The Unification of Asymmetry Signatures of Type Ia Supernovae
J.R. Maund, P.A. Hoeflich, F. Patat, J.C. Wheeler, P. Zelaya, D. Baade, L. Wang, A. Clocchiatti, J. Quinn
DD raft version O ctober
22, 2018
Preprint typeset using L A TEX style emulateapj v. 08 / / THE UNIFICATION OF ASYMMETRY SIGNATURES OF TYPE IA SUPERNOVAE J ustyn R. M aund , P eter H¨ oflich , F erdinando P atat , J. C raig W heeler , P aula Z elaya ,D ietrich B aade , L ifan W ang , A lejandro C locchiatti and J ason Q uinn Draft version October 22, 2018
ABSTRACTWe present a compilation of the geometry measures acquired using optical and IR spectroscopy and opticalspectropolarimetry to probe the explosion geometry of Type Ia SNe. Polarization measurements are sensitiveto asymmetries in the plane of the sky, whereas line profiles in nebular phase spectra are expected to traceasymmetries perpendicular to the plane of the sky. The combination of these two measures can overcometheir respective projection e ff ects, completely probing the 3D structures of these events. For 9 normal TypeIa SNe, we find that the polarization of Si II λ p S i II ) is well correlated withits velocity evolution (˙v
S i II ), implying ˙v
S i II is predominantly due to the asymmetry of the SNe. We find onlya weak correlation between the polarization of Si II and the reported velocities (v neb ) for peak emission ofoptical Fe II and Ni II lines in nebular spectra. Our sample is biased, with polarization measurements beingonly available for normal SNe which subsequently exhibited positive (i.e. redshifted) v neb . In unison theseindicators are consistent with an o ff -centre delayed detonation, in which the outer layers are dominated by aspherical oxygen layer, mixed with an asymmetric distribution of intermediate mass elements. The combinationof spectroscopic and spectropolarimetric indicators suggests a single geometric configuration for normal TypeIa SNe, with some of the diversity of observed properties arising from orientation e ff ects. Subject headings: supernovae: general — techniques: spectroscopic — techniques: polarimetric INTRODUCTION
The importance of Type Ia Supernovae (SNe Ia) as cosmo-logical distance indicators has been demonstrated over the lasttwelve years (Riess et al. 1998; Perlmutter et al. 1999). Keyto their utilisation as probes of the Universe’s expansion is anunderstanding of their apparent homogeneity (with a magni-tude dispersion at maximum light of σ B ∼ .
8; Phillips 1993)and the role of the underlying explosion mechanism. Modelsto describe these events, as the explosion of a carbon-oxygenChandrasekhar mass white dwarf in a single or double degen-erate binary system (Branch et al. 1995), proceed with burn-ing as deflagrations (Gamezo et al. 2004; R¨opke et al. 2006),detonations or a combination of the two: delayed detonations(Khokhlov 1991). These di ff erent mechanisms will imprint asignature on the resulting chemical abundances and geome-tries of the ejecta (H¨oflich et al. 2006).The geometries of SNe Ia have been probed using di ff er-ent techniques. At early times, spectropolarimetry of theseevents has shown a wide array of asymmetries (see Wang &Wheeler 2008, for a review), in the plane of the sky. In gen- Dark Cosmology Centre, Niels Bohr Institute, University of Copen-hagen, Juliane Maries Vej, DK-2100 Copenhagen Ø, Denmark; [email protected] Department of Astronomy & Astrophysics, University of California,Santa Cruz, 95064, U.S.A. Sophie & Tycho Brahe Fellow Department of Physics, Florida State University, Tallahassee, Florida32306-4350, U.S.A.; [email protected] ESO - European Organisation for Astronomical Research in the SouthernHemisphere, Karl-Schwarzschild-Str. 2, 85748 Garching b. M¨unchen, Ger-many; [email protected]; [email protected] Department of Astronomy and McDonald Observatory, The Universityof Texas, 1 University Station C1402, Austin, Texas 78712-0259, U.S.A.;[email protected] Departamento de Astronom´ia y Astrof´isica, PUC Casilla 306, Santiago22, Chile; [email protected]; [email protected] Department of Physics, Texas A&M University, College Station, Texas77843-4242, U.S.A.; [email protected] eral, the continuum polarization, which probes the shape ofthe photosphere, is of order a few tenths of a percent, indicat-ing departures from a spherical symmetry of <
10% (H¨oflich1991; Wang et al. 1997). Significant line polarization is ob-served, showing the distribution of elements in the ejecta, pre-dominantly for Ca II (e.g. for SN 2001el Wang et al. 2003;Kasen et al. 2003) and the principal classification feature forSNe Ia, Si II λ II and the lightcurve decline rate parameter ∆ m (B).Nebular phase spectroscopy of SNe Ia in the optical and in-frared (IR), when the ejecta are optically thin and the spectraare dominated by the Fe-group elements, can be used to revealthe structure at the centre of the explosion; albeit projectedonto the radial velocity direction (H¨oflich et al. 2006; Moto-hara et al. 2006; Gerardy et al. 2007; Maeda et al. 2010a). Thealmost-unblended Fe line at 1.6 µ m shows peculiar line pro-files such as flat topped profiles indicating material with cen-tral cavities, and Gerardy et al. (2007) observed identical pro-files and features in mid-IR lines. This strongly supports thatthe line asymmetries are kinematic in nature and all but ex-clude explanations due to uncertainties in the atomic physics,which may lead to an underestimation of blends, or opticalthickness e ff ects.Maeda et al. (2010b) linked the asymmetries of the centrallayers with expansion velocities less than ∼ < − withthe outer layers. They found that the peak of the nebular emis-sion line profiles, in the optical, exhibited blue or red shifts,consistent with an asymmetric distribution along the line ofsight. It was found that this kinematical o ff set was relatedto the evolution of the velocity of the absorption minimum ofthe Si II λ S i II ) at early times; suggesting both are signaturesof the departure of SNe Ia from simple spherical symmetry.Probing the inner structure and correlations with outer lay-ers is central to answering key questions of supernova theory a r X i v : . [ a s t r o - ph . S R ] A ug (for a review see H¨oflich 2006), such as whether the asym-metries in the inner and outer layers have a common physicalorigin. In this letter we discuss the correlations between ge-ometry indicators derived from early and late spectroscopicobservations and spectropolarimetric observations of SNe Ia,to approach a unified model of the behaviour of these events. POLARIZATION AND THE VELOCITY GRADIENT OF SI II Wang et al. (2007) observed that the polarization of theSi II line for normal SNe Ia peaks at ∼ − S i II for a range ofSNe Ia were compiled from those reported by Benetti et al.(2005) and Maeda et al. (2010b). These were cross checkedagainst the list of SNe Ia with the polarization of Si II λ − II polarization was esti-mated from spectropolarimetric observations at -9 and -4 days(Zelaya et al., in prep.), with ˙v S i II = ± − day − (Si-mon et al. 2009). The polarization of Si II λ + + . S iII givenby Leonard et al. 2005 and Yamanaka et al. 2009). Giventhat the observations of SN 2002bf and 2009dc were con-ducted later than − lower limits of p S i II at − S i II = ± − day − (usingthe data of Matheson et al. 2008) , which is higher than the20 km s − day − measured for the similar SN 1991T (Phillipset al. 1992; Taubenberger et al. 2008).The Si II polarizations for the peculiarly faint 1991bg-likeSNe 1999by and 2005ke were taken from Howell et al. (2001)and Patat et al., (in prep.), respectively. The velocity evolu-tion of Si II in this particular sub-class has been found to berelatively homogeneous (Taubenberger et al. 2008). We de-rive an average value ˙v S i II = . ± . − day − fromthe reported velocities of 4 members of the subluminous sub-class (SNe 1991bg Turatto et al. 1996; 1999by Vink´o et al.2001; Howell et al. 2001; 2005bl Taubenberger et al. 2008;and 2005ke Patat et al., in prep.). In the absence of furtherspectroscopic data, we assume that this value of the decelera-tion of Si II is valid for SN 2005ke.The relation between the velocity gradient of Si II and theassociated polarization is shown on Fig. 1. There is a distinctcorrelation for normal SNe Ia and, excluding SN 2004dt, wefind p S iII = . + . × ˙v S iII (with χ ν = .
07, Pearson cor-relation coe ffi cient r = . p S i II and the light-curve de-cline rate parameter ∆ m ( B ), Wang et al. (2007) found thatthese SNe are significant outliers from the linear correlationfound between these two parameters. SN 2004dt is not asignificant outlier ( < σ ), and including it in the fit slightlychanges the form of the relation. It may indicate that the rela-tion at very high ˙v S i II is no longer linear.The dividing line between the High Velocity Gradient(HVG) and Low Velocity Gradient (LVG) SNe is a deceler-ation of ∼
70 km s − day − in Fig. 1 (see Benetti et al. 2005).The relative numbers in each group, however, are dictatedby the selection e ff ect of the number of SNe with both suf- The data were acquired from the SUSPECT archive:http: // suspect.nhn.ou.edu / ∼ suspect / F ig . 1.— The observed line polarization of Si II λ − (cid:4) ),whereas the peculiarly faint SNe 1999by and SN 2005ke are indicated ingrey ( (cid:63) ) and the single representative of the SN 1991T-like SNe 2001V isindicated in green ( (cid:78) ). Lower limits on the polarization of Si II for two SNeare presented in purple ( • ). The dashed grey line at 70 km s − day − separatesthe HVG and LVG SNe Ia (Benetti et al. 2005). The solid black line indicatesthe best-fit straight line to the normal SNe Ia, while the dashed black line isfor the best-fit including SN 2004dt. ficiently dense early spectroscopy, to derive the Si II veloc-ity evolution, and early spectropolarimetry. The gap betweenthe HVG and LVG SNe may not, therefore, be representativeof significant di ff erentiation between the two types of normalSNe Ia. The lower polarization limits provided by SNe 2002bfand 2009dc may fill this gap and indicate that SNe Ia form acontinuous distribution on this diagram. Caution is required,however, as Leonard et al. (2005) suggest SN 2002bf maybe related to the peculiar SN 2004dt, due to its high veloci-ties, and Yamanaka et al. (2009) and Tanaka et al. (2010) sug-gest SN 2009dc may have arisen from a super-Chandrasekharmass White Dwarf progenitor. POLARIZATION AND THE NEBULAR PHASE VELOCITY OFFE-GROUP ELEMENTS
We examined the compilation of line-of-sight velocity o ff -sets v neb for the peaks of the emission lines of [Fe II ] λ II ] λ t > ff sets of [Ni III ] 7.35 and 11.002 µ m and [Ni IV ]8.41 µ m at 118 days (although Gerardy et al. do not observean o ff set associated with [Co III ] 11.89 µ m ).The velocity o ff sets for SN 2001V were determined for theoptical Fe II and Ni II lines, from a spectrum acquired at 106days (Matheson et al. 2008). We measured, however, di ff erentvelocity o ff sets for the two species. This may indicate a morecomplicated di ff erentiation of the distributions of Fe and Nithan observed for the normal HVG and LVG SNe Ia; and maybe commensurate with SN 2001V belonging to the subclassof 1991T-like SNe Ia.In Fig. 2 the polarization of Si II λ ff set for events for whichboth data exist. There is an obvious sample bias, however, asthe only SN Ia with a suitable polarization measurement and ablue shifted emission line profile is SN 2004dt. The remainingSNe Ia from our sample all have redshifted emission lines intheir optical nebular spectra.The apparent separation between the normal HVG and LVGSNe Ia in Fig. 2 approximately reflects the correlation be-tween ˙v S i II and p S i II established in §
2, and the observationof Maeda et al. (2010b) that the nebular lines of SNe Ia withhigher values of ˙v
S i II exhibit larger red displacements. Thecorrelation between v neb and p S i II for normal SNe Ia, exclud-ing SN 2004dt and 2001V, is much weaker ( r = .
54) than thecorrelation of Fig. 1. In this sample, based on measurementsof v neb alone, it is di ffi cult to distinguish between LVG andHVG normal SNe Ia (e.g. SNe 2001el and 2006X). Withoutfurther SNe with blue-shifted nebular velocities and polariza-tion measurements, it is di ffi cult to ascertain the behaviour inthe blue-shifted portion of the diagram. Given the relationshipestablished in §
2, those LVG SNe with negative v neb shouldappear in the lower left-hand quadrant of Fig. 2, tending to-wards zero polarization.Leloudas et al. (2009) measured a velocity gradient of˙v
S i II = ± − day − for the normal SN 2003hv. Thiswould suggest p S i II ∼ .
5% at -5 days, which is consistentwith adopting the observed polarization at + . ± .
05% as a lower limit. The implieddecrease in the p S i II between − + II ] line profiles for SN 2003hv as blue-shifted( − ± − ; Maeda et al. 2010b). Depending onthe strength of Si II polarization at -5 days, if the polarizationdecreased dramatically, SN 2003hv may occupy a similar lo-cus to that of SN 2004dt on Fig. 2. This would support atrend relating p S i II and v neb orthogonal to that apparent fromthe central group of points on Fig. 2, with SN 2004dt andSN 2001V at the extremes of a general trend that runs throughthe cloud of normal HVG and LVG SNe Ia. If the polariza-tion of SN 2003hv did not decrease substantially from -5 to + DISCUSSION & CONCLUSIONS
Maeda et al. (2010b) observed a relationship between ˙v
S i II and v neb , and interpreted it as indicating a single asymmet-ric geometry for Type Ia SNe. This asymmetry gives rise tothe apparent diversity amongst this SN class due to the dif-ferent orientations at which individual SNe are observed. Inalso considering polarimetric observations, we find a goodlinear correlation between p S i II , an established indicator ofgeometry, and ˙v
S i II , implying the latter is a signature of thegeometry. The observed tight correlation between these twoparameters for normal SNe Ia implies that the asymmetriesprobed by p S i II are unlikely to be due to a random, clumpyline forming region, rather it indicates the role of a large scaleasymmetry in the ejecta. We find that p S i II shows a weakcorrelation with the later nebular velocity, that may indicatea possible correlation between the asymmetries inferred forthe layers observed at early and late times. This correlation isexpected to be weak, however, as p S i II and v neb probe orthog-onal projections of the geometry. This suggests that in trying F ig . 2.— The observed line polarization of Si II λ − ff set. HVG SNe are indicated in orange( (cid:78) ), LVG SNe are in black ( (cid:4) ), SN 2005df (without a measurement of ˙v Si II )is shown in grey ( • ) and the two measurements for SN 2001V are shown ingreen ( (cid:63) ). An approximate trend between p Si II and v neb for normal SNe Iais indicated by the grey line. to understand the influences of geometry on measurable pho-tometric and spectroscopic parameters, such as ∆ m ( B ), it ispreferable to use the early time indicators ˙v S i II and p S i II .Benetti et al. (2005) suggested the di ff erence between HVGand LVG SNe Ia arises from the orientation at which the ejectaare viewed. They hypothesised that SNe Ia will be observed asHVG SNe if the ejecta approaching the observer were mixedwith heavy elements, increasing the opacity and keeping thephotosphere at high velocities at early times (H¨oflich et al.1993). In their model, LVG SNe Ia are observed from otherangles, dominated by intermediate mass elements (IMEs),through which the photosphere has already receded even atearly times. A key consequence of this model is asymmet-ric excitation of the ejecta leading to an asymmetric photo-sphere, in particular for HVG SNe. This is in contrast tothe low limits placed on the departure of the photospheresof SNe Ia from spherical symmetry from polarimetry of theseevents ( p cont ∼ . − . II , amongst other line features, isgenerally observed to be polarized in SNe Ia, whereas thecontinuum is not. This requires that the line forming regionbe asymmetrically distributed across a spherical photosphere(Kasen et al. 2003; Leonard et al. 2005; Maund et al. 2010).Benetti et al. (2005) and Maeda et al. (2010b) interpret the ve-locity of Si II λ II “W” feature at 5640Å, whichshows lower velocities and a less severe velocity gradient thanSi II , more accurately reflects the true photospheric velocity.Hachinger et al. (2006) observed that the velocity determinedfor S II is lower than that determined from Si II λ ff erence is largest for the HVGSNe. This implies Si II λ II is dependent on themass contained in the Si II line forming region above the pho-tosphere (deposited by the protrusions of IMEs into the outerlayers). Conversely to the correlation of Benetti et al. (2005),an increase in opacity due to Fe-group elements mixed intothe outer ejecta layers would slow the apparent recession ofthe photosphere leading to an LVG rather than HVG SN.For SN 2004dt, Wang et al. (2006) and Leonard et al.(2005) observed negligible polarization associated with theO I λ II and O I occupied the same velocity space,but had very di ff erent polarization properties, suggested thatportions of the oxygen layer were mixed with IMEs (such asSi) with an asymmetric distribution (H¨oflich et al. 2006). Thecombination of low polarization for oxygen and the contin-uum, but with polarized silicon has been observed for otherSNe Ia of both LVG and HVG classes such as SNe 2001el,2006X and 2009dc (Wang et al. 2003; Patat et al. 2009;Tanaka et al. 2010). The uniformity of oxygen in the outerejecta of SNe Ia is further demonstrated by Hachinger et al.(2006), who find almost constant line strength for the absorp-tion of O I λ ff ected by other processes such as the progenitor’s rota-tion and its binary companion (Howell et al. 2001; H¨oflichet al. 2006). In the model of Maeda et al. (2010b), significantheavy elements from the deflagration are mixed into the outerlayers, on the side of the ejecta from which an HVG SN Ia isobserved.The polarization of Si II , and the inferred asymmetries, andthe correlation with ˙v S i II suggest a single o ff set, with HVGSNe Ia being those with an o ff set Si distribution mixed intothe outer O layer in the direction of the observer and the prod-ucts of deflagration receding. At significant angles away from the o ff set direction, the Si is found in a thinner layer, moreevenly excited by the underlying Ni substrate, leading to anLVG SN Ia with a lower polarization for Si II λ Ni are spherically distributed in the ejecta.The two subluminous SNe, 1999by and 2005ke, show lowline polarization. Within the framework of single degener-ate scenarios, however, such SNe Ia produce Ni in the de-flagration, whereas during the detonation phase Si-group ele-ments are produced at the expense of Ni (Hoflich et al. 1995;H¨oflich et al. 2002). In models of normal SNe Ia, the major-ity of the Ni is produced during the detonation phase. Thefact that line polarization is observed to be strongest in normalSNe Ia and small in subluminous SNe Ia argues against defla-gration instabilities as the origin of the observed correlations.As polarization probes asymmetries in the plane of the skyand nebular phase velocities probe asymmetries perpendicu-lar to the plane of the sky, the combination of both these mea-sures provides the opportunity to overcome projection e ff ectsand completely probe the three-dimensional structure of theseevents. By bringing together the measurements of p S i II , ˙v
S i II and v neb , a portion of the spectral diversity of normal SNe Iamay be understood in terms of simple orientation e ff ects con-cerning a single geometry and explosion by an o ff centre de-layed detonation. ACKNOWLEDGEMENTS
The research of JRM is funded through the Sophie & TychoBrahe Fellowship. The Dark Cosmology Centre is supportedby the DNRF. The research of JCW is supported in part byNSF grant AST-0707769. AC, JQ, and PZ thank the supportof Basal CATA PFB 06 /
09, FONDAP No. 15010003, andP06-045-F (ICM / MIDEPLAN / Chile). The authors are grate-ful to Stefan Taubenberger for useful discussions concerningthe velocity evolution of the faint subclass of SNe Ia.
REFERENCESBenetti, S., et al. 2005, ApJ, 623, 1011Branch, D., Livio, M., Yungelson, L. R., Bo ffi , F. R., & Baron, E. 1995,PASP, 107, 1019Gamezo, V. N., Khokhlov, A. M., & Oran, E. S. 2004, Physical ReviewLetters, 92, 211102Gerardy, C. L., et al. 2007, ApJ, 661, 995Hachinger, S., Mazzali, P. A., & Benetti, S. 2006, MNRAS, 370, 299H¨oflich, P. 1991, A&A, 246, 481H¨oflich, P. 2006, Nuclear Physics A, 777, 579H¨oflich, P., Gerardy, C. L., Fesen, R. A., & Sakai, S. 2002, ApJ, 568, 791H¨oflich, P., Gerardy, C. L., Marion, H., & Quimby, R. 2006, NewAstronomy Review, 50, 470Hoflich, P., Khokhlov, A. M., & Wheeler, J. C. 1995, ApJ, 444, 831H¨oflich, P., Mueller, E., & Khokhlov, A. 1993, A&A, 268, 570Howell, D. A., H¨oflich, P., Wang, L., & Wheeler, J. C. 2001, ApJ, 556, 302Kasen, D., et al. 2003, ApJ, 593, 788Khokhlov, A. M. 1991, A&A, 245, 114Leloudas, G., et al. 2009, A&A, 505, 265Leonard, D. C., Li, W., Filippenko, A. V., Foley, R. J., & Chornock, R. 2005,ApJ, 632, 450Maeda, K., Taubenberger, S., Sollerman, J., Mazzali, P. A., Leloudas, G.,Nomoto, K., & Motohara, K. 2010a, ApJ, 708, 1703Maeda, K., et al. 2010b, Nature, 466, 82Matheson, T., et al. 2008, AJ, 135, 1598Maund, J., et al. 2010, ApJ, subm.Motohara, K., et al. 2006, ApJL, 652, L101 Patat, F., Baade, D., H¨oflich, P., Maund, J. R., Wang, L., & Wheeler, J. C.2009, A&A, 508, 229Patat, F., Benetti, S., Cappellaro, E., Danziger, I. J., della Valle, M., Mazzali,P. A., & Turatto, M. 1996, MNRAS, 278, 111Perlmutter, S., et al. 1999, ApJ, 517, 565Phillips, M. M. 1993, ApJL, 413, L105Phillips, M. M., Wells, L. A., Suntze ff , N. B., Hamuy, M., Leibundgut, B.,Kirshner, R. P., & Foltz, C. B. 1992, AJ, 103, 1632Riess, A. G., et al. 1998, AJ, 116, 1009R¨opke, F. K., Gieseler, M., Reinecke, M., Travaglio, C., & Hillebrandt, W.2006, A&A, 453, 203Simon, J. D., et al. 2009, ApJ, 702, 1157Tanaka, M., et al. 2010, ApJ, 714, 1209Taubenberger, S., et al. 2008, MNRAS, 385, 75Turatto, M., Benetti, S., Cappellaro, E., Danziger, I. J., Della Valle, M.,Goui ff es, C., Mazzali, P. A., & Patat, F. 1996, MNRAS, 283, 1Vink´o, J., Kiss, L. L., Cs´ak, B., F˝ur´esz, G., Szab´o, R., Thomson, J. R., &Mochnacki, S. W. 2001, AJ, 121, 3127Wang, L., Baade, D., H¨oflich, P., Khokhlov, A., Wheeler, J. C., Kasen, D.,Nugent, P. E., & et al. 2003, ApJ, 591, 1110Wang, L., Baade, D., H¨oflich, P., Wheeler, J. C., Kawabata, K., Khokhlov,A., Nomoto, K., & Patat, F. 2006, ApJ, 653, 490Wang, L., Baade, D., & Patat, F. 2007, Science, 315, 212Wang, L., & Wheeler, J. C. 2008, ARAA, 46, 433Wang, L., Wheeler, J. C., & H¨oflich, P. 1997, ApJL, 476, L27 ++