Physical Structures of the Type Ia Supernova Remnant N103B

The HST Hα images of N103B were obtained in Program 13282 (PI: Chu) on 2013 July 11, using the UVIS channel of Wide Field Camera 3 (WFC3) and the F656N filter. The 162″ × 162″ field of view is large enough to cover not only the entire N103B SNR but also part of the nearby superbubble around the NGC1850 cluster. The pixel size, 004, corresponds to 0.01 pc in the LMC (at 50 kpc distance).

The Hα imaging observations were dithered with the WFC3-UVIS-GAP-LINE pattern for 3 points and point spacings of 2414. The total exposure time for each position in N103B was 1350 s. All exposures used the FLASH = 11 option to make sure a background of roughly (10–12)e⁻ was present to compensate for charge transfer efficiency (CTE) issues in WFC3.

Using HST, we have also obtained F475W $(B)$, F555W $(V)$, and F814W $(I)$ band images of N103B to study its underlying stellar population. The UVIS channel of WFC3 was used for continuum-band images. All continuum-band observations have a 162″ × 162″ field of view as does the F656N observation. The F475W and F814W observations each had a total exposure time of 1050 s. The F555W observation had a total exposure time of 1117 s. The FLASH = 5 option was used for the F475W observation, and the FLASH = 4 option was used for the F814W observation to mitigate the CTE issues in WFC3.

We used DOLPHOT, a version of HSTPHOT with HST-specific modules (Dolphin 2000), to carry out point-spread function photometry on each of the HST images. The photometric parameters were selected following the recommendations of Williams et al. (2014). The catalog of detected objects was filtered to include only well-measured stars using the criteria of signal-to-noise ratio (S/N) > 5, ${{sharp}}^{2}\lt 0.1$, and ${crowd}\lt 1.0$, as recommended by Williams et al. (2014). Definitions of these parameters can be found in Dolphin (2000). These photometric measurements were used to construct color–magnitude diagrams (CMDs) in order to study the stellar population.

2.2.1. Imaging

2.2.2. Long-slit Echelle Spectroscopic Observations

2.2.3. Multi-order Echelle Spectroscopic Observations

The echelle line images in Figures 5 and 6 show a narrow, nearly constant-velocity Hα component that is present both inside and outside the SNR boundaries. The narrow velocity profile and the spatial distribution indicate that this ionized gas component has an extended distribution and is not interacting with the SNR N103B. The surface brightness gradient along the slit (Figure 5) is consistent with the large-scale Hα surface brightness variation shown in Figure 1, brighter toward the superbubble of NGC 1850.

The SNR N103B, however, may not be physically associated with this ionized gas component, as its Balmer-dominated shell (discussed in the next subsection) requires the SNR to be in a mostly neutral ISM (Chevalier et al. 1980). We identify this narrow, nearly constant-velocity Hα component as the background interstellar gas. The average [N ii]/Hα ratio of this background gas is 0.14 ± 0.01, and the average Hα velocity profiles have FWHM ∼ 30 km s⁻¹.

Within the SNR boundary, there exists enhanced Hα emission at velocities similar to that of the background ionized interstellar gas, but its velocity profiles have broader FWHM, ∼40 km s⁻¹, and its surface brightness variations correlate well with that of the filamentary elliptical Hα shell of N103B. Unlike the background interstellar gas, this broad Hα shell component does not have [N ii] counterparts (Figures 5, 6, 7). This is a characteristic of Balmer-dominated Type Ia SNRs (Tuohy et al. 1982) and has been reproduced by models of collisionless SNR shocks into a mostly neutral medium (Chevalier et al. 1980). A Balmer-dominated component should exhibit a core and very broad wings, as shown in low-dispersion spectra (Smith et al. 1991); however, our high-dispersion echelle spectra are not sensitive enough to detect the broad wings. Our intermediate-dispersion spectra of N103B did detect the broad wings, and these will be presented in a future paper.

Zoom In Zoom Out Reset image size

The most spectacular kinematic features in the echelle line images are discrete regions of Hα emission over velocity spreads up to ∼500 km s⁻¹ (Figure 5). These regions correspond to the nebular knots resolved in the HST Hα images. These broad Hα components have [N ii] counterparts, although the velocity spreads of the [N ii] line appear smaller owing to their lower S/Ns. The [N ii]/Hα ratios of these broad components are up to 0.18.

It is remarkable that most of the shocked material in the nebular knots is redshifted with respect to the background or the Balmer shell component. The distribution of the CSM is clearly not spherically symmetric; it is one-sided along the line of sight, as well as along the east–west direction.

To illustrate the asymmetric distribution of the CSM, we display in Figure 8 the high-dispersion CHIRON spectrum at the Hα line of the bright nebular knots in Group I. This Hα line profile can be fitted with two Gaussian components, a narrow one centered near ${V}_{\mathrm{hel}}\sim 250$ km s⁻¹ with FWHM ∼ 40 km s⁻¹ and a broad one centered near ${V}_{\mathrm{hel}}\sim 370$ km s⁻¹ with FWHM ∼ 190 km s⁻¹. The two Gaussian components are also plotted in Figure 8. It can be seen that the broad Gaussian component can describe the bulk of shocked gas but not near the highest velocities, as there is still emission near ${V}_{\mathrm{hel}}=600\mbox{--}700$ km s⁻¹.

Zoom In Zoom Out Reset image size

The [N ii] λ6583 line, due to nitrogen’s larger atomic weight, has a smaller thermal width than the Hα line and thus reveals slowly expanding structures more clearly. Figure 9 shows a stretched echelle image of the [N ii] line along the NS slit. It can be seen that the small velocity variations in the interstellar component follow a curve that is characteristic for an expanding shell. While the receding side of the shell is regular and complete, the approaching side is incomplete by more than 50% of the spatial extent. The expansion velocity of this shell is ∼10 km s⁻¹. This incomplete shell structure is reminiscent of the incomplete filamentary elliptical shell of N103B, and the preferential existence of redshifted material is similar to what is observed in the shocked nebular knots. However, a comparison of its spatial extent and the features in the Hα image shows that this expanding shell structure is bounded by two curved filaments extending outward from the superbubble around NGC 1850. Therefore, this expanding shell is likely a blister on the rim of the superbubble, and its large size precludes its physical association with the SNR N103B.

Zoom In Zoom Out Reset image size

To formulate a feasible model for the structure and formation of the SNR N103B, we need to explain the following observed features: (1) the apparently elliptical Balmer-dominated shell, (2) dense knots projected within the Balmer shell interior, and (3) the eastern extension of diffuse X-ray and radio emission that does not have optical counterparts.

The presence of the Balmer-dominated shell is indicative of collisionless shocks in a mostly neutral ISM (Chevalier et al. 1980). N103B’s Balmer shell thus suggests that the SNR is not in the ionized ISM surrounding NGC 1850's superbubble. However, the Balmer line emission peaks at velocities that are similar to those of the background ionized gas, suggesting that the medium surrounding N103B and the ionized superbubble of NGC 1850 are likely associated with the same interstellar complex.

The Balmer filaments represent the shock fronts, and their curvatures can be used to assess the SN explosion center. As the projection of a sphere is always a sphere, the apparently elliptical shape of the Balmer shell indicates that the shock fronts have run into a nonuniform medium and are no longer spherically symmetric. The elliptical shape can be produced by the projection of an elliptical shell or a tilted round torus (or short cylinder). As the Balmer shell is open to the east, it is unlikely that the Balmer shell is a uniform elliptical shell. Instead, it is more likely that the Balmer “shell” is actually a torus or a short cylinder with the eastern section missing. The projection of a short cylinder appears elliptical, and the center of the cylinder is projected at the center of the ellipse. Therefore, we identify the geometric center of the Balmer shell as the site of SN explosion in N103B.

The dense knots seen in the interior of the SNR shell consist of CSM (Williams et al. 2014) that has been impacted by the SNR shocks. The distribution of the knots shows clearly that the circumstellar material is not spherically symmetric; furthermore, the CSM appears to be distributed in a plane. This geometry is similar to that expected from the model of Hachisu et al. (2008) for Type Ia SNe. In this model, a WD accretes mass from a main-sequence companion and generates a fast wind that evacuates material in the polar directions and strips the surface material of the companion into a torus in the orbital plane. This model naturally explains the distribution of the CSM in N103B.

The CSM is missing toward the east, and the preferentially redshifted velocities of the shocked dense knots indicate a lack of CSM in the direction toward us as well. This highly asymmetric distribution of CSM can be caused by a large proper motion of a mass-losing system through the ISM, such as a fast-moving asymptotic giant branch (AGB) star (Villaver et al. 2012). As described above, the Balmer shell opens to the east with no Hα emission, but X-ray and radio emission extends much farther east. This asymmetric distribution of the mostly neutral ISM can also be naturally explained by the proper motion of the progenitor binary system. In fact, the radio morphology of N103B is reminiscent of the Hα image of the planetary nebula Sh 2–188, whose morphology has been explained by the nebula’s high-velocity motion through the ISM (Wareing et al. 2007).

We propose that the progenitor of N103B’s Type Ia SN has a normal star companion, and the binary system has a proper motion through the ISM. The WD’s powerful wind strips the companion’s envelope to form a circumstellar torus that is compressed to higher densities in the direction of proper motion. The compressed CSM breaks up and forms knots owing to Rayleigh–Taylor instabilities, while the CSM in the trailing side has low densities and is thus undetected.

The surviving companion of N103B’s SN progenitor, if it exists, should be at a finite distance from the explosion center, and the distance depends on its kick velocity at the explosion. In a previous surviving companion search, Pagnotta & Schaefer (2015) adopted the average geometric center of N103B’s images in radio (Dickel & Milne 1995) and X-ray (Lewis et al. 2003) wavelengths, although these images show very uneven surface brightnesses, with the eastern half much fainter than the western half. They found eight stars within $4^{\prime\prime}$ from this explosion center; however, an unambiguous identification of the surviving companion still awaits spectroscopic data demonstrating anomalous chemical abundance or kinematic properties, such as high velocity or fast rotation.

Our analysis of the physical structure of N103B, taking into account both the CSM and ISM, concludes a different explosion center. In our study, the Balmer-dominated filamentary shell in the HST Hα image is used to determine the site of SN explosion. The Balmer-dominated shell is visually fitted by an ellipse with a dimension of ∼28″ × 18″ and a center at 05^h08^m588, −68° 43′347 (J2000) (see left panel in Figure 10). The apparent ellipticity can be reproduced by a short round cylinder inclined by ∼50° against the sky plane.

Zoom In Zoom Out Reset image size

Adopting the center of the Balmer-dominated elliptical shell as the explosion center, we have compiled all stars with $V\lt$ 22.7 within $4^{\prime\prime}$ from the center. Adopting the age of 860 yr derived from the SN light echoes (Rest et al. 2005), these limits correspond to an F8V—G0V star and a kick velocity of (1 pc)/(860 yr) ∼ 1140 km s⁻¹, respectively. These stars are listed in Table 3 in order of increasing distance from the explosion center.

Table 3. Stars Brighter than V = 22.7 and within $4^{\prime\prime}$ from the Explosion Center in the SNR N103B

Star	R.A. (J2000)	Decl. (J2000)	B	V	I	B − V	V − I	Θ
1	05:08:58.95	−68:43:34.7	22.33 ± 0.01	21.81 ± 0.01	20.24 ± 0.01	0.53 ± 0.01	1.56 ± 0.01	08
2	05:08:58.56	−68:43:34.3	20.95 ± 0.00	20.87 ± 0.00	20.59 ± 0.01	0.07 ± 0.01	0.28 ± 0.01	13
3	05:08:59.06	−68:43:33.3	22.64 ± 0.01	22.28 ± 0.01	21.18 ± 0.01	0.36 ± 0.02	1.10 ± 0.01	20
4	05:08:58.46	−68:43:35.8	21.51 ± 0.01	21.40 ± 0.01	21.12 ± 0.01	0.11 ± 0.01	0.28 ± 0.01	22
5	05:08:58.57	−68:43:32.6	22.55 ± 0.01	22.38 ± 0.01	21.75 ± 0.02	0.18 ± 0.02	0.63 ± 0.07	24
6	05:08:58.33	−68:43:35.0	20.01 ± 0.00	19.50 ± 0.00	18.12 ± 0.00	0.51 ± 0.00	1.38 ± 0.00	25
7	05:08:58.56	−68:43:32.2	22.58 ± 0.01	22.34 ± 0.01	21.70 ± 0.02	0.24 ± 0.02	0.64 ± 0.02	29
8	05:08:58.54	−68:43:32.1	22.66 ± 0.01	22.56 ± 0.01	21.97 ± 0.02	0.11 ± 0.02	0.63 ± 0.02	29
9	05:08:58.27	−68:43:33.4	20.89 ± 0.00	20.61 ± 0.00	19.81 ± 0.00	0.27 ± 0.01	0.80 ± 0.00	30
10	05:08:58.69	−68:43:31.6	20.16 ± 0.00	20.11 ± 0.00	19.91 ± 0.00	0.05 ± 0.00	0.20 ± 0.01	32
11	05:08:58.80	−68:43:38.4	20.40 ± 0.00	19.97 ± 0.00	18.76 ± 0.00	0.43 ± 0.00	1.21 ± 0.00	36
12	05:08:59.40	−68:43:33.3	22.82 ± 0.01	22.58 ± 0.01	21.82 ± 0.01	0.24 ± 0.01	0.76 ± 0.02	36
13	05:08:59.24	−68:43:31.8	22.07 ± 0.01	21.87 ± 0.01	21.31 ± 0.01	0.20 ± 0.01	0.55 ± 0.01	38
14	05:08:59.38	−68:43:32.5	22.56 ± 0.01	22.33 ± 0.01	21.63 ± 0.01	0.23 ± 0.01	0.71 ± 0.01	39
15	05:08:59.10	−68:43:31.1	21.29 ± 0.01	21.19 ± 0.01	20.89 ± 0.01	0.10 ± 0.01	0.30 ± 0.01	40

Download table as:  ASCIITypeset image

To identify candidates for the surviving companion of N103B’s SN progenitor, we have plotted in Figure 11 all stars with $V\lt 22.7$ within $15^{\prime\prime}$ from the explosion center in the V versus B − V and I versus V − I CMDs, with the stars within $4^{\prime\prime}$ from the center marked and numbered in the CMDs, as well as in the Hα and color-composite images in Figure 10. We have also overplotted the post-impact evolutionary tracks of surviving companions for cases of He stars and main-sequence stars (Pan et al. 2014).

Zoom In Zoom Out Reset image size

It can be seen that within 4″ from the explosion center, no stars match the expectation of models of Pan et al. (2014). Only star 9 is close to the base of the tracks for MS companions in the V versus B − V CMD, but not in the I versus V − I CMD; furthermore, star 9 is $3^{\prime\prime}$ from the explosion center, and it would need a very high kick velocity, 850 km s⁻¹, which is unlikely.

In the SD scenario for Type Ia SNe, the companions may be MS stars, RGs, or He stars. Pan et al. (2014) modeled post-impact evolution of MS and He companions, but not RG companions. Podsiadlowski (2003), on the other hand, considered a 1 ${M}_{\odot }$ subgiant companion that had been stripped of 0.2–0.8 ${M}_{\odot }$ envelope mass and heated by the impact of SN ejecta and found that the surviving companion would be a ∼5000 K subgiant with 1–100 ${L}_{\odot }$. Interestingly, the star closest to the explosion center, star 1, appears to have temperature and luminosity consistent with this prediction. Assuming negligible extinction, star 1's $B-V\sim 0.53$ suggests a spectral type of late F to early G type, and its $V-I\sim 1.56$ suggests a spectral type of late K type; in the V band, star 1 is 4 times brighter than the Sun, while in the I band star 1 is 9 times as bright as the Sun. The mass of the CSM in N103B, as determined in Section 4, is ∼1 ${M}_{\odot }$, of similar order of magnitude as that of the stripped envelope of a subgiant or a giant. Furthermore, as shown in Figure 11, the location of star 1 in the CMDs is not well populated by LMC or Galactic stars, implying a less likelihood to be a popular background star. Thus, we suggest that star 1 is the most likely candidate for the surviving companion of the SN progenitor of N103B. Physical properties of post-impact companion stars have been modeled, as shown in Table 4 (Pan et al. 2013). Spectroscopic observations of star 1 in N103B should be made to search for high radial velocity and/or fast rotation to confirm or reject its identification as a surviving companion of the SN progenitor.

Table 4. Properties of Post-impact Companion Stars

Parameter	He Star	MS Star	Giant Star
Linear velocity (km s−1)	500–1000	100–200	<100
Rotational speed (km s−1)	>50–200	<100	very low
Color	Very blue
Effective temperature (K)	10,000–100,000	5000–10,000	Uncertain
Stripped mass (${M}_{\odot }$)	<0.1	0.1–0.4	0.5–1.0

Reference. Pan et al. (2013).

Download table as:  ASCIITypeset image

Kepler’s SNR, in the Milky Way at a distance of ∼5 kpc, resulted from the SN discovered by Johannes Kepler in 1604. Optical images of Kepler’s SNR display a complex network of knots and filaments within a radius of ∼100″. Although the type of Kepler’s SN was debated at first, X-ray, optical, and IR observations of the SNR have pointed toward a Type Ia origin (Blair et al. 2007; Reynolds et al. 2007) and demonstrated the existence of a dense CSM from the progenitor system (Van den Bergh et al. 1973; Van den Bergh & Kamper 1977; Dennefeld 1982; Blair et al. 1991).

N103B and Kepler’s SNR share many similar properties. In X-rays, both SNRs show prominent Si, S, Ar, Ca, and Fe Kα emission and a notable absence of O in the spectra of their SN ejecta, as expected for Type Ia SNe (Hughes et al. 1995; Reynolds et al. 2007). In nebular lines at optical wavelengths, both SNRs show similar knots-and-filaments morphological characteristics: both have Balmer-dominated filaments delineating the outer boundary of the SNR but only along 1/3 to 1/2 of the periphery, and both exhibit dense knots distributed within the SNR shell, as well as close to the Balmer filaments (Sankrit et al. 2008). The dense knots in Kepler’s SNR have electron densities of 2000–10,000 cm⁻³ (Blair et al. 1991), and those in N103B have ∼5300 cm⁻³. The dense knots in Kepler’s SNR have been suggested to be CSM ejected by the SN progenitor system because of the enhanced N abundance implied by the high [N ii]/Hα ratio, ∼2 (Dennefeld 1982; Dodorico et al. 1986). The high densities of N103B’s knots have been used to argue for a CSM origin, similar to Kepler’s knots; however, the [N ii]/Hα ratios of the dense knots in N103B are only ∼0.2, not as high as those of Kepler’s SNR. This factor of 10 difference in [N ii]/Hα ratio cannot be explained completely by the low N abundance in the LMC, 0.16 times that of the Sun or 0.37 times that of the solar neighborhood (Russell & Dopita 1992). The CSM originates from the envelope of the nondegenerate companion in the SN progenitor system (Hachisu et al. 2008). As the CNO process can build up N abundance and the processed material can be mixed into the envelope through convection, it is possible that the nondegenerate companion of Kepler’s SN progenitor has a higher mass than that for N103B, and Kepler’s CSM contains CNO-processed material.

A search for a surviving companion of Kepler’s SN progenitor has been conducted by Kerzendorf et al. (2014) using spectroscopic observations of the 24 brightest stars near the center of Kepler’s SNR. From the radial velocities of these stars, they ruled out an RG companion with high certainty and found no candidates for a donor star with luminosity >10 ${L}_{\odot }$, although follow-up astrometric observations for proper motion and high-dispersion spectra for stellar rotation are needed for further search. In this paper, we identify a possible candidate for a surviving companion of N103B’s SN progenitor near the explosion center of the SNR. This candidate was likely a solar-mass subgiant that had no significant CNO cycle burning of H in its interior (Podsiadlowski 2003). It also needs follow-up spectroscopic observations of radial velocity, rotational velocity, and abundances to verify its association with N103B’s SN progenitor.