently, Akras & Gonçalves (2016) employed the [OIII]/H$\alpha$ versus [SII]/H$\alpha$ excitation diagnostic diagram from Raga et al. (2008) to study a number of PNe with LISs.

Figure: Excitation diagnostic diagram and BPT diagram of the inner region of NGC 5189, covering a $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region (see Fig. 1). Top panel: Excitation diagnostic diagram presents logarithmic ratio maps of [OIII]/H$\alpha$ and [SII]/H$\alpha$. Bottom panel: BPT diagram presents logarithmic ratio maps of [OIII]/H$\beta$ and [SII]/H$\alpha$. Solid black lines show the boundaries of LINER-like and Seyfert-like activities from (Kewley et al., 2006). The solid red line depicts the nebular photon-shock dividing line in each panel chosen based on the shock models from Raga et al. (2008), as described in the text. The star points ($\ast$) show the mean flux ratios of the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region, while the cross ($\times$) and plus ($+$) points depict the flux ratios from García-Rojas et al. (2012) and Kingsburgh & Barlow (1994), respectively.

 

Figure: Spatially-resolved diagnostic map of the inner region of NGC 5189, covering a $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region centered on its [WO] central star, as shown in Fig. 1. The two pixel groups are color-coded according to their locations on the excitation diagnostic diagram in Fig. 2. Red pixels correspond to fast, low-ionization regions, and green pixels correspond to photo-ionized regions. Black pixels have either only one diagnostic line ([OIII] or [SII]), or one or both of diagnostic lines without at least $1\sigma$ of the mean value of the sky region.

Figure 2 (top) presents an excitation diagnostic diagram with a 2-D histogram of log([OIII]/H$\alpha$) versus log([SII]/H$\alpha$) plotted for $1\sigma$-masked WFC3 pixels extracted from the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region, centered on the [WO] central star of NGC 5189 (RADEC/J2000: 13h33m32.9s $-$65$^{\circ}$58 $\hbox{$^\prime$}$27.1 $\hbox{$^{\prime\prime}$}$). We selected a $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ extraction region, oriented with a position angle of $\sim 45^{\circ}$ (from the north toward the east in the equatorial coordinate system) to focus on the filamentary structures around the central star. The spatial resolution of each WFC3 pixel is $0.0396\hbox{$^{\prime\prime}$}\times0.0396\hbox{$^{\prime\prime}$}$, corresponds to $\sim 6.75 \times 10^6$ pixels in the extracted region, but not all pixels are statistically $1\sigma$ significant and have both [SII]/H$\alpha$ and [OIII]/H$\alpha$ flux ratios. To disentangle fast LISs from photoionized regions, we adopted two different regions of the excitation diagnostic diagram guided by the calculations presented in Raga et al. (2008). Specifically, we delineated fast LISs and photoionized regions. This diagram is similar to the Baldwin-Phillips-Terlevich (BPT) diagram (Baldwin et al., 1981), which is used to distinguish Seyfert-type and low ionization emission line region (LINER) classifications of starbursts and active galactic nuclei (AGN) galaxies (Kewley et al., 2006; Kewley et al., 2001). Recently, the BPT diagram has been used to spatially resolve Seyfert-type and LINER-type activities of the inner region of the extended narrowline region (ENLR) of NGC 3393 (Maksym et al., 2016; Maksym et al., 2017). For comparison, in Figure 2 (bottom) we also show the corresponding BPT diagram of NGC 5189, including the shock-excited and photo-ionized regions according to the excitation classification from Raga et al. (2008).

We used the excitation diagnostic diagrams presented in Figure 2 to delineate the fast, dense LISs from the photo-ionized medium of the inner region of the nebula NGC 5189. All the valid pixels, which possess both [SII]/H$\alpha$ and [OIII]/H$\alpha$ flux ratios, are included in the excitation diagnostic diagrams. We adopted a nebular photon-shock dividing line according to shock models (Raga et al., 2008). Although this division was not clearly defined by Raga et al. (2008), we adopted a photon-shock dividing line that is parallel with the Seyfert-LINER classification line, $1.89 \log($[SII]/H$\alpha$ $)+0.76 = \log($[OIII]/H$\beta$$)$, defined by Kewley et al. (2006). We use this line as an empirical division between photo-ionization and potential shock-ionization regimes within the nebula, hereafter referred to as the `nebular photon-shock dividing line'. Our empirical estimate based on the shock models for the location of this dividing line is $1.89 \log($[SII]/H$\alpha$ $)+2.46 = \log($[OIII]/H$\beta$$)$ for the BPT diagram (See Figure 2). For the excitation diagnostic diagram shown in Figure 2 (top), the same dividing line corresponds to $1.89 \log($[SII]/H$\alpha$ $)+2.0 = \log($[OIII]/H$\alpha$$)$.

Figure 3 shows the results of classifying pixels based on the diagnostic map of NGC5189. Red indicates fast, low-ionization activity, and green is typical of photo-ionized regions. We excluded pixels without at least $\sim 1\sigma$ of the mean value of the sky region, and without both the [SII]/H$\alpha$ and [OIII]/H$\alpha$ diagnostic ratios.

Based on the distributions in Figure 2, mean flux ratios measured from the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ extracted region would be [SII]/H$\alpha$ =  $0.13 \pm 0.11$, [OIII]/H$\alpha$ =  $3.98 \pm 1.63$ and [OIII]/H$\beta$ =  $11.32 \pm 4.62$ (indicated by $\ast$ in Fig. 2; errors correspond to the average absolute deviations). Flux ratios from the literature are [SII]/H$\alpha$ = 0.36, [OIII]/H$\alpha$ = 4.05 and [OIII]/H$\beta$ = 11.97 (indicated by $\times$ in Fig. 2; García-Rojas et al., 2012), [SII]/H$\alpha$ = 0.16, [OIII]/H$\alpha$ = 3.95 and [OIII]/H$\beta$ = 11.34 (indicated by $+$ in Fig. 2; Kingsburgh & Barlow, 1994). We note that García-Rojas et al. (2012) used a $1 \hbox{$^{\prime\prime}$}\times 5 \hbox{$^{\prime\prime}$}$ slit, whereas Kingsburgh & Barlow (1994) employed different slits with a total length of $18.4\hbox{$^{\prime\prime}$}$ and widths of $1\hbox{$^{\prime\prime}$}$ and $6.7\hbox{$^{\prime\prime}$}$. Despite the different slit configurations, the [OIII]/H$\alpha$ flux ratios are approximately the same. Our mean [SII]/H$\alpha$ flux ratio of the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region is roughly similar to what is reported by Kingsburgh & Barlow (1994), but a factor of $\sim 3$ lower than the flux ratio derived by García-Rojas et al. (2012). This discrepancy could be due to the shorter slit used by García-Rojas et al. (2012), and it is possible that the slit was placed on one of the nebular envelopes (see Figure 4) which is dominated by the fast, low-ionization regime. The longer and wider slits ( $18.4\hbox{$^{\prime\prime}$}\times1\hbox{$^{\prime\prime}$}$ and $18.4\hbox{$^{\prime\prime}$}\times6.7\hbox{$^{\prime\prime}$}$) employed by Kingsburgh & Barlow (1994) presented the flux ratios close to our results.

Figure: Logarithmic flux ratio map of [SII]/H$\alpha$ produced from the HST observations for the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region shown in Fig. 1. The contour lines illustrate the boundaries of photo-ionized region and fast, low-ionization region based on the pixel classification seen in Fig. 3. The main morphological features, large low-ionization envelope, and its smaller counterpart are labeled.

Figure: Logarithmic flux ratio map of [OIII]/H$\alpha$ produced from the HST observations for the $120\hbox {$^{\prime \prime }$}\times 90\hbox {$^{\prime \prime }$}$ region shown in Fig. 1. The contour lines illustrate the boundaries of photo-ionized region and fast, low-ionization region based on the pixel classification seen in Fig. 3. The main morphological features, large low-ionization envelope, and its smaller counterpart are labeled.

The [SII]/H$\alpha$ and [OIII]/H$\alpha$ logarithmic flux ratio maps are presented in Figures 4 and 5, respectively, which were produced from the dereddened, continuum-subtracted flux maps shown in Figure 1. The boundaries of the photo-ionized and the fast, low-ionization regions based on Figure 3 are illustrated as contour lines in both the flux ratio figures. The shock criterion /H$\alpha$ $)\gtrsim -0.4$;][]Mathewson1973,Fesen1985 is satisfied in several places within the fast, low-ionization region, suggesting that the dense, filamentary structures are interacting with the surrounding low-density medium.

We can easily identify the following main morphological features of the inner excitation regions:

(1) A large dense, low-ionization envelope with a maximum diameter of $\sim 55$ arcsec extended from the central star wherein the structures are expanding toward the northeast (see e.g. long-slit observation; Sabin et al., 2012).

(2) A smaller compact, dense low-ionization envelope with a maximum radius of $\sim 40 \hbox{$^{\prime\prime}$}$ is extended from the central star and is expanding toward the southwest, likely the large low-ionization envelope's counterpart.

(3) These two low-ionization envelopes are surrounded by the highly ionized, low-density gas (showing $\log($[OIII]/H$\alpha$ $)\geqslant 0.8$ in Figure 5).

(4) Multiple low-ionization filamentary and knotty structures in both of the two low-ionization envelopes.

As only one long-slit kinematic observation is available for this region (see Sabin et al., 2012), we could not constrain the 3-D geometry of these low-ionization envelopes through morphokinematic modeling. Additional long-slit observations are necessary in order to disentangle their 3-D morphological structures.