arXiv:0807.1695v1 [astro-ph] 10 Jul 2008

The m etam orphosisofSupernova SN 2008D /XRF 080109:a link between Supernovae and G RBs/Hypernovae

Accepted by Science Paolo A .M azzali1;2;3;4 ,Stefano Valenti5;6,M assim o D ella Valle 7;8;9, G uido C hincarini10;11,D anielN .Sauer1,Stefano Benetti2,Elena Pian12, T sviPiran13,Valerio D ’Elia14,N ancy Elias-R osa1,R a aella M argutti10, Francesco Pasotti10,L.A ngelo A ntonelli14,Filom ena Bufano2, Sergio C am pana11,Enrico C appellaro2,Stefano C ovino11, Paolo D ’A vanzo11,Fabrizio Fiore14,D ino Fugazza11,R oberto G ilm ozzi8, D eborah H unter5,K ate M aguire5,Elisabetta M aiorano15,Paola M arziani2, N icola M asetti15,Felix M irabel16,H ripsim e N avasardyan2, K en’ichiN om oto3;4;17,Eliana Palazzi15,A ndrea Pastorello5, N ino Panagia18;19,L.J.Pellizza20,R e’em Sari13,Stephen Sm artt5, G ianpiero Tagliaferri11,M asaom iTanaka3,Stefan Taubenberger1, N ozom u Tom inaga3,C arrie Trundle5,M assim o Turatto19

1

1

M ax-Planck Institut fur A strophysik, K arl-Schwarzschild-Str.1, 85748 G arching, G erm any 2 Istituto N azionale diA stro sica-O A Pd,vicolo dell’O sservatorio,2,I-35122 Padova,Italy 3 D epartm ent of A stronom y, Schoolof Science, U niversity of Tokyo, Bunkyo-ku, Tokyo 113-0033,Japan 4 R esearch C enterforthe Early U niverse,SchoolofScience,U niversity ofTokyo,Bunkyoku,Tokyo 113-0033,Japan 5 A strophysicsR esearch C entre,SchoolofM athsand Physics,Q ueen’sU niversity,Belfast, BT 7 1N N ,N orthern Ireland,U K 6 D ipartim ento diFisica,U niversita’diFerrara,via G .Saragat1,44100 Ferrara,Italy 7 Istituto N azionalediA stro sica,C apodim onteA stronom icalO bservatory,Salita M oiariello 16,I-80131 N apoli,Italy 8 European Southern O bservatory,K arl-Schwarzschild-Str.2,D -85748 G arching,G erm any 9 InternationalC enterforR elativisticA strophysicsN etwork,Piazzaledella R epubblica 10, I-65122 Pescara,Italy 10 D epartm ent ofPhysics,U niversita’diM ilano-Bicocca,Piazza delle Scienze 3,I-20126 M ilano,Italy 11 Istituto N azionale diA stro sica,Brera A stronom icalO bservatory,V ia E.Bianchi46, I-23807 M erate (LC ),Italy 12 Istituto N azionale diA stro sica-O AT s,V ia T iepolo 11,I-34131 Trieste,Italy 13 T he R acah Institute ofPhysics,H ebrew U niversity,Jerusalem 91904,Israel 14 Istituto N azionale diA stro sica,R om e A stronom icalO bservatory,V ia diFrascati33, I-00040 M onte Porzio C atone,Italy 15 Istituto N azionale diA stro sica,Istituto diA stro sica Spaziale e Fisica C osm ica,V ia P.G obetti101,I-40129 Bologna,Italy 16 European Southern O bservatory,A lonso de C ordova 3107,Santiago,C hile 17 InstituteforthePhysicsand M athem aticsoftheU niverse,U niversity ofTokyo,K ashiwa, C hiba 277-8582,Japan 18 Space Telescope Science Institute,Baltim ore,M D ,U SA 19 Istituto N azionale diA stro sica, C atania A stronom icalO bservatory, V ia S.So a 78, I-95123 C atania,Italy 20 Instituto de A stronom a y F sica delEspacio,C .C .67,Buenos A ires,A rgentina To w hom correspondence should be addressed;E-m ail: m azzali@ m pa-garching.m pg.de

2

T he only supernovae (SN e) to have show n early -ray or X -ray em ission thus far are overenergetic, broad-lined T ype Ic SN e (H ypernovae - H N e). R ecently,SN 2008D show s severalnovelfeatures: (i) w eak X R F ,(ii) an early,narrow opticalpeak,(iii) disappearance of the broad lines typicalofSN Ic H N e,(iv) developm ent ofH e lines as in SN e Ib. D etailed analysis show s that SN 2008D w as not a norm al 51 SN : its explosion energy (E 6 10 erg) and ejected m ass ( 7M ) are interm ediate betw een norm alSN eIbc and H N e. W e derive that SN 2008D w as originally a 30M star. W hen it collapsed a black hole form ed and a w eak,m ildly relativistic jet w as produced,w hich caused the X R F . SN 2008D is probably am ong the w eakest explosions that produce relativistic jets. Inner engine activity appears to be present w henever m assive stars collapse to black holes. O n 2008 January 9.57U T the X -R ay Telescope (X RT )on board Sw iftdetected a weak X -ray Flash (X R F 080109) in the galaxy N G C 2770(1). O ptical follow -up revealed the presence ofa supernova coincident w ith the X R F [SN 2008D ;R A (2000) = 09 09 30.625; D ec (2000)= + 33 08 20.16](2).W e detected SN 2008D photom etrically from A siago O bservatory on 10.01 January 2008U T , only 10.5 hours after the Sw ift detection. Early spectra showed broad absorption lines superposed on a blue continuum ,and lacked hydrogen orhelium lines(3). A ccordingly,SN 2008D wasclassi ed asa broad-lined SN Ic(4). SN e ofthis type are som etim es associated w ith G am m a-ray Bursts (G R B,R efs. 5,6) or X R Fs (7,8). T he spectra resem bled those of the X R F-SN 2006aj (8) or the non-G R B H N SN 2002ap (9) (Fig. 3,top),but a com parison suggests that SN 2008D was highly reddened: we estim ate that E (B V )tot = 0:65 m ag (see SO M ). T he hostgalaxy ofX R F 080109/SN 2008D ,N G C 2770 [redshiftz = 0:006494,distance 31 M pc],is a spiralgalaxy sim ilar to the M ilky W ay,M 31,or ESO 184-G 82,the host of SN 1998bw /G R B 980425. N G C 2770 has roughly solar m etallicity and a m oderate starform ation rate, 0:5M yr 1 (see SO M ).In contrast,typicalhost galaxies ofG R Bs are sm all,com pact,som ew hat m etal-poor,and highly star-form ing(10). In addition to the weak X R F,SN 2008D show s a num ber ofpeculiarfeatures,m ostof w hich are new . T he opticallight curve had two peaks (Fig. 1): a rst,dim m axim um (V 18:4)wasreached lessthan 2 daysaftertheX R F.A ftera briefdeclinethelum inosity increased again,reaching principalm axim um (V = 17:37) 19 days after the X R F.A n 18-20 day risetim e is typicalofG R B-H N e: norm alSN eIc reach m axim um in 10-12 days. Few stripped-envelope SN e have very early data,and in G R B-H N e a rst peak m ay be m asked by the afterglow light. A rst narrow opticalpeak was only seen in the Type Ib SN 1999ex (SN eIb are sim ilar to SN eIc but show strong helium lines,R ef. 4),the Type IIb SN 1993J (SN eIIb are sim ilar to SN eIb but stillhave som e hydrogen),and the Type Ic X R F-SN 2006aj.W hen itwasdiscovered,SN 1999ex wasdropping from a phase ofhigh lum inosity(11). It reached principalm axim um 20 days later,as did SN 2008D . A nother novelfeature is the spectralm etam orphosis (Fig. 2). U nlike SN e2006ajand 2002ap,the broad absorptions did not persist. A s they disappeared,H eilines developed (12). By principalm axim um SN 2008D had a narrow -lined,Type Ib spectrum (Fig. 3, bottom ). 3

Broad lines require m aterialm oving w ith velocity v > 0:1c,w here c is the speed of light(13).T heirdisappearance im pliesthatthe m assm oving athigh velocitieswassm all. Late developm ent of H ei lines, previously seen only in SN 2005bf (14),is predicted by theory (15). H elium levels have high excitation potentials, exceeding the energy of therm alphotonsand electrons. Excitation can be provided by the fastparticlesproduced as the -rays em itted in the decay chain of56N itherm alize (16).T his isthe process that m akes SN e shine. In the rst few days after explosion therm alization is e cient because ofthe high densities and not enough particles are available to excite helium . O nly w hen density drops su ciently can m ore particles escape the 56N izone and excite helium . W e reproduced the spectralevolution and the light curve ofSN 2008D after the rst 51 narrow peak using a m odelw ith M ej 7M and spherically sym m etric E 6 10 erg,of 50 w hich 0:03M ,w ith energy 5 10 erg,are at v > 0:1c (Figs. 1,3,and SO M ).O ur light curve ts indicate that SN 2008D synthesised 0:09M of 56N i,like the non-G R B H N SN Ic 2002ap (9) and the norm alSN Ic 1994I (17) but m uch less than the lum inous G R B-H N SN 1998bw (6). T he rapid rise in lum inosity follow ing the rst peak requires that som e 56N i(0:02M ) was m ixed uniform ly at allvelocities > 9000km s 1. T his is a typicalfeature ofH N e,and indicatesan asphericalexplosion (18).A sphericity m ay a ect our estim ate ofthe energy,but not the 56N im ass (19). C om paring them assoftheexploding H e-starthatwederived w ith evolutionary m odels ofm assive stars,we nd thatthe progenitorhad m ain sequence m ass 30M .A starof thism ass islikely to collapse to a black hole,asdo G R B/SN e (20).So,SN 2008D shared severalfeaturesofG R B/H N e.H owever,allSN e w ith G R Bsorstrong X R Fsinitially had velocities higher than SN 2008D or SN 2002ap (Fig.S3) and never showed helium . H ad the H e layer not been present in SN 2008D ,the explosion energy would have accelerated the inner core to higher velocities,and broad lines m ay have survived. T hecharacterizing featuresofSN 2008D (weak X R F, rstnarrow opticalpeak,initially broad-lined SN Ic spectrum that later transform ed into a narrow -lined SN Ib spectrum ) m ay be com m on to allSN eIb,orat leasta signi cant fraction ofthem ,and m aybe som e SN eIc,w hich however contain little or no helium . T he light curves ofvarious SN eIb are rather sim ilar (21). T he rst peak was observed only for SN 1999ex,but lack ofX -ray m onitoring probably prevented the detection ofm ore weak X R Fsand the early discovery ofthe associated SN e. O n the otherhand SN 2008D ,and possibly m ostSN eIb,wasm ore energetic than norm alcore-collapse SN e,including m ost SN eIc. Type II SN e in late Spiral/Irr galaxies (the typicalH ubble type of G R B hosts) are about 6 tim es m ore frequent than SN eIb (22). A lthough the serendipitous discovery of an SN Ib by X RT m ay be a statistical uctuation,itm ay also suggestthatthe softX -ray em ission accom panying SN 2008D is typicalofoverenergetic SN eIb,and absent (or very weak) in norm alcore-collapse SN e. T he X -ray spectrum ofSN 2008D (in total 500 photons)can be tted w ith either a sim ple power-law indicating a non-therm alem ission m echanism ora com bination ofa hot 6 black body (T = 3:8 10 K )and a powerlaw .In thelattercase,theunabsorbed lum inosity ofthe black-body com ponent is a sm allfraction ofthe totalX -ray lum inosity. T he high 43 tem perature and low lum inosity (L = 1:1 10 ergs 1) ofthe black-body com ponent at rst peak ( 100s after the onset ofthe X R F) im ply an em itting radius R ph 1010 cm (see SO M ,Section 4). T his is at least one order ofm agnitude sm aller than the size of 4

W olf-R ayet stars,the likely progenitors ofSN eIbc. T heX -ray areand the rstopticalpeak arem ostlikely associated (23).T hetim escale ofthe rst opticalpeak m ay suggest that it was related to shock breakout. A signature ofshock breakout is a hot black-body X -ray spectrum im m ediately after the explosion. T herm alX -ray em ission wassuggested forSN 2006aj(24),w hile no X -ray data are available forSN 1999ex. T he m odelofR ef.23 usesa sphericalcon guration and a black-body com ponent at 0.1keV ,below the X RT energy range. T his yelds a large radius,w hich they explain invoking the presence ofa dense surrounding m edium thatbulk-C om ptonize the shock breakout em ission to higher energies, producing the power-law spectrum observed by X RT between 0.3 and 10keV . O n the other hand,the angular size ofan em itting area w ith radius R ph 1010 cm is typicalofG R B jets.T hisleadsnaturally to an alternative scenario,thatwe propose here: X R F 080109 wasthe breakoutofa failed relativistic jetpowered by a centralengine asin G R Bs.T he jetfailed because itsenergy wasinitially low orbecause itwasdam ped by the H e layer,w hich isabsentin G R B-H N e,orboth.T he presence ofa jetissupported by our conclusion thata black hole was probably form ed w hen the star collapsed. T he m arginal breakout of the jet produced therm al X -rays and relativistic particles that caused the power-law X -ray com ponent. It also caused the rst opticalpeak: the tim escale ofthe rstpeak and the X -ray are and the corresponding radiiand tem perature are consistent w ith em ission from rapidly expanding,adiabatically cooling m aterial. T he weakness of the jet resulted in the low X -ray ux and the sm allam ount ofm aterialw ith v > 0:1c. T he failed jet contributed anisotropically to the SN kinetic energy. Lateralspreading of the ejecta w ith v > 0:1c leads to an angular size larger than the X -ray-em itting region, w hich is needed to produce the observed broad lines. T he sm allam ount ofhigh-velocity m aterialm oving along our line-of-sight m ay also indicate that we viewed the explosion signi cantly o -axis. T his can be tested by polarization or line pro les studies at late tim es, as in SN 2003jd (25). T he jet w ill spread further after breakout and it could dom inate the radio em ission at later tim es. T he scenario we propose im plies that G R B-like inner engine activity exists in all black hole-form ing SN eIbc (26).SN 2008D (and probably other SN eIb) has signi cantly higher energy than norm alcore-collapse SN e,although less than G R B/H N e. T herefore, it is unlikely that allSN e Ibc, and even m ore so allcore-collapse SN e produce a weak X -ray ash sim ilar to X R F 080109. T he presence ofhigh-energy em ission (G R B,X R F) depends on the jet energy and the stellar properties. O nly m assive,energetic,stripped SN eIc (H N e) have show n G R Bs. In borderline events like SN 2008D only a weak,m ildly relativistic jet m ay em erge, because the collapsing m ass is too sm all and a H e layer dam ps the jet. For even less m assive stars that stillcollapse to a black hole producing a less energetic explosion (e.g. SN 2002ap) no jet m ay em erge at all. Stars that only collapse to a neutron star are notexpected to have jets. SN 2008D thus links events that are physically related but have di erent observationalproperties.

R eferences and N otes 1. Berger,E.,& Soderberg,A .M .,G C N 7159 (2008). 5

2. D eng,J.,& Zhu,Y .,G C N 7160 (2008). 3. Valenti,S.,etal.,G C N 7171 (2008). 4. Filippenko,A .V .,A nn.Rev.A stron.A strophys.35,309 (1997). 5. G alam a,T .J.,etal.,N ature 395,670 (1998). 6. Iwam oto,K .,etal.,N ature 395,672 (1998). 7. Pian,E.,etal.,N ature 442,1011 (2006). 8. M azzali,P.A .,etal.,N ature 442,1018 (2006). 9. M azzali,P.A .,etal.,A strophys.J.572,L61 (2002). 10. Fruchter,A .S.,etal.,N ature 441,463 (2006). 11. Stritzinger,M .,etal.,A stron.J.124,2100 (2002). 12. M odjaz,M .,C hornock,R .,Foley,R .J.,Filippenko,A .V .,LiW .,& Stringfellow G ., G C N 7212 (2008). 13. M azzali,P.A .,Iwam oto,K .,N om oto,K .,A strophys.J.545,407 (2002). 14. Tom inaga,N .,etal.,A strophys.J.633,L97 (2005). 15. M azzali,P.A .,& Lucy,L.B.,M on.N ot.R .A stron.Soc.295,428 (1998). 16. Lucy,L.B.,A strophys.J.383,308 (1991). 17. Sauer,D .,etal.,M on.N ot.R .A stron.Soc.369,1939 (2006). 18. M aeda,K .,etal.,A strophys.J.593,931 (2003). 19. M aeda,K .,M azzali,P.A .,,& N om oto,K .,A strophys.J.645,1331 (2006). 20. M acFadyen,A .E.,& W oosley,S.E.,A strophys.J.524,262 (1999). 21. C locchiatti,A .,& W heeler,J.C .,A strophys.J.491,375 (1997). 22. C appellaro,E.,Evans,R .,Turatto,M .,A & A 351,459 (1999). 23. Soderberg,A .M .,etal.,N ature 453,469 (2008). 24. C am pana,S.,etal.,N ature 442,1008 (2006). 25. M azzali,P.A .,etal.,Science 308,1284 (2005). 26. M aeda,K .,et al.,Science 319,1220 (2008) 27. A rnett,W .D .,A strophys.J.253,785 (1982). 28. M azzali,P.A .,etal.,A strophys.J.645,1323 (2006). 29. C appellaro,E.,etal.,A stron.A strophys.328,203 (1997). 6

Figure 1: T he light curves of SN 2008D and of other Type Ibc SN e. T he shape of the light curve of SN 2008D is sim ilar to that of SN 1998bw and other G R B /H N e, and com parable to the non-G R B H N SN 1997ef,but m uch broader than the X R F/SN 2006ajor the norm alSN Ic 1994I.T his sim ilarity suggests a com parable value ofthe quantity M ej3=E ,w here M ej is the m ass ejected and E the explosion kinetic energy(27).A llknow n SN eIc w ith a broad lightcurve ejected a large m assofm aterial(28).Large valuesofM ej and E are also suggested by the presence ofH e m oving atv 10000km s 1: the velocity of H e in SN 2005bfwas lower(14). T he light curve ofSN 1999ex,w hich is sim ilar to that ofSN 2008D ,was 51 tted reasonably wellby a H e-star explosion m odelw ith M ej 5M ,E 3 10 erg(11). Such a m odel would also m atch the light curve ofSN 2008D ,but it probably would not reproduce the broad lines that characterizetheearly spectra.T hiswould requirea m odelcontaining som ehigh-velocity m aterial,leading to a larger E w ithout signi cantly a ecting the value ofMej or the light curve shape. T he line show s a synthetic bolom etric light curve com puted w ith a M ontecarlo code(29) for a m odelw ith M ej 7M , 51 E 6 10 erg.T he m odeldoesnotaddressthe physicsthatm ay be responsible forthe rstnarrow light curve peak,but only the m ain peak,w hich is due to di usion ofradiation in the SN envelope follow ing the deposition of -raysand positrons em itted in the decay chain 56 N ito 56 C o and 56 Fe.

7

Figure 2: Spectral evolution of SN 2008D . In the early phase the strongest features are broad Fe com plexes in the blue ( 4000 5000 A ),the Siii-dom inated feature near 6000A ,and C aiilines,both in the near U V (H & K ) and in the near IR (the IR triplet near 8500A ).O n the other hand O i7774 A , w hich isstrong in allH N e aswellasin allSN e Ibc,isconspicuously m issing. Starting 15 Jan,linesbegin to becom e narrow . In the later spectra,taken near m axim um ,H eilines have developed. T he strongest isolated lines are 6678 A ,seen near 6500A ,and 7065 A ,seen near 6900A .B oth lines indicate a helium velocity of 10000km s 1. T he other strong opticallines ofH eiare blended: 5876 A is blended w ith N aiD ,near 5600 A ,and 4471 A is blended w ith the broad Feiitrough near 4200 A .

8

Log fλ

1994I +7d (Ic) 1983N +4d (Ib) 2008D +4d (Ic−BL)

2002ap +3d (Ic−BL)

Log fλ

1994I +16d (Ic)

1983N +15d (Ib) 2008D +15d (Ib) 2002ap +14d (Ic−BL)

4000

5000

6000

7000

8000

9000

10000

Rest wavelength [Å] Figure 3: T he spectra ofSN 2008D com pared to those ofotherType Ibc SN e and to sim ulations. N ear the rst peak (top),SN 2008D has a broad-lined spectrum sim ilar to that ofSN 2002ap,a broad-lined SN w ithout a G R B (9), but di erent from both the norm alSN Ic 1994I and the SN Ib 1983N .A t the tim e ofthe m ain light curve peak (bottom ),the spectrum ofSN 2008D has narrow lines like SN e1994I and 1983N , w hile SN 2002ap and other H N e retain broad features throughout their evolution. A lso, SN 2008D developed H e lines (vertical ticks). A t this epoch, the spectra of SN e2008D , 1983N , and 1999ex are sim ilar. Synthetic spectra are overlaid on the two SN 2008D spectra (see SO M ).

9

Supplem entary Inform ation 1

O pticaland infrared observations

SN 2008D was observed photom etrically and spectroscopically w ith a num ber oftelescopes. Low -resolution spectra of SN 2008D have been obtained w ith the European Southern O bservatory (ESO ) 8.2m Very Large Telescope (V LT )-U T 2,equipped w ith FO R S2; the 4.2m W illiam H erschel Telescope (W H T ) equipped w ith ISIS; the 3.6m Telescopio N azionale G alileo (T N G ), equipped w ith D O LO R ES for optical observations and w ith N IC S for near-IR observations; the 3.6m ESO -N ew Technology Telescope (N T T ), equipped w ith EM M I; the 2.5m N ordic O ptical Telescope (N O T ), equipped w ith A LFO SC ;the 2.2m C alar A lto telescope (C A ),equipped w ith C A FO S;the 2.0m Liverpool Telescope (LT ),equipped w ith R AT C am ; the 1.82m A siago-Ekar telescope (A s1.82m ), equipped w ith A FO SC ; the 0.60m R apid Eye M ount telescope (R EM ), equipped w ith RO SS for opticalobservations and w ith R EM IR for near-IR observations. Tables S1 and S2 report the log ofthe observations covering the rst m onth afterthe X R F,w ith details on the acquisition,and the photom etric results. C olum n 1 gives the observation date,C olum n 2 the telescope and instrum ent used,C olum n 3 the observing setup,C olum n 4 theaverage seeing during thephotom etric acquisition (TableS1)and the spectralresolution (Table S2),C olum n 5 the observed V m agnitude (Table S1) and the spectrophotom etric standard used to calibrate the spectra (Table S2). T he photom etric errors are 1 uncertainties. M onochrom atic light curves are show n in Fig. S1,and the spectralevolution in Fig. 2 in the m ain paper).

10

Supplem entary Table S1: P hotom etric observations of SN 2008D D ate Telescope+ Setup Seeing V (2008 U T ) Instrum ent (arcsec) 7.005 Jan A s1.82m + A FO SC U BV R I 2.0 < 19:0 10.012 Jan A s1.82m + A FO SC U BV R I 2.5 19:10 0:06 11.213 Jan T N G + N IC S JH K 0.9 { 12.168 Jan N O T + A LFO SC U BV R I 2.0 18:41 0:10 13.006 Jan T N G + N IC S JH K 0.8 { 13.212 Jan T N G + D O LO R ES U BV R I 0.9 18:49 0:05 14.174 Jan N O T + A LFO SC U BV R I 1.6 18:29 0:05 16.205 Jan LT + R AT C am U BV R I 1.9 18:03 0:04 17.270 Jan LT + R AT C am U BV R I 1.6 17:96 0:02 18.267 Jan LT + R AT C am U BV R I 1.9 17:79 0:04 20.024 Jan LT + R AT C am U BV R I 2.2 17:64 0:06 20.153 Jan R EM + R EM IR H 3.5 { 21.310 Jan R EM + R EM IR H 2.9 { 22.309 Jan R EM + R EM IR H 2.9 { 23.123 Jan R EM + R EM IR JK 1.4 { 25.197 Jan V LT + FO R S2 BV R I 1.0 17:36 0:04 25.906 Jan LT + R AT C am U BV R I 1.9 17:36 0:03 28.021 Jan C A + C A FO SC U BV R I 1.2 17:33 0:05 28.987 Jan T N G + N IC S JH K 1.4 { 29.128 Jan LT + R AT C am U BV R I 1.2 17:34 0:03 30.127 Jan LT + R AT C am U BV R I 2.9 17:35 0:04 31.239 Jan R EM + RO S+ R EM IR R JH K 3.0 { 31.947 Jan T N G + N IC S JH K 3.1 { 31.964 Jan LT + R AT C am U BV R I 1.3 17:38 0:05

Supplem entary Table S1: P hotom etric observations of SN 2008D ,continued D ate Telescope+ Setup Seeing V (2008 U T ) Instrum ent (arcsec) 01.160 Feb LT + R AT C am U BV R I 1.3 17:43 0:03 05.111 Feb R EM + RO S+ R EM IR R JH K 3.2 { 06.114 Feb LT + R AT C am BV R I 1.3 17:58 0:04 08.162 Feb LT + R AT C am U BV R I 2.1 17:74 0:03 11.091 Feb T N G + D O LO R ES U BV R I 1.8 18:05 0:04 13.091 Feb R EM + RO S+ R EM IR JK 3.2 {

11

Supplem entary Table S2: Spectroscopic observations of SN 2008D D ate Telescope+ Setup R esolution standard (2008 U T ) Instrum ent (A ) 11.097 Jan T N G + D O LO R ES LR -B 12 H D 93521 12.045 Jan N O T + A LFO SC gm 4 14 Feige34 13.056 Jan T N G + N IC S IJ 7 H ip10559 13.187 Jan T N G + D O LO R ES LR -R 13 Feige34 13.392 Jan V LT + FO R S2 300V ;300I 10 H ILT 600 14.194 Jan N O T + A LFO SC gm 4 14 G 191-B2B 14.944 Jan T N G + D O LO R ES LR -B 13 Feige34 16.280 Jan V LT + FO R S2 300V ;300I 10 LT T 3864 25.212 Jan V LT + FO R S2 300V -300I 10 H ILT 600 28.033 Jan C A + C A FO SC b200 13 Feige34 28.270 Jan N T T + EM M I gm 2 12 Feige34 28.935 Jan T N G + N IC S IJ+ H K 7 H ip10559 28.966 Jan A s1.82m + A FO SC gm 4+ gm 2 25 Feige34 29.958 Jan A s1.82m + A FO SC gm 4+ gm 2 25 Feige34 31.957 Jan T N G + N IC S IJ+ H K 7 H ip10559 01.002 Jan W H T + ISIS R 300B+ R 158R 10 H D 93521 04.107 Feb T N G + D O LO R ES LR -B+ LR -R 12 Feige34 11.100 Feb T N G + D O LO R ES LR -B+ LR -R 11 G 191-B2B 12.185 Feb C A + C A FO SC b200 13 Feige34

12

K-2.8 H-2.5 J-2.2

12

14

I-1.9 magnitude

R-1.6 16

V-1.3 18

B-1 U

20

22

-5

0

5

10

15

20

25

30

35

days from explosion

Figure 4: T he m onochrom atic light curves ofSN 2008D .

13

40

2. T he properties of N G C 2770,the host of SN 2008D /X R F 080109. N G C 2770 isan Scgalaxy oflum inosity classIII-IV ,sim ilarto theM ilky W ay (S1),and very di erentfrom thehostsoflong G R Bs(S2).Itsnucleusisratherdi use,and according to FIR ST radio m aps it does not coincide w ith any radio peak. T he SD SS spectrum of the nucleus is typical of star-form ing nuclei of late-type spiral galaxies: relatively narrow lines (FW H M < 200km s 1) and diagnostic em ission line ratios are consistent w ith photoionization by hot stars. A ratio [O iii] 5007/H = 0:68,uncorrected because of the clearly visible H absorption, m akes N G C 2770 a galaxy w ith an H ii nuclear spectrum (S3). Since the Balm erlinesare m ainly due to photoionization by hotstars,the star form ation rate can be estim ated from indicators such as the H ux (0:2M yr 1), 1 IR A S uxes (0:45M yr ),or the radio lum inosity at 1.4G H z (0:7M yr 1). From the FIR Lum inosity wederive(S4)a SN rate 0:01SN eyr 1.T hesevaluesindicatethatN G C 2770 does nothave a very high star form ation lum inosity. T here is no strong evidence of a hidden non-therm alsource,even atradio frequencies,asalready show n by ref. S5,w ho failed to nd a high surface brightness nuclear radio source in N G C 2770.

14

3. E stim ate ofthe reddening to SN 2008D and derivation ofthe bolom etric light curve. W e estim ate dust extinction towards SN 2008D from spectralcom parison w ith other SN eIbc.U sing E (B V )= 0:16 forSN 1983N (S6),an extinction E (B V )tot = 0:65m ag forSN 2008D m akes the two SN e alm ostidentical(Fig.3 in the m ain paper). T his value is con rm ed by two independent checks: 1) it is com patible w ith the neutralhydrogen 21 colum n density 6 10 cm 2 estim ated from X -ray spectral tsassum ing a G alacticgas-todustratio (7);2)ityieldsopticaland infrared colourlightcurvessim ilarto those ofother SN eIb/c.C onsidering theuncertainties,weadoptforSN 2008D E (B V )tot = 0:65 0:15. M ost ofthe extinction occurs in the host galaxy,since G alactic extinction is very weak [E (B V )G al = 0.02 m ag;ref. 8]. W e com puted the bolom etric light curve using the data in Table 1 together w ith the m easurem ents reported in refs. 9 and 10.

15

4. X -ray data and m odels T he X RT light curve (Figure S2,top panel) was tted w ith a function C ountR ate = N (t t0) e (t t0 ),w here N = 14:5 2:2 and = 1:09 0:13,yielding 2=d:o:f:= 13:1=16. From this t, the onset of the X R F can be estim ated to have occurred on JD = 2454475:06413. T he X -ray ux reaches m axim um in 92s,w ith an uncertainty of 30s. T he decaying part of the X -ray curve can also be tted w ith a power-law w ith index = 1:52 0:16. T he num ber of photons is too sm all to perform tim eresolved spectroscopy. Instead,we show the tim e evolution ofthe hardness ratio H R = C R (1:5 10keV )=C R (0:3 1:5keV (Fig. S2, bottom panel). T he spectrum hardens during the rise to m axim um and softensafterwards,a typicalbehaviourofX -ray aresin G R Bs and in FR ED pro les. T he source is a ected by pile-up only w ithin a 3-pixelradius. T herefore we extracted thephotonsfrom an annularregion w ith a 3-pixelinnerradiusand a 30-pixelouterradius, yielding a totalof539 photons. T he spectrum can be tted w ith a sim ple power-law typicalofG R Bs,w ith photon in0:26 + 1:42 21 2 dex = 2:29+ 0: 24 and a hostgalaxy hydrogen colum n density ofN H = 6:42 1:18 10 cm ( 2=d:o:f: = 21:1=23). A lternatively, a m odel w ith a black body w ith tem perature 0:30 48 kT = 0:20+ 0: = 2:21+ 0: 20 keV superim posed on a power-law w ith photon index 0:93,and a 7:03 21 2 hostgalaxy hydrogen colum n density N H = 6:96+ 2: yields 2=d:o:f:= 20:6=21. 71 10 cm T he totalintensity is1:85 10 10 erg cm 2 s 1,and the black body intensity isno m ore than 14% ofit,thus the presence ofa black body rem ains uncertain (11). T he X -ray lum inosity can be com puted using the conversion from count rate to ux 1counts 1 = 1:17 1010 erg cm 2 s 1,derived from the spectralpower-law t corrected for N H absorption. O ver the 605 sec ofthe X -ray event,after applying a PSF correction factorof1.88,this yields a uence of1:1 107 ergcm 2,corresponding to a totalenergy 46 43 1:3 10 erg. T he peak lum inosity is 8 10 erg s 1. A ssum ing that the black-body com ponentisa constantfraction (14% )ofthe totallum inosity atany tim e,itslum inosity, 6 com bined w ith the high black-body tem perature (3:8 10 K ),im plies an em itting radius 10 R ph 10 cm ,an order ofm agnitude sm aller than the size ofW olf-R ayet stars.

16

8

Count Rate

6

4

2

0 1.0

1.5

2.0

2.5 3.0 Time Hs  100L

3.5

4.0

4.5

H 1.5 - 10. keVL  H0.3 - 1.5 keVL

8

6

4

2

0 0

1

2

3 Time Hs  100L

4

5

6

Figure 5: Top: the decaying part of the X -ray curve and a t w ith a power-law w ith index = 1:52 0:16;Bottom : tim e evolution ofthe hardness ratio H R = C R (1:5 10keV )=C R (0:3 1:5keV ).

17

5. Spectralm odelling. To estim ate the physical param eters that describe the supernova explosion we construct a series ofradiative transfer m odels to derive synthetic spectra. A s tim e elapses after the explosion ofthe supernova the ejecta expand and becom e diluted,progressively exposing deeperlayers. M odelling the spectra asthey evolve enablesusto inferthe structure ofthe ejected m aterial. A m ong the param eters that can be determ ined in this way are the lum inosity,the postion ofthe m om entary photosphere,and the com position and velocity ofthe line-form ing layers ofthe ejecta at each epoch. Forthe spectralm odelswe use a M onte C arlo radiative transfercode (S12,S13,S14). T he code em ploys an approxim ate description of non-LT E suitable for the analysis of supernova spectra during the photospheric epochs w hile retaining physically m eaningful relationships between m odel param eters. H ere we use a version of the code that uses a depth-dependent com position structure (S15). For allm odels in the series a density structure (v) is adopted (for SN e a H ubble-like expansion law r = v t allow s us to use the ejecta velocity as a tim e-independent radialvariable). Input param eters for each individualm odelarethelum inosity atthegiven epoch,a lowerboundary velocity,and the com position above this velocity. T his lower boundary represents the pseudo-photosphere from w hich allradiation is assum ed to em erge,w hich is a good approxim ation at early tim es. Lacking self-consistent explosion m odels we use a param eterized density structure w hich we constrain iteratively w ith the help of light curve and spectral m odels. O ur m odeldoes not take into account non-therm alexcitation and ionization by fast particles from the radioactive decay of56N i. T his is a fair approxim ation for m ost elem ents but it fails for helium . Because ofthe high ionization potentialofH ei,non-therm alprocesses are the m ain contribution to the excitation ofthis elem ents. T herefore,we cannot selfconsistently derive the opacity ofH eilines,w hich would anyway require a detailed m odel ofthe distribution of56N iand ofthe geom etry ofthe ejecta. T he very broad line featuresatearly epochs,up to day 5 afterexplosion,indicate the presence of appreciable am ounts of m atter at high velocities above v 30;000km s 1. Starting approxim ately 2008 January 15,(t= 6d)thelinesbecom enarrowerand theblueshifted high-velocity com ponentsdisappears. T hissuggestsa signi cantsteepening ofthe density gradient at lower velocities. T he last spectrum considered here was obtained on 2008 February 11 (t = 33d). For this m odel we use a lower boundary velocity of 7500km s 1. T he line w idths of this last spectrum suggest that the density gradient attens below a velocity of about 9000km s 1. D eeper layers are not yet accessible to observations. Fig. S3 show s the tim e evolution ofthe photospheric velocity. A rough estim ate ofthe param eters is obtained from the sim ilarity ofSN 2008D to SN e2002ap and 2006ajatthe earliesttim es and to SN 1994Iatlaterphases. T he m odels ofthese objects (S16,S17,and S18) are used as a guideline. T he evolution ofsynthetic spectra is show n in Fig. S4. Forthe m odels show n here we use a broken power-law w ith a power-law index ofn = 2:0 in theregion insideof9000km s 1,n = 7:5 between 9000km s 1 and 17,000km s 1, and n = 5:5 above 17,000km s 1. W e set the absolute density to m atch the observed spectral features, and obtain a total ejected m ass of 7M w ith a kinetic energy of 6 1051 erg, of w hich a m ass 0:03M w ith a E 5 1050 erg is located at v > 18

30;000km s 1. T he rst spectrum we consider here was taken on 12 January,3 days after the explosion. A n interesting feature ofthe spectra is the absence ofa strong O iabsorption at 7300 A .T he O ilines at 7774 A that are norm ally responsible for this feature are very strong. T herefore we can only accom m odate sm allam ounts ofoxygen in the outerlayers ofthe ejecta. A t early epochs oxygen could be so highly ionized that O ilines disappear, but the presence ofC aiiand Feiiin the spectra m akes this hypothesis unlikely because those species should not be present in an environm ent w here oxygen is m ostly ionized. O verall,we can m odelthe spectra w ith only a m ild variation ofthe com position w ith depth. Spectra before m axim um exhibit broad but shallow features of C a,Si,and Fegroup elem ents suggesting the presence ofonly sm allam ounts ofabsorbing m aterial. In the outerpartofthe ejecta above v 14;000km s 1 we assum e a com position dom inated by H e. T he attenuation ofthe ux in the blue requiresthe presence ofsom e T iand other Fe-group elem entsthatdo notproduce strong individualline featuresbutblock radiation in the blue and U V through a large num ber ofoverlapping weaker lines. A t epochs around and after m axim um the spectra show distinctly narrower line features that can be attributed the H ei,O i,C aiiand Feii. T here is also som e indication for C iin the red and infrared region (S19). In the inner part we assum e a com position w here C and O dom inate w ith a slightly enhanced contribution from heavy elem ents, likely including som e decay products ofradioactive 56N i.

19

Figure 6: T he tem poralevolution ofthe lower boundary (photosphere) velocity used in the m odelsforSN 2008D com pared to the casesofotherSN eIc. SN 2008D startsoutw ith a very high velocity, like SN 2002ap, then transitions to lower velocities like SN 1994I. T his isthe phase w hen broad lines disappear. A tlatertim es the evolution is slower than thatofSN 1994I,indicating a large m assw ith a sm alldensity gradientin the innerlayers, sim ilar to SN 1997ef. O nly SN e w ith velocities larger than SN e2008D or 2002ap were accom panied by a G R B or an X R F.

20

Figure 7:Seriesofm odelspectra forthe sequence ofearly tim e spectra ofSN 2008D .T he opticaldepth in H eilines in the m odelof24 January has been enhanced to m im ic the e ect ofnon-therm alexcitations by fast electrons generated in the decay of 56N i. W hile we cannot constrain the abundance of H e this way this m ethod allow s us to identify H e lines in the spectrum . T he rst two spectra are contam inated by the em ission ofthe afterglow w hich isnotdescribed by them odel.T herefore,thelum inosity needed to m atch those spectra is too high to give a consistent description leading to an over-ionization of m ost species. T his a ects in particularly the C aii IR triplet near 8000 A ,w hich is not reproduced by the m odels in the rst two epochs.

21

R eferences. 1. van den Bergh,S.,Li,W .,& Filippenko,A .V . PA SP,115,1280 (2003). 2. Fruchter,A .,et al.,N ature441,463-468 (2006). 3. Veilleux,S.,& O sterbrock,D .E.,A pJS,63,295 (1987). 4. M annucci,F.,et al.,A & A ,401,519 (2003). 5. U lvestad,J.S.,& H o,L.C .A pJ,581,925 (2002). 6. C locchiatti,A .,& W heeler,J.C .A pJ,491,375 (1997). 7. Page,K .L.,etal. G C N 7164 (2008). 8. Schlegel,D .J.,Finkbeiner,D .P.,& D avis,M .,A pJ,500,525 (1998). 9. Im m ler,S.,et al.G R B C oordinates N etwork,7168,1 (2008). 10. Li,W .,C hornock,R .,Foley,R .J.,Filippenko,A .V .,M odjaz,M .,Poznanski,D ., & Bloom ,J.S. G R B C oordinates N etwork,7176,1 (2008). 11. Li,L.-X .,astro-ph 0803.0079 (2008). 12. M azzali,P.A .& Lucy,L.B.,A & A ,279,447 (1993). 13. Lucy,L.B.,A & A ,345,211 (1999). 14. M azzali,P.A .,A & A ,363,705 (2000). 15. Stehle, M ., M azzali, P. A ., Benetti, S., & H illebrandt, W ., M N R A S, 360, 1231 (2005). 16. M azzali,P.A .,et al.,A pJ,572,L61 (2002). 17. M azzali,P.A .,et al.,N ature,442,1018 (2006). 18. Sauer,D .N .,et al.,M N R A S,369,1939 (2006). 19. Valenti,S.,et al.,A pJ,673,L155 (2008).

22