Table Of Content

Accepted forpublicationin ApJ Letters PreprinttypesetusingLATEXstyleemulateapjv.6/22/04 RELICS OF METAL-FREE LOW MASS STARS EXPLODING AS THERMONUCLEAR SUPERNOVAE Takuji Tsujimoto1 and Toshikazu Shigeyama2 Accepted for publication in ApJ Letters ABSTRACT Renewed interest in the first stars that were formed in the universe has led to the discovery of extremely iron-poor stars. Since several competing scenarios exist, our understanding of the mass rangethat determinesthe observedelementalabundancesremainsunclear. Inthis study,we consider three well-studied metal-poor stars in terms of the theoretical supernovae (SNe) model. Our results suggest that the observed abundance patterns in the metal-poor star BD +80◦245 and the pair of 6 stars HD 134439/40 agree strongly with the theoretical possibility that these stars inherited their 0 0 heavy element abundance patterns from SNe initiated by thermonuclear runaways in the degenerate 2 carbon-oxygencoresofprimordialasymptoticgiantbranchstarswithmassesof∼3.5−5M⊙. Recent theoreticalcalculations havepredicted that such SNe could be originatedfrommetal-free starsin the n intermediate mass range. On the other hand, intermediate mass stars containing some metals would a end their lives as white dwarfs after expelling their envelopes in the wind due to intense momentum J transport from outgoing photons to heavy elements. This new pathway for the formation of SNe 6 requires that stars are formed from the primordial gas. Thus, we suggest that stars of a few solar 1 masses were formed from the primordial gas and that some of them caused thermonuclear explosions whenthe mass oftheir degeneratecarbon-oxygencoresincreasedto the Chandrasekharlimitwithout 1 v experiencing efficient mass loss. 4 Subject headings: Galaxy: abundances — stars: abundances — stars: AGB and post-AGB — stars: 2 Population II — supernovae 3 1 0 1. INTRODUCTION sult, these stars are less opaque and more compact than 6 stars that have experienced the third dredge-up. These Elementalabundance ofrecently discoveredextremely 0 features also result in stars with lower radiation pres- iron-deficient stars has raised questions about the mass / h range of the first stars that were formed in the uni- sures and stronger surface gravities. Therefore, it is ex- p verse (Christlieb et al. 2002; Umeda & Nomoto 2003; pected that the mass loss from an extremely metal-poor - star will be substantially suppressed (Fujimoto et al. Schneider et al. 2003; Shigeyama et al. 2003; Suda et al. o 1984). Thus, a degeneratecarbon-oxygencorecanreach r 2004; Frebel et al. 2005; Iwamoto et al. 2005). Gen- t eral theoretical considerations seem to suggest that, the Chandrasekhar limit (Mch) while the star is in the s TPAGB phase. Eventually the star undergoes a ther- a due to inefficeint cooling of the primordial gas, the monuclear explosion to become an SN. This type of SN v: first stars were at least as massive as 100 M⊙ (e.g., can be classified as a type Ia in terms of nucleosynthesis Abel et al. 2002; Bromm et al. 2002). However, other i but as a type II in terms of spectral type. If the mass X studies claim that molecular or atomic hydrogen cool- r ing is sufficient to form low-mass stars from the primor- ofa proto-galacticcloudis 106 M⊙ correspondingto the a dialgas(Nakamura & Umemura2001;Omukai & Yoshii Jeans mass soon after recombination, a typical Type II SN(SN II) enrichesthe cloudup to [Fe/H]∼−4. There- 2003). Therefore, the detection of signatures constrain- fore only the primordial (population III) stars are likely ing the mass range of the first stars limits the possible to end up as such SNe. Hence, we define them as SNe formation mechanisms of these objects. IIIa. The key parameters controllingthe structure andevo- The lifetime of an SN IIIa progenitor is much shorter lution of stars are mass and metallicity. A given com- than that of an SN Ia progenitor, though these two bination of these two parameters establishes new path- types of SNe supply similar amounts of nuclear prod- ways for supernova (SN) formation from “primordial in- ucts. In a white dwarf, an SN Ia event needs sub- termediate mass” stars. According to recent calcula- stantial replenishment of the gas by accretion from a tions (Fujimoto et al. 2000;Suda et al.2004), stars with [Fe/H] ∼< −5 and masses in the range of 3.5M⊙ ∼< M lhoawv-emmausschcolomnpgaernieovno.lutAiosnaaryretsiumlte,-scSaNlesIathparnogthenositeoorsf ∼< 7M⊙ donotexperiencemixingepisodesincludingthe SNe II −, typically longer than 109 yrs −. This pre- thirddredge-up,intheirconvectiveenvelopeinthe ther- ventedSNe Ia fromcontributing to chemicalenrichment mally pulsing asymptotic giant branch (TPAGB) phase. in the early epoch of galaxy evolution. On the other These stars will end up with a small amount of carbon hand, an SN IIIa event occurs when an intermediate as wellas s-process elements in their envelopes. As a re- mass star evolves into the AGB phase in a period rang- ing from several 107 yrs to a few 108 yrs, depending 1 National Astronomical Observatory, Mitaka-shi, Tokyo 181- on the progenitor mass. Such short time scales could 8588,Japan;[email protected] 2 Research Center for the Early Universe, Graduate School of enable us to detect evidence of the above in the ele- Science,UniversityofTokyo,7-3-1Hongo,Bunkyo-ku,Tokyo113- mental abundances of old stars. Recent studies of the 0033,Japan;[email protected]. chemical composition of metal-poor stars have revealed 2 &$% &$% & & .!&$% .!&$% - - , , + + * * ) 0 ( !# ( !# !#$% !#$% !" !" !' !"$% !" !#$% !# !&$% & &$% !' !"$% !" !#$% !# !&$% & &$% (,-+/. (,-+/. Fig. 1.—Left panel: Correlationof[Na/Mg]with[Mg/H]forBD+80◦245andHD134439/40 (filledcircles;Ivans etal.2003,Fulbright 2000) together with other Galactic stars (open triangles; Cayreletal. 2004, crosses; Fulbright 2000). No abundance anomaly is seen for thethreestars. Rightpanel: Thesameasleftpanelbutfor[Ba/Mg]. ThedatashownbyopentrianglesaretakenfromMcWilliam(1998). that they have inherited the abundance pattern of the by the combination of heavy elements ejected from an ejecta of the preceding SNe (e.g., Audouze & Silk 1995; SN itself and those elements that are already present Shigeyama & Tsujimoto 1998; Umeda & Nomoto 2003). in the interstellar gas due to previous SN II explo- If this is the case, there might be some later stars that sions (Tsujimoto et al. 1999). Therefore, the deficiency inherited the nuclear products of SNe IIIa, though their not only of α-elements but also of n-capture and odd- abundancepatternscouldbemodifiedtosomedegreeby numberedelementsshouldbeattributedtoanadditional products from SNe II. Fe supply from an SN IIIa. Since an SN IIIa produces The most characteristic abundance feature of stars in- few α-elements, if we take the abundance ratios of these heriting the nuclear products of an SNe IIIa event is elements to Mg instead of Fe, both BD+80◦245 and expected to be the deficiency of α-elements, compared HD134439/40becomeindistinguishable fromotherhalo with the enhanced α/Fe ratios commonly seen in halo stars (see Fig. 1). In other words, the abundance of the stars. In fact, there exist so-called low-α stars in the interstellar medium (ISM) [Mg/H] at the birth epoch gas Galactic halo with low α/Fe ratios. It has been pro- of this star must be equal to the observed abundance posed (Nissen & Schuster 1997; Gilmore & Wyse 1998) of BD+80◦245, [Mg/H]= −2.29 (Ivans et al. 2003). As- that these low-α stars were born from accreted dwarf suming that the abundance ratio is equal to the average galaxies or protogalactic fragments, where a prolonged ratio for halo stars [Mg/Fe] =0.4, the Fe abundance gas lowstarformationrateenablesSNeIaeventstooccurin in the ISM is estimated to be [Fe/H] =−2.69. Adding gas a low-metallicity regime. The specific mechanism which another 0.63M⊙ of the Fe supplied by the preceding SN explainsthelowα/Feratiosinanindividualstariseasily IIIa to the 6.5×104 M⊙ of the gas swept up by the SN determinedbytheabundanceofs-processelements,since remnant(Shigeyama & Tsujimoto 1998), the metallicity thetimescalesassociatedwiththesetwotypesofSNeare of a star formed from this gas becomes [Fe/H]=−1.97. substantiallydifferent. ThelonglifetimeofSNeIaallows This is in good agreement with the observed value of AGBstarstosubstantiallypollutethegaswiths-process [Fe/H]=−2.07 (Ivans et al. 2003). elementsbeforelow-αstarsareformed,whereasSNeIIIa Another reason for connecting stellar abundance fea- preceded most AGB stars with masses ∼< 3.5M⊙. tures with a Pop III SN IIIa is the deficiency of Intheliterature,wefoundthreestars(BD+80◦245and odd-numbered elements (Mn, Co) compared to even- HD 134439/40)that exhibit the abundance patterns ex- numbered ones (Cr, Ni) among the Fe-group. This fea- pected from SNe IIIa. Elemental features of these three ture may favor nucleosynthesis in primordial SNe, since starsarediscussedinthesubsequenttwosections. Their odd nuclear charge numbered elements require the pre- origin is investigated in the context of a chemical evolu- existence of heavy element seeds in the SN progenitor. tion model developed by Tsujimoto et al. (1999) in §4. AmongtheseFe-groupelements,Mnabundancebestdis- tinguishes the products of SNe IIIa from those of SNe 2. ELEMENTALFEATURESOFBD+80◦245 Ia. If an SN Ia, instead of an SN IIIa, supplied the Mn and the Fe to the gas from which this star formed, then The conspicuous abundance feature of the star BD +80◦245 is the deficiency of α-elements and n-capture a theoretical SN Ia model (Iwamoto et al. 1999) would predict that the [Mn/Fe] ratio in the gas,which was ini- elements,suchasBa orEu,andlowratiosofNa,Al, Sc, tially ∼ −0.4, would become ∼ −0.07 after being swept ZnrelativetoFe(Fulbright2000;Stephens & Boesgaard up by the SN. In fact, the observed [Mn/Fe] =−0.26 2002; Ivans et al. 2003). In the supernova-induced star (Ivans et al.2003)implies thatthe MnyieldfromanSN formationscenarioproposedbyShigeyama & Tsujimoto IIIa is about one-third that of an SN Ia. Likewise, the (1998), the stellar abundance pattern is determined 3 !& "0*sun #0*<=> $("0*<=> first-generationstarsisintroducedasafreeparameterin this model, and this value is assumed to be 2.5×10−4 (Tsujimoto et al.2000). Detailsofthis processaregiven in Tsujimoto et al. (1999). -60&$##$7 First, we show the evolution of [Mg/H] in the gas !% (Fig. 2). Since an SN IIIa supplies only a small amount ofMg,the Mgabundancesofthe starsintroducedin the 8609:';%#" preceding sections coincide with the abundance of the . - interstellar gas at the birth epochs of the stars. Thus, +, !$ this age-[Mg/H]relationcanbe usedto date these stars. * Thatis,foragiven[Mg/H]value,theprogenitormassof ) the SN IIIa for each star can be deduced. The resultant masses are 5 M⊙ for BD +80◦245 and 3.5 M⊙ for HD !# 134439. These inferred masses are within the allowed mass range (∼3.5−7M⊙) for SN IIIa progenitors. The lower bound mass is required for the inhibition of mixing processes in the convective envelope (Fujimoto et al. 2000), while the upper bound mass is the lower limit at !" ' '(& '(% '($ '(# whichhydrostaticcarbonburningwilloccurinthestellar core (Siess et al. 2002). /012345 Incontrastto Mgabundance,stellarFe abundancein- herited from SN IIIa remnants differs from that in the Fig. 2.— The predicted age-[Mg/H] relationship for the gas in interstellar gas,which leads to unusual elemental [X/Fe] the protogalactic cloud. SN IIIa progenitor masses corresponding ratiosforthesestars. Sincetheamountofheavyelements to different stellar lifetimes are also indicated. The shaded areas denote [Mg/H] abundances with errors for BD+80◦245 and HD containedin the gas increases with time, the major con- 134439. The [Mg/H]abundance of HD134440 is not shownsince tributor to stellar metallicity switches from SN ejecta itonlydiffersby0.05dexfromthatofHD134439. to the metallicity in the interstellar gas. Therefore, the delay in SN IIIa explosions reduces the anomalies of the [X/Fe]ratios. Asaresult,the[X/Fe]ratiosofstarstrace ejecta of an SN Ia would result in [Co/Fe]∼ 0, whereas thecontinuousevolutionarytrackinFigure3,whichcon- observations show [Co/Fe]=−0.18 (Ivans et al. 2003). siderstheassemblageofelementsintheejectaofindivid- ual SNe IIIa, with different progenitor masses, in addi- 3. ELEMENTALFEATURESOFHD134439/40 tiontotheelementsintheISM.Aswell,thisfigureshows The pair of stars HD 134439/40 (King 1997) also ap- the modelcurves for the evolutionarychange in[Mg/Fe] peartoinheritthe abundancepatternofthe ejectaofan (leftpanel)and[Si/Fe](rightpanel)ratiostogetherwith SN IIIa. They both exhibit the deficiency of α- and n- the observed data for the three stars (Ivans et al. 2003; capture elements similar to BD +80◦245,but to a lesser Fulbright 2000) and numerous other Galactic stars with extent (Ryan et al. 1991; Fulbright 2000; Gratton et al. Hipparcos parallaxes (Fulbright 2000). For the nucle- 2003). Thismilddeficiencyisnaturallyinterpretedinthe osynthetic yields of Mg, Si, and Fe in SNe IIIa, those framework of the evolution of stellar abundances in the predicted by the W7 model of SNe Ia (Iwamoto et al. Galactic halo. The relatively small contribution of SNe 1999) are adopted. A small Mg yield in SN IIIa leads to IIIa to stellar abundance is attributed to a less massive theformationofstarswithanabundanceratio,[Mg/Fe], progenitor star. Its longer lifetime allows more numer- as low as ∼−1.6. In contrast, a relatively large Si yield ousSNe IItoenrichtheinterstellarmatteruntilthestar in SN IIIa results in stars with an [Si/Fe] ratio > −0.4. is formed. In §4, it will be shown that the unusual ele- Theoretically, the mass swept up by an SNR is depen- mental abundances of these three stars are linked to the dent on the local gas density (Shigeyama & Tsujimoto unique evolutionary track of stellar abundances in the 1998) and is found to lie within a factor of ∼3 about ejecta of SNe IIIa with different progenitor masses. themeanof6.5×104M⊙ (Nakasato & Shigeyama2000). In this study, we consider a factor of 1.25 larger than 4. THEEVOLUTIONOFELEMENTALABUNDANCESOF this(dashedcurves)togetherwiththetypicalcase(solid STARSTRIGGERED BYSNIIIA EXPLOSIONS curves) (see Figure 3). Taking into account the incom- Inthissection,wewilldiscusshowtheunusualelemen- plete mixing of elements from different SNe inside the tal abundances seen in the three stars under discussion inhomogeneous ISM, we calculated the abundance ratio can be modeled theoretically. The chemical evolution ofstarsformedfromSNeIIIawithtwodifferentvaluesof considered here is based on a scenario (Tsujimoto et al. thegaseous[Mg(Si)/Fe],i.e.,0.4(uppertwocurves)and 1999) in which the stellar abundances change through 0.25(lowertwocurves). Theseresultsrevealthattheob- successive sequences of the SN explosion, shell forma- servedlower[Mg/Fe]and[Si/Fe]ratiosforthethreestars tion, and star formation from the matter swept up by are well described by the unified scheme. This scheme the individual SN remnants (SNRs). Heavy elements proposesthatthesestarswereformedfromthegasswept ejected from an SN are assumed to be trapped and well up by SNe IIIa with different progenitor masses during mixed within the SNR shell. A fraction of this heavy different epochs. element materialis then incorporatedinto the next gen- Moreover,the predictedrelationshipsbetween[Ba/Fe] erationstars,whilethematerialremaininginthegaswill and [Fe/H], as well as [Eu/Fe] and [Fe/H], for the three mix with the ambientmedium. The mass fractionofthe starsalsoagreewiththeobservedratios,asshowninthe 4 # # &$% &$% & & )*+,-.!&$% 01+,-.!&$% ( ( !# !# !#$% !#$% !" !" !' !"$% !" !#$% !# !' !"$% !" !#$% !# (,-+/. (,-+/. Fig. 3.—Left panel: Thepredictedcorrelationof[Mg/Fe]with[Fe/H]insidetheshellsweptupbytheejectaofanSNIIIafordifferent progenitormasses,togetherwiththeobservationaldataforBD+80◦245andHD134439/40(filledcircles;Ivans etal.2003,Fulbright2000) aswellasotherGalacticstars(crosses;Fulbright2000). Individualmodelcurvesshowfourcaseswithcombinations oftheswept-upmass andthe[Mg/Fe]ofthegas,thatis(8×104M⊙,0.4)(upperdashedcurve),(6.5×104 M⊙,0.4)(uppersolidcurve),(8×104M⊙,0.25)(lower dashedcurve), and(6.5×104 M⊙,0.25)(lower solidcurve), respectively. Seetext fordetails. Right panel: Thesameasleftpanel butfor [Si/Fe]. Noobservational errorfor[Si/Fe]ofBD+80◦245isprovided. two panels of Figure 4. The evolution of [Ba/Fe] in the theTPAGBphase,theprogenitorsneedtobe metal-free gas is calculated assuming that the production site for stars. Since the three stars exhibit these two character- early-epoch s-process elements is 3−5 M⊙ AGB stars, istics, “a sufficient Fe supply on a short time-scale” and whichisshownbythedash-dottedcurve. Itwasassumed “primordiality” in their elemental abundance patterns, that the [Eu/Fe] ratio in the gas is 0.5. The predicted thisstronglysuggeststhattheremustbemetal-freestars [Ba/Fe]and[Eu/Fe]sequences ofstarsformedfromSNe of a few solar masses. IIIa with different progenitor masses are completely dif- We also predict that stars inheriting heavy ele- ferent from the evolution of gaseous abundances, as is ments from SNe IIIa populate distinct branches in the shown by the solid and dashed curves. Thus, the re- [α/Fe]−[Fe/H] and [n-capture/Fe]−[Fe/H] planes. The markablylow[Ba/Fe]and[Eu/Fe]ratiosforBD+80◦245 mannerinwhichthese starspopulatethis branchshould are attributable to an early SN IIIa explosion in very give valuable information on the initial mass func- low-metallicity gas due to a relatively massive (∼ 5M⊙) tion of metal-free stars. Stars descended from SNe progenitor. On the other hand, the SN IIIa progenitor, IIIa are predicted to have metallicities in the range of havingimprintedtherelativelyhigh[Ba/Fe]and[Eu/Fe] −2.3∼< [Fe/H]∼< −1.3. Thus, the eight low-α halo stars ratios in HD 134439/40, was less massive (∼ 3.5M⊙) with −1.3 ≤ [Fe/H] ≤ −0.8 (Nissen & Schuster 1997), than the SN IIIa progenitor forming BD+80◦245. At and D119 with [Fe/H] =−2.95 exhibiting extreme defi- the same time, the model curves in Figures 3 and 4 pre- ciencies of Ca, Sr, and Ba in the Draco dwarfspheroidal dict the location of stars formed from SNe IIIa in the galaxy (Fulbright et al. 2004), are unlikely to be associ- [α/Fe]-[Fe/H] and [n-capture/Fe]-[Fe/H] planes. Finally ated with the proposed SNe IIIa. itshouldbe notedthatheavyelementsejectedfromSNe Theoretically during the formation of galaxies, stars IIIa are expected to have little influence on the overall formin the protogalacticclouds assembling to construct enrichmentofinterstellargas,sinceSNeIIIa,whosepro- the Galactic halo. A spread of several billion years in genitors are first-generation stars, are underpopulated, the stellar age among the halo stars (Schuster & Nissen andSNeIIhadalreadyenrichedtheinterstellargaswith 1989) suggests that heavy elements dispersed into the heavy elements before the emergence of SN IIIa. halo from the protoclouds should have polluted other cloudsbeforestarformationhadstartedthere. Thus,the 5. CONCLUSIONSANDDISCUSSION formationofmetal-freestarsmusthavestartedatavery The unusually low abundances of α- and neutron- early epoch and, accordingly, very far from the Galactic capture elements relative to Fe seen in the metal-poor center. Thus, stars inheriting heavy elements from Pop star BD +80◦245 and the pair of stars HD 134439/40 III SNe are expected to share kinematic features of the are shown to be fossil records of SN explosions triggered outermosthalopopulation,suchaslargeapo-galacticor- byathermonuclearrunawayinametalfreeintermediate bital radii with high eccentricities. Observations reveal mass star. These SNe have two distinct characteristics. thatthe threestarsexaminedinthe currentletter doin- First, they can be classifiedas type Ia in terms of nucle- deedpossesssuchkinematicfeatures(Carney et al.1997; osynthesis, that is, they synthesize a large amount of Fe King 1997). as a result of the thermonuclear explosions. Second, for their cores to grow to the Chandrasekhar limit during 5 # # &$% &$% & & . . - - +,!&$% +,!&$% * 1 ) 0 ( ( !# !# !#$% !#$% !" !" !' !"$% !" !#$% !# !' !"$% !" !#$% !# (,-+/. (,-+/. Fig. 4.— Left panel: The sameas Figure3 but for[Ba/Fe]; solidand dashed curves correspond to the swept-up mass of 6.5×104M⊙ and 8×104M⊙, respectively. Dashed-dotted line shows the evolution of the abundance ratios in the gas. Right panel: The same as left panelbutfor[Eu/Fe]. TheobservationaldataforBD+80◦245aretakenfromFulbright(2000)insteadofIvans etal.(2003)sinceFulbright (2000)givesaconsistentr-processBa/Euratio. We are grateful to the anonymous referee for the use- and suggestions for improvement by J. B. Carr. This fulcommentsthathelpedimprovethispaper. TSthanks work has been partially supported by a grant-in-aid Masayuki Fujimoto for the valuable conversation on the (16540213) of the Japanese Ministry of Education, Sci- evolutionofmetal-freestars,whichstimulatedthiswork. ence, Culture, and Sports. We alsoappreciatethe carefulreadingofthe manuscript REFERENCES Abel,T.,Bryan,G.L.,&Norman,M.L.2002,Science, 295,93 King,J.R.1997,AJ,113,2302 Audouze,J.,&Silk,J.1995,ApJ,451,L49 McWilliam,A.1998,AJ,115,1640 Bromm,V.,Coppi,P.S.,&Larson,R.B.2002, ApJ,564,23 Nakamura,F.,&Umemura,M.2001,ApJ,548,19 Carney,B.W.,Wright,J.S.,Sneden,C.,Laird,J.B.,Aguilar,L. Nakasato,N.,&Shigeyama, T.2000,ApJ,541,L59 A.,&Latham, D.W.1997,AJ,114,363 Nissen,P.E.,&Schuster,W.J.1997,A&A,326,751 Cayrel,R.etal.2004, A&A,416,1117 Omukai,K.,&Yoshii,Y.2003, ApJ.599,746 Christlieb,N.etal.2002,Nature,419,904 Ryan,S.G.,Norris,J.E.,&Bessell,M.S.1991,AJ,102,303 Frebel,N.etal.2005, Nature,434,871 Schneider,R.,Ferrara,A.,Salvaterra,R.,Omukai,K.,&Bromm, Fujimoto, M. Y., Iben, I. Jr., Chieffi, A., & Tornambé, A. 1984, V.2003,Nature,422,869 ApJ,287,749 Schuster,W.J.,&Nissen,P.E.1989,A&A,222,69 Fujimoto,M.Y.,Ikeda,Y.,&Iben, I.Jr.2000,ApJ,529,L25 Shigeyama,T.,&Tsujimoto,T.1998, ApJ,507,L135 Fulbright,J.P.2000,AJ,120,1841 Shigeyama,T.,Tsujimoto,T.,&Yoshii,Y.2003,ApJ.586,57 Fulbright,J.P.,Rich,R.M.,&Castro,S.2004,ApJ,612,447 Siess,L.,Livio,M.,&Lattanzio, J.2002,ApJ,570,329 Gilmore,G.,&Wyse,R.F.G.1998,AJ,116,748 Stephens, A.,&Boesgaard,A.M.2002, AJ,123,1647 Gratton,R.G.,Carretta,E.,Claudi,R.,Lucatello,S.,&Barbieri, Suda, T., Aikawa, M.,Machida, M. N., Fujimoto, M. Y.,& Iben, M.2003,A&A,404,187 I.Jr.2004,ApJ,611,476 Iben,I.Jr.,&Renzini,A.1983,ARA&A,21,271 Tsujimoto,T.,Shigeyama, T.,&Yoshii,Y.1999,ApJ,519,L63 Ivans,I.I.etal.2003, ApJ,592,906 Tsujimoto,T.,Shigeyama, T.,&Yoshii,Y.2000,ApJ,531,L33 Iwamoto, K., Brachwitz, F., Nomoto, K., Kishimoto, N., Umeda, Umeda,H.,&Nomoto,K.2003, Nature,422,871 H.,Hix,W.R.,&Thielemann,F.-K.1999, ApJS,125,439 Iwamoto,N.,Umeda,H.,Tominaga,N.,Nomoto,K.,&Maeda,K. 2005, Science,309,451