A Detailed Kinematic Map of Cassiopeia A's Optical Main Shell and Outer High-Velocity Ejecta
aa r X i v : . [ a s t r o - ph . H E ] J un Submitted to ApJ October 31, 2012; accepted June 7, 2013
Preprint typeset using L A TEX style emulateapj v. 5/2/11
A DETAILED KINEMATIC MAP OF CASSIOPEIA A’S OPTICAL MAIN SHELLAND OUTER HIGH-VELOCITY EJECTA
Dan Milisavljevic and Robert A. Fesen Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA, 02138email: [email protected] Submitted to ApJ October 31, 2012; accepted June 7, 2013
ABSTRACTWe present three-dimensional kinematic reconstructions of optically emitting material in the youngGalactic supernova remnant Cassiopeia A (Cas A). These Doppler maps have the highest spectraland spatial resolutions of any previous survey of Cas A and represent the most complete catalog ofits optically emitting material to date. We confirm that the bulk of Cas A’s optically bright ejectapopulate a torus-like geometry tilted approximately 30 ◦ with respect to the plane of the sky with a − − radial velocity asymmetry. Near-tangent viewing angle effects and an inho-mogeneous surrounding CSM/ISM environment suggest that this geometry and velocity asymmetrymay not be faithfully representative of the remnant’s true 3D structure or the kinematic propertiesof the original explosion. The majority of the optical ejecta are arranged in several well-defined andnearly circular ring-like structures with diameters between approximately 30 ′′ (0.5 pc) and 2 ′ (2 pc).These ejecta rings appear to be a common phenomenon of young core-collapse remnants and maybe associated with post-explosion input of energy from plumes of radioactive Ni-rich ejecta thatrise, expand, and compress non-radioactive material. Our optical survey also encompassed Cas A’sfaint outlying ejecta knots and exceptionally high-velocity NE and SW streams of S-rich debris oftenreferred to as ‘jets’. These outer knots, which exhibit a chemical make-up suggestive of an origin deepwithin the progenitor star, appear to be arranged in opposing and wide-angle outflows with openinghalf-angles of ≈ ◦ . Subject headings:
ISM: supernova remnants — ISM: individual objects (Cassiopeia A) — supernovae:general INTRODUCTION
A variety of observations and hydrodynamic modelingmake a compelling case that high-mass, core-collapsesupernovae (SNe) are intrinsically aspherical eventswith highly clumped ejecta caused by dynamicalinstabilities (Wang & Wheeler 2008; Maeda et al.2008; Nordhaus et al. 2010; Janka 2012; Tanaka et al.2012). The origin of the expansion asphericities iscurrently uncertain. Possible causes include asym-metrical neutrino heating and accretion-shock in-stabilities (Kifonidis et al. 2000; Blondin et al. 2003;Kifonidis et al. 2003; Burrows et al. 2006; Scheck et al.2006; Marek & Janka 2009; Hanke et al. 2012), and theinfluences of rotation and magnetic fields (Wheeler et al.2002; Akiyama et al. 2003; Fryer & Warren 2004;Shibata et al. 2006; Masada et al. 2012).Observations of the kinematic and chemical prop-erties of SN ejecta can help investigate which ofthe aforementioned explosion mechanisms may dom-inate. For example, late-time optical spectra ob-tained t & ±
19, Cas A is the youngest Galac-tic core-collapse SNR known (Thorstensen et al. 2001;Fesen et al. 2006b). At an estimated distance of 3.4 kpc(Reed et al. 1995), it is also among the closest.Cas A is the only historical core-collapse SNR with asecure SN subtype classification. The detection of lightechoes of the supernova outburst (Rest et al. 2008, 2011;Besel & Krause 2012) has enabled follow-up optical spec-tral observations which indicate that the original super-nova associated with Cas A exhibited an optical spec-trum at maximum light similar to those seen for the TypeIIb events SN 1993J and SN 2003bg (Krause et al. 2008;Rest et al. 2011).Cas A is inferred to have undergone extensive mass Milisavljevic & Fesenloss from its original 20 −
25 M ⊙ progenitor to only3 − ⊙ upon explosion, leaving a relatively dense andslow moving remnant stellar wind (Chevalier & Oishi2003). Such conditions do not generally arise with aradiatively driven wind from a 20 −
25 M ⊙ progenitor,but require the existence of a binary companion to aidthe mass loss (Woosley et al. 1993; Young et al. 2006;Claeys et al. 2011).Several Doppler reconstructions of Cas A’s main shellejecta have been conducted in the past using optical, in-frared, and X-ray data. Results of these studies haverevealed significant ejecta asymmetries potentially re-lated to the explosion dynamics. For example, op-tical studies have shown that much of the remnant’sejecta are arranged in large rings on a spherical shellexhibiting an overall velocity asymmetry of − − (Minkowski 1959; Lawrence et al. 1995;Reed et al. 1995). Spectra of surrounding light echosalso indicate strong asymmetry in the Cas A supernova’sphotosphere (Rest et al. 2011).The recent study of the Cas A remnant byDeLaney et al. (2010) presented an extensive multi-wavelength Doppler reconstruction using new velocitymeasurements from Spitzer infrared and
Chandra
X-ray observations combined with previous optical data onthe remnant’s highest velocity, outer ejecta from Fesen(2001). Besides confirming the presence of several largeejecta rings in the infrared, they interpreted the structureof Cas A’s bright main shell of ejecta knots to consist of aspherical component, a tilted thick disk, multiple ejectajets/pistons, and optical fast-moving knots all populat-ing the thick disk plane. They concluded that the bulkof the symmetries and asymmetries seen in Cas A areproperties intrinsic to the supernova explosion.One often cited piece of observational evidence thatthe Cas A progenitor underwent a highly aspherical ex-plosion has been the presence of a ‘flare’ or ‘jet’ of un-usually high-velocity SN debris extending some 300 ′′ outalong the northeast (NE) limb from the remnant cen-ter (about twice that of main shell ejecta) at a po-sition angle of ∼ ◦ and visible even in the earliestphotographic images of the remnant (Minkowski 1968;van den Bergh & Dodd 1970). Ejecta knots in this NEregion exhibit proper motions indicating expansion veloc-ities extending up to 14 ,
000 (D/3.4 kpc) km s − , some8000 km s − faster than the fastest main shell ejectaknots (Fesen 2001; Fesen et al. 2006b).A fainter and considerably sparser southwest (SW)so-called ‘counterjet’ of equally high-velocity ejecta wasdiscovered optically (Fesen 2001) and subsequently con-firmed in X-rays and the infrared (Hwang et al. 2004;Hines et al. 2004). Chandra
X-ray images show the SWjet to be much fainter and not as extended radially asthe NE one, and in the optical only about 200 knotshave been identified compared to over 1000 in the NEjet. One reason for this substantial difference, at leastin the optical, may lie in greater line of sight extinction(A V ≈ − ∼ OBSERVATIONS
A series of observing runs starting September 2007 andcontinuing to November 2010 were conducted to obtainlow-dispersion long-slit optical spectra across the entireremnant. Most observations were carried out at MDMObservatory, on Kitt Peak, AZ using the 2.4m Hiltnertelescope. These observations were later supplementedby both multi-slit and long-slit spectra obtained withthe Kitt Peak Mayall 4m telescope. A summary of allobservations is given in Table 1. Additional relevant de-tails of these observations are discussed below. A findingchart of all slit positions is presented in Figure 1.For Cas A’s main shell ejecta, the MDM Modular Spec-trograph (Modspec) with a SITe 2048 × ′′ × ′ long-slit oriented north–south was spaced every3 ′′ . Exposures were generally 2 ×
500 s. The effectivewavelength coverage was approximately 4500 − < . . ′′ × . ′ slit andusing a SITe 1024 × − − ′′ × . ′ slit and either the (i)blue-sensitive 150 line mm − − Figure 1.
Finding charts of all long-slit positions. Background image is a mosaic created from 2004
HST /ACS observations sensitiveto oxygen and sulfur emissions (Proposal 10286; PI: R. Fesen; see Fesen et al. 2006a). Bottom right is montage of all locations. Refer toTable 1 for telescope and instrument details.
Table 1
Summary of ObservationsDate Telescope Instrument Region (No. of Positions)2007 Sep 01-08 MDM 2.4 m Modspec Main Shell (58)2008 Sep 18-25 MDM 2.4 m Modspec Main Shell (45); NE Jet (18)2008 Oct 27-30 MDM 2.4 m MKIII Main Shell (19); SW Jet (10)2009 Oct 19-22 MDM 2.4 m CCDS Main Shell (9); NE Jet (16)2009 Dec 11-14 MDM 2.4 m CCDS NE Jet (6)2010 Oct 1-5 KPNO 4 m MARS NE/SW Jets + Outer Knots (81)2010 Oct 8-13 MDM 2.4 m MKIII NE/SW Jets + Outer Knots (38)2010 Oct 28-Nov 1 MDM 2.4 m MKIII NE/SW Jets + Outer Knots (26) lamps and flats were taken to track telescope flexure andaid in the subtraction of fringing. These long-slit spectrawere taken using longer exposures of 1000 − − , ≈ DATA REDUCTION
Data were reduced homogeneously through a seriesof scripts written to automate
IRAF/PyRAF procedures.The 2D images of each position were first trimmed, bias- IRAF is distributed by the National Optical Astronomy Ob-servatories, which are operated by the Association of Universitiesfor Research in Astronomy, Inc., under cooperative agreement withthe National Science Foundation. PyRAF is a product of the SpaceTelescope Science Institute, which is operated by AURA for NASA.
Milisavljevic & Fesen
Figure 2.
Example of a fully reduced and cleaned 2D spectrum of a slit position along the main shell from which 1D extractions weremade. Main shell fast-moving knots (FMKs) and a CSM-related quasi-stationary material (QSF) are labeled. Wavelength positions for thebrightest emission lines are marked.
Figure 3.
Examples of 1D spectra of ejecta knots of Cas A used in the Doppler reconstruction. The top left spectrum from the mainshell is an example template used to identify knots and determine radial velocities. Top right shows a spectrum with multiple overlappingknots. Bottom left is a Mixed Ejecta Knot (MEK) from the NE Jet region. Bottom right is an oxygen-rich knot also from the NE Jetregion. subtracted, flattened, and co-added to remove cosmicrays. Images were then wavelength calibrated in thedispersion axis using the comparison lamp images andstraightened in the spatial axis using tracings of stellarcontinua.Excellent background subtraction of the images wasachieved through the IRAF task background using afifth-order chebyshev function fit along the spatial di-rection that was sampled from the median of 50 pixelbins. This provided sky emission clean 2D images fromwhich 1D spectra could be extracted. An example of areduced and cleaned 2D image is shown in Figure 2.Each line of the 2D spectra was extracted andadded in weighted triplet groupings. No effort wasmade to catalog the H α , [N II ] λλ I emissions associated with circumstellar material (CSM) often referred to as Quasi-Stationary Flocculi(QSF; van den Bergh & Dodd 1970) frequently encoun-tered (see Fig. 2). Investigation of this pre-SN mass lossmaterial is left for future work.From these 1D extractions, knot velocities were mea-sured through cross-correlation of templates using thetask xcsao . The template was based on actual datafrom a representative knot exhibiting the common linesof [O III ] λλ I ] λλ II ] λλ III ] λ II ] λλ xcsao was run on all extracted 1Dspectra for Doppler velocities running between − − to find individual knots. This procedurewas found to work well even in cases with multiple knots(Figure 3, top right).D Kinematics of Cas A 5 Figure 4.
Measured main shell knot velocities. Projected radiifrom the COE have been converted to velocities using scale factor S = 0.022 km s − arcsec − . Left : Data from this paper. Best fitsemi-circle is shown.
Right : Data from DeLaney et al. (2010) forinfrared Ar II emissions shown over top data from this paper. Results of the preliminary selection were then manu-ally inspected for false positives. The remnant’s faintjet and other outlying ejecta knots required additionalcare owing to the range of emission lines expressed,such as mixed emission knots (Figure 3, bottom left),and oxygen-rich knots (Figure 3, bottom right). SeeFesen & Gunderson (1996) for additional details aboutthe spectroscopic properties of these types of knots. Insome cases velocities were determined manually using de-convolving techniques available in the task splot .Radial velocities of individual knots were then as-signed coordinates in right ascension (RA) and decli-nation (DEC). A variety of cross-checks ensured consis-tent and accurate coordinates. Stars with well-measuredcoordinates encountered during the progression of long-slit positions provided fudicial reference points. In somecases, bright stars were intentionally observed to pro-vide stellar continua by which accurate positional off-sets could be determined. Long-slit positions orientedeast-west were checked against overlapping slits orientednorth-south to further check positions and flux levels. Afinal check to all positions was aided through comparisonwith high-resolution
Hubble Space Telescope ( HST ) im-ages (Proposal 10286; PI: R. Fesen). Measured velocitiesare believed accurate to ±
40 km s − at the 68% confi-dence level. Positional uncertainties are estimated to beno more than 2 ′′ ; that is, of order the slit width.
3D DOPPLER RECONSTRUCTION
Cas A’s relatively young age and nearby distancemake it a prime candidate for Doppler velocity re-construction. Previous studies show that its ejectaknots lie roughly on a spherical shell traveling radi-ally outward from a unique center of expansion (COE;Reed et al. 1995; Lawrence et al. 1995; DeLaney et al.2010). The COE is known to within one arcsecond preci-sion and lies at the coordinates α (2000 .
0) = 23 h m . s δ (2000 .
0) = +58 ◦ ′ . ′′ v = r × S, (1)where v is the radial velocity, r is the angular distancefrom the explosion center, and S is the scaling factor.The value of S can be found by fitting the mea-sured Doppler velocities to a spherical expansion model.We follow procedures reviewed in DeLaney et al. (2010),which itself closely followed the work of Reed et al.(1995). The model is a semi-circle of the velocity dis-tribution, which can be parameterized in terms of thecenter of the velocity distribution, v c , the minimum ve-locity at which the semi-circle crosses the velocity axis, v m , and S , which relates the velocity axis to the spatialaxis as: ( r p /S ) + ( v D − v c ) = ( v c − v m ) , (2)where r p is the observed projected radius and v D is theobserved Doppler velocity.In Figure 4, the results of our least-squares fit to thedata are shown. The calculated Doppler velocities are v c = 760 ±
100 km s − , v m = − ±
200 km s − , and thescaling factor is S = 0 . ′′ ± . ′′
003 per km s − . The re-sults are consistent with those determined by Reed et al.(1995) using optical data, as well as with DeLaney et al.(2010) using Spitzer
IR data of [Ar II ] emission. Knotpositions in RA and DEC were scaled to velocities us-ing Equation 1. We used the COE and the calculatedDoppler velocity v c to define a three-dimensional centerof expansion in velocity space from which all reportedvector trajectories originate. RESULTS
The results from our survey encompassing 13,769 in-dividual data points are presented in Figure 5. Incorpo-rated into the data set are an additional 73 outer ejectaknots with measured radial velocities reported in Fesen(2001). The resulting Doppler maps have the highestspectral and spatial resolutions of any previous survey ofCas A, and represent the most complete catalog of itsoptically emitting ejecta material to date.Measured knot radial velocities are represented in Fig-ure 5 with a color-coded gradient. A surface recon-struction has been performed using a ball-pivoting algo-rithm to interpolate and smooth the original point cloud.The Doppler map is plotted against a background im-age of Cas A created from the
HST
ACS/WFC images(GO 10286; PI Fesen) obtained in filters F625W (coloredblue), F775W (green), and F850LP (red) as a visual aidto relate the 3D kinematic structure to Cas A’s appear-ance on the plane of the sky. Three viewing orientationsare shown: a face-on view, a rear view, and an angledview from above shown with the background
HST imageoffset. An animation has also been created showing theindividual data points and the surface reconstruction, ro-tated along the north-south and east-west axes, with andwithout the outer high-velocity ejecta knots (Movie 1).Below we discuss some specific properties of Cas Ahighlighting our high-resolution 3D reconstructions ofthe remnant’s kinematic structure.
Kinematic Properties of the Main Shell
Milisavljevic & Fesen
Figure 5.
3D Doppler reconstruction of Cas A’s optically emitting ejecta. The main shell and all sampled high velocity outer material isrepresented. Top panel shows the reconstruction with respect to the plane of the sky as observed from Earth with north up and east tothe left. The velocity gradient is color-coded and only blueshifted material is visible from this perspective. Middle panel shows the sameperspective rotated 180 ◦ with respect to the north-south axis. In this representation, the vantage point is from behind Cas A with onlyredshifted material shown. Bottom panel is an angled perspective showing the full range of velocities. The plane of the sky is shown offsetfor reference. A translucent sphere is a visual aid to help distinguish between front and back material. Refer to Movie 1 for an animationof these data. The large-scale distribution of Cas A’s main shell ofejecta is shown in Figure 6 where knots having an ex-pansion velocity outside a 6000 km s − sphere have beenexcluded. Consistent with the recent 3D models pre-sented in DeLaney et al. (2010) based on infrared data,we find the main shell to be dominated by morphologicalstructures in the form of partial or complete rings. Themain difference of our 3D reconstructions shown here tothose of DeLaney et al. (2010) is increased spatial resolu- tion (by a factor of ≈
4) and more precise radial velocitydata (by a factor of ≈ ′′ (0.5 pc) and 2 ′ (2 pc). In Figure 7,a Mercator projection of the main shell knots is shownto illustrate the relative scale and distribution of therings. These rings dominate the large-scale structure ofthe main shell and are present on the front, rear, andD Kinematics of Cas A 7 Figure 6.
Various perspectives of the Cas A reconstruction illustrating the ubiquity of large ejecta rings. Only ejecta knots in the mainshell are shown. Top left panel shows some of the more prominent ejecta structures. Noticeable rings of ejecta include the blueshiftedring in the north, and the much larger neighboring redshifted ring. Also seen is the extent to which the rings are extended radially andhence more crown-like than simple flat rings. The top right panel shows a better perspective of the large northern redshifted ring. Bottomleft panel shows a side profile viewing towards the base of the NE jet, while the bottom right panel shows Cas A from below where asouthwestern ring can be seen as well as a number of partial or broken ring structures. Refer to Movie 1 for an animation of these data.
Figure 7.
The main shell of Cas A’s optically-emitting ejecta asrepresented in a Mercator projection. The linear scale is equal in alldirections around any point and conformal, but the cylindrical mapprojection distorts the size and shape of large objects, especiallytowards the poles. both east and west side hemispheres.The largest ejecta ring is located along the north-ern limb and is nearly entirely redshifted (Figure 6, topright). Connected to it by a small bridge of ejecta knotsis a much smaller but thicker ring in the north that isentirely blueshifted (Figure 6, top left). These two rings contain nearly all of the optical and infrared emissionalong the remnant’s northern limb. A separate but alsovery large, outwardly extending broken ring is locatednear the base of the NE jet (Figure 6, bottom left). An-other ring of comparable size and radial extension liesimmediately below it.Five of the six most obvious ejecta rings are actu-ally short cylinders giving them a crown-like appearance.The bases of these crowns are sometimes gently curvedupward as seen in the smaller northern blueshifted ring(Figure 6, top left), and the large northern ring (Figure 6,top left and bottom left). The height in velocity space ofthese crowns that extend radially away from the COE isup to 1000 km s − . The radial extent of the broken ringlocated near the base of the NE jet is largest (Figure 6,top right).Besides large-scale ejecta rings and crowned cylinders,the data show areas with a frothy ejecta substructuredown to sizes of 10 ′′ ( ≈ × cm). These small-scale features are often observed at the boundaries oflarger rings in tightly arranged groupings. They can bedistinguished in the Mercator projection in Figure 7 asthe smallest rings and ellipses. Similar small-scale ring-like structures can be seen in HST images (Fig. 5 inFesen et al. 2001).The bulk of the remnant’s optically bright main shellejecta lies fairly close to the plane of the sky and canbe contained within a torus or thick disk structure (seeFigure 6, bottom left). With respect to a unit vector Milisavljevic & Fesen
Figure 8.
Vector plots of all sampled outer ejecta knots. V x is east–west, V y is north–south, and V z is radial velocity. The top left panelshows the perspective as observed from Earth, with north up and east to the left. All knots to the east are colored green, and all knots tothe west are colored blue. The top right panel shows what the jets look like as viewed along the north-south axis looking toward south.Two angled perspectives are shown in the bottom left and right panels. normal to the plane of the sky directed from the remnantto the observer, the vector normal to the torus’ equatorialplane is tilted approximately 30 ◦ to the west and 30 ◦ tothe north, in agreement to the findings of DeLaney et al.(2010). We note, however, that conspicuous deviationsfrom a simple disk structure are observed. For example,the compact, entirely blueshifted northern ring extendsbeyond the torus defined by the bulk of the main shell.We discuss this and other deviations from a thick diskstructure more in Section 6.2. The NE and SW High-Velocity Streams of Ejecta
Vector plots of outer high velocity ejecta are shownin Figure 8 where ejecta knots are represented by green(east) and blue (west) vectors from the remnant’s COE.An animation has been made that shows these data ro-tated completely about the north-south axis (Movie 2).Considerably more knots are detected in the NE than inthe SW. This imbalance is may be attributable in partto higher levels of extinction in the SW compared to theNE. Fortunately, detections in the SW jet features arenumerous enough to draw some conclusions about therelative properties of both high-velocity ejecta features.Previous studies of the remnant’s jets lacked the kine-matic resolution and depth required to ascertain whetherthey form a true bipolar structure. Our data indicatethat this is indeed the case. The NE and SW jet regionsshow a broad scattering of outlying, high-velocity knotsthat appear to be directed in nearly opposite directionsand is suggestive of opposing flows of SN debris. Al- though the NE material exhibits a slightly larger rangeof radial velocities compared to the SW, we find no clearkinematic distinction between the two distributions interms of opening half-angle and maximum expansion ve-locity.The top left panel of Figure 8 shows the normal ob-server perspective of north up and east to the left.The distribution seen is much like the ‘bowtie’ plot ofFesen et al. (2006b) (their Fig. 4) created from measuredproper motions of 1825 knots seen on
HST images takenover a 9 month period. A straight line can be drawnthat runs through the approximate center of both dis-tributions, and this is maintained as the perspective ischanged in the various panels of Figure 8.These data reveal the brighter and richer NE jet’skinematic structure in much greater detail than previousstudies. We find that instead of being a few thin streamsof ejecta, the NE region actually encompasses a broadrange of knot radial velocities including large redshiftedvelocities up to +5000 km s − . Prior spectroscopic dataof the NE jet detected only a a handful of knots at thesevelocities (Fesen & Gunderson 1996; Fesen 2001). Con-sequently, although the fastest knots lie close to the planeof the sky and thus exhibit relatively modest radial ve-locities, the NE jet’s true structure is actually a ratherbroad fan of ejecta knots with an opening half-angle be-tween 35 − ◦ .In Figure 9, four perspectives of the outer ejecta vec-tors with respect to the COE are shown along with themain shell reconstruction. The top two panels are inD Kinematics of Cas A 9 Figure 9.
Vector plots of all NE and SW ejecta knots (colored green and blue, respectively) shown together with the main shell ejecta(red). Four angled perspectives are shown with the upper left being as seen in the sky. A black arrow shows the inferred motion of theXPS with respect to the center of expansion (Fesen et al. 2006b). Refer to Movie 2 for an animation of these data. the same orientation as the top two panels of Figure 8,and use the same color scheme to distinguish the eastand west material. The possible connection between themain shell geometry and distribution of outer ejecta isdiscussed in Section 6.3. Also shown in Figure 9, isthe inferred motion of the X-ray point source (XPS) be-lieved to be a neutron star with a carbon atmosphere(Ho & Heinke 2009). The direction the XPS determinedoriginally in Fesen et al. (2006b) is towards an obvioushole in the distribution of outer ejecta and is briefly dis-cussed in Section 6.4. DISCUSSION
Although optically-emitting debris may constitute onlya small fraction of the total ejected mass in the Cas Aremnant, our observations offer superior spatial and kine-matic resolution compared to previous surveys obtainedvia
Chandra
X-ray and
Spitzer infrared observations.Moreover, Cas A’s fastest moving material is best probedthrough optical studies. For example, optical emissionfrom the NE and SW jets can be traced some 90 ′′ far-ther out than in X-rays or infrared, and only a handfulof outer optical ejecta knots around the remainder of theremnant are detected in even the deepest X-ray imagesof Cas A (see Fesen et al. 2006b).It is now abundantly clear from the results of our sur-vey and previous studies that the distribution of Cas A’sejecta is far from being random. At least half a dozen large and coherent ejecta ring-like structures are ob-served and their size and arrangement may be informingus about important properties of the explosion dynamicsand subsequent evolution of the expanding debris. How-ever, interpretation of these structures is complicated inthat one must disentangle the observed remnant proper-ties that may originate in the explosion from later influ-ences related to possible post-explosion radioactive heat-ing, ejecta interaction with the surrounding CSM and in-terstellar medium (ISM), and effects of the reverse shockon the ejecta.An important additional caveat is that Cas A con-tinues to evolve and has changed considerably in itsoptical appearance over the last 50 years that ithas been monitored (Kamper & van den Bergh 1976;van den Bergh & Kamper 1985). On-going propagationof the remnant’s reverse shock continues to excite newejecta, revealing more of its entire distribution. Hence,any conclusions made in the present must acknowledgethis evolution and recognize that we are probably not see-ing the entire remnant but only those parts which haverecently undergone shock heating.With these limitations in mind, below we attempt todescribe and understand the nature of the observed prop-erties of Cas A’s main shell and outlying high-velocityejecta.0 Milisavljevic & Fesen Figure 10.
Location of iron-rich X-ray emitting ejecta (blue) with respect to the main shell sulfur- and oxygen-rich optically-emittingejecta (red). The X-ray data shown is from DeLaney et al. (2010). Refer to Movie 3 for an animation of these data.
Ejecta Rings = Ni Bubbles?
Reed et al. (1995) and Lawrence et al. (1995) were thefirst to establish the existence of conspicuous rings ofreverse shocked ejecta in Cas A. These rings, clearlyvisible in the recent infrared survey of DeLaney et al.(2010), are large (diameter ∼ ∼ radius of Cas A)and are most certainly structures that are related to thetrue structure of the remnant’s debris field and indepen-dent of any strong influences of the remnant’s CSM/ISMenvironment. The optical data presented here does notalter this overall picture, but provides an improved, moredetailed examination of Cas A’s expansion properties.In Figure 10, we show the optically-emitting main shellejecta along with the location of X-ray emitting iron-rich material measured from Chandra data presented inDeLaney et al. (2010). An animation showing these datasets rotated about the north-south and east-west axes hasbeen provided (Movie 3). Each of the three largest con-centrations of Fe-rich ejecta (i.e., along the west, north,and southeast limbs) lie within and bounded by ringstructures, strongly suggesting a causal relationship.DeLaney et al. (2010) noted the coincidence of largeejecta rings with the three regions of Fe–K X-ray emis-sion and argued that these and other less prominent fea-tures are regions where the ejecta have emerged fromthe explosion as ‘pistons’ of faster than average ejecta.In this view, the remnant’s main shell rings representthe intersection points of these pistons with the reverseshock, similar to the bow-shock structures described byBraun et al. (1987).An alternative explanation first suggested byBlondin et al. (2001) that we favor is that the ob-served ejecta rings represent cross-sections of largecavities in the expanding ejecta created in part by apost-explosion input of energy from plumes of radioac-tive Ni-rich ejecta. Li et al. (1993) have describedhow this input of energy might account for the high-volume filling factor of Fe in SN 1987A despite its smallmass, and Basko (1994) and Blondin et al. (2001) haveinvestigated hydrodynamic simulations based on thismodel. One possible consequence of this scenario is thecompression of surrounding non-radioactive material bythe rising and expanding bubbles of radioactive Ni-rich ejecta, ultimately giving way to a “Swiss cheese” ejectastructure. It is important to note that the main shellejecta structures are observed as rings because we arebiased by the reverse shock that only excites an thinshell of material. Thus, it is reasonable to suspect theobserved ejecta rings to be cross sections of what areactually larger spherical geometries; i.e., bubbles.The turbulent motions that would initiate this Ni bub-ble structure in Cas A are not unlike recent 3D sim-ulations of the large-scale mixing that takes place inthe shock-heated stellar layers ejected in the explo-sion of a 15.5 M ⊙ blue supergiant star presented inHammer et al. (2010). As shown in their Figure 2, theprogenitor’s metal-rich core is partially turned over withnickel-dominated fingers overtaking oxygen-rich bullets.Although the evolution of these simulations is stronglydependent on the internal structure of the progenitorstar (Ugliano et al. 2012), it is still tempting to drawan association between the Ni-rich outflows seen in theHammer et al. (2010) models and the rings of Cas A.However, there are difficulties with invoking a Ni bub-ble picture to explain how the X-ray emitting Fe ejectaare framed by rings of optical ejecta. Fe associated withthe bubble effect should be characterized by diffuse mor-phologies and low ionization ages (Blondin et al. 2001;Hwang & Laming 2003). The X-ray bright Fe we cur-rently see, however, is actually at an advanced ionizationage relative to the other elements and thus inconsistentwith a Ni bubble origin.It is possible that Fe associated with Ni bubbles re-mains undetected. The undetected Fe would likelybe of small mass compared to what is currently ob-servable since estimates of Cas A’s chemical abun-dances are close to model predictions ( M Fe ∼ . − . M ⊙ ; Hwang & Laming 2012), and be of low den-sity (Eriksen et al. 2009). Presently, however, there isweak evidence of unshocked Fe ejecta in infrared data(DeLaney et al. 2010). Moreover, an exhaustive X-raysurvey of Cas A by Hwang & Laming (2012) led them toconclude that almost all of the Fe ejected by the super-nova is now well outside the reverse shock and visible inX-rays, with very little left in the center of the remnant.Indeed, any Fe associated with the bubble effect mayD Kinematics of Cas A 11be too faint and underionized to be readily identifiable(Hwang & Laming 2003; Hwang & Laming 2012).Thus, while considerable evidence exists that arguesagainst associating the X-ray bright Fe-rich material withNi bubbles, the large-scale and coherent ejecta rings ob-served in Cas A are obviously not random filamentaryejecta structures, and their sizes, near spherical shapes,and ubiquity throughout the remnant are intriguinglysuggestive of a Ni bubble origin. The strikingly tightspatial coincidence between some optical ejecta rings andthe boundaries of Fe-rich ejecta seen in X-rays, perhapsbest seen in the remnant’s large northern ring of ejecta,lends support to this picture. Future efforts to locateadditional material internal and/or external to Cas A’smain shell may help clarify our understanding of the pro-cesses responsible for the ring-shaped morphology of themajority of Cas A’s optically emitting ejecta.Finally, we note that there is some evidence that ex-pansion of radioactive Ni-rich ejecta may have also in-fluenced the ejecta structure at smaller scales. A frothyring-like substructure at scales of 10 ′′ (0 . HST images (Fesen et al. 2001) is observedaround some of the large rings. These small featuresmay be due to fragments of Ni-rich material pushing outsurrounding material (e.g., blueshifted ring in Figure 6,top left panel). However, this substructure should notbe confused with possible influence from Rayleigh-Taylorand Kelvin-Helmhotz instabilities that can develop inclumpy ejecta due to their interaction with the reverseshock front. These processes may be the dominant onesbehind the pronounced crown-like structure of the rings(cf. Basko 1994; Blondin et al. 2001).
Main Shell Geometry and Velocity Asymmetry
In addition to the large-scale ring structures of themain shell, there are two other observed properties of themain shell ejecta that may be related to the explosion dy-namics. The first is the − − radial ve-locity asymmetry. This asymmetry has long been recog-nized in even the earliest optical observations of the rem-nant made by Minkowski (1968), and later more fully ap-preciated by observations of Lawrence et al. (1995) andReed et al. (1995). The second is the torus-like geome-try observed for the bulk of the remnant’s optical and in-frared bright ejecta, discussed at length by Markert et al.(1983) and more recently by DeLaney et al. (2010).Interpreting these properties is complicated given theproblem of distinguishing whether the remnant’s ob-served geometry is a faithful representation of the ex-plosion dynamics or has been significantly influencedby an inhomogeneous local CSM and ISM. Reed et al.(1995) argued that environmental effects were significantin understanding the remnant’s blue and redshift veloc-ity asymmetry. They concluded that the density of thesurrounding medium is greater on the blueshifted nearside, which would inhibit forward expansion and resultin an apparently redshifted COE. In support of this con-clusion, they cite the observed asymmetrical distributionof radial velocities of the circumstellar QSF knots, a ma-jority of which (76%) are blueshifted.However, DeLaney et al. (2010) and Isensee et al.(2010) note that there are major structural differencesbetween the front and back interior surfaces of Cas A.These differences in brightness and structure could be due to different masses, densities, and energies in thetwo different directions suggesting the observed velocityasymmetry might be intrinsic to the explosion itself.DeLaney et al. (2010) further argue that the torus-likedistribution of bright main shell ejecta is a result of theexplosion dynamics. They conclude that although inter-action with the CSM affects the detailed appearance ofthe remnant, the bulk of the symmetries and asymme-tries in Cas A are intrinsic to the explosion. In theirframework, the remnant is the result of a flattened ex-plosion where the highest velocity ejecta were expelledin a thick torus tilted not far off the plane of sky in anumber of large-scale pistons.In possible support of their picture, DeLaney et al.(2010) noted that Fesen (2001) found that relativelysmall angles off the sky plane (i.e., <
30 deg) seem tobe the rule for the remnant’s small, fastest-moving out-lying ejecta knots. That is, although many hundreds ofoutlying ejecta knots can be seen around most of Cas A’speriphery, none have been detected close to its center.Our extensive survey also was unable to find any small,very high-velocity knots within the main shell and pro-jected near the remnant’s center. Rather than support-ing a torus-like ejecta arrangement, however, this couldindicate that a near tangent viewing angle may be impor-tant for a knot’s optical visibility. Fesen (2001) originallyproposed that a near tangent viewing factor might holdfor both the remnant’s outer knots and its main shell.If true, this would mean some portion of the remnant’sfacing and rear sides may remain undetected, and thusthe currently observed torus-like configuration of brightemission not reflective of the remnant’s true 3D struc-ture.The notion of a disk geometry for the Cas A rem-nant that happens to lie fairly near to the plane of thesky is also inconsistent with recent infrared observationsthat detected some of the remnant’s interior, unshockedejecta. Isensee et al. (2010) observed redshifted [Si
III ]and [O IV ] line emissions from interior debris along asight line close to remnant center exhibiting velocitiesapproaching 5000 km s − . These velocities are compa-rable to the − − maximum radialvelocities seen for main shell ejecta, as well as the in-ferred maximum transverse velocity of around 6300 kms − derived from proper motions ( ≈ . ′′
39 yr − ) of mainshell ejecta knots assuming a remnant distance of 3.4kpc (van den Bergh & Kamper 1983; Thorstensen et al.2001).Further support of this viewpoint are some conspicuousdepartures from a thick-disk geometry in the distributionof main shell ejecta. Rear-facing ejecta seem to cut offabruptly along an angled plane, while the front-facingejecta do not. In particular, the northern, blueshiftedring forms a conspicuous bulge from the apparent torus(see bottom left panel of Figure 6). There are aspects ofthe asymmetric distribution of the remnant’s outermostejecta that also indicate possible influence of an inho-mogeneous environment. In Figure 11, the distributionof material as observed from two sides of Cas A lookingtoward the base of the NE and SW jets is shown. Thematerial in the SW jet region largely follows the abruptboundary cutoff, as does the main shell ejecta. Ejecta inthe the NE jet region, however, do not follow this cut-off.Deep H α images of Cas A show evidence for much more2 Milisavljevic & Fesen Figure 11.
Side-looking perspectives of Cas A showing non-uniform distribution of material. The SW material (blue) and main shellejecta (gray) cut off at a boundary defined by an angled plane. The NE material (green) does not cut off at the boundary, and this maydue to interaction with nearby ISM/CSM H-rich material seen in H α . H-rich CSM/ISM material in the eastern limb than in thewestern limb (Reynoso et al. 1997; Fesen 2001), which isconsistent with the cut-off being due to environmentaleffects.
Opposing High-Velocity Ejecta Outflows
The nature of the anomalously high velocities of ejectain the NE and SW regions has long been a puzzle.Minkowski (1968) suggested the NE jet might be the solesurviving part of an outer, high-velocity shell that hasbeen subsequently decelerated in all other directions. Al-ternatively, because the distribution of emitting gas maynot necessarily be the same as the actual distribution ofgas in Cas A, the jets may simply be the most visiblepart of a larger population of the remnant’s outer, highvelocity knots (Fesen & Gunderson 1996).Along similar lines, asymmetric debris structures pro-duced by circumstellar interaction has been theoreticallymodeled by Blondin et al. (1996) who show how a jet-like feature of SN ejecta can be generated in the progen-itor’s equatorial plane from pole/equator density gradi-ents in the local CSM. Thus, these high-velocity regionsmay be secondary features caused by instability-poweredflows from an equatorial torus, where the explosion axisis loosely defined by the X-ray iron-rich regions foundin the southeast and northwest (Burrows et al. 2005;Wheeler et al. 2008; Rest et al. 2011).However, the very prominent rupture-like features inthe main shell near the base of the NE jet visible inX-ray, optical, IR, and radio images (cf. Hughes et al.2000; van den Bergh & Kamper 1985; Ennis et al. 2006;Anderson & Rudnick 1995; DeLaney et al. 2010) are cer-tainly suggestive of an explosive formation process. Al-though perhaps not indicative of a jet-induced explosion,there is nonetheless substantial evidence that the NE andSW jets are somehow associated with core-collapse explo-sion dynamics in some way.For example, analyses of
Chandra
X-ray spectra indi-cate that the NE jet could not be formed through inter-action with a cavity or lower density region in the CSM(Laming & Hwang 2003; Laming et al. 2006). The rem-nant’s ejecta would have expanded rapidly into a CSM cavity, and led to a density too low for electron-ion equi-libration to match the observed values. Laming et al.(2006) estimate that a factor of two more explosion en-ergy went into the NE jet direction compared to otherregions.The location of chemically distinct optically-emittingknots in both jets is also consistent with an unusual,high-velocity ejection of underlying mantle material.Knots exhibiting a mix of H α , [N II ], [O II ], [S II ], and[Ar III ] line emissions are only observed in the jet re-gions, suggesting an eruptive and turbulent mixing ofthe underlying S, O, and Ar rich material with pho-tospheric H- and N-rich layers (Fesen & Becker 1991;Fesen & Gunderson 1996; Fesen 2001). Furthermore, op-tical knots in the NE and SW jets lying farthest outand possessing the highest ejection velocities show nodetectable emission lines other than those of [S II ], sug-gesting an origin from the S-Si-Ca-Ar rich layer deep in-side the progenitor star (van den Bergh & Kamper 1983;Fesen & Gunderson 1996; Fesen 2001; Fesen et al. 2001).As shown in Figure 8, we find that the overallconical distributions of the NE and SW jet regionsexhibit comparable maximum expansion velocities andare broadly anti-parallel in an orientation consistentwith a bipolar jet-counterjet structure. However, theobserved opening half-angles of these two flows ( ∼ ◦ )is quite broad and is not what would be anticipatedin a highly-collimated (opening half-angle < ◦ ) jet-induced explosion. Thus, the broadness and estimatedlow energies of the NE and SW jets make it unlikelythat they are signatures of a jet-induced explosionscenario (e.g., Khokhlov et al. 1999; Wheeler et al.2002; Akiyama et al. 2003; Fryer & Warren 2004;Laming et al. 2006).There appears to be only a tenuous connection betweenthe NE and SW jet regions and any of the main shell’sejecta rings. The centers of cones broadly defining theNE and SW ejecta streams as sampled in our surveyonly loosely point to centers of ejecta rings. The respec-tive distributions are illustrated in Figure 9, where theouter knot vectors are plotted with the main shell surfaceD Kinematics of Cas A 13reconstruction. Particularly suggestive is the NE jet re-gion, where the bulk of high-velocity material appears tobe emanating from a large main shell ring that has a pro-nounced radial extension. In this regard, the main shellrings and the jet-like streams of higher velocity debriscould be dynamically related.However, for both the NE and SW jets, many knotshave trajectories that do not pass through any main shellrings. In light of the large number of faint, outermosthigh-velocity jet knots, it is possible that additional ma-terial at even higher positive and negative radial veloci-ties may lie undetected, broadening the jets even more.Future work mapping the chemical abundances of jetknots to their kinematics might contribute to a betterunderstanding of their nature and possible relationshipto the main ejecta shell. Motion of X-ray Point Source
An additional clue to the nature of the explosion dy-namics may come from understanding the properties ofthe inferred motion of the central XPS. The location ofthe XPS ≈ ′′ to the southeast of the estimated COE(P.A. ≈ ◦ ) indicates a transverse velocity ‘kick’ of ≈
350 km s − imparted to the compact object duringthe explosion (Fesen et al. 2006b).Intriguingly, the projected line connecting the NE andSW jets lies nearly perpendicular to the inferred directionof the XPS (see Figure 8). Assuming the NE/SW axisto be the most significant in the original core-collapse,this runs counter to most jet-induced explosion modelsthat predict that the neutron star will undergo a kickroughly aligned with the jet axis (Burrows & Hayes 1996;Fryer & Warren 2004).There is a noticeable absence of outlying ejecta alongthe axis of the XPS’s inferred motion (Fesen 2001;Fesen et al. 2006b). With the possible exception of un-likely and extremely localized extinction conditions, deepimaging of Cas A surveying the entire remnant ensuresthat there is no observational bias that could explain thelack of detection of ejecta in these regions. Althoughpossibly just a coincidence, it is interesting that the XPSshould be moving with a trajectory in line with the onlytwo significant gaps of the outer ejecta which are locatednearly perpendicular to the NE-SW jet axis.We note that Smith et al. (2009) showed that strong[Ne II ] emission is observed symmetrically around theXPS, with an axis also close to the direction of the in-ferred kick. Although Smith et al. (2009) determinedthat there is not enough mass/energy in the Ne-crescentsto affect the XPS motion, they may perhaps arise due tothe same dynamical asymmetries that led to the XPSmotion (DeLaney et al. 2010). Comparison to Other Young SNRs
Recent 3D reconstructions of other young CCSN rem-nants have suggested that many of the kinematic prop-erties observed in Cas A are not unique. Although onlya handful of known SNRs are appropriate for this kindof analysis, those that have been studied have revealedevidence of the same large-scale ejecta rings, velocityasymmetries, and high-velocity bipolar asymmetries asobserved in Cas A.For example,Vogt & Dopita (2010) mapped the [O
III ] λ III ] λ ∼
12 pc in diameter inclined at anangle of ∼ ◦ to the line of sight, and evidence of a po-lar jet associated with a very fast oxygen-rich knot. Thesurvey led Vogt & Dopita (2011) to speculate that theoverall observed shape of the SNR to have been stronglyinfluenced by the pre-supernova mass loss from the pro-genitor star.Work by Winkler et al. (2009) on the core-collapseSNR G292.0+1.8 shows that it too has properties likethose seen in Cas A. Proper motion measurements ofG292.0+1.8’s fast filaments exhibit systematic motionsoutward from a point near the center of the radio/X-ray shell that lies 46 ′′ northwest from the young pulsarPSR J1124-5916. The inferred motion of the pulsar hasa transverse velocity of 440 km s − , which is close to thatobserved in Cas A’s XPS. Additionally, the fastest ejectain G292.0+1.8 form a bipolar outflow along an axis ori-ented roughly north–south in the plane of the sky. Theremnant appears to have undergone a complex evolu-tion resulting from a possibly asymmetric SN explosionand interaction with non-uniform ambient ISM (see alsoBraun et al. 1986; Park et al. 2002), which is a scenarionot too different from the one we envision for Cas A. Connections to Extragalactic SN Observations
Finally, we consider how the kinematic properties ofCas A discussed here may be linked to phenomena ob-served in the late-time optical emissions of extragalac-tic supernovae. Milisavljevic et al. (2012) summed allmain shell spectra presented in Section 2 into a single,integrated spectrum mimicking what the remnant wouldappear as as an unresolved extragalactic source. Similar-ities were seen between this integrated Cas A spectrumand several late-time optical spectra of decades-old ex-tragalactic SNe.Particularly well-matched with Cas A were SN 1979C,SN 1993J, SN 1980K, and the ultra-luminous supernovaremnant in NGC 4449. Milisavljevic et al. (2012) foundthat both Cas A and decades old extragalactic SNe showpronounced blueshifted emission with conspicuous linesubstructure in [O I ], [O II ], [O III ], [S II ], and [Ar III ].Since the emission line substructure observed in the for-bidden oxygen emission line profiles of Cas A are associ-ated with the large-scale rings of ejecta, we suggest thatidentical features in the intermediate-aged SNe whichhave often been interpreted as ejecta ‘clumps’ or ‘blobs’are, in fact, probable signs that large-scale rings of ejectaare common in SNe. The link is strongest between Cas Aand SN 1993J, given that these were both Type IIb explo-sions and that they exhibit extremely similar forbiddenoxygen profiles despite the three centuries of evolutionthat separate them.4 Milisavljevic & Fesen CONCLUSIONS
We have presented 3D kinematic reconstructions ofthe optical emission from the supernova remnant Cas Abased on radial velocity measurements extracted fromlong-slit and multi-slit spectra. This data set is of highspatial and kinematic resolution and encompasses the NEand SW streams of high velocity ejecta jets that until nowhave never been surveyed to this depth.The major results and conclusions of this study are asfollows:1) We confirm the findings of several previous kine-matic studies of Cas A that show the bulk of the rem-nant’s main shell ejecta to be arranged in several well-defined complete or broken ring-like structures. Thesering structures have diameters that can be comparableto the radius of the remnant ( ∼ ∼ . ◦ with respect to the plane of the sky and exhibit a − − radial velocity asymmetry. Unlike theconclusion reached by DeLaney et al. (2010), we suggestthat this observed geometry and velocity asymmetry isnot representative of the true overall kinematic proper-ties of the original explosion. Instead, an observationalbias caused by a near tangent viewing angle effect andinteraction with an inhomogeneous CSM/ISM environ-ment has likely contributed to some of Cas A’s presentlyobserved kinematic properties.3) The size, shape, and ubiquity of large-scale, reverseshock heated ejecta rings in Cas A are consistent with abubble-like interior structure. It is possible that some ofthese structures were generated by the input of energyfrom radioactive Ni rich ejecta that produced low den-sity bubbles surrounded by non-radioactive, intermediatemass element-rich ejecta. While the spatial coincidencebetween X-ray emitting Fe-rich material and some of theremnant’s large ejecta rings is suggestive of such a origin,the high ionization age of the X-ray bright Fe suggeststhat this material is not directly associated with Ni bub-bles.4) Our deep optical survey shows the remnant’s NEand SW jet features, which contain Cas A’s highest ve-locity ejecta, to be unexpectedly broad streams of ejectaknots with comparable conical opening half-angles of ap-proximately 40 ◦ . The jets appear to lie in an orientationconsistent with an opposing and wide bipolar outflow.The broadness together with low energy estimates of theNE and SW jets argue against them being associatedwith a jet-induced explosion. However, several kinematicand chemical properties support the view that they were formed during core collapse and independent of effectsfrom the local CSM or ISM. Faint ejecta in both jetsat even higher expansion velocities may lie as yet unde-tected.We end by noting that similar radial velocity datasets spanning X-ray, optical, and infrared wavelengthsof other young SN remnants may provide additionalinsights and tests of various core-collapse SN mod-els. A comparison of Cas A’s properties with thoseof other young remnants, especially ones supplementedwith chemical abundance analyses in addition to kine-matics, may also help to further constrain details of theexplosion mechanism, post-explosion dynamics, ejectainstabilities, and explosive nucleosynthesis. With thegrowing success obtaining multi-perspective optical spec-tra of light echoes associated with other historical Galac-tic supernovae e.g., SN 1572 (Krause et al. 2008), andthose nearby in the LMC and SMC (Rest et al. 2005),the possibility of developing 3D reconstructions of otheryoung CCSN supernova remnants with known SN sub-types may be possible in the future. ACKNOWLEDGMENTS
REFERENCESAkiyama, S., Wheeler, J. C., Meier, D. L., & Lichtenstadt, I.2003, ApJ, 584, 954Anderson, M. C., & Rudnick, L. 1995, ApJ, 441, 307Basko, M. 1994, ApJ, 425, 264Besel, M.-A., & Krause, O. 2012, A&A, 541, L3Blondin, J. M., Borkowski, K. J., & Reynolds, S. P. 2001,ApJ, 557, 782Blondin, J. M., Lundqvist, P., & Chevalier, R. A. 1996,ApJ, 472, 257Blondin, J. M., Mezzacappa, A., & DeMarino, C. 2003, ApJ, 584,971Braun, R., Goss, W. M., Caswell, J. L., & Roger, R. S. 1986,A&A, 162, 259Braun, R., Gull, S. F., & Perley, R. A. 1987, Nature, 327, 395Burrows, A., Dessart, L., Ott, C. D., & Livne, E. 2007,Phys. Rep., 442, 23Burrows, A., & Hayes, J. 1996, Physical Review Letters, 76, 352Burrows, A., Hayes, J., & Fryxell, B. A. 1995, ApJ, 450, 830Burrows, A., Livne, E., Dessart, L., Ott, C. D., & Murphy, J.2006, ApJ, 640, 878Burrows, A., Walder, R., Ott, C. D., & Livne, E. 2005, The Fateof the Most Massive Stars, 332, 350Chevalier, R. A. 2005, ApJ, 619, 839Chevalier, R. A., & Kirshner, R. P. 1979, ApJ, 233, 154Chevalier, R. A., & Oishi, J. 2003, ApJ, 593, L23Claeys, J. S. W., de Mink, S. E., Pols, O. R., Eldridge, J. J., &Baes, M. 2011, A&A, 528, A131Delaney, T. A. 2004, PhD thesis, Univ. of Minnesota