Far-Infrared Luminous Supernova Remnant Kes 17

The broad-band infrared (IR) observations are well suited for the study of the evolution of SNRs. Besides the general advantage of suffering less extinction effects, there are diverse important phenomena related to supernova remnants (SNRs) which are bright in the IR wavebands. In the near- and mid-IR regimes, for example, one can study the distributions of dense ejecta from core-collapse supernova explosions (e.g., Koo et al., 2007; Moon et al., 2009), circumstellar material produced by mass loss of the progenitors (e.g., Dwek et al., 2008; Lee et al., 2009a, b), shock interactions between SNRs and nearby molecular clouds (e.g., Burton et al., 1988; Oliva et al., 1990; Reach et al., 2005; Neufeld et al., 2007; Hewitt et al., 2009; Shinn et al., 2009), and hot dust synthesized in ejecta (e.g., Arendt et al., 1999; Rho et al., 2008), along with the emission from the central compact objects (e.g., Moon et al., 2004; Kaplan et al., 2006; Wang et al., 2006).

The far-IR observations of SNRs, on the other hand, are relatively rare but very efficient to investigate cold dust which often contributes to a significant portion of the total mass of the dust in SNRs. First, it is studied that the swept-up dust in SNRs is responsible for the most of cooling of SNRs and plays a critical role in the evolutions of SNRs in dense medium (Dwek et al., 1987). Secondly, the far-IR observations of SNRs appear to provide better means of studying the dust formed during the process of supernova explosion (e.g., Sibthorpe et al., 2010; Barlow et al., 2010). Finally, if an SNR is interacting with a nearby molecular cloud, its far-IR emission is expected to be significantly enhanced by the swept-up, inter-clump medium in the cloud or by the evaporation of dense clumps (e.g., Draine, 1981; Dwek, 1981), although the reality of these processes is yet to be confirmed observationally.

Kes 17, known as G304.6$+$0.1, is such an object in which broad-band IR observations can play an important role in our understanding of SNRs. Until recently this SNR has only been studied in radio continuum and OH line observations (Shaver & Goss, 1970; Milne & Dickel, 1975; Whiteoak & Green, 1996; Frail et al., 1996), revealing broken radio shell structures in the west and south. The recent $Spitzer$ observations, however, detected very bright IR emission in the 3–8 $\mu$m range (Lee, 2005; Reach et al., 2006), which makes it in fact one of the brightest SNRs in the wavelength range. The follow-up spectroscopic observations in the 5–30 $\mu$m range also detected many strong H${}_2$ emission lines (Hewitt et al., 2009). Very recently, the XMM-Newton X-ray observations detected centrally filled thermal X-ray emission inside the non-thermal radio shell, listing it as a member of the class of mixed-morphology SNRs (Combi et al., 2010). All these indicate that Kes 17 is interacting with nearby molecular gas and that more IR observations are needed to better understand its nature. In this paper, we present extensive broad-band IR observations of Kes 17 using the $AKARI$ satellite, together with new analyses of archival $Spitzer$ and radio continuum data. For the distance to Kes 17, only a lower limit of 8 kpc, corrected for the distance between the Sun and the Galactic center of 8.5 kpc, is known from the observations of H i line (Caswell et al., 1975). We adopt a distance of 8 $d_8$ kpc where $d_8$ is a scaling factor in this paper.

Figure 1 presents the $AKARI$ and $Spitzer$ near- to far-IR (4.5–160 $\mu$m) band images of Kes 17, together with the $ATCA$ 20 cm radio-continuum image. The most conspicuous feature in the IR images is the bright emission of $\sim$ 5′ in the western shell apparent in all of the IR bands. The emission appears to be clumpy and filamentary in the near- and mid-IR bands up to 24 $\mu$m, whereas it becomes unresolved in the longer (65–160 $\mu$m) wavelength images due to increasingly large beam sizes. This IR emission in the western shell partly overlaps that of the radio continuum emission. In addition to the bright western shell, there is relatively weak mid-IR (15 and 24 $\mu$m) emission in the southern shell overlapping the radio continuum emission in this region.

There exists noticeable background emission in the direction of Kes 17 in all the images longer than 4.5 $\mu$m (Figure 1), especially in the northeast. In order to identify the southern shell emission in the 15 and 24 $\mu$m images more clearly, we subtract the background emission from both the images as follows: first, given that both the 4.5 and 8 $\mu$m emission of Kes 17 is due to the same origin (i.e., emission from shocked H${}_2$; see § 4 for the details) in the western shell, we obtain a linear correlation with a correlation coefficient $r$ $\simeq$ 0.86 between the 4.5 and 8 $\mu$m emission of the shell. Based on the correlation, we scale the 4.5 $\mu$m emission to 8 $\mu$m and subtract the scaled emission from the observed 8 $\mu$m emission to create a background image at 8 $\mu$m. Next, we obtain a linear correlation of $r$ $\simeq$ 0.89 between the background 8 $\mu$m emission and the 24 $\mu$m emission outside the boundaries of Kes 17, and then scale the 8 $\mu$m emission to 24 $\mu$m using the correlation. We subtract the scaled emission from the observed emission at 24 $\mu$m, which finally gives the background-removed 24 $\mu$m image of Kes 17 (Figure 2). The peak surface brightnesses of the western shell after the background subtraction are 6.1 $\pm$ 1.5 and 6.4 $\pm$ 1.5 MJy sr${}^{-1}$ for 15 and 24 $\mu$m bands, respectively. We apply the same method to obtain the background-removed $AKARI$ 15 and 24 $\mu$m images. In all these processes, we convolved the images of different beam sizes with that of the 24 $\mu$m images. In Figure 2 we can now easily identify the existence of the southern shell at 15 and 24 $\mu$m which overlaps that of the radio continuum. For far-IR images of 65, 90, 140, and 160 $\mu$m, we estimate the background emission by taking the median value in the region between 4${}^{\prime }$ and 6${}^{\prime }$ radius from the center of Kes 17. After background subtraction, the peak surface brightnesses of the western shell are 108 $\pm$ 24, 135 $\pm$ 45, 173 $\pm$ 84, and 117 $\pm$ 55 MJy sr${}^{-1}$ for 65, 90, 140 and 160 $\mu$m bands, respectively.

Table 4 presents the measured (or upper limits) IR fluxes of the western and southern shells between 15 and 160 $\mu$m, together with the estimation of contributions of line emission in the bands. The errors in the flux measurements are mainly caused by the uncertainties in the background measurements and also by those associated with the calibrations of the $AKARI$ and $Spitzer$ observations. We estimate the line contributions to the observed fluxes of the 15 and 24 $\mu$m bands using the $Spitzer$ IRS 8–38 $\mu$m spectra. We construct the IRS spectra by subtracting the background obtained outside the shell structure and integrate them over the bandpass. The estimated line contributions are 75 $\pm$ 21 (15 $\mu$m) and 38 $\pm$ 10 % (24 $\mu$m for both $AKARI$ and $Spitzer$), respectively. The line emission in these bands is dominated by H${}_2$ and ionic lines. We also estimate the contribution of line emission of the 60 and 90 $\mu$m bands to be 16 $\pm$ 8 and 3 $\pm$ 2 %, respectively, using $Spitzer$ MIPS SED 52–97 $\mu$m spectra. The line contributions in these bands are mostly due to [O i] at 63 $\mu$m. For the 140 and 160 $\mu$m bands, we expect the line contributions to be less than 5 % by scaling the flux of the [O i] 63 $\mu$m line that we obtain above to those of [O i] 145 $\mu$m and [C ii] 153 $\mu$m lines in a shocked region (Hollenbach & McKee, 1989; Allen et al., 2008). Figure 3 shows the flux distribution of the western shell excluding the line contributions.

The IR synchrotron flux expected from an extrapolation of the radio fluxes applying synchrotron power-law index of $-$0.54 (Shaver & Goss, 1970) is less than a few percents of the measured IR values, supporting that the observed continuum is dominated by dust emission. In order to calculate temperatures of the dusts responsible for the mid- and far-IR emission of Kes 17, we perform the spectral energy distribution (SED) fits of the flux distribution (Figure 3) with two modified blackbody components. For the emissivity of the dusts in the SED fits, we adopt a mixture of carbonaceous and silicate interstellar grains (Draine, 2003). As a result, we obtain acceptable fit (reduced chi-square $<$2) with dust temperature components of 79 $\pm$ 6 K and 27 $\pm$ 3 K (Figure 3). The mass of the hot (79 K) dust is $(6.2\pm 4.6)\times {10}^{-4}$ $d_8^2$ $M_{\odot }$ and that of the cold (27 K) dust is 6.7 $\pm$ 4.0 $d_8^2$ $M_{\odot }$. The luminosities of the dust emission are (9.5 $\pm$ 6.1) $\times$ 10${}^2$ $d_8^2$ $L_{\odot }$ and (8.1 $\pm$ 5.0) $\times$ 10${}^3$ $d_8^2$ $L_{\odot }$ for the hot and cold components, respectively. It is worthwhile to note that the hot dust temperature could have been overestimated if there is any contribution from stochastically heated small grains to the 15 and 24 $\mu$m continuum fluxes (Dwek, 1986).

Figure 4 shows near-IR (2.5–5.0 $\mu$m) $AKARI$ spectra of five points (A, B1, B2, C1, C2 in Figure 1) in the western shell of Kes 17 obtained after subtraction of the background spectrum. (See Table 4 for their coordinates.) We detect several H${}_2$ lines in the spectra including the pure rotational transitions of 0-0 S(9–14) and ro-vibrational transitions of 1-0 O(3, 5, 6). Table 5 lists the observed intensities of the line emission.

Figure 5 presents the excitation diagram of the pure rotational and ro-vibrational H${}_2$ lines of B2 between the upper energy level of 6,000 K and 20,000 K. For this we correct the extinction effect using the H column density of 3.6$\times$ 10${}^{22}$ cm${}^{-2}$ estimated in previous X-ray observations of Kes 17 and the Galactic extinction curve (Hewitt et al., 2009; Draine, 2003). The linear correlation between the upper energy level and the upper column density represented by a straight line in Figure 5 indicates the level population of the local thermodynamic equilibrium at the excitation temperature of 2200 $\pm$ 400 K. The spectra of other positions, which have smaller number of detected lines, show similar excitation temperatures to that of B2.

We show in this paper that Kes 17 is one of the most luminous SNRs in the IR wavebands, reaching the far-IR luminosity of $\sim$ 8100 $L_{\odot }$. Its IR emission is concentrated on the western and southern shell structures. The western shell is bright in the mid- and far-IR continuum and the near-IR H${}_2$ line emission, whereas the southern shell is visible only in the mid-IR continuum emission. The far-IR continuum and near-IR H${}_2$ line emission of the western shell is related to its interaction with nearby molecular gas. It is apparent that there exists dense shocked molecular gas, excited to $\sim$ 2,000 K, in this region that produces the observed H${}_2$ line emission. If the molecular gas is clumpy, the observed far-IR emission can be produced by the dust associated with inter-clump molecular gas of lower densities. The far-IR emission may originate from the dust emission of the swept-up inter-clump molecular gas, the dust emission from an evaporating cloud, or the dust emission exposed to strong radiative shocks. The mid-IR emission of Kes 17 is bright in both the western and southern shells and overlaps radio continuum emission which forms a partly-broken circular shell structure. This suggests that the mid-IR and radio continuum emission delineates the boundary of Kes 17 made by the swept-up material.

The SED distribution of the IR emission of Kes 17 indicates the existence of two dust components: hot (79 $\pm$ 6 K) component of $(6.2\pm 4.6)\times {10}^{-4}$ $d_8^2$ $M_{\odot }$ and cold (27 $\pm$ 3 K) component of 6.7 $\pm$ 4.0 $d_8^2$ $M_{\odot }$. The former is responsible for most of the mid-IR continuum emission, while the latter is for the far-IR emission. The observed flux distribution of the H${}_2$ lines, on the other hand, can be explained with a thermal admixture of the shocked molecular gas of different temperatures in the 100–4,000 K range. This thermal admixture of the shocked molecular gas may be a natural consequence of the interactions between Kes 17 and clumpy molecular gas in which various shock velocities can easily produce thermal admixture of the molecular gas. Overall, it appears that combined far-IR continuum and near-IR spectroscopic observations are very useful to studying SNRs interacting with molecular gas, as was the case for Kes 17 in this paper, providing information for both the gas and dust components associated with the shock interactions.

This work is based on observations with $AKARI$, a JAXA project with the participation of ESA. This work is based in part on observations made with the $Spitzer$ Space Telescope, which is operated by the Jet Propulsion Laboratory, California Institute of Technology under a contract with NASA. The Australia Telescope Compact Array is part of the Australia Telescope which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO. We thank all the members of the $AKARI$ project. We also thank an anonymous referee for constructive comments. H.-G.L. was supported by the Early Research Award Program (ERA07-03-270) to D.-S.M. from Ministry of Research and Innovation of the Ontario Provincial Government. D.-S.M. acknowledges the support by NSERC through Discovery program 327277. B.-C.K. was supported by the National Research Foundation of Korea (NRF) Grant (NRF-2010-616-C00020).

Facility: $AKARI$, $Spitzer$, $ATCA$