The bubble-like interior of the core-collapse supernova remnant Cassiopeia A

See allHide authors and affiliations

Science  30 Jan 2015:
Vol. 347, Issue 6221, pp. 526-530
DOI: 10.1126/science.1261949

Burbling explosions blow metallic bubbles

Stars more than about eight times the mass of the Sun don't go out quietly. Asymmetric explosions are kicked off by the collapse of their iron cores when no more fusion energy can sustain them. Exactly how this stellar catastrophe proceeds is difficult to probe. Milisavljevic et al. have now peered into the supernova remnant Cas A in the near-infrared and present a three-dimensional map of its interior unshocked ejecta. The bubble-like structure points to turbulent mixing, which may help us understand other supernova remnants whose structure cannot be seen in such detail.

Science, this issue p. 526


The death of massive stars is believed to involve aspheric explosions initiated by the collapse of an iron core. The specifics of these catastrophic explosions remain uncertain, due partly to limited observational constraints on asymmetries deep inside the star. Here we present near-infrared observations of the young supernova remnant Cassiopeia A, descendant of a type IIb core-collapse explosion, and a three-dimensional map of its interior unshocked ejecta. The remnant’s interior has a bubble-like morphology that smoothly connects to and helps explain the multiringed structures seen in the remnant's bright reverse-shocked main shell of expanding debris. This internal structure may originate from turbulent mixing processes that encouraged outwardly expanding plumes of radioactive 56Ni-rich ejecta. If this is true, substantial amounts of its decay product, 56Fe, may still reside in these interior cavities.

Computer simulations have long shown that the collapse of a high-mass star’s dense iron-rich core into a neutron star generates an outward-moving shock wave that is unable to disrupt the star into a supernova (SN) under spherical conditions (1). This is because the outgoing shock wave is too weak to overcome the infalling outer layers of the star and stalls, thus requiring some postbounce revival (2, 3). It is now believed that explosion asymmetries introduced by dynamical instabilities and the influences of rotation and magnetic fields must contribute to the core-collapse process (4, 5), but their relative contributions are uncertain. Some aspects of these explosion processes have been successfully probed by observations of extragalactic supernovae (SNe) (6, 7). However, even with the Hubble Space Telescope (HST), distant SNe appear as unresolved point sources, which limits our ability to unravel the three-dimensional (3D) structure of the expanding ejecta that can reveal key properties of the explosion mechanism.

An alternative approach to understanding core-collapse SN explosions is through studies of their remnants in our own Milky Way galaxy that are still young enough to be in free expansion and near enough to permit detailed studies of the expanding debris field. Such investigations can provide information on the explosion-driven mixing of the progenitor star’s chemically distinct layers, the star’s mass loss history before explosion, and the fate of its remnant core. With an age of around 340 years (8), a distance of only 11,000 light years (9), and its status as a known descendant of a type IIb explosion (10), Cassiopeia A (Cas A) is one of the best specimens for postmortem examination. Cas A is visible today through the heating effects of a reflected or “reverse” shock generated when the original SN’s high-velocity blast wave ran into the surrounding interstellar medium. Cas A’s metal-rich debris is arranged in large ringlike structures that together form a roughly spherical shell (commonly referred to as the main shell) approximately 6 light years in radius, with an overall velocity range of −4000 to +6000 km s−1 (11, 12). These and many other properties of Cas A, such as its wide-angle (≈40°) opposing streams of Si- and S-rich ejecta seen in the northeast and southwest regions traveling at unusually high velocities of up to 15,000 km s−1 (13, 14), an uneven expansion of the photosphere at the time of outburst (15), and an overall high chemical abundance ratio of 44Ti/56Ni with a nonuniform distribution (16, 17), all point to an asymmetric explosion.

In order to investigate possible asymmetries in Cas A’s interior debris, we obtained near-infrared spectra of the remnant in 2011 and 2013 using the Mayall 4-m telescope at Kitt Peak National Observatory, in combination with instrumental setups particularly sensitive to the wavelength region from 900 to 1000 nm covering the [S III] 906.9- and 953.1-nm line emissions from Cas A’s sulfur-rich ejecta. Closely spaced long-slit spectra were taken across the remnant’s central regions (18). The resulting data were transformed into three-dimensional coordinates and incorporated into our existing reconstruction of the optical main shell and high-velocity outer knots (12).

Faint and patchy [S III] emission can be seen interior to Cas A’s bright main shell, with radial velocities spanning –3000 to +4000 km s−1 in a continuous manner (Fig. 1). The emission is diffuse, in sharp contrast with the bright main-shell ejecta that are compressed into small knots and filaments by the reverse shock. Most of our detections are located in the southern half of the remnant, but some are found in the northern half as well (Fig. 2). Presumably, the [S III] emission arises from unshocked interior material being photoionized by ultraviolet and x-ray flux from the main-shell ejecta that has been heated to temperatures up to several million degrees kelvin by the reverse shock.

Fig. 1 Representative near-infrared observation of the [S III] 906.9- and 953.1-nm line emission from Cas A.

(Left) A finding chart of a single long-slit position, which was rotated to optimize coverage of particular interior emission regions. The background image is a mosaic created from 2004 HST observations sensitive to O and S emissions (31). ACS/WFC, Advanced Camera for Surveys/Wide Field Channel. (Right) The corresponding 2D spectrum. Square brackets highlight regions where interior unshocked ejecta have been detected. See fig. S1 for a map of all slit positions used in the survey. KPNO, Kitt Peak National Observatory; MARS, Multi-Aperture Red Spectrograph.

Fig. 2 Map of [S III] 906.9- and 953.1-nm line emission detected by our survey.

The blue-to-red color gradient represents the range of measured Doppler velocities. Spheres mark individual measurements. (A) All blueshifted emission (<0 km s−1) and (B) all redshifted emission (>0 km s−1). The background is a composite HST image sensitive to the remnant’s O and S emissions retrieved from The center of expansion is shown as a white cross. (C) A perspective rotated 90° toward the west along the north-south axis. The north and south interior cavities are highlighted, as well as the wall of ejecta where the two cavities intersect.

The majority of the unshocked ejecta exhibit coherent structure (Fig. 3) in the form of large cavities or “bubbles” that appear to be physically connected to the main-shell rings. At least two cavities are well defined. A cavity seen in the southeast with the largest range of blueshifted velocities extends outward from just below the remnant’s approximate center and smoothly connects with a pair of optically bright curved filaments sometimes referred to as the “Parentheses.” Opposite this, in the northwest and immediately below the remnant’s largest main-shell ring of redshifted material, is the largest internal cavity. These two cavities intersect along a concentration of central emission that runs from the front to the back of the remnant.

Fig. 3 Doppler reconstruction of Cas A made from the [S III] 906.9- and 953.1-nm emission map presented here and previous optical observations of main-shell ejecta (12).

The blue-to-red color gradient corresponds to Doppler velocities that range from –4000 to 6000 km s−1. [S III] measurements are individual spheres, and previous optical data are smoothed with a surface reconstruction. (A) A side perspective of a portion of the remnant spanning all material located between 15″ east of the center of expansion to 50″ west of the center of expansion to emphasize the two conspicuous interior cavities and their connections to main-shell ejecta. The translucent sphere centered on the center of expansion is a visual aid to differentiate between front and back material. (B) Two angled perspectives highlighting the south cavity. The first perspective angled 20° away from the observer’s line of sight shows all data, and the second perspective angled 70° away from the observer’s line of sight shows the same portion of the remnant as displayed in (A). The background image representing the plane of the sky as seen from Earth is the same shown in Fig. 2. An animation of the entire reconstruction is provided in movie S1.

The larger northern cavity dominates the remnant’s interior volume and has a radius of approximately 3 light years, whereas the southeast cavity is approximately half as large. Both cavities exhibit a few S-rich clumps or filaments, indicating that neither cavity is completely empty of ejecta. Although it is difficult to accurately assess the total number of cavities, the diameters of the main-shell rings are comparable to the diameters of the cavities, which suggests that they are approximately equal in number (about six). These properties, along with the fact that the main-shell rings extend radially outward along gently sloped paths that follow the circumference of the cavities, support the notion that the reverse-shocked rings and unshocked cavities of ejecta share a common formation origin.

Portions of Cas A’s interior unshocked ejecta that were surveyed by our near-infrared observations are also visible in previous observations taken at longer wavelengths. Infrared images of Cas A taken with the Spitzer spacecraft show [S III] 33.48-μm and [S IV] 10.51-μm emission inside the boundary of the main shell at locations coincident with regions where we detect the strongest [S III] 906.9- and 953.1-nm emissions (19). Follow-up Spitzer infrared spectra in the central region showed line emission from interior O, Si, and S ejecta in sheetlike structures and filaments with inferred radial velocities approaching ±5000 km s−1 (20). Those spectroscopic Spitzer observations covered a smaller 50″ × 40″ area of Cas A, whereas our survey encompasses the entire remnant and reveals a far larger extent of the remnant’s internal debris.

We interpret Cas A’s main-shell rings of ejecta to be the cross sections of reverse-shock–heated cavities in the remnant’s internal ejecta now made visible by our survey. A cavity-filled interior is in line with prior predictions for the arrangement of expanding debris created by a postexplosion input of energy from plumes of radioactive 56Ni-rich ejecta (21, 22). Such plumes can push the nuclear burning zones located around the Fe core outward, creating dense shells separating zones rich in O, S, and Si from the Ni-rich material. Compression of surrounding nonradioactive material by hot expanding plumes of radioactive 56Ni-rich ejecta generates a “Swiss cheese”–like structure that is frozen into the homologous expansion during the first few weeks after the SN explosion, when the radioactive power of 56Ni is strongest.

In this scenario, the decay chain of 56Ni → 56Co → 56Fe should eventually make these bubble-like structures enriched in Fe. Doppler reconstruction of Chandra x-ray observations sensitive to Fe K emissions shows that the three most significant regions of Fe-rich ejecta are located within three of Cas A’s main-shell rings (11, 12). Thus, the coincidence of Fe-rich material with rings of O- and S-rich debris is consistent with the notion of 56Ni bubbles.

However, Fe-rich ejecta associated with the Ni bubble effect should be characterized by diffuse morphologies and low ionization ages, and yet the x-ray bright Fe emissions we currently see are at an advanced ionization age relative to the other elements (23, 24). Furthermore, not all main-shell rings have associated x-ray–emitting Fe-rich ejecta, and there is no clear relationship between the locations of the internal bubbles we have detected and the spatial distribution of 44Ti recently mapped by NASA’s NuSTAR (Nuclear Spectroscopic Telescope Array) (17).

One solution is that Fe-rich ejecta associated with some of the remnant’s internal cavities remain undetected. Low-density Fe could be present in ionization states between those detectable by optical or infrared line emission and those in x-rays. The total mass of unshocked Fe that is potentially contained in the bubbles is constrained by the total nucleosynthetic yield of Fe in the original SN explosion, which is estimated to be less than ∼0.2 solar mass (M) (25), and the amount of shocked Fe that is observed today, which is estimated to be 0.09 to 0.13 M (24). Together these estimates imply that no more than an additional ∼0.1 M of Fe could potentially be located within Cas A’s reverse shock.

Whatever their true cause, Cas A’s bubble-like interior and outer ringlike structures observed in the main shell suggest that large-scale mixing greatly influences the overall arrangement of ejecta in core-collapse SNe. Presently, the extent of such mixing and how it takes place are not well known. A variety of potential dynamical processes may contribute to the redistribution of chemical layers, including uneven neutrino heating, axisymmetric magnetorotational effects, and Rayleigh-Taylor and Kelvin-Helmholtz instabilities [e.g. (26, 27)].

Compelling evidence for large-scale mixing involving considerable nonradial flow was first observed in the nearest and brightest SN seen in modern times, SN 1987A. In that case, high-energy gamma rays and x-rays with broad emission line widths from the decay of 56Ni were detected only months after the explosion, implying that Ni-rich material was near the star’s surface well before 1D progenitor models had predicted, assuming spherical symmetry (28).

Since SN 1987A, state-of-the-art 3D computer simulations of core-collapse explosions have confirmed that large-scale mixing can lead to Ni-dominated plumes overtaking the star’s outer oxygen- and carbon-rich layers with velocities up to 4000 km s−1 (29). However, the majority of these simulations show that the mass density should essentially be unaffected. Although mixing can affect the species distribution, the bulk of the Ni mass should remain inside the remnant with velocities below 2000 km s−1. This is, in fact, opposite to what we currently see in Cas A, where the x-ray bright Fe has velocities around the 4000 km s−1 limit (11). Thus, either the simulations are not adequately following the dynamics of mixing or, as we suspect, more Fe remains to be detected in Cas A’s interior.

An additional consideration in interpreting a SN debris field is the chemical makeup of the star at the time of outburst. The evolution of massive stars toward the ends of their life cycles is likely to be nonspherical and may produce extensive intershell mixing. If strong enough, these dynamical interactions lead to Rayleigh-Taylor instabilities in the progenitor structure that can contribute to the formation of Ni-rich bubbles and influence the overall progression of the explosion (30). Thus, asymmetries introduced by a turbulent progenitor star interior, in addition to those initiated by the explosion mechanism, could contribute to the bubble-like morphology observed in Cas A.

Because Cas A’s opposing streams of Si- and S-rich debris have kinematic and chemical properties indicative of an origin deep within the progenitor star (12), we searched for evidence of any structure joining the high-velocity material with the interior ejecta mapped in our survey. However, we were not able to find any clear relationship between them. An indirect association is hinted at by the inferred projected motion of Cas A’s central x-ray point source (XPS) that is thought to be the remnant neutron star. Its motion toward the southwest of the center of expansion is (i) roughly opposite to and moving away from the direction of the largest internal cavity in our reconstruction that is coincident with a sizable concentration of reverse-shocked Fe and (ii) nearly perpendicular to the axis of the high-velocity jets (31). The XPS, conserving momentum, could have been kicked in a direction opposite the largest plume of Fe-rich material (15) and released an energetic proto–neutron star wind that shaped the jets shortly after the core-collapse explosion (32).

The apparent mismatch of Ti-rich and Fe-rich ejecta regions uncovered by NuSTAR is a reminder that unresolved key issues surrounding Cas A still linger despite decades of scrutiny. Our 3D map of its interior is an important step forward, as it represents a rare look at the geometry of a SN remnant’s inner volume of debris unmodified by reverse-shock instabilities. Because Cas A shows many striking similarities with SNe young and old (33, 34), its dynamical properties described in this work are probably not unique and can be used to help interpret other SN explosions and remnants that cannot be resolved.

Our data make it clear that Cas A’s dominant ejecta structure is in the form of large internal cavities whose cross sections are the prominent rings of the reverse-shock–heated main shell. What is not clear, however, is why only half a dozen bubbles—not dozens—are present. SN explosion models can explore this issue, as well as better understand how the remnant’s interior bubbles fit into a single coherent picture with its opposing high-velocity jets. A crucial test of the origin of these cavities would be a confirmation of the “missing” internal Fe we predict to be located in the remnant’s interior, but conclusive observations may not be possible until the next generation of infrared and x-ray space telescopes comes online.

Supplementary Materials

Materials and Methods

Fig. S1

References (3537)

Movie S1

References and Notes

  1. Acknowledgments: This material is based on work supported by the National Science Foundation under grant no. AST-0908237, as well as observations made with the NASA/European Space Agency HST associated with Guest Observer program 10286 (Principal Investigator, R. Fesen) and obtained from the data archive at the Space Telescope Science Institute (STScI). STScI is operated by the Association of Universities for Research in Astronomy under NASA contract NAS 5-26555. Visual modeling of our observations was aided with the use of MeshLab (, a tool developed with the support of the 3D-CoForm project. We thank anonymous reviewers for providing suggestions that improved the content and presentation of the manuscript and D. Patnaude for helpful discussions.

Stay Connected to Science

Navigate This Article