SNR 0453-68.5: AN ASYMMETRIC REMNANT AND ITS PLERION IN THE LARGE MAGELLANIC CLOUD

Upon examination of the fit parameters it becomes obvious that SNR 0453-68.5 is dominated by ISM emission outside of the central PWN containing region. In general, the elemental abundances match those expected from the LMC. The only elements found to vary are O and Fe. Significant departures from accepted LMC values are seen in regions A (>5σ), E (>5σ), G (4σ), and H (>5σ) as a depletion of oxygen while region E is also underabundant in iron (>3σ). There is no direct evidence that any of the regions outside of the PWN display significant ejecta emission. The departure from the LMC value of oxygen may not be as significant as suggested in Figure 5. When deriving abundances for the Russell & Dopita (1992) paper, the authors used data from an earlier survey of H ii regions and SNR in the Magellanic Clouds (Russell & Dopita 1990). Both classes of objects displayed large variations in oxygen abundance. In H ii regions the value was found to be 0.34^+0.22_{− 0.14} relative to solar with 1σ errors. Furthermore, N4A, the closest sampled H ii region to SNR 0453-68.5, is on the lower end of that range with an oxygen abundance of 0.22. When surveying SNR the authors found a relatively lower value of 0.26^+0.20_{− 0.12} (1σ). This is consistent with the LMC SNR spectral analysis performed by Hughes et al. (1998) using ASCA data. A sample of seven remnants in the LMC, including SNR 0453-68.5, displays reduced oxygen abundances averaging 0.26 ± 0.05 (the value found for SNR 0453-68.5 is 0.24 ± 0.07, 90% confidence). These authors find no evidence for ongoing dust destruction and suggest that the overall reduced abundances found in LMC remnants must reflect the actual abundances of the ISM. Therefore, the depleted oxygen abundance we find here may simply be consistent with LMC SNR in general, just on the lower end. Hughes et al. (1998) find that SNR 0453-68.5 has the lowest metallicity of the sampled SNR across the board suggesting that the local environment is itself depleted, which is supported by the Russell & Dopita (1990) findings on N4A. In fact, SNR 0453-68.5 lies in a region of the LMC that is somewhat devoid in tracers of star formation such as OB associations, CO emission, and H ii regions (Hughes et al. 1998). This lack of star formation may be responsible for the overly depleted abundances seen in SNR 0453-68.5.

Most of the best-fit models contain an equilibrium plasma. The state of CIE suggests that the remnant is either very old or the emission is dominated by shocked material in overdense regions. The latter is often seen in SNR that result from a cavity explosion (McCray & Snow 1979). The massive progenitor star has a strong wind which blows a cavity in the ISM. The supernova (SN) initially sends a blast wave through this vacated volume but then eventually catches up to the cavity wall. The wall consists of a highly anisotropic, clumpy medium with the densest clumps quickly reaching CIE. In these regions the ejecta quickly mixes with the dominant, shocked ISM. This scenario matches the observation with bright filaments of emission near the limb, most notably in regions A, G, and H. These contain dense regions of the cavity wall that have been shock heated by the SN blast wave. In contrast, the diffuse, yet bright, interior emission of regions such as B, C, D, and F appear closer to the center of the remnant. The presence of bright filaments in these regions may be due to line-of-sight confusion where the most intense emission is still coming from dense areas of the cavity wall but in projection. This is supported by the fact that these regions are best fit with multiple temperature components; a low temperature equilibrium plasma that has cooled in the dense cavity wall, and a higher temperature non-equilibrium plasma from interior, more rarefied gas that has been reheated by a reflected shock from the cavity wall. This hotter interior gas is evident in regions E and I. Similar interaction geometries occur for other shell-type SNR such as the Cygnus Loop (McEntaffer & Brantseg 2011). The limb-brightened morphology could also be due to the concentration of swept-up matter over time by the SN blast wave. This scenario is supported by the observation that hot interior gas as seen in regions E, and most notably I, is also ISM dominated. If the initial explosion expanded into a cavity then we might expect that interior emission originates from ejecta. However, we see no signs of ejecta-dominated plasma in these regions.

Calculations of shock velocity and density support this conclusion as shown in Table 3. Solving the Rankine–Hugoniot relations in the strong shock case with γ = 5/3 gives the post-shock temperature as a function of shock velocity, kT = (3/16)μm_pv², with μ = 0.6. We assume that T_e ∼ T_ion as expected through Coulomb collisions in a plasma with a temperature of ∼10⁶ K and n_e ∼ 1 cm⁻³. Such a plasma can equilibrate within a few hundred years or even more rapidly given higher densities (see Equation (36.37) in Draine 2011). If the electrons and ions are not in temperature equilibrium then the calculated velocities signify lower bounds. Our resulting shock velocities are typically a few hundred km s⁻¹, showing a significant deceleration from the initial blast wave velocity (>5000 km s⁻¹) and signifying a dense medium. This density is found using the norm parameter for each component, where norm = [10⁻¹⁴/(4πD²)]∫n_en_HdV. The integral contains the emission measure for the plasma which is dependent on the density and total emitting volume. A value of 50 kpc is used for D and we assume that n_e = 1.2n_H. The volume is calculated for each region assuming that the emission is confined within an annulus. It is within this annulus that the SN blast wave has interacted with the surrounding medium and it is assumed that the bulk of the X-ray emission, especially for the CIE plasmas, lies within this volume. A contribution from high-temperature, low-density gas from the interior of the remnant may exist, but this component is not observed to be significant from the spectral fits. If this component is indeed a significant contribution to the higher temperature vnei components then the densities calculated for these plasmas can be considered as upper limits. As depicted in Figure 8, the volume calculation is then $V = 2Af(\ssty\sqrt{R_o^2 - a^2} - \sqrt{R_i^2 - a^2})/{\rm cos}\theta$, where A is the projected area of the region on the sky, R_o is the outer limit of the X-ray emission (16.8 pc), R_i is an estimated inner radius of the emission annulus (15.8 pc), a is the radial distance to the center of the region, θ is the azimuthal offset of the region from the center of the remnant, and f is the filling factor of the X-ray-emitting gas which ranges from 0 to 1. Using these calculated volumes, the resulting densities range from 1.9/f − 7.6/f cm⁻³ in the equilibrium plasmas indicating that dense regions of interaction are responsible for the X-ray emission. In the regions with multiple temperature components (B, C, D and F) the equilibrium plasma is always at lower temperature and higher density than the nonequilibrium component. This is consistent with the above scenario where the SN blast wave interacts with a dense cloud near the limb of the remnant. The shock decelerates in the density enhancement and equilibrium is reached quickly. An RS propagates back into the lower density interior reheating this gas to higher temperatures and non-equilibrium conditions.

Zoom In Zoom Out Reset image size

Table 3. Calculated Region Parameters

Region	vvnei	vvapec	Volume × f	$n_{e_{vnei}} / f$	$n_{e_{vapec}} /f$	Age
	(km s−1)	(km s−1)	(pc3)	(cm−3)	(cm−3)	(years)
A	...	383	73	...	5.0	17200
B	480	296	130	2.3	4.5	22200
C	460	293	130	3.1	6.4	22500
D	846	380	94	1.1	4.9	17300
E	533	...	105	2.6	...	...
F	442	278	78	3.8	7.6	23600
G	...	382	318	...	1.9	17200
H	...	382	79	...	5.1	17200
I	584	...	237	0.75	...	...

Download table as:  ASCIITypeset image

We can also estimate the age of the remnant from the equilibrium shock velocities. Assuming that the remnant can be described by a Sedov solution (Sedov 1993), the time since shock heating can be calculated as t = (2/5)R/v, where R is the estimated shock distance from the center, 16.8 pc. As shown in Table 3 this results in age estimates between 17,200 and 23,600 years, significantly older than the 13,000 year age used in Lopez et al. (2011) and Gaensler et al. (2003). The spectral fits support an older age for the remnant given that we see no evidence for synchrotron emission near the limb, no evidence for fast shocks, and no evidence for ejecta-dominated emission (outside of the J regions), all of which are typically seen in young SNR. Given the Type II identification of the source there could be leftovers of the molecular cloud that formed the precursor star. The limb-brightened X-ray morphology, ISM-dominated emission, and the middle-aged nature of this remnant suggest that the X-ray emission results from the interaction of the SN blast wave with this cloud.

To further investigate this possibility we reanalyzed 24μm Spitzer data as shown in Figure 9. Williams et al. (2006) have previously analyzed Spitzer data at 24μm and 70μm. They attribute the emission to warm dust radiating a thermal continuum. This is consistent with our assumption that the limb-brightened X-ray emission is due to interactions of the SN blast wave with dense ISM clouds. We then use this assumption to compare the morphologies in the X-ray and infrared (IR). Figure 9 shows the data sets superimposed on one another. The image on the left is the raw IR data tracing the warm dust in the shock-heated ISM. The image on the right shows the X-ray data in green and the IR data superimposed in red. Generally, the IR data are well correlated with the X-ray data. The shell morphology of the remnant is easily noticeable in the IR with enhancements occurring in the same vicinities as those in the X-ray. The long, smooth shocked region in A appears to be a significant wall of material which shows up in the IR. There appear to be more discrete IR features in the vicinity as well, but these could be due to line-of-sight source confusion. The high-density clumps found in the X-ray emission of region B also appear in the IR data. These bright, dense features in regions A and B are consistent with the equilibrium conditions found from their X-ray spectra. Filaments seen in the X-ray, such as those in the north corresponding to regions C and F are also present in the IR. The only region where X-ray emission is bright but IR emission is lacking is the central core of emission surrounding the PWN. Also, the bright IR cloud in the southwest does not appear to correlate well with X-ray emission. There is X-ray emission present, but it shows no sign of interaction with a cloud such as a large-scale indentation at the limb, a bright knot of emission, or a distinct lack of X-rays due to absorption. Therefore, it is assumed that this IR feature lies behind the SNR and is not physically related.

Zoom In Zoom Out Reset image size

The IR limb brightening noticeable around most of the remnant suggests a nearly spherical distribution of the emitting clouds, thus supporting the idea of swept-up ISM or cavity explosion. The IR data at the limb tend to be outside (at larger radii) the X-ray emission indicating that the shock has preferentially heated the inner parts of these clouds. A lack of X-ray emission at slightly larger radii could be due extinction in the dense cloud or due to shock deceleration in the deeper recesses of the cloud leading to velocities that are too low to produce X-ray emission. In addition, there appears to be a general lack of IR emission in the south–southwest region of this remnant. When compared to the northern hemisphere of the remnant, the south is more dim in the IR. This relation is mirrored in the X-ray where the northern hemisphere also contains most of the emission (see Figure 1). If we assume that the plerion marks the origin of the SN blast wave then this distribution of IR/X-ray emission seems reasonable. For instance, the eastern limb is somewhat closer to the plerion than the western limb, but this is probably due to the larger, denser cloud in the east slowing the shock more. Also, the plerion appears to be located north of the center of emission. However, the low density in the south and southwest has allowed the blast wave to propagate further from the explosion origin thus increasing the distance from the limb to the center in comparison to the north. Therefore, the off-center position of the plerion may not be due to an asymmetric explosion imparting velocity to the neutron star. Instead, this analysis of the IR morphology would argue that the pulsar has very little proper motion relative to the remnant, especially when considering the >17,000 year age. The pulsar may be located very close to the original explosion origin unless its direction of motion is along the line of sight.

The fact that SNR 0453-68.5 is an older, ISM-dominated remnant is an interesting result because the only other previous study on the shell X-ray morphology of this remnant, Lopez et al. (2009, 2011), could not discern between explosion or environment as the cause of the asymmetry. The Lopez et al. studies use a "power ratio method" which measures X-ray surface brightness asymmetries using multipole moments seen in the soft X-ray (0.5–2.1 keV) and in Si xiii (data between 1.75 and 2.0 keV). A comparison of the quadrupole versus the octupole ratios of the emission allow the authors to categorize 24 SNR in the Milky Way and LMC into two distinct groups which are consistent with their explosion type—Type Ia versus CC. SNR 0453-68.5 is quite close to the dividing line in the broadband X-ray analysis. It is more distinctly classified as a CC remnant when considering Si xiii emission. Therefore, their result is that the distribution of X-ray surface brightness, which as an initial condition is presumed to be dominated by ejecta emission, leads to a CC classification for SNR 0453-68.5. We agree with the classification of this object and build on their result by showing that in the case of this particular Type II SNR, the morphology is not determined by the asymmetric distribution of ejecta arising from an asymmetric CC explosion, but rather inhomogeneity in the surrounding ISM. When we extract data from the 1.75 to 2.0 keV energy range, which is where the dominant Si xiii triplet lies, we see a surface brightness distribution that matches the low-energy X-rays. These data trace the interaction at the limb of the remnant which is dominated by the ISM, not clumps of ejecta. Furthermore, given the near constant position of the pulsar, the distribution of ISM seen in the IR, and the nearly symmetric blast wave propagation (after consideration of the cloud distribution suggested from the IR) it appears that the explosion that created SNR 0453-68.5 was quite likely very symmetric. Any evidence from the ejecta has been lost through mixing as seen from the ISM-dominated spectral fits. There appears to be no evidence from this remnant that the explosion mechanism that led to this CC SNR was inherently asymmetric.

The pulsar and its associated emission are detected in region J. The spectral index of the high-energy emission, α = 1.0  ±  0.2, is consistent with the range from other remnants containing a PWN in the LMC: α = 0.4–1.4 for a series of annular regions encompassing the PWN of SNR 0540-69.3 (Petre et al. 2007), α = 1.2 ± 0.3 in N206 (SNR 0532-71.0) (Williams et al. 2005), α = 0.57^+0.05_{− 0.06} for DEM L241 (Bamba et al. 2006), and α ∼ 1.2–1.6 over several regions from N157B (Chen et al. 2006).

Furthermore, it is consistent with a previous fit quoted in Gaensler et al. (2003), α = 0.9 ± 0.4. However, Gaensler et al. (2003) do not find the presence of a point source to be significant. To test the presence of a point source they model the central emission using a combination of an elliptical Gaussian, a point source, and background. They then state that the fit is equal or better if the point source is resolvable, thus leading to a conclusion that there is no strong detection of a point source. However, Lopez et al. (2011) also search for a pulsar in SNR 0453-68.5 and find a "bright pulsar" using wavdetect, consistent with the findings here. Even without a point-source detection the presence of a pulsar is likely given the detected synchrotron nebula which is powered by the embedded pulsar.

SNR 0453-68.5 exhibits a morphology that is consistent with a composite remnant. It has a thermal shell of emission from interaction with the ambient ISM and a central nebula of synchrotron emission surrounding a pulsar. Theoretical evolution of a PWN suggests that the expanding nebula will inevitably interact with an RS from the SNR (Gaensler & Slane 2006). The RS is formed during the Sedov phase of SNR evolution when the forward shock has swept up significant ISM mass in comparison to the ejecta mass thus forming a contact discontinuity. The RS will eventually propagate inward and reach the center of the remnant within several thousand years. However, it will encounter the PWN before this time resulting in compression of the nebula. The PWN will react with an increase in pressure and subsequent expansion. This battle continues for several thousand years resulting in complex morphology and mixing at the unstable interface. This mixing will occur between the synchrotron-emitting plasma of the PWN and the hot, shocked, thermally emitting ejecta near the interface. Any asymmetry in these variables such as ISM distribution, RS velocity, or pulsar velocity adds significant additional complexity. The obvious presence of the PWN in SNR 0453-68.5 and its age of >17, 000 years suggest that this interaction has or is currently taking place. The best-fit spectrum from region J supports this conclusion (Figure 2). The spectrum shows the presence of obvious emission lines at soft energies, most notably O viii and Ne ix, which leads to a significant thermal contribution to the fit. Furthermore, the abundances determined from this fit are significantly enhanced relative to the local ISM (O = 4 ± 1 and Ne < 7 relative to solar) suggesting that the plasma is ejecta dominated. Therefore, the spectral properties of region J combined with the expected evolutionary stage of the PWN, given a 17,000 year age, suggest that we are observing this remnant during the interaction of the PWN with the RS of the SNR.

Inspection of Figure 6 gives hints at complexities within the PWN. It is somewhat elongated along the northeast to southwest direction. There also appears to be a clump of emission located to the south–southwest of the point source. Although these features are intriguing, a longer Chandra observation would be necessary for investigating the structure in more detail. Given the short observation, we are not able to perform detailed extractions to exhaustively probe this RS/PWN interaction. However, using regions that are 1.5×, 2×, 3×, and 4× the area of region J we are still able to investigate the interaction and make comparisons to other PWN.

There are several trends that are noticeable from the spectral fits to the regions surrounding the point source (see Table 2). First, the abundances of O and Ne remain enriched relative to the expected abundances in the LMC, which is consistent with the region being ejecta dominated. Second, as the size of the region increases the relative contribution from the power-law component decreases in comparison to the thermal component. This is interpreted as an increase in the filling factor of hot, shocked, thermal ejecta as the regions encompass more of this plasma relative to the confined PWN emission closer to the center. It appears that the relative contributions become equal just above a region size of 1.5×J. Next, the photon index remains constant at Γ ∼ 2. If we assume that the density of the PWN is constant, then the increasing norm is due to the increased volume of nebula encompassed with each region. Therefore, subsequent regions sample non-thermal plasma at increasing radii from the pulsar, although the modest increase in the synchrotron emitting volume covers a radial increase of less than a factor of two in the PWN plasma. The constant photon index suggests that the distribution of non-thermal particles remains the same throughout the regions probed, i.e., the electron distribution is constant over these radii given Γ = α + 1 and p = 2α + 1, where p is the power-law index for the electron distribution. At the LMC distance of 50 kpc the semi-major axes of these ellipses probe scales from 0.4 to 1.2 pc. In contrast, other PWN containing remnants show a significant change in the photon index over these scales. The Crab Nebula varies from Γ = 1.8–2.4 over a 1.2 pc scale (Willingale et al. 2001) while SNR 0540-69.3 has a variation of Γ = 1.4–2.4 over 1.2 pc as well (Petre et al. 2007). Both remnants show a spectral hardening at smaller radii which is attributed to a radial dependence on the magnetic field. It is important to note that at this scale these remnants are dominated by the PWN emission versus thermal emission. In SNR 0453-68.5 we see a substantial contribution (>22%) from thermal plasma even at the smallest scales. Therefore, even though we probe to a similar spatial scale we see a much smaller physical volume dominated by the PWN in comparison to other remnants. We believe that this can be attributed to the age of SNR 0453-68.5. The interaction between the PWN and the RS has matured during the >17,000 year timescale. Over this time efficient mixing has occurred between the hot ejecta and the synchrotron plasma. If there is a PWN-dominated volume then it is contained within 0.4 pc. Outside of this the oscillations of the PWN/RS interaction have mixed the thermal and non-thermal plasmas. The constant temperature displayed by these regions may be indicative of this as well. Any variation in Γ or spectral hardening would exist on too fine of a spatial scale to be determined from this observation.

Another LMC PWN, contained within DEM L241, displays a similarly limited spatial extent as SNR 0453-68.5. Bamba et al. (2006) observed DEM L241 with XMM-Newton and resolved X-ray structures that consist of a hard central source surrounded by diffuse emission along with a tail of extended diffuse emission. The hard central source is determined to be a PWN well fit by a power law with Γ = 1.57. The extent of the central source is consistent with the XMM PSF of 5'' which translates to 1.2 pc at 50 kpc. Therefore, determining variability of the spectral index is limited by the spatial resolution. However, the authors determine that this remnant is very old, >1 Myr, which would suggest that any RS/PWN interaction would have taken place long ago and input power from the pulsar has significantly decreased. Furthermore, the situation for DEM L241 is quite different given its environment. Despite its old age, the diffuse emission from this remnant is ejecta dominated and centrally filled; it lacks any shell of emission. This may result from the fact that DEM L241 is contained within an OB association. The massive star winds have created a large, low-density bubble in which the SNR is expanding. The lack of ISM-dominated shell emission suggests that an RS is weak or non-existent, thus complicating the evolution story for the DEM L241 PWN and making connections with SNR 0453-68.5 difficult. Similar to DEM L241, N157B, another LMC PWN containing SNR, is located within a massive star association, but exhibits a resolved RS/PWN interaction (Chen et al. 2006). In the case of N157B the PWN exhibits a cometary shape with a pulsar power-law index of Γ = 1.73, a surrounding non-thermal coma with a power-law index of Γ = 2.18 and a non-thermal tail with Γ = 2.62. Chen et al. (2006) put forth the possibility that the shape of the nebula is not determined by pulsar proper motion but rather by an RS propagating from a dense cloud close to the explosion. The PWN and SNR have expanded freely in the low-density medium close to the massive stars, but are now encountering the edges of the bubble in this very complicated star formation region. Optical and IR data reveal the presence of a probable cloud responsible for this RS. Therefore, there are many unique morphologies between the LMC PWNe, all of which are highly influenced by the local environment. SNR 0453-68.5 is no exception to this and is a member of a class characterized by middle-aged, composite remnants with evolved RS/PWN interactions.

The morphology of the X-ray and infrared radiation as well as the central location of the detected point source indicate that there is little to no proper motion of the SNR 0453-68.5 pulsar transverse to the line of sight. Further evidence is the lack of a relic PWN. If the pulsar experienced a typical kick from the SN of a few hundred km s⁻¹ then we would expect it to be ∼25% of the way to the limb. It would generate a new PWN in its current position which would be observable in X-rays while leaving behind the original nebula which would still glow in the radio, hence the relic PWN. Therefore, any proper motion should show a displacement between the X-ray and radio emission. However, Gaensler et al. (2003) show that the radio and X-ray centroids coincide. This lack of offset, the >17,000 year age, and the mature PWN/RS interaction all argue that the pulsar could still be centrally located with very little, if any, velocity imparted from the SN. An LMC PWN that exhibits a pulsar with significant proper motion is N206. Williams et al. (2005) performed a combined X-ray, optical, and radio analysis of this remnant. They classify N206 as a mixed-composite SNR because it exhibits the composite nature with a PWN and shell while also showing a mixed morphology with central X-rays not associated with the PWN. They find a compact hard X-ray source near the limb of emission at the tip of an elongated radio feature. They interpret this as a pulsar with significant proper motion moving with respect to the SNR. The radio feature is the trail of synchrotron emission in the pulsar's wake and a bow shock structure in the X-ray is also evident. This morphology is in stark contrast to that shown by SNR 0453-68.5, which shows none of these features.