MAPPING HIGH-VELOCITY Hα AND Lyα EMISSION FROM SUPERNOVA 1987A

Fransson et al. (2013) noted that the large observed velocity of the Hα at late times (V_obs $\gtrsim$ 11 × 10³ km s⁻¹ at t > 20 yr) suggests that the RS front has expanded beyond the ring plane. They conclude that the RS extent in the polar direction is ∼15% larger than the ER radius and ∼40% larger than the RS radius in the ring plane. We find deprojected velocities of 10–12 × 10³ km s⁻¹ in the B-b and B-r regions, corresponding to RS radii 40–70% larger than the 6.1 × 10¹⁷ cm ring radius. The C regions extend to 2–2.5 times the ER radius.

To interpret the high-velocity Hα images, we recall that surfaces of constant Doppler shift are planar sections of the freely expanding supernova debris. Therefore, the pass bands of the "Hα Blue" and "Hα Red" filters map to slabs in physical space, as illustrated in Figure 5. Portions of the RS surface visible through the "Hα Blue" filter (colored blue) reside within the slab delineated by the blue vertical lines, while those visible through the "Hα Red" filter (colored red) reside within the slab delineated by the red vertical lines. The density of the outer debris crossing the RS front, and therefore the mass flux, is ${{\rho }_{{\rm sn}}}$ $\propto$ ${{\rho }_{o}}$ ${{t}^{-3}}$ ${{V}^{-9}}$ (Luo et al. 1994). The much higher density of the debris crossing the RS in the equatorial direction relative to the polar direction (${{\rho }_{{\rm ER}}}$/${{\rho }_{{\rm pol}}}$ ∼10⁴) makes the A region much brighter (detectable spectroscopically within ∼10 yr of the explosion). The lower density material flowing out of the ring plane (B and C regions) is fainter, only now visible with the dedicated high-velocity Hα imaging and sufficient time for the polar shock structure to be angularly separated from the ring plane.

Zoom In Zoom Out Reset image size

While much of the RS surface is not imaged because of the discrete velocity surfaces transmitted through the filters (shaded gray in Figure 5), the emerging picture for the extended Hα emission in the off-plane regions suggests a bipolar morphology. This is expected due to the high pressure of the equatorial plane, owing to the presence of the ring (Blondin et al. 1996). Assuming the outer ejecta pressure is proportional to ${{\rho }_{{\rm sn}}}$ ${{V}^{2}}$, the C polar region pressure is ∼1/1800 of the equatorial region pressure. This general picture resembles the bipolar surfaces suggested for the pre-supernova interacting wind picture by Blondin & Lundqvist (1993), however, recent studies of a "pre-explosion twin" to the SN 1987 A system (SBW1; Smith et al. 2013) suggest that interacting winds may not be required to explain the nebular morphology.

A second possible explanation for the C region emission is a conical or helical structure (e.g., Smith 2007; Smith et al. 2013) connecting the ER with the extended outer rings (ORs). In this scenario, (1) high-velocity ejecta crossing this structure would be limb-brightened, explaining the bright edges in the C-b and C-r emission, and (2) the southern C-b streaks would be offset to the east compared to the northern C-r streaks by an amount comparable to the E–W offset of the ORs (e.g, Sugerman et al. 2005), which is indeed what is observed in Figure 2. In this picture, the C emission would essentially be a new SN 1987 A intermediate ring system. Deep Hα imaging spectroscopy to map the region could distinguish between these scenarios and may shed light on the progenitor structure, including the possibility of a binary merger prior to the supernova explosion (Morris & Podsiadlowski 2007).

Neutral hydrogen mass flux and the Lyα/Hα ratio—if all of the high-velocity hydrogen emission originates from collisional excitation as neutral hydrogen atoms cross the RS front, the total flux is proportional to the fluence of hydrogen atoms across the shock boundary. Integrating the individual band images, we calculate the total number of atoms in a given velocity range. We take the total (Galactic + LMC) reddening toward SN 1987 A as a standard ISM curve with R_V = 3.1 and $E(B\;-\;V)$ = 0.19 (see France et al. 2011 for a discussion). The total rate of hydrogen atoms crossing the RS front is approximately

$$\begin{eqnarray}&&\dot{{{N}_{H}}}=4\pi {{d}^{2}}\times F({\Delta }\lambda )\frac{\lambda }{hc}{\Delta }\lambda \times P{{(H)}^{-1}}\times {{C}_{{\rm ISM}}},\end{eqnarray} \tag{ 1 }$$

where d ∼50 kpc (∼10% uncertainty; Panagia et al. 1991), $F({\Delta }\lambda )$ is the image-integrated flux density in a given band (∼5% uncertainty), ${\Delta }\lambda$ is the effective bandpass of the filter, $P(H)$ is the number of line photons emitted per atom, and C_ISM is the interstellar attenuation correction factor. $P(H)$ ≈ 0.2 for Hα and 1 for Lyα (Michael et al. 2003; Heng & McCray 2007).

For the "Hα Blue" image, ${\Delta }\lambda$ ≈ 100 Å and C_ISM = 1.57 ± 0.08, giving a total hydrogen fluence of 4.1 (±0.7) × 10⁴⁶ atoms s⁻¹. For the "Hα Red" image, ${\Delta }\lambda$ ≈ 120 Å and C_ISM = 1.53 ± 0.08, giving a total hydrogen fluence of 3.3 (±0.5) × 10⁴⁵ atoms s⁻¹. Taken together, we estimate the total mass flux of high-velocity hydrogen atoms crossing the RS is $\dot{{{M}_{H}}}$ = 7.4 (±1.2) × 10²² g s⁻¹ (1.2 × 10⁻³ M $_{\odot }$ yr⁻¹).

The analogous calculation using the F122M Lyα image has ${\Delta }\lambda$ ≈ 100 Å and C_ISM = 6.7 × 1.63, where 6.7 (±1.7) is the correction for dust attenuation and 1.63 (±0.08) is the empirically determined correction for the interstellar Lyα line core attenuation. This calculation shows a Lyα-to-Hα photon ratio, $R({\rm L}\alpha /{\rm H}\alpha )$, of ≈17 ± 6, a factor of 3.5 larger than attributable to excitation in the RS alone. Therefore, an additional Lyα emission mechanism is present. Note that the reddening correction does not account for differential extinction within the supernova debris, so the intrinsic $R({\rm L}\alpha /{\rm H}\alpha )$ may be greater. Lyα excess was observed by previous authors using spectroscopic observations of small spatial regions of the SN 1987 A inner ring region (Heng et al. 2006; France et al. 2010).

Combining the "Hα Blue" and "Hα Red" images with the Lyα map, we find the baseline $R({\rm L}\alpha /{\rm H}\alpha )$ interior to the ER is 8–10, at the center and southeastern sides of the inner remnant. $R({\rm L}\alpha /{\rm H}\alpha )$ reaches a maximum on the western and northwestern side of the inner remnant where the Chandra emission is brightest (Figure 3). The maximum $R({\rm L}\alpha /{\rm H}\alpha )$ is ∼35 near PA = 270°, just interior to the brightest X-ray emission from the ER. $R({\rm L}\alpha /{\rm H}\alpha )$ values ∼20 are found toward PA ∼ 60°, the location of the northeast X-ray maximum.

The Role of X-rays—the observed offset between the X-ray and Lyα emitting regions is consistent with the expected separation between the ER and the outer extent of the RS. Following the observation that the dominant energy input in the unshocked supernova debris has evolved from internal heating by radioactive decay to external heating from X-rays produced at the ER (Larsson et al. 2011), Fransson et al. (2013) presented calculations of the X-ray energy deposition into the ejecta. The heating of the outer debris is dominated by X-rays with E < 0.5 keV while higher energy X-rays penetrate deeper to heat the interior. X-rays propagating into the outer debris produce fast photoelectrons, which deposit most of their kinetic energy as heat, provided that the debris has fractional ionization ${{n}_{e}}/{{n}_{H}}$ >0.03 (Xu & McCray 1991). Radiative cooling by thermal excitation of Lyα limits the debris temperature to T $\lesssim$ 10⁴ K, at which temperature thermal excitation of Hα is negligible.¹⁴ Therefore, we expect the Lyα/Hα ratio to be at a maximum nearest the source of the X-ray heating. The spatial stratification of the Lyα and X-ray emission and the enhanced Lyα/Hα ratio support the scenario where X-ray heating of the outer debris is responsible for the majority of the observed Lyα emission from SN 1987 A at present.

The total Lyα luminosity in the F122M band is ${{L}_{Ly\alpha }}$ = 2.5 (±0.7) × 10³⁶ erg s⁻¹, ∼70% of which is in excess of the expected RS emission. The total flux of the X-ray ring is 1.04 × 10⁻¹¹ erg cm⁻² s⁻¹, or L_X(0.3–8.0 keV) = 3.1 (±0.3) × 10³⁶ erg s⁻¹. However, most of the heating radiation is in the soft X-ray/EUV band (0.01–0.5 keV) which is not observed at earth because of interstellar neutral hydrogen and dust attenuation toward SN 1987 A. Scaling the Zhekov et al. (2006) model X-ray spectrum to the observed days 9885 0.3–8.0 keV flux, the total 0.01–8.0 keV luminosity is L_X(0.01–8.0 keV) = 5.6 × 10³⁶ erg s⁻¹. Assuming that the only source of EUV/soft X-rays is that observed by Chandra and that the neutral hydrogen in the outer ejecta intercepts half of the soft X-rays emitted by the ring, the Lyα heating efficiency must be ∼60% to reproduce the F122M image of SN 1987 A. However, slower shocks that contribute more to the optical/UV emission of the ER than to the X-rays will also be a source of EUV irradiance of the outer debris (Fransson et al. 2013), meaning that the actual soft X-ray heating efficiency is likely substantially lower.